From a9f4c6c999008c92e2aba6d4f50f2b2d10ed7fd0 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 31 Jan 2024 17:57:54 -0300 Subject: perf trace: Collect sys_nanosleep first argument That is a 'struct timespec' passed from userspace to the kernel as we can see with a system wide syscall tracing: root@number:~# perf trace -e nanosleep 0.000 (10.102 ms): podman/9150 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 38.924 (10.077 ms): podman/2195174 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 100.177 (10.107 ms): podman/9150 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 139.171 (10.063 ms): podman/2195174 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 200.603 (10.105 ms): podman/9150 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 239.399 (10.064 ms): podman/2195174 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 300.994 (10.096 ms): podman/9150 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 339.584 (10.067 ms): podman/2195174 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 401.335 (10.057 ms): podman/9150 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 439.758 (10.166 ms): podman/2195174 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 501.814 (10.110 ms): podman/9150 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 539.983 (10.227 ms): podman/2195174 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 602.284 (10.199 ms): podman/9150 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 640.208 (10.105 ms): podman/2195174 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 702.662 (10.163 ms): podman/9150 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 740.440 (10.107 ms): podman/2195174 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 802.993 (10.159 ms): podman/9150 nanosleep(rqtp: { .tv_sec: 0, .tv_nsec: 10000000 }) = 0 ^Croot@number:~# strace -p 9150 -e nanosleep If we then use the ptrace method to look at that podman process: root@number:~# strace -p 9150 -e nanosleep strace: Process 9150 attached nanosleep({tv_sec=0, tv_nsec=10000000}, NULL) = 0 nanosleep({tv_sec=0, tv_nsec=10000000}, NULL) = 0 nanosleep({tv_sec=0, tv_nsec=10000000}, NULL) = 0 nanosleep({tv_sec=0, tv_nsec=10000000}, NULL) = 0 nanosleep({tv_sec=0, tv_nsec=10000000}, NULL) = 0 nanosleep({tv_sec=0, tv_nsec=10000000}, NULL) = 0 nanosleep({tv_sec=0, tv_nsec=10000000}, NULL) = 0 ^Cstrace: Process 9150 detached root@number:~# With some changes we can get something closer to the strace output, still in system wide mode: root@number:~# perf config trace.show_arg_names=false root@number:~# perf config trace.show_duration=false root@number:~# perf config trace.show_timestamp=false root@number:~# perf config trace.show_zeros=true root@number:~# perf config trace.args_alignment=0 root@number:~# perf trace -e nanosleep --max-events=10 podman/2195174 nanosleep({ .tv_sec: 0, .tv_nsec: 10000000 }, NULL) = 0 podman/9150 nanosleep({ .tv_sec: 0, .tv_nsec: 10000000 }, NULL) = 0 podman/2195174 nanosleep({ .tv_sec: 0, .tv_nsec: 10000000 }, NULL) = 0 podman/9150 nanosleep({ .tv_sec: 0, .tv_nsec: 10000000 }, NULL) = 0 podman/2195174 nanosleep({ .tv_sec: 0, .tv_nsec: 10000000 }, NULL) = 0 podman/9150 nanosleep({ .tv_sec: 0, .tv_nsec: 10000000 }, NULL) = 0 podman/2195174 nanosleep({ .tv_sec: 0, .tv_nsec: 10000000 }, NULL) = 0 podman/9150 nanosleep({ .tv_sec: 0, .tv_nsec: 10000000 }, NULL) = 0 podman/2195174 nanosleep({ .tv_sec: 0, .tv_nsec: 10000000 }, NULL) = 0 podman/9150 nanosleep({ .tv_sec: 0, .tv_nsec: 10000000 }, NULL) = 0 root@number:~# root@number:~# perf config trace.show_arg_names=false trace.show_duration=false trace.show_timestamp=false trace.show_zeros=true trace.args_alignment=0 root@number:~# cat ~/.perfconfig # this file is auto-generated. [trace] show_arg_names = false show_duration = false show_timestamp = false show_zeros = true args_alignment = 0 root@number:~# This will not get reused by any other syscall as nanosleep is the only one to have as its first argument a 'struct timespec" pointer argument passed from userspace to the kernel: root@number:~# grep timespec /sys/kernel/tracing/events/syscalls/sys_enter_*/format | grep offset:16 /sys/kernel/tracing/events/syscalls/sys_enter_nanosleep/format: field:struct __kernel_timespec * rqtp; offset:16; size:8; signed:0; root@number:~# BTF based pretty printing will simplify all this, but then lets just get the low hanging fruits first. Reviewed-by: Ian Rogers Cc: Adrian Hunter Cc: Jiri Olsa Cc: Namhyung Kim Link: https://lore.kernel.org/lkml/Zbq72dJRpOlfRWnf@kernel.org/ Signed-off-by: Arnaldo Carvalho de Melo --- .../perf/util/bpf_skel/augmented_raw_syscalls.bpf.c | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c b/tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c index 2872f9bc0785..0acbd74e8c76 100644 --- a/tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +++ b/tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c @@ -341,6 +341,27 @@ failure: return 1; /* Failure: don't filter */ } +SEC("tp/syscalls/sys_enter_nanosleep") +int sys_enter_nanosleep(struct syscall_enter_args *args) +{ + struct augmented_args_payload *augmented_args = augmented_args_payload(); + const void *req_arg = (const void *)args->args[0]; + unsigned int len = sizeof(augmented_args->args); + __u32 size = sizeof(struct timespec64); + + if (augmented_args == NULL) + goto failure; + + if (size > sizeof(augmented_args->__data)) + goto failure; + + bpf_probe_read_user(&augmented_args->__data, size, req_arg); + + return augmented__output(args, augmented_args, len + size); +failure: + return 1; /* Failure: don't filter */ +} + static pid_t getpid(void) { return bpf_get_current_pid_tgid(); -- cgit v1.2.3 From bb69c912c4e8005cf1ee6c63782d2fc28838dee2 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Fri, 15 Mar 2024 09:13:34 +0200 Subject: perf auxtrace: Fix multiple use of --itrace option If the --itrace option is used more than once, the options are combined, but "i" and "y" (sub-)options can be corrupted because itrace_do_parse_synth_opts() incorrectly overwrites the period type and period with default values. For example, with: --itrace=i0ns --itrace=e The processing of "--itrace=e", resets the "i" period from 0 nanoseconds to the default 100 microseconds. Fix by performing the default setting of period type and period only if "i" or "y" are present in the currently processed --itrace value. Fixes: f6986c95af84ff2a ("perf session: Add instruction tracing options") Signed-off-by: Adrian Hunter Cc: Adrian Hunter Cc: Andi Kleen Cc: Ian Rogers Cc: Jiri Olsa Cc: Namhyung Kim Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20240315071334.3478-2-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/auxtrace.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/auxtrace.c b/tools/perf/util/auxtrace.c index 3684e6009b63..ef314a5797e3 100644 --- a/tools/perf/util/auxtrace.c +++ b/tools/perf/util/auxtrace.c @@ -1466,6 +1466,7 @@ int itrace_do_parse_synth_opts(struct itrace_synth_opts *synth_opts, char *endptr; bool period_type_set = false; bool period_set = false; + bool iy = false; synth_opts->set = true; @@ -1484,6 +1485,7 @@ int itrace_do_parse_synth_opts(struct itrace_synth_opts *synth_opts, switch (*p++) { case 'i': case 'y': + iy = true; if (p[-1] == 'y') synth_opts->cycles = true; else @@ -1649,7 +1651,7 @@ int itrace_do_parse_synth_opts(struct itrace_synth_opts *synth_opts, } } out: - if (synth_opts->instructions || synth_opts->cycles) { + if (iy) { if (!period_type_set) synth_opts->period_type = PERF_ITRACE_DEFAULT_PERIOD_TYPE; -- cgit v1.2.3 From 3e5deb708c8f3f7d645f567b2ba38d8045fa11ba Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Fri, 2 Feb 2024 15:40:54 -0800 Subject: perf cpumap: Clean up use of perf_cpu_map__has_any_cpu_or_is_empty MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Most uses of what was perf_cpu_map__empty but is now perf_cpu_map__has_any_cpu_or_is_empty want to do something with the CPU map if it contains CPUs. Replace uses of perf_cpu_map__has_any_cpu_or_is_empty with other helpers so that CPUs within the map can be handled. Reviewed-by: James Clark Signed-off-by: Ian Rogers Acked-by: Namhyung Kim Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Alexandre Ghiti Cc: Andrew Jones Cc: AndrĂ© Almeida Cc: Athira Rajeev Cc: Atish Patra Cc: Changbin Du Cc: Darren Hart Cc: Davidlohr Bueso Cc: Huacai Chen Cc: Ingo Molnar Cc: Jiri Olsa Cc: John Garry Cc: K Prateek Nayak Cc: Kajol Jain Cc: Kan Liang Cc: Leo Yan Cc: Mark Rutland Cc: Mike Leach Cc: Nick Desaulniers Cc: Paolo Bonzini Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Sandipan Das Cc: Sean Christopherson Cc: Steinar H. Gunderson Cc: Suzuki Poulouse Cc: Thomas Gleixner Cc: Will Deacon Cc: Yang Jihong Cc: Yang Li Cc: Yanteng Si Link: https://lore.kernel.org/r/20240202234057.2085863-6-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-c2c.c | 6 +----- tools/perf/builtin-stat.c | 9 ++++----- tools/perf/util/auxtrace.c | 4 ++-- tools/perf/util/record.c | 2 +- tools/perf/util/stat.c | 2 +- 5 files changed, 9 insertions(+), 14 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-c2c.c b/tools/perf/builtin-c2c.c index 14f228c7c869..1f1d17df9b9a 100644 --- a/tools/perf/builtin-c2c.c +++ b/tools/perf/builtin-c2c.c @@ -2319,11 +2319,7 @@ static int setup_nodes(struct perf_session *session) nodes[node] = set; - /* empty node, skip */ - if (perf_cpu_map__has_any_cpu_or_is_empty(map)) - continue; - - perf_cpu_map__for_each_cpu(cpu, idx, map) { + perf_cpu_map__for_each_cpu_skip_any(cpu, idx, map) { __set_bit(cpu.cpu, set); if (WARN_ONCE(cpu2node[cpu.cpu] != -1, "node/cpu topology bug")) diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index 6bba1a89d030..a47ced077fa6 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -1319,10 +1319,9 @@ static int cpu__get_cache_id_from_map(struct perf_cpu cpu, char *map) * be the first online CPU in the cache domain else use the * first online CPU of the cache domain as the ID. */ - if (perf_cpu_map__has_any_cpu_or_is_empty(cpu_map)) + id = perf_cpu_map__min(cpu_map).cpu; + if (id == -1) id = cpu.cpu; - else - id = perf_cpu_map__cpu(cpu_map, 0).cpu; /* Free the perf_cpu_map used to find the cache ID */ perf_cpu_map__put(cpu_map); @@ -1642,7 +1641,7 @@ static int perf_stat_init_aggr_mode(void) * taking the highest cpu number to be the size of * the aggregation translate cpumap. */ - if (!perf_cpu_map__has_any_cpu_or_is_empty(evsel_list->core.user_requested_cpus)) + if (!perf_cpu_map__is_any_cpu_or_is_empty(evsel_list->core.user_requested_cpus)) nr = perf_cpu_map__max(evsel_list->core.user_requested_cpus).cpu; else nr = 0; @@ -2334,7 +2333,7 @@ int process_stat_config_event(struct perf_session *session, perf_event__read_stat_config(&stat_config, &event->stat_config); - if (perf_cpu_map__has_any_cpu_or_is_empty(st->cpus)) { + if (perf_cpu_map__is_empty(st->cpus)) { if (st->aggr_mode != AGGR_UNSET) pr_warning("warning: processing task data, aggregation mode not set\n"); } else if (st->aggr_mode != AGGR_UNSET) { diff --git a/tools/perf/util/auxtrace.c b/tools/perf/util/auxtrace.c index ef314a5797e3..cfa2153d4611 100644 --- a/tools/perf/util/auxtrace.c +++ b/tools/perf/util/auxtrace.c @@ -174,7 +174,7 @@ void auxtrace_mmap_params__set_idx(struct auxtrace_mmap_params *mp, struct evlist *evlist, struct evsel *evsel, int idx) { - bool per_cpu = !perf_cpu_map__has_any_cpu_or_is_empty(evlist->core.user_requested_cpus); + bool per_cpu = !perf_cpu_map__has_any_cpu(evlist->core.user_requested_cpus); mp->mmap_needed = evsel->needs_auxtrace_mmap; @@ -648,7 +648,7 @@ int auxtrace_parse_snapshot_options(struct auxtrace_record *itr, static int evlist__enable_event_idx(struct evlist *evlist, struct evsel *evsel, int idx) { - bool per_cpu_mmaps = !perf_cpu_map__has_any_cpu_or_is_empty(evlist->core.user_requested_cpus); + bool per_cpu_mmaps = !perf_cpu_map__has_any_cpu(evlist->core.user_requested_cpus); if (per_cpu_mmaps) { struct perf_cpu evlist_cpu = perf_cpu_map__cpu(evlist->core.all_cpus, idx); diff --git a/tools/perf/util/record.c b/tools/perf/util/record.c index 87e817b3cf7e..e867de8ddaaa 100644 --- a/tools/perf/util/record.c +++ b/tools/perf/util/record.c @@ -237,7 +237,7 @@ bool evlist__can_select_event(struct evlist *evlist, const char *str) evsel = evlist__last(temp_evlist); - if (!evlist || perf_cpu_map__has_any_cpu_or_is_empty(evlist->core.user_requested_cpus)) { + if (!evlist || perf_cpu_map__is_any_cpu_or_is_empty(evlist->core.user_requested_cpus)) { struct perf_cpu_map *cpus = perf_cpu_map__new_online_cpus(); if (cpus) diff --git a/tools/perf/util/stat.c b/tools/perf/util/stat.c index b0bcf92f0f9c..0bd5467389e4 100644 --- a/tools/perf/util/stat.c +++ b/tools/perf/util/stat.c @@ -315,7 +315,7 @@ static int check_per_pkg(struct evsel *counter, struct perf_counts_values *vals, if (!counter->per_pkg) return 0; - if (perf_cpu_map__has_any_cpu_or_is_empty(cpus)) + if (perf_cpu_map__is_any_cpu_or_is_empty(cpus)) return 0; if (!mask) { -- cgit v1.2.3 From 71bc3ac8e8c93f769d1a2040153fe6d6b8093fa7 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Fri, 2 Feb 2024 15:40:57 -0800 Subject: perf cpumap: Use perf_cpu_map__for_each_cpu when possible MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Rather than manually iterating the CPU map, use perf_cpu_map__for_each_cpu(). When possible tidy local variables. Reviewed-by: James Clark Signed-off-by: Ian Rogers Acked-by: Namhyung Kim Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Alexandre Ghiti Cc: Andrew Jones Cc: AndrĂ© Almeida Cc: Athira Rajeev Cc: Atish Patra Cc: Changbin Du Cc: Darren Hart Cc: Davidlohr Bueso Cc: Huacai Chen Cc: Ingo Molnar Cc: Jiri Olsa Cc: John Garry Cc: K Prateek Nayak Cc: Kajol Jain Cc: Kan Liang Cc: Leo Yan Cc: Mark Rutland Cc: Mike Leach Cc: Nick Desaulniers Cc: Paolo Bonzini Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Sandipan Das Cc: Sean Christopherson Cc: Steinar H. Gunderson Cc: Suzuki Poulouse Cc: Thomas Gleixner Cc: Will Deacon Cc: Yang Jihong Cc: Yang Li Cc: Yanteng Si Link: https://lore.kernel.org/r/20240202234057.2085863-9-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/arch/arm64/util/header.c | 10 ++--- tools/perf/tests/bitmap.c | 13 +++--- tools/perf/tests/topology.c | 46 +++++++++++----------- tools/perf/util/bpf_kwork.c | 16 ++++---- tools/perf/util/bpf_kwork_top.c | 12 +++--- tools/perf/util/cpumap.c | 12 +++--- .../util/scripting-engines/trace-event-python.c | 12 +++--- tools/perf/util/session.c | 5 +-- tools/perf/util/svghelper.c | 20 +++++----- 9 files changed, 72 insertions(+), 74 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/arch/arm64/util/header.c b/tools/perf/arch/arm64/util/header.c index a9de0b5187dd..741df3614a09 100644 --- a/tools/perf/arch/arm64/util/header.c +++ b/tools/perf/arch/arm64/util/header.c @@ -4,8 +4,6 @@ #include #include #include -#include -#include #include #include #include "debug.h" @@ -19,18 +17,18 @@ static int _get_cpuid(char *buf, size_t sz, struct perf_cpu_map *cpus) { const char *sysfs = sysfs__mountpoint(); - int cpu; - int ret = EINVAL; + struct perf_cpu cpu; + int idx, ret = EINVAL; if (!sysfs || sz < MIDR_SIZE) return EINVAL; - for (cpu = 0; cpu < perf_cpu_map__nr(cpus); cpu++) { + perf_cpu_map__for_each_cpu(cpu, idx, cpus) { char path[PATH_MAX]; FILE *file; scnprintf(path, PATH_MAX, "%s/devices/system/cpu/cpu%d" MIDR, - sysfs, RC_CHK_ACCESS(cpus)->map[cpu].cpu); + sysfs, cpu.cpu); file = fopen(path, "r"); if (!file) { diff --git a/tools/perf/tests/bitmap.c b/tools/perf/tests/bitmap.c index 0173f5402a35..98956e0e0765 100644 --- a/tools/perf/tests/bitmap.c +++ b/tools/perf/tests/bitmap.c @@ -11,18 +11,19 @@ static unsigned long *get_bitmap(const char *str, int nbits) { struct perf_cpu_map *map = perf_cpu_map__new(str); - unsigned long *bm = NULL; - int i; + unsigned long *bm; bm = bitmap_zalloc(nbits); if (map && bm) { - for (i = 0; i < perf_cpu_map__nr(map); i++) - __set_bit(perf_cpu_map__cpu(map, i).cpu, bm); + int i; + struct perf_cpu cpu; + + perf_cpu_map__for_each_cpu(cpu, i, map) + __set_bit(cpu.cpu, bm); } - if (map) - perf_cpu_map__put(map); + perf_cpu_map__put(map); return bm; } diff --git a/tools/perf/tests/topology.c b/tools/perf/tests/topology.c index 2a842f53fbb5..a8cb5ba898ab 100644 --- a/tools/perf/tests/topology.c +++ b/tools/perf/tests/topology.c @@ -68,6 +68,7 @@ static int check_cpu_topology(char *path, struct perf_cpu_map *map) }; int i; struct aggr_cpu_id id; + struct perf_cpu cpu; session = perf_session__new(&data, NULL); TEST_ASSERT_VAL("can't get session", !IS_ERR(session)); @@ -113,8 +114,7 @@ static int check_cpu_topology(char *path, struct perf_cpu_map *map) TEST_ASSERT_VAL("Session header CPU map not set", session->header.env.cpu); for (i = 0; i < session->header.env.nr_cpus_avail; i++) { - struct perf_cpu cpu = { .cpu = i }; - + cpu.cpu = i; if (!perf_cpu_map__has(map, cpu)) continue; pr_debug("CPU %d, core %d, socket %d\n", i, @@ -123,48 +123,48 @@ static int check_cpu_topology(char *path, struct perf_cpu_map *map) } // Test that CPU ID contains socket, die, core and CPU - for (i = 0; i < perf_cpu_map__nr(map); i++) { - id = aggr_cpu_id__cpu(perf_cpu_map__cpu(map, i), NULL); + perf_cpu_map__for_each_cpu(cpu, i, map) { + id = aggr_cpu_id__cpu(cpu, NULL); TEST_ASSERT_VAL("Cpu map - CPU ID doesn't match", - perf_cpu_map__cpu(map, i).cpu == id.cpu.cpu); + cpu.cpu == id.cpu.cpu); TEST_ASSERT_VAL("Cpu map - Core ID doesn't match", - session->header.env.cpu[perf_cpu_map__cpu(map, i).cpu].core_id == id.core); + session->header.env.cpu[cpu.cpu].core_id == id.core); TEST_ASSERT_VAL("Cpu map - Socket ID doesn't match", - session->header.env.cpu[perf_cpu_map__cpu(map, i).cpu].socket_id == + session->header.env.cpu[cpu.cpu].socket_id == id.socket); TEST_ASSERT_VAL("Cpu map - Die ID doesn't match", - session->header.env.cpu[perf_cpu_map__cpu(map, i).cpu].die_id == id.die); + session->header.env.cpu[cpu.cpu].die_id == id.die); TEST_ASSERT_VAL("Cpu map - Node ID is set", id.node == -1); TEST_ASSERT_VAL("Cpu map - Thread IDX is set", id.thread_idx == -1); } // Test that core ID contains socket, die and core - for (i = 0; i < perf_cpu_map__nr(map); i++) { - id = aggr_cpu_id__core(perf_cpu_map__cpu(map, i), NULL); + perf_cpu_map__for_each_cpu(cpu, i, map) { + id = aggr_cpu_id__core(cpu, NULL); TEST_ASSERT_VAL("Core map - Core ID doesn't match", - session->header.env.cpu[perf_cpu_map__cpu(map, i).cpu].core_id == id.core); + session->header.env.cpu[cpu.cpu].core_id == id.core); TEST_ASSERT_VAL("Core map - Socket ID doesn't match", - session->header.env.cpu[perf_cpu_map__cpu(map, i).cpu].socket_id == + session->header.env.cpu[cpu.cpu].socket_id == id.socket); TEST_ASSERT_VAL("Core map - Die ID doesn't match", - session->header.env.cpu[perf_cpu_map__cpu(map, i).cpu].die_id == id.die); + session->header.env.cpu[cpu.cpu].die_id == id.die); TEST_ASSERT_VAL("Core map - Node ID is set", id.node == -1); TEST_ASSERT_VAL("Core map - Thread IDX is set", id.thread_idx == -1); } // Test that die ID contains socket and die - for (i = 0; i < perf_cpu_map__nr(map); i++) { - id = aggr_cpu_id__die(perf_cpu_map__cpu(map, i), NULL); + perf_cpu_map__for_each_cpu(cpu, i, map) { + id = aggr_cpu_id__die(cpu, NULL); TEST_ASSERT_VAL("Die map - Socket ID doesn't match", - session->header.env.cpu[perf_cpu_map__cpu(map, i).cpu].socket_id == + session->header.env.cpu[cpu.cpu].socket_id == id.socket); TEST_ASSERT_VAL("Die map - Die ID doesn't match", - session->header.env.cpu[perf_cpu_map__cpu(map, i).cpu].die_id == id.die); + session->header.env.cpu[cpu.cpu].die_id == id.die); TEST_ASSERT_VAL("Die map - Node ID is set", id.node == -1); TEST_ASSERT_VAL("Die map - Core is set", id.core == -1); @@ -173,10 +173,10 @@ static int check_cpu_topology(char *path, struct perf_cpu_map *map) } // Test that socket ID contains only socket - for (i = 0; i < perf_cpu_map__nr(map); i++) { - id = aggr_cpu_id__socket(perf_cpu_map__cpu(map, i), NULL); + perf_cpu_map__for_each_cpu(cpu, i, map) { + id = aggr_cpu_id__socket(cpu, NULL); TEST_ASSERT_VAL("Socket map - Socket ID doesn't match", - session->header.env.cpu[perf_cpu_map__cpu(map, i).cpu].socket_id == + session->header.env.cpu[cpu.cpu].socket_id == id.socket); TEST_ASSERT_VAL("Socket map - Node ID is set", id.node == -1); @@ -187,10 +187,10 @@ static int check_cpu_topology(char *path, struct perf_cpu_map *map) } // Test that node ID contains only node - for (i = 0; i < perf_cpu_map__nr(map); i++) { - id = aggr_cpu_id__node(perf_cpu_map__cpu(map, i), NULL); + perf_cpu_map__for_each_cpu(cpu, i, map) { + id = aggr_cpu_id__node(cpu, NULL); TEST_ASSERT_VAL("Node map - Node ID doesn't match", - cpu__get_node(perf_cpu_map__cpu(map, i)) == id.node); + cpu__get_node(cpu) == id.node); TEST_ASSERT_VAL("Node map - Socket is set", id.socket == -1); TEST_ASSERT_VAL("Node map - Die ID is set", id.die == -1); TEST_ASSERT_VAL("Node map - Core is set", id.core == -1); diff --git a/tools/perf/util/bpf_kwork.c b/tools/perf/util/bpf_kwork.c index 6eb2c78fd7f4..44f0f708a15d 100644 --- a/tools/perf/util/bpf_kwork.c +++ b/tools/perf/util/bpf_kwork.c @@ -147,12 +147,12 @@ static bool valid_kwork_class_type(enum kwork_class_type type) static int setup_filters(struct perf_kwork *kwork) { - u8 val = 1; - int i, nr_cpus, key, fd; - struct perf_cpu_map *map; - if (kwork->cpu_list != NULL) { - fd = bpf_map__fd(skel->maps.perf_kwork_cpu_filter); + int idx, nr_cpus; + struct perf_cpu_map *map; + struct perf_cpu cpu; + int fd = bpf_map__fd(skel->maps.perf_kwork_cpu_filter); + if (fd < 0) { pr_debug("Invalid cpu filter fd\n"); return -1; @@ -165,8 +165,8 @@ static int setup_filters(struct perf_kwork *kwork) } nr_cpus = libbpf_num_possible_cpus(); - for (i = 0; i < perf_cpu_map__nr(map); i++) { - struct perf_cpu cpu = perf_cpu_map__cpu(map, i); + perf_cpu_map__for_each_cpu(cpu, idx, map) { + u8 val = 1; if (cpu.cpu >= nr_cpus) { perf_cpu_map__put(map); @@ -181,6 +181,8 @@ static int setup_filters(struct perf_kwork *kwork) } if (kwork->profile_name != NULL) { + int key, fd; + if (strlen(kwork->profile_name) >= MAX_KWORKNAME) { pr_err("Requested name filter %s too large, limit to %d\n", kwork->profile_name, MAX_KWORKNAME - 1); diff --git a/tools/perf/util/bpf_kwork_top.c b/tools/perf/util/bpf_kwork_top.c index 035e02272790..22a3b00a1e23 100644 --- a/tools/perf/util/bpf_kwork_top.c +++ b/tools/perf/util/bpf_kwork_top.c @@ -122,11 +122,11 @@ static bool valid_kwork_class_type(enum kwork_class_type type) static int setup_filters(struct perf_kwork *kwork) { - u8 val = 1; - int i, nr_cpus, fd; - struct perf_cpu_map *map; - if (kwork->cpu_list) { + int idx, nr_cpus, fd; + struct perf_cpu_map *map; + struct perf_cpu cpu; + fd = bpf_map__fd(skel->maps.kwork_top_cpu_filter); if (fd < 0) { pr_debug("Invalid cpu filter fd\n"); @@ -140,8 +140,8 @@ static int setup_filters(struct perf_kwork *kwork) } nr_cpus = libbpf_num_possible_cpus(); - for (i = 0; i < perf_cpu_map__nr(map); i++) { - struct perf_cpu cpu = perf_cpu_map__cpu(map, i); + perf_cpu_map__for_each_cpu(cpu, idx, map) { + u8 val = 1; if (cpu.cpu >= nr_cpus) { perf_cpu_map__put(map); diff --git a/tools/perf/util/cpumap.c b/tools/perf/util/cpumap.c index 356e30c42cd8..6a270d640acb 100644 --- a/tools/perf/util/cpumap.c +++ b/tools/perf/util/cpumap.c @@ -655,10 +655,10 @@ static char hex_char(unsigned char val) size_t cpu_map__snprint_mask(struct perf_cpu_map *map, char *buf, size_t size) { - int i, cpu; + int idx; char *ptr = buf; unsigned char *bitmap; - struct perf_cpu last_cpu = perf_cpu_map__cpu(map, perf_cpu_map__nr(map) - 1); + struct perf_cpu c, last_cpu = perf_cpu_map__max(map); if (buf == NULL) return 0; @@ -669,12 +669,10 @@ size_t cpu_map__snprint_mask(struct perf_cpu_map *map, char *buf, size_t size) return 0; } - for (i = 0; i < perf_cpu_map__nr(map); i++) { - cpu = perf_cpu_map__cpu(map, i).cpu; - bitmap[cpu / 8] |= 1 << (cpu % 8); - } + perf_cpu_map__for_each_cpu(c, idx, map) + bitmap[c.cpu / 8] |= 1 << (c.cpu % 8); - for (cpu = last_cpu.cpu / 4 * 4; cpu >= 0; cpu -= 4) { + for (int cpu = last_cpu.cpu / 4 * 4; cpu >= 0; cpu -= 4) { unsigned char bits = bitmap[cpu / 8]; if (cpu % 8) diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c index b4f0f60e60a6..8aa301948de5 100644 --- a/tools/perf/util/scripting-engines/trace-event-python.c +++ b/tools/perf/util/scripting-engines/trace-event-python.c @@ -1699,13 +1699,15 @@ static void python_process_stat(struct perf_stat_config *config, { struct perf_thread_map *threads = counter->core.threads; struct perf_cpu_map *cpus = counter->core.cpus; - int cpu, thread; - for (thread = 0; thread < perf_thread_map__nr(threads); thread++) { - for (cpu = 0; cpu < perf_cpu_map__nr(cpus); cpu++) { - process_stat(counter, perf_cpu_map__cpu(cpus, cpu), + for (int thread = 0; thread < perf_thread_map__nr(threads); thread++) { + int idx; + struct perf_cpu cpu; + + perf_cpu_map__for_each_cpu(cpu, idx, cpus) { + process_stat(counter, cpu, perf_thread_map__pid(threads, thread), tstamp, - perf_counts(counter->counts, cpu, thread)); + perf_counts(counter->counts, idx, thread)); } } } diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 06d0bd7fb459..02a932a83c51 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -2749,6 +2749,7 @@ int perf_session__cpu_bitmap(struct perf_session *session, int i, err = -1; struct perf_cpu_map *map; int nr_cpus = min(session->header.env.nr_cpus_avail, MAX_NR_CPUS); + struct perf_cpu cpu; for (i = 0; i < PERF_TYPE_MAX; ++i) { struct evsel *evsel; @@ -2770,9 +2771,7 @@ int perf_session__cpu_bitmap(struct perf_session *session, return -1; } - for (i = 0; i < perf_cpu_map__nr(map); i++) { - struct perf_cpu cpu = perf_cpu_map__cpu(map, i); - + perf_cpu_map__for_each_cpu(cpu, i, map) { if (cpu.cpu >= nr_cpus) { pr_err("Requested CPU %d too large. " "Consider raising MAX_NR_CPUS\n", cpu.cpu); diff --git a/tools/perf/util/svghelper.c b/tools/perf/util/svghelper.c index 1892e9b6aa7f..2b04f47f4db0 100644 --- a/tools/perf/util/svghelper.c +++ b/tools/perf/util/svghelper.c @@ -725,26 +725,24 @@ static void scan_core_topology(int *map, struct topology *t, int nr_cpus) static int str_to_bitmap(char *s, cpumask_t *b, int nr_cpus) { - int i; - int ret = 0; - struct perf_cpu_map *m; - struct perf_cpu c; + int idx, ret = 0; + struct perf_cpu_map *map; + struct perf_cpu cpu; - m = perf_cpu_map__new(s); - if (!m) + map = perf_cpu_map__new(s); + if (!map) return -1; - for (i = 0; i < perf_cpu_map__nr(m); i++) { - c = perf_cpu_map__cpu(m, i); - if (c.cpu >= nr_cpus) { + perf_cpu_map__for_each_cpu(cpu, idx, map) { + if (cpu.cpu >= nr_cpus) { ret = -1; break; } - __set_bit(c.cpu, cpumask_bits(b)); + __set_bit(cpu.cpu, cpumask_bits(b)); } - perf_cpu_map__put(m); + perf_cpu_map__put(map); return ret; } -- cgit v1.2.3 From b508965d35321534e84daf8946b2cc5f64517db9 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:50:53 -0700 Subject: perf dwarf-aux: Remove unused pc argument It's not used, let's get rid of it. Signed-off-by: Namhyung Kim Reviewed-by: Arnaldo Carvalho de Melo Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-2-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 4 ++-- tools/perf/util/dwarf-aux.c | 7 ++----- tools/perf/util/dwarf-aux.h | 6 ++---- 3 files changed, 6 insertions(+), 11 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 30c4d19fcf11..59ce5f4f4a40 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -263,7 +263,7 @@ static int find_data_type_die(struct debuginfo *di, u64 pc, u64 addr, offset = loc->offset; if (reg == DWARF_REG_PC) { - if (die_find_variable_by_addr(&cu_die, pc, addr, &var_die, &offset)) { + if (die_find_variable_by_addr(&cu_die, addr, &var_die, &offset)) { ret = check_variable(&var_die, type_die, offset, /*is_pointer=*/false); loc->offset = offset; @@ -312,7 +312,7 @@ retry: /* Search from the inner-most scope to the outer */ for (i = nr_scopes - 1; i >= 0; i--) { if (reg == DWARF_REG_PC) { - if (!die_find_variable_by_addr(&scopes[i], pc, addr, + if (!die_find_variable_by_addr(&scopes[i], addr, &var_die, &offset)) continue; } else { diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c index 2791126069b4..e84d0d6a7750 100644 --- a/tools/perf/util/dwarf-aux.c +++ b/tools/perf/util/dwarf-aux.c @@ -1456,7 +1456,6 @@ static int __die_find_var_addr_cb(Dwarf_Die *die_mem, void *arg) /** * die_find_variable_by_addr - Find variable located at given address * @sc_die: a scope DIE - * @pc: the program address to find * @addr: the data address to find * @die_mem: a buffer to save the resulting DIE * @offset: the offset in the resulting type @@ -1464,12 +1463,10 @@ static int __die_find_var_addr_cb(Dwarf_Die *die_mem, void *arg) * Find the variable DIE located at the given address (in PC-relative mode). * This is usually for global variables. */ -Dwarf_Die *die_find_variable_by_addr(Dwarf_Die *sc_die, Dwarf_Addr pc, - Dwarf_Addr addr, Dwarf_Die *die_mem, - int *offset) +Dwarf_Die *die_find_variable_by_addr(Dwarf_Die *sc_die, Dwarf_Addr addr, + Dwarf_Die *die_mem, int *offset) { struct find_var_data data = { - .pc = pc, .addr = addr, }; Dwarf_Die *result; diff --git a/tools/perf/util/dwarf-aux.h b/tools/perf/util/dwarf-aux.h index 85dd527ae1f7..9973801a20c1 100644 --- a/tools/perf/util/dwarf-aux.h +++ b/tools/perf/util/dwarf-aux.h @@ -146,9 +146,8 @@ Dwarf_Die *die_find_variable_by_reg(Dwarf_Die *sc_die, Dwarf_Addr pc, int reg, Dwarf_Die *die_mem); /* Find a (global) variable located in the 'addr' */ -Dwarf_Die *die_find_variable_by_addr(Dwarf_Die *sc_die, Dwarf_Addr pc, - Dwarf_Addr addr, Dwarf_Die *die_mem, - int *offset); +Dwarf_Die *die_find_variable_by_addr(Dwarf_Die *sc_die, Dwarf_Addr addr, + Dwarf_Die *die_mem, int *offset); #else /* HAVE_DWARF_GETLOCATIONS_SUPPORT */ @@ -170,7 +169,6 @@ static inline Dwarf_Die *die_find_variable_by_reg(Dwarf_Die *sc_die __maybe_unus } static inline Dwarf_Die *die_find_variable_by_addr(Dwarf_Die *sc_die __maybe_unused, - Dwarf_Addr pc __maybe_unused, Dwarf_Addr addr __maybe_unused, Dwarf_Die *die_mem __maybe_unused, int *offset __maybe_unused) -- cgit v1.2.3 From 932dcc2c39aedf54ef291bc0b4129a54f5fe1e84 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:50:54 -0700 Subject: perf dwarf-aux: Add die_collect_vars() The die_collect_vars() is to find all variable information in the scope including function parameters. The struct die_var_type is to save the type of the variable with the location (reg and offset) as well as where it's defined in the code (addr). Signed-off-by: Namhyung Kim Acked-by: Masami Hiramatsu Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-3-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dwarf-aux.c | 118 +++++++++++++++++++++++++++++++++----------- tools/perf/util/dwarf-aux.h | 17 +++++++ 2 files changed, 107 insertions(+), 28 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c index e84d0d6a7750..785aa7a3d725 100644 --- a/tools/perf/util/dwarf-aux.c +++ b/tools/perf/util/dwarf-aux.c @@ -1136,6 +1136,40 @@ int die_get_varname(Dwarf_Die *vr_die, struct strbuf *buf) return ret < 0 ? ret : strbuf_addf(buf, "\t%s", dwarf_diename(vr_die)); } +#if defined(HAVE_DWARF_GETLOCATIONS_SUPPORT) || defined(HAVE_DWARF_CFI_SUPPORT) +static int reg_from_dwarf_op(Dwarf_Op *op) +{ + switch (op->atom) { + case DW_OP_reg0 ... DW_OP_reg31: + return op->atom - DW_OP_reg0; + case DW_OP_breg0 ... DW_OP_breg31: + return op->atom - DW_OP_breg0; + case DW_OP_regx: + case DW_OP_bregx: + return op->number; + default: + break; + } + return -1; +} + +static int offset_from_dwarf_op(Dwarf_Op *op) +{ + switch (op->atom) { + case DW_OP_reg0 ... DW_OP_reg31: + case DW_OP_regx: + return 0; + case DW_OP_breg0 ... DW_OP_breg31: + return op->number; + case DW_OP_bregx: + return op->number2; + default: + break; + } + return -1; +} +#endif /* HAVE_DWARF_GETLOCATIONS_SUPPORT || HAVE_DWARF_CFI_SUPPORT */ + #ifdef HAVE_DWARF_GETLOCATIONS_SUPPORT /** * die_get_var_innermost_scope - Get innermost scope range of given variable DIE @@ -1476,41 +1510,69 @@ Dwarf_Die *die_find_variable_by_addr(Dwarf_Die *sc_die, Dwarf_Addr addr, *offset = data.offset; return result; } -#endif /* HAVE_DWARF_GETLOCATIONS_SUPPORT */ -#ifdef HAVE_DWARF_CFI_SUPPORT -static int reg_from_dwarf_op(Dwarf_Op *op) +static int __die_collect_vars_cb(Dwarf_Die *die_mem, void *arg) { - switch (op->atom) { - case DW_OP_reg0 ... DW_OP_reg31: - return op->atom - DW_OP_reg0; - case DW_OP_breg0 ... DW_OP_breg31: - return op->atom - DW_OP_breg0; - case DW_OP_regx: - case DW_OP_bregx: - return op->number; - default: - break; - } - return -1; + struct die_var_type **var_types = arg; + Dwarf_Die type_die; + int tag = dwarf_tag(die_mem); + Dwarf_Attribute attr; + Dwarf_Addr base, start, end; + Dwarf_Op *ops; + size_t nops; + struct die_var_type *vt; + + if (tag != DW_TAG_variable && tag != DW_TAG_formal_parameter) + return DIE_FIND_CB_SIBLING; + + if (dwarf_attr(die_mem, DW_AT_location, &attr) == NULL) + return DIE_FIND_CB_SIBLING; + + /* + * Only collect the first location as it can reconstruct the + * remaining state by following the instructions. + * start = 0 means it covers the whole range. + */ + if (dwarf_getlocations(&attr, 0, &base, &start, &end, &ops, &nops) <= 0) + return DIE_FIND_CB_SIBLING; + + if (die_get_real_type(die_mem, &type_die) == NULL) + return DIE_FIND_CB_SIBLING; + + vt = malloc(sizeof(*vt)); + if (vt == NULL) + return DIE_FIND_CB_END; + + vt->die_off = dwarf_dieoffset(&type_die); + vt->addr = start; + vt->reg = reg_from_dwarf_op(ops); + vt->offset = offset_from_dwarf_op(ops); + vt->next = *var_types; + *var_types = vt; + + return DIE_FIND_CB_SIBLING; } -static int offset_from_dwarf_op(Dwarf_Op *op) +/** + * die_collect_vars - Save all variables and parameters + * @sc_die: a scope DIE + * @var_types: a pointer to save the resulting list + * + * Save all variables and parameters in the @sc_die and save them to @var_types. + * The @var_types is a singly-linked list containing type and location info. + * Actual type can be retrieved using dwarf_offdie() with 'die_off' later. + * + * Callers should free @var_types. + */ +void die_collect_vars(Dwarf_Die *sc_die, struct die_var_type **var_types) { - switch (op->atom) { - case DW_OP_reg0 ... DW_OP_reg31: - case DW_OP_regx: - return 0; - case DW_OP_breg0 ... DW_OP_breg31: - return op->number; - case DW_OP_bregx: - return op->number2; - default: - break; - } - return -1; + Dwarf_Die die_mem; + + die_find_child(sc_die, __die_collect_vars_cb, (void *)var_types, &die_mem); } +#endif /* HAVE_DWARF_GETLOCATIONS_SUPPORT */ +#ifdef HAVE_DWARF_CFI_SUPPORT /** * die_get_cfa - Get frame base information * @dwarf: a Dwarf info diff --git a/tools/perf/util/dwarf-aux.h b/tools/perf/util/dwarf-aux.h index 9973801a20c1..cd171b06fd4c 100644 --- a/tools/perf/util/dwarf-aux.h +++ b/tools/perf/util/dwarf-aux.h @@ -135,6 +135,15 @@ void die_skip_prologue(Dwarf_Die *sp_die, Dwarf_Die *cu_die, /* Get the list of including scopes */ int die_get_scopes(Dwarf_Die *cu_die, Dwarf_Addr pc, Dwarf_Die **scopes); +/* Variable type information */ +struct die_var_type { + struct die_var_type *next; + u64 die_off; + u64 addr; + int reg; + int offset; +}; + #ifdef HAVE_DWARF_GETLOCATIONS_SUPPORT /* Get byte offset range of given variable DIE */ @@ -149,6 +158,9 @@ Dwarf_Die *die_find_variable_by_reg(Dwarf_Die *sc_die, Dwarf_Addr pc, int reg, Dwarf_Die *die_find_variable_by_addr(Dwarf_Die *sc_die, Dwarf_Addr addr, Dwarf_Die *die_mem, int *offset); +/* Save all variables and parameters in this scope */ +void die_collect_vars(Dwarf_Die *sc_die, struct die_var_type **var_types); + #else /* HAVE_DWARF_GETLOCATIONS_SUPPORT */ static inline int die_get_var_range(Dwarf_Die *sp_die __maybe_unused, @@ -176,6 +188,11 @@ static inline Dwarf_Die *die_find_variable_by_addr(Dwarf_Die *sc_die __maybe_unu return NULL; } +static inline void die_collect_vars(Dwarf_Die *sc_die __maybe_unused, + struct die_var_type **var_types __maybe_unused) +{ +} + #endif /* HAVE_DWARF_GETLOCATIONS_SUPPORT */ #ifdef HAVE_DWARF_CFI_SUPPORT -- cgit v1.2.3 From 437683a9941891c17059d99561eb3ce85e0f51fa Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:50:55 -0700 Subject: perf dwarf-aux: Handle type transfer for memory access We want to track type states as instructions are executed. Each instruction can access compound types like struct or union and load/ store its members to a different location. The die_deref_ptr_type() is to find a type of memory access with a pointer variable. If it points to a compound type like struct, the target memory is a member in the struct. The access will happen with an offset indicating which member it refers. Let's follow the DWARF info to figure out the type of the pointer target. For example, say we have the following code. struct foo { int a; int b; }; struct foo *p = malloc(sizeof(*p)); p->b = 0; The last pointer access should produce x86 asm like below: mov 0x0, 4(%rbx) And we know %rbx register has a pointer to struct foo. Then offset 4 should return the debug info of member 'b'. Also variables of compound types can be accessed directly without a pointer. The die_get_member_type() is to handle a such case. Signed-off-by: Namhyung Kim Acked-by: Masami Hiramatsu Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-4-namhyung@kernel.org [ Check if die_get_real_type() returned NULL ] Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dwarf-aux.c | 113 ++++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/dwarf-aux.h | 6 +++ 2 files changed, 119 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c index 785aa7a3d725..09fd6f1f0ed8 100644 --- a/tools/perf/util/dwarf-aux.c +++ b/tools/perf/util/dwarf-aux.c @@ -1838,3 +1838,116 @@ int die_get_scopes(Dwarf_Die *cu_die, Dwarf_Addr pc, Dwarf_Die **scopes) *scopes = data.scopes; return data.nr; } + +static int __die_find_member_offset_cb(Dwarf_Die *die_mem, void *arg) +{ + Dwarf_Die type_die; + Dwarf_Word size, loc; + Dwarf_Word offset = (long)arg; + int tag = dwarf_tag(die_mem); + + if (tag != DW_TAG_member) + return DIE_FIND_CB_SIBLING; + + /* Unions might not have location */ + if (die_get_data_member_location(die_mem, &loc) < 0) + loc = 0; + + if (offset == loc) + return DIE_FIND_CB_END; + + if (die_get_real_type(die_mem, &type_die) == NULL) { + // TODO: add a pr_debug_dtp() later for this unlikely failure + return DIE_FIND_CB_SIBLING; + } + + if (dwarf_aggregate_size(&type_die, &size) < 0) + size = 0; + + if (loc < offset && offset < (loc + size)) + return DIE_FIND_CB_END; + + return DIE_FIND_CB_SIBLING; +} + +/** + * die_get_member_type - Return type info of struct member + * @type_die: a type DIE + * @offset: offset in the type + * @die_mem: a buffer to save the resulting DIE + * + * This function returns a type of a member in @type_die where it's located at + * @offset if it's a struct. For now, it just returns the first matching + * member in a union. For other types, it'd return the given type directly + * if it's within the size of the type or NULL otherwise. + */ +Dwarf_Die *die_get_member_type(Dwarf_Die *type_die, int offset, + Dwarf_Die *die_mem) +{ + Dwarf_Die *member; + Dwarf_Die mb_type; + int tag; + + tag = dwarf_tag(type_die); + /* If it's not a compound type, return the type directly */ + if (tag != DW_TAG_structure_type && tag != DW_TAG_union_type) { + Dwarf_Word size; + + if (dwarf_aggregate_size(type_die, &size) < 0) + size = 0; + + if ((unsigned)offset >= size) + return NULL; + + *die_mem = *type_die; + return die_mem; + } + + mb_type = *type_die; + /* TODO: Handle union types better? */ + while (tag == DW_TAG_structure_type || tag == DW_TAG_union_type) { + member = die_find_child(&mb_type, __die_find_member_offset_cb, + (void *)(long)offset, die_mem); + if (member == NULL) + return NULL; + + if (die_get_real_type(member, &mb_type) == NULL) + return NULL; + + tag = dwarf_tag(&mb_type); + + if (tag == DW_TAG_structure_type || tag == DW_TAG_union_type) { + Dwarf_Word loc; + + /* Update offset for the start of the member struct */ + if (die_get_data_member_location(member, &loc) == 0) + offset -= loc; + } + } + *die_mem = mb_type; + return die_mem; +} + +/** + * die_deref_ptr_type - Return type info for pointer access + * @ptr_die: a pointer type DIE + * @offset: access offset for the pointer + * @die_mem: a buffer to save the resulting DIE + * + * This function follows the pointer in @ptr_die with given @offset + * and saves the resulting type in @die_mem. If the pointer points + * a struct type, actual member at the offset would be returned. + */ +Dwarf_Die *die_deref_ptr_type(Dwarf_Die *ptr_die, int offset, + Dwarf_Die *die_mem) +{ + Dwarf_Die type_die; + + if (dwarf_tag(ptr_die) != DW_TAG_pointer_type) + return NULL; + + if (die_get_real_type(ptr_die, &type_die) == NULL) + return NULL; + + return die_get_member_type(&type_die, offset, die_mem); +} diff --git a/tools/perf/util/dwarf-aux.h b/tools/perf/util/dwarf-aux.h index cd171b06fd4c..16c916311bc0 100644 --- a/tools/perf/util/dwarf-aux.h +++ b/tools/perf/util/dwarf-aux.h @@ -144,6 +144,12 @@ struct die_var_type { int offset; }; +/* Return type info of a member at offset */ +Dwarf_Die *die_get_member_type(Dwarf_Die *type_die, int offset, Dwarf_Die *die_mem); + +/* Return type info where the pointer and offset point to */ +Dwarf_Die *die_deref_ptr_type(Dwarf_Die *ptr_die, int offset, Dwarf_Die *die_mem); + #ifdef HAVE_DWARF_GETLOCATIONS_SUPPORT /* Get byte offset range of given variable DIE */ -- cgit v1.2.3 From 7a838c2fd2ac81e10bedcd912153a74ca662b309 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:50:56 -0700 Subject: perf dwarf-aux: Add die_find_func_rettype() The die_find_func_rettype() is to find a debug entry for the given function name and sets the type information of the return value. By convention, it'd return the pointer to the type die (should be the same as the given mem_die argument) if found, or NULL otherwise. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-5-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dwarf-aux.c | 43 +++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/dwarf-aux.h | 4 ++++ 2 files changed, 47 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c index 09fd6f1f0ed8..d4ec239c5adb 100644 --- a/tools/perf/util/dwarf-aux.c +++ b/tools/perf/util/dwarf-aux.c @@ -696,6 +696,49 @@ Dwarf_Die *die_find_inlinefunc(Dwarf_Die *sp_die, Dwarf_Addr addr, return die_mem; } +static int __die_find_func_rettype_cb(Dwarf_Die *die_mem, void *data) +{ + const char *func_name; + + if (dwarf_tag(die_mem) != DW_TAG_subprogram) + return DIE_FIND_CB_SIBLING; + + func_name = dwarf_diename(die_mem); + if (func_name && !strcmp(func_name, data)) + return DIE_FIND_CB_END; + + return DIE_FIND_CB_SIBLING; +} + +/** + * die_find_func_rettype - Search a return type of function + * @cu_die: a CU DIE + * @name: target function name + * @die_mem: a buffer for result DIE + * + * Search a non-inlined function which matches to @name and stores the + * return type of the function to @die_mem and returns it if found. + * Returns NULL if failed. Note that it doesn't needs to find a + * definition of the function, so it doesn't match with address. + * Most likely, it can find a declaration at the top level. Thus the + * callback function continues to sibling entries only. + */ +Dwarf_Die *die_find_func_rettype(Dwarf_Die *cu_die, const char *name, + Dwarf_Die *die_mem) +{ + Dwarf_Die tmp_die; + + cu_die = die_find_child(cu_die, __die_find_func_rettype_cb, + (void *)name, &tmp_die); + if (!cu_die) + return NULL; + + if (die_get_real_type(&tmp_die, die_mem) == NULL) + return NULL; + + return die_mem; +} + struct __instance_walk_param { void *addr; int (*callback)(Dwarf_Die *, void *); diff --git a/tools/perf/util/dwarf-aux.h b/tools/perf/util/dwarf-aux.h index 16c916311bc0..b0f25fbf9668 100644 --- a/tools/perf/util/dwarf-aux.h +++ b/tools/perf/util/dwarf-aux.h @@ -94,6 +94,10 @@ Dwarf_Die *die_find_top_inlinefunc(Dwarf_Die *sp_die, Dwarf_Addr addr, Dwarf_Die *die_find_inlinefunc(Dwarf_Die *sp_die, Dwarf_Addr addr, Dwarf_Die *die_mem); +/* Search a non-inlined function by name and returns its return type */ +Dwarf_Die *die_find_func_rettype(Dwarf_Die *sp_die, const char *name, + Dwarf_Die *die_mem); + /* Walk on the instances of given DIE */ int die_walk_instances(Dwarf_Die *in_die, int (*callback)(Dwarf_Die *, void *), void *data); -- cgit v1.2.3 From 52a09bc24c6a4deeeb8030476a7663b6696d73f0 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:50:57 -0700 Subject: perf map: Add map__objdump_2rip() Sometimes we want to convert an address in objdump output to map-relative address to match with a sample data. Let's add map__objdump_2rip() for that. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-6-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/map.c | 17 +++++++++++++++++ tools/perf/util/map.h | 3 +++ 2 files changed, 20 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index 14a5ea70d81e..a5d57c201a30 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -587,6 +587,23 @@ u64 map__objdump_2mem(struct map *map, u64 ip) return ip + map__reloc(map); } +/* convert objdump address to relative address. (To be removed) */ +u64 map__objdump_2rip(struct map *map, u64 ip) +{ + const struct dso *dso = map__dso(map); + + if (!dso->adjust_symbols) + return ip; + + if (dso->rel) + return ip + map__pgoff(map); + + if (dso->kernel == DSO_SPACE__USER) + return ip - dso->text_offset; + + return map__map_ip(map, ip + map__reloc(map)); +} + bool map__contains_symbol(const struct map *map, const struct symbol *sym) { u64 ip = map__unmap_ip(map, sym->start); diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h index 49756716cb13..65e2609fa1b1 100644 --- a/tools/perf/util/map.h +++ b/tools/perf/util/map.h @@ -132,6 +132,9 @@ u64 map__rip_2objdump(struct map *map, u64 rip); /* objdump address -> memory address */ u64 map__objdump_2mem(struct map *map, u64 ip); +/* objdump address -> rip */ +u64 map__objdump_2rip(struct map *map, u64 ip); + struct symbol; struct thread; -- cgit v1.2.3 From a3f4d5b57dd8fd2a749be261d7253616f15671b5 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:50:58 -0700 Subject: perf annotate-data: Introduce 'struct data_loc_info' The find_data_type() needs many information to describe the location of the data. Add the new 'struct data_loc_info' to pass those information at once. No functional changes intended. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-7-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 83 ++++++++++++++++++++++------------------- tools/perf/util/annotate-data.h | 38 ++++++++++++++++--- tools/perf/util/annotate.c | 30 +++++++-------- 3 files changed, 91 insertions(+), 60 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 59ce5f4f4a40..ff81d164aa57 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -239,21 +239,28 @@ static int check_variable(Dwarf_Die *var_die, Dwarf_Die *type_die, int offset, } /* The result will be saved in @type_die */ -static int find_data_type_die(struct debuginfo *di, u64 pc, u64 addr, - const char *var_name, struct annotated_op_loc *loc, - Dwarf_Die *type_die) +static int find_data_type_die(struct data_loc_info *dloc, Dwarf_Die *type_die) { + struct annotated_op_loc *loc = dloc->op; Dwarf_Die cu_die, var_die; Dwarf_Die *scopes = NULL; int reg, offset; int ret = -1; int i, nr_scopes; int fbreg = -1; - bool is_fbreg = false; int fb_offset = 0; + bool is_fbreg = false; + u64 pc; + + /* + * IP is a relative instruction address from the start of the map, as + * it can be randomized/relocated, it needs to translate to PC which is + * a file address for DWARF processing. + */ + pc = map__rip_2objdump(dloc->ms->map, dloc->ip); /* Get a compile_unit for this address */ - if (!find_cu_die(di, pc, &cu_die)) { + if (!find_cu_die(dloc->di, pc, &cu_die)) { pr_debug("cannot find CU for address %" PRIx64 "\n", pc); ann_data_stat.no_cuinfo++; return -1; @@ -263,18 +270,19 @@ static int find_data_type_die(struct debuginfo *di, u64 pc, u64 addr, offset = loc->offset; if (reg == DWARF_REG_PC) { - if (die_find_variable_by_addr(&cu_die, addr, &var_die, &offset)) { + if (die_find_variable_by_addr(&cu_die, dloc->var_addr, &var_die, + &offset)) { ret = check_variable(&var_die, type_die, offset, /*is_pointer=*/false); - loc->offset = offset; + dloc->type_offset = offset; goto out; } - if (var_name && die_find_variable_at(&cu_die, var_name, pc, - &var_die)) { - ret = check_variable(&var_die, type_die, 0, + if (dloc->var_name && + die_find_variable_at(&cu_die, dloc->var_name, pc, &var_die)) { + ret = check_variable(&var_die, type_die, dloc->type_offset, /*is_pointer=*/false); - /* loc->offset will be updated by the caller */ + /* dloc->type_offset was updated by the caller */ goto out; } } @@ -291,10 +299,11 @@ static int find_data_type_die(struct debuginfo *di, u64 pc, u64 addr, dwarf_formblock(&attr, &block) == 0 && block.length == 1) { switch (*block.data) { case DW_OP_reg0 ... DW_OP_reg31: - fbreg = *block.data - DW_OP_reg0; + fbreg = dloc->fbreg = *block.data - DW_OP_reg0; break; case DW_OP_call_frame_cfa: - if (die_get_cfa(di->dbg, pc, &fbreg, + dloc->fb_cfa = true; + if (die_get_cfa(dloc->di->dbg, pc, &fbreg, &fb_offset) < 0) fbreg = -1; break; @@ -312,7 +321,7 @@ retry: /* Search from the inner-most scope to the outer */ for (i = nr_scopes - 1; i >= 0; i--) { if (reg == DWARF_REG_PC) { - if (!die_find_variable_by_addr(&scopes[i], addr, + if (!die_find_variable_by_addr(&scopes[i], dloc->var_addr, &var_die, &offset)) continue; } else { @@ -325,7 +334,7 @@ retry: /* Found a variable, see if it's correct */ ret = check_variable(&var_die, type_die, offset, reg != DWARF_REG_PC && !is_fbreg); - loc->offset = offset; + dloc->type_offset = offset; goto out; } @@ -344,50 +353,46 @@ out: /** * find_data_type - Return a data type at the location - * @ms: map and symbol at the location - * @ip: instruction address of the memory access - * @loc: instruction operand location - * @addr: data address of the memory access - * @var_name: global variable name + * @dloc: data location * * This functions searches the debug information of the binary to get the data - * type it accesses. The exact location is expressed by (@ip, reg, offset) - * for pointer variables or (@ip, @addr) for global variables. Note that global - * variables might update the @loc->offset after finding the start of the variable. - * If it cannot find a global variable by address, it tried to fine a declaration - * of the variable using @var_name. In that case, @loc->offset won't be updated. + * type it accesses. The exact location is expressed by (ip, reg, offset) + * for pointer variables or (ip, addr) for global variables. Note that global + * variables might update the @dloc->type_offset after finding the start of the + * variable. If it cannot find a global variable by address, it tried to find + * a declaration of the variable using var_name. In that case, @dloc->offset + * won't be updated. * * It return %NULL if not found. */ -struct annotated_data_type *find_data_type(struct map_symbol *ms, u64 ip, - struct annotated_op_loc *loc, u64 addr, - const char *var_name) +struct annotated_data_type *find_data_type(struct data_loc_info *dloc) { struct annotated_data_type *result = NULL; - struct dso *dso = map__dso(ms->map); - struct debuginfo *di; + struct dso *dso = map__dso(dloc->ms->map); Dwarf_Die type_die; - u64 pc; - di = debuginfo__new(dso->long_name); - if (di == NULL) { + dloc->di = debuginfo__new(dso->long_name); + if (dloc->di == NULL) { pr_debug("cannot get the debug info\n"); return NULL; } /* - * IP is a relative instruction address from the start of the map, as - * it can be randomized/relocated, it needs to translate to PC which is - * a file address for DWARF processing. + * The type offset is the same as instruction offset by default. + * But when finding a global variable, the offset won't be valid. */ - pc = map__rip_2objdump(ms->map, ip); - if (find_data_type_die(di, pc, addr, var_name, loc, &type_die) < 0) + if (dloc->var_name == NULL) + dloc->type_offset = dloc->op->offset; + + dloc->fbreg = -1; + + if (find_data_type_die(dloc, &type_die) < 0) goto out; result = dso__findnew_data_type(dso, &type_die); out: - debuginfo__delete(di); + debuginfo__delete(dloc->di); return result; } diff --git a/tools/perf/util/annotate-data.h b/tools/perf/util/annotate-data.h index 1b0db8e8c40e..ad6493ea2c8e 100644 --- a/tools/perf/util/annotate-data.h +++ b/tools/perf/util/annotate-data.h @@ -8,6 +8,7 @@ #include struct annotated_op_loc; +struct debuginfo; struct evsel; struct map_symbol; @@ -72,6 +73,35 @@ struct annotated_data_type { extern struct annotated_data_type unknown_type; extern struct annotated_data_type stackop_type; +/** + * struct data_loc_info - Data location information + * @ms: Map and Symbol info + * @ip: Instruction address + * @var_addr: Data address (for global variables) + * @var_name: Variable name (for global variables) + * @op: Instruction operand location (regs and offset) + * @di: Debug info + * @fbreg: Frame base register + * @fb_cfa: Whether the frame needs to check CFA + * @type_offset: Final offset in the type + */ +struct data_loc_info { + /* These are input field, should be filled by caller */ + struct map_symbol *ms; + u64 ip; + u64 var_addr; + const char *var_name; + struct annotated_op_loc *op; + + /* These are used internally */ + struct debuginfo *di; + int fbreg; + bool fb_cfa; + + /* This is for the result */ + int type_offset; +}; + /** * struct annotated_data_stat - Debug statistics * @total: Total number of entry @@ -106,9 +136,7 @@ extern struct annotated_data_stat ann_data_stat; #ifdef HAVE_DWARF_SUPPORT /* Returns data type at the location (ip, reg, offset) */ -struct annotated_data_type *find_data_type(struct map_symbol *ms, u64 ip, - struct annotated_op_loc *loc, u64 addr, - const char *var_name); +struct annotated_data_type *find_data_type(struct data_loc_info *dloc); /* Update type access histogram at the given offset */ int annotated_data_type__update_samples(struct annotated_data_type *adt, @@ -121,9 +149,7 @@ void annotated_data_type__tree_delete(struct rb_root *root); #else /* HAVE_DWARF_SUPPORT */ static inline struct annotated_data_type * -find_data_type(struct map_symbol *ms __maybe_unused, u64 ip __maybe_unused, - struct annotated_op_loc *loc __maybe_unused, - u64 addr __maybe_unused, const char *var_name __maybe_unused) +find_data_type(struct data_loc_info *dloc __maybe_unused) { return NULL; } diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index ac002d907d81..a15ff6758210 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -3816,9 +3816,7 @@ struct annotated_data_type *hist_entry__get_data_type(struct hist_entry *he) struct annotated_op_loc *op_loc; struct annotated_data_type *mem_type; struct annotated_item_stat *istat; - u64 ip = he->ip, addr = 0; - const char *var_name = NULL; - int var_offset; + u64 ip = he->ip; int i; ann_data_stat.total++; @@ -3871,51 +3869,53 @@ retry: } for_each_insn_op_loc(&loc, i, op_loc) { + struct data_loc_info dloc = { + .ms = ms, + /* Recalculate IP for LOCK prefix or insn fusion */ + .ip = ms->sym->start + dl->al.offset, + .op = op_loc, + }; + if (!op_loc->mem_ref) continue; /* Recalculate IP because of LOCK prefix or insn fusion */ ip = ms->sym->start + dl->al.offset; - var_offset = op_loc->offset; - /* PC-relative addressing */ if (op_loc->reg1 == DWARF_REG_PC) { struct addr_location al; struct symbol *var; u64 map_addr; - addr = annotate_calc_pcrel(ms, ip, op_loc->offset, dl); + dloc.var_addr = annotate_calc_pcrel(ms, ip, op_loc->offset, dl); /* Kernel symbols might be relocated */ - map_addr = addr + map__reloc(ms->map); + map_addr = dloc.var_addr + map__reloc(ms->map); addr_location__init(&al); var = thread__find_symbol_fb(he->thread, he->cpumode, map_addr, &al); if (var) { - var_name = var->name; + dloc.var_name = var->name; /* Calculate type offset from the start of variable */ - var_offset = map_addr - map__unmap_ip(al.map, var->start); + dloc.type_offset = map_addr - map__unmap_ip(al.map, var->start); } addr_location__exit(&al); } - mem_type = find_data_type(ms, ip, op_loc, addr, var_name); + mem_type = find_data_type(&dloc); if (mem_type) istat->good++; else istat->bad++; - if (mem_type && var_name) - op_loc->offset = var_offset; - if (symbol_conf.annotate_data_sample) { annotated_data_type__update_samples(mem_type, evsel, - op_loc->offset, + dloc.type_offset, he->stat.nr_events, he->stat.period); } - he->mem_type_off = op_loc->offset; + he->mem_type_off = dloc.type_offset; return mem_type; } -- cgit v1.2.3 From 5cdd3fd7995a7a07b1654ea37e0fcc3c64f0cc44 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:50:59 -0700 Subject: perf annotate: Add annotate_get_basic_blocks() The annotate_get_basic_blocks() is to find a list of basic blocks from the source instruction to the destination instruction in a function. It'll be used to find variables in a scope. Use BFS (Breadth First Search) to find a shortest path to carry the variable/register state minimally. Also change find_disasm_line() to be used in annotate_get_basic_blocks() and add 'allow_update' argument to control if it can update the IP. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-8-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 222 ++++++++++++++++++++++++++++++++++++++++++++- tools/perf/util/annotate.h | 16 ++++ 2 files changed, 235 insertions(+), 3 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index a15ff6758210..aa005c13ff67 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -3714,7 +3714,8 @@ static void symbol__ensure_annotate(struct map_symbol *ms, struct evsel *evsel) } } -static struct disasm_line *find_disasm_line(struct symbol *sym, u64 ip) +static struct disasm_line *find_disasm_line(struct symbol *sym, u64 ip, + bool allow_update) { struct disasm_line *dl; struct annotation *notes; @@ -3727,7 +3728,8 @@ static struct disasm_line *find_disasm_line(struct symbol *sym, u64 ip) * llvm-objdump places "lock" in a separate line and * in that case, we want to get the next line. */ - if (!strcmp(dl->ins.name, "lock") && *dl->ops.raw == '\0') { + if (!strcmp(dl->ins.name, "lock") && + *dl->ops.raw == '\0' && allow_update) { ip++; continue; } @@ -3843,7 +3845,7 @@ struct annotated_data_type *hist_entry__get_data_type(struct hist_entry *he) * Get a disasm to extract the location from the insn. * This is too slow... */ - dl = find_disasm_line(ms->sym, ip); + dl = find_disasm_line(ms->sym, ip, /*allow_update=*/true); if (dl == NULL) { ann_data_stat.no_insn++; return NULL; @@ -3937,3 +3939,217 @@ retry: istat->bad++; return NULL; } + +/* Basic block traversal (BFS) data structure */ +struct basic_block_data { + struct list_head queue; + struct list_head visited; +}; + +/* + * During the traversal, it needs to know the parent block where the current + * block block started from. Note that single basic block can be parent of + * two child basic blocks (in case of condition jump). + */ +struct basic_block_link { + struct list_head node; + struct basic_block_link *parent; + struct annotated_basic_block *bb; +}; + +/* Check any of basic block in the list already has the offset */ +static bool basic_block_has_offset(struct list_head *head, s64 offset) +{ + struct basic_block_link *link; + + list_for_each_entry(link, head, node) { + s64 begin_offset = link->bb->begin->al.offset; + s64 end_offset = link->bb->end->al.offset; + + if (begin_offset <= offset && offset <= end_offset) + return true; + } + return false; +} + +static bool is_new_basic_block(struct basic_block_data *bb_data, + struct disasm_line *dl) +{ + s64 offset = dl->al.offset; + + if (basic_block_has_offset(&bb_data->visited, offset)) + return false; + if (basic_block_has_offset(&bb_data->queue, offset)) + return false; + return true; +} + +/* Add a basic block starting from dl and link it to the parent */ +static int add_basic_block(struct basic_block_data *bb_data, + struct basic_block_link *parent, + struct disasm_line *dl) +{ + struct annotated_basic_block *bb; + struct basic_block_link *link; + + if (dl == NULL) + return -1; + + if (!is_new_basic_block(bb_data, dl)) + return 0; + + bb = zalloc(sizeof(*bb)); + if (bb == NULL) + return -1; + + bb->begin = dl; + bb->end = dl; + INIT_LIST_HEAD(&bb->list); + + link = malloc(sizeof(*link)); + if (link == NULL) { + free(bb); + return -1; + } + + link->bb = bb; + link->parent = parent; + list_add_tail(&link->node, &bb_data->queue); + return 0; +} + +/* Returns true when it finds the target in the current basic block */ +static bool process_basic_block(struct basic_block_data *bb_data, + struct basic_block_link *link, + struct symbol *sym, u64 target) +{ + struct disasm_line *dl, *next_dl, *last_dl; + struct annotation *notes = symbol__annotation(sym); + bool found = false; + + dl = link->bb->begin; + /* Check if it's already visited */ + if (basic_block_has_offset(&bb_data->visited, dl->al.offset)) + return false; + + last_dl = list_last_entry(¬es->src->source, + struct disasm_line, al.node); + + list_for_each_entry_from(dl, ¬es->src->source, al.node) { + /* Found the target instruction */ + if (sym->start + dl->al.offset == target) { + found = true; + break; + } + /* End of the function, finish the block */ + if (dl == last_dl) + break; + /* 'return' instruction finishes the block */ + if (dl->ins.ops == &ret_ops) + break; + /* normal instructions are part of the basic block */ + if (dl->ins.ops != &jump_ops) + continue; + /* jump to a different function, tail call or return */ + if (dl->ops.target.outside) + break; + /* jump instruction creates new basic block(s) */ + next_dl = find_disasm_line(sym, sym->start + dl->ops.target.offset, + /*allow_update=*/false); + add_basic_block(bb_data, link, next_dl); + + /* + * FIXME: determine conditional jumps properly. + * Conditional jumps create another basic block with the + * next disasm line. + */ + if (!strstr(dl->ins.name, "jmp")) { + next_dl = list_next_entry(dl, al.node); + add_basic_block(bb_data, link, next_dl); + } + break; + + } + link->bb->end = dl; + return found; +} + +/* + * It founds a target basic block, build a proper linked list of basic blocks + * by following the link recursively. + */ +static void link_found_basic_blocks(struct basic_block_link *link, + struct list_head *head) +{ + while (link) { + struct basic_block_link *parent = link->parent; + + list_move(&link->bb->list, head); + list_del(&link->node); + free(link); + + link = parent; + } +} + +static void delete_basic_blocks(struct basic_block_data *bb_data) +{ + struct basic_block_link *link, *tmp; + + list_for_each_entry_safe(link, tmp, &bb_data->queue, node) { + list_del(&link->node); + free(link->bb); + free(link); + } + + list_for_each_entry_safe(link, tmp, &bb_data->visited, node) { + list_del(&link->node); + free(link->bb); + free(link); + } +} + +/** + * annotate_get_basic_blocks - Get basic blocks for given address range + * @sym: symbol to annotate + * @src: source address + * @dst: destination address + * @head: list head to save basic blocks + * + * This function traverses disasm_lines from @src to @dst and save them in a + * list of annotated_basic_block to @head. It uses BFS to find the shortest + * path between two. The basic_block_link is to maintain parent links so + * that it can build a list of blocks from the start. + */ +int annotate_get_basic_blocks(struct symbol *sym, s64 src, s64 dst, + struct list_head *head) +{ + struct basic_block_data bb_data = { + .queue = LIST_HEAD_INIT(bb_data.queue), + .visited = LIST_HEAD_INIT(bb_data.visited), + }; + struct basic_block_link *link; + struct disasm_line *dl; + int ret = -1; + + dl = find_disasm_line(sym, src, /*allow_update=*/false); + if (dl == NULL) + return -1; + + if (add_basic_block(&bb_data, /*parent=*/NULL, dl) < 0) + return -1; + + /* Find shortest path from src to dst using BFS */ + while (!list_empty(&bb_data.queue)) { + link = list_first_entry(&bb_data.queue, struct basic_block_link, node); + + if (process_basic_block(&bb_data, link, sym, dst)) { + link_found_basic_blocks(link, head); + ret = 0; + break; + } + list_move(&link->node, &bb_data.visited); + } + delete_basic_blocks(&bb_data); + return ret; +} diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h index 13cc659e508c..0928663fddee 100644 --- a/tools/perf/util/annotate.h +++ b/tools/perf/util/annotate.h @@ -561,4 +561,20 @@ extern struct list_head ann_insn_stat; u64 annotate_calc_pcrel(struct map_symbol *ms, u64 ip, int offset, struct disasm_line *dl); +/** + * struct annotated_basic_block - Basic block of instructions + * @list: List node + * @begin: start instruction in the block + * @end: end instruction in the block + */ +struct annotated_basic_block { + struct list_head list; + struct disasm_line *begin; + struct disasm_line *end; +}; + +/* Get a list of basic blocks from src to dst addresses */ +int annotate_get_basic_blocks(struct symbol *sym, s64 src, s64 dst, + struct list_head *head); + #endif /* __PERF_ANNOTATE_H */ -- cgit v1.2.3 From 90429524f3e7342e98fd3a14deaa46f1dd3f6ffe Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:00 -0700 Subject: perf annotate-data: Add debug messages Add a new debug option "type-profile" to enable the detailed info during the type analysis especially for instruction tracking. You can use this before the command name like 'report' or 'annotate'. $ perf --debug type-profile annotate --data-type Committer testing: First get some memory events: $ perf mem record ls Then, without data-type profiling debug: $ perf annotate --data-type | head Annotate type: 'struct rtld_global' in /usr/lib64/ld-linux-x86-64.so.2 (1 samples): ============================================================================ samples offset size field 1 0 4336 struct rtld_global { 0 0 0 struct link_namespaces* _dl_ns; 0 2560 8 size_t _dl_nns; 0 2568 40 __rtld_lock_recursive_t _dl_load_lock { 0 2568 40 pthread_mutex_t mutex { 0 2568 40 struct __pthread_mutex_s __data { 0 2568 4 int __lock; $ And with only data-type profiling: $ perf --debug type-profile annotate --data-type | head ----------------------------------------------------------- find_data_type_die [1e67] for reg13873052 (PC) offset=0x150e2 in dl_main CU die offset: 0x29cd3 found PC-rel by addr=0x34020 offset=0x20 ----------------------------------------------------------- find_data_type_die [2e] for reg12 offset=0 in __GI___readdir64 CU die offset: 0x137a45 frame base: cfa=1 fbreg=-1 found "__futex" in scope=2/2 (die: 0x137ad5) 0(reg12) type=int (die:2a) ----------------------------------------------------------- find_data_type_die [52] for reg5 offset=0 in __memmove_avx_unaligned_erms CU die offset: 0x1124ed no variable found Annotate type: 'struct rtld_global' in /usr/lib64/ld-linux-x86-64.so.2 (1 samples): ============================================================================ samples offset size field 1 0 4336 struct rtld_global { 0 0 0 struct link_namespaces* _dl_ns; 0 2560 8 size_t _dl_nns; 0 2568 40 __rtld_lock_recursive_t _dl_load_lock { 0 2568 40 pthread_mutex_t mutex { 0 2568 40 struct __pthread_mutex_s __data { 0 2568 4 int __lock; $ Signed-off-by: Namhyung Kim Tested-by: Arnaldo Carvalho de Melo Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-9-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 74 +++++++++++++++++++++++++++++++++++++---- tools/perf/util/debug.c | 3 ++ tools/perf/util/debug.h | 1 + 3 files changed, 71 insertions(+), 7 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index ff81d164aa57..f482ccfdaa91 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -23,6 +23,29 @@ #include "symbol.h" #include "symbol_conf.h" +#define pr_debug_dtp(fmt, ...) \ +do { \ + if (debug_type_profile) \ + pr_info(fmt, ##__VA_ARGS__); \ + else \ + pr_debug3(fmt, ##__VA_ARGS__); \ +} while (0) + +static void pr_debug_type_name(Dwarf_Die *die) +{ + struct strbuf sb; + char *str; + + if (!debug_type_profile && verbose < 3) + return; + + strbuf_init(&sb, 32); + die_get_typename_from_type(die, &sb); + str = strbuf_detach(&sb, NULL); + pr_info(" type=%s (die:%lx)\n", str, (long)dwarf_dieoffset(die)); + free(str); +} + /* * Compare type name and size to maintain them in a tree. * I'm not sure if DWARF would have information of a single type in many @@ -201,7 +224,7 @@ static int check_variable(Dwarf_Die *var_die, Dwarf_Die *type_die, int offset, /* Get the type of the variable */ if (die_get_real_type(var_die, type_die) == NULL) { - pr_debug("variable has no type\n"); + pr_debug_dtp("variable has no type\n"); ann_data_stat.no_typeinfo++; return -1; } @@ -215,7 +238,7 @@ static int check_variable(Dwarf_Die *var_die, Dwarf_Die *type_die, int offset, if ((dwarf_tag(type_die) != DW_TAG_pointer_type && dwarf_tag(type_die) != DW_TAG_array_type) || die_get_real_type(type_die, type_die) == NULL) { - pr_debug("no pointer or no type\n"); + pr_debug_dtp("no pointer or no type\n"); ann_data_stat.no_typeinfo++; return -1; } @@ -223,14 +246,15 @@ static int check_variable(Dwarf_Die *var_die, Dwarf_Die *type_die, int offset, /* Get the size of the actual type */ if (dwarf_aggregate_size(type_die, &size) < 0) { - pr_debug("type size is unknown\n"); + pr_debug_dtp("type size is unknown\n"); ann_data_stat.invalid_size++; return -1; } /* Minimal sanity check */ if ((unsigned)offset >= size) { - pr_debug("offset: %d is bigger than size: %" PRIu64 "\n", offset, size); + pr_debug_dtp("offset: %d is bigger than size: %"PRIu64"\n", + offset, size); ann_data_stat.bad_offset++; return -1; } @@ -251,6 +275,19 @@ static int find_data_type_die(struct data_loc_info *dloc, Dwarf_Die *type_die) int fb_offset = 0; bool is_fbreg = false; u64 pc; + char buf[64]; + + if (dloc->op->multi_regs) + snprintf(buf, sizeof(buf), " or reg%d", dloc->op->reg2); + else if (dloc->op->reg1 == DWARF_REG_PC) + snprintf(buf, sizeof(buf), " (PC)"); + else + buf[0] = '\0'; + + pr_debug_dtp("-----------------------------------------------------------\n"); + pr_debug_dtp("%s [%"PRIx64"] for reg%d%s offset=%#x in %s\n", + __func__, dloc->ip - dloc->ms->sym->start, + dloc->op->reg1, buf, dloc->op->offset, dloc->ms->sym->name); /* * IP is a relative instruction address from the start of the map, as @@ -261,7 +298,7 @@ static int find_data_type_die(struct data_loc_info *dloc, Dwarf_Die *type_die) /* Get a compile_unit for this address */ if (!find_cu_die(dloc->di, pc, &cu_die)) { - pr_debug("cannot find CU for address %" PRIx64 "\n", pc); + pr_debug_dtp("cannot find CU for address %"PRIx64"\n", pc); ann_data_stat.no_cuinfo++; return -1; } @@ -269,12 +306,17 @@ static int find_data_type_die(struct data_loc_info *dloc, Dwarf_Die *type_die) reg = loc->reg1; offset = loc->offset; + pr_debug_dtp("CU die offset: %#lx\n", (long)dwarf_dieoffset(&cu_die)); + if (reg == DWARF_REG_PC) { if (die_find_variable_by_addr(&cu_die, dloc->var_addr, &var_die, &offset)) { ret = check_variable(&var_die, type_die, offset, /*is_pointer=*/false); dloc->type_offset = offset; + + pr_debug_dtp("found PC-rel by addr=%#"PRIx64" offset=%#x\n", + dloc->var_addr, offset); goto out; } @@ -310,6 +352,9 @@ static int find_data_type_die(struct data_loc_info *dloc, Dwarf_Die *type_die) default: break; } + + pr_debug_dtp("frame base: cfa=%d fbreg=%d\n", + dloc->fb_cfa, fbreg); } } @@ -334,6 +379,19 @@ retry: /* Found a variable, see if it's correct */ ret = check_variable(&var_die, type_die, offset, reg != DWARF_REG_PC && !is_fbreg); + if (ret == 0) { + pr_debug_dtp("found \"%s\" in scope=%d/%d (die: %#lx) ", + dwarf_diename(&var_die), i+1, nr_scopes, + (long)dwarf_dieoffset(&scopes[i])); + if (reg == DWARF_REG_PC) + pr_debug_dtp("%#x(PC) offset=%#x", loc->offset, offset); + else if (reg == DWARF_REG_FB || is_fbreg) + pr_debug_dtp("%#x(reg%d) stack fb_offset=%#x offset=%#x", + loc->offset, reg, fb_offset, offset); + else + pr_debug_dtp("%#x(reg%d)", loc->offset, reg); + pr_debug_type_name(type_die); + } dloc->type_offset = offset; goto out; } @@ -343,8 +401,10 @@ retry: goto retry; } - if (ret < 0) + if (ret < 0) { + pr_debug_dtp("no variable found\n"); ann_data_stat.no_var++; + } out: free(scopes); @@ -373,7 +433,7 @@ struct annotated_data_type *find_data_type(struct data_loc_info *dloc) dloc->di = debuginfo__new(dso->long_name); if (dloc->di == NULL) { - pr_debug("cannot get the debug info\n"); + pr_debug_dtp("cannot get the debug info\n"); return NULL; } diff --git a/tools/perf/util/debug.c b/tools/perf/util/debug.c index c39ee0fcb8cf..d633d15329fa 100644 --- a/tools/perf/util/debug.c +++ b/tools/perf/util/debug.c @@ -41,6 +41,7 @@ static int redirect_to_stderr; int debug_data_convert; static FILE *_debug_file; bool debug_display_time; +int debug_type_profile; FILE *debug_file(void) { @@ -231,6 +232,7 @@ static struct sublevel_option debug_opts[] = { { .name = "data-convert", .value_ptr = &debug_data_convert }, { .name = "perf-event-open", .value_ptr = &debug_peo_args }, { .name = "kmaps", .value_ptr = &debug_kmaps }, + { .name = "type-profile", .value_ptr = &debug_type_profile }, { .name = NULL, } }; @@ -270,6 +272,7 @@ int perf_quiet_option(void) redirect_to_stderr = 0; debug_peo_args = 0; debug_kmaps = 0; + debug_type_profile = 0; return 0; } diff --git a/tools/perf/util/debug.h b/tools/perf/util/debug.h index 35a7a5ae762e..a4026d1fd6a3 100644 --- a/tools/perf/util/debug.h +++ b/tools/perf/util/debug.h @@ -14,6 +14,7 @@ extern int debug_peo_args; extern bool quiet, dump_trace; extern int debug_ordered_events; extern int debug_data_convert; +extern int debug_type_profile; #ifndef pr_fmt #define pr_fmt(fmt) fmt -- cgit v1.2.3 From 06b2ce75386df04b7aea53818ed3c42ee50ec426 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:01 -0700 Subject: perf annotate-data: Maintain variable type info As it collected basic block and variable information in each scope, it now can build a state table to find matching variable at the location. The struct type_state is to keep the type info saved in each register and stack slot. The update_var_state() updates the table when it finds variables in the current address. It expects die_collect_vars() filled a list of variables with type info and starting address. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-10-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 173 ++++++++++++++++++++++++++++++++++++++++ tools/perf/util/dwarf-aux.c | 4 + 2 files changed, 177 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index f482ccfdaa91..8eaa06f1cee5 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -46,6 +46,62 @@ static void pr_debug_type_name(Dwarf_Die *die) free(str); } +/* Type information in a register, valid when ok is true */ +struct type_state_reg { + Dwarf_Die type; + bool ok; +}; + +/* Type information in a stack location, dynamically allocated */ +struct type_state_stack { + struct list_head list; + Dwarf_Die type; + int offset; + int size; + bool compound; +}; + +/* FIXME: This should be arch-dependent */ +#define TYPE_STATE_MAX_REGS 16 + +/* + * State table to maintain type info in each register and stack location. + * It'll be updated when new variable is allocated or type info is moved + * to a new location (register or stack). As it'd be used with the + * shortest path of basic blocks, it only maintains a single table. + */ +struct type_state { + struct type_state_reg regs[TYPE_STATE_MAX_REGS]; + struct list_head stack_vars; +}; + +static bool has_reg_type(struct type_state *state, int reg) +{ + return (unsigned)reg < ARRAY_SIZE(state->regs); +} + +/* These declarations will be remove once they are changed to static */ +void init_type_state(struct type_state *state, struct arch *arch __maybe_unused); +void exit_type_state(struct type_state *state); +void update_var_state(struct type_state *state, struct data_loc_info *dloc, + u64 addr, u64 insn_offset, struct die_var_type *var_types); + +void init_type_state(struct type_state *state, struct arch *arch __maybe_unused) +{ + memset(state, 0, sizeof(*state)); + INIT_LIST_HEAD(&state->stack_vars); +} + +void exit_type_state(struct type_state *state) +{ + struct type_state_stack *stack, *tmp; + + list_for_each_entry_safe(stack, tmp, &state->stack_vars, list) { + list_del(&stack->list); + free(stack); + } +} + /* * Compare type name and size to maintain them in a tree. * I'm not sure if DWARF would have information of a single type in many @@ -262,6 +318,123 @@ static int check_variable(Dwarf_Die *var_die, Dwarf_Die *type_die, int offset, return 0; } +static struct type_state_stack *find_stack_state(struct type_state *state, + int offset) +{ + struct type_state_stack *stack; + + list_for_each_entry(stack, &state->stack_vars, list) { + if (offset == stack->offset) + return stack; + + if (stack->compound && stack->offset < offset && + offset < stack->offset + stack->size) + return stack; + } + return NULL; +} + +static void set_stack_state(struct type_state_stack *stack, int offset, + Dwarf_Die *type_die) +{ + int tag; + Dwarf_Word size; + + if (dwarf_aggregate_size(type_die, &size) < 0) + size = 0; + + tag = dwarf_tag(type_die); + + stack->type = *type_die; + stack->size = size; + stack->offset = offset; + + switch (tag) { + case DW_TAG_structure_type: + case DW_TAG_union_type: + stack->compound = true; + break; + default: + stack->compound = false; + break; + } +} + +static struct type_state_stack *findnew_stack_state(struct type_state *state, + int offset, Dwarf_Die *type_die) +{ + struct type_state_stack *stack = find_stack_state(state, offset); + + if (stack) { + set_stack_state(stack, offset, type_die); + return stack; + } + + stack = malloc(sizeof(*stack)); + if (stack) { + set_stack_state(stack, offset, type_die); + list_add(&stack->list, &state->stack_vars); + } + return stack; +} + +/** + * update_var_state - Update type state using given variables + * @state: type state table + * @dloc: data location info + * @addr: instruction address to match with variable + * @insn_offset: instruction offset (for debug) + * @var_types: list of variables with type info + * + * This function fills the @state table using @var_types info. Each variable + * is used only at the given location and updates an entry in the table. + */ +void update_var_state(struct type_state *state, struct data_loc_info *dloc, + u64 addr, u64 insn_offset, struct die_var_type *var_types) +{ + Dwarf_Die mem_die; + struct die_var_type *var; + int fbreg = dloc->fbreg; + int fb_offset = 0; + + if (dloc->fb_cfa) { + if (die_get_cfa(dloc->di->dbg, addr, &fbreg, &fb_offset) < 0) + fbreg = -1; + } + + for (var = var_types; var != NULL; var = var->next) { + if (var->addr != addr) + continue; + /* Get the type DIE using the offset */ + if (!dwarf_offdie(dloc->di->dbg, var->die_off, &mem_die)) + continue; + + if (var->reg == DWARF_REG_FB) { + findnew_stack_state(state, var->offset, &mem_die); + + pr_debug_dtp("var [%"PRIx64"] -%#x(stack) type=", + insn_offset, -var->offset); + pr_debug_type_name(&mem_die); + } else if (var->reg == fbreg) { + findnew_stack_state(state, var->offset - fb_offset, &mem_die); + + pr_debug_dtp("var [%"PRIx64"] -%#x(stack) type=", + insn_offset, -var->offset + fb_offset); + pr_debug_type_name(&mem_die); + } else if (has_reg_type(state, var->reg) && var->offset == 0) { + struct type_state_reg *reg; + + reg = &state->regs[var->reg]; + reg->type = mem_die; + reg->ok = true; + + pr_debug_dtp("var [%"PRIx64"] reg%d type=", + insn_offset, var->reg); + pr_debug_type_name(&mem_die); + } + } +} + /* The result will be saved in @type_die */ static int find_data_type_die(struct data_loc_info *dloc, Dwarf_Die *type_die) { diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c index d4ec239c5adb..7dad99ee3ff3 100644 --- a/tools/perf/util/dwarf-aux.c +++ b/tools/perf/util/dwarf-aux.c @@ -9,6 +9,7 @@ #include #include "debug.h" #include "dwarf-aux.h" +#include "dwarf-regs.h" #include "strbuf.h" #include "string2.h" @@ -1190,6 +1191,8 @@ static int reg_from_dwarf_op(Dwarf_Op *op) case DW_OP_regx: case DW_OP_bregx: return op->number; + case DW_OP_fbreg: + return DWARF_REG_FB; default: break; } @@ -1203,6 +1206,7 @@ static int offset_from_dwarf_op(Dwarf_Op *op) case DW_OP_regx: return 0; case DW_OP_breg0 ... DW_OP_breg31: + case DW_OP_fbreg: return op->number; case DW_OP_bregx: return op->number2; -- cgit v1.2.3 From 4f903455befa257c50422a4570c4dca0020a1fc8 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:02 -0700 Subject: perf annotate-data: Add update_insn_state() The update_insn_state() function is to update the type state table after processing each instruction. For now, it handles MOV (on x86) insn to transfer type info from the source location to the target. The location can be a register or a stack slot. Check carefully when memory reference happens and fetch the type correctly. It basically ignores write to a memory since it doesn't change the type info. One exception is writes to (new) stack slots for register spilling. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-11-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 161 +++++++++++++++++++++++++++++++++++++++- tools/perf/util/annotate-data.h | 2 + tools/perf/util/annotate.c | 1 + 3 files changed, 161 insertions(+), 3 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 8eaa06f1cee5..592437b6c097 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -71,7 +71,9 @@ struct type_state_stack { * shortest path of basic blocks, it only maintains a single table. */ struct type_state { + /* state of general purpose registers */ struct type_state_reg regs[TYPE_STATE_MAX_REGS]; + /* state of stack location */ struct list_head stack_vars; }; @@ -85,6 +87,8 @@ void init_type_state(struct type_state *state, struct arch *arch __maybe_unused) void exit_type_state(struct type_state *state); void update_var_state(struct type_state *state, struct data_loc_info *dloc, u64 addr, u64 insn_offset, struct die_var_type *var_types); +void update_insn_state(struct type_state *state, struct data_loc_info *dloc, + struct disasm_line *dl); void init_type_state(struct type_state *state, struct arch *arch __maybe_unused) { @@ -412,13 +416,13 @@ void update_var_state(struct type_state *state, struct data_loc_info *dloc, if (var->reg == DWARF_REG_FB) { findnew_stack_state(state, var->offset, &mem_die); - pr_debug_dtp("var [%"PRIx64"] -%#x(stack) type=", + pr_debug_dtp("var [%"PRIx64"] -%#x(stack)", insn_offset, -var->offset); pr_debug_type_name(&mem_die); } else if (var->reg == fbreg) { findnew_stack_state(state, var->offset - fb_offset, &mem_die); - pr_debug_dtp("var [%"PRIx64"] -%#x(stack) type=", + pr_debug_dtp("var [%"PRIx64"] -%#x(stack)", insn_offset, -var->offset + fb_offset); pr_debug_type_name(&mem_die); } else if (has_reg_type(state, var->reg) && var->offset == 0) { @@ -428,13 +432,164 @@ void update_var_state(struct type_state *state, struct data_loc_info *dloc, reg->type = mem_die; reg->ok = true; - pr_debug_dtp("var [%"PRIx64"] reg%d type=", + pr_debug_dtp("var [%"PRIx64"] reg%d", insn_offset, var->reg); pr_debug_type_name(&mem_die); } } } +static void update_insn_state_x86(struct type_state *state, + struct data_loc_info *dloc, + struct disasm_line *dl) +{ + struct annotated_insn_loc loc; + struct annotated_op_loc *src = &loc.ops[INSN_OP_SOURCE]; + struct annotated_op_loc *dst = &loc.ops[INSN_OP_TARGET]; + struct type_state_reg *tsr; + Dwarf_Die type_die; + u32 insn_offset = dl->al.offset; + int fbreg = dloc->fbreg; + int fboff = 0; + + if (annotate_get_insn_location(dloc->arch, dl, &loc) < 0) + return; + + if (strncmp(dl->ins.name, "mov", 3)) + return; + + if (dloc->fb_cfa) { + u64 ip = dloc->ms->sym->start + dl->al.offset; + u64 pc = map__rip_2objdump(dloc->ms->map, ip); + + if (die_get_cfa(dloc->di->dbg, pc, &fbreg, &fboff) < 0) + fbreg = -1; + } + + /* Case 1. register to register transfers */ + if (!src->mem_ref && !dst->mem_ref) { + if (!has_reg_type(state, dst->reg1)) + return; + + tsr = &state->regs[dst->reg1]; + if (!has_reg_type(state, src->reg1) || + !state->regs[src->reg1].ok) { + tsr->ok = false; + return; + } + + tsr->type = state->regs[src->reg1].type; + tsr->ok = true; + + pr_debug_dtp("mov [%x] reg%d -> reg%d", + insn_offset, src->reg1, dst->reg1); + pr_debug_type_name(&tsr->type); + } + /* Case 2. memory to register transers */ + if (src->mem_ref && !dst->mem_ref) { + int sreg = src->reg1; + + if (!has_reg_type(state, dst->reg1)) + return; + + tsr = &state->regs[dst->reg1]; + +retry: + /* Check stack variables with offset */ + if (sreg == fbreg) { + struct type_state_stack *stack; + int offset = src->offset - fboff; + + stack = find_stack_state(state, offset); + if (stack == NULL) { + tsr->ok = false; + return; + } else if (!stack->compound) { + tsr->type = stack->type; + tsr->ok = true; + } else if (die_get_member_type(&stack->type, + offset - stack->offset, + &type_die)) { + tsr->type = type_die; + tsr->ok = true; + } else { + tsr->ok = false; + return; + } + + pr_debug_dtp("mov [%x] -%#x(stack) -> reg%d", + insn_offset, -offset, dst->reg1); + pr_debug_type_name(&tsr->type); + } + /* And then dereference the pointer if it has one */ + else if (has_reg_type(state, sreg) && state->regs[sreg].ok && + die_deref_ptr_type(&state->regs[sreg].type, + src->offset, &type_die)) { + tsr->type = type_die; + tsr->ok = true; + + pr_debug_dtp("mov [%x] %#x(reg%d) -> reg%d", + insn_offset, src->offset, sreg, dst->reg1); + pr_debug_type_name(&tsr->type); + } + /* Or try another register if any */ + else if (src->multi_regs && sreg == src->reg1 && + src->reg1 != src->reg2) { + sreg = src->reg2; + goto retry; + } + /* It failed to get a type info, mark it as invalid */ + else { + tsr->ok = false; + } + } + /* Case 3. register to memory transfers */ + if (!src->mem_ref && dst->mem_ref) { + if (!has_reg_type(state, src->reg1) || + !state->regs[src->reg1].ok) + return; + + /* Check stack variables with offset */ + if (dst->reg1 == fbreg) { + struct type_state_stack *stack; + int offset = dst->offset - fboff; + + stack = find_stack_state(state, offset); + if (stack) { + /* + * The source register is likely to hold a type + * of member if it's a compound type. Do not + * update the stack variable type since we can + * get the member type later by using the + * die_get_member_type(). + */ + if (!stack->compound) + set_stack_state(stack, offset, + &state->regs[src->reg1].type); + } else { + findnew_stack_state(state, offset, + &state->regs[src->reg1].type); + } + + pr_debug_dtp("mov [%x] reg%d -> -%#x(stack)", + insn_offset, src->reg1, -offset); + pr_debug_type_name(&state->regs[src->reg1].type); + } + /* + * Ignore other transfers since it'd set a value in a struct + * and won't change the type. + */ + } + /* Case 4. memory to memory transfers (not handled for now) */ +} + +void update_insn_state(struct type_state *state, struct data_loc_info *dloc, + struct disasm_line *dl) +{ + if (arch__is(dloc->arch, "x86")) + update_insn_state_x86(state, dloc, dl); +} + /* The result will be saved in @type_die */ static int find_data_type_die(struct data_loc_info *dloc, Dwarf_Die *type_die) { diff --git a/tools/perf/util/annotate-data.h b/tools/perf/util/annotate-data.h index ad6493ea2c8e..7324cafe2c7b 100644 --- a/tools/perf/util/annotate-data.h +++ b/tools/perf/util/annotate-data.h @@ -75,6 +75,7 @@ extern struct annotated_data_type stackop_type; /** * struct data_loc_info - Data location information + * @arch: CPU architecture info * @ms: Map and Symbol info * @ip: Instruction address * @var_addr: Data address (for global variables) @@ -87,6 +88,7 @@ extern struct annotated_data_type stackop_type; */ struct data_loc_info { /* These are input field, should be filled by caller */ + struct arch *arch; struct map_symbol *ms; u64 ip; u64 var_addr; diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index aa005c13ff67..9777df5dc2e3 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -3872,6 +3872,7 @@ retry: for_each_insn_op_loc(&loc, i, op_loc) { struct data_loc_info dloc = { + .arch = arch, .ms = ms, /* Recalculate IP for LOCK prefix or insn fusion */ .ip = ms->sym->start + dl->al.offset, -- cgit v1.2.3 From 1ebb5e17ef21b492ee60654d4e22cbfb3763661f Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:03 -0700 Subject: perf annotate-data: Add get_global_var_type() Accessing global variable is common when it tracks execution later. Factor out the common code into a function for later use. It adds thread and cpumode to struct data_loc_info to find (global) symbols if needed. Also remove var_name as it's retrieved in the helper function. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-12-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 62 +++++++++++++++++++++++++++++++---------- tools/perf/util/annotate-data.h | 7 +++-- tools/perf/util/annotate.c | 21 +++----------- 3 files changed, 57 insertions(+), 33 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 592437b6c097..3b661e693410 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -22,6 +22,7 @@ #include "strbuf.h" #include "symbol.h" #include "symbol_conf.h" +#include "thread.h" #define pr_debug_dtp(fmt, ...) \ do { \ @@ -382,6 +383,50 @@ static struct type_state_stack *findnew_stack_state(struct type_state *state, return stack; } +static bool get_global_var_type(Dwarf_Die *cu_die, struct data_loc_info *dloc, + u64 ip, u64 var_addr, int *var_offset, + Dwarf_Die *type_die) +{ + u64 pc, mem_addr; + int offset; + bool is_pointer = false; + const char *var_name = NULL; + Dwarf_Die var_die; + struct addr_location al; + struct symbol *sym; + + /* Try to get the variable by address first */ + if (die_find_variable_by_addr(cu_die, var_addr, &var_die, &offset) && + check_variable(&var_die, type_die, offset, is_pointer) == 0) { + *var_offset = offset; + return true; + } + + /* Kernel symbols might be relocated */ + mem_addr = var_addr + map__reloc(dloc->ms->map); + + addr_location__init(&al); + sym = thread__find_symbol_fb(dloc->thread, dloc->cpumode, + mem_addr, &al); + if (sym) { + var_name = sym->name; + /* Calculate type offset from the start of variable */ + *var_offset = mem_addr - map__unmap_ip(al.map, sym->start); + } + addr_location__exit(&al); + if (var_name == NULL) + return false; + + pc = map__rip_2objdump(dloc->ms->map, ip); + + /* Try to get the name of global variable */ + if (die_find_variable_at(cu_die, var_name, pc, &var_die) && + check_variable(&var_die, type_die, *var_offset, is_pointer) == 0) + return true; + + return false; +} + /** * update_var_state - Update type state using given variables * @state: type state table @@ -637,24 +682,14 @@ static int find_data_type_die(struct data_loc_info *dloc, Dwarf_Die *type_die) pr_debug_dtp("CU die offset: %#lx\n", (long)dwarf_dieoffset(&cu_die)); if (reg == DWARF_REG_PC) { - if (die_find_variable_by_addr(&cu_die, dloc->var_addr, &var_die, - &offset)) { - ret = check_variable(&var_die, type_die, offset, - /*is_pointer=*/false); + if (get_global_var_type(&cu_die, dloc, dloc->ip, dloc->var_addr, + &offset, type_die)) { dloc->type_offset = offset; pr_debug_dtp("found PC-rel by addr=%#"PRIx64" offset=%#x\n", dloc->var_addr, offset); goto out; } - - if (dloc->var_name && - die_find_variable_at(&cu_die, dloc->var_name, pc, &var_die)) { - ret = check_variable(&var_die, type_die, dloc->type_offset, - /*is_pointer=*/false); - /* dloc->type_offset was updated by the caller */ - goto out; - } } /* Get a list of nested scopes - i.e. (inlined) functions and blocks. */ @@ -769,8 +804,7 @@ struct annotated_data_type *find_data_type(struct data_loc_info *dloc) * The type offset is the same as instruction offset by default. * But when finding a global variable, the offset won't be valid. */ - if (dloc->var_name == NULL) - dloc->type_offset = dloc->op->offset; + dloc->type_offset = dloc->op->offset; dloc->fbreg = -1; diff --git a/tools/perf/util/annotate-data.h b/tools/perf/util/annotate-data.h index 7324cafe2c7b..acfbd1748d02 100644 --- a/tools/perf/util/annotate-data.h +++ b/tools/perf/util/annotate-data.h @@ -11,6 +11,7 @@ struct annotated_op_loc; struct debuginfo; struct evsel; struct map_symbol; +struct thread; /** * struct annotated_member - Type of member field @@ -76,10 +77,11 @@ extern struct annotated_data_type stackop_type; /** * struct data_loc_info - Data location information * @arch: CPU architecture info + * @thread: Thread info * @ms: Map and Symbol info * @ip: Instruction address * @var_addr: Data address (for global variables) - * @var_name: Variable name (for global variables) + * @cpumode: CPU execution mode * @op: Instruction operand location (regs and offset) * @di: Debug info * @fbreg: Frame base register @@ -89,10 +91,11 @@ extern struct annotated_data_type stackop_type; struct data_loc_info { /* These are input field, should be filled by caller */ struct arch *arch; + struct thread *thread; struct map_symbol *ms; u64 ip; u64 var_addr; - const char *var_name; + u8 cpumode; struct annotated_op_loc *op; /* These are used internally */ diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 9777df5dc2e3..abb641aa8ec0 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -3873,9 +3873,11 @@ retry: for_each_insn_op_loc(&loc, i, op_loc) { struct data_loc_info dloc = { .arch = arch, + .thread = he->thread, .ms = ms, /* Recalculate IP for LOCK prefix or insn fusion */ .ip = ms->sym->start + dl->al.offset, + .cpumode = he->cpumode, .op = op_loc, }; @@ -3887,23 +3889,8 @@ retry: /* PC-relative addressing */ if (op_loc->reg1 == DWARF_REG_PC) { - struct addr_location al; - struct symbol *var; - u64 map_addr; - - dloc.var_addr = annotate_calc_pcrel(ms, ip, op_loc->offset, dl); - /* Kernel symbols might be relocated */ - map_addr = dloc.var_addr + map__reloc(ms->map); - - addr_location__init(&al); - var = thread__find_symbol_fb(he->thread, he->cpumode, - map_addr, &al); - if (var) { - dloc.var_name = var->name; - /* Calculate type offset from the start of variable */ - dloc.type_offset = map_addr - map__unmap_ip(al.map, var->start); - } - addr_location__exit(&al); + dloc.var_addr = annotate_calc_pcrel(ms, dloc.ip, + op_loc->offset, dl); } mem_type = find_data_type(&dloc); -- cgit v1.2.3 From 0a41e5d6849b4f705c377d34799485b546764970 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:04 -0700 Subject: perf annotate-data: Handle global variable access When updating the instruction states, it also needs to handle global variable accesses. Same as it does for PC-relative addressing, it can look up the type by address (if it's defined in the same file), or by name after finding the symbol by address (for declarations). Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-13-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 46 +++++++++++++++++++++++++++++++++++++---- 1 file changed, 42 insertions(+), 4 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 3b661e693410..2cc9f56e3eea 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -89,7 +89,7 @@ void exit_type_state(struct type_state *state); void update_var_state(struct type_state *state, struct data_loc_info *dloc, u64 addr, u64 insn_offset, struct die_var_type *var_types); void update_insn_state(struct type_state *state, struct data_loc_info *dloc, - struct disasm_line *dl); + Dwarf_Die *cu_die, struct disasm_line *dl); void init_type_state(struct type_state *state, struct arch *arch __maybe_unused) { @@ -485,7 +485,7 @@ void update_var_state(struct type_state *state, struct data_loc_info *dloc, } static void update_insn_state_x86(struct type_state *state, - struct data_loc_info *dloc, + struct data_loc_info *dloc, Dwarf_Die *cu_die, struct disasm_line *dl) { struct annotated_insn_loc loc; @@ -577,6 +577,29 @@ retry: insn_offset, src->offset, sreg, dst->reg1); pr_debug_type_name(&tsr->type); } + /* Or check if it's a global variable */ + else if (sreg == DWARF_REG_PC) { + struct map_symbol *ms = dloc->ms; + u64 ip = ms->sym->start + dl->al.offset; + u64 addr; + int offset; + + addr = annotate_calc_pcrel(ms, ip, src->offset, dl); + + if (!get_global_var_type(cu_die, dloc, ip, addr, &offset, + &type_die) || + !die_get_member_type(&type_die, offset, &type_die)) { + tsr->ok = false; + return; + } + + tsr->type = type_die; + tsr->ok = true; + + pr_debug_dtp("mov [%x] global addr=%"PRIx64" -> reg%d", + insn_offset, addr, dst->reg1); + pr_debug_type_name(&type_die); + } /* Or try another register if any */ else if (src->multi_regs && sreg == src->reg1 && src->reg1 != src->reg2) { @@ -628,11 +651,26 @@ retry: /* Case 4. memory to memory transfers (not handled for now) */ } +/** + * update_insn_state - Update type state for an instruction + * @state: type state table + * @dloc: data location info + * @cu_die: compile unit debug entry + * @dl: disasm line for the instruction + * + * This function updates the @state table for the target operand of the + * instruction at @dl if it transfers the type like MOV on x86. Since it + * tracks the type, it won't care about the values like in arithmetic + * instructions like ADD/SUB/MUL/DIV and INC/DEC. + * + * Note that ops->reg2 is only available when both mem_ref and multi_regs + * are true. + */ void update_insn_state(struct type_state *state, struct data_loc_info *dloc, - struct disasm_line *dl) + Dwarf_Die *cu_die, struct disasm_line *dl) { if (arch__is(dloc->arch, "x86")) - update_insn_state_x86(state, dloc, dl); + update_insn_state_x86(state, dloc, cu_die, dl); } /* The result will be saved in @type_die */ -- cgit v1.2.3 From cffb7910afbdf73ca5b2fdf8a617584a7dcecf14 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:05 -0700 Subject: perf annotate-data: Handle call instructions When updating instruction states, the call instruction should play a role since it changes the register states. For simplicity, mark some registers as caller-saved registers (should be arch-dependent), and invalidate them all after a function call. If the function returns something, the designated register (ret_reg) will have the type info. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-14-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 54 +++++++++++++++++++++++++++++++++++++++-- 1 file changed, 52 insertions(+), 2 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 2cc9f56e3eea..6bcf22e523cb 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -47,10 +47,14 @@ static void pr_debug_type_name(Dwarf_Die *die) free(str); } -/* Type information in a register, valid when ok is true */ +/* + * Type information in a register, valid when @ok is true. + * The @caller_saved registers are invalidated after a function call. + */ struct type_state_reg { Dwarf_Die type; bool ok; + bool caller_saved; }; /* Type information in a stack location, dynamically allocated */ @@ -76,6 +80,8 @@ struct type_state { struct type_state_reg regs[TYPE_STATE_MAX_REGS]; /* state of stack location */ struct list_head stack_vars; + /* return value register */ + int ret_reg; }; static bool has_reg_type(struct type_state *state, int reg) @@ -91,10 +97,23 @@ void update_var_state(struct type_state *state, struct data_loc_info *dloc, void update_insn_state(struct type_state *state, struct data_loc_info *dloc, Dwarf_Die *cu_die, struct disasm_line *dl); -void init_type_state(struct type_state *state, struct arch *arch __maybe_unused) +void init_type_state(struct type_state *state, struct arch *arch) { memset(state, 0, sizeof(*state)); INIT_LIST_HEAD(&state->stack_vars); + + if (arch__is(arch, "x86")) { + state->regs[0].caller_saved = true; + state->regs[1].caller_saved = true; + state->regs[2].caller_saved = true; + state->regs[4].caller_saved = true; + state->regs[5].caller_saved = true; + state->regs[8].caller_saved = true; + state->regs[9].caller_saved = true; + state->regs[10].caller_saved = true; + state->regs[11].caller_saved = true; + state->ret_reg = 0; + } } void exit_type_state(struct type_state *state) @@ -500,6 +519,37 @@ static void update_insn_state_x86(struct type_state *state, if (annotate_get_insn_location(dloc->arch, dl, &loc) < 0) return; + if (ins__is_call(&dl->ins)) { + struct symbol *func = dl->ops.target.sym; + + if (func == NULL) + return; + + /* __fentry__ will preserve all registers */ + if (!strcmp(func->name, "__fentry__")) + return; + + pr_debug_dtp("call [%x] %s\n", insn_offset, func->name); + + /* Otherwise invalidate caller-saved registers after call */ + for (unsigned i = 0; i < ARRAY_SIZE(state->regs); i++) { + if (state->regs[i].caller_saved) + state->regs[i].ok = false; + } + + /* Update register with the return type (if any) */ + if (die_find_func_rettype(cu_die, func->name, &type_die)) { + tsr = &state->regs[state->ret_reg]; + tsr->type = type_die; + tsr->ok = true; + + pr_debug_dtp("call [%x] return -> reg%d", + insn_offset, state->ret_reg); + pr_debug_type_name(&type_die); + } + return; + } + if (strncmp(dl->ins.name, "mov", 3)) return; -- cgit v1.2.3 From eb8a55e01de9a671e130b4c0e5b483997f4f491f Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:06 -0700 Subject: perf annotate-data: Implement instruction tracking If it failed to find a variable for the location directly, it might be due to a missing variable in the source code. For example, accessing pointer variables in a chain can result in the case like below: struct foo *foo = ...; int i = foo->bar->baz; The DWARF debug information is created for each variable so it'd have one for 'foo'. But there's no variable for 'foo->bar' and then it cannot know the type of 'bar' and 'baz'. The above source code can be compiled to the follow x86 instructions: mov 0x8(%rax), %rcx mov 0x4(%rcx), %rdx <=== PMU sample mov %rdx, -4(%rbp) Let's say 'foo' is located in the %rax and it has a pointer to struct foo. But perf sample is captured in the second instruction and there is no variable or type info for the %rcx. It'd be great if compiler could generate debug info for %rcx, but we should handle it on our side. So this patch implements the logic to iterate instructions and update the type table for each location. As it already collected a list of scopes including the target instruction, we can use it to construct the type table smartly. +---------------- scope[0] subprogram | | +-------------- scope[1] lexical_block | | | | +------------ scope[2] inlined_subroutine | | | | | | +---------- scope[3] inlined_subroutine | | | | | | | | +-------- scope[4] lexical_block | | | | | | | | | | *** target instruction ... Image the target instruction has 5 scopes, each scope will have its own variables and parameters. Then it can start with the innermost scope (4). So it'd search the shortest path from the start of scope[4] to the target address and build a list of basic blocks. Then it iterates the basic blocks with the variables in the scope and update the table. If it finds a type at the target instruction, then returns it. Otherwise, it moves to the upper scope[3]. Now it'd search the shortest path from the start of scope[3] to the start of scope[4]. Then connect it to the existing basic block list. Then it'd iterate the blocks with variables for both scopes. It can repeat this until it finds a type at the target instruction or reaches to the top scope[0]. As the basic blocks contain the shortest path, it won't worry about branches and can update the table simply. The final check will be done by find_matching_type() in the next patch. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-15-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-annotate.c | 1 + tools/perf/util/annotate-data.c | 223 +++++++++++++++++++++++++++++++++++++--- tools/perf/util/annotate-data.h | 1 + 3 files changed, 211 insertions(+), 14 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 6c1cc797692d..f677671409b1 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -430,6 +430,7 @@ static void print_annotate_data_stat(struct annotated_data_stat *s) PRINT_STAT(no_typeinfo); PRINT_STAT(invalid_size); PRINT_STAT(bad_offset); + PRINT_STAT(insn_track); printf("\n"); #undef PRINT_STAT diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 6bcf22e523cb..13ba65693367 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -89,15 +89,7 @@ static bool has_reg_type(struct type_state *state, int reg) return (unsigned)reg < ARRAY_SIZE(state->regs); } -/* These declarations will be remove once they are changed to static */ -void init_type_state(struct type_state *state, struct arch *arch __maybe_unused); -void exit_type_state(struct type_state *state); -void update_var_state(struct type_state *state, struct data_loc_info *dloc, - u64 addr, u64 insn_offset, struct die_var_type *var_types); -void update_insn_state(struct type_state *state, struct data_loc_info *dloc, - Dwarf_Die *cu_die, struct disasm_line *dl); - -void init_type_state(struct type_state *state, struct arch *arch) +static void init_type_state(struct type_state *state, struct arch *arch) { memset(state, 0, sizeof(*state)); INIT_LIST_HEAD(&state->stack_vars); @@ -116,7 +108,7 @@ void init_type_state(struct type_state *state, struct arch *arch) } } -void exit_type_state(struct type_state *state) +static void exit_type_state(struct type_state *state) { struct type_state_stack *stack, *tmp; @@ -457,8 +449,8 @@ static bool get_global_var_type(Dwarf_Die *cu_die, struct data_loc_info *dloc, * This function fills the @state table using @var_types info. Each variable * is used only at the given location and updates an entry in the table. */ -void update_var_state(struct type_state *state, struct data_loc_info *dloc, - u64 addr, u64 insn_offset, struct die_var_type *var_types) +static void update_var_state(struct type_state *state, struct data_loc_info *dloc, + u64 addr, u64 insn_offset, struct die_var_type *var_types) { Dwarf_Die mem_die; struct die_var_type *var; @@ -716,13 +708,207 @@ retry: * Note that ops->reg2 is only available when both mem_ref and multi_regs * are true. */ -void update_insn_state(struct type_state *state, struct data_loc_info *dloc, - Dwarf_Die *cu_die, struct disasm_line *dl) +static void update_insn_state(struct type_state *state, struct data_loc_info *dloc, + Dwarf_Die *cu_die, struct disasm_line *dl) { if (arch__is(dloc->arch, "x86")) update_insn_state_x86(state, dloc, cu_die, dl); } +/* + * Prepend this_blocks (from the outer scope) to full_blocks, removing + * duplicate disasm line. + */ +static void prepend_basic_blocks(struct list_head *this_blocks, + struct list_head *full_blocks) +{ + struct annotated_basic_block *first_bb, *last_bb; + + last_bb = list_last_entry(this_blocks, typeof(*last_bb), list); + first_bb = list_first_entry(full_blocks, typeof(*first_bb), list); + + if (list_empty(full_blocks)) + goto out; + + /* Last insn in this_blocks should be same as first insn in full_blocks */ + if (last_bb->end != first_bb->begin) { + pr_debug("prepend basic blocks: mismatched disasm line %"PRIx64" -> %"PRIx64"\n", + last_bb->end->al.offset, first_bb->begin->al.offset); + goto out; + } + + /* Is the basic block have only one disasm_line? */ + if (last_bb->begin == last_bb->end) { + list_del(&last_bb->list); + free(last_bb); + goto out; + } + + /* Point to the insn before the last when adding this block to full_blocks */ + last_bb->end = list_prev_entry(last_bb->end, al.node); + +out: + list_splice(this_blocks, full_blocks); +} + +static void delete_basic_blocks(struct list_head *basic_blocks) +{ + struct annotated_basic_block *bb, *tmp; + + list_for_each_entry_safe(bb, tmp, basic_blocks, list) { + list_del(&bb->list); + free(bb); + } +} + +/* Make sure all variables have a valid start address */ +static void fixup_var_address(struct die_var_type *var_types, u64 addr) +{ + while (var_types) { + /* + * Some variables have no address range meaning it's always + * available in the whole scope. Let's adjust the start + * address to the start of the scope. + */ + if (var_types->addr == 0) + var_types->addr = addr; + + var_types = var_types->next; + } +} + +static void delete_var_types(struct die_var_type *var_types) +{ + while (var_types) { + struct die_var_type *next = var_types->next; + + free(var_types); + var_types = next; + } +} + +/* It's at the target address, check if it has a matching type */ +static bool find_matching_type(struct type_state *state __maybe_unused, + struct data_loc_info *dloc __maybe_unused, + int reg __maybe_unused, + Dwarf_Die *type_die __maybe_unused) +{ + /* TODO */ + return false; +} + +/* Iterate instructions in basic blocks and update type table */ +static bool find_data_type_insn(struct data_loc_info *dloc, int reg, + struct list_head *basic_blocks, + struct die_var_type *var_types, + Dwarf_Die *cu_die, Dwarf_Die *type_die) +{ + struct type_state state; + struct symbol *sym = dloc->ms->sym; + struct annotation *notes = symbol__annotation(sym); + struct annotated_basic_block *bb; + bool found = false; + + init_type_state(&state, dloc->arch); + + list_for_each_entry(bb, basic_blocks, list) { + struct disasm_line *dl = bb->begin; + + pr_debug_dtp("bb: [%"PRIx64" - %"PRIx64"]\n", + bb->begin->al.offset, bb->end->al.offset); + + list_for_each_entry_from(dl, ¬es->src->source, al.node) { + u64 this_ip = sym->start + dl->al.offset; + u64 addr = map__rip_2objdump(dloc->ms->map, this_ip); + + /* Update variable type at this address */ + update_var_state(&state, dloc, addr, dl->al.offset, var_types); + + if (this_ip == dloc->ip) { + found = find_matching_type(&state, dloc, reg, + type_die); + goto out; + } + + /* Update type table after processing the instruction */ + update_insn_state(&state, dloc, cu_die, dl); + if (dl == bb->end) + break; + } + } + +out: + exit_type_state(&state); + return found; +} + +/* + * Construct a list of basic blocks for each scope with variables and try to find + * the data type by updating a type state table through instructions. + */ +static int find_data_type_block(struct data_loc_info *dloc, int reg, + Dwarf_Die *cu_die, Dwarf_Die *scopes, + int nr_scopes, Dwarf_Die *type_die) +{ + LIST_HEAD(basic_blocks); + struct die_var_type *var_types = NULL; + u64 src_ip, dst_ip, prev_dst_ip; + int ret = -1; + + /* TODO: other architecture support */ + if (!arch__is(dloc->arch, "x86")) + return -1; + + prev_dst_ip = dst_ip = dloc->ip; + for (int i = nr_scopes - 1; i >= 0; i--) { + Dwarf_Addr base, start, end; + LIST_HEAD(this_blocks); + + if (dwarf_ranges(&scopes[i], 0, &base, &start, &end) < 0) + break; + + pr_debug_dtp("scope: [%d/%d] (die:%lx)\n", + i + 1, nr_scopes, (long)dwarf_dieoffset(&scopes[i])); + src_ip = map__objdump_2rip(dloc->ms->map, start); + +again: + /* Get basic blocks for this scope */ + if (annotate_get_basic_blocks(dloc->ms->sym, src_ip, dst_ip, + &this_blocks) < 0) { + /* Try previous block if they are not connected */ + if (prev_dst_ip != dst_ip) { + dst_ip = prev_dst_ip; + goto again; + } + + pr_debug_dtp("cannot find a basic block from %"PRIx64" to %"PRIx64"\n", + src_ip - dloc->ms->sym->start, + dst_ip - dloc->ms->sym->start); + continue; + } + prepend_basic_blocks(&this_blocks, &basic_blocks); + + /* Get variable info for this scope and add to var_types list */ + die_collect_vars(&scopes[i], &var_types); + fixup_var_address(var_types, start); + + /* Find from start of this scope to the target instruction */ + if (find_data_type_insn(dloc, reg, &basic_blocks, var_types, + cu_die, type_die)) { + ret = 0; + break; + } + + /* Go up to the next scope and find blocks to the start */ + prev_dst_ip = dst_ip; + dst_ip = src_ip; + } + + delete_basic_blocks(&basic_blocks); + delete_var_types(var_types); + return ret; +} + /* The result will be saved in @type_die */ static int find_data_type_die(struct data_loc_info *dloc, Dwarf_Die *type_die) { @@ -847,6 +1033,15 @@ retry: goto out; } + if (reg != DWARF_REG_PC) { + ret = find_data_type_block(dloc, reg, &cu_die, scopes, + nr_scopes, type_die); + if (ret == 0) { + ann_data_stat.insn_track++; + goto out; + } + } + if (loc->multi_regs && reg == loc->reg1 && loc->reg1 != loc->reg2) { reg = loc->reg2; goto retry; diff --git a/tools/perf/util/annotate-data.h b/tools/perf/util/annotate-data.h index acfbd1748d02..ae0f87aed804 100644 --- a/tools/perf/util/annotate-data.h +++ b/tools/perf/util/annotate-data.h @@ -135,6 +135,7 @@ struct annotated_data_stat { int no_typeinfo; int invalid_size; int bad_offset; + int insn_track; }; extern struct annotated_data_stat ann_data_stat; -- cgit v1.2.3 From bdc80ace07106a62b51d1752869df29dbd65ad2c Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:07 -0700 Subject: perf annotate-data: Check register state for type As instruction tracking updates the type state for each register, check the final type info for the target register at the given instruction. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-16-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 88 +++++++++++++++++++++++++++++++++++++---- 1 file changed, 81 insertions(+), 7 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 13ba65693367..f5329a78a97d 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -788,12 +788,83 @@ static void delete_var_types(struct die_var_type *var_types) } /* It's at the target address, check if it has a matching type */ -static bool find_matching_type(struct type_state *state __maybe_unused, - struct data_loc_info *dloc __maybe_unused, - int reg __maybe_unused, - Dwarf_Die *type_die __maybe_unused) +static bool check_matching_type(struct type_state *state, + struct data_loc_info *dloc, int reg, + Dwarf_Die *type_die) { - /* TODO */ + Dwarf_Word size; + u32 insn_offset = dloc->ip - dloc->ms->sym->start; + + pr_debug_dtp("chk [%x] reg%d offset=%#x ok=%d", + insn_offset, reg, dloc->op->offset, state->regs[reg].ok); + + if (state->regs[reg].ok) { + int tag = dwarf_tag(&state->regs[reg].type); + + pr_debug_dtp("\n"); + + /* + * Normal registers should hold a pointer (or array) to + * dereference a memory location. + */ + if (tag != DW_TAG_pointer_type && tag != DW_TAG_array_type) + return false; + + /* Remove the pointer and get the target type */ + if (die_get_real_type(&state->regs[reg].type, type_die) == NULL) + return false; + + dloc->type_offset = dloc->op->offset; + + /* Get the size of the actual type */ + if (dwarf_aggregate_size(type_die, &size) < 0 || + (unsigned)dloc->type_offset >= size) + return false; + + return true; + } + + if (reg == dloc->fbreg) { + struct type_state_stack *stack; + + pr_debug_dtp(" fbreg\n"); + + stack = find_stack_state(state, dloc->type_offset); + if (stack == NULL) + return false; + + *type_die = stack->type; + /* Update the type offset from the start of slot */ + dloc->type_offset -= stack->offset; + + return true; + } + + if (dloc->fb_cfa) { + struct type_state_stack *stack; + u64 pc = map__rip_2objdump(dloc->ms->map, dloc->ip); + int fbreg, fboff; + + pr_debug_dtp(" cfa\n"); + + if (die_get_cfa(dloc->di->dbg, pc, &fbreg, &fboff) < 0) + fbreg = -1; + + if (reg != fbreg) + return false; + + stack = find_stack_state(state, dloc->type_offset - fboff); + if (stack == NULL) + return false; + + *type_die = stack->type; + /* Update the type offset from the start of slot */ + dloc->type_offset -= fboff + stack->offset; + + return true; + } + + pr_debug_dtp("\n"); return false; } @@ -825,8 +896,8 @@ static bool find_data_type_insn(struct data_loc_info *dloc, int reg, update_var_state(&state, dloc, addr, dl->al.offset, var_types); if (this_ip == dloc->ip) { - found = find_matching_type(&state, dloc, reg, - type_die); + found = check_matching_type(&state, dloc, reg, + type_die); goto out; } @@ -896,6 +967,9 @@ again: if (find_data_type_insn(dloc, reg, &basic_blocks, var_types, cu_die, type_die)) { ret = 0; + pr_debug_dtp("found by insn track: %#x(reg%d) type-offset=%#x", + dloc->op->offset, reg, dloc->type_offset); + pr_debug_type_name(type_die); break; } -- cgit v1.2.3 From cbaf89a8c5b467f99dd930f24a698f3fa338b012 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:08 -0700 Subject: perf annotate: Parse x86 segment register location Add a segment field in the struct annotated_insn_loc and save it for the segment based addressing like %gs:0x28. For simplicity it now handles %gs register only. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-17-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 36 ++++++++++++++++++++++++++++++++---- tools/perf/util/annotate.h | 15 +++++++++++++++ 2 files changed, 47 insertions(+), 4 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index abb641aa8ec0..3aa3a3b987ad 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -94,6 +94,7 @@ struct arch { char skip_functions_char; char register_char; char memory_ref_char; + char imm_char; } objdump; }; @@ -211,6 +212,7 @@ static struct arch architectures[] = { .comment_char = '#', .register_char = '%', .memory_ref_char = '(', + .imm_char = '$', }, }, { @@ -3585,6 +3587,12 @@ static int extract_reg_offset(struct arch *arch, const char *str, * %gs:0x18(%rbx). In that case it should skip the part. */ if (*str == arch->objdump.register_char) { + if (arch__is(arch, "x86")) { + /* FIXME: Handle other segment registers */ + if (!strncmp(str, "%gs:", 4)) + op_loc->segment = INSN_SEG_X86_GS; + } + while (*str && !isdigit(*str) && *str != arch->objdump.memory_ref_char) str++; @@ -3681,12 +3689,32 @@ int annotate_get_insn_location(struct arch *arch, struct disasm_line *dl, op_loc->multi_regs = multi_regs; extract_reg_offset(arch, insn_str, op_loc); } else { - char *s = strdup(insn_str); + char *s, *p = NULL; + + if (arch__is(arch, "x86")) { + /* FIXME: Handle other segment registers */ + if (!strncmp(insn_str, "%gs:", 4)) { + op_loc->segment = INSN_SEG_X86_GS; + op_loc->offset = strtol(insn_str + 4, + &p, 0); + if (p && p != insn_str + 4) + op_loc->imm = true; + continue; + } + } + + s = strdup(insn_str); + if (s == NULL) + return -1; - if (s) { + if (*s == arch->objdump.register_char) op_loc->reg1 = get_dwarf_regnum(s, 0); - free(s); + else if (*s == arch->objdump.imm_char) { + op_loc->offset = strtol(s + 1, &p, 0); + if (p && p != s + 1) + op_loc->imm = true; } + free(s); } } @@ -3881,7 +3909,7 @@ retry: .op = op_loc, }; - if (!op_loc->mem_ref) + if (!op_loc->mem_ref && op_loc->segment == INSN_SEG_NONE) continue; /* Recalculate IP because of LOCK prefix or insn fusion */ diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h index 0928663fddee..14980b65f812 100644 --- a/tools/perf/util/annotate.h +++ b/tools/perf/util/annotate.h @@ -511,15 +511,19 @@ int annotate_check_args(void); * @reg1: First register in the operand * @reg2: Second register in the operand * @offset: Memory access offset in the operand + * @segment: Segment selector register * @mem_ref: Whether the operand accesses memory * @multi_regs: Whether the second register is used + * @imm: Whether the operand is an immediate value (in offset) */ struct annotated_op_loc { int reg1; int reg2; int offset; + u8 segment; bool mem_ref; bool multi_regs; + bool imm; }; enum annotated_insn_ops { @@ -529,6 +533,17 @@ enum annotated_insn_ops { INSN_OP_MAX, }; +enum annotated_x86_segment { + INSN_SEG_NONE = 0, + + INSN_SEG_X86_CS, + INSN_SEG_X86_DS, + INSN_SEG_X86_ES, + INSN_SEG_X86_FS, + INSN_SEG_X86_GS, + INSN_SEG_X86_SS, +}; + /** * struct annotated_insn_loc - Location info of instruction * @ops: Array of location info for source and target operands -- cgit v1.2.3 From 02e17ca917423c622da10ac6bd0924c17462962e Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:09 -0700 Subject: perf annotate-data: Handle this-cpu variables in kernel On x86, the kernel gets the current task using the current macro like below: #define current get_current() static __always_inline struct task_struct *get_current(void) { return this_cpu_read_stable(pcpu_hot.current_task); } So it returns the current_task field of struct pcpu_hot which is the first member. On my build, it's located at 0x32940. $ nm vmlinux | grep pcpu_hot 0000000000032940 D pcpu_hot And the current macro generates the instructions like below: mov %gs:0x32940, %rcx So the %gs segment register points to the beginning of the per-cpu region of this cpu and it points the variable with a constant. Let's update the instruction location info to have a segment register and handle %gs in kernel to look up a global variable. Pretend it as a global variable by changing the register number to DWARF_REG_PC. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-18-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 21 +++++++++++++++++++-- tools/perf/util/annotate.c | 7 +++++++ 2 files changed, 26 insertions(+), 2 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index f5329a78a97d..d57622ddd5d3 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -790,7 +790,7 @@ static void delete_var_types(struct die_var_type *var_types) /* It's at the target address, check if it has a matching type */ static bool check_matching_type(struct type_state *state, struct data_loc_info *dloc, int reg, - Dwarf_Die *type_die) + Dwarf_Die *cu_die, Dwarf_Die *type_die) { Dwarf_Word size; u32 insn_offset = dloc->ip - dloc->ms->sym->start; @@ -864,6 +864,23 @@ static bool check_matching_type(struct type_state *state, return true; } + if (map__dso(dloc->ms->map)->kernel && arch__is(dloc->arch, "x86")) { + u64 addr; + int offset; + + if (dloc->op->segment == INSN_SEG_X86_GS && dloc->op->imm) { + pr_debug_dtp(" this-cpu var\n"); + + addr = dloc->op->offset; + + if (get_global_var_type(cu_die, dloc, dloc->ip, addr, + &offset, type_die)) { + dloc->type_offset = offset; + return true; + } + } + } + pr_debug_dtp("\n"); return false; } @@ -897,7 +914,7 @@ static bool find_data_type_insn(struct data_loc_info *dloc, int reg, if (this_ip == dloc->ip) { found = check_matching_type(&state, dloc, reg, - type_die); + cu_die, type_die); goto out; } diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 3aa3a3b987ad..e4121acb4f88 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -3921,6 +3921,13 @@ retry: op_loc->offset, dl); } + /* This CPU access in kernel - pretend PC-relative addressing */ + if (map__dso(ms->map)->kernel && arch__is(arch, "x86") && + op_loc->segment == INSN_SEG_X86_GS && op_loc->imm) { + dloc.var_addr = op_loc->offset; + op_loc->reg1 = DWARF_REG_PC; + } + mem_type = find_data_type(&dloc); if (mem_type) istat->good++; -- cgit v1.2.3 From ad62edbfc55b23347539c8c6fff9a70de17c4b95 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:10 -0700 Subject: perf annotate-data: Track instructions with a this-cpu variable Like global variables, this per-cpu variables should be tracked correctly. Factor our get_global_var_type() to handle both global and per-cpu (for this cpu) variables in the same manner. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-19-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 31 ++++++++++++++++++++++++++++++- 1 file changed, 30 insertions(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index d57622ddd5d3..48fea0c716ef 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -553,12 +553,41 @@ static void update_insn_state_x86(struct type_state *state, fbreg = -1; } - /* Case 1. register to register transfers */ + /* Case 1. register to register or segment:offset to register transfers */ if (!src->mem_ref && !dst->mem_ref) { if (!has_reg_type(state, dst->reg1)) return; tsr = &state->regs[dst->reg1]; + if (map__dso(dloc->ms->map)->kernel && + src->segment == INSN_SEG_X86_GS && src->imm) { + u64 ip = dloc->ms->sym->start + dl->al.offset; + u64 var_addr; + int offset; + + /* + * In kernel, %gs points to a per-cpu region for the + * current CPU. Access with a constant offset should + * be treated as a global variable access. + */ + var_addr = src->offset; + + if (!get_global_var_type(cu_die, dloc, ip, var_addr, + &offset, &type_die) || + !die_get_member_type(&type_die, offset, &type_die)) { + tsr->ok = false; + return; + } + + tsr->type = type_die; + tsr->ok = true; + + pr_debug_dtp("mov [%x] this-cpu addr=%#"PRIx64" -> reg%d", + insn_offset, var_addr, dst->reg1); + pr_debug_type_name(&tsr->type); + return; + } + if (!has_reg_type(state, src->reg1) || !state->regs[src->reg1].ok) { tsr->ok = false; -- cgit v1.2.3 From f5b095924d0c1c2f0698df07c54887d92c39fd3a Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:11 -0700 Subject: perf annotate-data: Support general per-cpu access This is to support per-cpu variable access often without a matching DWARF entry. For some reason, I cannot find debug info of per-cpu variables sometimes. They have more complex pattern to calculate the address of per-cpu variables like below. 2b7d: mov -0x1e0(%rbp),%rax ; rax = cpu 2b84: mov -0x7da0f7e0(,%rax,8),%rcx ; rcx = __per_cpu_offset[cpu] * 2b8c: mov 0x34870(%rcx),%rax ; *(__per_cpu_offset[cpu] + 0x34870) Let's assume the rax register has a number for a CPU at 2b7d. The next instruction is to get the per-cpu offset' for that cpu. The offset -0x7da0f7e0 is 0xffffffff825f0820 in u64 which is the address of the '__per_cpu_offset' array in my system. So it'd get the actual offset of that CPU's per-cpu region and save it to the rcx register. Then, at 2b8c, accesses using rcx can be handled same as the global variable access. To handle this case, it should check if the offset of the instruction matches to the address of '__per_cpu_offset'. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-20-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 213 +++++++++++++++++++++++++++++++--------- 1 file changed, 169 insertions(+), 44 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 48fea0c716ef..83b5aa00f01c 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -24,6 +24,12 @@ #include "symbol_conf.h" #include "thread.h" +enum type_state_kind { + TSR_KIND_INVALID = 0, + TSR_KIND_TYPE, + TSR_KIND_PERCPU_BASE, +}; + #define pr_debug_dtp(fmt, ...) \ do { \ if (debug_type_profile) \ @@ -32,7 +38,7 @@ do { \ pr_debug3(fmt, ##__VA_ARGS__); \ } while (0) -static void pr_debug_type_name(Dwarf_Die *die) +static void pr_debug_type_name(Dwarf_Die *die, enum type_state_kind kind) { struct strbuf sb; char *str; @@ -40,6 +46,18 @@ static void pr_debug_type_name(Dwarf_Die *die) if (!debug_type_profile && verbose < 3) return; + switch (kind) { + case TSR_KIND_INVALID: + pr_info("\n"); + return; + case TSR_KIND_PERCPU_BASE: + pr_info(" percpu base\n"); + return; + case TSR_KIND_TYPE: + default: + break; + } + strbuf_init(&sb, 32); die_get_typename_from_type(die, &sb); str = strbuf_detach(&sb, NULL); @@ -53,8 +71,10 @@ static void pr_debug_type_name(Dwarf_Die *die) */ struct type_state_reg { Dwarf_Die type; + u32 imm_value; bool ok; bool caller_saved; + u8 kind; }; /* Type information in a stack location, dynamically allocated */ @@ -64,6 +84,7 @@ struct type_state_stack { int offset; int size; bool compound; + u8 kind; }; /* FIXME: This should be arch-dependent */ @@ -82,6 +103,8 @@ struct type_state { struct list_head stack_vars; /* return value register */ int ret_reg; + /* stack pointer register */ + int stack_reg; }; static bool has_reg_type(struct type_state *state, int reg) @@ -105,6 +128,7 @@ static void init_type_state(struct type_state *state, struct arch *arch) state->regs[10].caller_saved = true; state->regs[11].caller_saved = true; state->ret_reg = 0; + state->stack_reg = 7; } } @@ -350,7 +374,7 @@ static struct type_state_stack *find_stack_state(struct type_state *state, return NULL; } -static void set_stack_state(struct type_state_stack *stack, int offset, +static void set_stack_state(struct type_state_stack *stack, int offset, u8 kind, Dwarf_Die *type_die) { int tag; @@ -364,6 +388,7 @@ static void set_stack_state(struct type_state_stack *stack, int offset, stack->type = *type_die; stack->size = size; stack->offset = offset; + stack->kind = kind; switch (tag) { case DW_TAG_structure_type: @@ -377,34 +402,60 @@ static void set_stack_state(struct type_state_stack *stack, int offset, } static struct type_state_stack *findnew_stack_state(struct type_state *state, - int offset, Dwarf_Die *type_die) + int offset, u8 kind, + Dwarf_Die *type_die) { struct type_state_stack *stack = find_stack_state(state, offset); if (stack) { - set_stack_state(stack, offset, type_die); + set_stack_state(stack, offset, kind, type_die); return stack; } stack = malloc(sizeof(*stack)); if (stack) { - set_stack_state(stack, offset, type_die); + set_stack_state(stack, offset, kind, type_die); list_add(&stack->list, &state->stack_vars); } return stack; } +static bool get_global_var_info(struct data_loc_info *dloc, u64 addr, + const char **var_name, int *var_offset) +{ + struct addr_location al; + struct symbol *sym; + u64 mem_addr; + + /* Kernel symbols might be relocated */ + mem_addr = addr + map__reloc(dloc->ms->map); + + addr_location__init(&al); + sym = thread__find_symbol_fb(dloc->thread, dloc->cpumode, + mem_addr, &al); + if (sym) { + *var_name = sym->name; + /* Calculate type offset from the start of variable */ + *var_offset = mem_addr - map__unmap_ip(al.map, sym->start); + } else { + *var_name = NULL; + } + addr_location__exit(&al); + if (*var_name == NULL) + return false; + + return true; +} + static bool get_global_var_type(Dwarf_Die *cu_die, struct data_loc_info *dloc, u64 ip, u64 var_addr, int *var_offset, Dwarf_Die *type_die) { - u64 pc, mem_addr; + u64 pc; int offset; bool is_pointer = false; - const char *var_name = NULL; + const char *var_name; Dwarf_Die var_die; - struct addr_location al; - struct symbol *sym; /* Try to get the variable by address first */ if (die_find_variable_by_addr(cu_die, var_addr, &var_die, &offset) && @@ -413,19 +464,7 @@ static bool get_global_var_type(Dwarf_Die *cu_die, struct data_loc_info *dloc, return true; } - /* Kernel symbols might be relocated */ - mem_addr = var_addr + map__reloc(dloc->ms->map); - - addr_location__init(&al); - sym = thread__find_symbol_fb(dloc->thread, dloc->cpumode, - mem_addr, &al); - if (sym) { - var_name = sym->name; - /* Calculate type offset from the start of variable */ - *var_offset = mem_addr - map__unmap_ip(al.map, sym->start); - } - addr_location__exit(&al); - if (var_name == NULL) + if (!get_global_var_info(dloc, var_addr, &var_name, var_offset)) return false; pc = map__rip_2objdump(dloc->ms->map, ip); @@ -470,27 +509,30 @@ static void update_var_state(struct type_state *state, struct data_loc_info *dlo continue; if (var->reg == DWARF_REG_FB) { - findnew_stack_state(state, var->offset, &mem_die); + findnew_stack_state(state, var->offset, TSR_KIND_TYPE, + &mem_die); pr_debug_dtp("var [%"PRIx64"] -%#x(stack)", insn_offset, -var->offset); - pr_debug_type_name(&mem_die); + pr_debug_type_name(&mem_die, TSR_KIND_TYPE); } else if (var->reg == fbreg) { - findnew_stack_state(state, var->offset - fb_offset, &mem_die); + findnew_stack_state(state, var->offset - fb_offset, + TSR_KIND_TYPE, &mem_die); pr_debug_dtp("var [%"PRIx64"] -%#x(stack)", insn_offset, -var->offset + fb_offset); - pr_debug_type_name(&mem_die); + pr_debug_type_name(&mem_die, TSR_KIND_TYPE); } else if (has_reg_type(state, var->reg) && var->offset == 0) { struct type_state_reg *reg; reg = &state->regs[var->reg]; reg->type = mem_die; + reg->kind = TSR_KIND_TYPE; reg->ok = true; pr_debug_dtp("var [%"PRIx64"] reg%d", insn_offset, var->reg); - pr_debug_type_name(&mem_die); + pr_debug_type_name(&mem_die, TSR_KIND_TYPE); } } } @@ -533,11 +575,12 @@ static void update_insn_state_x86(struct type_state *state, if (die_find_func_rettype(cu_die, func->name, &type_die)) { tsr = &state->regs[state->ret_reg]; tsr->type = type_die; + tsr->kind = TSR_KIND_TYPE; tsr->ok = true; pr_debug_dtp("call [%x] return -> reg%d", insn_offset, state->ret_reg); - pr_debug_type_name(&type_die); + pr_debug_type_name(&type_die, tsr->kind); } return; } @@ -580,11 +623,12 @@ static void update_insn_state_x86(struct type_state *state, } tsr->type = type_die; + tsr->kind = TSR_KIND_TYPE; tsr->ok = true; pr_debug_dtp("mov [%x] this-cpu addr=%#"PRIx64" -> reg%d", insn_offset, var_addr, dst->reg1); - pr_debug_type_name(&tsr->type); + pr_debug_type_name(&tsr->type, tsr->kind); return; } @@ -595,11 +639,12 @@ static void update_insn_state_x86(struct type_state *state, } tsr->type = state->regs[src->reg1].type; + tsr->kind = state->regs[src->reg1].kind; tsr->ok = true; pr_debug_dtp("mov [%x] reg%d -> reg%d", insn_offset, src->reg1, dst->reg1); - pr_debug_type_name(&tsr->type); + pr_debug_type_name(&tsr->type, tsr->kind); } /* Case 2. memory to register transers */ if (src->mem_ref && !dst->mem_ref) { @@ -622,11 +667,13 @@ retry: return; } else if (!stack->compound) { tsr->type = stack->type; + tsr->kind = stack->kind; tsr->ok = true; } else if (die_get_member_type(&stack->type, offset - stack->offset, &type_die)) { tsr->type = type_die; + tsr->kind = TSR_KIND_TYPE; tsr->ok = true; } else { tsr->ok = false; @@ -635,18 +682,20 @@ retry: pr_debug_dtp("mov [%x] -%#x(stack) -> reg%d", insn_offset, -offset, dst->reg1); - pr_debug_type_name(&tsr->type); + pr_debug_type_name(&tsr->type, tsr->kind); } /* And then dereference the pointer if it has one */ else if (has_reg_type(state, sreg) && state->regs[sreg].ok && + state->regs[sreg].kind == TSR_KIND_TYPE && die_deref_ptr_type(&state->regs[sreg].type, src->offset, &type_die)) { tsr->type = type_die; + tsr->kind = TSR_KIND_TYPE; tsr->ok = true; pr_debug_dtp("mov [%x] %#x(reg%d) -> reg%d", insn_offset, src->offset, sreg, dst->reg1); - pr_debug_type_name(&tsr->type); + pr_debug_type_name(&tsr->type, tsr->kind); } /* Or check if it's a global variable */ else if (sreg == DWARF_REG_PC) { @@ -665,11 +714,37 @@ retry: } tsr->type = type_die; + tsr->kind = TSR_KIND_TYPE; tsr->ok = true; pr_debug_dtp("mov [%x] global addr=%"PRIx64" -> reg%d", insn_offset, addr, dst->reg1); - pr_debug_type_name(&type_die); + pr_debug_type_name(&type_die, tsr->kind); + } + /* And check percpu access with base register */ + else if (has_reg_type(state, sreg) && + state->regs[sreg].kind == TSR_KIND_PERCPU_BASE) { + u64 ip = dloc->ms->sym->start + dl->al.offset; + int offset; + + /* + * In kernel, %gs points to a per-cpu region for the + * current CPU. Access with a constant offset should + * be treated as a global variable access. + */ + if (get_global_var_type(cu_die, dloc, ip, src->offset, + &offset, &type_die) && + die_get_member_type(&type_die, offset, &type_die)) { + tsr->type = type_die; + tsr->kind = TSR_KIND_TYPE; + tsr->ok = true; + + pr_debug_dtp("mov [%x] percpu %#x(reg%d) -> reg%d type=", + insn_offset, src->offset, sreg, dst->reg1); + pr_debug_type_name(&tsr->type, tsr->kind); + } else { + tsr->ok = false; + } } /* Or try another register if any */ else if (src->multi_regs && sreg == src->reg1 && @@ -677,8 +752,22 @@ retry: sreg = src->reg2; goto retry; } - /* It failed to get a type info, mark it as invalid */ else { + int offset; + const char *var_name = NULL; + + /* it might be per-cpu variable (in kernel) access */ + if (src->offset < 0) { + if (get_global_var_info(dloc, (s64)src->offset, + &var_name, &offset) && + !strcmp(var_name, "__per_cpu_offset")) { + tsr->kind = TSR_KIND_PERCPU_BASE; + + pr_debug_dtp("mov [%x] percpu base reg%d\n", + insn_offset, dst->reg1); + } + } + tsr->ok = false; } } @@ -693,6 +782,8 @@ retry: struct type_state_stack *stack; int offset = dst->offset - fboff; + tsr = &state->regs[src->reg1]; + stack = find_stack_state(state, offset); if (stack) { /* @@ -703,16 +794,16 @@ retry: * die_get_member_type(). */ if (!stack->compound) - set_stack_state(stack, offset, - &state->regs[src->reg1].type); + set_stack_state(stack, offset, tsr->kind, + &tsr->type); } else { - findnew_stack_state(state, offset, - &state->regs[src->reg1].type); + findnew_stack_state(state, offset, tsr->kind, + &tsr->type); } pr_debug_dtp("mov [%x] reg%d -> -%#x(stack)", insn_offset, src->reg1, -offset); - pr_debug_type_name(&state->regs[src->reg1].type); + pr_debug_type_name(&tsr->type, tsr->kind); } /* * Ignore other transfers since it'd set a value in a struct @@ -824,10 +915,11 @@ static bool check_matching_type(struct type_state *state, Dwarf_Word size; u32 insn_offset = dloc->ip - dloc->ms->sym->start; - pr_debug_dtp("chk [%x] reg%d offset=%#x ok=%d", - insn_offset, reg, dloc->op->offset, state->regs[reg].ok); + pr_debug_dtp("chk [%x] reg%d offset=%#x ok=%d kind=%d", + insn_offset, reg, dloc->op->offset, + state->regs[reg].ok, state->regs[reg].kind); - if (state->regs[reg].ok) { + if (state->regs[reg].ok && state->regs[reg].kind == TSR_KIND_TYPE) { int tag = dwarf_tag(&state->regs[reg].type); pr_debug_dtp("\n"); @@ -893,10 +985,25 @@ static bool check_matching_type(struct type_state *state, return true; } + if (state->regs[reg].kind == TSR_KIND_PERCPU_BASE) { + u64 var_addr = dloc->op->offset; + int var_offset; + + pr_debug_dtp(" percpu var\n"); + + if (get_global_var_type(cu_die, dloc, dloc->ip, var_addr, + &var_offset, type_die)) { + dloc->type_offset = var_offset; + return true; + } + return false; + } + if (map__dso(dloc->ms->map)->kernel && arch__is(dloc->arch, "x86")) { u64 addr; int offset; + /* Direct this-cpu access like "%gs:0x34740" */ if (dloc->op->segment == INSN_SEG_X86_GS && dloc->op->imm) { pr_debug_dtp(" this-cpu var\n"); @@ -907,6 +1014,24 @@ static bool check_matching_type(struct type_state *state, dloc->type_offset = offset; return true; } + return false; + } + + /* Access to per-cpu base like "-0x7dcf0500(,%rdx,8)" */ + if (dloc->op->offset < 0 && reg != state->stack_reg) { + const char *var_name = NULL; + + addr = (s64) dloc->op->offset; + + if (get_global_var_info(dloc, addr, &var_name, &offset) && + !strcmp(var_name, "__per_cpu_offset") && offset == 0 && + get_global_var_type(cu_die, dloc, dloc->ip, addr, + &offset, type_die)) { + pr_debug_dtp(" percpu base\n"); + + dloc->type_offset = offset; + return true; + } } } @@ -1015,7 +1140,7 @@ again: ret = 0; pr_debug_dtp("found by insn track: %#x(reg%d) type-offset=%#x", dloc->op->offset, reg, dloc->type_offset); - pr_debug_type_name(type_die); + pr_debug_type_name(type_die, TSR_KIND_TYPE); break; } @@ -1147,7 +1272,7 @@ retry: loc->offset, reg, fb_offset, offset); else pr_debug_dtp("%#x(reg%d)", loc->offset, reg); - pr_debug_type_name(type_die); + pr_debug_type_name(type_die, TSR_KIND_TYPE); } dloc->type_offset = offset; goto out; -- cgit v1.2.3 From eb9190afaed6afd5ed54bb0a3269eec338663858 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:12 -0700 Subject: perf annotate-data: Handle ADD instructions MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit There are different patterns for percpu variable access using a constant value added to the base.  2aeb:  mov    -0x7da0f7e0(,%rax,8),%r14 # r14 = __per_cpu_offset[cpu]  2af3:  mov    $0x34740,%rax # rax = address of runqueues * 2afa:  add    %rax,%r14 # r14 = &per_cpu(runqueues, cpu)  2bfd:  cmpl   $0x0,0x10(%r14) # cpu_rq(cpu)->has_blocked_load  2b03:  je     0x2b36 At the first instruction, r14 has the __per_cpu_offset. And then rax has an immediate value and then added to r14 to calculate the address of a per-cpu variable. So it needs to track the immediate values and ADD instructions. Similar but a little different case is to use "this_cpu_off" instead of "__per_cpu_offset" for the current CPU. This time the variable address comes with PC-rel addressing. 89: mov $0x34740,%rax # rax = address of runqueues * 90: add %gs:0x7f015f60(%rip),%rax # 19a78 98: incl 0xd8c(%rax) # cpu_rq(cpu)->sched_count Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-21-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 107 +++++++++++++++++++++++++++++++++++++++- 1 file changed, 105 insertions(+), 2 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 83b5aa00f01c..bd10a576cfbf 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -28,6 +28,8 @@ enum type_state_kind { TSR_KIND_INVALID = 0, TSR_KIND_TYPE, TSR_KIND_PERCPU_BASE, + TSR_KIND_CONST, + TSR_KIND_POINTER, }; #define pr_debug_dtp(fmt, ...) \ @@ -53,6 +55,13 @@ static void pr_debug_type_name(Dwarf_Die *die, enum type_state_kind kind) case TSR_KIND_PERCPU_BASE: pr_info(" percpu base\n"); return; + case TSR_KIND_CONST: + pr_info(" constant\n"); + return; + case TSR_KIND_POINTER: + pr_info(" pointer"); + /* it also prints the type info */ + break; case TSR_KIND_TYPE: default: break; @@ -393,7 +402,7 @@ static void set_stack_state(struct type_state_stack *stack, int offset, u8 kind, switch (tag) { case DW_TAG_structure_type: case DW_TAG_union_type: - stack->compound = true; + stack->compound = (kind != TSR_KIND_POINTER); break; default: stack->compound = false; @@ -585,6 +594,58 @@ static void update_insn_state_x86(struct type_state *state, return; } + if (!strncmp(dl->ins.name, "add", 3)) { + u64 imm_value = -1ULL; + int offset; + const char *var_name = NULL; + struct map_symbol *ms = dloc->ms; + u64 ip = ms->sym->start + dl->al.offset; + + if (!has_reg_type(state, dst->reg1)) + return; + + tsr = &state->regs[dst->reg1]; + + if (src->imm) + imm_value = src->offset; + else if (has_reg_type(state, src->reg1) && + state->regs[src->reg1].kind == TSR_KIND_CONST) + imm_value = state->regs[src->reg1].imm_value; + else if (src->reg1 == DWARF_REG_PC) { + u64 var_addr = annotate_calc_pcrel(dloc->ms, ip, + src->offset, dl); + + if (get_global_var_info(dloc, var_addr, + &var_name, &offset) && + !strcmp(var_name, "this_cpu_off") && + tsr->kind == TSR_KIND_CONST) { + tsr->kind = TSR_KIND_PERCPU_BASE; + imm_value = tsr->imm_value; + } + } + else + return; + + if (tsr->kind != TSR_KIND_PERCPU_BASE) + return; + + if (get_global_var_type(cu_die, dloc, ip, imm_value, &offset, + &type_die) && offset == 0) { + /* + * This is not a pointer type, but it should be treated + * as a pointer. + */ + tsr->type = type_die; + tsr->kind = TSR_KIND_POINTER; + tsr->ok = true; + + pr_debug_dtp("add [%x] percpu %#"PRIx64" -> reg%d", + insn_offset, imm_value, dst->reg1); + pr_debug_type_name(&tsr->type, tsr->kind); + } + return; + } + if (strncmp(dl->ins.name, "mov", 3)) return; @@ -632,6 +693,16 @@ static void update_insn_state_x86(struct type_state *state, return; } + if (src->imm) { + tsr->kind = TSR_KIND_CONST; + tsr->imm_value = src->offset; + tsr->ok = true; + + pr_debug_dtp("mov [%x] imm=%#x -> reg%d\n", + insn_offset, tsr->imm_value, dst->reg1); + return; + } + if (!has_reg_type(state, src->reg1) || !state->regs[src->reg1].ok) { tsr->ok = false; @@ -739,13 +810,26 @@ retry: tsr->kind = TSR_KIND_TYPE; tsr->ok = true; - pr_debug_dtp("mov [%x] percpu %#x(reg%d) -> reg%d type=", + pr_debug_dtp("mov [%x] percpu %#x(reg%d) -> reg%d", insn_offset, src->offset, sreg, dst->reg1); pr_debug_type_name(&tsr->type, tsr->kind); } else { tsr->ok = false; } } + /* And then dereference the calculated pointer if it has one */ + else if (has_reg_type(state, sreg) && state->regs[sreg].ok && + state->regs[sreg].kind == TSR_KIND_POINTER && + die_get_member_type(&state->regs[sreg].type, + src->offset, &type_die)) { + tsr->type = type_die; + tsr->kind = TSR_KIND_TYPE; + tsr->ok = true; + + pr_debug_dtp("mov [%x] pointer %#x(reg%d) -> reg%d", + insn_offset, src->offset, sreg, dst->reg1); + pr_debug_type_name(&tsr->type, tsr->kind); + } /* Or try another register if any */ else if (src->multi_regs && sreg == src->reg1 && src->reg1 != src->reg2) { @@ -999,6 +1083,25 @@ static bool check_matching_type(struct type_state *state, return false; } + if (state->regs[reg].ok && state->regs[reg].kind == TSR_KIND_POINTER) { + pr_debug_dtp(" percpu ptr\n"); + + /* + * It's actaully pointer but the address was calculated using + * some arithmetic. So it points to the actual type already. + */ + *type_die = state->regs[reg].type; + + dloc->type_offset = dloc->op->offset; + + /* Get the size of the actual type */ + if (dwarf_aggregate_size(type_die, &size) < 0 || + (unsigned)dloc->type_offset >= size) + return false; + + return true; + } + if (map__dso(dloc->ms->map)->kernel && arch__is(dloc->arch, "x86")) { u64 addr; int offset; -- cgit v1.2.3 From b3c95109c131fcc959d2473e7c384d8cc62d23d0 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:13 -0700 Subject: perf annotate-data: Add stack canary type When the stack protector is enabled, compiler would generate code to check stack overflow with a special value called 'stack carary' at runtime. On x86_64, GCC hard-codes the stack canary as %gs:40. While there's a definition of fixed_percpu_data in asm/processor.h, it seems that the header is not included everywhere and many places it cannot find the type info. As it's in the well-known location (at %gs:40), let's add a pseudo stack canary type to handle it specially. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-22-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 46 +++++++++++++++++++++++++++++++++++++++++ tools/perf/util/annotate-data.h | 1 + tools/perf/util/annotate.c | 25 ++++++++++++++++++++++ 3 files changed, 72 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index bd10a576cfbf..633fe125fcd8 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -30,6 +30,7 @@ enum type_state_kind { TSR_KIND_PERCPU_BASE, TSR_KIND_CONST, TSR_KIND_POINTER, + TSR_KIND_CANARY, }; #define pr_debug_dtp(fmt, ...) \ @@ -62,6 +63,9 @@ static void pr_debug_type_name(Dwarf_Die *die, enum type_state_kind kind) pr_info(" pointer"); /* it also prints the type info */ break; + case TSR_KIND_CANARY: + pr_info(" stack canary\n"); + return; case TSR_KIND_TYPE: default: break; @@ -676,6 +680,15 @@ static void update_insn_state_x86(struct type_state *state, */ var_addr = src->offset; + if (var_addr == 40) { + tsr->kind = TSR_KIND_CANARY; + tsr->ok = true; + + pr_debug_dtp("mov [%x] stack canary -> reg%d\n", + insn_offset, dst->reg1); + return; + } + if (!get_global_var_type(cu_die, dloc, ip, var_addr, &offset, &type_die) || !die_get_member_type(&type_die, offset, &type_die)) { @@ -991,6 +1004,16 @@ static void delete_var_types(struct die_var_type *var_types) } } +/* should match to is_stack_canary() in util/annotate.c */ +static void setup_stack_canary(struct data_loc_info *dloc) +{ + if (arch__is(dloc->arch, "x86")) { + dloc->op->segment = INSN_SEG_X86_GS; + dloc->op->imm = true; + dloc->op->offset = 40; + } +} + /* It's at the target address, check if it has a matching type */ static bool check_matching_type(struct type_state *state, struct data_loc_info *dloc, int reg, @@ -1038,6 +1061,11 @@ static bool check_matching_type(struct type_state *state, if (stack == NULL) return false; + if (stack->kind == TSR_KIND_CANARY) { + setup_stack_canary(dloc); + return false; + } + *type_die = stack->type; /* Update the type offset from the start of slot */ dloc->type_offset -= stack->offset; @@ -1062,6 +1090,11 @@ static bool check_matching_type(struct type_state *state, if (stack == NULL) return false; + if (stack->kind == TSR_KIND_CANARY) { + setup_stack_canary(dloc); + return false; + } + *type_die = stack->type; /* Update the type offset from the start of slot */ dloc->type_offset -= fboff + stack->offset; @@ -1102,6 +1135,19 @@ static bool check_matching_type(struct type_state *state, return true; } + if (state->regs[reg].ok && state->regs[reg].kind == TSR_KIND_CANARY) { + pr_debug_dtp(" stack canary\n"); + + /* + * This is a saved value of the stack canary which will be handled + * in the outer logic when it returns failure here. Pretend it's + * from the stack canary directly. + */ + setup_stack_canary(dloc); + + return false; + } + if (map__dso(dloc->ms->map)->kernel && arch__is(dloc->arch, "x86")) { u64 addr; int offset; diff --git a/tools/perf/util/annotate-data.h b/tools/perf/util/annotate-data.h index ae0f87aed804..1b5a152163b5 100644 --- a/tools/perf/util/annotate-data.h +++ b/tools/perf/util/annotate-data.h @@ -73,6 +73,7 @@ struct annotated_data_type { extern struct annotated_data_type unknown_type; extern struct annotated_data_type stackop_type; +extern struct annotated_data_type canary_type; /** * struct data_loc_info - Data location information diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index e4121acb4f88..64e54ff1aa1d 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -118,6 +118,13 @@ struct annotated_data_type stackop_type = { }, }; +struct annotated_data_type canary_type = { + .self = { + .type_name = (char *)"(stack canary)", + .children = LIST_HEAD_INIT(canary_type.self.children), + }, +}; + static int arch__grow_instructions(struct arch *arch) { struct ins *new_instructions; @@ -3803,6 +3810,18 @@ static bool is_stack_operation(struct arch *arch, struct disasm_line *dl) return false; } +static bool is_stack_canary(struct arch *arch, struct annotated_op_loc *loc) +{ + /* On x86_64, %gs:40 is used for stack canary */ + if (arch__is(arch, "x86")) { + if (loc->segment == INSN_SEG_X86_GS && loc->imm && + loc->offset == 40) + return true; + } + + return false; +} + u64 annotate_calc_pcrel(struct map_symbol *ms, u64 ip, int offset, struct disasm_line *dl) { @@ -3929,6 +3948,12 @@ retry: } mem_type = find_data_type(&dloc); + + if (mem_type == NULL && is_stack_canary(arch, op_loc)) { + mem_type = &canary_type; + dloc.type_offset = 0; + } + if (mem_type) istat->good++; else -- cgit v1.2.3 From 55ee3d005d62279d3951a26d5c211a4d9aebc222 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:14 -0700 Subject: perf annotate-data: Add a cache for global variable types They are often searched by many different places. Let's add a cache for them to reduce the duplicate DWARF access. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-23-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 107 ++++++++++++++++++++++++++++++++++++++-- tools/perf/util/annotate-data.h | 7 +++ tools/perf/util/dso.c | 2 + tools/perf/util/dso.h | 6 ++- 4 files changed, 118 insertions(+), 4 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 633fe125fcd8..969e2f82079c 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -433,6 +433,91 @@ static struct type_state_stack *findnew_stack_state(struct type_state *state, return stack; } +/* Maintain a cache for quick global variable lookup */ +struct global_var_entry { + struct rb_node node; + char *name; + u64 start; + u64 end; + u64 die_offset; +}; + +static int global_var_cmp(const void *_key, const struct rb_node *node) +{ + const u64 addr = (uintptr_t)_key; + struct global_var_entry *gvar; + + gvar = rb_entry(node, struct global_var_entry, node); + + if (gvar->start <= addr && addr < gvar->end) + return 0; + return gvar->start > addr ? -1 : 1; +} + +static bool global_var_less(struct rb_node *node_a, const struct rb_node *node_b) +{ + struct global_var_entry *gvar_a, *gvar_b; + + gvar_a = rb_entry(node_a, struct global_var_entry, node); + gvar_b = rb_entry(node_b, struct global_var_entry, node); + + return gvar_a->start < gvar_b->start; +} + +static struct global_var_entry *global_var__find(struct data_loc_info *dloc, u64 addr) +{ + struct dso *dso = map__dso(dloc->ms->map); + struct rb_node *node; + + node = rb_find((void *)(uintptr_t)addr, &dso->global_vars, global_var_cmp); + if (node == NULL) + return NULL; + + return rb_entry(node, struct global_var_entry, node); +} + +static bool global_var__add(struct data_loc_info *dloc, u64 addr, + const char *name, Dwarf_Die *type_die) +{ + struct dso *dso = map__dso(dloc->ms->map); + struct global_var_entry *gvar; + Dwarf_Word size; + + if (dwarf_aggregate_size(type_die, &size) < 0) + return false; + + gvar = malloc(sizeof(*gvar)); + if (gvar == NULL) + return false; + + gvar->name = strdup(name); + if (gvar->name == NULL) { + free(gvar); + return false; + } + + gvar->start = addr; + gvar->end = addr + size; + gvar->die_offset = dwarf_dieoffset(type_die); + + rb_add(&gvar->node, &dso->global_vars, global_var_less); + return true; +} + +void global_var_type__tree_delete(struct rb_root *root) +{ + struct global_var_entry *gvar; + + while (!RB_EMPTY_ROOT(root)) { + struct rb_node *node = rb_first(root); + + rb_erase(node, root); + gvar = rb_entry(node, struct global_var_entry, node); + free(gvar->name); + free(gvar); + } +} + static bool get_global_var_info(struct data_loc_info *dloc, u64 addr, const char **var_name, int *var_offset) { @@ -467,14 +552,25 @@ static bool get_global_var_type(Dwarf_Die *cu_die, struct data_loc_info *dloc, u64 pc; int offset; bool is_pointer = false; - const char *var_name; + const char *var_name = NULL; + struct global_var_entry *gvar; Dwarf_Die var_die; + gvar = global_var__find(dloc, var_addr); + if (gvar) { + if (!dwarf_offdie(dloc->di->dbg, gvar->die_offset, type_die)) + return false; + + *var_offset = var_addr - gvar->start; + return true; + } + /* Try to get the variable by address first */ if (die_find_variable_by_addr(cu_die, var_addr, &var_die, &offset) && check_variable(&var_die, type_die, offset, is_pointer) == 0) { + var_name = dwarf_diename(&var_die); *var_offset = offset; - return true; + goto ok; } if (!get_global_var_info(dloc, var_addr, &var_name, var_offset)) @@ -485,9 +581,14 @@ static bool get_global_var_type(Dwarf_Die *cu_die, struct data_loc_info *dloc, /* Try to get the name of global variable */ if (die_find_variable_at(cu_die, var_name, pc, &var_die) && check_variable(&var_die, type_die, *var_offset, is_pointer) == 0) - return true; + goto ok; return false; + +ok: + /* The address should point to the start of the variable */ + global_var__add(dloc, var_addr - *var_offset, var_name, type_die); + return true; } /** diff --git a/tools/perf/util/annotate-data.h b/tools/perf/util/annotate-data.h index 1b5a152163b5..fe1e53d6e8c7 100644 --- a/tools/perf/util/annotate-data.h +++ b/tools/perf/util/annotate-data.h @@ -153,6 +153,9 @@ int annotated_data_type__update_samples(struct annotated_data_type *adt, /* Release all data type information in the tree */ void annotated_data_type__tree_delete(struct rb_root *root); +/* Release all global variable information in the tree */ +void global_var_type__tree_delete(struct rb_root *root); + #else /* HAVE_DWARF_SUPPORT */ static inline struct annotated_data_type * @@ -175,6 +178,10 @@ static inline void annotated_data_type__tree_delete(struct rb_root *root __maybe { } +static inline void global_var_type__tree_delete(struct rb_root *root __maybe_unused) +{ +} + #endif /* HAVE_DWARF_SUPPORT */ #endif /* _PERF_ANNOTATE_DATA_H */ diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c index 22fd5fa806ed..6e2a7198b382 100644 --- a/tools/perf/util/dso.c +++ b/tools/perf/util/dso.c @@ -1329,6 +1329,7 @@ struct dso *dso__new_id(const char *name, struct dso_id *id) dso->inlined_nodes = RB_ROOT_CACHED; dso->srclines = RB_ROOT_CACHED; dso->data_types = RB_ROOT; + dso->global_vars = RB_ROOT; dso->data.fd = -1; dso->data.status = DSO_DATA_STATUS_UNKNOWN; dso->symtab_type = DSO_BINARY_TYPE__NOT_FOUND; @@ -1373,6 +1374,7 @@ void dso__delete(struct dso *dso) dso->symbol_names_len = 0; zfree(&dso->symbol_names); annotated_data_type__tree_delete(&dso->data_types); + global_var_type__tree_delete(&dso->global_vars); if (dso->short_name_allocated) { zfree((char **)&dso->short_name); diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h index ce9f3849a773..2cdcd1e2ef8b 100644 --- a/tools/perf/util/dso.h +++ b/tools/perf/util/dso.h @@ -154,7 +154,8 @@ struct dso { size_t symbol_names_len; struct rb_root_cached inlined_nodes; struct rb_root_cached srclines; - struct rb_root data_types; + struct rb_root data_types; + struct rb_root global_vars; struct { u64 addr; @@ -411,4 +412,7 @@ int dso__strerror_load(struct dso *dso, char *buf, size_t buflen); void reset_fd_limit(void); +u64 dso__find_global_type(struct dso *dso, u64 addr); +u64 dso__findnew_global_type(struct dso *dso, u64 addr, u64 offset); + #endif /* __PERF_DSO */ -- cgit v1.2.3 From bd62de08084c24a4ff1b1875d53cc4cc1ea2312d Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Mar 2024 22:51:15 -0700 Subject: perf annotate-data: Do not retry for invalid types In some cases, it was able to find a type or location info (for per-cpu variable) but cannot match because of invalid offset or missing global information. In those cases, it's meaningless to go to the outer scope and retry because there will be no additional information. Let's change the return type of find_matching_type() and bail out if it returns -1 for the cases. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240319055115.4063940-24-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 83 ++++++++++++++++++++++++----------------- 1 file changed, 48 insertions(+), 35 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 969e2f82079c..043d80791bd0 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -1115,10 +1115,15 @@ static void setup_stack_canary(struct data_loc_info *dloc) } } -/* It's at the target address, check if it has a matching type */ -static bool check_matching_type(struct type_state *state, - struct data_loc_info *dloc, int reg, - Dwarf_Die *cu_die, Dwarf_Die *type_die) +/* + * It's at the target address, check if it has a matching type. + * It returns 1 if found, 0 if not or -1 if not found but no need to + * repeat the search. The last case is for per-cpu variables which + * are similar to global variables and no additional info is needed. + */ +static int check_matching_type(struct type_state *state, + struct data_loc_info *dloc, int reg, + Dwarf_Die *cu_die, Dwarf_Die *type_die) { Dwarf_Word size; u32 insn_offset = dloc->ip - dloc->ms->sym->start; @@ -1137,20 +1142,20 @@ static bool check_matching_type(struct type_state *state, * dereference a memory location. */ if (tag != DW_TAG_pointer_type && tag != DW_TAG_array_type) - return false; + return -1; /* Remove the pointer and get the target type */ if (die_get_real_type(&state->regs[reg].type, type_die) == NULL) - return false; + return -1; dloc->type_offset = dloc->op->offset; /* Get the size of the actual type */ if (dwarf_aggregate_size(type_die, &size) < 0 || (unsigned)dloc->type_offset >= size) - return false; + return -1; - return true; + return 1; } if (reg == dloc->fbreg) { @@ -1160,18 +1165,18 @@ static bool check_matching_type(struct type_state *state, stack = find_stack_state(state, dloc->type_offset); if (stack == NULL) - return false; + return 0; if (stack->kind == TSR_KIND_CANARY) { setup_stack_canary(dloc); - return false; + return -1; } *type_die = stack->type; /* Update the type offset from the start of slot */ dloc->type_offset -= stack->offset; - return true; + return 1; } if (dloc->fb_cfa) { @@ -1185,22 +1190,22 @@ static bool check_matching_type(struct type_state *state, fbreg = -1; if (reg != fbreg) - return false; + return 0; stack = find_stack_state(state, dloc->type_offset - fboff); if (stack == NULL) - return false; + return 0; if (stack->kind == TSR_KIND_CANARY) { setup_stack_canary(dloc); - return false; + return -1; } *type_die = stack->type; /* Update the type offset from the start of slot */ dloc->type_offset -= fboff + stack->offset; - return true; + return 1; } if (state->regs[reg].kind == TSR_KIND_PERCPU_BASE) { @@ -1212,9 +1217,10 @@ static bool check_matching_type(struct type_state *state, if (get_global_var_type(cu_die, dloc, dloc->ip, var_addr, &var_offset, type_die)) { dloc->type_offset = var_offset; - return true; + return 1; } - return false; + /* No need to retry per-cpu (global) variables */ + return -1; } if (state->regs[reg].ok && state->regs[reg].kind == TSR_KIND_POINTER) { @@ -1231,9 +1237,9 @@ static bool check_matching_type(struct type_state *state, /* Get the size of the actual type */ if (dwarf_aggregate_size(type_die, &size) < 0 || (unsigned)dloc->type_offset >= size) - return false; + return -1; - return true; + return 1; } if (state->regs[reg].ok && state->regs[reg].kind == TSR_KIND_CANARY) { @@ -1246,7 +1252,7 @@ static bool check_matching_type(struct type_state *state, */ setup_stack_canary(dloc); - return false; + return -1; } if (map__dso(dloc->ms->map)->kernel && arch__is(dloc->arch, "x86")) { @@ -1262,9 +1268,9 @@ static bool check_matching_type(struct type_state *state, if (get_global_var_type(cu_die, dloc, dloc->ip, addr, &offset, type_die)) { dloc->type_offset = offset; - return true; + return 1; } - return false; + return -1; } /* Access to per-cpu base like "-0x7dcf0500(,%rdx,8)" */ @@ -1280,26 +1286,28 @@ static bool check_matching_type(struct type_state *state, pr_debug_dtp(" percpu base\n"); dloc->type_offset = offset; - return true; + return 1; } + pr_debug_dtp(" negative offset\n"); + return -1; } } pr_debug_dtp("\n"); - return false; + return 0; } /* Iterate instructions in basic blocks and update type table */ -static bool find_data_type_insn(struct data_loc_info *dloc, int reg, - struct list_head *basic_blocks, - struct die_var_type *var_types, - Dwarf_Die *cu_die, Dwarf_Die *type_die) +static int find_data_type_insn(struct data_loc_info *dloc, int reg, + struct list_head *basic_blocks, + struct die_var_type *var_types, + Dwarf_Die *cu_die, Dwarf_Die *type_die) { struct type_state state; struct symbol *sym = dloc->ms->sym; struct annotation *notes = symbol__annotation(sym); struct annotated_basic_block *bb; - bool found = false; + int ret = 0; init_type_state(&state, dloc->arch); @@ -1317,8 +1325,8 @@ static bool find_data_type_insn(struct data_loc_info *dloc, int reg, update_var_state(&state, dloc, addr, dl->al.offset, var_types); if (this_ip == dloc->ip) { - found = check_matching_type(&state, dloc, reg, - cu_die, type_die); + ret = check_matching_type(&state, dloc, reg, + cu_die, type_die); goto out; } @@ -1331,7 +1339,7 @@ static bool find_data_type_insn(struct data_loc_info *dloc, int reg, out: exit_type_state(&state); - return found; + return ret; } /* @@ -1355,6 +1363,7 @@ static int find_data_type_block(struct data_loc_info *dloc, int reg, for (int i = nr_scopes - 1; i >= 0; i--) { Dwarf_Addr base, start, end; LIST_HEAD(this_blocks); + int found; if (dwarf_ranges(&scopes[i], 0, &base, &start, &end) < 0) break; @@ -1385,15 +1394,19 @@ again: fixup_var_address(var_types, start); /* Find from start of this scope to the target instruction */ - if (find_data_type_insn(dloc, reg, &basic_blocks, var_types, - cu_die, type_die)) { - ret = 0; + found = find_data_type_insn(dloc, reg, &basic_blocks, var_types, + cu_die, type_die); + if (found > 0) { pr_debug_dtp("found by insn track: %#x(reg%d) type-offset=%#x", dloc->op->offset, reg, dloc->type_offset); pr_debug_type_name(type_die, TSR_KIND_TYPE); + ret = 0; break; } + if (found < 0) + break; + /* Go up to the next scope and find blocks to the start */ prev_dst_ip = dst_ip; dst_ip = src_ip; -- cgit v1.2.3 From 525615ef6df476d5fe277d311c645b4c179d56db Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 7 Mar 2024 16:19:10 -0800 Subject: perf list: Add tracepoint encoding to detailed output The tracepoint id holds the config value and is probed in determining what an event is. Add reading of the id so that we can display the event encoding as: $ perf list --details ... alarmtimer:alarmtimer_cancel [Tracepoint event] tracepoint/config=0x18c/ ... Signed-off-by: Ian Rogers Tested-by: Arnaldo Carvalho de Melo Tested-by: Kan Liang Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Yang Jihong Link: https://lore.kernel.org/r/20240308001915.4060155-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/print-events.c | 35 +++++++++++++++++++++++++---------- 1 file changed, 25 insertions(+), 10 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/print-events.c b/tools/perf/util/print-events.c index 7b54e9385442..e0d2b49bab66 100644 --- a/tools/perf/util/print-events.c +++ b/tools/perf/util/print-events.c @@ -9,6 +9,7 @@ #include #include +#include #include #include #include @@ -92,34 +93,48 @@ void print_tracepoint_events(const struct print_callbacks *print_cb __maybe_unus evt_items = scandirat(events_fd, sys_dirent->d_name, &evt_namelist, NULL, alphasort); for (int j = 0; j < evt_items; j++) { + /* + * Buffer sized at twice the max filename length + 1 + * separator + 1 \0 terminator. + */ + char buf[NAME_MAX * 2 + 2]; + /* 16 possible hex digits and 22 other characters and \0. */ + char encoding[16 + 22]; struct dirent *evt_dirent = evt_namelist[j]; - char evt_path[MAXPATHLEN]; - int evt_fd; + struct io id; + __u64 config; if (evt_dirent->d_type != DT_DIR || !strcmp(evt_dirent->d_name, ".") || !strcmp(evt_dirent->d_name, "..")) goto next_evt; - snprintf(evt_path, sizeof(evt_path), "%s/id", evt_dirent->d_name); - evt_fd = openat(dir_fd, evt_path, O_RDONLY); - if (evt_fd < 0) + snprintf(buf, sizeof(buf), "%s/id", evt_dirent->d_name); + io__init(&id, openat(dir_fd, buf, O_RDONLY), buf, sizeof(buf)); + + if (id.fd < 0) + goto next_evt; + + if (io__get_dec(&id, &config) < 0) { + close(id.fd); goto next_evt; - close(evt_fd); + } + close(id.fd); - snprintf(evt_path, MAXPATHLEN, "%s:%s", + snprintf(buf, sizeof(buf), "%s:%s", sys_dirent->d_name, evt_dirent->d_name); + snprintf(encoding, sizeof(encoding), "tracepoint/config=0x%llx/", config); print_cb->print_event(print_state, /*topic=*/NULL, - /*pmu_name=*/NULL, - evt_path, + /*pmu_name=*/NULL, /* really "tracepoint" */ + /*event_name=*/buf, /*event_alias=*/NULL, /*scale_unit=*/NULL, /*deprecated=*/false, "Tracepoint event", /*desc=*/NULL, /*long_desc=*/NULL, - /*encoding_desc=*/NULL); + encoding); next_evt: free(evt_namelist[j]); } -- cgit v1.2.3 From 39aa4ff61fd310704d04063fc3f2842011e03b4d Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 7 Mar 2024 16:19:11 -0800 Subject: perf pmu: Drop "default_core" from alias names "default_core" is used by jevents.py for json events' PMU name when none is specified. On x86 the "default_core" is typically the PMU "cpu". When creating an alias see if the event's PMU name is "default_core" in which case don't record it. This means in places like "perf list" the PMU's name will be used in its place. Before: $ perf list --details ... cache: l1d.replacement [Counts the number of cache lines replaced in L1 data cache] default_core/event=0x51,period=0x186a3,umask=0x1/ ... After: $ perf list --details ... cache: l1d.replacement [Counts the number of cache lines replaced in L1 data cache. Unit: cpu] cpu/event=0x51,period=0x186a3,umask=0x1/ ... Signed-off-by: Ian Rogers Tested-by: Kan Liang Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Yang Jihong Link: https://lore.kernel.org/r/20240308001915.4060155-3-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/pmu.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index f39cbbc1a7ec..24be587e3537 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -518,7 +518,8 @@ static int perf_pmu__new_alias(struct perf_pmu *pmu, const char *name, unit = pe->unit; perpkg = pe->perpkg; deprecated = pe->deprecated; - pmu_name = pe->pmu; + if (pe->pmu && strcmp(pe->pmu, "default_core")) + pmu_name = pe->pmu; } alias = zalloc(sizeof(*alias)); -- cgit v1.2.3 From 4ccf3bb703ed01a872de7d39db53f477d44ade0b Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 7 Mar 2024 16:19:13 -0800 Subject: perf list: Give more details about raw event encodings List all the PMUs, not just the first core one, and list real format specifiers with value ranges. Before: $ perf list ... rNNN [Raw hardware event descriptor] cpu/t1=v1[,t2=v2,t3 ...]/modifier [Raw hardware event descriptor] [(see 'man perf-list' on how to encode it)] mem:[/len][:access] [Hardware breakpoint] ... After: $ perf list ... rNNN [Raw event descriptor] cpu/event=0..255,pc,edge,.../modifier [Raw event descriptor] [(see 'man perf-list' or 'man perf-record' on how to encode it)] breakpoint//modifier [Raw event descriptor] cstate_core/event=0..0xffffffffffffffff/modifier [Raw event descriptor] cstate_pkg/event=0..0xffffffffffffffff/modifier [Raw event descriptor] i915/i915_eventid=0..0x1fffff/modifier [Raw event descriptor] intel_bts//modifier [Raw event descriptor] intel_pt/ptw,event,cyc_thresh=0..15,.../modifier [Raw event descriptor] kprobe/retprobe/modifier [Raw event descriptor] msr/event=0..0xffffffffffffffff/modifier [Raw event descriptor] power/event=0..255/modifier [Raw event descriptor] software//modifier [Raw event descriptor] tracepoint//modifier [Raw event descriptor] uncore_arb/event=0..255,edge,inv,.../modifier [Raw event descriptor] uncore_cbox/event=0..255,edge,inv,.../modifier [Raw event descriptor] uncore_clock/event=0..255/modifier [Raw event descriptor] uncore_imc_free_running/event=0..255,umask=0..255/modifier[Raw event descriptor] uprobe/ref_ctr_offset=0..0xffffffff,retprobe/modifier[Raw event descriptor] mem:[/len][:access] [Hardware breakpoint] ... With '--details' provide more details on the formats encoding: cpu/event=0..255,pc,edge,.../modifier [Raw event descriptor] [(see 'man perf-list' or 'man perf-record' on how to encode it)] cpu/event=0..255,pc,edge,offcore_rsp=0..0xffffffffffffffff,ldlat=0..0xffff,inv, umask=0..255,frontend=0..0xffffff,cmask=0..255,config=0..0xffffffffffffffff, config1=0..0xffffffffffffffff,config2=0..0xffffffffffffffff,config3=0..0xffffffffffffffff, name=string,period=number,freq=number,branch_type=(u|k|hv|any|...),time, call-graph=(fp|dwarf|lbr),stack-size=number,max-stack=number,nr=number,inherit,no-inherit, overwrite,no-overwrite,percore,aux-output,aux-sample-size=number/modifier breakpoint//modifier [Raw event descriptor] breakpoint//modifier cstate_core/event=0..0xffffffffffffffff/modifier [Raw event descriptor] cstate_core/event=0..0xffffffffffffffff/modifier cstate_pkg/event=0..0xffffffffffffffff/modifier [Raw event descriptor] cstate_pkg/event=0..0xffffffffffffffff/modifier i915/i915_eventid=0..0x1fffff/modifier [Raw event descriptor] i915/i915_eventid=0..0x1fffff/modifier intel_bts//modifier [Raw event descriptor] intel_bts//modifier intel_pt/ptw,event,cyc_thresh=0..15,.../modifier [Raw event descriptor] intel_pt/ptw,event,cyc_thresh=0..15,pt,notnt,branch,tsc,pwr_evt,fup_on_ptw,cyc,noretcomp, mtc,psb_period=0..15,mtc_period=0..15/modifier kprobe/retprobe/modifier [Raw event descriptor] kprobe/retprobe/modifier msr/event=0..0xffffffffffffffff/modifier [Raw event descriptor] msr/event=0..0xffffffffffffffff/modifier power/event=0..255/modifier [Raw event descriptor] power/event=0..255/modifier software//modifier [Raw event descriptor] software//modifier tracepoint//modifier [Raw event descriptor] tracepoint//modifier uncore_arb/event=0..255,edge,inv,.../modifier [Raw event descriptor] uncore_arb/event=0..255,edge,inv,umask=0..255,cmask=0..31/modifier uncore_cbox/event=0..255,edge,inv,.../modifier [Raw event descriptor] uncore_cbox/event=0..255,edge,inv,umask=0..255,cmask=0..31/modifier uncore_clock/event=0..255/modifier [Raw event descriptor] uncore_clock/event=0..255/modifier uncore_imc_free_running/event=0..255,umask=0..255/modifier[Raw event descriptor] uncore_imc_free_running/event=0..255,umask=0..255/modifier uprobe/ref_ctr_offset=0..0xffffffff,retprobe/modifier[Raw event descriptor] uprobe/ref_ctr_offset=0..0xffffffff,retprobe/modifier Committer notes: Address this build error in various distros: 55 58.44 ubuntu:24.04 : FAIL gcc version 13.2.0 (Ubuntu 13.2.0-17ubuntu2) util/pmu.c:1638:70: error: '_Static_assert' with no message is a C2x extension [-Werror,-Wc2x-extensions] 1638 | _Static_assert(ARRAY_SIZE(terms) == __PARSE_EVENTS__TERM_TYPE_NR - 6); | ^ | , "" 1 error generated. Signed-off-by: Ian Rogers Tested-by: Arnaldo Carvalho de Melo Tested-by: Kan Liang Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Yang Jihong Link: https://lore.kernel.org/r/20240308001915.4060155-5-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/pmu.c | 62 +++++++++++++++++++++++++++- tools/perf/util/pmu.h | 3 ++ tools/perf/util/pmus.c | 94 ++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/pmus.h | 1 + tools/perf/util/print-events.c | 20 +-------- 5 files changed, 161 insertions(+), 19 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index 24be587e3537..81952c6cd3c6 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -1603,6 +1603,62 @@ bool perf_pmu__has_format(const struct perf_pmu *pmu, const char *name) return false; } +int perf_pmu__for_each_format(struct perf_pmu *pmu, void *state, pmu_format_callback cb) +{ + static const char *const terms[] = { + "config=0..0xffffffffffffffff", + "config1=0..0xffffffffffffffff", + "config2=0..0xffffffffffffffff", + "config3=0..0xffffffffffffffff", + "name=string", + "period=number", + "freq=number", + "branch_type=(u|k|hv|any|...)", + "time", + "call-graph=(fp|dwarf|lbr)", + "stack-size=number", + "max-stack=number", + "nr=number", + "inherit", + "no-inherit", + "overwrite", + "no-overwrite", + "percore", + "aux-output", + "aux-sample-size=number", + }; + struct perf_pmu_format *format; + int ret; + + /* + * max-events and driver-config are missing above as are the internal + * types user, metric-id, raw, legacy cache and hardware. Assert against + * the enum parse_events__term_type so they are kept in sync. + */ + _Static_assert(ARRAY_SIZE(terms) == __PARSE_EVENTS__TERM_TYPE_NR - 6, + "perf_pmu__for_each_format()'s terms must be kept in sync with enum parse_events__term_type"); + list_for_each_entry(format, &pmu->format, list) { + perf_pmu_format__load(pmu, format); + ret = cb(state, format->name, (int)format->value, format->bits); + if (ret) + return ret; + } + if (!pmu->is_core) + return 0; + + for (size_t i = 0; i < ARRAY_SIZE(terms); i++) { + int config = PERF_PMU_FORMAT_VALUE_CONFIG; + + if (i < PERF_PMU_FORMAT_VALUE_CONFIG_END) + config = i; + + ret = cb(state, terms[i], config, /*bits=*/NULL); + if (ret) + return ret; + } + return 0; +} + bool is_pmu_core(const char *name) { return !strcmp(name, "cpu") || !strcmp(name, "cpum_cf") || is_sysfs_pmu_core(name); @@ -1697,8 +1753,12 @@ int perf_pmu__for_each_event(struct perf_pmu *pmu, bool skip_duplicate_pmus, pmu_add_cpu_aliases(pmu); list_for_each_entry(event, &pmu->aliases, list) { size_t buf_used; + int pmu_name_len; info.pmu_name = event->pmu_name ?: pmu->name; + pmu_name_len = skip_duplicate_pmus + ? pmu_name_len_no_suffix(info.pmu_name, /*num=*/NULL) + : (int)strlen(info.pmu_name); info.alias = NULL; if (event->desc) { info.name = event->name; @@ -1723,7 +1783,7 @@ int perf_pmu__for_each_event(struct perf_pmu *pmu, bool skip_duplicate_pmus, info.encoding_desc = buf + buf_used; parse_events_terms__to_strbuf(&event->terms, &sb); buf_used += snprintf(buf + buf_used, sizeof(buf) - buf_used, - "%s/%s/", info.pmu_name, sb.buf) + 1; + "%.*s/%s/", pmu_name_len, info.pmu_name, sb.buf) + 1; info.topic = event->topic; info.str = sb.buf; info.deprecated = event->deprecated; diff --git a/tools/perf/util/pmu.h b/tools/perf/util/pmu.h index e35d985206db..9f5284b29ecf 100644 --- a/tools/perf/util/pmu.h +++ b/tools/perf/util/pmu.h @@ -196,6 +196,8 @@ struct pmu_event_info { }; typedef int (*pmu_event_callback)(void *state, struct pmu_event_info *info); +typedef int (*pmu_format_callback)(void *state, const char *name, int config, + const unsigned long *bits); void pmu_add_sys_aliases(struct perf_pmu *pmu); int perf_pmu__config(struct perf_pmu *pmu, struct perf_event_attr *attr, @@ -215,6 +217,7 @@ int perf_pmu__find_event(struct perf_pmu *pmu, const char *event, void *state, p int perf_pmu__format_parse(struct perf_pmu *pmu, int dirfd, bool eager_load); void perf_pmu_format__set_value(void *format, int config, unsigned long *bits); bool perf_pmu__has_format(const struct perf_pmu *pmu, const char *name); +int perf_pmu__for_each_format(struct perf_pmu *pmu, void *state, pmu_format_callback cb); bool is_pmu_core(const char *name); bool perf_pmu__supports_legacy_cache(const struct perf_pmu *pmu); diff --git a/tools/perf/util/pmus.c b/tools/perf/util/pmus.c index 16505071d362..2fd369e45832 100644 --- a/tools/perf/util/pmus.c +++ b/tools/perf/util/pmus.c @@ -16,6 +16,7 @@ #include "pmus.h" #include "pmu.h" #include "print-events.h" +#include "strbuf.h" /* * core_pmus: A PMU belongs to core_pmus if it's name is "cpu" or it's sysfs @@ -503,6 +504,99 @@ void perf_pmus__print_pmu_events(const struct print_callbacks *print_cb, void *p zfree(&aliases); } +struct build_format_string_args { + struct strbuf short_string; + struct strbuf long_string; + int num_formats; +}; + +static int build_format_string(void *state, const char *name, int config, + const unsigned long *bits) +{ + struct build_format_string_args *args = state; + unsigned int num_bits; + int ret1, ret2 = 0; + + (void)config; + args->num_formats++; + if (args->num_formats > 1) { + strbuf_addch(&args->long_string, ','); + if (args->num_formats < 4) + strbuf_addch(&args->short_string, ','); + } + num_bits = bits ? bitmap_weight(bits, PERF_PMU_FORMAT_BITS) : 0; + if (num_bits <= 1) { + ret1 = strbuf_addf(&args->long_string, "%s", name); + if (args->num_formats < 4) + ret2 = strbuf_addf(&args->short_string, "%s", name); + } else if (num_bits > 8) { + ret1 = strbuf_addf(&args->long_string, "%s=0..0x%llx", name, + ULLONG_MAX >> (64 - num_bits)); + if (args->num_formats < 4) { + ret2 = strbuf_addf(&args->short_string, "%s=0..0x%llx", name, + ULLONG_MAX >> (64 - num_bits)); + } + } else { + ret1 = strbuf_addf(&args->long_string, "%s=0..%llu", name, + ULLONG_MAX >> (64 - num_bits)); + if (args->num_formats < 4) { + ret2 = strbuf_addf(&args->short_string, "%s=0..%llu", name, + ULLONG_MAX >> (64 - num_bits)); + } + } + return ret1 < 0 ? ret1 : (ret2 < 0 ? ret2 : 0); +} + +void perf_pmus__print_raw_pmu_events(const struct print_callbacks *print_cb, void *print_state) +{ + bool skip_duplicate_pmus = print_cb->skip_duplicate_pmus(print_state); + struct perf_pmu *(*scan_fn)(struct perf_pmu *); + struct perf_pmu *pmu = NULL; + + if (skip_duplicate_pmus) + scan_fn = perf_pmus__scan_skip_duplicates; + else + scan_fn = perf_pmus__scan; + + while ((pmu = scan_fn(pmu)) != NULL) { + struct build_format_string_args format_args = { + .short_string = STRBUF_INIT, + .long_string = STRBUF_INIT, + .num_formats = 0, + }; + int len = pmu_name_len_no_suffix(pmu->name, /*num=*/NULL); + const char *desc = "(see 'man perf-list' or 'man perf-record' on how to encode it)"; + + if (!pmu->is_core) + desc = NULL; + + strbuf_addf(&format_args.short_string, "%.*s/", len, pmu->name); + strbuf_addf(&format_args.long_string, "%.*s/", len, pmu->name); + perf_pmu__for_each_format(pmu, &format_args, build_format_string); + + if (format_args.num_formats > 3) + strbuf_addf(&format_args.short_string, ",.../modifier"); + else + strbuf_addf(&format_args.short_string, "/modifier"); + + strbuf_addf(&format_args.long_string, "/modifier"); + print_cb->print_event(print_state, + /*topic=*/NULL, + /*pmu_name=*/NULL, + format_args.short_string.buf, + /*event_alias=*/NULL, + /*scale_unit=*/NULL, + /*deprecated=*/false, + "Raw event descriptor", + desc, + /*long_desc=*/NULL, + format_args.long_string.buf); + + strbuf_release(&format_args.short_string); + strbuf_release(&format_args.long_string); + } +} + bool perf_pmus__have_event(const char *pname, const char *name) { struct perf_pmu *pmu = perf_pmus__find(pname); diff --git a/tools/perf/util/pmus.h b/tools/perf/util/pmus.h index 94d2a08d894b..eec599d8aebd 100644 --- a/tools/perf/util/pmus.h +++ b/tools/perf/util/pmus.h @@ -18,6 +18,7 @@ struct perf_pmu *perf_pmus__scan_core(struct perf_pmu *pmu); const struct perf_pmu *perf_pmus__pmu_for_pmu_filter(const char *str); void perf_pmus__print_pmu_events(const struct print_callbacks *print_cb, void *print_state); +void perf_pmus__print_raw_pmu_events(const struct print_callbacks *print_cb, void *print_state); bool perf_pmus__have_event(const char *pname, const char *name); int perf_pmus__num_core_pmus(void); bool perf_pmus__supports_extended_type(void); diff --git a/tools/perf/util/print-events.c b/tools/perf/util/print-events.c index e0d2b49bab66..3f38c27f0157 100644 --- a/tools/perf/util/print-events.c +++ b/tools/perf/util/print-events.c @@ -39,7 +39,7 @@ static const char * const event_type_descriptors[] = { "Software event", "Tracepoint event", "Hardware cache event", - "Raw hardware event descriptor", + "Raw event descriptor", "Hardware breakpoint", }; @@ -416,8 +416,6 @@ void print_symbol_events(const struct print_callbacks *print_cb, void *print_sta */ void print_events(const struct print_callbacks *print_cb, void *print_state) { - char *tmp; - print_symbol_events(print_cb, print_state, PERF_TYPE_HARDWARE, event_symbols_hw, PERF_COUNT_HW_MAX); print_symbol_events(print_cb, print_state, PERF_TYPE_SOFTWARE, @@ -441,21 +439,7 @@ void print_events(const struct print_callbacks *print_cb, void *print_state) /*long_desc=*/NULL, /*encoding_desc=*/NULL); - if (asprintf(&tmp, "%s/t1=v1[,t2=v2,t3 ...]/modifier", - perf_pmus__scan_core(/*pmu=*/NULL)->name) > 0) { - print_cb->print_event(print_state, - /*topic=*/NULL, - /*pmu_name=*/NULL, - tmp, - /*event_alias=*/NULL, - /*scale_unit=*/NULL, - /*deprecated=*/false, - event_type_descriptors[PERF_TYPE_RAW], - "(see 'man perf-list' on how to encode it)", - /*long_desc=*/NULL, - /*encoding_desc=*/NULL); - free(tmp); - } + perf_pmus__print_raw_pmu_events(print_cb, print_state); print_cb->print_event(print_state, /*topic=*/NULL, -- cgit v1.2.3 From 7093882067e2e2f88d3449c35c5f0f3f566c8a26 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 7 Mar 2024 16:19:14 -0800 Subject: perf tools: Use pmus to describe type from attribute When dumping a perf_event_attr, use pmus to find the PMU and its name by the type number. This allows dynamically added PMUs to be described. Before: $ perf stat -vv -e data_read true ... perf_event_attr: type 24 size 136 config 0x20ff sample_type IDENTIFIER read_format TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING disabled 1 inherit 1 exclude_guest 1 ... After: $ perf stat -vv -e data_read true ... perf_event_attr: type 24 (uncore_imc_free_running_0) size 136 config 0x20ff sample_type IDENTIFIER read_format TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING disabled 1 inherit 1 exclude_guest 1 ... However, it also means that when we have a PMU name we prefer it to a hard coded name: Before: $ perf stat -vv -e faults true ... perf_event_attr: type 1 (PERF_TYPE_SOFTWARE) size 136 config 0x2 (PERF_COUNT_SW_PAGE_FAULTS) sample_type IDENTIFIER read_format TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING disabled 1 inherit 1 enable_on_exec 1 exclude_guest 1 ... After: $ perf stat -vv -e faults true ... perf_event_attr: type 1 (software) size 136 config 0x2 (PERF_COUNT_SW_PAGE_FAULTS) sample_type IDENTIFIER read_format TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING disabled 1 inherit 1 enable_on_exec 1 exclude_guest 1 ... It feels more consistent to do this, rather than only prefer a PMU name when a hard coded name isn't available. Signed-off-by: Ian Rogers Tested-by: Arnaldo Carvalho de Melo Tested-by: Kan Liang Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Yang Jihong Link: https://lore.kernel.org/r/20240308001915.4060155-6-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/perf_event_attr_fprintf.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/perf_event_attr_fprintf.c b/tools/perf/util/perf_event_attr_fprintf.c index 8f04d3b7f3ec..29e66835da3a 100644 --- a/tools/perf/util/perf_event_attr_fprintf.c +++ b/tools/perf/util/perf_event_attr_fprintf.c @@ -7,6 +7,8 @@ #include #include #include "util/evsel_fprintf.h" +#include "util/pmu.h" +#include "util/pmus.h" #include "trace-event.h" struct bit_names { @@ -75,9 +77,12 @@ static void __p_read_format(char *buf, size_t size, u64 value) } #define ENUM_ID_TO_STR_CASE(x) case x: return (#x); -static const char *stringify_perf_type_id(u64 value) +static const char *stringify_perf_type_id(struct perf_pmu *pmu, u32 type) { - switch (value) { + if (pmu) + return pmu->name; + + switch (type) { ENUM_ID_TO_STR_CASE(PERF_TYPE_HARDWARE) ENUM_ID_TO_STR_CASE(PERF_TYPE_SOFTWARE) ENUM_ID_TO_STR_CASE(PERF_TYPE_TRACEPOINT) @@ -175,9 +180,9 @@ do { \ #define print_id_unsigned(_s) PRINT_ID(_s, "%"PRIu64) #define print_id_hex(_s) PRINT_ID(_s, "%#"PRIx64) -static void __p_type_id(char *buf, size_t size, u64 value) +static void __p_type_id(struct perf_pmu *pmu, char *buf, size_t size, u64 value) { - print_id_unsigned(stringify_perf_type_id(value)); + print_id_unsigned(stringify_perf_type_id(pmu, value)); } static void __p_config_hw_id(char *buf, size_t size, u64 value) @@ -246,7 +251,7 @@ static void __p_config_id(char *buf, size_t size, u32 type, u64 value) #define p_sample_type(val) __p_sample_type(buf, BUF_SIZE, val) #define p_branch_sample_type(val) __p_branch_sample_type(buf, BUF_SIZE, val) #define p_read_format(val) __p_read_format(buf, BUF_SIZE, val) -#define p_type_id(val) __p_type_id(buf, BUF_SIZE, val) +#define p_type_id(val) __p_type_id(pmu, buf, BUF_SIZE, val) #define p_config_id(val) __p_config_id(buf, BUF_SIZE, attr->type, val) #define PRINT_ATTRn(_n, _f, _p, _a) \ @@ -262,6 +267,7 @@ do { \ int perf_event_attr__fprintf(FILE *fp, struct perf_event_attr *attr, attr__fprintf_f attr__fprintf, void *priv) { + struct perf_pmu *pmu = perf_pmus__find_by_type(attr->type); char buf[BUF_SIZE]; int ret = 0; -- cgit v1.2.3 From 67ee8e71daabb8632931b7559e5c8a4b69a427f8 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 7 Mar 2024 16:19:15 -0800 Subject: perf tools: Add/use PMU reverse lookup from config to name Add perf_pmu__name_from_config that does a reverse lookup from a config number to an alias name. The lookup is expensive as the config is computed for every alias by filling in a perf_event_attr, but this is only done when verbose output is enabled. The lookup also only considers config, and not config1, config2 or config3. An example of the output: $ perf stat -vv -e data_read true ... perf_event_attr: type 24 (uncore_imc_free_running_0) size 136 config 0x20ff (data_read) sample_type IDENTIFIER read_format TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING disabled 1 inherit 1 exclude_guest 1 ... Committer notes: Fix the python binding build by adding dummies for not strictly needed perf_pmu__name_from_config() and perf_pmus__find_by_type(). Signed-off-by: Ian Rogers Tested-by: Arnaldo Carvalho de Melo Tested-by: Kan Liang Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Yang Jihong Link: https://lore.kernel.org/r/20240308001915.4060155-7-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/perf_event_attr_fprintf.c | 10 ++++++++-- tools/perf/util/pmu.c | 18 ++++++++++++++++++ tools/perf/util/pmu.h | 1 + tools/perf/util/python.c | 10 ++++++++++ 4 files changed, 37 insertions(+), 2 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/perf_event_attr_fprintf.c b/tools/perf/util/perf_event_attr_fprintf.c index 29e66835da3a..59fbbba79697 100644 --- a/tools/perf/util/perf_event_attr_fprintf.c +++ b/tools/perf/util/perf_event_attr_fprintf.c @@ -222,8 +222,14 @@ static void __p_config_tracepoint_id(char *buf, size_t size, u64 value) } #endif -static void __p_config_id(char *buf, size_t size, u32 type, u64 value) +static void __p_config_id(struct perf_pmu *pmu, char *buf, size_t size, u32 type, u64 value) { + const char *name = perf_pmu__name_from_config(pmu, value); + + if (name) { + print_id_hex(name); + return; + } switch (type) { case PERF_TYPE_HARDWARE: return __p_config_hw_id(buf, size, value); @@ -252,7 +258,7 @@ static void __p_config_id(char *buf, size_t size, u32 type, u64 value) #define p_branch_sample_type(val) __p_branch_sample_type(buf, BUF_SIZE, val) #define p_read_format(val) __p_read_format(buf, BUF_SIZE, val) #define p_type_id(val) __p_type_id(pmu, buf, BUF_SIZE, val) -#define p_config_id(val) __p_config_id(buf, BUF_SIZE, attr->type, val) +#define p_config_id(val) __p_config_id(pmu, buf, BUF_SIZE, attr->type, val) #define PRINT_ATTRn(_n, _f, _p, _a) \ do { \ diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index 81952c6cd3c6..ab30f22eaf10 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -2146,3 +2146,21 @@ void perf_pmu__delete(struct perf_pmu *pmu) zfree(&pmu->id); free(pmu); } + +const char *perf_pmu__name_from_config(struct perf_pmu *pmu, u64 config) +{ + struct perf_pmu_alias *event; + + if (!pmu) + return NULL; + + pmu_add_cpu_aliases(pmu); + list_for_each_entry(event, &pmu->aliases, list) { + struct perf_event_attr attr = {.config = 0,}; + int ret = perf_pmu__config(pmu, &attr, &event->terms, NULL); + + if (ret == 0 && config == attr.config) + return event->name; + } + return NULL; +} diff --git a/tools/perf/util/pmu.h b/tools/perf/util/pmu.h index 9f5284b29ecf..152700f78455 100644 --- a/tools/perf/util/pmu.h +++ b/tools/perf/util/pmu.h @@ -276,5 +276,6 @@ struct perf_pmu *perf_pmu__lookup(struct list_head *pmus, int dirfd, const char struct perf_pmu *perf_pmu__create_placeholder_core_pmu(struct list_head *core_pmus); void perf_pmu__delete(struct perf_pmu *pmu); struct perf_pmu *perf_pmus__find_core_pmu(void); +const char *perf_pmu__name_from_config(struct perf_pmu *pmu, u64 config); #endif /* __PMU_H */ diff --git a/tools/perf/util/python.c b/tools/perf/util/python.c index 075c0f79b1b9..0aeb97c11c03 100644 --- a/tools/perf/util/python.c +++ b/tools/perf/util/python.c @@ -103,6 +103,16 @@ int perf_pmu__scan_file(const struct perf_pmu *pmu, const char *name, const char return EOF; } +const char *perf_pmu__name_from_config(struct perf_pmu *pmu __maybe_unused, u64 config __maybe_unused) +{ + return NULL; +} + +struct perf_pmu *perf_pmus__find_by_type(unsigned int type __maybe_unused) +{ + return NULL; +} + int perf_pmus__num_core_pmus(void) { return 1; -- cgit v1.2.3 From f664d5159de275d3453ee5699b3376575c6aa156 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Fri, 1 Mar 2024 12:13:05 -0800 Subject: perf tools: Suggest inbuilt commands for unknown command The existing unknown command code looks for perf scripts like perf-archive.sh and perf-iostat.sh, however, inbuilt commands aren't suggested. Add the inbuilt commands so they may be suggested too. Before: $ perf reccord perf: 'reccord' is not a perf-command. See 'perf --help'. $ After: $ perf reccord perf: 'reccord' is not a perf-command. See 'perf --help'. Did you mean this? record $ Signed-off-by: Ian Rogers Tested-by: Arnaldo Carvalho de Melo Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240301201306.2680986-1-irogers@google.com [ Added some fixes from Ian to problems I noticed while testing ] Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin.h | 4 +++- tools/perf/perf.c | 23 ++++++++++++++----- tools/perf/util/help-unknown-cmd.c | 45 ++++++++++++++++++-------------------- 3 files changed, 41 insertions(+), 31 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/builtin.h b/tools/perf/builtin.h index f2ab5bae2150..f4375deabfa3 100644 --- a/tools/perf/builtin.h +++ b/tools/perf/builtin.h @@ -2,8 +2,10 @@ #ifndef BUILTIN_H #define BUILTIN_H +struct cmdnames; + void list_common_cmds_help(void); -const char *help_unknown_cmd(const char *cmd); +const char *help_unknown_cmd(const char *cmd, struct cmdnames *main_cmds); int cmd_annotate(int argc, const char **argv); int cmd_bench(int argc, const char **argv); diff --git a/tools/perf/perf.c b/tools/perf/perf.c index 921bee0a6437..bd3f80b5bb46 100644 --- a/tools/perf/perf.c +++ b/tools/perf/perf.c @@ -18,6 +18,7 @@ #include #include "util/parse-events.h" #include +#include #include "util/debug.h" #include "util/event.h" #include "util/util.h" // usage() @@ -458,7 +459,7 @@ static int libperf_print(enum libperf_print_level level, int main(int argc, const char **argv) { - int err; + int err, done_help = 0; const char *cmd; char sbuf[STRERR_BUFSIZE]; @@ -557,22 +558,32 @@ int main(int argc, const char **argv) pthread__block_sigwinch(); while (1) { - static int done_help; - run_argv(&argc, &argv); if (errno != ENOENT) break; if (!done_help) { - cmd = argv[0] = help_unknown_cmd(cmd); + struct cmdnames main_cmds = {}; + + for (unsigned int i = 0; i < ARRAY_SIZE(commands); i++) { + add_cmdname(&main_cmds, + commands[i].cmd, + strlen(commands[i].cmd)); + } + cmd = argv[0] = help_unknown_cmd(cmd, &main_cmds); + clean_cmdnames(&main_cmds); done_help = 1; + if (!cmd) + break; } else break; } - fprintf(stderr, "Failed to run command '%s': %s\n", - cmd, str_error_r(errno, sbuf, sizeof(sbuf))); + if (cmd) { + fprintf(stderr, "Failed to run command '%s': %s\n", + cmd, str_error_r(errno, sbuf, sizeof(sbuf))); + } out: if (debug_fp) fclose(debug_fp); diff --git a/tools/perf/util/help-unknown-cmd.c b/tools/perf/util/help-unknown-cmd.c index eab99ea6ac01..2ba3369f1620 100644 --- a/tools/perf/util/help-unknown-cmd.c +++ b/tools/perf/util/help-unknown-cmd.c @@ -52,46 +52,44 @@ static int add_cmd_list(struct cmdnames *cmds, struct cmdnames *old) return 0; } -const char *help_unknown_cmd(const char *cmd) +const char *help_unknown_cmd(const char *cmd, struct cmdnames *main_cmds) { unsigned int i, n = 0, best_similarity = 0; - struct cmdnames main_cmds, other_cmds; + struct cmdnames other_cmds; - memset(&main_cmds, 0, sizeof(main_cmds)); - memset(&other_cmds, 0, sizeof(main_cmds)); + memset(&other_cmds, 0, sizeof(other_cmds)); perf_config(perf_unknown_cmd_config, NULL); - load_command_list("perf-", &main_cmds, &other_cmds); + load_command_list("perf-", main_cmds, &other_cmds); - if (add_cmd_list(&main_cmds, &other_cmds) < 0) { + if (add_cmd_list(main_cmds, &other_cmds) < 0) { fprintf(stderr, "ERROR: Failed to allocate command list for unknown command.\n"); goto end; } - qsort(main_cmds.names, main_cmds.cnt, - sizeof(main_cmds.names), cmdname_compare); - uniq(&main_cmds); + qsort(main_cmds->names, main_cmds->cnt, + sizeof(main_cmds->names), cmdname_compare); + uniq(main_cmds); - if (main_cmds.cnt) { + if (main_cmds->cnt) { /* This reuses cmdname->len for similarity index */ - for (i = 0; i < main_cmds.cnt; ++i) - main_cmds.names[i]->len = - levenshtein(cmd, main_cmds.names[i]->name, 0, 2, 1, 4); + for (i = 0; i < main_cmds->cnt; ++i) + main_cmds->names[i]->len = + levenshtein(cmd, main_cmds->names[i]->name, 0, 2, 1, 4); - qsort(main_cmds.names, main_cmds.cnt, - sizeof(*main_cmds.names), levenshtein_compare); + qsort(main_cmds->names, main_cmds->cnt, + sizeof(*main_cmds->names), levenshtein_compare); - best_similarity = main_cmds.names[0]->len; + best_similarity = main_cmds->names[0]->len; n = 1; - while (n < main_cmds.cnt && best_similarity == main_cmds.names[n]->len) + while (n < main_cmds->cnt && best_similarity == main_cmds->names[n]->len) ++n; } if (autocorrect && n == 1) { - const char *assumed = main_cmds.names[0]->name; + const char *assumed = main_cmds->names[0]->name; - main_cmds.names[0] = NULL; - clean_cmdnames(&main_cmds); + main_cmds->names[0] = NULL; clean_cmdnames(&other_cmds); fprintf(stderr, "WARNING: You called a perf program named '%s', " "which does not exist.\n" @@ -107,15 +105,14 @@ const char *help_unknown_cmd(const char *cmd) fprintf(stderr, "perf: '%s' is not a perf-command. See 'perf --help'.\n", cmd); - if (main_cmds.cnt && best_similarity < 6) { + if (main_cmds->cnt && best_similarity < 6) { fprintf(stderr, "\nDid you mean %s?\n", n < 2 ? "this": "one of these"); for (i = 0; i < n; i++) - fprintf(stderr, "\t%s\n", main_cmds.names[i]->name); + fprintf(stderr, "\t%s\n", main_cmds->names[i]->name); } end: - clean_cmdnames(&main_cmds); clean_cmdnames(&other_cmds); - exit(1); + return NULL; } -- cgit v1.2.3 From 7aea01eaf4f301bd652a25d543bcba6c23cb4b85 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Fri, 1 Mar 2024 12:13:06 -0800 Subject: perf help: Lower levenshtein penality for deleting character The levenshtein penalty for deleting a character was far higher than subsituting or inserting a character. Lower the penalty to match that of inserting a character. Before: $ perf recccord perf: 'recccord' is not a perf-command. See 'perf --help'. $ After: $ perf recccord perf: 'recccord' is not a perf-command. See 'perf --help'. Did you mean this? record $ Signed-off-by: Ian Rogers Tested-by: Arnaldo Carvalho de Melo Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240301201306.2680986-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/help-unknown-cmd.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/help-unknown-cmd.c b/tools/perf/util/help-unknown-cmd.c index 2ba3369f1620..a0a46e34f8d1 100644 --- a/tools/perf/util/help-unknown-cmd.c +++ b/tools/perf/util/help-unknown-cmd.c @@ -73,10 +73,14 @@ const char *help_unknown_cmd(const char *cmd, struct cmdnames *main_cmds) if (main_cmds->cnt) { /* This reuses cmdname->len for similarity index */ - for (i = 0; i < main_cmds->cnt; ++i) + for (i = 0; i < main_cmds->cnt; ++i) { main_cmds->names[i]->len = - levenshtein(cmd, main_cmds->names[i]->name, 0, 2, 1, 4); - + levenshtein(cmd, main_cmds->names[i]->name, + /*swap_penalty=*/0, + /*substition_penality=*/2, + /*insertion_penality=*/1, + /*deletion_penalty=*/1); + } qsort(main_cmds->names, main_cmds->cnt, sizeof(*main_cmds->names), levenshtein_compare); -- cgit v1.2.3 From 581037151910126a7934e369e4b6ac70eda9a703 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 21 Mar 2024 11:13:30 -0300 Subject: perf probe: Add missing libgen.h header needed for using basename() This prototype is obtained indirectly, by luck, from some other header in probe-event.c in most systems, but recently exploded on alpine:edge: 8 13.39 alpine:edge : FAIL gcc version 13.2.1 20240309 (Alpine 13.2.1_git20240309) util/probe-event.c: In function 'convert_exec_to_group': util/probe-event.c:225:16: error: implicit declaration of function 'basename' [-Werror=implicit-function-declaration] 225 | ptr1 = basename(exec_copy); | ^~~~~~~~ util/probe-event.c:225:14: error: assignment to 'char *' from 'int' makes pointer from integer without a cast [-Werror=int-conversion] 225 | ptr1 = basename(exec_copy); | ^ cc1: all warnings being treated as errors make[3]: *** [/git/perf-6.8.0/tools/build/Makefile.build:158: util] Error 2 Fix it by adding the libgen.h header where basename() is prototyped. Fixes: fb7345bbf7fad9bf ("perf probe: Support basic dwarf-based operations on uprobe events") Cc: Masami Hiramatsu Cc: Adrian Hunter Cc: Ian Rogers Cc: Jiri Olsa Cc: Namhyung Kim Link: https://lore.kernel.org/lkml/ Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/probe-event.c | 1 + 1 file changed, 1 insertion(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index 2a0ad9ecf0a2..5c12459e9765 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -11,6 +11,7 @@ #include #include #include +#include #include #include #include -- cgit v1.2.3 From 2a5049b75d22c971e73501784f10548c1d69c407 Mon Sep 17 00:00:00 2001 From: Anne Macedo Date: Tue, 19 Mar 2024 14:36:26 +0000 Subject: perf lock contention: Trim backtrace by skipping traceiter functions The 'perf lock contention' program currently shows the caller of the locks as __traceiter_contention_begin+0x??. This caller can be ignored, as it is from the traceiter itself. Instead, it should show the real callers for the locks. When fiddling with the --stack-skip parameter, the actual callers for the locks start to show up. However, just ignore the __traceiter_contention_begin and the __traceiter_contention_end symbols so the actual callers will show up. Before this patch is applied: sudo perf lock con -a -b -- sleep 3 contended total wait max wait avg wait type caller 8 2.33 s 2.28 s 291.18 ms rwlock:W __traceiter_contention_begin+0x44 4 2.33 s 2.28 s 582.35 ms rwlock:W __traceiter_contention_begin+0x44 7 140.30 ms 46.77 ms 20.04 ms rwlock:W __traceiter_contention_begin+0x44 2 63.35 ms 33.76 ms 31.68 ms mutex trace_contention_begin+0x84 2 46.74 ms 46.73 ms 23.37 ms rwlock:W __traceiter_contention_begin+0x44 1 13.54 us 13.54 us 13.54 us mutex trace_contention_begin+0x84 1 3.67 us 3.67 us 3.67 us rwsem:R __traceiter_contention_begin+0x44 Before this patch is applied - using --stack-skip 5 sudo perf lock con --stack-skip 5 -a -b -- sleep 3 contended total wait max wait avg wait type caller 2 2.24 s 2.24 s 1.12 s rwlock:W do_epoll_wait+0x5a0 4 1.65 s 824.21 ms 412.08 ms rwlock:W do_exit+0x338 2 824.35 ms 824.29 ms 412.17 ms spinlock get_signal+0x108 2 824.14 ms 824.14 ms 412.07 ms rwlock:W release_task+0x68 1 25.22 ms 25.22 ms 25.22 ms mutex cgroup_kn_lock_live+0x58 1 24.71 us 24.71 us 24.71 us spinlock do_exit+0x44 1 22.04 us 22.04 us 22.04 us rwsem:R lock_mm_and_find_vma+0xb0 After this patch is applied: sudo ./perf lock con -a -b -- sleep 3 contended total wait max wait avg wait type caller 4 4.13 s 2.07 s 1.03 s rwlock:W release_task+0x68 2 2.07 s 2.07 s 1.03 s rwlock:R mm_update_next_owner+0x50 2 2.07 s 2.07 s 1.03 s rwlock:W do_exit+0x338 1 41.56 ms 41.56 ms 41.56 ms mutex cgroup_kn_lock_live+0x58 2 36.12 us 18.83 us 18.06 us rwlock:W do_exit+0x338 Signed-off-by: Anne Macedo Acked-by: Namhyung Kim Tested-by: Arnaldo Carvalho de Melo Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mark Rutland Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240319143629.3422590-1-retpolanne@posteo.net Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 24 ++++++++++++++++++++++++ tools/perf/util/machine.h | 2 +- 2 files changed, 25 insertions(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index 527517db3182..5eb9044bc223 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -3266,6 +3266,17 @@ bool machine__is_lock_function(struct machine *machine, u64 addr) sym = machine__find_kernel_symbol_by_name(machine, "__lock_text_end", &kmap); machine->lock.text_end = map__unmap_ip(kmap, sym->start); + + sym = machine__find_kernel_symbol_by_name(machine, "__traceiter_contention_begin", &kmap); + if (sym) { + machine->traceiter.text_start = map__unmap_ip(kmap, sym->start); + machine->traceiter.text_end = map__unmap_ip(kmap, sym->end); + } + sym = machine__find_kernel_symbol_by_name(machine, "trace_contention_begin", &kmap); + if (sym) { + machine->trace.text_start = map__unmap_ip(kmap, sym->start); + machine->trace.text_end = map__unmap_ip(kmap, sym->end); + } } /* failed to get kernel symbols */ @@ -3280,5 +3291,18 @@ bool machine__is_lock_function(struct machine *machine, u64 addr) if (machine->lock.text_start <= addr && addr < machine->lock.text_end) return true; + /* traceiter functions currently don't have their own section + * but we consider them lock functions + */ + if (machine->traceiter.text_start != 0) { + if (machine->traceiter.text_start <= addr && addr < machine->traceiter.text_end) + return true; + } + + if (machine->trace.text_start != 0) { + if (machine->trace.text_start <= addr && addr < machine->trace.text_end) + return true; + } + return false; } diff --git a/tools/perf/util/machine.h b/tools/perf/util/machine.h index e28c787616fe..4312f6db6de0 100644 --- a/tools/perf/util/machine.h +++ b/tools/perf/util/machine.h @@ -49,7 +49,7 @@ struct machine { struct { u64 text_start; u64 text_end; - } sched, lock; + } sched, lock, traceiter, trace; pid_t *current_tid; size_t current_tid_sz; union { /* Tool specific area */ -- cgit v1.2.3 From b3ad832d8da583ff4237b04a1ba23cdbf8918907 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 21 Mar 2024 09:02:48 -0700 Subject: perf dso: Reorder members to save space in 'struct dso' MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Save 40 bytes and move from 8 to 7 cache lines. Make member dwfl dependent on being a powerpc build. Squeeze bits of int/enum types when appropriate. Remove holes/padding by reordering variables. Before: struct dso { struct mutex lock; /* 0 40 */ struct list_head node; /* 40 16 */ struct rb_node rb_node __attribute__((__aligned__(8))); /* 56 24 */ /* --- cacheline 1 boundary (64 bytes) was 16 bytes ago --- */ struct rb_root * root; /* 80 8 */ struct rb_root_cached symbols; /* 88 16 */ struct symbol * * symbol_names; /* 104 8 */ size_t symbol_names_len; /* 112 8 */ struct rb_root_cached inlined_nodes; /* 120 16 */ /* --- cacheline 2 boundary (128 bytes) was 8 bytes ago --- */ struct rb_root_cached srclines; /* 136 16 */ struct { u64 addr; /* 152 8 */ struct symbol * symbol; /* 160 8 */ } last_find_result; /* 152 16 */ void * a2l; /* 168 8 */ char * symsrc_filename; /* 176 8 */ unsigned int a2l_fails; /* 184 4 */ enum dso_space_type kernel; /* 188 4 */ /* --- cacheline 3 boundary (192 bytes) --- */ _Bool is_kmod; /* 192 1 */ /* XXX 3 bytes hole, try to pack */ enum dso_swap_type needs_swap; /* 196 4 */ enum dso_binary_type symtab_type; /* 200 4 */ enum dso_binary_type binary_type; /* 204 4 */ enum dso_load_errno load_errno; /* 208 4 */ u8 adjust_symbols:1; /* 212: 0 1 */ u8 has_build_id:1; /* 212: 1 1 */ u8 header_build_id:1; /* 212: 2 1 */ u8 has_srcline:1; /* 212: 3 1 */ u8 hit:1; /* 212: 4 1 */ u8 annotate_warned:1; /* 212: 5 1 */ u8 auxtrace_warned:1; /* 212: 6 1 */ u8 short_name_allocated:1; /* 212: 7 1 */ u8 long_name_allocated:1; /* 213: 0 1 */ u8 is_64_bit:1; /* 213: 1 1 */ /* XXX 6 bits hole, try to pack */ _Bool sorted_by_name; /* 214 1 */ _Bool loaded; /* 215 1 */ u8 rel; /* 216 1 */ /* XXX 7 bytes hole, try to pack */ struct build_id bid; /* 224 32 */ /* --- cacheline 4 boundary (256 bytes) --- */ u64 text_offset; /* 256 8 */ u64 text_end; /* 264 8 */ const char * short_name; /* 272 8 */ const char * long_name; /* 280 8 */ u16 long_name_len; /* 288 2 */ u16 short_name_len; /* 290 2 */ /* XXX 4 bytes hole, try to pack */ void * dwfl; /* 296 8 */ struct auxtrace_cache * auxtrace_cache; /* 304 8 */ int comp; /* 312 4 */ /* XXX 4 bytes hole, try to pack */ /* --- cacheline 5 boundary (320 bytes) --- */ struct { struct rb_root cache; /* 320 8 */ int fd; /* 328 4 */ int status; /* 332 4 */ u32 status_seen; /* 336 4 */ /* XXX 4 bytes hole, try to pack */ u64 file_size; /* 344 8 */ struct list_head open_entry; /* 352 16 */ u64 elf_base_addr; /* 368 8 */ u64 debug_frame_offset; /* 376 8 */ /* --- cacheline 6 boundary (384 bytes) --- */ u64 eh_frame_hdr_addr; /* 384 8 */ u64 eh_frame_hdr_offset; /* 392 8 */ } data; /* 320 80 */ struct { u32 id; /* 400 4 */ u32 sub_id; /* 404 4 */ struct perf_env * env; /* 408 8 */ } bpf_prog; /* 400 16 */ union { void * priv; /* 416 8 */ u64 db_id; /* 416 8 */ }; /* 416 8 */ struct nsinfo * nsinfo; /* 424 8 */ struct dso_id id; /* 432 24 */ /* --- cacheline 7 boundary (448 bytes) was 8 bytes ago --- */ refcount_t refcnt; /* 456 4 */ char name[]; /* 460 0 */ /* size: 464, cachelines: 8, members: 49 */ /* sum members: 440, holes: 4, sum holes: 18 */ /* sum bitfield members: 10 bits, bit holes: 1, sum bit holes: 6 bits */ /* padding: 4 */ /* forced alignments: 1 */ /* last cacheline: 16 bytes */ } __attribute__((__aligned__(8))); After: struct dso { struct mutex lock; /* 0 40 */ struct list_head node; /* 40 16 */ struct rb_node rb_node __attribute__((__aligned__(8))); /* 56 24 */ /* --- cacheline 1 boundary (64 bytes) was 16 bytes ago --- */ struct rb_root * root; /* 80 8 */ struct rb_root_cached symbols; /* 88 16 */ struct symbol * * symbol_names; /* 104 8 */ size_t symbol_names_len; /* 112 8 */ struct rb_root_cached inlined_nodes; /* 120 16 */ /* --- cacheline 2 boundary (128 bytes) was 8 bytes ago --- */ struct rb_root_cached srclines; /* 136 16 */ struct { u64 addr; /* 152 8 */ struct symbol * symbol; /* 160 8 */ } last_find_result; /* 152 16 */ struct build_id bid; /* 168 32 */ /* --- cacheline 3 boundary (192 bytes) was 8 bytes ago --- */ u64 text_offset; /* 200 8 */ u64 text_end; /* 208 8 */ const char * short_name; /* 216 8 */ const char * long_name; /* 224 8 */ void * a2l; /* 232 8 */ char * symsrc_filename; /* 240 8 */ struct nsinfo * nsinfo; /* 248 8 */ /* --- cacheline 4 boundary (256 bytes) --- */ struct auxtrace_cache * auxtrace_cache; /* 256 8 */ union { void * priv; /* 264 8 */ u64 db_id; /* 264 8 */ }; /* 264 8 */ struct { struct perf_env * env; /* 272 8 */ u32 id; /* 280 4 */ u32 sub_id; /* 284 4 */ } bpf_prog; /* 272 16 */ struct { struct rb_root cache; /* 288 8 */ struct list_head open_entry; /* 296 16 */ u64 file_size; /* 312 8 */ /* --- cacheline 5 boundary (320 bytes) --- */ u64 elf_base_addr; /* 320 8 */ u64 debug_frame_offset; /* 328 8 */ u64 eh_frame_hdr_addr; /* 336 8 */ u64 eh_frame_hdr_offset; /* 344 8 */ int fd; /* 352 4 */ int status; /* 356 4 */ u32 status_seen; /* 360 4 */ } data; /* 288 80 */ /* XXX last struct has 4 bytes of padding */ struct dso_id id; /* 368 24 */ /* --- cacheline 6 boundary (384 bytes) was 8 bytes ago --- */ unsigned int a2l_fails; /* 392 4 */ int comp; /* 396 4 */ refcount_t refcnt; /* 400 4 */ enum dso_load_errno load_errno; /* 404 4 */ u16 long_name_len; /* 408 2 */ u16 short_name_len; /* 410 2 */ enum dso_binary_type symtab_type:8; /* 412: 0 4 */ enum dso_binary_type binary_type:8; /* 412: 8 4 */ enum dso_space_type kernel:2; /* 412:16 4 */ enum dso_swap_type needs_swap:2; /* 412:18 4 */ /* Bitfield combined with next fields */ _Bool is_kmod:1; /* 414: 4 1 */ u8 adjust_symbols:1; /* 414: 5 1 */ u8 has_build_id:1; /* 414: 6 1 */ u8 header_build_id:1; /* 414: 7 1 */ u8 has_srcline:1; /* 415: 0 1 */ u8 hit:1; /* 415: 1 1 */ u8 annotate_warned:1; /* 415: 2 1 */ u8 auxtrace_warned:1; /* 415: 3 1 */ u8 short_name_allocated:1; /* 415: 4 1 */ u8 long_name_allocated:1; /* 415: 5 1 */ u8 is_64_bit:1; /* 415: 6 1 */ /* XXX 1 bit hole, try to pack */ _Bool sorted_by_name; /* 416 1 */ _Bool loaded; /* 417 1 */ u8 rel; /* 418 1 */ char name[]; /* 419 0 */ /* size: 424, cachelines: 7, members: 48 */ /* sum members: 415 */ /* sum bitfield members: 31 bits, bit holes: 1, sum bit holes: 1 bits */ /* padding: 5 */ /* paddings: 1, sum paddings: 4 */ /* forced alignments: 1 */ /* last cacheline: 40 bytes */ } __attribute__((__aligned__(8))); Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Ahelenia ZiemiaÅ„ska Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Ben Gainey Cc: Changbin Du Cc: Chengen Du Cc: Colin Ian King Cc: Ilkka Koskinen Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kan Liang Cc: Leo Yan Cc: Li Dong Cc: Liam Howlett Cc: Mark Rutland Cc: Markus Elfring Cc: Masami Hiramatsu Cc: Miguel Ojeda Cc: Namhyung Kim Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Song Liu Cc: Sun Haiyong Cc: Yanteng Si Cc: zhaimingbing Link: https://lore.kernel.org/r/20240321160300.1635121-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dso.h | 84 +++++++++++++++++++++++++-------------------------- 1 file changed, 42 insertions(+), 42 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h index 2cdcd1e2ef8b..17dab230a2ca 100644 --- a/tools/perf/util/dso.h +++ b/tools/perf/util/dso.h @@ -161,66 +161,66 @@ struct dso { u64 addr; struct symbol *symbol; } last_find_result; - void *a2l; - char *symsrc_filename; - unsigned int a2l_fails; - enum dso_space_type kernel; - bool is_kmod; - enum dso_swap_type needs_swap; - enum dso_binary_type symtab_type; - enum dso_binary_type binary_type; - enum dso_load_errno load_errno; - u8 adjust_symbols:1; - u8 has_build_id:1; - u8 header_build_id:1; - u8 has_srcline:1; - u8 hit:1; - u8 annotate_warned:1; - u8 auxtrace_warned:1; - u8 short_name_allocated:1; - u8 long_name_allocated:1; - u8 is_64_bit:1; - bool sorted_by_name; - bool loaded; - u8 rel; struct build_id bid; u64 text_offset; u64 text_end; const char *short_name; const char *long_name; - u16 long_name_len; - u16 short_name_len; + void *a2l; + char *symsrc_filename; +#if defined(__powerpc__) void *dwfl; /* DWARF debug info */ +#endif + struct nsinfo *nsinfo; struct auxtrace_cache *auxtrace_cache; - int comp; - + union { /* Tool specific area */ + void *priv; + u64 db_id; + }; + /* bpf prog information */ + struct { + struct perf_env *env; + u32 id; + u32 sub_id; + } bpf_prog; /* dso data file */ struct { struct rb_root cache; - int fd; - int status; - u32 status_seen; - u64 file_size; struct list_head open_entry; + u64 file_size; u64 elf_base_addr; u64 debug_frame_offset; u64 eh_frame_hdr_addr; u64 eh_frame_hdr_offset; + int fd; + int status; + u32 status_seen; } data; - /* bpf prog information */ - struct { - u32 id; - u32 sub_id; - struct perf_env *env; - } bpf_prog; - - union { /* Tool specific area */ - void *priv; - u64 db_id; - }; - struct nsinfo *nsinfo; struct dso_id id; + unsigned int a2l_fails; + int comp; refcount_t refcnt; + enum dso_load_errno load_errno; + u16 long_name_len; + u16 short_name_len; + enum dso_binary_type symtab_type:8; + enum dso_binary_type binary_type:8; + enum dso_space_type kernel:2; + enum dso_swap_type needs_swap:2; + bool is_kmod:1; + u8 adjust_symbols:1; + u8 has_build_id:1; + u8 header_build_id:1; + u8 has_srcline:1; + u8 hit:1; + u8 annotate_warned:1; + u8 auxtrace_warned:1; + u8 short_name_allocated:1; + u8 long_name_allocated:1; + u8 is_64_bit:1; + bool sorted_by_name; + bool loaded; + u8 rel; char name[]; }; -- cgit v1.2.3 From 7d8296b250f2eed73f1758607926d4d258dea5d4 Mon Sep 17 00:00:00 2001 From: Alexander Lobakin Date: Wed, 27 Mar 2024 16:23:42 +0100 Subject: bitops: make BYTES_TO_BITS() treewide-available Avoid open-coding that simple expression each time by moving BYTES_TO_BITS() from the probes code to to export it to the rest of the kernel. Simplify the macro while at it. `BITS_PER_LONG / sizeof(long)` always equals to %BITS_PER_BYTE, regardless of the target architecture. Do the same for the tools ecosystem as well (incl. its version of bitops.h). The previous implementation had its implicit type of long, while the new one is int, so adjust the format literal accordingly in the perf code. Suggested-by: Andy Shevchenko Reviewed-by: Przemek Kitszel Acked-by: Yury Norov Signed-off-by: Alexander Lobakin Signed-off-by: David S. Miller --- include/linux/bitops.h | 2 ++ kernel/trace/trace_probe.c | 2 -- tools/include/linux/bitops.h | 2 ++ tools/perf/util/probe-finder.c | 4 +--- 4 files changed, 5 insertions(+), 5 deletions(-) (limited to 'tools/perf/util') diff --git a/include/linux/bitops.h b/include/linux/bitops.h index f7f5a783da2a..e0cd09eb91cd 100644 --- a/include/linux/bitops.h +++ b/include/linux/bitops.h @@ -21,6 +21,8 @@ #define BITS_TO_U32(nr) __KERNEL_DIV_ROUND_UP(nr, BITS_PER_TYPE(u32)) #define BITS_TO_BYTES(nr) __KERNEL_DIV_ROUND_UP(nr, BITS_PER_TYPE(char)) +#define BYTES_TO_BITS(nb) ((nb) * BITS_PER_BYTE) + extern unsigned int __sw_hweight8(unsigned int w); extern unsigned int __sw_hweight16(unsigned int w); extern unsigned int __sw_hweight32(unsigned int w); diff --git a/kernel/trace/trace_probe.c b/kernel/trace/trace_probe.c index dfe3ee6035ec..87337a0c8e03 100644 --- a/kernel/trace/trace_probe.c +++ b/kernel/trace/trace_probe.c @@ -1180,8 +1180,6 @@ parse_probe_arg(char *arg, const struct fetch_type *type, return ret; } -#define BYTES_TO_BITS(nb) ((BITS_PER_LONG * (nb)) / sizeof(long)) - /* Bitfield type needs to be parsed into a fetch function */ static int __parse_bitfield_probe_arg(const char *bf, const struct fetch_type *t, diff --git a/tools/include/linux/bitops.h b/tools/include/linux/bitops.h index 7319f6ced108..272f15d0e434 100644 --- a/tools/include/linux/bitops.h +++ b/tools/include/linux/bitops.h @@ -20,6 +20,8 @@ #define BITS_TO_U32(nr) DIV_ROUND_UP(nr, BITS_PER_TYPE(u32)) #define BITS_TO_BYTES(nr) DIV_ROUND_UP(nr, BITS_PER_TYPE(char)) +#define BYTES_TO_BITS(nb) ((nb) * BITS_PER_BYTE) + extern unsigned int __sw_hweight8(unsigned int w); extern unsigned int __sw_hweight16(unsigned int w); extern unsigned int __sw_hweight32(unsigned int w); diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c index c8923375e30d..630e16c54ed5 100644 --- a/tools/perf/util/probe-finder.c +++ b/tools/perf/util/probe-finder.c @@ -186,8 +186,6 @@ static_var: return ret2; } -#define BYTES_TO_BITS(nb) ((nb) * BITS_PER_LONG / sizeof(long)) - static int convert_variable_type(Dwarf_Die *vr_die, struct probe_trace_arg *tvar, const char *cast, bool user_access) @@ -217,7 +215,7 @@ static int convert_variable_type(Dwarf_Die *vr_die, total = dwarf_bytesize(vr_die); if (boffs < 0 || total < 0) return -ENOENT; - ret = snprintf(buf, 16, "b%d@%d/%zd", bsize, boffs, + ret = snprintf(buf, 16, "b%d@%d/%d", bsize, boffs, BYTES_TO_BITS(total)); goto formatted; } -- cgit v1.2.3 From ad399baa06931a62d1166d215eaad6f3b0dcd3d5 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 29 Mar 2024 14:58:08 -0700 Subject: perf annotate: Use ins__is_xxx() if possible This is to prepare separation of disasm related code. Use the public ins API instead of checking the internal data structure. Signed-off-by: Namhyung Kim Tested-by: Ian Rogers Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240329215812.537846-2-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 64e54ff1aa1d..986c499150ef 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -3665,7 +3665,7 @@ int annotate_get_insn_location(struct arch *arch, struct disasm_line *dl, struct annotated_op_loc *op_loc; int i; - if (!strcmp(dl->ins.name, "lock")) + if (ins__is_lock(&dl->ins)) ops = dl->ops.locked.ops; else ops = &dl->ops; @@ -3763,7 +3763,7 @@ static struct disasm_line *find_disasm_line(struct symbol *sym, u64 ip, * llvm-objdump places "lock" in a separate line and * in that case, we want to get the next line. */ - if (!strcmp(dl->ins.name, "lock") && + if (ins__is_lock(&dl->ins) && *dl->ops.raw == '\0' && allow_update) { ip++; continue; @@ -4093,10 +4093,10 @@ static bool process_basic_block(struct basic_block_data *bb_data, if (dl == last_dl) break; /* 'return' instruction finishes the block */ - if (dl->ins.ops == &ret_ops) + if (ins__is_ret(&dl->ins)) break; /* normal instructions are part of the basic block */ - if (dl->ins.ops != &jump_ops) + if (!ins__is_jump(&dl->ins)) continue; /* jump to a different function, tail call or return */ if (dl->ops.target.outside) -- cgit v1.2.3 From 10adbf777622e22323abbf9f7861c26deb373199 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 29 Mar 2024 14:58:09 -0700 Subject: perf annotate: Add and use ins__is_nop() Likewise, add ins__is_nop() to check if the current instruction is NOP. Signed-off-by: Namhyung Kim Tested-by: Ian Rogers Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240329215812.537846-3-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 7 ++++++- tools/perf/util/annotate.h | 1 + 2 files changed, 7 insertions(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 986c499150ef..5d0ca004dcfb 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -757,6 +757,11 @@ static struct ins_ops ret_ops = { .scnprintf = ins__raw_scnprintf, }; +bool ins__is_nop(const struct ins *ins) +{ + return ins->ops == &nop_ops; +} + bool ins__is_ret(const struct ins *ins) { return ins->ops == &ret_ops; @@ -1785,7 +1790,7 @@ static void delete_last_nop(struct symbol *sym) dl = list_entry(list->prev, struct disasm_line, al.node); if (dl->ins.ops) { - if (dl->ins.ops != &nop_ops) + if (!ins__is_nop(&dl->ins)) return; } else { if (!strstr(dl->al.line, " nop ") && diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h index 14980b65f812..98f556af637c 100644 --- a/tools/perf/util/annotate.h +++ b/tools/perf/util/annotate.h @@ -74,6 +74,7 @@ struct ins_ops { bool ins__is_jump(const struct ins *ins); bool ins__is_call(const struct ins *ins); +bool ins__is_nop(const struct ins *ins); bool ins__is_ret(const struct ins *ins); bool ins__is_lock(const struct ins *ins); int ins__scnprintf(struct ins *ins, char *bf, size_t size, struct ins_operands *ops, int max_ins_name); -- cgit v1.2.3 From 98f69a573c668a18cfda41b3976118e04521a196 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 29 Mar 2024 14:58:10 -0700 Subject: perf annotate: Split out util/disasm.c MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The util/annotate.c code has both disassembly and sample annotation related codes. Factor out the disasm part so that it can be handled more easily. No functional changes intended. Committer notes: Add missing include env.h, util.h, bpf-event.h and bpf-util.h to disasm.c, to fix things like: util/disasm.c: In function ‘symbol__disassemble_bpf’: util/disasm.c:1203:9: error: implicit declaration of function ‘perf_exe’ [-Werror=implicit-function-declaration] 1203 | perf_exe(tpath, sizeof(tpath)); | ^~~~~~~~ Signed-off-by: Namhyung Kim Tested-by: Ian Rogers Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240329215812.537846-4-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/Build | 1 + tools/perf/util/annotate.c | 1708 ++------------------------------------------ tools/perf/util/annotate.h | 60 +- tools/perf/util/disasm.c | 1591 +++++++++++++++++++++++++++++++++++++++++ tools/perf/util/disasm.h | 112 +++ 5 files changed, 1762 insertions(+), 1710 deletions(-) create mode 100644 tools/perf/util/disasm.c create mode 100644 tools/perf/util/disasm.h (limited to 'tools/perf/util') diff --git a/tools/perf/util/Build b/tools/perf/util/Build index e0a723e24503..aec5a590e349 100644 --- a/tools/perf/util/Build +++ b/tools/perf/util/Build @@ -12,6 +12,7 @@ perf-y += config.o perf-y += copyfile.o perf-y += ctype.o perf-y += db-export.o +perf-y += disasm.o perf-y += env.o perf-y += event.o perf-y += evlist.o diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 5d0ca004dcfb..b795f27f2602 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -16,6 +16,7 @@ #include "build-id.h" #include "color.h" #include "config.h" +#include "disasm.h" #include "dso.h" #include "env.h" #include "map.h" @@ -64,48 +65,6 @@ /* global annotation options */ struct annotation_options annotate_opts; -static regex_t file_lineno; - -static struct ins_ops *ins__find(struct arch *arch, const char *name); -static void ins__sort(struct arch *arch); -static int disasm_line__parse(char *line, const char **namep, char **rawp); -static int call__scnprintf(struct ins *ins, char *bf, size_t size, - struct ins_operands *ops, int max_ins_name); -static int jump__scnprintf(struct ins *ins, char *bf, size_t size, - struct ins_operands *ops, int max_ins_name); - -struct arch { - const char *name; - struct ins *instructions; - size_t nr_instructions; - size_t nr_instructions_allocated; - struct ins_ops *(*associate_instruction_ops)(struct arch *arch, const char *name); - bool sorted_instructions; - bool initialized; - const char *insn_suffix; - void *priv; - unsigned int model; - unsigned int family; - int (*init)(struct arch *arch, char *cpuid); - bool (*ins_is_fused)(struct arch *arch, const char *ins1, - const char *ins2); - struct { - char comment_char; - char skip_functions_char; - char register_char; - char memory_ref_char; - char imm_char; - } objdump; -}; - -static struct ins_ops call_ops; -static struct ins_ops dec_ops; -static struct ins_ops jump_ops; -static struct ins_ops mov_ops; -static struct ins_ops nop_ops; -static struct ins_ops lock_ops; -static struct ins_ops ret_ops; - /* Data type collection debug statistics */ struct annotated_data_stat ann_data_stat; LIST_HEAD(ann_insn_stat); @@ -125,759 +84,6 @@ struct annotated_data_type canary_type = { }, }; -static int arch__grow_instructions(struct arch *arch) -{ - struct ins *new_instructions; - size_t new_nr_allocated; - - if (arch->nr_instructions_allocated == 0 && arch->instructions) - goto grow_from_non_allocated_table; - - new_nr_allocated = arch->nr_instructions_allocated + 128; - new_instructions = realloc(arch->instructions, new_nr_allocated * sizeof(struct ins)); - if (new_instructions == NULL) - return -1; - -out_update_instructions: - arch->instructions = new_instructions; - arch->nr_instructions_allocated = new_nr_allocated; - return 0; - -grow_from_non_allocated_table: - new_nr_allocated = arch->nr_instructions + 128; - new_instructions = calloc(new_nr_allocated, sizeof(struct ins)); - if (new_instructions == NULL) - return -1; - - memcpy(new_instructions, arch->instructions, arch->nr_instructions); - goto out_update_instructions; -} - -static int arch__associate_ins_ops(struct arch* arch, const char *name, struct ins_ops *ops) -{ - struct ins *ins; - - if (arch->nr_instructions == arch->nr_instructions_allocated && - arch__grow_instructions(arch)) - return -1; - - ins = &arch->instructions[arch->nr_instructions]; - ins->name = strdup(name); - if (!ins->name) - return -1; - - ins->ops = ops; - arch->nr_instructions++; - - ins__sort(arch); - return 0; -} - -#include "arch/arc/annotate/instructions.c" -#include "arch/arm/annotate/instructions.c" -#include "arch/arm64/annotate/instructions.c" -#include "arch/csky/annotate/instructions.c" -#include "arch/loongarch/annotate/instructions.c" -#include "arch/mips/annotate/instructions.c" -#include "arch/x86/annotate/instructions.c" -#include "arch/powerpc/annotate/instructions.c" -#include "arch/riscv64/annotate/instructions.c" -#include "arch/s390/annotate/instructions.c" -#include "arch/sparc/annotate/instructions.c" - -static struct arch architectures[] = { - { - .name = "arc", - .init = arc__annotate_init, - }, - { - .name = "arm", - .init = arm__annotate_init, - }, - { - .name = "arm64", - .init = arm64__annotate_init, - }, - { - .name = "csky", - .init = csky__annotate_init, - }, - { - .name = "mips", - .init = mips__annotate_init, - .objdump = { - .comment_char = '#', - }, - }, - { - .name = "x86", - .init = x86__annotate_init, - .instructions = x86__instructions, - .nr_instructions = ARRAY_SIZE(x86__instructions), - .insn_suffix = "bwlq", - .objdump = { - .comment_char = '#', - .register_char = '%', - .memory_ref_char = '(', - .imm_char = '$', - }, - }, - { - .name = "powerpc", - .init = powerpc__annotate_init, - }, - { - .name = "riscv64", - .init = riscv64__annotate_init, - }, - { - .name = "s390", - .init = s390__annotate_init, - .objdump = { - .comment_char = '#', - }, - }, - { - .name = "sparc", - .init = sparc__annotate_init, - .objdump = { - .comment_char = '#', - }, - }, - { - .name = "loongarch", - .init = loongarch__annotate_init, - .objdump = { - .comment_char = '#', - }, - }, -}; - -static void ins__delete(struct ins_operands *ops) -{ - if (ops == NULL) - return; - zfree(&ops->source.raw); - zfree(&ops->source.name); - zfree(&ops->target.raw); - zfree(&ops->target.name); -} - -static int ins__raw_scnprintf(struct ins *ins, char *bf, size_t size, - struct ins_operands *ops, int max_ins_name) -{ - return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, ops->raw); -} - -int ins__scnprintf(struct ins *ins, char *bf, size_t size, - struct ins_operands *ops, int max_ins_name) -{ - if (ins->ops->scnprintf) - return ins->ops->scnprintf(ins, bf, size, ops, max_ins_name); - - return ins__raw_scnprintf(ins, bf, size, ops, max_ins_name); -} - -bool ins__is_fused(struct arch *arch, const char *ins1, const char *ins2) -{ - if (!arch || !arch->ins_is_fused) - return false; - - return arch->ins_is_fused(arch, ins1, ins2); -} - -static int call__parse(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms) -{ - char *endptr, *tok, *name; - struct map *map = ms->map; - struct addr_map_symbol target = { - .ms = { .map = map, }, - }; - - ops->target.addr = strtoull(ops->raw, &endptr, 16); - - name = strchr(endptr, '<'); - if (name == NULL) - goto indirect_call; - - name++; - - if (arch->objdump.skip_functions_char && - strchr(name, arch->objdump.skip_functions_char)) - return -1; - - tok = strchr(name, '>'); - if (tok == NULL) - return -1; - - *tok = '\0'; - ops->target.name = strdup(name); - *tok = '>'; - - if (ops->target.name == NULL) - return -1; -find_target: - target.addr = map__objdump_2mem(map, ops->target.addr); - - if (maps__find_ams(ms->maps, &target) == 0 && - map__rip_2objdump(target.ms.map, map__map_ip(target.ms.map, target.addr)) == ops->target.addr) - ops->target.sym = target.ms.sym; - - return 0; - -indirect_call: - tok = strchr(endptr, '*'); - if (tok != NULL) { - endptr++; - - /* Indirect call can use a non-rip register and offset: callq *0x8(%rbx). - * Do not parse such instruction. */ - if (strstr(endptr, "(%r") == NULL) - ops->target.addr = strtoull(endptr, NULL, 16); - } - goto find_target; -} - -static int call__scnprintf(struct ins *ins, char *bf, size_t size, - struct ins_operands *ops, int max_ins_name) -{ - if (ops->target.sym) - return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, ops->target.sym->name); - - if (ops->target.addr == 0) - return ins__raw_scnprintf(ins, bf, size, ops, max_ins_name); - - if (ops->target.name) - return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, ops->target.name); - - return scnprintf(bf, size, "%-*s *%" PRIx64, max_ins_name, ins->name, ops->target.addr); -} - -static struct ins_ops call_ops = { - .parse = call__parse, - .scnprintf = call__scnprintf, -}; - -bool ins__is_call(const struct ins *ins) -{ - return ins->ops == &call_ops || ins->ops == &s390_call_ops || ins->ops == &loongarch_call_ops; -} - -/* - * Prevents from matching commas in the comment section, e.g.: - * ffff200008446e70: b.cs ffff2000084470f4 // b.hs, b.nlast - * - * and skip comma as part of function arguments, e.g.: - * 1d8b4ac - */ -static inline const char *validate_comma(const char *c, struct ins_operands *ops) -{ - if (ops->jump.raw_comment && c > ops->jump.raw_comment) - return NULL; - - if (ops->jump.raw_func_start && c > ops->jump.raw_func_start) - return NULL; - - return c; -} - -static int jump__parse(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms) -{ - struct map *map = ms->map; - struct symbol *sym = ms->sym; - struct addr_map_symbol target = { - .ms = { .map = map, }, - }; - const char *c = strchr(ops->raw, ','); - u64 start, end; - - ops->jump.raw_comment = strchr(ops->raw, arch->objdump.comment_char); - ops->jump.raw_func_start = strchr(ops->raw, '<'); - - c = validate_comma(c, ops); - - /* - * Examples of lines to parse for the _cpp_lex_token@@Base - * function: - * - * 1159e6c: jne 115aa32 <_cpp_lex_token@@Base+0xf92> - * 1159e8b: jne c469be - * - * The first is a jump to an offset inside the same function, - * the second is to another function, i.e. that 0xa72 is an - * offset in the cpp_named_operator2name@@base function. - */ - /* - * skip over possible up to 2 operands to get to address, e.g.: - * tbnz w0, #26, ffff0000083cd190 - */ - if (c++ != NULL) { - ops->target.addr = strtoull(c, NULL, 16); - if (!ops->target.addr) { - c = strchr(c, ','); - c = validate_comma(c, ops); - if (c++ != NULL) - ops->target.addr = strtoull(c, NULL, 16); - } - } else { - ops->target.addr = strtoull(ops->raw, NULL, 16); - } - - target.addr = map__objdump_2mem(map, ops->target.addr); - start = map__unmap_ip(map, sym->start); - end = map__unmap_ip(map, sym->end); - - ops->target.outside = target.addr < start || target.addr > end; - - /* - * FIXME: things like this in _cpp_lex_token (gcc's cc1 program): - - cpp_named_operator2name@@Base+0xa72 - - * Point to a place that is after the cpp_named_operator2name - * boundaries, i.e. in the ELF symbol table for cc1 - * cpp_named_operator2name is marked as being 32-bytes long, but it in - * fact is much larger than that, so we seem to need a symbols__find() - * routine that looks for >= current->start and < next_symbol->start, - * possibly just for C++ objects? - * - * For now lets just make some progress by marking jumps to outside the - * current function as call like. - * - * Actual navigation will come next, with further understanding of how - * the symbol searching and disassembly should be done. - */ - if (maps__find_ams(ms->maps, &target) == 0 && - map__rip_2objdump(target.ms.map, map__map_ip(target.ms.map, target.addr)) == ops->target.addr) - ops->target.sym = target.ms.sym; - - if (!ops->target.outside) { - ops->target.offset = target.addr - start; - ops->target.offset_avail = true; - } else { - ops->target.offset_avail = false; - } - - return 0; -} - -static int jump__scnprintf(struct ins *ins, char *bf, size_t size, - struct ins_operands *ops, int max_ins_name) -{ - const char *c; - - if (!ops->target.addr || ops->target.offset < 0) - return ins__raw_scnprintf(ins, bf, size, ops, max_ins_name); - - if (ops->target.outside && ops->target.sym != NULL) - return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, ops->target.sym->name); - - c = strchr(ops->raw, ','); - c = validate_comma(c, ops); - - if (c != NULL) { - const char *c2 = strchr(c + 1, ','); - - c2 = validate_comma(c2, ops); - /* check for 3-op insn */ - if (c2 != NULL) - c = c2; - c++; - - /* mirror arch objdump's space-after-comma style */ - if (*c == ' ') - c++; - } - - return scnprintf(bf, size, "%-*s %.*s%" PRIx64, max_ins_name, - ins->name, c ? c - ops->raw : 0, ops->raw, - ops->target.offset); -} - -static void jump__delete(struct ins_operands *ops __maybe_unused) -{ - /* - * The ops->jump.raw_comment and ops->jump.raw_func_start belong to the - * raw string, don't free them. - */ -} - -static struct ins_ops jump_ops = { - .free = jump__delete, - .parse = jump__parse, - .scnprintf = jump__scnprintf, -}; - -bool ins__is_jump(const struct ins *ins) -{ - return ins->ops == &jump_ops || ins->ops == &loongarch_jump_ops; -} - -static int comment__symbol(char *raw, char *comment, u64 *addrp, char **namep) -{ - char *endptr, *name, *t; - - if (strstr(raw, "(%rip)") == NULL) - return 0; - - *addrp = strtoull(comment, &endptr, 16); - if (endptr == comment) - return 0; - name = strchr(endptr, '<'); - if (name == NULL) - return -1; - - name++; - - t = strchr(name, '>'); - if (t == NULL) - return 0; - - *t = '\0'; - *namep = strdup(name); - *t = '>'; - - return 0; -} - -static int lock__parse(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms) -{ - ops->locked.ops = zalloc(sizeof(*ops->locked.ops)); - if (ops->locked.ops == NULL) - return 0; - - if (disasm_line__parse(ops->raw, &ops->locked.ins.name, &ops->locked.ops->raw) < 0) - goto out_free_ops; - - ops->locked.ins.ops = ins__find(arch, ops->locked.ins.name); - - if (ops->locked.ins.ops == NULL) - goto out_free_ops; - - if (ops->locked.ins.ops->parse && - ops->locked.ins.ops->parse(arch, ops->locked.ops, ms) < 0) - goto out_free_ops; - - return 0; - -out_free_ops: - zfree(&ops->locked.ops); - return 0; -} - -static int lock__scnprintf(struct ins *ins, char *bf, size_t size, - struct ins_operands *ops, int max_ins_name) -{ - int printed; - - if (ops->locked.ins.ops == NULL) - return ins__raw_scnprintf(ins, bf, size, ops, max_ins_name); - - printed = scnprintf(bf, size, "%-*s ", max_ins_name, ins->name); - return printed + ins__scnprintf(&ops->locked.ins, bf + printed, - size - printed, ops->locked.ops, max_ins_name); -} - -static void lock__delete(struct ins_operands *ops) -{ - struct ins *ins = &ops->locked.ins; - - if (ins->ops && ins->ops->free) - ins->ops->free(ops->locked.ops); - else - ins__delete(ops->locked.ops); - - zfree(&ops->locked.ops); - zfree(&ops->target.raw); - zfree(&ops->target.name); -} - -static struct ins_ops lock_ops = { - .free = lock__delete, - .parse = lock__parse, - .scnprintf = lock__scnprintf, -}; - -/* - * Check if the operand has more than one registers like x86 SIB addressing: - * 0x1234(%rax, %rbx, 8) - * - * But it doesn't care segment selectors like %gs:0x5678(%rcx), so just check - * the input string after 'memory_ref_char' if exists. - */ -static bool check_multi_regs(struct arch *arch, const char *op) -{ - int count = 0; - - if (arch->objdump.register_char == 0) - return false; - - if (arch->objdump.memory_ref_char) { - op = strchr(op, arch->objdump.memory_ref_char); - if (op == NULL) - return false; - } - - while ((op = strchr(op, arch->objdump.register_char)) != NULL) { - count++; - op++; - } - - return count > 1; -} - -static int mov__parse(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms __maybe_unused) -{ - char *s = strchr(ops->raw, ','), *target, *comment, prev; - - if (s == NULL) - return -1; - - *s = '\0'; - - /* - * x86 SIB addressing has something like 0x8(%rax, %rcx, 1) - * then it needs to have the closing parenthesis. - */ - if (strchr(ops->raw, '(')) { - *s = ','; - s = strchr(ops->raw, ')'); - if (s == NULL || s[1] != ',') - return -1; - *++s = '\0'; - } - - ops->source.raw = strdup(ops->raw); - *s = ','; - - if (ops->source.raw == NULL) - return -1; - - ops->source.multi_regs = check_multi_regs(arch, ops->source.raw); - - target = skip_spaces(++s); - comment = strchr(s, arch->objdump.comment_char); - - if (comment != NULL) - s = comment - 1; - else - s = strchr(s, '\0') - 1; - - while (s > target && isspace(s[0])) - --s; - s++; - prev = *s; - *s = '\0'; - - ops->target.raw = strdup(target); - *s = prev; - - if (ops->target.raw == NULL) - goto out_free_source; - - ops->target.multi_regs = check_multi_regs(arch, ops->target.raw); - - if (comment == NULL) - return 0; - - comment = skip_spaces(comment); - comment__symbol(ops->source.raw, comment + 1, &ops->source.addr, &ops->source.name); - comment__symbol(ops->target.raw, comment + 1, &ops->target.addr, &ops->target.name); - - return 0; - -out_free_source: - zfree(&ops->source.raw); - return -1; -} - -static int mov__scnprintf(struct ins *ins, char *bf, size_t size, - struct ins_operands *ops, int max_ins_name) -{ - return scnprintf(bf, size, "%-*s %s,%s", max_ins_name, ins->name, - ops->source.name ?: ops->source.raw, - ops->target.name ?: ops->target.raw); -} - -static struct ins_ops mov_ops = { - .parse = mov__parse, - .scnprintf = mov__scnprintf, -}; - -static int dec__parse(struct arch *arch __maybe_unused, struct ins_operands *ops, struct map_symbol *ms __maybe_unused) -{ - char *target, *comment, *s, prev; - - target = s = ops->raw; - - while (s[0] != '\0' && !isspace(s[0])) - ++s; - prev = *s; - *s = '\0'; - - ops->target.raw = strdup(target); - *s = prev; - - if (ops->target.raw == NULL) - return -1; - - comment = strchr(s, arch->objdump.comment_char); - if (comment == NULL) - return 0; - - comment = skip_spaces(comment); - comment__symbol(ops->target.raw, comment + 1, &ops->target.addr, &ops->target.name); - - return 0; -} - -static int dec__scnprintf(struct ins *ins, char *bf, size_t size, - struct ins_operands *ops, int max_ins_name) -{ - return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, - ops->target.name ?: ops->target.raw); -} - -static struct ins_ops dec_ops = { - .parse = dec__parse, - .scnprintf = dec__scnprintf, -}; - -static int nop__scnprintf(struct ins *ins __maybe_unused, char *bf, size_t size, - struct ins_operands *ops __maybe_unused, int max_ins_name) -{ - return scnprintf(bf, size, "%-*s", max_ins_name, "nop"); -} - -static struct ins_ops nop_ops = { - .scnprintf = nop__scnprintf, -}; - -static struct ins_ops ret_ops = { - .scnprintf = ins__raw_scnprintf, -}; - -bool ins__is_nop(const struct ins *ins) -{ - return ins->ops == &nop_ops; -} - -bool ins__is_ret(const struct ins *ins) -{ - return ins->ops == &ret_ops; -} - -bool ins__is_lock(const struct ins *ins) -{ - return ins->ops == &lock_ops; -} - -static int ins__key_cmp(const void *name, const void *insp) -{ - const struct ins *ins = insp; - - return strcmp(name, ins->name); -} - -static int ins__cmp(const void *a, const void *b) -{ - const struct ins *ia = a; - const struct ins *ib = b; - - return strcmp(ia->name, ib->name); -} - -static void ins__sort(struct arch *arch) -{ - const int nmemb = arch->nr_instructions; - - qsort(arch->instructions, nmemb, sizeof(struct ins), ins__cmp); -} - -static struct ins_ops *__ins__find(struct arch *arch, const char *name) -{ - struct ins *ins; - const int nmemb = arch->nr_instructions; - - if (!arch->sorted_instructions) { - ins__sort(arch); - arch->sorted_instructions = true; - } - - ins = bsearch(name, arch->instructions, nmemb, sizeof(struct ins), ins__key_cmp); - if (ins) - return ins->ops; - - if (arch->insn_suffix) { - char tmp[32]; - char suffix; - size_t len = strlen(name); - - if (len == 0 || len >= sizeof(tmp)) - return NULL; - - suffix = name[len - 1]; - if (strchr(arch->insn_suffix, suffix) == NULL) - return NULL; - - strcpy(tmp, name); - tmp[len - 1] = '\0'; /* remove the suffix and check again */ - - ins = bsearch(tmp, arch->instructions, nmemb, sizeof(struct ins), ins__key_cmp); - } - return ins ? ins->ops : NULL; -} - -static struct ins_ops *ins__find(struct arch *arch, const char *name) -{ - struct ins_ops *ops = __ins__find(arch, name); - - if (!ops && arch->associate_instruction_ops) - ops = arch->associate_instruction_ops(arch, name); - - return ops; -} - -static int arch__key_cmp(const void *name, const void *archp) -{ - const struct arch *arch = archp; - - return strcmp(name, arch->name); -} - -static int arch__cmp(const void *a, const void *b) -{ - const struct arch *aa = a; - const struct arch *ab = b; - - return strcmp(aa->name, ab->name); -} - -static void arch__sort(void) -{ - const int nmemb = ARRAY_SIZE(architectures); - - qsort(architectures, nmemb, sizeof(struct arch), arch__cmp); -} - -static struct arch *arch__find(const char *name) -{ - const int nmemb = ARRAY_SIZE(architectures); - static bool sorted; - - if (!sorted) { - arch__sort(); - sorted = true; - } - - return bsearch(name, architectures, nmemb, sizeof(struct arch), arch__key_cmp); -} - -bool arch__is(struct arch *arch, const char *name) -{ - return !strcmp(arch->name, name); -} - /* symbol histogram: key = offset << 16 | evsel->core.idx */ static size_t sym_hist_hash(long key, void *ctx __maybe_unused) { @@ -1214,212 +420,76 @@ static void annotation__count_and_fill(struct annotation *notes, u64 start, u64 branch->cover_insn += cover_insn; } } -} - -static int annotation__compute_ipc(struct annotation *notes, size_t size) -{ - int err = 0; - s64 offset; - - if (!notes->branch || !notes->branch->cycles_hist) - return 0; - - notes->branch->total_insn = annotation__count_insn(notes, 0, size - 1); - notes->branch->hit_cycles = 0; - notes->branch->hit_insn = 0; - notes->branch->cover_insn = 0; - - annotation__lock(notes); - for (offset = size - 1; offset >= 0; --offset) { - struct cyc_hist *ch; - - ch = ¬es->branch->cycles_hist[offset]; - if (ch && ch->cycles) { - struct annotation_line *al; - - al = notes->src->offsets[offset]; - if (al && al->cycles == NULL) { - al->cycles = zalloc(sizeof(*al->cycles)); - if (al->cycles == NULL) { - err = ENOMEM; - break; - } - } - if (ch->have_start) - annotation__count_and_fill(notes, ch->start, offset, ch); - if (al && ch->num_aggr) { - al->cycles->avg = ch->cycles_aggr / ch->num_aggr; - al->cycles->max = ch->cycles_max; - al->cycles->min = ch->cycles_min; - } - } - } - - if (err) { - while (++offset < (s64)size) { - struct cyc_hist *ch = ¬es->branch->cycles_hist[offset]; - - if (ch && ch->cycles) { - struct annotation_line *al = notes->src->offsets[offset]; - if (al) - zfree(&al->cycles); - } - } - } - - annotation__unlock(notes); - return 0; -} - -int addr_map_symbol__inc_samples(struct addr_map_symbol *ams, struct perf_sample *sample, - struct evsel *evsel) -{ - return symbol__inc_addr_samples(&ams->ms, evsel, ams->al_addr, sample); -} - -int hist_entry__inc_addr_samples(struct hist_entry *he, struct perf_sample *sample, - struct evsel *evsel, u64 ip) -{ - return symbol__inc_addr_samples(&he->ms, evsel, ip, sample); -} - -static void disasm_line__init_ins(struct disasm_line *dl, struct arch *arch, struct map_symbol *ms) -{ - dl->ins.ops = ins__find(arch, dl->ins.name); - - if (!dl->ins.ops) - return; - - if (dl->ins.ops->parse && dl->ins.ops->parse(arch, &dl->ops, ms) < 0) - dl->ins.ops = NULL; -} - -static int disasm_line__parse(char *line, const char **namep, char **rawp) -{ - char tmp, *name = skip_spaces(line); - - if (name[0] == '\0') - return -1; - - *rawp = name + 1; - - while ((*rawp)[0] != '\0' && !isspace((*rawp)[0])) - ++*rawp; - - tmp = (*rawp)[0]; - (*rawp)[0] = '\0'; - *namep = strdup(name); - - if (*namep == NULL) - goto out; - - (*rawp)[0] = tmp; - *rawp = strim(*rawp); - - return 0; - -out: - return -1; -} - -struct annotate_args { - struct arch *arch; - struct map_symbol ms; - struct evsel *evsel; - struct annotation_options *options; - s64 offset; - char *line; - int line_nr; - char *fileloc; -}; - -static void annotation_line__init(struct annotation_line *al, - struct annotate_args *args, - int nr) -{ - al->offset = args->offset; - al->line = strdup(args->line); - al->line_nr = args->line_nr; - al->fileloc = args->fileloc; - al->data_nr = nr; -} - -static void annotation_line__exit(struct annotation_line *al) -{ - zfree_srcline(&al->path); - zfree(&al->line); - zfree(&al->cycles); -} - -static size_t disasm_line_size(int nr) -{ - struct annotation_line *al; - - return (sizeof(struct disasm_line) + (sizeof(al->data[0]) * nr)); -} - -/* - * Allocating the disasm annotation line data with - * following structure: - * - * ------------------------------------------- - * struct disasm_line | struct annotation_line - * ------------------------------------------- - * - * We have 'struct annotation_line' member as last member - * of 'struct disasm_line' to have an easy access. - */ -static struct disasm_line *disasm_line__new(struct annotate_args *args) +} + +static int annotation__compute_ipc(struct annotation *notes, size_t size) { - struct disasm_line *dl = NULL; - int nr = 1; + int err = 0; + s64 offset; - if (evsel__is_group_event(args->evsel)) - nr = args->evsel->core.nr_members; + if (!notes->branch || !notes->branch->cycles_hist) + return 0; - dl = zalloc(disasm_line_size(nr)); - if (!dl) - return NULL; + notes->branch->total_insn = annotation__count_insn(notes, 0, size - 1); + notes->branch->hit_cycles = 0; + notes->branch->hit_insn = 0; + notes->branch->cover_insn = 0; - annotation_line__init(&dl->al, args, nr); - if (dl->al.line == NULL) - goto out_delete; + annotation__lock(notes); + for (offset = size - 1; offset >= 0; --offset) { + struct cyc_hist *ch; - if (args->offset != -1) { - if (disasm_line__parse(dl->al.line, &dl->ins.name, &dl->ops.raw) < 0) - goto out_free_line; + ch = ¬es->branch->cycles_hist[offset]; + if (ch && ch->cycles) { + struct annotation_line *al; - disasm_line__init_ins(dl, args->arch, &args->ms); + al = notes->src->offsets[offset]; + if (al && al->cycles == NULL) { + al->cycles = zalloc(sizeof(*al->cycles)); + if (al->cycles == NULL) { + err = ENOMEM; + break; + } + } + if (ch->have_start) + annotation__count_and_fill(notes, ch->start, offset, ch); + if (al && ch->num_aggr) { + al->cycles->avg = ch->cycles_aggr / ch->num_aggr; + al->cycles->max = ch->cycles_max; + al->cycles->min = ch->cycles_min; + } + } } - return dl; + if (err) { + while (++offset < (s64)size) { + struct cyc_hist *ch = ¬es->branch->cycles_hist[offset]; -out_free_line: - zfree(&dl->al.line); -out_delete: - free(dl); - return NULL; + if (ch && ch->cycles) { + struct annotation_line *al = notes->src->offsets[offset]; + if (al) + zfree(&al->cycles); + } + } + } + + annotation__unlock(notes); + return 0; } -void disasm_line__free(struct disasm_line *dl) +int addr_map_symbol__inc_samples(struct addr_map_symbol *ams, struct perf_sample *sample, + struct evsel *evsel) { - if (dl->ins.ops && dl->ins.ops->free) - dl->ins.ops->free(&dl->ops); - else - ins__delete(&dl->ops); - zfree(&dl->ins.name); - annotation_line__exit(&dl->al); - free(dl); + return symbol__inc_addr_samples(&ams->ms, evsel, ams->al_addr, sample); } -int disasm_line__scnprintf(struct disasm_line *dl, char *bf, size_t size, bool raw, int max_ins_name) +int hist_entry__inc_addr_samples(struct hist_entry *he, struct perf_sample *sample, + struct evsel *evsel, u64 ip) { - if (raw || !dl->ins.ops) - return scnprintf(bf, size, "%-*s %s", max_ins_name, dl->ins.name, dl->ops.raw); - - return ins__scnprintf(&dl->ins, bf, size, &dl->ops, max_ins_name); + return symbol__inc_addr_samples(&he->ms, evsel, ip, sample); } + void annotation__exit(struct annotation *notes) { annotated_source__delete(notes->src); @@ -1478,8 +548,7 @@ bool annotation__trylock(struct annotation *notes) return mutex_trylock(mutex); } - -static void annotation_line__add(struct annotation_line *al, struct list_head *head) +void annotation_line__add(struct annotation_line *al, struct list_head *head) { list_add_tail(&al->node, head); } @@ -1689,673 +758,6 @@ annotation_line__print(struct annotation_line *al, struct symbol *sym, u64 start return 0; } -/* - * symbol__parse_objdump_line() parses objdump output (with -d --no-show-raw) - * which looks like following - * - * 0000000000415500 <_init>: - * 415500: sub $0x8,%rsp - * 415504: mov 0x2f5ad5(%rip),%rax # 70afe0 <_DYNAMIC+0x2f8> - * 41550b: test %rax,%rax - * 41550e: je 415515 <_init+0x15> - * 415510: callq 416e70 <__gmon_start__@plt> - * 415515: add $0x8,%rsp - * 415519: retq - * - * it will be parsed and saved into struct disasm_line as - * - * - * The offset will be a relative offset from the start of the symbol and -1 - * means that it's not a disassembly line so should be treated differently. - * The ops.raw part will be parsed further according to type of the instruction. - */ -static int symbol__parse_objdump_line(struct symbol *sym, - struct annotate_args *args, - char *parsed_line, int *line_nr, char **fileloc) -{ - struct map *map = args->ms.map; - struct annotation *notes = symbol__annotation(sym); - struct disasm_line *dl; - char *tmp; - s64 line_ip, offset = -1; - regmatch_t match[2]; - - /* /filename:linenr ? Save line number and ignore. */ - if (regexec(&file_lineno, parsed_line, 2, match, 0) == 0) { - *line_nr = atoi(parsed_line + match[1].rm_so); - free(*fileloc); - *fileloc = strdup(parsed_line); - return 0; - } - - /* Process hex address followed by ':'. */ - line_ip = strtoull(parsed_line, &tmp, 16); - if (parsed_line != tmp && tmp[0] == ':' && tmp[1] != '\0') { - u64 start = map__rip_2objdump(map, sym->start), - end = map__rip_2objdump(map, sym->end); - - offset = line_ip - start; - if ((u64)line_ip < start || (u64)line_ip >= end) - offset = -1; - else - parsed_line = tmp + 1; - } - - args->offset = offset; - args->line = parsed_line; - args->line_nr = *line_nr; - args->fileloc = *fileloc; - args->ms.sym = sym; - - dl = disasm_line__new(args); - (*line_nr)++; - - if (dl == NULL) - return -1; - - if (!disasm_line__has_local_offset(dl)) { - dl->ops.target.offset = dl->ops.target.addr - - map__rip_2objdump(map, sym->start); - dl->ops.target.offset_avail = true; - } - - /* kcore has no symbols, so add the call target symbol */ - if (dl->ins.ops && ins__is_call(&dl->ins) && !dl->ops.target.sym) { - struct addr_map_symbol target = { - .addr = dl->ops.target.addr, - .ms = { .map = map, }, - }; - - if (!maps__find_ams(args->ms.maps, &target) && - target.ms.sym->start == target.al_addr) - dl->ops.target.sym = target.ms.sym; - } - - annotation_line__add(&dl->al, ¬es->src->source); - return 0; -} - -static __attribute__((constructor)) void symbol__init_regexpr(void) -{ - regcomp(&file_lineno, "^/[^:]+:([0-9]+)", REG_EXTENDED); -} - -static void delete_last_nop(struct symbol *sym) -{ - struct annotation *notes = symbol__annotation(sym); - struct list_head *list = ¬es->src->source; - struct disasm_line *dl; - - while (!list_empty(list)) { - dl = list_entry(list->prev, struct disasm_line, al.node); - - if (dl->ins.ops) { - if (!ins__is_nop(&dl->ins)) - return; - } else { - if (!strstr(dl->al.line, " nop ") && - !strstr(dl->al.line, " nopl ") && - !strstr(dl->al.line, " nopw ")) - return; - } - - list_del_init(&dl->al.node); - disasm_line__free(dl); - } -} - -int symbol__strerror_disassemble(struct map_symbol *ms, int errnum, char *buf, size_t buflen) -{ - struct dso *dso = map__dso(ms->map); - - BUG_ON(buflen == 0); - - if (errnum >= 0) { - str_error_r(errnum, buf, buflen); - return 0; - } - - switch (errnum) { - case SYMBOL_ANNOTATE_ERRNO__NO_VMLINUX: { - char bf[SBUILD_ID_SIZE + 15] = " with build id "; - char *build_id_msg = NULL; - - if (dso->has_build_id) { - build_id__sprintf(&dso->bid, bf + 15); - build_id_msg = bf; - } - scnprintf(buf, buflen, - "No vmlinux file%s\nwas found in the path.\n\n" - "Note that annotation using /proc/kcore requires CAP_SYS_RAWIO capability.\n\n" - "Please use:\n\n" - " perf buildid-cache -vu vmlinux\n\n" - "or:\n\n" - " --vmlinux vmlinux\n", build_id_msg ?: ""); - } - break; - case SYMBOL_ANNOTATE_ERRNO__NO_LIBOPCODES_FOR_BPF: - scnprintf(buf, buflen, "Please link with binutils's libopcode to enable BPF annotation"); - break; - case SYMBOL_ANNOTATE_ERRNO__ARCH_INIT_REGEXP: - scnprintf(buf, buflen, "Problems with arch specific instruction name regular expressions."); - break; - case SYMBOL_ANNOTATE_ERRNO__ARCH_INIT_CPUID_PARSING: - scnprintf(buf, buflen, "Problems while parsing the CPUID in the arch specific initialization."); - break; - case SYMBOL_ANNOTATE_ERRNO__BPF_INVALID_FILE: - scnprintf(buf, buflen, "Invalid BPF file: %s.", dso->long_name); - break; - case SYMBOL_ANNOTATE_ERRNO__BPF_MISSING_BTF: - scnprintf(buf, buflen, "The %s BPF file has no BTF section, compile with -g or use pahole -J.", - dso->long_name); - break; - default: - scnprintf(buf, buflen, "Internal error: Invalid %d error code\n", errnum); - break; - } - - return 0; -} - -static int dso__disassemble_filename(struct dso *dso, char *filename, size_t filename_size) -{ - char linkname[PATH_MAX]; - char *build_id_filename; - char *build_id_path = NULL; - char *pos; - int len; - - if (dso->symtab_type == DSO_BINARY_TYPE__KALLSYMS && - !dso__is_kcore(dso)) - return SYMBOL_ANNOTATE_ERRNO__NO_VMLINUX; - - build_id_filename = dso__build_id_filename(dso, NULL, 0, false); - if (build_id_filename) { - __symbol__join_symfs(filename, filename_size, build_id_filename); - free(build_id_filename); - } else { - if (dso->has_build_id) - return ENOMEM; - goto fallback; - } - - build_id_path = strdup(filename); - if (!build_id_path) - return ENOMEM; - - /* - * old style build-id cache has name of XX/XXXXXXX.. while - * new style has XX/XXXXXXX../{elf,kallsyms,vdso}. - * extract the build-id part of dirname in the new style only. - */ - pos = strrchr(build_id_path, '/'); - if (pos && strlen(pos) < SBUILD_ID_SIZE - 2) - dirname(build_id_path); - - if (dso__is_kcore(dso)) - goto fallback; - - len = readlink(build_id_path, linkname, sizeof(linkname) - 1); - if (len < 0) - goto fallback; - - linkname[len] = '\0'; - if (strstr(linkname, DSO__NAME_KALLSYMS) || - access(filename, R_OK)) { -fallback: - /* - * If we don't have build-ids or the build-id file isn't in the - * cache, or is just a kallsyms file, well, lets hope that this - * DSO is the same as when 'perf record' ran. - */ - if (dso->kernel && dso->long_name[0] == '/') - snprintf(filename, filename_size, "%s", dso->long_name); - else - __symbol__join_symfs(filename, filename_size, dso->long_name); - - mutex_lock(&dso->lock); - if (access(filename, R_OK) && errno == ENOENT && dso->nsinfo) { - char *new_name = dso__filename_with_chroot(dso, filename); - if (new_name) { - strlcpy(filename, new_name, filename_size); - free(new_name); - } - } - mutex_unlock(&dso->lock); - } - - free(build_id_path); - return 0; -} - -#if defined(HAVE_LIBBFD_SUPPORT) && defined(HAVE_LIBBPF_SUPPORT) -#define PACKAGE "perf" -#include -#include -#include -#include -#include -#include -#include - -static int symbol__disassemble_bpf(struct symbol *sym, - struct annotate_args *args) -{ - struct annotation *notes = symbol__annotation(sym); - struct bpf_prog_linfo *prog_linfo = NULL; - struct bpf_prog_info_node *info_node; - int len = sym->end - sym->start; - disassembler_ftype disassemble; - struct map *map = args->ms.map; - struct perf_bpil *info_linear; - struct disassemble_info info; - struct dso *dso = map__dso(map); - int pc = 0, count, sub_id; - struct btf *btf = NULL; - char tpath[PATH_MAX]; - size_t buf_size; - int nr_skip = 0; - char *buf; - bfd *bfdf; - int ret; - FILE *s; - - if (dso->binary_type != DSO_BINARY_TYPE__BPF_PROG_INFO) - return SYMBOL_ANNOTATE_ERRNO__BPF_INVALID_FILE; - - pr_debug("%s: handling sym %s addr %" PRIx64 " len %" PRIx64 "\n", __func__, - sym->name, sym->start, sym->end - sym->start); - - memset(tpath, 0, sizeof(tpath)); - perf_exe(tpath, sizeof(tpath)); - - bfdf = bfd_openr(tpath, NULL); - if (bfdf == NULL) - abort(); - - if (!bfd_check_format(bfdf, bfd_object)) - abort(); - - s = open_memstream(&buf, &buf_size); - if (!s) { - ret = errno; - goto out; - } - init_disassemble_info_compat(&info, s, - (fprintf_ftype) fprintf, - fprintf_styled); - info.arch = bfd_get_arch(bfdf); - info.mach = bfd_get_mach(bfdf); - - info_node = perf_env__find_bpf_prog_info(dso->bpf_prog.env, - dso->bpf_prog.id); - if (!info_node) { - ret = SYMBOL_ANNOTATE_ERRNO__BPF_MISSING_BTF; - goto out; - } - info_linear = info_node->info_linear; - sub_id = dso->bpf_prog.sub_id; - - info.buffer = (void *)(uintptr_t)(info_linear->info.jited_prog_insns); - info.buffer_length = info_linear->info.jited_prog_len; - - if (info_linear->info.nr_line_info) - prog_linfo = bpf_prog_linfo__new(&info_linear->info); - - if (info_linear->info.btf_id) { - struct btf_node *node; - - node = perf_env__find_btf(dso->bpf_prog.env, - info_linear->info.btf_id); - if (node) - btf = btf__new((__u8 *)(node->data), - node->data_size); - } - - disassemble_init_for_target(&info); - -#ifdef DISASM_FOUR_ARGS_SIGNATURE - disassemble = disassembler(info.arch, - bfd_big_endian(bfdf), - info.mach, - bfdf); -#else - disassemble = disassembler(bfdf); -#endif - if (disassemble == NULL) - abort(); - - fflush(s); - do { - const struct bpf_line_info *linfo = NULL; - struct disasm_line *dl; - size_t prev_buf_size; - const char *srcline; - u64 addr; - - addr = pc + ((u64 *)(uintptr_t)(info_linear->info.jited_ksyms))[sub_id]; - count = disassemble(pc, &info); - - if (prog_linfo) - linfo = bpf_prog_linfo__lfind_addr_func(prog_linfo, - addr, sub_id, - nr_skip); - - if (linfo && btf) { - srcline = btf__name_by_offset(btf, linfo->line_off); - nr_skip++; - } else - srcline = NULL; - - fprintf(s, "\n"); - prev_buf_size = buf_size; - fflush(s); - - if (!annotate_opts.hide_src_code && srcline) { - args->offset = -1; - args->line = strdup(srcline); - args->line_nr = 0; - args->fileloc = NULL; - args->ms.sym = sym; - dl = disasm_line__new(args); - if (dl) { - annotation_line__add(&dl->al, - ¬es->src->source); - } - } - - args->offset = pc; - args->line = buf + prev_buf_size; - args->line_nr = 0; - args->fileloc = NULL; - args->ms.sym = sym; - dl = disasm_line__new(args); - if (dl) - annotation_line__add(&dl->al, ¬es->src->source); - - pc += count; - } while (count > 0 && pc < len); - - ret = 0; -out: - free(prog_linfo); - btf__free(btf); - fclose(s); - bfd_close(bfdf); - return ret; -} -#else // defined(HAVE_LIBBFD_SUPPORT) && defined(HAVE_LIBBPF_SUPPORT) -static int symbol__disassemble_bpf(struct symbol *sym __maybe_unused, - struct annotate_args *args __maybe_unused) -{ - return SYMBOL_ANNOTATE_ERRNO__NO_LIBOPCODES_FOR_BPF; -} -#endif // defined(HAVE_LIBBFD_SUPPORT) && defined(HAVE_LIBBPF_SUPPORT) - -static int -symbol__disassemble_bpf_image(struct symbol *sym, - struct annotate_args *args) -{ - struct annotation *notes = symbol__annotation(sym); - struct disasm_line *dl; - - args->offset = -1; - args->line = strdup("to be implemented"); - args->line_nr = 0; - args->fileloc = NULL; - dl = disasm_line__new(args); - if (dl) - annotation_line__add(&dl->al, ¬es->src->source); - - zfree(&args->line); - return 0; -} - -/* - * Possibly create a new version of line with tabs expanded. Returns the - * existing or new line, storage is updated if a new line is allocated. If - * allocation fails then NULL is returned. - */ -static char *expand_tabs(char *line, char **storage, size_t *storage_len) -{ - size_t i, src, dst, len, new_storage_len, num_tabs; - char *new_line; - size_t line_len = strlen(line); - - for (num_tabs = 0, i = 0; i < line_len; i++) - if (line[i] == '\t') - num_tabs++; - - if (num_tabs == 0) - return line; - - /* - * Space for the line and '\0', less the leading and trailing - * spaces. Each tab may introduce 7 additional spaces. - */ - new_storage_len = line_len + 1 + (num_tabs * 7); - - new_line = malloc(new_storage_len); - if (new_line == NULL) { - pr_err("Failure allocating memory for tab expansion\n"); - return NULL; - } - - /* - * Copy regions starting at src and expand tabs. If there are two - * adjacent tabs then 'src == i', the memcpy is of size 0 and the spaces - * are inserted. - */ - for (i = 0, src = 0, dst = 0; i < line_len && num_tabs; i++) { - if (line[i] == '\t') { - len = i - src; - memcpy(&new_line[dst], &line[src], len); - dst += len; - new_line[dst++] = ' '; - while (dst % 8 != 0) - new_line[dst++] = ' '; - src = i + 1; - num_tabs--; - } - } - - /* Expand the last region. */ - len = line_len - src; - memcpy(&new_line[dst], &line[src], len); - dst += len; - new_line[dst] = '\0'; - - free(*storage); - *storage = new_line; - *storage_len = new_storage_len; - return new_line; - -} - -static int symbol__disassemble(struct symbol *sym, struct annotate_args *args) -{ - struct annotation_options *opts = &annotate_opts; - struct map *map = args->ms.map; - struct dso *dso = map__dso(map); - char *command; - FILE *file; - char symfs_filename[PATH_MAX]; - struct kcore_extract kce; - bool delete_extract = false; - bool decomp = false; - int lineno = 0; - char *fileloc = NULL; - int nline; - char *line; - size_t line_len; - const char *objdump_argv[] = { - "/bin/sh", - "-c", - NULL, /* Will be the objdump command to run. */ - "--", - NULL, /* Will be the symfs path. */ - NULL, - }; - struct child_process objdump_process; - int err = dso__disassemble_filename(dso, symfs_filename, sizeof(symfs_filename)); - - if (err) - return err; - - pr_debug("%s: filename=%s, sym=%s, start=%#" PRIx64 ", end=%#" PRIx64 "\n", __func__, - symfs_filename, sym->name, map__unmap_ip(map, sym->start), - map__unmap_ip(map, sym->end)); - - pr_debug("annotating [%p] %30s : [%p] %30s\n", - dso, dso->long_name, sym, sym->name); - - if (dso->binary_type == DSO_BINARY_TYPE__BPF_PROG_INFO) { - return symbol__disassemble_bpf(sym, args); - } else if (dso->binary_type == DSO_BINARY_TYPE__BPF_IMAGE) { - return symbol__disassemble_bpf_image(sym, args); - } else if (dso__is_kcore(dso)) { - kce.kcore_filename = symfs_filename; - kce.addr = map__rip_2objdump(map, sym->start); - kce.offs = sym->start; - kce.len = sym->end - sym->start; - if (!kcore_extract__create(&kce)) { - delete_extract = true; - strlcpy(symfs_filename, kce.extract_filename, - sizeof(symfs_filename)); - } - } else if (dso__needs_decompress(dso)) { - char tmp[KMOD_DECOMP_LEN]; - - if (dso__decompress_kmodule_path(dso, symfs_filename, - tmp, sizeof(tmp)) < 0) - return -1; - - decomp = true; - strcpy(symfs_filename, tmp); - } - - err = asprintf(&command, - "%s %s%s --start-address=0x%016" PRIx64 - " --stop-address=0x%016" PRIx64 - " %s -d %s %s %s %c%s%c %s%s -C \"$1\"", - opts->objdump_path ?: "objdump", - opts->disassembler_style ? "-M " : "", - opts->disassembler_style ?: "", - map__rip_2objdump(map, sym->start), - map__rip_2objdump(map, sym->end), - opts->show_linenr ? "-l" : "", - opts->show_asm_raw ? "" : "--no-show-raw-insn", - opts->annotate_src ? "-S" : "", - opts->prefix ? "--prefix " : "", - opts->prefix ? '"' : ' ', - opts->prefix ?: "", - opts->prefix ? '"' : ' ', - opts->prefix_strip ? "--prefix-strip=" : "", - opts->prefix_strip ?: ""); - - if (err < 0) { - pr_err("Failure allocating memory for the command to run\n"); - goto out_remove_tmp; - } - - pr_debug("Executing: %s\n", command); - - objdump_argv[2] = command; - objdump_argv[4] = symfs_filename; - - /* Create a pipe to read from for stdout */ - memset(&objdump_process, 0, sizeof(objdump_process)); - objdump_process.argv = objdump_argv; - objdump_process.out = -1; - objdump_process.err = -1; - objdump_process.no_stderr = 1; - if (start_command(&objdump_process)) { - pr_err("Failure starting to run %s\n", command); - err = -1; - goto out_free_command; - } - - file = fdopen(objdump_process.out, "r"); - if (!file) { - pr_err("Failure creating FILE stream for %s\n", command); - /* - * If we were using debug info should retry with - * original binary. - */ - err = -1; - goto out_close_stdout; - } - - /* Storage for getline. */ - line = NULL; - line_len = 0; - - nline = 0; - while (!feof(file)) { - const char *match; - char *expanded_line; - - if (getline(&line, &line_len, file) < 0 || !line) - break; - - /* Skip lines containing "filename:" */ - match = strstr(line, symfs_filename); - if (match && match[strlen(symfs_filename)] == ':') - continue; - - expanded_line = strim(line); - expanded_line = expand_tabs(expanded_line, &line, &line_len); - if (!expanded_line) - break; - - /* - * The source code line number (lineno) needs to be kept in - * across calls to symbol__parse_objdump_line(), so that it - * can associate it with the instructions till the next one. - * See disasm_line__new() and struct disasm_line::line_nr. - */ - if (symbol__parse_objdump_line(sym, args, expanded_line, - &lineno, &fileloc) < 0) - break; - nline++; - } - free(line); - free(fileloc); - - err = finish_command(&objdump_process); - if (err) - pr_err("Error running %s\n", command); - - if (nline == 0) { - err = -1; - pr_err("No output from %s\n", command); - } - - /* - * kallsyms does not have symbol sizes so there may a nop at the end. - * Remove it. - */ - if (dso__is_kcore(dso)) - delete_last_nop(sym); - - fclose(file); - -out_close_stdout: - close(objdump_process.out); - -out_free_command: - free(command); - -out_remove_tmp: - if (decomp) - unlink(symfs_filename); - - if (delete_extract) - kcore_extract__delete(&kce); - - return err; -} - static void calc_percent(struct annotation *notes, struct evsel *evsel, struct annotation_data *data, diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h index 98f556af637c..b3007c9966fd 100644 --- a/tools/perf/util/annotate.h +++ b/tools/perf/util/annotate.h @@ -13,10 +13,10 @@ #include "mutex.h" #include "spark.h" #include "hashmap.h" +#include "disasm.h" struct hist_browser_timer; struct hist_entry; -struct ins_ops; struct map; struct map_symbol; struct addr_map_symbol; @@ -26,60 +26,6 @@ struct evsel; struct symbol; struct annotated_data_type; -struct ins { - const char *name; - struct ins_ops *ops; -}; - -struct ins_operands { - char *raw; - struct { - char *raw; - char *name; - struct symbol *sym; - u64 addr; - s64 offset; - bool offset_avail; - bool outside; - bool multi_regs; - } target; - union { - struct { - char *raw; - char *name; - u64 addr; - bool multi_regs; - } source; - struct { - struct ins ins; - struct ins_operands *ops; - } locked; - struct { - char *raw_comment; - char *raw_func_start; - } jump; - }; -}; - -struct arch; - -bool arch__is(struct arch *arch, const char *name); - -struct ins_ops { - void (*free)(struct ins_operands *ops); - int (*parse)(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms); - int (*scnprintf)(struct ins *ins, char *bf, size_t size, - struct ins_operands *ops, int max_ins_name); -}; - -bool ins__is_jump(const struct ins *ins); -bool ins__is_call(const struct ins *ins); -bool ins__is_nop(const struct ins *ins); -bool ins__is_ret(const struct ins *ins); -bool ins__is_lock(const struct ins *ins); -int ins__scnprintf(struct ins *ins, char *bf, size_t size, struct ins_operands *ops, int max_ins_name); -bool ins__is_fused(struct arch *arch, const char *ins1, const char *ins2); - #define ANNOTATION__IPC_WIDTH 6 #define ANNOTATION__CYCLES_WIDTH 6 #define ANNOTATION__MINMAX_CYCLES_WIDTH 19 @@ -172,6 +118,8 @@ struct disasm_line { struct annotation_line al; }; +void annotation_line__add(struct annotation_line *al, struct list_head *head); + static inline double annotation_data__percent(struct annotation_data *data, unsigned int which) { @@ -213,7 +161,6 @@ static inline bool disasm_line__has_local_offset(const struct disasm_line *dl) */ bool disasm_line__is_valid_local_jump(struct disasm_line *dl, struct symbol *sym); -void disasm_line__free(struct disasm_line *dl); struct annotation_line * annotation_line__next(struct annotation_line *pos, struct list_head *head); @@ -236,7 +183,6 @@ int __annotation__scnprintf_samples_period(struct annotation *notes, struct evsel *evsel, bool show_freq); -int disasm_line__scnprintf(struct disasm_line *dl, char *bf, size_t size, bool raw, int max_ins_name); size_t disasm__fprintf(struct list_head *head, FILE *fp); void symbol__calc_percent(struct symbol *sym, struct evsel *evsel); diff --git a/tools/perf/util/disasm.c b/tools/perf/util/disasm.c new file mode 100644 index 000000000000..21a43b07e8b5 --- /dev/null +++ b/tools/perf/util/disasm.c @@ -0,0 +1,1591 @@ +// SPDX-License-Identifier: GPL-2.0-only +#include +#include +#include +#include +#include +#include +#include + +#include +#include + +#include "annotate.h" +#include "build-id.h" +#include "debug.h" +#include "disasm.h" +#include "dso.h" +#include "env.h" +#include "evsel.h" +#include "map.h" +#include "maps.h" +#include "srcline.h" +#include "symbol.h" +#include "util.h" + +static regex_t file_lineno; + +/* These can be referred from the arch-dependent code */ +static struct ins_ops call_ops; +static struct ins_ops dec_ops; +static struct ins_ops jump_ops; +static struct ins_ops mov_ops; +static struct ins_ops nop_ops; +static struct ins_ops lock_ops; +static struct ins_ops ret_ops; + +static int jump__scnprintf(struct ins *ins, char *bf, size_t size, + struct ins_operands *ops, int max_ins_name); +static int call__scnprintf(struct ins *ins, char *bf, size_t size, + struct ins_operands *ops, int max_ins_name); + +static void ins__sort(struct arch *arch); +static int disasm_line__parse(char *line, const char **namep, char **rawp); + +static __attribute__((constructor)) void symbol__init_regexpr(void) +{ + regcomp(&file_lineno, "^/[^:]+:([0-9]+)", REG_EXTENDED); +} + +static int arch__grow_instructions(struct arch *arch) +{ + struct ins *new_instructions; + size_t new_nr_allocated; + + if (arch->nr_instructions_allocated == 0 && arch->instructions) + goto grow_from_non_allocated_table; + + new_nr_allocated = arch->nr_instructions_allocated + 128; + new_instructions = realloc(arch->instructions, new_nr_allocated * sizeof(struct ins)); + if (new_instructions == NULL) + return -1; + +out_update_instructions: + arch->instructions = new_instructions; + arch->nr_instructions_allocated = new_nr_allocated; + return 0; + +grow_from_non_allocated_table: + new_nr_allocated = arch->nr_instructions + 128; + new_instructions = calloc(new_nr_allocated, sizeof(struct ins)); + if (new_instructions == NULL) + return -1; + + memcpy(new_instructions, arch->instructions, arch->nr_instructions); + goto out_update_instructions; +} + +static int arch__associate_ins_ops(struct arch* arch, const char *name, struct ins_ops *ops) +{ + struct ins *ins; + + if (arch->nr_instructions == arch->nr_instructions_allocated && + arch__grow_instructions(arch)) + return -1; + + ins = &arch->instructions[arch->nr_instructions]; + ins->name = strdup(name); + if (!ins->name) + return -1; + + ins->ops = ops; + arch->nr_instructions++; + + ins__sort(arch); + return 0; +} + +#include "arch/arc/annotate/instructions.c" +#include "arch/arm/annotate/instructions.c" +#include "arch/arm64/annotate/instructions.c" +#include "arch/csky/annotate/instructions.c" +#include "arch/loongarch/annotate/instructions.c" +#include "arch/mips/annotate/instructions.c" +#include "arch/x86/annotate/instructions.c" +#include "arch/powerpc/annotate/instructions.c" +#include "arch/riscv64/annotate/instructions.c" +#include "arch/s390/annotate/instructions.c" +#include "arch/sparc/annotate/instructions.c" + +static struct arch architectures[] = { + { + .name = "arc", + .init = arc__annotate_init, + }, + { + .name = "arm", + .init = arm__annotate_init, + }, + { + .name = "arm64", + .init = arm64__annotate_init, + }, + { + .name = "csky", + .init = csky__annotate_init, + }, + { + .name = "mips", + .init = mips__annotate_init, + .objdump = { + .comment_char = '#', + }, + }, + { + .name = "x86", + .init = x86__annotate_init, + .instructions = x86__instructions, + .nr_instructions = ARRAY_SIZE(x86__instructions), + .insn_suffix = "bwlq", + .objdump = { + .comment_char = '#', + .register_char = '%', + .memory_ref_char = '(', + .imm_char = '$', + }, + }, + { + .name = "powerpc", + .init = powerpc__annotate_init, + }, + { + .name = "riscv64", + .init = riscv64__annotate_init, + }, + { + .name = "s390", + .init = s390__annotate_init, + .objdump = { + .comment_char = '#', + }, + }, + { + .name = "sparc", + .init = sparc__annotate_init, + .objdump = { + .comment_char = '#', + }, + }, + { + .name = "loongarch", + .init = loongarch__annotate_init, + .objdump = { + .comment_char = '#', + }, + }, +}; + +static int arch__key_cmp(const void *name, const void *archp) +{ + const struct arch *arch = archp; + + return strcmp(name, arch->name); +} + +static int arch__cmp(const void *a, const void *b) +{ + const struct arch *aa = a; + const struct arch *ab = b; + + return strcmp(aa->name, ab->name); +} + +static void arch__sort(void) +{ + const int nmemb = ARRAY_SIZE(architectures); + + qsort(architectures, nmemb, sizeof(struct arch), arch__cmp); +} + +struct arch *arch__find(const char *name) +{ + const int nmemb = ARRAY_SIZE(architectures); + static bool sorted; + + if (!sorted) { + arch__sort(); + sorted = true; + } + + return bsearch(name, architectures, nmemb, sizeof(struct arch), arch__key_cmp); +} + +bool arch__is(struct arch *arch, const char *name) +{ + return !strcmp(arch->name, name); +} + +static void ins_ops__delete(struct ins_operands *ops) +{ + if (ops == NULL) + return; + zfree(&ops->source.raw); + zfree(&ops->source.name); + zfree(&ops->target.raw); + zfree(&ops->target.name); +} + +static int ins__raw_scnprintf(struct ins *ins, char *bf, size_t size, + struct ins_operands *ops, int max_ins_name) +{ + return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, ops->raw); +} + +int ins__scnprintf(struct ins *ins, char *bf, size_t size, + struct ins_operands *ops, int max_ins_name) +{ + if (ins->ops->scnprintf) + return ins->ops->scnprintf(ins, bf, size, ops, max_ins_name); + + return ins__raw_scnprintf(ins, bf, size, ops, max_ins_name); +} + +bool ins__is_fused(struct arch *arch, const char *ins1, const char *ins2) +{ + if (!arch || !arch->ins_is_fused) + return false; + + return arch->ins_is_fused(arch, ins1, ins2); +} + +static int call__parse(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms) +{ + char *endptr, *tok, *name; + struct map *map = ms->map; + struct addr_map_symbol target = { + .ms = { .map = map, }, + }; + + ops->target.addr = strtoull(ops->raw, &endptr, 16); + + name = strchr(endptr, '<'); + if (name == NULL) + goto indirect_call; + + name++; + + if (arch->objdump.skip_functions_char && + strchr(name, arch->objdump.skip_functions_char)) + return -1; + + tok = strchr(name, '>'); + if (tok == NULL) + return -1; + + *tok = '\0'; + ops->target.name = strdup(name); + *tok = '>'; + + if (ops->target.name == NULL) + return -1; +find_target: + target.addr = map__objdump_2mem(map, ops->target.addr); + + if (maps__find_ams(ms->maps, &target) == 0 && + map__rip_2objdump(target.ms.map, map__map_ip(target.ms.map, target.addr)) == ops->target.addr) + ops->target.sym = target.ms.sym; + + return 0; + +indirect_call: + tok = strchr(endptr, '*'); + if (tok != NULL) { + endptr++; + + /* Indirect call can use a non-rip register and offset: callq *0x8(%rbx). + * Do not parse such instruction. */ + if (strstr(endptr, "(%r") == NULL) + ops->target.addr = strtoull(endptr, NULL, 16); + } + goto find_target; +} + +static int call__scnprintf(struct ins *ins, char *bf, size_t size, + struct ins_operands *ops, int max_ins_name) +{ + if (ops->target.sym) + return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, ops->target.sym->name); + + if (ops->target.addr == 0) + return ins__raw_scnprintf(ins, bf, size, ops, max_ins_name); + + if (ops->target.name) + return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, ops->target.name); + + return scnprintf(bf, size, "%-*s *%" PRIx64, max_ins_name, ins->name, ops->target.addr); +} + +static struct ins_ops call_ops = { + .parse = call__parse, + .scnprintf = call__scnprintf, +}; + +bool ins__is_call(const struct ins *ins) +{ + return ins->ops == &call_ops || ins->ops == &s390_call_ops || ins->ops == &loongarch_call_ops; +} + +/* + * Prevents from matching commas in the comment section, e.g.: + * ffff200008446e70: b.cs ffff2000084470f4 // b.hs, b.nlast + * + * and skip comma as part of function arguments, e.g.: + * 1d8b4ac + */ +static inline const char *validate_comma(const char *c, struct ins_operands *ops) +{ + if (ops->jump.raw_comment && c > ops->jump.raw_comment) + return NULL; + + if (ops->jump.raw_func_start && c > ops->jump.raw_func_start) + return NULL; + + return c; +} + +static int jump__parse(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms) +{ + struct map *map = ms->map; + struct symbol *sym = ms->sym; + struct addr_map_symbol target = { + .ms = { .map = map, }, + }; + const char *c = strchr(ops->raw, ','); + u64 start, end; + + ops->jump.raw_comment = strchr(ops->raw, arch->objdump.comment_char); + ops->jump.raw_func_start = strchr(ops->raw, '<'); + + c = validate_comma(c, ops); + + /* + * Examples of lines to parse for the _cpp_lex_token@@Base + * function: + * + * 1159e6c: jne 115aa32 <_cpp_lex_token@@Base+0xf92> + * 1159e8b: jne c469be + * + * The first is a jump to an offset inside the same function, + * the second is to another function, i.e. that 0xa72 is an + * offset in the cpp_named_operator2name@@base function. + */ + /* + * skip over possible up to 2 operands to get to address, e.g.: + * tbnz w0, #26, ffff0000083cd190 + */ + if (c++ != NULL) { + ops->target.addr = strtoull(c, NULL, 16); + if (!ops->target.addr) { + c = strchr(c, ','); + c = validate_comma(c, ops); + if (c++ != NULL) + ops->target.addr = strtoull(c, NULL, 16); + } + } else { + ops->target.addr = strtoull(ops->raw, NULL, 16); + } + + target.addr = map__objdump_2mem(map, ops->target.addr); + start = map__unmap_ip(map, sym->start); + end = map__unmap_ip(map, sym->end); + + ops->target.outside = target.addr < start || target.addr > end; + + /* + * FIXME: things like this in _cpp_lex_token (gcc's cc1 program): + + cpp_named_operator2name@@Base+0xa72 + + * Point to a place that is after the cpp_named_operator2name + * boundaries, i.e. in the ELF symbol table for cc1 + * cpp_named_operator2name is marked as being 32-bytes long, but it in + * fact is much larger than that, so we seem to need a symbols__find() + * routine that looks for >= current->start and < next_symbol->start, + * possibly just for C++ objects? + * + * For now lets just make some progress by marking jumps to outside the + * current function as call like. + * + * Actual navigation will come next, with further understanding of how + * the symbol searching and disassembly should be done. + */ + if (maps__find_ams(ms->maps, &target) == 0 && + map__rip_2objdump(target.ms.map, map__map_ip(target.ms.map, target.addr)) == ops->target.addr) + ops->target.sym = target.ms.sym; + + if (!ops->target.outside) { + ops->target.offset = target.addr - start; + ops->target.offset_avail = true; + } else { + ops->target.offset_avail = false; + } + + return 0; +} + +static int jump__scnprintf(struct ins *ins, char *bf, size_t size, + struct ins_operands *ops, int max_ins_name) +{ + const char *c; + + if (!ops->target.addr || ops->target.offset < 0) + return ins__raw_scnprintf(ins, bf, size, ops, max_ins_name); + + if (ops->target.outside && ops->target.sym != NULL) + return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, ops->target.sym->name); + + c = strchr(ops->raw, ','); + c = validate_comma(c, ops); + + if (c != NULL) { + const char *c2 = strchr(c + 1, ','); + + c2 = validate_comma(c2, ops); + /* check for 3-op insn */ + if (c2 != NULL) + c = c2; + c++; + + /* mirror arch objdump's space-after-comma style */ + if (*c == ' ') + c++; + } + + return scnprintf(bf, size, "%-*s %.*s%" PRIx64, max_ins_name, + ins->name, c ? c - ops->raw : 0, ops->raw, + ops->target.offset); +} + +static void jump__delete(struct ins_operands *ops __maybe_unused) +{ + /* + * The ops->jump.raw_comment and ops->jump.raw_func_start belong to the + * raw string, don't free them. + */ +} + +static struct ins_ops jump_ops = { + .free = jump__delete, + .parse = jump__parse, + .scnprintf = jump__scnprintf, +}; + +bool ins__is_jump(const struct ins *ins) +{ + return ins->ops == &jump_ops || ins->ops == &loongarch_jump_ops; +} + +static int comment__symbol(char *raw, char *comment, u64 *addrp, char **namep) +{ + char *endptr, *name, *t; + + if (strstr(raw, "(%rip)") == NULL) + return 0; + + *addrp = strtoull(comment, &endptr, 16); + if (endptr == comment) + return 0; + name = strchr(endptr, '<'); + if (name == NULL) + return -1; + + name++; + + t = strchr(name, '>'); + if (t == NULL) + return 0; + + *t = '\0'; + *namep = strdup(name); + *t = '>'; + + return 0; +} + +static int lock__parse(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms) +{ + ops->locked.ops = zalloc(sizeof(*ops->locked.ops)); + if (ops->locked.ops == NULL) + return 0; + + if (disasm_line__parse(ops->raw, &ops->locked.ins.name, &ops->locked.ops->raw) < 0) + goto out_free_ops; + + ops->locked.ins.ops = ins__find(arch, ops->locked.ins.name); + + if (ops->locked.ins.ops == NULL) + goto out_free_ops; + + if (ops->locked.ins.ops->parse && + ops->locked.ins.ops->parse(arch, ops->locked.ops, ms) < 0) + goto out_free_ops; + + return 0; + +out_free_ops: + zfree(&ops->locked.ops); + return 0; +} + +static int lock__scnprintf(struct ins *ins, char *bf, size_t size, + struct ins_operands *ops, int max_ins_name) +{ + int printed; + + if (ops->locked.ins.ops == NULL) + return ins__raw_scnprintf(ins, bf, size, ops, max_ins_name); + + printed = scnprintf(bf, size, "%-*s ", max_ins_name, ins->name); + return printed + ins__scnprintf(&ops->locked.ins, bf + printed, + size - printed, ops->locked.ops, max_ins_name); +} + +static void lock__delete(struct ins_operands *ops) +{ + struct ins *ins = &ops->locked.ins; + + if (ins->ops && ins->ops->free) + ins->ops->free(ops->locked.ops); + else + ins_ops__delete(ops->locked.ops); + + zfree(&ops->locked.ops); + zfree(&ops->target.raw); + zfree(&ops->target.name); +} + +static struct ins_ops lock_ops = { + .free = lock__delete, + .parse = lock__parse, + .scnprintf = lock__scnprintf, +}; + +/* + * Check if the operand has more than one registers like x86 SIB addressing: + * 0x1234(%rax, %rbx, 8) + * + * But it doesn't care segment selectors like %gs:0x5678(%rcx), so just check + * the input string after 'memory_ref_char' if exists. + */ +static bool check_multi_regs(struct arch *arch, const char *op) +{ + int count = 0; + + if (arch->objdump.register_char == 0) + return false; + + if (arch->objdump.memory_ref_char) { + op = strchr(op, arch->objdump.memory_ref_char); + if (op == NULL) + return false; + } + + while ((op = strchr(op, arch->objdump.register_char)) != NULL) { + count++; + op++; + } + + return count > 1; +} + +static int mov__parse(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms __maybe_unused) +{ + char *s = strchr(ops->raw, ','), *target, *comment, prev; + + if (s == NULL) + return -1; + + *s = '\0'; + + /* + * x86 SIB addressing has something like 0x8(%rax, %rcx, 1) + * then it needs to have the closing parenthesis. + */ + if (strchr(ops->raw, '(')) { + *s = ','; + s = strchr(ops->raw, ')'); + if (s == NULL || s[1] != ',') + return -1; + *++s = '\0'; + } + + ops->source.raw = strdup(ops->raw); + *s = ','; + + if (ops->source.raw == NULL) + return -1; + + ops->source.multi_regs = check_multi_regs(arch, ops->source.raw); + + target = skip_spaces(++s); + comment = strchr(s, arch->objdump.comment_char); + + if (comment != NULL) + s = comment - 1; + else + s = strchr(s, '\0') - 1; + + while (s > target && isspace(s[0])) + --s; + s++; + prev = *s; + *s = '\0'; + + ops->target.raw = strdup(target); + *s = prev; + + if (ops->target.raw == NULL) + goto out_free_source; + + ops->target.multi_regs = check_multi_regs(arch, ops->target.raw); + + if (comment == NULL) + return 0; + + comment = skip_spaces(comment); + comment__symbol(ops->source.raw, comment + 1, &ops->source.addr, &ops->source.name); + comment__symbol(ops->target.raw, comment + 1, &ops->target.addr, &ops->target.name); + + return 0; + +out_free_source: + zfree(&ops->source.raw); + return -1; +} + +static int mov__scnprintf(struct ins *ins, char *bf, size_t size, + struct ins_operands *ops, int max_ins_name) +{ + return scnprintf(bf, size, "%-*s %s,%s", max_ins_name, ins->name, + ops->source.name ?: ops->source.raw, + ops->target.name ?: ops->target.raw); +} + +static struct ins_ops mov_ops = { + .parse = mov__parse, + .scnprintf = mov__scnprintf, +}; + +static int dec__parse(struct arch *arch __maybe_unused, struct ins_operands *ops, struct map_symbol *ms __maybe_unused) +{ + char *target, *comment, *s, prev; + + target = s = ops->raw; + + while (s[0] != '\0' && !isspace(s[0])) + ++s; + prev = *s; + *s = '\0'; + + ops->target.raw = strdup(target); + *s = prev; + + if (ops->target.raw == NULL) + return -1; + + comment = strchr(s, arch->objdump.comment_char); + if (comment == NULL) + return 0; + + comment = skip_spaces(comment); + comment__symbol(ops->target.raw, comment + 1, &ops->target.addr, &ops->target.name); + + return 0; +} + +static int dec__scnprintf(struct ins *ins, char *bf, size_t size, + struct ins_operands *ops, int max_ins_name) +{ + return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, + ops->target.name ?: ops->target.raw); +} + +static struct ins_ops dec_ops = { + .parse = dec__parse, + .scnprintf = dec__scnprintf, +}; + +static int nop__scnprintf(struct ins *ins __maybe_unused, char *bf, size_t size, + struct ins_operands *ops __maybe_unused, int max_ins_name) +{ + return scnprintf(bf, size, "%-*s", max_ins_name, "nop"); +} + +static struct ins_ops nop_ops = { + .scnprintf = nop__scnprintf, +}; + +static struct ins_ops ret_ops = { + .scnprintf = ins__raw_scnprintf, +}; + +bool ins__is_nop(const struct ins *ins) +{ + return ins->ops == &nop_ops; +} + +bool ins__is_ret(const struct ins *ins) +{ + return ins->ops == &ret_ops; +} + +bool ins__is_lock(const struct ins *ins) +{ + return ins->ops == &lock_ops; +} + +static int ins__key_cmp(const void *name, const void *insp) +{ + const struct ins *ins = insp; + + return strcmp(name, ins->name); +} + +static int ins__cmp(const void *a, const void *b) +{ + const struct ins *ia = a; + const struct ins *ib = b; + + return strcmp(ia->name, ib->name); +} + +static void ins__sort(struct arch *arch) +{ + const int nmemb = arch->nr_instructions; + + qsort(arch->instructions, nmemb, sizeof(struct ins), ins__cmp); +} + +static struct ins_ops *__ins__find(struct arch *arch, const char *name) +{ + struct ins *ins; + const int nmemb = arch->nr_instructions; + + if (!arch->sorted_instructions) { + ins__sort(arch); + arch->sorted_instructions = true; + } + + ins = bsearch(name, arch->instructions, nmemb, sizeof(struct ins), ins__key_cmp); + if (ins) + return ins->ops; + + if (arch->insn_suffix) { + char tmp[32]; + char suffix; + size_t len = strlen(name); + + if (len == 0 || len >= sizeof(tmp)) + return NULL; + + suffix = name[len - 1]; + if (strchr(arch->insn_suffix, suffix) == NULL) + return NULL; + + strcpy(tmp, name); + tmp[len - 1] = '\0'; /* remove the suffix and check again */ + + ins = bsearch(tmp, arch->instructions, nmemb, sizeof(struct ins), ins__key_cmp); + } + return ins ? ins->ops : NULL; +} + +struct ins_ops *ins__find(struct arch *arch, const char *name) +{ + struct ins_ops *ops = __ins__find(arch, name); + + if (!ops && arch->associate_instruction_ops) + ops = arch->associate_instruction_ops(arch, name); + + return ops; +} + +static void disasm_line__init_ins(struct disasm_line *dl, struct arch *arch, struct map_symbol *ms) +{ + dl->ins.ops = ins__find(arch, dl->ins.name); + + if (!dl->ins.ops) + return; + + if (dl->ins.ops->parse && dl->ins.ops->parse(arch, &dl->ops, ms) < 0) + dl->ins.ops = NULL; +} + +static int disasm_line__parse(char *line, const char **namep, char **rawp) +{ + char tmp, *name = skip_spaces(line); + + if (name[0] == '\0') + return -1; + + *rawp = name + 1; + + while ((*rawp)[0] != '\0' && !isspace((*rawp)[0])) + ++*rawp; + + tmp = (*rawp)[0]; + (*rawp)[0] = '\0'; + *namep = strdup(name); + + if (*namep == NULL) + goto out; + + (*rawp)[0] = tmp; + *rawp = strim(*rawp); + + return 0; + +out: + return -1; +} + +static void annotation_line__init(struct annotation_line *al, + struct annotate_args *args, + int nr) +{ + al->offset = args->offset; + al->line = strdup(args->line); + al->line_nr = args->line_nr; + al->fileloc = args->fileloc; + al->data_nr = nr; +} + +static void annotation_line__exit(struct annotation_line *al) +{ + zfree_srcline(&al->path); + zfree(&al->line); + zfree(&al->cycles); +} + +static size_t disasm_line_size(int nr) +{ + struct annotation_line *al; + + return (sizeof(struct disasm_line) + (sizeof(al->data[0]) * nr)); +} + +/* + * Allocating the disasm annotation line data with + * following structure: + * + * ------------------------------------------- + * struct disasm_line | struct annotation_line + * ------------------------------------------- + * + * We have 'struct annotation_line' member as last member + * of 'struct disasm_line' to have an easy access. + */ +struct disasm_line *disasm_line__new(struct annotate_args *args) +{ + struct disasm_line *dl = NULL; + int nr = 1; + + if (evsel__is_group_event(args->evsel)) + nr = args->evsel->core.nr_members; + + dl = zalloc(disasm_line_size(nr)); + if (!dl) + return NULL; + + annotation_line__init(&dl->al, args, nr); + if (dl->al.line == NULL) + goto out_delete; + + if (args->offset != -1) { + if (disasm_line__parse(dl->al.line, &dl->ins.name, &dl->ops.raw) < 0) + goto out_free_line; + + disasm_line__init_ins(dl, args->arch, &args->ms); + } + + return dl; + +out_free_line: + zfree(&dl->al.line); +out_delete: + free(dl); + return NULL; +} + +void disasm_line__free(struct disasm_line *dl) +{ + if (dl->ins.ops && dl->ins.ops->free) + dl->ins.ops->free(&dl->ops); + else + ins_ops__delete(&dl->ops); + zfree(&dl->ins.name); + annotation_line__exit(&dl->al); + free(dl); +} + +int disasm_line__scnprintf(struct disasm_line *dl, char *bf, size_t size, bool raw, int max_ins_name) +{ + if (raw || !dl->ins.ops) + return scnprintf(bf, size, "%-*s %s", max_ins_name, dl->ins.name, dl->ops.raw); + + return ins__scnprintf(&dl->ins, bf, size, &dl->ops, max_ins_name); +} + +/* + * symbol__parse_objdump_line() parses objdump output (with -d --no-show-raw) + * which looks like following + * + * 0000000000415500 <_init>: + * 415500: sub $0x8,%rsp + * 415504: mov 0x2f5ad5(%rip),%rax # 70afe0 <_DYNAMIC+0x2f8> + * 41550b: test %rax,%rax + * 41550e: je 415515 <_init+0x15> + * 415510: callq 416e70 <__gmon_start__@plt> + * 415515: add $0x8,%rsp + * 415519: retq + * + * it will be parsed and saved into struct disasm_line as + * + * + * The offset will be a relative offset from the start of the symbol and -1 + * means that it's not a disassembly line so should be treated differently. + * The ops.raw part will be parsed further according to type of the instruction. + */ +static int symbol__parse_objdump_line(struct symbol *sym, + struct annotate_args *args, + char *parsed_line, int *line_nr, char **fileloc) +{ + struct map *map = args->ms.map; + struct annotation *notes = symbol__annotation(sym); + struct disasm_line *dl; + char *tmp; + s64 line_ip, offset = -1; + regmatch_t match[2]; + + /* /filename:linenr ? Save line number and ignore. */ + if (regexec(&file_lineno, parsed_line, 2, match, 0) == 0) { + *line_nr = atoi(parsed_line + match[1].rm_so); + free(*fileloc); + *fileloc = strdup(parsed_line); + return 0; + } + + /* Process hex address followed by ':'. */ + line_ip = strtoull(parsed_line, &tmp, 16); + if (parsed_line != tmp && tmp[0] == ':' && tmp[1] != '\0') { + u64 start = map__rip_2objdump(map, sym->start), + end = map__rip_2objdump(map, sym->end); + + offset = line_ip - start; + if ((u64)line_ip < start || (u64)line_ip >= end) + offset = -1; + else + parsed_line = tmp + 1; + } + + args->offset = offset; + args->line = parsed_line; + args->line_nr = *line_nr; + args->fileloc = *fileloc; + args->ms.sym = sym; + + dl = disasm_line__new(args); + (*line_nr)++; + + if (dl == NULL) + return -1; + + if (!disasm_line__has_local_offset(dl)) { + dl->ops.target.offset = dl->ops.target.addr - + map__rip_2objdump(map, sym->start); + dl->ops.target.offset_avail = true; + } + + /* kcore has no symbols, so add the call target symbol */ + if (dl->ins.ops && ins__is_call(&dl->ins) && !dl->ops.target.sym) { + struct addr_map_symbol target = { + .addr = dl->ops.target.addr, + .ms = { .map = map, }, + }; + + if (!maps__find_ams(args->ms.maps, &target) && + target.ms.sym->start == target.al_addr) + dl->ops.target.sym = target.ms.sym; + } + + annotation_line__add(&dl->al, ¬es->src->source); + return 0; +} + +static void delete_last_nop(struct symbol *sym) +{ + struct annotation *notes = symbol__annotation(sym); + struct list_head *list = ¬es->src->source; + struct disasm_line *dl; + + while (!list_empty(list)) { + dl = list_entry(list->prev, struct disasm_line, al.node); + + if (dl->ins.ops) { + if (!ins__is_nop(&dl->ins)) + return; + } else { + if (!strstr(dl->al.line, " nop ") && + !strstr(dl->al.line, " nopl ") && + !strstr(dl->al.line, " nopw ")) + return; + } + + list_del_init(&dl->al.node); + disasm_line__free(dl); + } +} + +int symbol__strerror_disassemble(struct map_symbol *ms, int errnum, char *buf, size_t buflen) +{ + struct dso *dso = map__dso(ms->map); + + BUG_ON(buflen == 0); + + if (errnum >= 0) { + str_error_r(errnum, buf, buflen); + return 0; + } + + switch (errnum) { + case SYMBOL_ANNOTATE_ERRNO__NO_VMLINUX: { + char bf[SBUILD_ID_SIZE + 15] = " with build id "; + char *build_id_msg = NULL; + + if (dso->has_build_id) { + build_id__sprintf(&dso->bid, bf + 15); + build_id_msg = bf; + } + scnprintf(buf, buflen, + "No vmlinux file%s\nwas found in the path.\n\n" + "Note that annotation using /proc/kcore requires CAP_SYS_RAWIO capability.\n\n" + "Please use:\n\n" + " perf buildid-cache -vu vmlinux\n\n" + "or:\n\n" + " --vmlinux vmlinux\n", build_id_msg ?: ""); + } + break; + case SYMBOL_ANNOTATE_ERRNO__NO_LIBOPCODES_FOR_BPF: + scnprintf(buf, buflen, "Please link with binutils's libopcode to enable BPF annotation"); + break; + case SYMBOL_ANNOTATE_ERRNO__ARCH_INIT_REGEXP: + scnprintf(buf, buflen, "Problems with arch specific instruction name regular expressions."); + break; + case SYMBOL_ANNOTATE_ERRNO__ARCH_INIT_CPUID_PARSING: + scnprintf(buf, buflen, "Problems while parsing the CPUID in the arch specific initialization."); + break; + case SYMBOL_ANNOTATE_ERRNO__BPF_INVALID_FILE: + scnprintf(buf, buflen, "Invalid BPF file: %s.", dso->long_name); + break; + case SYMBOL_ANNOTATE_ERRNO__BPF_MISSING_BTF: + scnprintf(buf, buflen, "The %s BPF file has no BTF section, compile with -g or use pahole -J.", + dso->long_name); + break; + default: + scnprintf(buf, buflen, "Internal error: Invalid %d error code\n", errnum); + break; + } + + return 0; +} + +static int dso__disassemble_filename(struct dso *dso, char *filename, size_t filename_size) +{ + char linkname[PATH_MAX]; + char *build_id_filename; + char *build_id_path = NULL; + char *pos; + int len; + + if (dso->symtab_type == DSO_BINARY_TYPE__KALLSYMS && + !dso__is_kcore(dso)) + return SYMBOL_ANNOTATE_ERRNO__NO_VMLINUX; + + build_id_filename = dso__build_id_filename(dso, NULL, 0, false); + if (build_id_filename) { + __symbol__join_symfs(filename, filename_size, build_id_filename); + free(build_id_filename); + } else { + if (dso->has_build_id) + return ENOMEM; + goto fallback; + } + + build_id_path = strdup(filename); + if (!build_id_path) + return ENOMEM; + + /* + * old style build-id cache has name of XX/XXXXXXX.. while + * new style has XX/XXXXXXX../{elf,kallsyms,vdso}. + * extract the build-id part of dirname in the new style only. + */ + pos = strrchr(build_id_path, '/'); + if (pos && strlen(pos) < SBUILD_ID_SIZE - 2) + dirname(build_id_path); + + if (dso__is_kcore(dso)) + goto fallback; + + len = readlink(build_id_path, linkname, sizeof(linkname) - 1); + if (len < 0) + goto fallback; + + linkname[len] = '\0'; + if (strstr(linkname, DSO__NAME_KALLSYMS) || + access(filename, R_OK)) { +fallback: + /* + * If we don't have build-ids or the build-id file isn't in the + * cache, or is just a kallsyms file, well, lets hope that this + * DSO is the same as when 'perf record' ran. + */ + if (dso->kernel && dso->long_name[0] == '/') + snprintf(filename, filename_size, "%s", dso->long_name); + else + __symbol__join_symfs(filename, filename_size, dso->long_name); + + mutex_lock(&dso->lock); + if (access(filename, R_OK) && errno == ENOENT && dso->nsinfo) { + char *new_name = dso__filename_with_chroot(dso, filename); + if (new_name) { + strlcpy(filename, new_name, filename_size); + free(new_name); + } + } + mutex_unlock(&dso->lock); + } + + free(build_id_path); + return 0; +} + +#if defined(HAVE_LIBBFD_SUPPORT) && defined(HAVE_LIBBPF_SUPPORT) +#define PACKAGE "perf" +#include +#include +#include +#include +#include +#include +#include + +#include "bpf-event.h" +#include "bpf-utils.h" + +static int symbol__disassemble_bpf(struct symbol *sym, + struct annotate_args *args) +{ + struct annotation *notes = symbol__annotation(sym); + struct bpf_prog_linfo *prog_linfo = NULL; + struct bpf_prog_info_node *info_node; + int len = sym->end - sym->start; + disassembler_ftype disassemble; + struct map *map = args->ms.map; + struct perf_bpil *info_linear; + struct disassemble_info info; + struct dso *dso = map__dso(map); + int pc = 0, count, sub_id; + struct btf *btf = NULL; + char tpath[PATH_MAX]; + size_t buf_size; + int nr_skip = 0; + char *buf; + bfd *bfdf; + int ret; + FILE *s; + + if (dso->binary_type != DSO_BINARY_TYPE__BPF_PROG_INFO) + return SYMBOL_ANNOTATE_ERRNO__BPF_INVALID_FILE; + + pr_debug("%s: handling sym %s addr %" PRIx64 " len %" PRIx64 "\n", __func__, + sym->name, sym->start, sym->end - sym->start); + + memset(tpath, 0, sizeof(tpath)); + perf_exe(tpath, sizeof(tpath)); + + bfdf = bfd_openr(tpath, NULL); + if (bfdf == NULL) + abort(); + + if (!bfd_check_format(bfdf, bfd_object)) + abort(); + + s = open_memstream(&buf, &buf_size); + if (!s) { + ret = errno; + goto out; + } + init_disassemble_info_compat(&info, s, + (fprintf_ftype) fprintf, + fprintf_styled); + info.arch = bfd_get_arch(bfdf); + info.mach = bfd_get_mach(bfdf); + + info_node = perf_env__find_bpf_prog_info(dso->bpf_prog.env, + dso->bpf_prog.id); + if (!info_node) { + ret = SYMBOL_ANNOTATE_ERRNO__BPF_MISSING_BTF; + goto out; + } + info_linear = info_node->info_linear; + sub_id = dso->bpf_prog.sub_id; + + info.buffer = (void *)(uintptr_t)(info_linear->info.jited_prog_insns); + info.buffer_length = info_linear->info.jited_prog_len; + + if (info_linear->info.nr_line_info) + prog_linfo = bpf_prog_linfo__new(&info_linear->info); + + if (info_linear->info.btf_id) { + struct btf_node *node; + + node = perf_env__find_btf(dso->bpf_prog.env, + info_linear->info.btf_id); + if (node) + btf = btf__new((__u8 *)(node->data), + node->data_size); + } + + disassemble_init_for_target(&info); + +#ifdef DISASM_FOUR_ARGS_SIGNATURE + disassemble = disassembler(info.arch, + bfd_big_endian(bfdf), + info.mach, + bfdf); +#else + disassemble = disassembler(bfdf); +#endif + if (disassemble == NULL) + abort(); + + fflush(s); + do { + const struct bpf_line_info *linfo = NULL; + struct disasm_line *dl; + size_t prev_buf_size; + const char *srcline; + u64 addr; + + addr = pc + ((u64 *)(uintptr_t)(info_linear->info.jited_ksyms))[sub_id]; + count = disassemble(pc, &info); + + if (prog_linfo) + linfo = bpf_prog_linfo__lfind_addr_func(prog_linfo, + addr, sub_id, + nr_skip); + + if (linfo && btf) { + srcline = btf__name_by_offset(btf, linfo->line_off); + nr_skip++; + } else + srcline = NULL; + + fprintf(s, "\n"); + prev_buf_size = buf_size; + fflush(s); + + if (!annotate_opts.hide_src_code && srcline) { + args->offset = -1; + args->line = strdup(srcline); + args->line_nr = 0; + args->fileloc = NULL; + args->ms.sym = sym; + dl = disasm_line__new(args); + if (dl) { + annotation_line__add(&dl->al, + ¬es->src->source); + } + } + + args->offset = pc; + args->line = buf + prev_buf_size; + args->line_nr = 0; + args->fileloc = NULL; + args->ms.sym = sym; + dl = disasm_line__new(args); + if (dl) + annotation_line__add(&dl->al, ¬es->src->source); + + pc += count; + } while (count > 0 && pc < len); + + ret = 0; +out: + free(prog_linfo); + btf__free(btf); + fclose(s); + bfd_close(bfdf); + return ret; +} +#else // defined(HAVE_LIBBFD_SUPPORT) && defined(HAVE_LIBBPF_SUPPORT) +static int symbol__disassemble_bpf(struct symbol *sym __maybe_unused, + struct annotate_args *args __maybe_unused) +{ + return SYMBOL_ANNOTATE_ERRNO__NO_LIBOPCODES_FOR_BPF; +} +#endif // defined(HAVE_LIBBFD_SUPPORT) && defined(HAVE_LIBBPF_SUPPORT) + +static int +symbol__disassemble_bpf_image(struct symbol *sym, + struct annotate_args *args) +{ + struct annotation *notes = symbol__annotation(sym); + struct disasm_line *dl; + + args->offset = -1; + args->line = strdup("to be implemented"); + args->line_nr = 0; + args->fileloc = NULL; + dl = disasm_line__new(args); + if (dl) + annotation_line__add(&dl->al, ¬es->src->source); + + zfree(&args->line); + return 0; +} + +/* + * Possibly create a new version of line with tabs expanded. Returns the + * existing or new line, storage is updated if a new line is allocated. If + * allocation fails then NULL is returned. + */ +static char *expand_tabs(char *line, char **storage, size_t *storage_len) +{ + size_t i, src, dst, len, new_storage_len, num_tabs; + char *new_line; + size_t line_len = strlen(line); + + for (num_tabs = 0, i = 0; i < line_len; i++) + if (line[i] == '\t') + num_tabs++; + + if (num_tabs == 0) + return line; + + /* + * Space for the line and '\0', less the leading and trailing + * spaces. Each tab may introduce 7 additional spaces. + */ + new_storage_len = line_len + 1 + (num_tabs * 7); + + new_line = malloc(new_storage_len); + if (new_line == NULL) { + pr_err("Failure allocating memory for tab expansion\n"); + return NULL; + } + + /* + * Copy regions starting at src and expand tabs. If there are two + * adjacent tabs then 'src == i', the memcpy is of size 0 and the spaces + * are inserted. + */ + for (i = 0, src = 0, dst = 0; i < line_len && num_tabs; i++) { + if (line[i] == '\t') { + len = i - src; + memcpy(&new_line[dst], &line[src], len); + dst += len; + new_line[dst++] = ' '; + while (dst % 8 != 0) + new_line[dst++] = ' '; + src = i + 1; + num_tabs--; + } + } + + /* Expand the last region. */ + len = line_len - src; + memcpy(&new_line[dst], &line[src], len); + dst += len; + new_line[dst] = '\0'; + + free(*storage); + *storage = new_line; + *storage_len = new_storage_len; + return new_line; +} + +int symbol__disassemble(struct symbol *sym, struct annotate_args *args) +{ + struct annotation_options *opts = &annotate_opts; + struct map *map = args->ms.map; + struct dso *dso = map__dso(map); + char *command; + FILE *file; + char symfs_filename[PATH_MAX]; + struct kcore_extract kce; + bool delete_extract = false; + bool decomp = false; + int lineno = 0; + char *fileloc = NULL; + int nline; + char *line; + size_t line_len; + const char *objdump_argv[] = { + "/bin/sh", + "-c", + NULL, /* Will be the objdump command to run. */ + "--", + NULL, /* Will be the symfs path. */ + NULL, + }; + struct child_process objdump_process; + int err = dso__disassemble_filename(dso, symfs_filename, sizeof(symfs_filename)); + + if (err) + return err; + + pr_debug("%s: filename=%s, sym=%s, start=%#" PRIx64 ", end=%#" PRIx64 "\n", __func__, + symfs_filename, sym->name, map__unmap_ip(map, sym->start), + map__unmap_ip(map, sym->end)); + + pr_debug("annotating [%p] %30s : [%p] %30s\n", + dso, dso->long_name, sym, sym->name); + + if (dso->binary_type == DSO_BINARY_TYPE__BPF_PROG_INFO) { + return symbol__disassemble_bpf(sym, args); + } else if (dso->binary_type == DSO_BINARY_TYPE__BPF_IMAGE) { + return symbol__disassemble_bpf_image(sym, args); + } else if (dso__is_kcore(dso)) { + kce.kcore_filename = symfs_filename; + kce.addr = map__rip_2objdump(map, sym->start); + kce.offs = sym->start; + kce.len = sym->end - sym->start; + if (!kcore_extract__create(&kce)) { + delete_extract = true; + strlcpy(symfs_filename, kce.extract_filename, + sizeof(symfs_filename)); + } + } else if (dso__needs_decompress(dso)) { + char tmp[KMOD_DECOMP_LEN]; + + if (dso__decompress_kmodule_path(dso, symfs_filename, + tmp, sizeof(tmp)) < 0) + return -1; + + decomp = true; + strcpy(symfs_filename, tmp); + } + + err = asprintf(&command, + "%s %s%s --start-address=0x%016" PRIx64 + " --stop-address=0x%016" PRIx64 + " %s -d %s %s %s %c%s%c %s%s -C \"$1\"", + opts->objdump_path ?: "objdump", + opts->disassembler_style ? "-M " : "", + opts->disassembler_style ?: "", + map__rip_2objdump(map, sym->start), + map__rip_2objdump(map, sym->end), + opts->show_linenr ? "-l" : "", + opts->show_asm_raw ? "" : "--no-show-raw-insn", + opts->annotate_src ? "-S" : "", + opts->prefix ? "--prefix " : "", + opts->prefix ? '"' : ' ', + opts->prefix ?: "", + opts->prefix ? '"' : ' ', + opts->prefix_strip ? "--prefix-strip=" : "", + opts->prefix_strip ?: ""); + + if (err < 0) { + pr_err("Failure allocating memory for the command to run\n"); + goto out_remove_tmp; + } + + pr_debug("Executing: %s\n", command); + + objdump_argv[2] = command; + objdump_argv[4] = symfs_filename; + + /* Create a pipe to read from for stdout */ + memset(&objdump_process, 0, sizeof(objdump_process)); + objdump_process.argv = objdump_argv; + objdump_process.out = -1; + objdump_process.err = -1; + objdump_process.no_stderr = 1; + if (start_command(&objdump_process)) { + pr_err("Failure starting to run %s\n", command); + err = -1; + goto out_free_command; + } + + file = fdopen(objdump_process.out, "r"); + if (!file) { + pr_err("Failure creating FILE stream for %s\n", command); + /* + * If we were using debug info should retry with + * original binary. + */ + err = -1; + goto out_close_stdout; + } + + /* Storage for getline. */ + line = NULL; + line_len = 0; + + nline = 0; + while (!feof(file)) { + const char *match; + char *expanded_line; + + if (getline(&line, &line_len, file) < 0 || !line) + break; + + /* Skip lines containing "filename:" */ + match = strstr(line, symfs_filename); + if (match && match[strlen(symfs_filename)] == ':') + continue; + + expanded_line = strim(line); + expanded_line = expand_tabs(expanded_line, &line, &line_len); + if (!expanded_line) + break; + + /* + * The source code line number (lineno) needs to be kept in + * across calls to symbol__parse_objdump_line(), so that it + * can associate it with the instructions till the next one. + * See disasm_line__new() and struct disasm_line::line_nr. + */ + if (symbol__parse_objdump_line(sym, args, expanded_line, + &lineno, &fileloc) < 0) + break; + nline++; + } + free(line); + free(fileloc); + + err = finish_command(&objdump_process); + if (err) + pr_err("Error running %s\n", command); + + if (nline == 0) { + err = -1; + pr_err("No output from %s\n", command); + } + + /* + * kallsyms does not have symbol sizes so there may a nop at the end. + * Remove it. + */ + if (dso__is_kcore(dso)) + delete_last_nop(sym); + + fclose(file); + +out_close_stdout: + close(objdump_process.out); + +out_free_command: + free(command); + +out_remove_tmp: + if (decomp) + unlink(symfs_filename); + + if (delete_extract) + kcore_extract__delete(&kce); + + return err; +} diff --git a/tools/perf/util/disasm.h b/tools/perf/util/disasm.h new file mode 100644 index 000000000000..3d381a043520 --- /dev/null +++ b/tools/perf/util/disasm.h @@ -0,0 +1,112 @@ +// SPDX-License-Identifier: GPL-2.0 +#ifndef __PERF_UTIL_DISASM_H +#define __PERF_UTIL_DISASM_H + +#include "map_symbol.h" + +struct annotation_options; +struct disasm_line; +struct ins; +struct evsel; +struct symbol; + +struct arch { + const char *name; + struct ins *instructions; + size_t nr_instructions; + size_t nr_instructions_allocated; + struct ins_ops *(*associate_instruction_ops)(struct arch *arch, const char *name); + bool sorted_instructions; + bool initialized; + const char *insn_suffix; + void *priv; + unsigned int model; + unsigned int family; + int (*init)(struct arch *arch, char *cpuid); + bool (*ins_is_fused)(struct arch *arch, const char *ins1, + const char *ins2); + struct { + char comment_char; + char skip_functions_char; + char register_char; + char memory_ref_char; + char imm_char; + } objdump; +}; + +struct ins { + const char *name; + struct ins_ops *ops; +}; + +struct ins_operands { + char *raw; + struct { + char *raw; + char *name; + struct symbol *sym; + u64 addr; + s64 offset; + bool offset_avail; + bool outside; + bool multi_regs; + } target; + union { + struct { + char *raw; + char *name; + u64 addr; + bool multi_regs; + } source; + struct { + struct ins ins; + struct ins_operands *ops; + } locked; + struct { + char *raw_comment; + char *raw_func_start; + } jump; + }; +}; + +struct ins_ops { + void (*free)(struct ins_operands *ops); + int (*parse)(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms); + int (*scnprintf)(struct ins *ins, char *bf, size_t size, + struct ins_operands *ops, int max_ins_name); +}; + +struct annotate_args { + struct arch *arch; + struct map_symbol ms; + struct evsel *evsel; + struct annotation_options *options; + s64 offset; + char *line; + int line_nr; + char *fileloc; +}; + +struct arch *arch__find(const char *name); +bool arch__is(struct arch *arch, const char *name); + +struct ins_ops *ins__find(struct arch *arch, const char *name); +int ins__scnprintf(struct ins *ins, char *bf, size_t size, + struct ins_operands *ops, int max_ins_name); + +bool ins__is_call(const struct ins *ins); +bool ins__is_jump(const struct ins *ins); +bool ins__is_fused(struct arch *arch, const char *ins1, const char *ins2); +bool ins__is_nop(const struct ins *ins); +bool ins__is_ret(const struct ins *ins); +bool ins__is_lock(const struct ins *ins); + +struct disasm_line *disasm_line__new(struct annotate_args *args); +void disasm_line__free(struct disasm_line *dl); + +int disasm_line__scnprintf(struct disasm_line *dl, char *bf, size_t size, + bool raw, int max_ins_name); + +int symbol__disassemble(struct symbol *sym, struct annotate_args *args); + +#endif /* __PERF_UTIL_DISASM_H */ -- cgit v1.2.3 From 6d17edc113de1e21fc66afa76be475a4f7c91826 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 29 Mar 2024 14:58:11 -0700 Subject: perf annotate: Use libcapstone to disassemble Now it can use the capstone library to disassemble the instructions. Let's use that (if available) for perf annotate to speed up. Currently it only supports x86 architecture. With this change I can see ~3x speed up in data type profiling. But note that capstone cannot give the source file and line number info. For now, users should use the external objdump for that by specifying the --objdump option explicitly. Signed-off-by: Namhyung Kim Tested-by: Ian Rogers Cc: Adrian Hunter Cc: Changbin Du Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240329215812.537846-5-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/disasm.c | 160 +++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 160 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/disasm.c b/tools/perf/util/disasm.c index 21a43b07e8b5..6552b45e37ec 100644 --- a/tools/perf/util/disasm.c +++ b/tools/perf/util/disasm.c @@ -1,6 +1,7 @@ // SPDX-License-Identifier: GPL-2.0-only #include #include +#include #include #include #include @@ -19,6 +20,7 @@ #include "evsel.h" #include "map.h" #include "maps.h" +#include "namespaces.h" #include "srcline.h" #include "symbol.h" #include "util.h" @@ -1346,6 +1348,158 @@ symbol__disassemble_bpf_image(struct symbol *sym, return 0; } +#ifdef HAVE_LIBCAPSTONE_SUPPORT +#include + +static int open_capstone_handle(struct annotate_args *args, bool is_64bit, + csh *handle) +{ + struct annotation_options *opt = args->options; + cs_mode mode = is_64bit ? CS_MODE_64 : CS_MODE_32; + + /* TODO: support more architectures */ + if (!arch__is(args->arch, "x86")) + return -1; + + if (cs_open(CS_ARCH_X86, mode, handle) != CS_ERR_OK) + return -1; + + if (!opt->disassembler_style || + !strcmp(opt->disassembler_style, "att")) + cs_option(*handle, CS_OPT_SYNTAX, CS_OPT_SYNTAX_ATT); + + return 0; +} + +struct find_file_offset_data { + u64 ip; + u64 offset; +}; + +/* This will be called for each PHDR in an ELF binary */ +static int find_file_offset(u64 start, u64 len, u64 pgoff, void *arg) +{ + struct find_file_offset_data *data = arg; + + if (start <= data->ip && data->ip < start + len) { + data->offset = pgoff + data->ip - start; + return 1; + } + return 0; +} + +static int symbol__disassemble_capstone(char *filename, struct symbol *sym, + struct annotate_args *args) +{ + struct annotation *notes = symbol__annotation(sym); + struct map *map = args->ms.map; + struct dso *dso = map__dso(map); + struct nscookie nsc; + u64 start = map__rip_2objdump(map, sym->start); + u64 end = map__rip_2objdump(map, sym->end); + u64 len = end - start; + u64 offset; + int i, fd, count; + bool is_64bit = false; + bool needs_cs_close = false; + u8 *buf = NULL; + struct find_file_offset_data data = { + .ip = start, + }; + csh handle; + cs_insn *insn; + char disasm_buf[512]; + struct disasm_line *dl; + + if (args->options->objdump_path) + return -1; + + nsinfo__mountns_enter(dso->nsinfo, &nsc); + fd = open(filename, O_RDONLY); + nsinfo__mountns_exit(&nsc); + if (fd < 0) + return -1; + + if (file__read_maps(fd, /*exe=*/true, find_file_offset, &data, + &is_64bit) == 0) + goto err; + + if (open_capstone_handle(args, is_64bit, &handle) < 0) + goto err; + + needs_cs_close = true; + + buf = malloc(len); + if (buf == NULL) + goto err; + + count = pread(fd, buf, len, data.offset); + close(fd); + fd = -1; + + if ((u64)count != len) + goto err; + + /* add the function address and name */ + scnprintf(disasm_buf, sizeof(disasm_buf), "%#"PRIx64" <%s>:", + start, sym->name); + + args->offset = -1; + args->line = disasm_buf; + args->line_nr = 0; + args->fileloc = NULL; + args->ms.sym = sym; + + dl = disasm_line__new(args); + if (dl == NULL) + goto err; + + annotation_line__add(&dl->al, ¬es->src->source); + + count = cs_disasm(handle, buf, len, start, len, &insn); + for (i = 0, offset = 0; i < count; i++) { + scnprintf(disasm_buf, sizeof(disasm_buf), + " %-7s %s", + insn[i].mnemonic, insn[i].op_str); + + args->offset = offset; + args->line = disasm_buf; + + dl = disasm_line__new(args); + if (dl == NULL) + goto err; + + annotation_line__add(&dl->al, ¬es->src->source); + + offset += insn[i].size; + } + +out: + if (needs_cs_close) + cs_close(&handle); + free(buf); + return count < 0 ? count : 0; + +err: + if (fd >= 0) + close(fd); + if (needs_cs_close) { + struct disasm_line *tmp; + + /* + * It probably failed in the middle of the above loop. + * Release any resources it might add. + */ + list_for_each_entry_safe(dl, tmp, ¬es->src->source, al.node) { + list_del(&dl->al.node); + free(dl); + } + } + count = -1; + goto out; +} +#endif + /* * Possibly create a new version of line with tabs expanded. Returns the * existing or new line, storage is updated if a new line is allocated. If @@ -1468,6 +1622,12 @@ int symbol__disassemble(struct symbol *sym, struct annotate_args *args) strcpy(symfs_filename, tmp); } +#ifdef HAVE_LIBCAPSTONE_SUPPORT + err = symbol__disassemble_capstone(symfs_filename, sym, args); + if (err == 0) + goto out_remove_tmp; +#endif + err = asprintf(&command, "%s %s%s --start-address=0x%016" PRIx64 " --stop-address=0x%016" PRIx64 -- cgit v1.2.3 From 92dfc59463d55b9abb47429591fd8a21b27930cf Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 29 Mar 2024 14:58:12 -0700 Subject: perf annotate: Add symbol name when using capstone This is to keep the existing behavior with objdump. It needs to show symbol information of global variables like below: Percent | Source code & Disassembly of elf for cycles:P (1 samples, percent: local period) ------------------------------------------------------------------------------------------------ : 0 0xffffffff81338f70 : 0.00 : ffffffff81338f70: endbr64 0.00 : ffffffff81338f74: callq 0xffffffff81083a40 0.00 : ffffffff81338f79: movq %rdi, %r8 0.00 : ffffffff81338f7c: movq %rdx, %rdi 0.00 : ffffffff81338f7f: callq *0x17021c3(%rip) # ffffffff82a3b148 0.00 : ffffffff81338f85: movq 0xffbf3c(%rip), %rdx # ffffffff82334ec8 0.00 : ffffffff81338f8c: testq %rax, %rax ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 0.00 : ffffffff81338f8f: je 0xffffffff81338fd0 here 0.00 : ffffffff81338f91: movq %rax, %rcx 0.00 : ffffffff81338f94: andl $1, %ecx Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Changbin Du Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240329215812.537846-6-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/disasm.c | 74 ++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 71 insertions(+), 3 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/disasm.c b/tools/perf/util/disasm.c index 6552b45e37ec..a1219eb930aa 100644 --- a/tools/perf/util/disasm.c +++ b/tools/perf/util/disasm.c @@ -1368,6 +1368,12 @@ static int open_capstone_handle(struct annotate_args *args, bool is_64bit, !strcmp(opt->disassembler_style, "att")) cs_option(*handle, CS_OPT_SYNTAX, CS_OPT_SYNTAX_ATT); + /* + * Resolving address operands to symbols is implemented + * on x86 by investigating instruction details. + */ + cs_option(*handle, CS_OPT_DETAIL, CS_OPT_ON); + return 0; } @@ -1388,6 +1394,63 @@ static int find_file_offset(u64 start, u64 len, u64 pgoff, void *arg) return 0; } +static void print_capstone_detail(cs_insn *insn, char *buf, size_t len, + struct annotate_args *args, u64 addr) +{ + int i; + struct map *map = args->ms.map; + struct symbol *sym; + + /* TODO: support more architectures */ + if (!arch__is(args->arch, "x86")) + return; + + if (insn->detail == NULL) + return; + + for (i = 0; i < insn->detail->x86.op_count; i++) { + cs_x86_op *op = &insn->detail->x86.operands[i]; + u64 orig_addr; + + if (op->type != X86_OP_MEM) + continue; + + /* only print RIP-based global symbols for now */ + if (op->mem.base != X86_REG_RIP) + continue; + + /* get the target address */ + orig_addr = addr + insn->size + op->mem.disp; + addr = map__objdump_2mem(map, orig_addr); + + if (map__dso(map)->kernel) { + /* + * The kernel maps can be splitted into sections, + * let's find the map first and the search the symbol. + */ + map = maps__find(map__kmaps(map), addr); + if (map == NULL) + continue; + } + + /* convert it to map-relative address for search */ + addr = map__map_ip(map, addr); + + sym = map__find_symbol(map, addr); + if (sym == NULL) + continue; + + if (addr == sym->start) { + scnprintf(buf, len, "\t# %"PRIx64" <%s>", + orig_addr, sym->name); + } else { + scnprintf(buf, len, "\t# %"PRIx64" <%s+%#"PRIx64">", + orig_addr, sym->name, addr - sym->start); + } + break; + } +} + static int symbol__disassemble_capstone(char *filename, struct symbol *sym, struct annotate_args *args) { @@ -1458,9 +1521,14 @@ static int symbol__disassemble_capstone(char *filename, struct symbol *sym, count = cs_disasm(handle, buf, len, start, len, &insn); for (i = 0, offset = 0; i < count; i++) { - scnprintf(disasm_buf, sizeof(disasm_buf), - " %-7s %s", - insn[i].mnemonic, insn[i].op_str); + int printed; + + printed = scnprintf(disasm_buf, sizeof(disasm_buf), + " %-7s %s", + insn[i].mnemonic, insn[i].op_str); + print_capstone_detail(&insn[i], disasm_buf + printed, + sizeof(disasm_buf) - printed, args, + start + offset); args->offset = offset; args->line = disasm_buf; -- cgit v1.2.3 From b6347cb5e04e9c1d17342ab46e2ace2d448de727 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 3 Apr 2024 11:44:19 -0300 Subject: perf annotate: Initialize 'arch' variable not to trip some -Werror=maybe-uninitialized In some older distros the build is failing due to -Werror=maybe-uninitialized, in this case we know that this isn't the case because 'arch' gets initialized by evsel__get_arch(), so make sure it is initialized to NULL before returning from evsel__get_arch(), as suggested by Ian Rogers. E.g.: 32 17.12 opensuse:15.5 : FAIL gcc version 7.5.0 (SUSE Linux) util/annotate.c: In function 'hist_entry__get_data_type': util/annotate.c:2269:15: error: 'arch' may be used uninitialized in this function [-Werror=maybe-uninitialized] struct arch *arch; ^~~~ cc1: all warnings being treated as errors 43 7.30 ubuntu:18.04-x-powerpc64el : FAIL gcc version 7.5.0 (Ubuntu 7.5.0-3ubuntu1~18.04) util/annotate.c: In function 'hist_entry__get_data_type': util/annotate.c:2351:36: error: 'arch' may be used uninitialized in this function [-Werror=maybe-uninitialized] if (map__dso(ms->map)->kernel && arch__is(arch, "x86") && ^~~~~~~~~~~~~~~~~~~~~ cc1: all warnings being treated as errors Suggested-by: Ian Rogers Reviewed-by: Ian Rogers Cc: Adrian Hunter Cc: Jiri Olsa Cc: Namhyung Kim Link: https://lore.kernel.org/lkml/CAP-5=fUqtjxAsmdGrnkjhUTLHs-JvV10TtxyocpYDJK_+LYTiQ@mail.gmail.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index b795f27f2602..35235147b111 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -838,8 +838,10 @@ static int evsel__get_arch(struct evsel *evsel, struct arch **parch) struct arch *arch; int err; - if (!arch_name) + if (!arch_name) { + *parch = NULL; return errno; + } *parch = arch = arch__find(arch_name); if (arch == NULL) { -- cgit v1.2.3 From 38ab60132b0d477e19d841daed701b491c20b465 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Mon, 1 Apr 2024 14:08:03 -0700 Subject: perf script: Support 32bit code under 64bit OS with capstone Use the DSO to resolve whether an IP is 32bit or 64bit and use that to configure capstone to the correct mode. This allows to correctly disassemble 32bit code under a 64bit OS. % cat > loop.c volatile int var; int main(void) { int i; for (i = 0; i < 100000; i++) var++; } % gcc -m32 -o loop loop.c % perf record -e cycles:u ./loop % perf script -F +disasm loop 82665 1833176.618023: 1 cycles:u: f7eed500 _start+0x0 (/usr/lib/ld-linux.so.2) movl %esp, %eax loop 82665 1833176.618029: 1 cycles:u: f7eed500 _start+0x0 (/usr/lib/ld-linux.so.2) movl %esp, %eax loop 82665 1833176.618031: 7 cycles:u: f7eed500 _start+0x0 (/usr/lib/ld-linux.so.2) movl %esp, %eax loop 82665 1833176.618034: 91 cycles:u: f7eed500 _start+0x0 (/usr/lib/ld-linux.so.2) movl %esp, %eax loop 82665 1833176.618036: 1242 cycles:u: f7eed500 _start+0x0 (/usr/lib/ld-linux.so.2) movl %esp, %eax Reviewed-by: Adrian Hunter Acked-by: Thomas Richter Signed-off-by: Andi Kleen Link: https://lore.kernel.org/r/20240401210925.209671-2-ak@linux.intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-script.c | 9 +++++---- tools/perf/util/print_insn.c | 27 ++++++++++++++++++++++----- tools/perf/util/print_insn.h | 2 +- 3 files changed, 28 insertions(+), 10 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index 6a274e27b108..a711bedace47 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -1517,7 +1517,8 @@ void script_fetch_insn(struct perf_sample *sample, struct thread *thread, static int perf_sample__fprintf_insn(struct perf_sample *sample, struct perf_event_attr *attr, struct thread *thread, - struct machine *machine, FILE *fp) + struct machine *machine, FILE *fp, + struct addr_location *al) { int printed = 0; @@ -1531,7 +1532,7 @@ static int perf_sample__fprintf_insn(struct perf_sample *sample, } if (PRINT_FIELD(DISASM) && sample->insn_len) { printed += fprintf(fp, "\t\t"); - printed += sample__fprintf_insn_asm(sample, thread, machine, fp); + printed += sample__fprintf_insn_asm(sample, thread, machine, fp, al); } if (PRINT_FIELD(BRSTACKINSN) || PRINT_FIELD(BRSTACKINSNLEN)) printed += perf_sample__fprintf_brstackinsn(sample, thread, attr, machine, fp); @@ -1606,7 +1607,7 @@ static int perf_sample__fprintf_bts(struct perf_sample *sample, if (print_srcline_last) printed += map__fprintf_srcline(al->map, al->addr, "\n ", fp); - printed += perf_sample__fprintf_insn(sample, attr, thread, machine, fp); + printed += perf_sample__fprintf_insn(sample, attr, thread, machine, fp, al); printed += fprintf(fp, "\n"); if (PRINT_FIELD(SRCCODE)) { int ret = map__fprintf_srccode(al->map, al->addr, stdout, @@ -2259,7 +2260,7 @@ static void process_event(struct perf_script *script, if (evsel__is_bpf_output(evsel) && PRINT_FIELD(BPF_OUTPUT)) perf_sample__fprintf_bpf_output(sample, fp); - perf_sample__fprintf_insn(sample, attr, thread, machine, fp); + perf_sample__fprintf_insn(sample, attr, thread, machine, fp, al); if (PRINT_FIELD(PHYS_ADDR)) fprintf(fp, "%16" PRIx64, sample->phys_addr); diff --git a/tools/perf/util/print_insn.c b/tools/perf/util/print_insn.c index 459e0e93d7b1..32dc9dad9cf2 100644 --- a/tools/perf/util/print_insn.c +++ b/tools/perf/util/print_insn.c @@ -12,6 +12,8 @@ #include "machine.h" #include "thread.h" #include "print_insn.h" +#include "map.h" +#include "dso.h" size_t sample__fprintf_insn_raw(struct perf_sample *sample, FILE *fp) { @@ -28,12 +30,12 @@ size_t sample__fprintf_insn_raw(struct perf_sample *sample, FILE *fp) #ifdef HAVE_LIBCAPSTONE_SUPPORT #include -static int capstone_init(struct machine *machine, csh *cs_handle) +static int capstone_init(struct machine *machine, csh *cs_handle, bool is64) { cs_arch arch; cs_mode mode; - if (machine__is(machine, "x86_64")) { + if (machine__is(machine, "x86_64") && is64) { arch = CS_ARCH_X86; mode = CS_MODE_64; } else if (machine__normalized_is(machine, "x86")) { @@ -93,17 +95,31 @@ static size_t print_insn_x86(struct perf_sample *sample, struct thread *thread, return printed; } +static bool is64bitip(struct machine *machine, struct addr_location *al) +{ + const struct dso *dso = al->map ? map__dso(al->map) : NULL; + + if (dso) + return dso->is_64_bit; + + return machine__is(machine, "x86_64") || + machine__normalized_is(machine, "arm64") || + machine__normalized_is(machine, "s390"); +} + size_t sample__fprintf_insn_asm(struct perf_sample *sample, struct thread *thread, - struct machine *machine, FILE *fp) + struct machine *machine, FILE *fp, + struct addr_location *al) { csh cs_handle; cs_insn *insn; size_t count; size_t printed = 0; int ret; + bool is64bit = is64bitip(machine, al); /* TODO: Try to initiate capstone only once but need a proper place. */ - ret = capstone_init(machine, &cs_handle); + ret = capstone_init(machine, &cs_handle, is64bit); if (ret < 0) { /* fallback */ return sample__fprintf_insn_raw(sample, fp); @@ -128,7 +144,8 @@ size_t sample__fprintf_insn_asm(struct perf_sample *sample, struct thread *threa size_t sample__fprintf_insn_asm(struct perf_sample *sample __maybe_unused, struct thread *thread __maybe_unused, struct machine *machine __maybe_unused, - FILE *fp __maybe_unused) + FILE *fp __maybe_unused, + struct addr_location *al __maybe_unused) { return 0; } diff --git a/tools/perf/util/print_insn.h b/tools/perf/util/print_insn.h index 465bdcfcc2fd..6447dd41b543 100644 --- a/tools/perf/util/print_insn.h +++ b/tools/perf/util/print_insn.h @@ -10,7 +10,7 @@ struct thread; struct machine; size_t sample__fprintf_insn_asm(struct perf_sample *sample, struct thread *thread, - struct machine *machine, FILE *fp); + struct machine *machine, FILE *fp, struct addr_location *al); size_t sample__fprintf_insn_raw(struct perf_sample *sample, FILE *fp); #endif /* PERF_PRINT_INSN_H */ -- cgit v1.2.3 From d812044688dfe73e1b309689d58d06f50a15e618 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Mon, 1 Apr 2024 14:08:04 -0700 Subject: perf script: Add capstone support for '-F +brstackdisasm' Support capstone output for the '-F +brstackinsn' branch dump. The new output is enabled with the new field 'brstackdisasm'. This was possible before with --xed, but now also allow it for users that don't have xed using the builtin capstone support. Before: perf record -b emacs -Q --batch '()' perf script -F +brstackinsn ... emacs 55778 1814366.755945: 151564 cycles:P: 7f0ab2d17192 intel_check_word.constprop.0+0x162 (/usr/lib64/ld-linux-x86-64.s> intel_check_word.constprop.0+237: 00007f0ab2d1711d insn: 75 e6 # PRED 3 cycles [3] 00007f0ab2d17105 insn: 73 51 00007f0ab2d17107 insn: 48 89 c1 00007f0ab2d1710a insn: 48 39 ca 00007f0ab2d1710d insn: 73 96 00007f0ab2d1710f insn: 48 8d 04 11 00007f0ab2d17113 insn: 48 d1 e8 00007f0ab2d17116 insn: 49 8d 34 c1 00007f0ab2d1711a insn: 44 3a 06 00007f0ab2d1711d insn: 75 e6 # PRED 3 cycles [6] 3.00 IPC 00007f0ab2d17105 insn: 73 51 # PRED 1 cycles [7] 1.00 IPC 00007f0ab2d17158 insn: 48 8d 50 01 00007f0ab2d1715c insn: eb 92 # PRED 1 cycles [8] 2.00 IPC 00007f0ab2d170f0 insn: 48 39 ca 00007f0ab2d170f3 insn: 73 b0 # PRED 1 cycles [9] 2.00 IPC After (perf must be compiled with capstone): perf script -F +brstackdisasm ... emacs 55778 1814366.755945: 151564 cycles:P: 7f0ab2d17192 intel_check_word.constprop.0+0x162 (/usr/lib64/ld-linux-x86-64.s> intel_check_word.constprop.0+237: 00007f0ab2d1711d jne intel_check_word.constprop.0+0xd5 # PRED 3 cycles [3] 00007f0ab2d17105 jae intel_check_word.constprop.0+0x128 00007f0ab2d17107 movq %rax, %rcx 00007f0ab2d1710a cmpq %rcx, %rdx 00007f0ab2d1710d jae intel_check_word.constprop.0+0x75 00007f0ab2d1710f leaq (%rcx, %rdx), %rax 00007f0ab2d17113 shrq $1, %rax 00007f0ab2d17116 leaq (%r9, %rax, 8), %rsi 00007f0ab2d1711a cmpb (%rsi), %r8b 00007f0ab2d1711d jne intel_check_word.constprop.0+0xd5 # PRED 3 cycles [6] 3.00 IPC 00007f0ab2d17105 jae intel_check_word.constprop.0+0x128 # PRED 1 cycles [7] 1.00 IPC 00007f0ab2d17158 leaq 1(%rax), %rdx 00007f0ab2d1715c jmp intel_check_word.constprop.0+0xc0 # PRED 1 cycles [8] 2.00 IPC 00007f0ab2d170f0 cmpq %rcx, %rdx 00007f0ab2d170f3 jae intel_check_word.constprop.0+0x75 # PRED 1 cycles [9] 2.00 IPC Reviewed-by: Adrian Hunter Signed-off-by: Andi Kleen Link: https://lore.kernel.org/r/20240401210925.209671-3-ak@linux.intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/Documentation/perf-script.txt | 7 +++-- tools/perf/builtin-script.c | 32 +++++++++++++++----- tools/perf/util/dump-insn.h | 1 + tools/perf/util/print_insn.c | 52 ++++++++++++++++++++++++++++++++ tools/perf/util/print_insn.h | 3 ++ 5 files changed, 86 insertions(+), 9 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/Documentation/perf-script.txt b/tools/perf/Documentation/perf-script.txt index 005e51df855e..ff086ef05a0c 100644 --- a/tools/perf/Documentation/perf-script.txt +++ b/tools/perf/Documentation/perf-script.txt @@ -132,9 +132,9 @@ OPTIONS Comma separated list of fields to print. Options are: comm, tid, pid, time, cpu, event, trace, ip, sym, dso, dsoff, addr, symoff, srcline, period, iregs, uregs, brstack, brstacksym, flags, bpf-output, - brstackinsn, brstackinsnlen, brstackoff, callindent, insn, disasm, + brstackinsn, brstackinsnlen, brstackdisasm, brstackoff, callindent, insn, disasm, insnlen, synth, phys_addr, metric, misc, srccode, ipc, data_page_size, - code_page_size, ins_lat, machine_pid, vcpu, cgroup, retire_lat. + code_page_size, ins_lat, machine_pid, vcpu, cgroup, retire_lat, Field list can be prepended with the type, trace, sw or hw, to indicate to which event type the field list applies. @@ -257,6 +257,9 @@ OPTIONS can’t know the next sequential instruction after an unconditional branch unless you calculate that based on its length. + brstackdisasm acts like brstackinsn, but will print disassembled instructions if + perf is built with the capstone library. + The brstackoff field will print an offset into a specific dso/binary. With the metric option perf script can compute metrics for diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index a711bedace47..dd10f158ed0c 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -136,6 +136,7 @@ enum perf_output_field { PERF_OUTPUT_RETIRE_LAT = 1ULL << 40, PERF_OUTPUT_DSOFF = 1ULL << 41, PERF_OUTPUT_DISASM = 1ULL << 42, + PERF_OUTPUT_BRSTACKDISASM = 1ULL << 43, }; struct perf_script { @@ -210,6 +211,7 @@ struct output_option { {.str = "vcpu", .field = PERF_OUTPUT_VCPU}, {.str = "cgroup", .field = PERF_OUTPUT_CGROUP}, {.str = "retire_lat", .field = PERF_OUTPUT_RETIRE_LAT}, + {.str = "brstackdisasm", .field = PERF_OUTPUT_BRSTACKDISASM}, }; enum { @@ -510,7 +512,8 @@ static int evsel__check_attr(struct evsel *evsel, struct perf_session *session) "selected. Hence, no address to lookup the source line number.\n"); return -EINVAL; } - if ((PRINT_FIELD(BRSTACKINSN) || PRINT_FIELD(BRSTACKINSNLEN)) && !allow_user_set && + if ((PRINT_FIELD(BRSTACKINSN) || PRINT_FIELD(BRSTACKINSNLEN) || PRINT_FIELD(BRSTACKDISASM)) + && !allow_user_set && !(evlist__combined_branch_type(session->evlist) & PERF_SAMPLE_BRANCH_ANY)) { pr_err("Display of branch stack assembler requested, but non all-branch filter set\n" "Hint: run 'perf record -b ...'\n"); @@ -1162,6 +1165,20 @@ out: return ret; } +static const char *any_dump_insn(struct perf_event_attr *attr __maybe_unused, + struct perf_insn *x, uint64_t ip, + u8 *inbuf, int inlen, int *lenp) +{ +#ifdef HAVE_LIBCAPSTONE_SUPPORT + if (PRINT_FIELD(BRSTACKDISASM)) { + const char *p = cs_dump_insn(x, ip, inbuf, inlen, lenp); + if (p) + return p; + } +#endif + return dump_insn(x, ip, inbuf, inlen, lenp); +} + static int ip__fprintf_jump(uint64_t ip, struct branch_entry *en, struct perf_insn *x, u8 *inbuf, int len, int insn, FILE *fp, int *total_cycles, @@ -1170,7 +1187,7 @@ static int ip__fprintf_jump(uint64_t ip, struct branch_entry *en, { int ilen = 0; int printed = fprintf(fp, "\t%016" PRIx64 "\t%-30s\t", ip, - dump_insn(x, ip, inbuf, len, &ilen)); + any_dump_insn(attr, x, ip, inbuf, len, &ilen)); if (PRINT_FIELD(BRSTACKINSNLEN)) printed += fprintf(fp, "ilen: %d\t", ilen); @@ -1262,6 +1279,7 @@ static int perf_sample__fprintf_brstackinsn(struct perf_sample *sample, nr = max_blocks + 1; x.thread = thread; + x.machine = machine; x.cpu = sample->cpu; printed += fprintf(fp, "%c", '\n'); @@ -1313,7 +1331,7 @@ static int perf_sample__fprintf_brstackinsn(struct perf_sample *sample, } else { ilen = 0; printed += fprintf(fp, "\t%016" PRIx64 "\t%s", ip, - dump_insn(&x, ip, buffer + off, len - off, &ilen)); + any_dump_insn(attr, &x, ip, buffer + off, len - off, &ilen)); if (PRINT_FIELD(BRSTACKINSNLEN)) printed += fprintf(fp, "\tilen: %d", ilen); printed += fprintf(fp, "\n"); @@ -1361,7 +1379,7 @@ static int perf_sample__fprintf_brstackinsn(struct perf_sample *sample, goto out; ilen = 0; printed += fprintf(fp, "\t%016" PRIx64 "\t%s", sample->ip, - dump_insn(&x, sample->ip, buffer, len, &ilen)); + any_dump_insn(attr, &x, sample->ip, buffer, len, &ilen)); if (PRINT_FIELD(BRSTACKINSNLEN)) printed += fprintf(fp, "\tilen: %d", ilen); printed += fprintf(fp, "\n"); @@ -1372,7 +1390,7 @@ static int perf_sample__fprintf_brstackinsn(struct perf_sample *sample, for (off = 0; off <= end - start; off += ilen) { ilen = 0; printed += fprintf(fp, "\t%016" PRIx64 "\t%s", start + off, - dump_insn(&x, start + off, buffer + off, len - off, &ilen)); + any_dump_insn(attr, &x, start + off, buffer + off, len - off, &ilen)); if (PRINT_FIELD(BRSTACKINSNLEN)) printed += fprintf(fp, "\tilen: %d", ilen); printed += fprintf(fp, "\n"); @@ -1534,7 +1552,7 @@ static int perf_sample__fprintf_insn(struct perf_sample *sample, printed += fprintf(fp, "\t\t"); printed += sample__fprintf_insn_asm(sample, thread, machine, fp, al); } - if (PRINT_FIELD(BRSTACKINSN) || PRINT_FIELD(BRSTACKINSNLEN)) + if (PRINT_FIELD(BRSTACKINSN) || PRINT_FIELD(BRSTACKINSNLEN) || PRINT_FIELD(BRSTACKDISASM)) printed += perf_sample__fprintf_brstackinsn(sample, thread, attr, machine, fp); return printed; @@ -3940,7 +3958,7 @@ int cmd_script(int argc, const char **argv) "Fields: comm,tid,pid,time,cpu,event,trace,ip,sym,dso,dsoff," "addr,symoff,srcline,period,iregs,uregs,brstack," "brstacksym,flags,data_src,weight,bpf-output,brstackinsn," - "brstackinsnlen,brstackoff,callindent,insn,disasm,insnlen,synth," + "brstackinsnlen,brstackdisasm,brstackoff,callindent,insn,disasm,insnlen,synth," "phys_addr,metric,misc,srccode,ipc,tod,data_page_size," "code_page_size,ins_lat,machine_pid,vcpu,cgroup,retire_lat", parse_output_fields), diff --git a/tools/perf/util/dump-insn.h b/tools/perf/util/dump-insn.h index 650125061530..4a7797dd6d09 100644 --- a/tools/perf/util/dump-insn.h +++ b/tools/perf/util/dump-insn.h @@ -11,6 +11,7 @@ struct thread; struct perf_insn { /* Initialized by callers: */ struct thread *thread; + struct machine *machine; u8 cpumode; bool is64bit; int cpu; diff --git a/tools/perf/util/print_insn.c b/tools/perf/util/print_insn.c index 32dc9dad9cf2..8825330d435f 100644 --- a/tools/perf/util/print_insn.c +++ b/tools/perf/util/print_insn.c @@ -12,6 +12,7 @@ #include "machine.h" #include "thread.h" #include "print_insn.h" +#include "dump-insn.h" #include "map.h" #include "dso.h" @@ -71,6 +72,57 @@ static int capstone_init(struct machine *machine, csh *cs_handle, bool is64) return 0; } +static void dump_insn_x86(struct thread *thread, cs_insn *insn, struct perf_insn *x) +{ + struct addr_location al; + bool printed = false; + + if (insn->detail && insn->detail->x86.op_count == 1) { + cs_x86_op *op = &insn->detail->x86.operands[0]; + + addr_location__init(&al); + if (op->type == X86_OP_IMM && + thread__find_symbol(thread, x->cpumode, op->imm, &al) && + al.sym && + al.addr < al.sym->end) { + snprintf(x->out, sizeof(x->out), "%s %s+%#" PRIx64 " [%#" PRIx64 "]", insn[0].mnemonic, + al.sym->name, al.addr - al.sym->start, op->imm); + printed = true; + } + addr_location__exit(&al); + } + + if (!printed) + snprintf(x->out, sizeof(x->out), "%s %s", insn[0].mnemonic, insn[0].op_str); +} + +const char *cs_dump_insn(struct perf_insn *x, uint64_t ip, + u8 *inbuf, int inlen, int *lenp) +{ + int ret; + int count; + cs_insn *insn; + csh cs_handle; + + ret = capstone_init(x->machine, &cs_handle, x->is64bit); + if (ret < 0) + return NULL; + + count = cs_disasm(cs_handle, (uint8_t *)inbuf, inlen, ip, 1, &insn); + if (count > 0) { + if (machine__normalized_is(x->machine, "x86")) + dump_insn_x86(x->thread, &insn[0], x); + else + snprintf(x->out, sizeof(x->out), "%s %s", + insn[0].mnemonic, insn[0].op_str); + *lenp = insn->size; + cs_free(insn, count); + } else { + return NULL; + } + return x->out; +} + static size_t print_insn_x86(struct perf_sample *sample, struct thread *thread, cs_insn *insn, FILE *fp) { diff --git a/tools/perf/util/print_insn.h b/tools/perf/util/print_insn.h index 6447dd41b543..c2a6391a45ce 100644 --- a/tools/perf/util/print_insn.h +++ b/tools/perf/util/print_insn.h @@ -8,9 +8,12 @@ struct perf_sample; struct thread; struct machine; +struct perf_insn; size_t sample__fprintf_insn_asm(struct perf_sample *sample, struct thread *thread, struct machine *machine, FILE *fp, struct addr_location *al); size_t sample__fprintf_insn_raw(struct perf_sample *sample, FILE *fp); +const char *cs_dump_insn(struct perf_insn *x, uint64_t ip, + u8 *inbuf, int inlen, int *lenp); #endif /* PERF_PRINT_INSN_H */ -- cgit v1.2.3 From 218c200f677d8af46a8540319e4d26c52b3277a5 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 1 Apr 2024 14:08:05 -0700 Subject: perf script: Consolidate capstone print functions Consolidate capstone print functions, to reduce duplication. Amend call sites to use a file pointer for output, which is consistent with most perf tools print functions. Add print_opts with an option to print also the hex value of a resolved symbol+offset. Signed-off-by: Adrian Hunter Link: https://lore.kernel.org/r/20240401210925.209671-4-ak@linux.intel.com Signed-off-by: Andi Kleen [ Added missing inttypes.h include to use PRIx64 in util/print_insn.c ] Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-script.c | 43 +++++++++++------- tools/perf/util/print_insn.c | 104 +++++++++++++++---------------------------- tools/perf/util/print_insn.h | 7 ++- 3 files changed, 68 insertions(+), 86 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index dd10f158ed0c..647cb31a47c8 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -1165,18 +1165,29 @@ out: return ret; } -static const char *any_dump_insn(struct perf_event_attr *attr __maybe_unused, - struct perf_insn *x, uint64_t ip, - u8 *inbuf, int inlen, int *lenp) +static int any_dump_insn(struct perf_event_attr *attr __maybe_unused, + struct perf_insn *x, uint64_t ip, + u8 *inbuf, int inlen, int *lenp, + FILE *fp) { #ifdef HAVE_LIBCAPSTONE_SUPPORT if (PRINT_FIELD(BRSTACKDISASM)) { - const char *p = cs_dump_insn(x, ip, inbuf, inlen, lenp); - if (p) - return p; + int printed = fprintf_insn_asm(x->machine, x->thread, x->cpumode, x->is64bit, + (uint8_t *)inbuf, inlen, ip, lenp, + PRINT_INSN_IMM_HEX, fp); + + if (printed > 0) + return printed; } #endif - return dump_insn(x, ip, inbuf, inlen, lenp); + return fprintf(fp, "%s", dump_insn(x, ip, inbuf, inlen, lenp)); +} + +static int add_padding(FILE *fp, int printed, int padding) +{ + if (printed >= 0 && printed < padding) + printed += fprintf(fp, "%*s", padding - printed, ""); + return printed; } static int ip__fprintf_jump(uint64_t ip, struct branch_entry *en, @@ -1186,8 +1197,10 @@ static int ip__fprintf_jump(uint64_t ip, struct branch_entry *en, struct thread *thread) { int ilen = 0; - int printed = fprintf(fp, "\t%016" PRIx64 "\t%-30s\t", ip, - any_dump_insn(attr, x, ip, inbuf, len, &ilen)); + int printed = fprintf(fp, "\t%016" PRIx64 "\t", ip); + + printed += add_padding(fp, any_dump_insn(attr, x, ip, inbuf, len, &ilen, fp), 30); + printed += fprintf(fp, "\t"); if (PRINT_FIELD(BRSTACKINSNLEN)) printed += fprintf(fp, "ilen: %d\t", ilen); @@ -1330,8 +1343,8 @@ static int perf_sample__fprintf_brstackinsn(struct perf_sample *sample, break; } else { ilen = 0; - printed += fprintf(fp, "\t%016" PRIx64 "\t%s", ip, - any_dump_insn(attr, &x, ip, buffer + off, len - off, &ilen)); + printed += fprintf(fp, "\t%016" PRIx64 "\t", ip); + printed += any_dump_insn(attr, &x, ip, buffer + off, len - off, &ilen, fp); if (PRINT_FIELD(BRSTACKINSNLEN)) printed += fprintf(fp, "\tilen: %d", ilen); printed += fprintf(fp, "\n"); @@ -1378,8 +1391,8 @@ static int perf_sample__fprintf_brstackinsn(struct perf_sample *sample, if (len <= 0) goto out; ilen = 0; - printed += fprintf(fp, "\t%016" PRIx64 "\t%s", sample->ip, - any_dump_insn(attr, &x, sample->ip, buffer, len, &ilen)); + printed += fprintf(fp, "\t%016" PRIx64 "\t", sample->ip); + printed += any_dump_insn(attr, &x, sample->ip, buffer, len, &ilen, fp); if (PRINT_FIELD(BRSTACKINSNLEN)) printed += fprintf(fp, "\tilen: %d", ilen); printed += fprintf(fp, "\n"); @@ -1389,8 +1402,8 @@ static int perf_sample__fprintf_brstackinsn(struct perf_sample *sample, } for (off = 0; off <= end - start; off += ilen) { ilen = 0; - printed += fprintf(fp, "\t%016" PRIx64 "\t%s", start + off, - any_dump_insn(attr, &x, start + off, buffer + off, len - off, &ilen)); + printed += fprintf(fp, "\t%016" PRIx64 "\t", start + off); + printed += any_dump_insn(attr, &x, start + off, buffer + off, len - off, &ilen, fp); if (PRINT_FIELD(BRSTACKINSNLEN)) printed += fprintf(fp, "\tilen: %d", ilen); printed += fprintf(fp, "\n"); diff --git a/tools/perf/util/print_insn.c b/tools/perf/util/print_insn.c index 8825330d435f..aab11d8e1b1d 100644 --- a/tools/perf/util/print_insn.c +++ b/tools/perf/util/print_insn.c @@ -4,6 +4,7 @@ * * Author(s): Changbin Du */ +#include #include #include #include "debug.h" @@ -72,59 +73,8 @@ static int capstone_init(struct machine *machine, csh *cs_handle, bool is64) return 0; } -static void dump_insn_x86(struct thread *thread, cs_insn *insn, struct perf_insn *x) -{ - struct addr_location al; - bool printed = false; - - if (insn->detail && insn->detail->x86.op_count == 1) { - cs_x86_op *op = &insn->detail->x86.operands[0]; - - addr_location__init(&al); - if (op->type == X86_OP_IMM && - thread__find_symbol(thread, x->cpumode, op->imm, &al) && - al.sym && - al.addr < al.sym->end) { - snprintf(x->out, sizeof(x->out), "%s %s+%#" PRIx64 " [%#" PRIx64 "]", insn[0].mnemonic, - al.sym->name, al.addr - al.sym->start, op->imm); - printed = true; - } - addr_location__exit(&al); - } - - if (!printed) - snprintf(x->out, sizeof(x->out), "%s %s", insn[0].mnemonic, insn[0].op_str); -} - -const char *cs_dump_insn(struct perf_insn *x, uint64_t ip, - u8 *inbuf, int inlen, int *lenp) -{ - int ret; - int count; - cs_insn *insn; - csh cs_handle; - - ret = capstone_init(x->machine, &cs_handle, x->is64bit); - if (ret < 0) - return NULL; - - count = cs_disasm(cs_handle, (uint8_t *)inbuf, inlen, ip, 1, &insn); - if (count > 0) { - if (machine__normalized_is(x->machine, "x86")) - dump_insn_x86(x->thread, &insn[0], x); - else - snprintf(x->out, sizeof(x->out), "%s %s", - insn[0].mnemonic, insn[0].op_str); - *lenp = insn->size; - cs_free(insn, count); - } else { - return NULL; - } - return x->out; -} - -static size_t print_insn_x86(struct perf_sample *sample, struct thread *thread, - cs_insn *insn, FILE *fp) +static size_t print_insn_x86(struct thread *thread, u8 cpumode, cs_insn *insn, + int print_opts, FILE *fp) { struct addr_location al; size_t printed = 0; @@ -134,9 +84,11 @@ static size_t print_insn_x86(struct perf_sample *sample, struct thread *thread, addr_location__init(&al); if (op->type == X86_OP_IMM && - thread__find_symbol(thread, sample->cpumode, op->imm, &al)) { + thread__find_symbol(thread, cpumode, op->imm, &al)) { printed += fprintf(fp, "%s ", insn[0].mnemonic); printed += symbol__fprintf_symname_offs(al.sym, &al, fp); + if (print_opts & PRINT_INSN_IMM_HEX) + printed += fprintf(fp, " [%#" PRIx64 "]", op->imm); addr_location__exit(&al); return printed; } @@ -159,39 +111,53 @@ static bool is64bitip(struct machine *machine, struct addr_location *al) machine__normalized_is(machine, "s390"); } -size_t sample__fprintf_insn_asm(struct perf_sample *sample, struct thread *thread, - struct machine *machine, FILE *fp, - struct addr_location *al) +ssize_t fprintf_insn_asm(struct machine *machine, struct thread *thread, u8 cpumode, + bool is64bit, const uint8_t *code, size_t code_size, + uint64_t ip, int *lenp, int print_opts, FILE *fp) { - csh cs_handle; + size_t printed; cs_insn *insn; + csh cs_handle; size_t count; - size_t printed = 0; int ret; - bool is64bit = is64bitip(machine, al); /* TODO: Try to initiate capstone only once but need a proper place. */ ret = capstone_init(machine, &cs_handle, is64bit); - if (ret < 0) { - /* fallback */ - return sample__fprintf_insn_raw(sample, fp); - } + if (ret < 0) + return ret; - count = cs_disasm(cs_handle, (uint8_t *)sample->insn, sample->insn_len, - sample->ip, 1, &insn); + count = cs_disasm(cs_handle, code, code_size, ip, 1, &insn); if (count > 0) { if (machine__normalized_is(machine, "x86")) - printed += print_insn_x86(sample, thread, &insn[0], fp); + printed = print_insn_x86(thread, cpumode, &insn[0], print_opts, fp); else - printed += fprintf(fp, "%s %s", insn[0].mnemonic, insn[0].op_str); + printed = fprintf(fp, "%s %s", insn[0].mnemonic, insn[0].op_str); + if (lenp) + *lenp = insn->size; cs_free(insn, count); } else { - printed += fprintf(fp, "illegal instruction"); + printed = -1; } cs_close(&cs_handle); return printed; } + +size_t sample__fprintf_insn_asm(struct perf_sample *sample, struct thread *thread, + struct machine *machine, FILE *fp, + struct addr_location *al) +{ + bool is64bit = is64bitip(machine, al); + ssize_t printed; + + printed = fprintf_insn_asm(machine, thread, sample->cpumode, is64bit, + (uint8_t *)sample->insn, sample->insn_len, + sample->ip, NULL, 0, fp); + if (printed < 0) + return sample__fprintf_insn_raw(sample, fp); + + return printed; +} #else size_t sample__fprintf_insn_asm(struct perf_sample *sample __maybe_unused, struct thread *thread __maybe_unused, diff --git a/tools/perf/util/print_insn.h b/tools/perf/util/print_insn.h index c2a6391a45ce..07d11af3fc1c 100644 --- a/tools/perf/util/print_insn.h +++ b/tools/perf/util/print_insn.h @@ -10,10 +10,13 @@ struct thread; struct machine; struct perf_insn; +#define PRINT_INSN_IMM_HEX (1<<0) + size_t sample__fprintf_insn_asm(struct perf_sample *sample, struct thread *thread, struct machine *machine, FILE *fp, struct addr_location *al); size_t sample__fprintf_insn_raw(struct perf_sample *sample, FILE *fp); -const char *cs_dump_insn(struct perf_insn *x, uint64_t ip, - u8 *inbuf, int inlen, int *lenp); +ssize_t fprintf_insn_asm(struct machine *machine, struct thread *thread, u8 cpumode, + bool is64bit, const uint8_t *code, size_t code_size, + uint64_t ip, int *lenp, int print_opts, FILE *fp); #endif /* PERF_PRINT_INSN_H */ -- cgit v1.2.3 From aaf494cf483a1a835c44e942861429b30a00cab0 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 4 Apr 2024 10:57:08 -0700 Subject: perf annotate: Fix annotation_calc_lines() to pass correct address to get_srcline() It should pass a proper address (i.e. suitable for objdump or addr2line) to get_srcline() in order to work correctly. It used to pass an address with map__rip_2objdump() as the second argument but later it's changed to use notes->start. It's ok in normal cases but it can be changed when annotate_opts.full_addr is set. So let's convert the address directly instead of using the notes->start. Also the last argument is an IP to print symbol offset if requested. So it should pass symbol-relative address. Fixes: 7d18a824b5e57ddd ("perf annotate: Toggle full address <-> offset display") Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240404175716.1225482-2-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 35235147b111..a330e92c2552 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -1440,7 +1440,7 @@ void annotation__toggle_full_addr(struct annotation *notes, struct map_symbol *m annotation__update_column_widths(notes); } -static void annotation__calc_lines(struct annotation *notes, struct map *map, +static void annotation__calc_lines(struct annotation *notes, struct map_symbol *ms, struct rb_root *root) { struct annotation_line *al; @@ -1448,6 +1448,7 @@ static void annotation__calc_lines(struct annotation *notes, struct map *map, list_for_each_entry(al, ¬es->src->source, node) { double percent_max = 0.0; + u64 addr; int i; for (i = 0; i < al->data_nr; i++) { @@ -1463,8 +1464,9 @@ static void annotation__calc_lines(struct annotation *notes, struct map *map, if (percent_max <= 0.5) continue; - al->path = get_srcline(map__dso(map), notes->start + al->offset, NULL, - false, true, notes->start + al->offset); + addr = map__rip_2objdump(ms->map, ms->sym->start); + al->path = get_srcline(map__dso(ms->map), addr + al->offset, NULL, + false, true, ms->sym->start + al->offset); insert_source_line(&tmp_root, al); } @@ -1475,7 +1477,7 @@ static void symbol__calc_lines(struct map_symbol *ms, struct rb_root *root) { struct annotation *notes = symbol__annotation(ms->sym); - annotation__calc_lines(notes, ms->map, root); + annotation__calc_lines(notes, ms, root); } int symbol__tty_annotate2(struct map_symbol *ms, struct evsel *evsel) -- cgit v1.2.3 From bfd98ceb6267712ae298f10144ba0576cc03c70f Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 4 Apr 2024 10:57:09 -0700 Subject: perf annotate: Staticize some local functions I found annotation__mark_jump_targets(), annotation__set_offsets() and annotation__init_column_widths() are only used in the same file. Let's make them static. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240404175716.1225482-3-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 8 +++++--- tools/perf/util/annotate.h | 3 --- 2 files changed, 5 insertions(+), 6 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index a330e92c2552..bbf4894b1309 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -1316,7 +1316,8 @@ bool disasm_line__is_valid_local_jump(struct disasm_line *dl, struct symbol *sym return true; } -void annotation__mark_jump_targets(struct annotation *notes, struct symbol *sym) +static void +annotation__mark_jump_targets(struct annotation *notes, struct symbol *sym) { u64 offset, size = symbol__size(sym); @@ -1347,7 +1348,7 @@ void annotation__mark_jump_targets(struct annotation *notes, struct symbol *sym) } } -void annotation__set_offsets(struct annotation *notes, s64 size) +static void annotation__set_offsets(struct annotation *notes, s64 size) { struct annotation_line *al; struct annotated_source *src = notes->src; @@ -1404,7 +1405,8 @@ static int annotation__max_ins_name(struct annotation *notes) return max_name; } -void annotation__init_column_widths(struct annotation *notes, struct symbol *sym) +static void +annotation__init_column_widths(struct annotation *notes, struct symbol *sym) { notes->widths.addr = notes->widths.target = notes->widths.min_addr = hex_width(symbol__size(sym)); diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h index b3007c9966fd..3f383f38f65f 100644 --- a/tools/perf/util/annotate.h +++ b/tools/perf/util/annotate.h @@ -340,10 +340,7 @@ static inline bool annotation_line__filter(struct annotation_line *al) return annotate_opts.hide_src_code && al->offset == -1; } -void annotation__set_offsets(struct annotation *notes, s64 size); -void annotation__mark_jump_targets(struct annotation *notes, struct symbol *sym); void annotation__update_column_widths(struct annotation *notes); -void annotation__init_column_widths(struct annotation *notes, struct symbol *sym); void annotation__toggle_full_addr(struct annotation *notes, struct map_symbol *ms); static inline struct sym_hist *annotated_source__histogram(struct annotated_source *src, int idx) -- cgit v1.2.3 From 6f157d9af1e42aafa469deb7c16203e39a2f9545 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 4 Apr 2024 10:57:10 -0700 Subject: perf annotate: Introduce annotated_source__get_line() It's a helper function to get annotation_line at the given offset without using the offsets array. The goal is to get rid of the offsets array altogether. It just does the linear search but I think it's better to save memory as it won't be called in a hot path. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240404175716.1225482-4-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browsers/annotate.c | 2 +- tools/perf/util/annotate.c | 26 +++++++++++++++++++++----- tools/perf/util/annotate.h | 3 +++ 3 files changed, 25 insertions(+), 6 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/ui/browsers/annotate.c b/tools/perf/ui/browsers/annotate.c index ec5e21932876..e72583f37972 100644 --- a/tools/perf/ui/browsers/annotate.c +++ b/tools/perf/ui/browsers/annotate.c @@ -186,7 +186,7 @@ static void annotate_browser__draw_current_jump(struct ui_browser *browser) * name right after the '<' token and probably treating this like a * 'call' instruction. */ - target = notes->src->offsets[cursor->ops.target.offset]; + target = annotated_source__get_line(notes->src, cursor->ops.target.offset); if (target == NULL) { ui_helpline__printf("WARN: jump target inconsistency, press 'o', notes->offsets[%#x] = NULL\n", cursor->ops.target.offset); diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index bbf4894b1309..2409d7424c71 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -369,13 +369,25 @@ int addr_map_symbol__account_cycles(struct addr_map_symbol *ams, return err; } +struct annotation_line *annotated_source__get_line(struct annotated_source *src, + s64 offset) +{ + struct annotation_line *al; + + list_for_each_entry(al, &src->source, node) { + if (al->offset == offset) + return al; + } + return NULL; +} + static unsigned annotation__count_insn(struct annotation *notes, u64 start, u64 end) { unsigned n_insn = 0; u64 offset; for (offset = start; offset <= end; offset++) { - if (notes->src->offsets[offset]) + if (annotated_source__get_line(notes->src, offset)) n_insn++; } return n_insn; @@ -405,8 +417,9 @@ static void annotation__count_and_fill(struct annotation *notes, u64 start, u64 return; for (offset = start; offset <= end; offset++) { - struct annotation_line *al = notes->src->offsets[offset]; + struct annotation_line *al; + al = annotated_source__get_line(notes->src, offset); if (al && al->cycles && al->cycles->ipc == 0.0) { al->cycles->ipc = ipc; cover_insn++; @@ -443,7 +456,7 @@ static int annotation__compute_ipc(struct annotation *notes, size_t size) if (ch && ch->cycles) { struct annotation_line *al; - al = notes->src->offsets[offset]; + al = annotated_source__get_line(notes->src, offset); if (al && al->cycles == NULL) { al->cycles = zalloc(sizeof(*al->cycles)); if (al->cycles == NULL) { @@ -466,7 +479,9 @@ static int annotation__compute_ipc(struct annotation *notes, size_t size) struct cyc_hist *ch = ¬es->branch->cycles_hist[offset]; if (ch && ch->cycles) { - struct annotation_line *al = notes->src->offsets[offset]; + struct annotation_line *al; + + al = annotated_source__get_line(notes->src, offset); if (al) zfree(&al->cycles); } @@ -1326,9 +1341,10 @@ annotation__mark_jump_targets(struct annotation *notes, struct symbol *sym) return; for (offset = 0; offset < size; ++offset) { - struct annotation_line *al = notes->src->offsets[offset]; + struct annotation_line *al; struct disasm_line *dl; + al = annotated_source__get_line(notes->src, offset); dl = disasm_line(al); if (!disasm_line__is_valid_local_jump(dl, sym)) diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h index 3f383f38f65f..aa3298c20300 100644 --- a/tools/perf/util/annotate.h +++ b/tools/perf/util/annotate.h @@ -270,6 +270,9 @@ struct annotated_source { u16 max_line_len; }; +struct annotation_line *annotated_source__get_line(struct annotated_source *src, + s64 offset); + /** * struct annotated_branch - basic block and IPC information for a symbol. * -- cgit v1.2.3 From 0c053ed27303660140ee5e9a82c06f923d4f9c73 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 4 Apr 2024 10:57:11 -0700 Subject: perf annotate: Check annotation lines more efficiently In some places, it checks annotated (disasm) lines for each byte. But as it already has a list of disasm lines, it'd be better to traverse the list entries instead of checking every offset with linear search (by annotated_source__get_line() helper). Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240404175716.1225482-5-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 56 +++++++++++++++++++++++++++++----------------- 1 file changed, 35 insertions(+), 21 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 2409d7424c71..d98fc248ba5b 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -383,12 +383,19 @@ struct annotation_line *annotated_source__get_line(struct annotated_source *src, static unsigned annotation__count_insn(struct annotation *notes, u64 start, u64 end) { + struct annotation_line *al; unsigned n_insn = 0; - u64 offset; - for (offset = start; offset <= end; offset++) { - if (annotated_source__get_line(notes->src, offset)) - n_insn++; + al = annotated_source__get_line(notes->src, start); + if (al == NULL) + return 0; + + list_for_each_entry_from(al, ¬es->src->source, node) { + if (al->offset == -1) + continue; + if ((u64)al->offset > end) + break; + n_insn++; } return n_insn; } @@ -405,10 +412,10 @@ static void annotation__count_and_fill(struct annotation *notes, u64 start, u64 { unsigned n_insn; unsigned int cover_insn = 0; - u64 offset; n_insn = annotation__count_insn(notes, start, end); if (n_insn && ch->num && ch->cycles) { + struct annotation_line *al; struct annotated_branch *branch; float ipc = n_insn / ((double)ch->cycles / (double)ch->num); @@ -416,11 +423,16 @@ static void annotation__count_and_fill(struct annotation *notes, u64 start, u64 if (ch->reset >= 0x7fff) return; - for (offset = start; offset <= end; offset++) { - struct annotation_line *al; + al = annotated_source__get_line(notes->src, start); + if (al == NULL) + return; - al = annotated_source__get_line(notes->src, offset); - if (al && al->cycles && al->cycles->ipc == 0.0) { + list_for_each_entry_from(al, ¬es->src->source, node) { + if (al->offset == -1) + continue; + if ((u64)al->offset > end) + break; + if (al->cycles && al->cycles->ipc == 0.0) { al->cycles->ipc = ipc; cover_insn++; } @@ -1268,13 +1280,16 @@ void symbol__annotate_decay_histogram(struct symbol *sym, int evidx) { struct annotation *notes = symbol__annotation(sym); struct sym_hist *h = annotation__histogram(notes, evidx); - int len = symbol__size(sym), offset; + struct annotation_line *al; h->nr_samples = 0; - for (offset = 0; offset < len; ++offset) { + list_for_each_entry(al, ¬es->src->source, node) { struct sym_hist_entry *entry; - entry = annotated_source__hist_entry(notes->src, evidx, offset); + if (al->offset == -1) + continue; + + entry = annotated_source__hist_entry(notes->src, evidx, al->offset); if (entry == NULL) continue; @@ -1334,33 +1349,32 @@ bool disasm_line__is_valid_local_jump(struct disasm_line *dl, struct symbol *sym static void annotation__mark_jump_targets(struct annotation *notes, struct symbol *sym) { - u64 offset, size = symbol__size(sym); + struct annotation_line *al; /* PLT symbols contain external offsets */ if (strstr(sym->name, "@plt")) return; - for (offset = 0; offset < size; ++offset) { - struct annotation_line *al; + list_for_each_entry(al, ¬es->src->source, node) { struct disasm_line *dl; + struct annotation_line *target; - al = annotated_source__get_line(notes->src, offset); dl = disasm_line(al); if (!disasm_line__is_valid_local_jump(dl, sym)) continue; - al = notes->src->offsets[dl->ops.target.offset]; - + target = annotated_source__get_line(notes->src, + dl->ops.target.offset); /* * FIXME: Oops, no jump target? Buggy disassembler? Or do we * have to adjust to the previous offset? */ - if (al == NULL) + if (target == NULL) continue; - if (++al->jump_sources > notes->max_jump_sources) - notes->max_jump_sources = al->jump_sources; + if (++target->jump_sources > notes->max_jump_sources) + notes->max_jump_sources = target->jump_sources; } } -- cgit v1.2.3 From cee9b86043d3690bc9154dabe13e7d53ca44ef9b Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 4 Apr 2024 10:57:12 -0700 Subject: perf annotate: Get rid of offsets array The struct annotated_source.offsets[] is to save pointers to annotation_line at each offset. We can use annotated_source__get_line() helper instead so let's get rid of the array. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240404175716.1225482-6-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browsers/annotate.c | 5 +---- tools/perf/util/annotate.c | 29 ++++++----------------------- tools/perf/util/annotate.h | 2 -- 3 files changed, 7 insertions(+), 29 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/ui/browsers/annotate.c b/tools/perf/ui/browsers/annotate.c index e72583f37972..c93da2ce727f 100644 --- a/tools/perf/ui/browsers/annotate.c +++ b/tools/perf/ui/browsers/annotate.c @@ -977,7 +977,7 @@ int symbol__tui_annotate(struct map_symbol *ms, struct evsel *evsel, dso->annotate_warned = true; symbol__strerror_disassemble(ms, err, msg, sizeof(msg)); ui__error("Couldn't annotate %s:\n%s", sym->name, msg); - goto out_free_offsets; + return -1; } } @@ -996,8 +996,5 @@ int symbol__tui_annotate(struct map_symbol *ms, struct evsel *evsel, if(not_annotated) annotated_source__purge(notes->src); -out_free_offsets: - if(not_annotated) - zfree(¬es->src->offsets); return ret; } diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index d98fc248ba5b..0e8319835986 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -1378,7 +1378,7 @@ annotation__mark_jump_targets(struct annotation *notes, struct symbol *sym) } } -static void annotation__set_offsets(struct annotation *notes, s64 size) +static void annotation__set_index(struct annotation *notes) { struct annotation_line *al; struct annotated_source *src = notes->src; @@ -1393,18 +1393,9 @@ static void annotation__set_offsets(struct annotation *notes, s64 size) if (src->max_line_len < line_len) src->max_line_len = line_len; al->idx = src->nr_entries++; - if (al->offset != -1) { + if (al->offset != -1) al->idx_asm = src->nr_asm_entries++; - /* - * FIXME: short term bandaid to cope with assembly - * routines that comes with labels in the same column - * as the address in objdump, sigh. - * - * E.g. copy_user_generic_unrolled - */ - if (al->offset < size) - notes->src->offsets[al->offset] = al; - } else + else al->idx_asm = -1; } } @@ -1835,25 +1826,21 @@ int symbol__annotate2(struct map_symbol *ms, struct evsel *evsel, size_t size = symbol__size(sym); int nr_pcnt = 1, err; - notes->src->offsets = zalloc(size * sizeof(struct annotation_line *)); - if (notes->src->offsets == NULL) - return ENOMEM; - if (evsel__is_group_event(evsel)) nr_pcnt = evsel->core.nr_members; err = symbol__annotate(ms, evsel, parch); if (err) - goto out_free_offsets; + return err; symbol__calc_percent(sym, evsel); - annotation__set_offsets(notes, size); + annotation__set_index(notes); annotation__mark_jump_targets(notes, sym); err = annotation__compute_ipc(notes, size); if (err) - goto out_free_offsets; + return err; annotation__init_column_widths(notes, sym); notes->nr_events = nr_pcnt; @@ -1862,10 +1849,6 @@ int symbol__annotate2(struct map_symbol *ms, struct evsel *evsel, sym->annotate2 = 1; return 0; - -out_free_offsets: - zfree(¬es->src->offsets); - return err; } static int annotation__config(const char *var, const char *value, void *data) diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h index aa3298c20300..d61184499bda 100644 --- a/tools/perf/util/annotate.h +++ b/tools/perf/util/annotate.h @@ -246,7 +246,6 @@ struct cyc_hist { * we have more than a group in a evlist, where we will want * to see each group separately, that is why symbol__annotate2() * sets src->nr_histograms to evsel->nr_members. - * @offsets: Array of annotation_line to be accessed by offset. * @samples: Hash map of sym_hist_entry. Keyed by event index and offset in symbol. * @nr_entries: Number of annotated_line in the source list. * @nr_asm_entries: Number of annotated_line with actual asm instruction in the @@ -262,7 +261,6 @@ struct cyc_hist { struct annotated_source { struct list_head source; struct sym_hist *histograms; - struct annotation_line **offsets; struct hashmap *samples; int nr_histograms; int nr_entries; -- cgit v1.2.3 From a46acc45673bc5537b0d9fc77b7a4edb5d068de6 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 4 Apr 2024 10:57:13 -0700 Subject: perf annotate: Move 'widths' struct to 'struct annotated_source' It's only used in 'perf annotate' output which means functions with actual samples. No need to consume memory for every symbol ('struct annotation'). Also move the 'max_line_len' field into it as it's related. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240404175716.1225482-7-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browsers/annotate.c | 6 +++--- tools/perf/util/annotate.c | 41 +++++++++++++++++++++------------------ tools/perf/util/annotate.h | 20 +++++++++---------- 3 files changed, 35 insertions(+), 32 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/ui/browsers/annotate.c b/tools/perf/ui/browsers/annotate.c index c93da2ce727f..032642a0b4b6 100644 --- a/tools/perf/ui/browsers/annotate.c +++ b/tools/perf/ui/browsers/annotate.c @@ -205,13 +205,13 @@ static void annotate_browser__draw_current_jump(struct ui_browser *browser) ui_browser__set_color(browser, HE_COLORSET_JUMP_ARROWS); __ui_browser__line_arrow(browser, - pcnt_width + 2 + notes->widths.addr + width, + pcnt_width + 2 + notes->src->widths.addr + width, from, to); diff = is_fused(ab, cursor); if (diff > 0) { ui_browser__mark_fused(browser, - pcnt_width + 3 + notes->widths.addr + width, + pcnt_width + 3 + notes->src->widths.addr + width, from - diff, diff, to > from); } } @@ -983,7 +983,7 @@ int symbol__tui_annotate(struct map_symbol *ms, struct evsel *evsel, ui_helpline__push("Press ESC to exit"); - browser.b.width = notes->src->max_line_len; + browser.b.width = notes->src->widths.max_line_len; browser.b.nr_entries = notes->src->nr_entries; browser.b.entries = ¬es->src->source, browser.b.width += 18; /* Percentage */ diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 0e8319835986..0be744bb529c 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -1383,15 +1383,15 @@ static void annotation__set_index(struct annotation *notes) struct annotation_line *al; struct annotated_source *src = notes->src; - src->max_line_len = 0; + src->widths.max_line_len = 0; src->nr_entries = 0; src->nr_asm_entries = 0; list_for_each_entry(al, &src->source, node) { size_t line_len = strlen(al->line); - if (src->max_line_len < line_len) - src->max_line_len = line_len; + if (src->widths.max_line_len < line_len) + src->widths.max_line_len = line_len; al->idx = src->nr_entries++; if (al->offset != -1) al->idx_asm = src->nr_asm_entries++; @@ -1429,26 +1429,26 @@ static int annotation__max_ins_name(struct annotation *notes) static void annotation__init_column_widths(struct annotation *notes, struct symbol *sym) { - notes->widths.addr = notes->widths.target = - notes->widths.min_addr = hex_width(symbol__size(sym)); - notes->widths.max_addr = hex_width(sym->end); - notes->widths.jumps = width_jumps(notes->max_jump_sources); - notes->widths.max_ins_name = annotation__max_ins_name(notes); + notes->src->widths.addr = notes->src->widths.target = + notes->src->widths.min_addr = hex_width(symbol__size(sym)); + notes->src->widths.max_addr = hex_width(sym->end); + notes->src->widths.jumps = width_jumps(notes->max_jump_sources); + notes->src->widths.max_ins_name = annotation__max_ins_name(notes); } void annotation__update_column_widths(struct annotation *notes) { if (annotate_opts.use_offset) - notes->widths.target = notes->widths.min_addr; + notes->src->widths.target = notes->src->widths.min_addr; else if (annotate_opts.full_addr) - notes->widths.target = BITS_PER_LONG / 4; + notes->src->widths.target = BITS_PER_LONG / 4; else - notes->widths.target = notes->widths.max_addr; + notes->src->widths.target = notes->src->widths.max_addr; - notes->widths.addr = notes->widths.target; + notes->src->widths.addr = notes->src->widths.target; if (annotate_opts.show_nr_jumps) - notes->widths.addr += notes->widths.jumps + 1; + notes->src->widths.addr += notes->src->widths.jumps + 1; } void annotation__toggle_full_addr(struct annotation *notes, struct map_symbol *ms) @@ -1625,7 +1625,8 @@ call_like: obj__printf(obj, " "); } - disasm_line__scnprintf(dl, bf, size, !annotate_opts.use_offset, notes->widths.max_ins_name); + disasm_line__scnprintf(dl, bf, size, !annotate_opts.use_offset, + notes->src->widths.max_ins_name); } static void ipc_coverage_string(char *bf, int size, struct annotation *notes) @@ -1753,9 +1754,11 @@ static void __annotation_line__write(struct annotation_line *al, struct annotati obj__printf(obj, "%-*s", width - pcnt_width - cycles_width, " "); else if (al->offset == -1) { if (al->line_nr && annotate_opts.show_linenr) - printed = scnprintf(bf, sizeof(bf), "%-*d ", notes->widths.addr + 1, al->line_nr); + printed = scnprintf(bf, sizeof(bf), "%-*d ", + notes->src->widths.addr + 1, al->line_nr); else - printed = scnprintf(bf, sizeof(bf), "%-*s ", notes->widths.addr, " "); + printed = scnprintf(bf, sizeof(bf), "%-*s ", + notes->src->widths.addr, " "); obj__printf(obj, bf); obj__printf(obj, "%-*s", width - printed - pcnt_width - cycles_width + 1, al->line); } else { @@ -1773,7 +1776,7 @@ static void __annotation_line__write(struct annotation_line *al, struct annotati if (annotate_opts.show_nr_jumps) { int prev; printed = scnprintf(bf, sizeof(bf), "%*d ", - notes->widths.jumps, + notes->src->widths.jumps, al->jump_sources); prev = obj__set_jumps_percent_color(obj, al->jump_sources, current_entry); @@ -1782,7 +1785,7 @@ static void __annotation_line__write(struct annotation_line *al, struct annotati } print_addr: printed = scnprintf(bf, sizeof(bf), "%*" PRIx64 ": ", - notes->widths.target, addr); + notes->src->widths.target, addr); } else if (ins__is_call(&disasm_line(al)->ins) && annotate_opts.offset_level >= ANNOTATION__OFFSET_CALL) { goto print_addr; @@ -1790,7 +1793,7 @@ print_addr: goto print_addr; } else { printed = scnprintf(bf, sizeof(bf), "%-*s ", - notes->widths.addr, " "); + notes->src->widths.addr, " "); } } diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h index d61184499bda..402ae774426b 100644 --- a/tools/perf/util/annotate.h +++ b/tools/perf/util/annotate.h @@ -250,7 +250,7 @@ struct cyc_hist { * @nr_entries: Number of annotated_line in the source list. * @nr_asm_entries: Number of annotated_line with actual asm instruction in the * source list. - * @max_line_len: Maximum length of objdump output in an annotated_line. + * @widths: Precalculated width of each column in the TUI output. * * disasm_lines are allocated, percentages calculated and all sorted by percentage * when the annotation is about to be presented, so the percentages are for @@ -265,7 +265,15 @@ struct annotated_source { int nr_histograms; int nr_entries; int nr_asm_entries; - u16 max_line_len; + struct { + u8 addr; + u8 jumps; + u8 target; + u8 min_addr; + u8 max_addr; + u8 max_ins_name; + u16 max_line_len; + } widths; }; struct annotation_line *annotated_source__get_line(struct annotated_source *src, @@ -302,14 +310,6 @@ struct LOCKABLE annotation { u64 start; int nr_events; int max_jump_sources; - struct { - u8 addr; - u8 jumps; - u8 target; - u8 min_addr; - u8 max_addr; - u8 max_ins_name; - } widths; struct annotated_source *src; struct annotated_branch *branch; }; -- cgit v1.2.3 From f6b18ababa5ea8d7f798aa452eae83058fd09c59 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 4 Apr 2024 10:57:14 -0700 Subject: perf annotate: Move 'max_jump_sources' struct to 'struct annotated_source' It's only used in 'perf annotate' output which means functions with actual samples. No need to consume memory for every symbol ('struct annotation'). Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240404175716.1225482-8-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browsers/annotate.c | 2 +- tools/perf/util/annotate.c | 6 +++--- tools/perf/util/annotate.h | 4 +++- 3 files changed, 7 insertions(+), 5 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/ui/browsers/annotate.c b/tools/perf/ui/browsers/annotate.c index 032642a0b4b6..0e16c268e329 100644 --- a/tools/perf/ui/browsers/annotate.c +++ b/tools/perf/ui/browsers/annotate.c @@ -49,7 +49,7 @@ static int ui_browser__jumps_percent_color(struct ui_browser *browser, int nr, b if (current && (!browser->use_navkeypressed || browser->navkeypressed)) return HE_COLORSET_SELECTED; - if (nr == notes->max_jump_sources) + if (nr == notes->src->max_jump_sources) return HE_COLORSET_TOP; if (nr > 1) return HE_COLORSET_MEDIUM; diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 0be744bb529c..1fd51856d78f 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -1373,8 +1373,8 @@ annotation__mark_jump_targets(struct annotation *notes, struct symbol *sym) if (target == NULL) continue; - if (++target->jump_sources > notes->max_jump_sources) - notes->max_jump_sources = target->jump_sources; + if (++target->jump_sources > notes->src->max_jump_sources) + notes->src->max_jump_sources = target->jump_sources; } } @@ -1432,7 +1432,7 @@ annotation__init_column_widths(struct annotation *notes, struct symbol *sym) notes->src->widths.addr = notes->src->widths.target = notes->src->widths.min_addr = hex_width(symbol__size(sym)); notes->src->widths.max_addr = hex_width(sym->end); - notes->src->widths.jumps = width_jumps(notes->max_jump_sources); + notes->src->widths.jumps = width_jumps(notes->src->max_jump_sources); notes->src->widths.max_ins_name = annotation__max_ins_name(notes); } diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h index 402ae774426b..382705311d28 100644 --- a/tools/perf/util/annotate.h +++ b/tools/perf/util/annotate.h @@ -250,6 +250,8 @@ struct cyc_hist { * @nr_entries: Number of annotated_line in the source list. * @nr_asm_entries: Number of annotated_line with actual asm instruction in the * source list. + * @max_jump_sources: Maximum number of jump instructions targeting to the same + * instruction. * @widths: Precalculated width of each column in the TUI output. * * disasm_lines are allocated, percentages calculated and all sorted by percentage @@ -265,6 +267,7 @@ struct annotated_source { int nr_histograms; int nr_entries; int nr_asm_entries; + int max_jump_sources; struct { u8 addr; u8 jumps; @@ -309,7 +312,6 @@ struct annotated_branch { struct LOCKABLE annotation { u64 start; int nr_events; - int max_jump_sources; struct annotated_source *src; struct annotated_branch *branch; }; -- cgit v1.2.3 From 6f94a72d45295741b8be64da40a86780c7681a3b Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 4 Apr 2024 10:57:15 -0700 Subject: perf annotate: Move nr_events struct to 'struct annotated_source' It's only used in 'perf annotate' output which means functions with actual samples. No need to consume memory for every symbol ('struct annotation'). Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240404175716.1225482-9-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 6 +++--- tools/perf/util/annotate.h | 5 +++-- 2 files changed, 6 insertions(+), 5 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 1fd51856d78f..5f79ae0bccfd 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -1584,7 +1584,7 @@ static double annotation_line__max_percent(struct annotation_line *al, double percent_max = 0.0; int i; - for (i = 0; i < notes->nr_events; i++) { + for (i = 0; i < notes->src->nr_events; i++) { double percent; percent = annotation_data__percent(&al->data[i], @@ -1674,7 +1674,7 @@ static void __annotation_line__write(struct annotation_line *al, struct annotati if (al->offset != -1 && percent_max != 0.0) { int i; - for (i = 0; i < notes->nr_events; i++) { + for (i = 0; i < notes->src->nr_events; i++) { double percent; percent = annotation_data__percent(&al->data[i], percent_type); @@ -1846,7 +1846,7 @@ int symbol__annotate2(struct map_symbol *ms, struct evsel *evsel, return err; annotation__init_column_widths(notes, sym); - notes->nr_events = nr_pcnt; + notes->src->nr_events = nr_pcnt; annotation__update_column_widths(notes); sym->annotate2 = 1; diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h index 382705311d28..d22b9e9a2fad 100644 --- a/tools/perf/util/annotate.h +++ b/tools/perf/util/annotate.h @@ -247,6 +247,7 @@ struct cyc_hist { * to see each group separately, that is why symbol__annotate2() * sets src->nr_histograms to evsel->nr_members. * @samples: Hash map of sym_hist_entry. Keyed by event index and offset in symbol. + * @nr_events: Number of events in the current output. * @nr_entries: Number of annotated_line in the source list. * @nr_asm_entries: Number of annotated_line with actual asm instruction in the * source list. @@ -265,6 +266,7 @@ struct annotated_source { struct sym_hist *histograms; struct hashmap *samples; int nr_histograms; + int nr_events; int nr_entries; int nr_asm_entries; int max_jump_sources; @@ -311,7 +313,6 @@ struct annotated_branch { struct LOCKABLE annotation { u64 start; - int nr_events; struct annotated_source *src; struct annotated_branch *branch; }; @@ -335,7 +336,7 @@ static inline int annotation__cycles_width(struct annotation *notes) static inline int annotation__pcnt_width(struct annotation *notes) { - return (symbol_conf.show_total_period ? 12 : 7) * notes->nr_events; + return (symbol_conf.show_total_period ? 12 : 7) * notes->src->nr_events; } static inline bool annotation_line__filter(struct annotation_line *al) -- cgit v1.2.3 From 8c004c7a608a278548e4a7f25df6bae0b0a04370 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 4 Apr 2024 10:57:16 -0700 Subject: perf annotate: Move 'start' field struct to 'struct annotated_source' It's only used in 'perf annotate' output which means functions with actual samples. No need to consume memory for every symbol ('struct annotation'). Signed-off-by: Namhyung Kim Link: https://lore.kernel.org/r/20240404175716.1225482-10-namhyung@kernel.org Cc: Ian Rogers Cc: Peter Zijlstra Cc: Adrian Hunter Cc: Arnaldo Carvalho de Melo Cc: Jiri Olsa Cc: Ingo Molnar Cc: Kan Liang Cc: LKML Cc: Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 10 +++++----- tools/perf/util/annotate.h | 2 +- 2 files changed, 6 insertions(+), 6 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 5f79ae0bccfd..4db49611c386 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -909,9 +909,9 @@ int symbol__annotate(struct map_symbol *ms, struct evsel *evsel, args.arch = arch; args.ms = *ms; if (annotate_opts.full_addr) - notes->start = map__objdump_2mem(ms->map, ms->sym->start); + notes->src->start = map__objdump_2mem(ms->map, ms->sym->start); else - notes->start = map__rip_2objdump(ms->map, ms->sym->start); + notes->src->start = map__rip_2objdump(ms->map, ms->sym->start); return symbol__disassemble(sym, &args); } @@ -1456,9 +1456,9 @@ void annotation__toggle_full_addr(struct annotation *notes, struct map_symbol *m annotate_opts.full_addr = !annotate_opts.full_addr; if (annotate_opts.full_addr) - notes->start = map__objdump_2mem(ms->map, ms->sym->start); + notes->src->start = map__objdump_2mem(ms->map, ms->sym->start); else - notes->start = map__rip_2objdump(ms->map, ms->sym->start); + notes->src->start = map__rip_2objdump(ms->map, ms->sym->start); annotation__update_column_widths(notes); } @@ -1766,7 +1766,7 @@ static void __annotation_line__write(struct annotation_line *al, struct annotati int color = -1; if (!annotate_opts.use_offset) - addr += notes->start; + addr += notes->src->start; if (!annotate_opts.use_offset) { printed = scnprintf(bf, sizeof(bf), "%" PRIx64 ": ", addr); diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h index d22b9e9a2fad..d5c821c22f79 100644 --- a/tools/perf/util/annotate.h +++ b/tools/perf/util/annotate.h @@ -270,6 +270,7 @@ struct annotated_source { int nr_entries; int nr_asm_entries; int max_jump_sources; + u64 start; struct { u8 addr; u8 jumps; @@ -312,7 +313,6 @@ struct annotated_branch { }; struct LOCKABLE annotation { - u64 start; struct annotated_source *src; struct annotated_branch *branch; }; -- cgit v1.2.3 From 657852135d39b75b1b5139839b7388c1d47f3ecc Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 5 Apr 2024 14:17:58 -0700 Subject: perf annotate-data: Fix global variable lookup The recent change in the global variable handling added a bug to miss setting the return value even if it found a data type. Also add the type name in the debug message. Fixes: 1ebb5e17ef21b492 ("perf annotate-data: Add get_global_var_type()") Reviewed-by: Ian Rogers Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240405211800.1412920-3-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 043d80791bd0..1047ea9d578c 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -1468,8 +1468,10 @@ static int find_data_type_die(struct data_loc_info *dloc, Dwarf_Die *type_die) &offset, type_die)) { dloc->type_offset = offset; - pr_debug_dtp("found PC-rel by addr=%#"PRIx64" offset=%#x\n", + pr_debug_dtp("found PC-rel by addr=%#"PRIx64" offset=%#x", dloc->var_addr, offset); + pr_debug_type_name(type_die, TSR_KIND_TYPE); + ret = 0; goto out; } } -- cgit v1.2.3 From 879ebf3c830dba437781d034c536af53b0bff0c0 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 5 Apr 2024 14:17:59 -0700 Subject: perf annotate-data: Do not delete non-asm lines For data type profiling, it removed non-instruction lines from the list of annotation lines. It was to simplify the implementation dealing with instructions like to calculate the PC-relative address and to search the shortest path to the target instruction or basic block. But it means that it removes all the comments and debug information in the annotate output like source file name and line numbers. To support both code annotation and data type annotation, it'd be better to keep the non-instruction lines as well. So this change is to skip those lines during the data type profiling and to display them in the normal perf annotate output. No function changes intended (other than having more lines). Reviewed-by: Ian Rogers Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240405211800.1412920-4-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 6 +++ tools/perf/util/annotate.c | 93 ++++++++++++++++++++++++++++++----------- 2 files changed, 74 insertions(+), 25 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 1047ea9d578c..b69a1cd1577a 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -1314,6 +1314,8 @@ static int find_data_type_insn(struct data_loc_info *dloc, int reg, list_for_each_entry(bb, basic_blocks, list) { struct disasm_line *dl = bb->begin; + BUG_ON(bb->begin->al.offset == -1 || bb->end->al.offset == -1); + pr_debug_dtp("bb: [%"PRIx64" - %"PRIx64"]\n", bb->begin->al.offset, bb->end->al.offset); @@ -1321,6 +1323,10 @@ static int find_data_type_insn(struct data_loc_info *dloc, int reg, u64 this_ip = sym->start + dl->al.offset; u64 addr = map__rip_2objdump(dloc->ms->map, this_ip); + /* Skip comment or debug info lines */ + if (dl->al.offset == -1) + continue; + /* Update variable type at this address */ update_var_state(&state, dloc, addr, dl->al.offset, var_types); diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 4db49611c386..5d9b79559c73 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -2159,23 +2159,10 @@ int annotate_get_insn_location(struct arch *arch, struct disasm_line *dl, static void symbol__ensure_annotate(struct map_symbol *ms, struct evsel *evsel) { - struct disasm_line *dl, *tmp_dl; - struct annotation *notes; - - notes = symbol__annotation(ms->sym); - if (!list_empty(¬es->src->source)) - return; - - if (symbol__annotate(ms, evsel, NULL) < 0) - return; + struct annotation *notes = symbol__annotation(ms->sym); - /* remove non-insn disasm lines for simplicity */ - list_for_each_entry_safe(dl, tmp_dl, ¬es->src->source, al.node) { - if (dl->al.offset == -1) { - list_del(&dl->al.node); - free(dl); - } - } + if (list_empty(¬es->src->source)) + symbol__annotate(ms, evsel, NULL); } static struct disasm_line *find_disasm_line(struct symbol *sym, u64 ip, @@ -2187,6 +2174,9 @@ static struct disasm_line *find_disasm_line(struct symbol *sym, u64 ip, notes = symbol__annotation(sym); list_for_each_entry(dl, ¬es->src->source, al.node) { + if (dl->al.offset == -1) + continue; + if (sym->start + dl->al.offset == ip) { /* * llvm-objdump places "lock" in a separate line and @@ -2251,6 +2241,46 @@ static bool is_stack_canary(struct arch *arch, struct annotated_op_loc *loc) return false; } +static struct disasm_line * +annotation__prev_asm_line(struct annotation *notes, struct disasm_line *curr) +{ + struct list_head *sources = ¬es->src->source; + struct disasm_line *prev; + + if (curr == list_first_entry(sources, struct disasm_line, al.node)) + return NULL; + + prev = list_prev_entry(curr, al.node); + while (prev->al.offset == -1 && + prev != list_first_entry(sources, struct disasm_line, al.node)) + prev = list_prev_entry(prev, al.node); + + if (prev->al.offset == -1) + return NULL; + + return prev; +} + +static struct disasm_line * +annotation__next_asm_line(struct annotation *notes, struct disasm_line *curr) +{ + struct list_head *sources = ¬es->src->source; + struct disasm_line *next; + + if (curr == list_last_entry(sources, struct disasm_line, al.node)) + return NULL; + + next = list_next_entry(curr, al.node); + while (next->al.offset == -1 && + next != list_last_entry(sources, struct disasm_line, al.node)) + next = list_next_entry(next, al.node); + + if (next->al.offset == -1) + return NULL; + + return next; +} + u64 annotate_calc_pcrel(struct map_symbol *ms, u64 ip, int offset, struct disasm_line *dl) { @@ -2266,12 +2296,12 @@ u64 annotate_calc_pcrel(struct map_symbol *ms, u64 ip, int offset, * disasm_line. If it's the last one, we can use symbol's end * address directly. */ - if (&dl->al.node == notes->src->source.prev) + next = annotation__next_asm_line(notes, dl); + if (next == NULL) addr = ms->sym->end + offset; - else { - next = list_next_entry(dl, al.node); + else addr = ip + (next->al.offset - dl->al.offset) + offset; - } + return map__rip_2objdump(ms->map, addr); } @@ -2403,10 +2433,13 @@ retry: * from the previous instruction. */ if (dl->al.offset > 0) { + struct annotation *notes; struct disasm_line *prev_dl; - prev_dl = list_prev_entry(dl, al.node); - if (ins__is_fused(arch, prev_dl->ins.name, dl->ins.name)) { + notes = symbol__annotation(ms->sym); + prev_dl = annotation__prev_asm_line(notes, dl); + + if (prev_dl && ins__is_fused(arch, prev_dl->ins.name, dl->ins.name)) { dl = prev_dl; goto retry; } @@ -2511,8 +2544,16 @@ static bool process_basic_block(struct basic_block_data *bb_data, last_dl = list_last_entry(¬es->src->source, struct disasm_line, al.node); + if (last_dl->al.offset == -1) + last_dl = annotation__prev_asm_line(notes, last_dl); + + if (last_dl == NULL) + return false; list_for_each_entry_from(dl, ¬es->src->source, al.node) { + /* Skip comment or debug info line */ + if (dl->al.offset == -1) + continue; /* Found the target instruction */ if (sym->start + dl->al.offset == target) { found = true; @@ -2533,7 +2574,8 @@ static bool process_basic_block(struct basic_block_data *bb_data, /* jump instruction creates new basic block(s) */ next_dl = find_disasm_line(sym, sym->start + dl->ops.target.offset, /*allow_update=*/false); - add_basic_block(bb_data, link, next_dl); + if (next_dl) + add_basic_block(bb_data, link, next_dl); /* * FIXME: determine conditional jumps properly. @@ -2541,8 +2583,9 @@ static bool process_basic_block(struct basic_block_data *bb_data, * next disasm line. */ if (!strstr(dl->ins.name, "jmp")) { - next_dl = list_next_entry(dl, al.node); - add_basic_block(bb_data, link, next_dl); + next_dl = annotation__next_asm_line(notes, dl); + if (next_dl) + add_basic_block(bb_data, link, next_dl); } break; -- cgit v1.2.3 From 0235abd89feaf29fe67d3abbe2c18b7119503537 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 5 Apr 2024 14:18:00 -0700 Subject: perf annotate: Get rid of symbol__ensure_annotate() Now symbol__annotate() is reentrant and it doesn't need to remove non-instruction lines. Let's get rid of symbol__ensure_annotate() and call symbol__annotate() directly. Also we can use it to get the arch pointer instead of calling evsel__get_arch() directly. Reviewed-by: Ian Rogers Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240405211800.1412920-5-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 14 ++------------ 1 file changed, 2 insertions(+), 12 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 5d9b79559c73..11da27801d88 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -2157,14 +2157,6 @@ int annotate_get_insn_location(struct arch *arch, struct disasm_line *dl, return 0; } -static void symbol__ensure_annotate(struct map_symbol *ms, struct evsel *evsel) -{ - struct annotation *notes = symbol__annotation(ms->sym); - - if (list_empty(¬es->src->source)) - symbol__annotate(ms, evsel, NULL); -} - static struct disasm_line *find_disasm_line(struct symbol *sym, u64 ip, bool allow_update) { @@ -2339,14 +2331,12 @@ struct annotated_data_type *hist_entry__get_data_type(struct hist_entry *he) return NULL; } - if (evsel__get_arch(evsel, &arch) < 0) { + /* Make sure it has the disasm of the function */ + if (symbol__annotate(ms, evsel, &arch) < 0) { ann_data_stat.no_insn++; return NULL; } - /* Make sure it runs objdump to get disasm of the function */ - symbol__ensure_annotate(ms, evsel); - /* * Get a disasm to extract the location from the insn. * This is too slow... -- cgit v1.2.3 From 4b5ee6db2d3cd6249919ae554552856de0e29791 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Wed, 3 Apr 2024 09:46:36 -0700 Subject: perf metrics: Remove the "No_group" metric group Rather than place metrics without a metric group in "No_group" place them in a a metric group that is their name. Still allow such metrics to be selected if "No_group" is passed, this change just impacts perf list. Reviewed-by: Kan Liang Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240403164636.3429091-1-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/metricgroup.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/metricgroup.c b/tools/perf/util/metricgroup.c index 79ef6095ab28..6ec083af14a1 100644 --- a/tools/perf/util/metricgroup.c +++ b/tools/perf/util/metricgroup.c @@ -455,7 +455,7 @@ static int metricgroup__add_to_mep_groups(const struct pmu_metric *pm, const char *g; char *omg, *mg; - mg = strdup(pm->metric_group ?: "No_group"); + mg = strdup(pm->metric_group ?: pm->metric_name); if (!mg) return -ENOMEM; omg = mg; @@ -466,7 +466,7 @@ static int metricgroup__add_to_mep_groups(const struct pmu_metric *pm, if (strlen(g)) me = mep_lookup(groups, g, pm->metric_name); else - me = mep_lookup(groups, "No_group", pm->metric_name); + me = mep_lookup(groups, pm->metric_name, pm->metric_name); if (me) { me->metric_desc = pm->desc; -- cgit v1.2.3 From df12e21d4e15e48a5e7d12e58f1a00742c4177d0 Mon Sep 17 00:00:00 2001 From: James Clark Date: Wed, 10 Apr 2024 11:34:54 +0100 Subject: perf map: Remove kernel map before updating start and end addresses In a debug build there is validation that mmap lists are sorted when taking a lock. In machine__update_kernel_mmap() the start and end addresses are updated resulting in an unsorted list before the map is removed from the list. When the map is removed, the lock is taken which triggers the validation and the failure: $ perf test "object code reading" --- start --- perf: util/maps.c:88: check_invariants: Assertion `map__start(prev) <= map__start(map)' failed. Aborted Fix it by updating the addresses after removal, but before insertion. The bug depends on the ordering and type of debug info on the system and doesn't reproduce everywhere. Fixes: 659ad3492b913c90 ("perf maps: Switch from rbtree to lazily sorted array for addresses") Reviewed-by: Ian Rogers Signed-off-by: James Clark Acked-by: Namhyung Kim Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Athira Rajeev Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Leo Yan Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Spoorthy S Link: https://lore.kernel.org/r/20240410103458.813656-4-james.clark@arm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index 5eb9044bc223..a26c8bea58d0 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -1549,8 +1549,8 @@ static int machine__update_kernel_mmap(struct machine *machine, updated = map__get(orig); machine->vmlinux_map = updated; - machine__set_kernel_mmap(machine, start, end); maps__remove(machine__kernel_maps(machine), orig); + machine__set_kernel_mmap(machine, start, end); err = maps__insert(machine__kernel_maps(machine), updated); map__put(orig); -- cgit v1.2.3 From eb833488631b171e693a6917eb17b41fdedda659 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 10 Apr 2024 20:32:50 -0700 Subject: perf annotate-data: Skip sample histogram for stack canary It's a pseudo data type and has no field. Fixes: b3c95109c131fcc9 ("perf annotate-data: Add stack canary type") Closes: https://lore.kernel.org/lkml/Zhb6jJneP36Z-or0@x1 Reported-by: Arnaldo Carvalho de Melo Signed-off-by: Namhyung Kim Tested-by: Arnaldo Carvalho de Melo Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240411033256.2099646-2-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 11da27801d88..ec79c120a7d2 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -2399,8 +2399,9 @@ retry: mem_type = find_data_type(&dloc); if (mem_type == NULL && is_stack_canary(arch, op_loc)) { - mem_type = &canary_type; - dloc.type_offset = 0; + istat->good++; + he->mem_type_off = 0; + return &canary_type; } if (mem_type) -- cgit v1.2.3 From 9b561be15febda6f9b314c9eab51e157f8f34dea Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 10 Apr 2024 20:32:52 -0700 Subject: perf annotate-data: Add hist_entry__annotate_data_tty() And move the related code into util/annotate-data.c file. Reviewed-by: Ian Rogers Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240411033256.2099646-4-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-annotate.c | 106 +------------------------------------ tools/perf/util/annotate-data.c | 112 ++++++++++++++++++++++++++++++++++++++++ tools/perf/util/annotate-data.h | 9 ++++ 3 files changed, 122 insertions(+), 105 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 332e1ddcacbd..0812664faa54 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -329,108 +329,6 @@ static int hist_entry__tty_annotate(struct hist_entry *he, return symbol__tty_annotate2(&he->ms, evsel); } -static void print_annotated_data_header(struct hist_entry *he, struct evsel *evsel) -{ - struct dso *dso = map__dso(he->ms.map); - int nr_members = 1; - int nr_samples = he->stat.nr_events; - int width = 7; - const char *val_hdr = "Percent"; - - if (evsel__is_group_event(evsel)) { - struct hist_entry *pair; - - list_for_each_entry(pair, &he->pairs.head, pairs.node) - nr_samples += pair->stat.nr_events; - } - - printf("Annotate type: '%s' in %s (%d samples):\n", - he->mem_type->self.type_name, dso->name, nr_samples); - - if (evsel__is_group_event(evsel)) { - struct evsel *pos; - int i = 0; - - for_each_group_evsel(pos, evsel) - printf(" event[%d] = %s\n", i++, pos->name); - - nr_members = evsel->core.nr_members; - } - - if (symbol_conf.show_total_period) { - width = 11; - val_hdr = "Period"; - } else if (symbol_conf.show_nr_samples) { - width = 7; - val_hdr = "Samples"; - } - - printf("============================================================================\n"); - printf("%*s %10s %10s %s\n", (width + 1) * nr_members, val_hdr, - "offset", "size", "field"); -} - -static void print_annotated_data_value(struct type_hist *h, u64 period, int nr_samples) -{ - double percent = h->period ? (100.0 * period / h->period) : 0; - const char *color = get_percent_color(percent); - - if (symbol_conf.show_total_period) - color_fprintf(stdout, color, " %11" PRIu64, period); - else if (symbol_conf.show_nr_samples) - color_fprintf(stdout, color, " %7d", nr_samples); - else - color_fprintf(stdout, color, " %7.2f", percent); -} - -static void print_annotated_data_type(struct annotated_data_type *mem_type, - struct annotated_member *member, - struct evsel *evsel, int indent) -{ - struct annotated_member *child; - struct type_hist *h = mem_type->histograms[evsel->core.idx]; - int i, nr_events = 1, samples = 0; - u64 period = 0; - int width = symbol_conf.show_total_period ? 11 : 7; - - for (i = 0; i < member->size; i++) { - samples += h->addr[member->offset + i].nr_samples; - period += h->addr[member->offset + i].period; - } - print_annotated_data_value(h, period, samples); - - if (evsel__is_group_event(evsel)) { - struct evsel *pos; - - for_each_group_member(pos, evsel) { - h = mem_type->histograms[pos->core.idx]; - - samples = 0; - period = 0; - for (i = 0; i < member->size; i++) { - samples += h->addr[member->offset + i].nr_samples; - period += h->addr[member->offset + i].period; - } - print_annotated_data_value(h, period, samples); - } - nr_events = evsel->core.nr_members; - } - - printf(" %10d %10d %*s%s\t%s", - member->offset, member->size, indent, "", member->type_name, - member->var_name ?: ""); - - if (!list_empty(&member->children)) - printf(" {\n"); - - list_for_each_entry(child, &member->children, node) - print_annotated_data_type(mem_type, child, evsel, indent + 4); - - if (!list_empty(&member->children)) - printf("%*s}", (width + 1) * nr_events + 24 + indent, ""); - printf(";\n"); -} - static void print_annotate_data_stat(struct annotated_data_stat *s) { #define PRINT_STAT(fld) if (s->fld) printf("%10d : %s\n", s->fld, #fld) @@ -571,9 +469,7 @@ find_next: goto find_next; } - print_annotated_data_header(he, evsel); - print_annotated_data_type(he->mem_type, &he->mem_type->self, evsel, 0); - printf("\n"); + hist_entry__annotate_data_tty(he, evsel); goto find_next; } diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index b69a1cd1577a..b150137a92dc 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -19,6 +19,7 @@ #include "evlist.h" #include "map.h" #include "map_symbol.h" +#include "sort.h" #include "strbuf.h" #include "symbol.h" #include "symbol_conf.h" @@ -1710,3 +1711,114 @@ int annotated_data_type__update_samples(struct annotated_data_type *adt, h->addr[offset].period += period; return 0; } + +static void print_annotated_data_header(struct hist_entry *he, struct evsel *evsel) +{ + struct dso *dso = map__dso(he->ms.map); + int nr_members = 1; + int nr_samples = he->stat.nr_events; + int width = 7; + const char *val_hdr = "Percent"; + + if (evsel__is_group_event(evsel)) { + struct hist_entry *pair; + + list_for_each_entry(pair, &he->pairs.head, pairs.node) + nr_samples += pair->stat.nr_events; + } + + printf("Annotate type: '%s' in %s (%d samples):\n", + he->mem_type->self.type_name, dso->name, nr_samples); + + if (evsel__is_group_event(evsel)) { + struct evsel *pos; + int i = 0; + + for_each_group_evsel(pos, evsel) + printf(" event[%d] = %s\n", i++, pos->name); + + nr_members = evsel->core.nr_members; + } + + if (symbol_conf.show_total_period) { + width = 11; + val_hdr = "Period"; + } else if (symbol_conf.show_nr_samples) { + width = 7; + val_hdr = "Samples"; + } + + printf("============================================================================\n"); + printf("%*s %10s %10s %s\n", (width + 1) * nr_members, val_hdr, + "offset", "size", "field"); +} + +static void print_annotated_data_value(struct type_hist *h, u64 period, int nr_samples) +{ + double percent = h->period ? (100.0 * period / h->period) : 0; + const char *color = get_percent_color(percent); + + if (symbol_conf.show_total_period) + color_fprintf(stdout, color, " %11" PRIu64, period); + else if (symbol_conf.show_nr_samples) + color_fprintf(stdout, color, " %7d", nr_samples); + else + color_fprintf(stdout, color, " %7.2f", percent); +} + +static void print_annotated_data_type(struct annotated_data_type *mem_type, + struct annotated_member *member, + struct evsel *evsel, int indent) +{ + struct annotated_member *child; + struct type_hist *h = mem_type->histograms[evsel->core.idx]; + int i, nr_events = 1, samples = 0; + u64 period = 0; + int width = symbol_conf.show_total_period ? 11 : 7; + + for (i = 0; i < member->size; i++) { + samples += h->addr[member->offset + i].nr_samples; + period += h->addr[member->offset + i].period; + } + print_annotated_data_value(h, period, samples); + + if (evsel__is_group_event(evsel)) { + struct evsel *pos; + + for_each_group_member(pos, evsel) { + h = mem_type->histograms[pos->core.idx]; + + samples = 0; + period = 0; + for (i = 0; i < member->size; i++) { + samples += h->addr[member->offset + i].nr_samples; + period += h->addr[member->offset + i].period; + } + print_annotated_data_value(h, period, samples); + } + nr_events = evsel->core.nr_members; + } + + printf(" %10d %10d %*s%s\t%s", + member->offset, member->size, indent, "", member->type_name, + member->var_name ?: ""); + + if (!list_empty(&member->children)) + printf(" {\n"); + + list_for_each_entry(child, &member->children, node) + print_annotated_data_type(mem_type, child, evsel, indent + 4); + + if (!list_empty(&member->children)) + printf("%*s}", (width + 1) * nr_events + 24 + indent, ""); + printf(";\n"); +} + +int hist_entry__annotate_data_tty(struct hist_entry *he, struct evsel *evsel) +{ + print_annotated_data_header(he, evsel); + print_annotated_data_type(he->mem_type, &he->mem_type->self, evsel, 0); + printf("\n"); + + return 0; +} diff --git a/tools/perf/util/annotate-data.h b/tools/perf/util/annotate-data.h index fe1e53d6e8c7..01489db267d4 100644 --- a/tools/perf/util/annotate-data.h +++ b/tools/perf/util/annotate-data.h @@ -10,6 +10,7 @@ struct annotated_op_loc; struct debuginfo; struct evsel; +struct hist_entry; struct map_symbol; struct thread; @@ -156,6 +157,8 @@ void annotated_data_type__tree_delete(struct rb_root *root); /* Release all global variable information in the tree */ void global_var_type__tree_delete(struct rb_root *root); +int hist_entry__annotate_data_tty(struct hist_entry *he, struct evsel *evsel); + #else /* HAVE_DWARF_SUPPORT */ static inline struct annotated_data_type * @@ -182,6 +185,12 @@ static inline void global_var_type__tree_delete(struct rb_root *root __maybe_unu { } +static inline int hist_entry__annotate_data_tty(struct hist_entry *he __maybe_unused, + struct evsel *evsel __maybe_unused) +{ + return -1; +} + #endif /* HAVE_DWARF_SUPPORT */ #endif /* _PERF_ANNOTATE_DATA_H */ -- cgit v1.2.3 From d001c7a7f473674353311a79cf140d054b26afcd Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 10 Apr 2024 20:32:53 -0700 Subject: perf annotate-data: Add hist_entry__annotate_data_tui() Support data type profiling output on TUI. Testing from Arnaldo: First make sure that the debug information for your workload binaries in embedded in them by building it with '-g' or install the debuginfo packages, since our workload is 'find': root@number:~# type find find is hashed (/usr/bin/find) root@number:~# rpm -qf /usr/bin/find findutils-4.9.0-5.fc39.x86_64 root@number:~# dnf debuginfo-install findutils root@number:~# Then collect some data: root@number:~# echo 1 > /proc/sys/vm/drop_caches root@number:~# perf mem record find / > /dev/null [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.331 MB perf.data (3982 samples) ] root@number:~# Finally do data-type annotation with the following command, that will default, as 'perf report' to the --tui mode, with lines colored to highlight the hotspots, etc. root@number:~# perf annotate --data-type Annotate type: 'struct predicate' (58 samples) Percent Offset Size Field 100.00 0 312 struct predicate { 0.00 0 8 PRED_FUNC pred_func; 0.00 8 8 char* p_name; 0.00 16 4 enum predicate_type p_type; 0.00 20 4 enum predicate_precedence p_prec; 0.00 24 1 _Bool side_effects; 0.00 25 1 _Bool no_default_print; 0.00 26 1 _Bool need_stat; 0.00 27 1 _Bool need_type; 0.00 28 1 _Bool need_inum; 0.00 32 4 enum EvaluationCost p_cost; 0.00 36 4 float est_success_rate; 0.00 40 1 _Bool literal_control_chars; 0.00 41 1 _Bool artificial; 0.00 48 8 char* arg_text; Reviewed-by: Ian Rogers Signed-off-by: Namhyung Kim Tested-by: Arnaldo Carvalho de Melo Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240411033256.2099646-5-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-annotate.c | 30 +++- tools/perf/ui/browsers/Build | 1 + tools/perf/ui/browsers/annotate-data.c | 282 +++++++++++++++++++++++++++++++++ tools/perf/util/annotate-data.c | 3 +- tools/perf/util/annotate-data.h | 13 ++ 5 files changed, 324 insertions(+), 5 deletions(-) create mode 100644 tools/perf/ui/browsers/annotate-data.c (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 0812664faa54..6f7104f06c42 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -469,8 +469,32 @@ find_next: goto find_next; } - hist_entry__annotate_data_tty(he, evsel); - goto find_next; + if (use_browser == 1) + key = hist_entry__annotate_data_tui(he, evsel, NULL); + else + key = hist_entry__annotate_data_tty(he, evsel); + + switch (key) { + case -1: + if (!ann->skip_missing) + return; + /* fall through */ + case K_RIGHT: + case '>': + next = rb_next(nd); + break; + case K_LEFT: + case '<': + next = rb_prev(nd); + break; + default: + return; + } + + if (next != NULL) + nd = next; + + continue; } if (use_browser == 2) { @@ -873,9 +897,7 @@ int cmd_annotate(int argc, const char **argv) use_browser = 2; #endif - /* FIXME: only support stdio for now */ if (annotate.data_type) { - use_browser = 0; annotate_opts.annotate_src = false; symbol_conf.annotate_data_member = true; symbol_conf.annotate_data_sample = true; diff --git a/tools/perf/ui/browsers/Build b/tools/perf/ui/browsers/Build index 7a1d5ddaf688..2608b5da3167 100644 --- a/tools/perf/ui/browsers/Build +++ b/tools/perf/ui/browsers/Build @@ -1,4 +1,5 @@ perf-y += annotate.o +perf-y += annotate-data.o perf-y += hists.o perf-y += map.o perf-y += scripts.o diff --git a/tools/perf/ui/browsers/annotate-data.c b/tools/perf/ui/browsers/annotate-data.c new file mode 100644 index 000000000000..fefacaaf16db --- /dev/null +++ b/tools/perf/ui/browsers/annotate-data.c @@ -0,0 +1,282 @@ +// SPDX-License-Identifier: GPL-2.0 +#include +#include +#include + +#include "ui/browser.h" +#include "ui/helpline.h" +#include "ui/keysyms.h" +#include "ui/ui.h" +#include "util/annotate.h" +#include "util/annotate-data.h" +#include "util/evsel.h" +#include "util/sort.h" + +struct annotated_data_browser { + struct ui_browser b; + struct list_head entries; +}; + +struct browser_entry { + struct list_head node; + struct annotated_member *data; + struct type_hist_entry hists; + int indent; +}; + +static void update_hist_entry(struct type_hist_entry *dst, + struct type_hist_entry *src) +{ + dst->nr_samples += src->nr_samples; + dst->period += src->period; +} + +static int get_member_overhead(struct annotated_data_type *adt, + struct browser_entry *entry, + struct evsel *evsel) +{ + struct annotated_member *member = entry->data; + int i; + + for (i = 0; i < member->size; i++) { + struct type_hist *h; + int offset = member->offset + i; + + h = adt->histograms[evsel->core.idx]; + update_hist_entry(&entry->hists, &h->addr[offset]); + } + return 0; +} + +static int add_child_entries(struct annotated_data_browser *browser, + struct annotated_data_type *adt, + struct annotated_member *member, + struct evsel *evsel, int indent) +{ + struct annotated_member *pos; + struct browser_entry *entry; + int nr_entries = 0; + + entry = zalloc(sizeof(*entry)); + if (entry == NULL) + return -1; + + entry->data = member; + entry->indent = indent; + if (get_member_overhead(adt, entry, evsel) < 0) { + free(entry); + return -1; + } + + list_add_tail(&entry->node, &browser->entries); + nr_entries++; + + list_for_each_entry(pos, &member->children, node) { + int nr = add_child_entries(browser, adt, pos, evsel, indent + 1); + + if (nr < 0) + return nr; + + nr_entries += nr; + } + + /* add an entry for the closing bracket ("}") */ + if (!list_empty(&member->children)) { + entry = zalloc(sizeof(*entry)); + if (entry == NULL) + return -1; + + entry->indent = indent; + list_add_tail(&entry->node, &browser->entries); + nr_entries++; + } + + return nr_entries; +} + +static int annotated_data_browser__collect_entries(struct annotated_data_browser *browser) +{ + struct hist_entry *he = browser->b.priv; + struct annotated_data_type *adt = he->mem_type; + struct evsel *evsel = hists_to_evsel(he->hists); + + INIT_LIST_HEAD(&browser->entries); + browser->b.entries = &browser->entries; + browser->b.nr_entries = add_child_entries(browser, adt, &adt->self, + evsel, /*indent=*/0); + return 0; +} + +static void annotated_data_browser__delete_entries(struct annotated_data_browser *browser) +{ + struct browser_entry *pos, *tmp; + + list_for_each_entry_safe(pos, tmp, &browser->entries, node) { + list_del_init(&pos->node); + free(pos); + } +} + +static unsigned int browser__refresh(struct ui_browser *uib) +{ + return ui_browser__list_head_refresh(uib); +} + +static int browser__show(struct ui_browser *uib) +{ + struct hist_entry *he = uib->priv; + struct annotated_data_type *adt = he->mem_type; + const char *help = "Press 'h' for help on key bindings"; + char title[256]; + + snprintf(title, sizeof(title), "Annotate type: '%s' (%d samples)", + adt->self.type_name, he->stat.nr_events); + + if (ui_browser__show(uib, title, help) < 0) + return -1; + + /* second line header */ + ui_browser__gotorc_title(uib, 0, 0); + ui_browser__set_color(uib, HE_COLORSET_ROOT); + + if (symbol_conf.show_total_period) + strcpy(title, "Period"); + else if (symbol_conf.show_nr_samples) + strcpy(title, "Samples"); + else + strcpy(title, "Percent"); + + ui_browser__printf(uib, " %10s %10s %10s %s", + title, "Offset", "Size", "Field"); + ui_browser__write_nstring(uib, "", uib->width); + return 0; +} + +static void browser__write_overhead(struct ui_browser *uib, + struct type_hist *total, + struct type_hist_entry *hist, int row) +{ + u64 period = hist->period; + double percent = total->period ? (100.0 * period / total->period) : 0; + bool current = ui_browser__is_current_entry(uib, row); + int nr_samples = 0; + + ui_browser__set_percent_color(uib, percent, current); + + if (symbol_conf.show_total_period) + ui_browser__printf(uib, " %10" PRIu64, period); + else if (symbol_conf.show_nr_samples) + ui_browser__printf(uib, " %10d", nr_samples); + else + ui_browser__printf(uib, " %10.2f", percent); + + ui_browser__set_percent_color(uib, 0, current); +} + +static void browser__write(struct ui_browser *uib, void *entry, int row) +{ + struct browser_entry *be = entry; + struct annotated_member *member = be->data; + struct hist_entry *he = uib->priv; + struct annotated_data_type *adt = he->mem_type; + struct evsel *evsel = hists_to_evsel(he->hists); + + if (member == NULL) { + bool current = ui_browser__is_current_entry(uib, row); + + /* print the closing bracket */ + ui_browser__set_percent_color(uib, 0, current); + ui_browser__write_nstring(uib, "", 11); + ui_browser__printf(uib, " %10s %10s %*s};", + "", "", be->indent * 4, ""); + ui_browser__write_nstring(uib, "", uib->width); + return; + } + + /* print the number */ + browser__write_overhead(uib, adt->histograms[evsel->core.idx], + &be->hists, row); + + /* print type info */ + if (be->indent == 0 && !member->var_name) { + ui_browser__printf(uib, " %10d %10d %s%s", + member->offset, member->size, + member->type_name, + list_empty(&member->children) ? ";" : " {"); + } else { + ui_browser__printf(uib, " %10d %10d %*s%s\t%s%s", + member->offset, member->size, + be->indent * 4, "", member->type_name, + member->var_name ?: "", + list_empty(&member->children) ? ";" : " {"); + } + /* fill the rest */ + ui_browser__write_nstring(uib, "", uib->width); +} + +static int annotated_data_browser__run(struct annotated_data_browser *browser, + struct evsel *evsel __maybe_unused, + struct hist_browser_timer *hbt) +{ + int delay_secs = hbt ? hbt->refresh : 0; + int key; + + if (browser__show(&browser->b) < 0) + return -1; + + while (1) { + key = ui_browser__run(&browser->b, delay_secs); + + switch (key) { + case K_TIMER: + if (hbt) + hbt->timer(hbt->arg); + continue; + case K_F1: + case 'h': + ui_browser__help_window(&browser->b, + "UP/DOWN/PGUP\n" + "PGDN/SPACE Navigate\n" + " Move to prev/next symbol\n" + "q/ESC/CTRL+C Exit\n\n"); + continue; + case K_LEFT: + case '<': + case '>': + case K_ESC: + case 'q': + case CTRL('c'): + goto out; + default: + continue; + } + } +out: + ui_browser__hide(&browser->b); + return key; +} + +int hist_entry__annotate_data_tui(struct hist_entry *he, struct evsel *evsel, + struct hist_browser_timer *hbt) +{ + struct annotated_data_browser browser = { + .b = { + .refresh = browser__refresh, + .seek = ui_browser__list_head_seek, + .write = browser__write, + .priv = he, + .extra_title_lines = 1, + }, + }; + int ret; + + ui_helpline__push("Press ESC to exit"); + + ret = annotated_data_browser__collect_entries(&browser); + if (ret == 0) + ret = annotated_data_browser__run(&browser, evsel, hbt); + + annotated_data_browser__delete_entries(&browser); + + return ret; +} diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index b150137a92dc..1cd857400038 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -1820,5 +1820,6 @@ int hist_entry__annotate_data_tty(struct hist_entry *he, struct evsel *evsel) print_annotated_data_type(he->mem_type, &he->mem_type->self, evsel, 0); printf("\n"); - return 0; + /* move to the next entry */ + return '>'; } diff --git a/tools/perf/util/annotate-data.h b/tools/perf/util/annotate-data.h index 01489db267d4..0a57d9f5ee78 100644 --- a/tools/perf/util/annotate-data.h +++ b/tools/perf/util/annotate-data.h @@ -10,6 +10,7 @@ struct annotated_op_loc; struct debuginfo; struct evsel; +struct hist_browser_timer; struct hist_entry; struct map_symbol; struct thread; @@ -193,4 +194,16 @@ static inline int hist_entry__annotate_data_tty(struct hist_entry *he __maybe_un #endif /* HAVE_DWARF_SUPPORT */ +#ifdef HAVE_SLANG_SUPPORT +int hist_entry__annotate_data_tui(struct hist_entry *he, struct evsel *evsel, + struct hist_browser_timer *hbt); +#else +static inline int hist_entry__annotate_data_tui(struct hist_entry *he __maybe_unused, + struct evsel *evsel __maybe_unused, + struct hist_browser_timer *hbt __maybe_unused) +{ + return -1; +} +#endif /* HAVE_SLANG_SUPPORT */ + #endif /* _PERF_ANNOTATE_DATA_H */ -- cgit v1.2.3 From 6cdd977ec24e1538b35a08bde823a74b69e829f2 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 10 Apr 2024 20:32:56 -0700 Subject: perf report: Do not collect sample histogram unnecessarily The data type profiling alone doesn't need the sample histogram for functions. It only needs the histogram for the types. Let's remove the condition in the report_callback to check if data type profiling is selected and make sure the annotation has the 'struct annotated_source' instantiated before calling symbol__disassemble(). Reviewed-by: Ian Rogers Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240411033256.2099646-8-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-report.c | 2 +- tools/perf/util/annotate.c | 7 +++++++ 2 files changed, 8 insertions(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index aaa6427a1224..dafba6e030ef 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -172,7 +172,7 @@ static int hist_iter__report_callback(struct hist_entry_iter *iter, struct mem_info *mi; struct branch_info *bi; - if (!ui__has_annotation() && !rep->symbol_ipc && !rep->data_type) + if (!ui__has_annotation() && !rep->symbol_ipc) return 0; if (sort__mode == SORT_MODE__BRANCH) { diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index ec79c120a7d2..7595c8fbc2c5 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -908,6 +908,13 @@ int symbol__annotate(struct map_symbol *ms, struct evsel *evsel, args.arch = arch; args.ms = *ms; + + if (notes->src == NULL) { + notes->src = annotated_source__new(); + if (notes->src == NULL) + return -1; + } + if (annotate_opts.full_addr) notes->src->start = map__objdump_2mem(ms->map, ms->sym->start); else -- cgit v1.2.3 From 873a83731f1cc85c8427fdc7a9e24fb270faca44 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 10 Apr 2024 11:51:17 -0700 Subject: perf annotate: Skip DSOs not found In some data file, I see the following messages repeated. It seems it doesn't have DSOs in the system and the dso->binary_type is set to DSO_BINARY_TYPE__NOT_FOUND. Let's skip them to avoid the followings. No output from objdump --start-address=0x0000000000000000 --stop-address=0x00000000000000d4 -d --no-show-raw-insn -C "$1" Error running objdump --start-address=0x0000000000000000 --stop-address=0x0000000000000631 -d --no-show-raw-insn -C "$1" ... Closes: https://lore.kernel.org/linux-perf-users/15e1a2847b8cebab4de57fc68e033086aa6980ce.camel@yandex.ru/ Reported-by: Konstantin Kharlamov Reviewed-by: Ian Rogers Signed-off-by: Namhyung Kim Tested-by: Konstantin Kharlamov Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240410185117.1987239-1-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/disasm.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/disasm.c b/tools/perf/util/disasm.c index a1219eb930aa..92937809be85 100644 --- a/tools/perf/util/disasm.c +++ b/tools/perf/util/disasm.c @@ -1669,6 +1669,8 @@ int symbol__disassemble(struct symbol *sym, struct annotate_args *args) return symbol__disassemble_bpf(sym, args); } else if (dso->binary_type == DSO_BINARY_TYPE__BPF_IMAGE) { return symbol__disassemble_bpf_image(sym, args); + } else if (dso->binary_type == DSO_BINARY_TYPE__NOT_FOUND) { + return -1; } else if (dso__is_kcore(dso)) { kce.kcore_filename = symfs_filename; kce.addr = map__rip_2objdump(map, sym->start); -- cgit v1.2.3 From 83acca9f90c700bafd81229f2c2d5debcf6b27bc Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 9 Apr 2024 23:42:03 -0700 Subject: perf dsos: Attempt to better abstract DSOs internals Move functions from machine and build-id to dsos. Pass 'struct dsos' rather than internal state. Rename some functions to better represent which data structure they operate on. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Anne Macedo Cc: Athira Rajeev Cc: Ben Gainey Cc: Changbin Du Cc: Chengen Du Cc: Colin Ian King Cc: Ilkka Koskinen Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kan Liang Cc: Leo Yan Cc: Li Dong Cc: Mark Rutland Cc: Markus Elfring Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Song Liu Cc: Sun Haiyong Cc: Thomas Richter Cc: Yang Jihong Cc: Yanteng Si Cc: Yicong Yang Cc: zhaimingbing Link: https://lore.kernel.org/r/20240410064214.2755936-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-inject.c | 2 +- tools/perf/builtin-record.c | 2 +- tools/perf/util/build-id.c | 38 +---------------------------------- tools/perf/util/build-id.h | 2 -- tools/perf/util/dso.h | 6 ------ tools/perf/util/dsos.c | 49 ++++++++++++++++++++++++++++++++++++++++++--- tools/perf/util/dsos.h | 19 ++++++++++++++---- tools/perf/util/machine.c | 40 ++++++++---------------------------- tools/perf/util/machine.h | 2 ++ tools/perf/util/session.c | 21 +++++++++++++++++++ tools/perf/util/session.h | 2 ++ 11 files changed, 97 insertions(+), 86 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c index eb3ef5c24b66..ef73317e6ae7 100644 --- a/tools/perf/builtin-inject.c +++ b/tools/perf/builtin-inject.c @@ -2122,7 +2122,7 @@ static int __cmd_inject(struct perf_inject *inject) */ if (perf_header__has_feat(&session->header, HEADER_BUILD_ID) && inject->have_auxtrace && !inject->itrace_synth_opts.set) - dsos__hit_all(session); + perf_session__dsos_hit_all(session); /* * The AUX areas have been removed and replaced with * synthesized hardware events, so clear the feature flag. diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 6aeae398ec28..2ff718d3e202 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -1788,7 +1788,7 @@ record__finish_output(struct record *rec) process_buildids(rec); if (rec->buildid_all) - dsos__hit_all(rec->session); + perf_session__dsos_hit_all(rec->session); } perf_session__write_header(rec->session, rec->evlist, fd, true); diff --git a/tools/perf/util/build-id.c b/tools/perf/util/build-id.c index 03c64b85383b..a617b1917e6b 100644 --- a/tools/perf/util/build-id.c +++ b/tools/perf/util/build-id.c @@ -390,42 +390,6 @@ int perf_session__write_buildid_table(struct perf_session *session, return err; } -static int __dsos__hit_all(struct list_head *head) -{ - struct dso *pos; - - list_for_each_entry(pos, head, node) - pos->hit = true; - - return 0; -} - -static int machine__hit_all_dsos(struct machine *machine) -{ - return __dsos__hit_all(&machine->dsos.head); -} - -int dsos__hit_all(struct perf_session *session) -{ - struct rb_node *nd; - int err; - - err = machine__hit_all_dsos(&session->machines.host); - if (err) - return err; - - for (nd = rb_first_cached(&session->machines.guests); nd; - nd = rb_next(nd)) { - struct machine *pos = rb_entry(nd, struct machine, rb_node); - - err = machine__hit_all_dsos(pos); - if (err) - return err; - } - - return 0; -} - void disable_buildid_cache(void) { no_buildid_cache = true; @@ -992,7 +956,7 @@ int perf_session__cache_build_ids(struct perf_session *session) static bool machine__read_build_ids(struct machine *machine, bool with_hits) { - return __dsos__read_build_ids(&machine->dsos.head, with_hits); + return __dsos__read_build_ids(&machine->dsos, with_hits); } bool perf_session__read_build_ids(struct perf_session *session, bool with_hits) diff --git a/tools/perf/util/build-id.h b/tools/perf/util/build-id.h index 4e3a1169379b..3fa8bffb07ca 100644 --- a/tools/perf/util/build-id.h +++ b/tools/perf/util/build-id.h @@ -39,8 +39,6 @@ int build_id__mark_dso_hit(struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, struct machine *machine); -int dsos__hit_all(struct perf_session *session); - int perf_event__inject_buildid(struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, struct machine *machine); diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h index 17dab230a2ca..3d4faad8d5dc 100644 --- a/tools/perf/util/dso.h +++ b/tools/perf/util/dso.h @@ -233,12 +233,6 @@ struct dso { #define dso__for_each_symbol(dso, pos, n) \ symbols__for_each_entry(&(dso)->symbols, pos, n) -#define dsos__for_each_with_build_id(pos, head) \ - list_for_each_entry(pos, head, node) \ - if (!pos->has_build_id) \ - continue; \ - else - static inline void dso__set_loaded(struct dso *dso) { dso->loaded = true; diff --git a/tools/perf/util/dsos.c b/tools/perf/util/dsos.c index cf80aa42dd07..e65ef6762bed 100644 --- a/tools/perf/util/dsos.c +++ b/tools/perf/util/dsos.c @@ -12,6 +12,35 @@ #include // filename__read_build_id #include +void dsos__init(struct dsos *dsos) +{ + INIT_LIST_HEAD(&dsos->head); + dsos->root = RB_ROOT; + init_rwsem(&dsos->lock); +} + +static void dsos__purge(struct dsos *dsos) +{ + struct dso *pos, *n; + + down_write(&dsos->lock); + + list_for_each_entry_safe(pos, n, &dsos->head, node) { + RB_CLEAR_NODE(&pos->rb_node); + pos->root = NULL; + list_del_init(&pos->node); + dso__put(pos); + } + + up_write(&dsos->lock); +} + +void dsos__exit(struct dsos *dsos) +{ + dsos__purge(dsos); + exit_rwsem(&dsos->lock); +} + static int __dso_id__cmp(struct dso_id *a, struct dso_id *b) { if (a->maj > b->maj) return -1; @@ -73,8 +102,9 @@ int dso__cmp_id(struct dso *a, struct dso *b) return __dso_id__cmp(&a->id, &b->id); } -bool __dsos__read_build_ids(struct list_head *head, bool with_hits) +bool __dsos__read_build_ids(struct dsos *dsos, bool with_hits) { + struct list_head *head = &dsos->head; bool have_build_id = false; struct dso *pos; struct nscookie nsc; @@ -303,9 +333,10 @@ struct dso *dsos__findnew_id(struct dsos *dsos, const char *name, struct dso_id return dso; } -size_t __dsos__fprintf_buildid(struct list_head *head, FILE *fp, +size_t __dsos__fprintf_buildid(struct dsos *dsos, FILE *fp, bool (skip)(struct dso *dso, int parm), int parm) { + struct list_head *head = &dsos->head; struct dso *pos; size_t ret = 0; @@ -320,8 +351,9 @@ size_t __dsos__fprintf_buildid(struct list_head *head, FILE *fp, return ret; } -size_t __dsos__fprintf(struct list_head *head, FILE *fp) +size_t __dsos__fprintf(struct dsos *dsos, FILE *fp) { + struct list_head *head = &dsos->head; struct dso *pos; size_t ret = 0; @@ -331,3 +363,14 @@ size_t __dsos__fprintf(struct list_head *head, FILE *fp) return ret; } + +int __dsos__hit_all(struct dsos *dsos) +{ + struct list_head *head = &dsos->head; + struct dso *pos; + + list_for_each_entry(pos, head, node) + pos->hit = true; + + return 0; +} diff --git a/tools/perf/util/dsos.h b/tools/perf/util/dsos.h index 5dbec2bc6966..1c81ddf07f8f 100644 --- a/tools/perf/util/dsos.h +++ b/tools/perf/util/dsos.h @@ -21,6 +21,15 @@ struct dsos { struct rw_semaphore lock; }; +#define dsos__for_each_with_build_id(pos, head) \ + list_for_each_entry(pos, head, node) \ + if (!pos->has_build_id) \ + continue; \ + else + +void dsos__init(struct dsos *dsos); +void dsos__exit(struct dsos *dsos); + void __dsos__add(struct dsos *dsos, struct dso *dso); void dsos__add(struct dsos *dsos, struct dso *dso); struct dso *__dsos__addnew(struct dsos *dsos, const char *name); @@ -28,13 +37,15 @@ struct dso *__dsos__find(struct dsos *dsos, const char *name, bool cmp_short); struct dso *dsos__findnew_id(struct dsos *dsos, const char *name, struct dso_id *id); +bool __dsos__read_build_ids(struct dsos *dsos, bool with_hits); + struct dso *__dsos__findnew_link_by_longname_id(struct rb_root *root, struct dso *dso, const char *name, struct dso_id *id); -bool __dsos__read_build_ids(struct list_head *head, bool with_hits); - -size_t __dsos__fprintf_buildid(struct list_head *head, FILE *fp, +size_t __dsos__fprintf_buildid(struct dsos *dsos, FILE *fp, bool (skip)(struct dso *dso, int parm), int parm); -size_t __dsos__fprintf(struct list_head *head, FILE *fp); +size_t __dsos__fprintf(struct dsos *dsos, FILE *fp); + +int __dsos__hit_all(struct dsos *dsos); #endif /* __PERF_DSOS */ diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index a26c8bea58d0..848a94b8602b 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -48,13 +48,6 @@ static struct dso *machine__kernel_dso(struct machine *machine) return map__dso(machine->vmlinux_map); } -static void dsos__init(struct dsos *dsos) -{ - INIT_LIST_HEAD(&dsos->head); - dsos->root = RB_ROOT; - init_rwsem(&dsos->lock); -} - static int machine__set_mmap_name(struct machine *machine) { if (machine__is_host(machine)) @@ -165,28 +158,6 @@ struct machine *machine__new_kallsyms(void) return machine; } -static void dsos__purge(struct dsos *dsos) -{ - struct dso *pos, *n; - - down_write(&dsos->lock); - - list_for_each_entry_safe(pos, n, &dsos->head, node) { - RB_CLEAR_NODE(&pos->rb_node); - pos->root = NULL; - list_del_init(&pos->node); - dso__put(pos); - } - - up_write(&dsos->lock); -} - -static void dsos__exit(struct dsos *dsos) -{ - dsos__purge(dsos); - exit_rwsem(&dsos->lock); -} - void machine__delete_threads(struct machine *machine) { threads__remove_all_threads(&machine->threads); @@ -907,11 +878,11 @@ out: size_t machines__fprintf_dsos(struct machines *machines, FILE *fp) { struct rb_node *nd; - size_t ret = __dsos__fprintf(&machines->host.dsos.head, fp); + size_t ret = __dsos__fprintf(&machines->host.dsos, fp); for (nd = rb_first_cached(&machines->guests); nd; nd = rb_next(nd)) { struct machine *pos = rb_entry(nd, struct machine, rb_node); - ret += __dsos__fprintf(&pos->dsos.head, fp); + ret += __dsos__fprintf(&pos->dsos, fp); } return ret; @@ -920,7 +891,7 @@ size_t machines__fprintf_dsos(struct machines *machines, FILE *fp) size_t machine__fprintf_dsos_buildid(struct machine *m, FILE *fp, bool (skip)(struct dso *dso, int parm), int parm) { - return __dsos__fprintf_buildid(&m->dsos.head, fp, skip, parm); + return __dsos__fprintf_buildid(&m->dsos, fp, skip, parm); } size_t machines__fprintf_dsos_buildid(struct machines *machines, FILE *fp, @@ -3306,3 +3277,8 @@ bool machine__is_lock_function(struct machine *machine, u64 addr) return false; } + +int machine__hit_all_dsos(struct machine *machine) +{ + return __dsos__hit_all(&machine->dsos); +} diff --git a/tools/perf/util/machine.h b/tools/perf/util/machine.h index 4312f6db6de0..82a47bac8023 100644 --- a/tools/perf/util/machine.h +++ b/tools/perf/util/machine.h @@ -306,4 +306,6 @@ int machine__map_x86_64_entry_trampolines(struct machine *machine, int machine__resolve(struct machine *machine, struct addr_location *al, struct perf_sample *sample); +int machine__hit_all_dsos(struct machine *machine); + #endif /* __PERF_MACHINE_H */ diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 02a932a83c51..a10343b9dcd4 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -2916,3 +2916,24 @@ int perf_event__process_id_index(struct perf_session *session, } return 0; } + +int perf_session__dsos_hit_all(struct perf_session *session) +{ + struct rb_node *nd; + int err; + + err = machine__hit_all_dsos(&session->machines.host); + if (err) + return err; + + for (nd = rb_first_cached(&session->machines.guests); nd; + nd = rb_next(nd)) { + struct machine *pos = rb_entry(nd, struct machine, rb_node); + + err = machine__hit_all_dsos(pos); + if (err) + return err; + } + + return 0; +} diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 5064c6ec11e7..3b0256e977a6 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -156,6 +156,8 @@ int perf_session__deliver_synth_event(struct perf_session *session, union perf_event *event, struct perf_sample *sample); +int perf_session__dsos_hit_all(struct perf_session *session); + int perf_event__process_id_index(struct perf_session *session, union perf_event *event); -- cgit v1.2.3 From f649ed80f3cabbf16b228894bb7ecd718da86e47 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 9 Apr 2024 23:42:04 -0700 Subject: perf dsos: Tidy reference counting and locking Move more functionality into dsos.c generally from machine.c, renaming functions to match their new usage. The find function is made to always "get" before returning a dso. Reduce the scope of locks in vdso to match the locking paradigm. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Anne Macedo Cc: Athira Rajeev Cc: Ben Gainey Cc: Changbin Du Cc: Chengen Du Cc: Colin Ian King Cc: Ilkka Koskinen Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kan Liang Cc: Leo Yan Cc: Li Dong Cc: Mark Rutland Cc: Markus Elfring Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Song Liu Cc: Sun Haiyong Cc: Thomas Richter Cc: Yang Jihong Cc: Yanteng Si Cc: Yicong Yang Cc: zhaimingbing Link: https://lore.kernel.org/r/20240410064214.2755936-3-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dsos.c | 73 ++++++++++++++++++++++++++++++++++++++++++----- tools/perf/util/dsos.h | 9 +++++- tools/perf/util/machine.c | 62 ++-------------------------------------- tools/perf/util/map.c | 4 +-- tools/perf/util/vdso.c | 48 +++++++++++++------------------ 5 files changed, 97 insertions(+), 99 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dsos.c b/tools/perf/util/dsos.c index e65ef6762bed..d269e09005a7 100644 --- a/tools/perf/util/dsos.c +++ b/tools/perf/util/dsos.c @@ -181,7 +181,7 @@ struct dso *__dsos__findnew_link_by_longname_id(struct rb_root *root, struct dso * at the end of the list of duplicates. */ if (!dso || (dso == this)) - return this; /* Find matching dso */ + return dso__get(this); /* Find matching dso */ /* * The core kernel DSOs may have duplicated long name. * In this case, the short name should be different. @@ -253,15 +253,20 @@ static struct dso *__dsos__find_id(struct dsos *dsos, const char *name, struct d if (cmp_short) { list_for_each_entry(pos, &dsos->head, node) if (__dso__cmp_short_name(name, id, pos) == 0) - return pos; + return dso__get(pos); return NULL; } return __dsos__findnew_by_longname_id(&dsos->root, name, id); } -struct dso *__dsos__find(struct dsos *dsos, const char *name, bool cmp_short) +struct dso *dsos__find(struct dsos *dsos, const char *name, bool cmp_short) { - return __dsos__find_id(dsos, name, NULL, cmp_short); + struct dso *res; + + down_read(&dsos->lock); + res = __dsos__find_id(dsos, name, NULL, cmp_short); + up_read(&dsos->lock); + return res; } static void dso__set_basename(struct dso *dso) @@ -303,8 +308,6 @@ static struct dso *__dsos__addnew_id(struct dsos *dsos, const char *name, struct if (dso != NULL) { __dsos__add(dsos, dso); dso__set_basename(dso); - /* Put dso here because __dsos_add already got it */ - dso__put(dso); } return dso; } @@ -328,7 +331,7 @@ struct dso *dsos__findnew_id(struct dsos *dsos, const char *name, struct dso_id { struct dso *dso; down_write(&dsos->lock); - dso = dso__get(__dsos__findnew_id(dsos, name, id)); + dso = __dsos__findnew_id(dsos, name, id); up_write(&dsos->lock); return dso; } @@ -374,3 +377,59 @@ int __dsos__hit_all(struct dsos *dsos) return 0; } + +struct dso *dsos__findnew_module_dso(struct dsos *dsos, + struct machine *machine, + struct kmod_path *m, + const char *filename) +{ + struct dso *dso; + + down_write(&dsos->lock); + + dso = __dsos__find_id(dsos, m->name, NULL, /*cmp_short=*/true); + if (!dso) { + dso = __dsos__addnew(dsos, m->name); + if (dso == NULL) + goto out_unlock; + + dso__set_module_info(dso, m, machine); + dso__set_long_name(dso, strdup(filename), true); + dso->kernel = DSO_SPACE__KERNEL; + } + +out_unlock: + up_write(&dsos->lock); + return dso; +} + +struct dso *dsos__find_kernel_dso(struct dsos *dsos) +{ + struct dso *dso, *res = NULL; + + down_read(&dsos->lock); + list_for_each_entry(dso, &dsos->head, node) { + /* + * The cpumode passed to is_kernel_module is not the cpumode of + * *this* event. If we insist on passing correct cpumode to + * is_kernel_module, we should record the cpumode when we adding + * this dso to the linked list. + * + * However we don't really need passing correct cpumode. We + * know the correct cpumode must be kernel mode (if not, we + * should not link it onto kernel_dsos list). + * + * Therefore, we pass PERF_RECORD_MISC_CPUMODE_UNKNOWN. + * is_kernel_module() treats it as a kernel cpumode. + */ + if (!dso->kernel || + is_kernel_module(dso->long_name, + PERF_RECORD_MISC_CPUMODE_UNKNOWN)) + continue; + + res = dso__get(dso); + break; + } + up_read(&dsos->lock); + return res; +} diff --git a/tools/perf/util/dsos.h b/tools/perf/util/dsos.h index 1c81ddf07f8f..a7c7f723c5ff 100644 --- a/tools/perf/util/dsos.h +++ b/tools/perf/util/dsos.h @@ -10,6 +10,8 @@ struct dso; struct dso_id; +struct kmod_path; +struct machine; /* * DSOs are put into both a list for fast iteration and rbtree for fast @@ -33,7 +35,7 @@ void dsos__exit(struct dsos *dsos); void __dsos__add(struct dsos *dsos, struct dso *dso); void dsos__add(struct dsos *dsos, struct dso *dso); struct dso *__dsos__addnew(struct dsos *dsos, const char *name); -struct dso *__dsos__find(struct dsos *dsos, const char *name, bool cmp_short); +struct dso *dsos__find(struct dsos *dsos, const char *name, bool cmp_short); struct dso *dsos__findnew_id(struct dsos *dsos, const char *name, struct dso_id *id); @@ -48,4 +50,9 @@ size_t __dsos__fprintf(struct dsos *dsos, FILE *fp); int __dsos__hit_all(struct dsos *dsos); +struct dso *dsos__findnew_module_dso(struct dsos *dsos, struct machine *machine, + struct kmod_path *m, const char *filename); + +struct dso *dsos__find_kernel_dso(struct dsos *dsos); + #endif /* __PERF_DSOS */ diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index 848a94b8602b..e2c800a1449b 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -646,31 +646,6 @@ int machine__process_lost_samples_event(struct machine *machine __maybe_unused, return 0; } -static struct dso *machine__findnew_module_dso(struct machine *machine, - struct kmod_path *m, - const char *filename) -{ - struct dso *dso; - - down_write(&machine->dsos.lock); - - dso = __dsos__find(&machine->dsos, m->name, true); - if (!dso) { - dso = __dsos__addnew(&machine->dsos, m->name); - if (dso == NULL) - goto out_unlock; - - dso__set_module_info(dso, m, machine); - dso__set_long_name(dso, strdup(filename), true); - dso->kernel = DSO_SPACE__KERNEL; - } - - dso__get(dso); -out_unlock: - up_write(&machine->dsos.lock); - return dso; -} - int machine__process_aux_event(struct machine *machine __maybe_unused, union perf_event *event) { @@ -854,7 +829,7 @@ static struct map *machine__addnew_module_map(struct machine *machine, u64 start if (kmod_path__parse_name(&m, filename)) return NULL; - dso = machine__findnew_module_dso(machine, &m, filename); + dso = dsos__findnew_module_dso(&machine->dsos, machine, &m, filename); if (dso == NULL) goto out; @@ -1663,40 +1638,7 @@ static int machine__process_kernel_mmap_event(struct machine *machine, * Should be there already, from the build-id table in * the header. */ - struct dso *kernel = NULL; - struct dso *dso; - - down_read(&machine->dsos.lock); - - list_for_each_entry(dso, &machine->dsos.head, node) { - - /* - * The cpumode passed to is_kernel_module is not the - * cpumode of *this* event. If we insist on passing - * correct cpumode to is_kernel_module, we should - * record the cpumode when we adding this dso to the - * linked list. - * - * However we don't really need passing correct - * cpumode. We know the correct cpumode must be kernel - * mode (if not, we should not link it onto kernel_dsos - * list). - * - * Therefore, we pass PERF_RECORD_MISC_CPUMODE_UNKNOWN. - * is_kernel_module() treats it as a kernel cpumode. - */ - - if (!dso->kernel || - is_kernel_module(dso->long_name, - PERF_RECORD_MISC_CPUMODE_UNKNOWN)) - continue; - - - kernel = dso__get(dso); - break; - } - - up_read(&machine->dsos.lock); + struct dso *kernel = dsos__find_kernel_dso(&machine->dsos); if (kernel == NULL) kernel = machine__findnew_dso(machine, machine->mmap_name); diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index a5d57c201a30..cca871959f87 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -196,9 +196,7 @@ struct map *map__new(struct machine *machine, u64 start, u64 len, * reading the header will have the build ID set and all future mmaps will * have it missing. */ - down_read(&machine->dsos.lock); - header_bid_dso = __dsos__find(&machine->dsos, filename, false); - up_read(&machine->dsos.lock); + header_bid_dso = dsos__find(&machine->dsos, filename, false); if (header_bid_dso && header_bid_dso->header_build_id) { dso__set_build_id(dso, &header_bid_dso->bid); dso->header_build_id = 1; diff --git a/tools/perf/util/vdso.c b/tools/perf/util/vdso.c index df8963796187..35532dcbff74 100644 --- a/tools/perf/util/vdso.c +++ b/tools/perf/util/vdso.c @@ -133,8 +133,6 @@ static struct dso *__machine__addnew_vdso(struct machine *machine, const char *s if (dso != NULL) { __dsos__add(&machine->dsos, dso); dso__set_long_name(dso, long_name, false); - /* Put dso here because __dsos_add already got it */ - dso__put(dso); } return dso; @@ -252,17 +250,15 @@ static struct dso *__machine__findnew_compat(struct machine *machine, const char *file_name; struct dso *dso; - dso = __dsos__find(&machine->dsos, vdso_file->dso_name, true); + dso = dsos__find(&machine->dsos, vdso_file->dso_name, true); if (dso) - goto out; + return dso; file_name = vdso__get_compat_file(vdso_file); if (!file_name) - goto out; + return NULL; - dso = __machine__addnew_vdso(machine, vdso_file->dso_name, file_name); -out: - return dso; + return __machine__addnew_vdso(machine, vdso_file->dso_name, file_name); } static int __machine__findnew_vdso_compat(struct machine *machine, @@ -308,21 +304,21 @@ static struct dso *machine__find_vdso(struct machine *machine, dso_type = machine__thread_dso_type(machine, thread); switch (dso_type) { case DSO__TYPE_32BIT: - dso = __dsos__find(&machine->dsos, DSO__NAME_VDSO32, true); + dso = dsos__find(&machine->dsos, DSO__NAME_VDSO32, true); if (!dso) { - dso = __dsos__find(&machine->dsos, DSO__NAME_VDSO, - true); + dso = dsos__find(&machine->dsos, DSO__NAME_VDSO, + true); if (dso && dso_type != dso__type(dso, machine)) dso = NULL; } break; case DSO__TYPE_X32BIT: - dso = __dsos__find(&machine->dsos, DSO__NAME_VDSOX32, true); + dso = dsos__find(&machine->dsos, DSO__NAME_VDSOX32, true); break; case DSO__TYPE_64BIT: case DSO__TYPE_UNKNOWN: default: - dso = __dsos__find(&machine->dsos, DSO__NAME_VDSO, true); + dso = dsos__find(&machine->dsos, DSO__NAME_VDSO, true); break; } @@ -334,37 +330,33 @@ struct dso *machine__findnew_vdso(struct machine *machine, { struct vdso_info *vdso_info; struct dso *dso = NULL; + char *file; - down_write(&machine->dsos.lock); if (!machine->vdso_info) machine->vdso_info = vdso_info__new(); vdso_info = machine->vdso_info; if (!vdso_info) - goto out_unlock; + return NULL; dso = machine__find_vdso(machine, thread); if (dso) - goto out_unlock; + return dso; #if BITS_PER_LONG == 64 if (__machine__findnew_vdso_compat(machine, thread, vdso_info, &dso)) - goto out_unlock; + return dso; #endif - dso = __dsos__find(&machine->dsos, DSO__NAME_VDSO, true); - if (!dso) { - char *file; + dso = dsos__find(&machine->dsos, DSO__NAME_VDSO, true); + if (dso) + return dso; - file = get_file(&vdso_info->vdso); - if (file) - dso = __machine__addnew_vdso(machine, DSO__NAME_VDSO, file); - } + file = get_file(&vdso_info->vdso); + if (!file) + return NULL; -out_unlock: - dso__get(dso); - up_write(&machine->dsos.lock); - return dso; + return __machine__addnew_vdso(machine, DSO__NAME_VDSO, file); } bool dso__is_vdso(struct dso *dso) -- cgit v1.2.3 From 73f3fea2e11dbd92f49807ee1c33429674f71f13 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 9 Apr 2024 23:42:05 -0700 Subject: perf dsos: Introduce dsos__for_each_dso() To better abstract the dsos internals, introduce dsos__for_each_dso that does a callback on each dso. This also means the read lock can be correctly held. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Anne Macedo Cc: Athira Rajeev Cc: Ben Gainey Cc: Changbin Du Cc: Chengen Du Cc: Colin Ian King Cc: Ilkka Koskinen Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kan Liang Cc: Leo Yan Cc: Li Dong Cc: Mark Rutland Cc: Markus Elfring Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Song Liu Cc: Sun Haiyong Cc: Thomas Richter Cc: Yang Jihong Cc: Yanteng Si Cc: Yicong Yang Cc: zhaimingbing Link: https://lore.kernel.org/r/20240410064214.2755936-4-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-inject.c | 25 +++++++++------ tools/perf/util/build-id.c | 76 +++++++++++++++++++++++++-------------------- tools/perf/util/dsos.c | 16 ++++++++++ tools/perf/util/dsos.h | 8 ++--- tools/perf/util/machine.c | 40 +++++++++++++++--------- 5 files changed, 100 insertions(+), 65 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c index ef73317e6ae7..ce5e28eaad90 100644 --- a/tools/perf/builtin-inject.c +++ b/tools/perf/builtin-inject.c @@ -1187,23 +1187,28 @@ static int synthesize_build_id(struct perf_inject *inject, struct dso *dso, pid_ process_build_id, machine); } +static int guest_session__add_build_ids_cb(struct dso *dso, void *data) +{ + struct guest_session *gs = data; + struct perf_inject *inject = container_of(gs, struct perf_inject, guest_session); + + if (!dso->has_build_id) + return 0; + + return synthesize_build_id(inject, dso, gs->machine_pid); + +} + static int guest_session__add_build_ids(struct guest_session *gs) { struct perf_inject *inject = container_of(gs, struct perf_inject, guest_session); - struct machine *machine = &gs->session->machines.host; - struct dso *dso; - int ret; /* Build IDs will be put in the Build ID feature section */ perf_header__set_feat(&inject->session->header, HEADER_BUILD_ID); - dsos__for_each_with_build_id(dso, &machine->dsos.head) { - ret = synthesize_build_id(inject, dso, gs->machine_pid); - if (ret) - return ret; - } - - return 0; + return dsos__for_each_dso(&gs->session->machines.host.dsos, + guest_session__add_build_ids_cb, + gs); } static int guest_session__ksymbol_event(struct perf_tool *tool, diff --git a/tools/perf/util/build-id.c b/tools/perf/util/build-id.c index a617b1917e6b..a6d3c253f19f 100644 --- a/tools/perf/util/build-id.c +++ b/tools/perf/util/build-id.c @@ -327,48 +327,56 @@ static int write_buildid(const char *name, size_t name_len, struct build_id *bid return write_padded(fd, name, name_len + 1, len); } -static int machine__write_buildid_table(struct machine *machine, - struct feat_fd *fd) +struct machine__write_buildid_table_cb_args { + struct machine *machine; + struct feat_fd *fd; + u16 kmisc, umisc; +}; + +static int machine__write_buildid_table_cb(struct dso *dso, void *data) { - int err = 0; - struct dso *pos; - u16 kmisc = PERF_RECORD_MISC_KERNEL, - umisc = PERF_RECORD_MISC_USER; + struct machine__write_buildid_table_cb_args *args = data; + const char *name; + size_t name_len; + bool in_kernel = false; - if (!machine__is_host(machine)) { - kmisc = PERF_RECORD_MISC_GUEST_KERNEL; - umisc = PERF_RECORD_MISC_GUEST_USER; - } + if (!dso->has_build_id) + return 0; - dsos__for_each_with_build_id(pos, &machine->dsos.head) { - const char *name; - size_t name_len; - bool in_kernel = false; + if (!dso->hit && !dso__is_vdso(dso)) + return 0; - if (!pos->hit && !dso__is_vdso(pos)) - continue; + if (dso__is_vdso(dso)) { + name = dso->short_name; + name_len = dso->short_name_len; + } else if (dso__is_kcore(dso)) { + name = args->machine->mmap_name; + name_len = strlen(name); + } else { + name = dso->long_name; + name_len = dso->long_name_len; + } - if (dso__is_vdso(pos)) { - name = pos->short_name; - name_len = pos->short_name_len; - } else if (dso__is_kcore(pos)) { - name = machine->mmap_name; - name_len = strlen(name); - } else { - name = pos->long_name; - name_len = pos->long_name_len; - } + in_kernel = dso->kernel || is_kernel_module(name, PERF_RECORD_MISC_CPUMODE_UNKNOWN); + return write_buildid(name, name_len, &dso->bid, args->machine->pid, + in_kernel ? args->kmisc : args->umisc, args->fd); +} - in_kernel = pos->kernel || - is_kernel_module(name, - PERF_RECORD_MISC_CPUMODE_UNKNOWN); - err = write_buildid(name, name_len, &pos->bid, machine->pid, - in_kernel ? kmisc : umisc, fd); - if (err) - break; +static int machine__write_buildid_table(struct machine *machine, struct feat_fd *fd) +{ + struct machine__write_buildid_table_cb_args args = { + .machine = machine, + .fd = fd, + .kmisc = PERF_RECORD_MISC_KERNEL, + .umisc = PERF_RECORD_MISC_USER, + }; + + if (!machine__is_host(machine)) { + args.kmisc = PERF_RECORD_MISC_GUEST_KERNEL; + args.umisc = PERF_RECORD_MISC_GUEST_USER; } - return err; + return dsos__for_each_dso(&machine->dsos, machine__write_buildid_table_cb, &args); } int perf_session__write_buildid_table(struct perf_session *session, diff --git a/tools/perf/util/dsos.c b/tools/perf/util/dsos.c index d269e09005a7..d43f64939b12 100644 --- a/tools/perf/util/dsos.c +++ b/tools/perf/util/dsos.c @@ -433,3 +433,19 @@ struct dso *dsos__find_kernel_dso(struct dsos *dsos) up_read(&dsos->lock); return res; } + +int dsos__for_each_dso(struct dsos *dsos, int (*cb)(struct dso *dso, void *data), void *data) +{ + struct dso *dso; + + down_read(&dsos->lock); + list_for_each_entry(dso, &dsos->head, node) { + int err; + + err = cb(dso, data); + if (err) + return err; + } + up_read(&dsos->lock); + return 0; +} diff --git a/tools/perf/util/dsos.h b/tools/perf/util/dsos.h index a7c7f723c5ff..317a263f0e37 100644 --- a/tools/perf/util/dsos.h +++ b/tools/perf/util/dsos.h @@ -23,12 +23,6 @@ struct dsos { struct rw_semaphore lock; }; -#define dsos__for_each_with_build_id(pos, head) \ - list_for_each_entry(pos, head, node) \ - if (!pos->has_build_id) \ - continue; \ - else - void dsos__init(struct dsos *dsos); void dsos__exit(struct dsos *dsos); @@ -55,4 +49,6 @@ struct dso *dsos__findnew_module_dso(struct dsos *dsos, struct machine *machine, struct dso *dsos__find_kernel_dso(struct dsos *dsos); +int dsos__for_each_dso(struct dsos *dsos, int (*cb)(struct dso *dso, void *data), void *data); + #endif /* __PERF_DSOS */ diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index e2c800a1449b..e3deef38405c 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -1562,16 +1562,14 @@ out_put: return ret; } -static bool machine__uses_kcore(struct machine *machine) +static int machine__uses_kcore_cb(struct dso *dso, void *data __maybe_unused) { - struct dso *dso; - - list_for_each_entry(dso, &machine->dsos.head, node) { - if (dso__is_kcore(dso)) - return true; - } + return dso__is_kcore(dso) ? 1 : 0; +} - return false; +static bool machine__uses_kcore(struct machine *machine) +{ + return dsos__for_each_dso(&machine->dsos, machine__uses_kcore_cb, NULL) != 0 ? true : false; } static bool perf_event__is_extra_kernel_mmap(struct machine *machine, @@ -3137,16 +3135,28 @@ char *machine__resolve_kernel_addr(void *vmachine, unsigned long long *addrp, ch return sym->name; } +struct machine__for_each_dso_cb_args { + struct machine *machine; + machine__dso_t fn; + void *priv; +}; + +static int machine__for_each_dso_cb(struct dso *dso, void *data) +{ + struct machine__for_each_dso_cb_args *args = data; + + return args->fn(dso, args->machine, args->priv); +} + int machine__for_each_dso(struct machine *machine, machine__dso_t fn, void *priv) { - struct dso *pos; - int err = 0; + struct machine__for_each_dso_cb_args args = { + .machine = machine, + .fn = fn, + .priv = priv, + }; - list_for_each_entry(pos, &machine->dsos.head, node) { - if (fn(pos, machine, priv)) - err = -1; - } - return err; + return dsos__for_each_dso(&machine->dsos, machine__for_each_dso_cb, &args); } int machine__for_each_kernel_map(struct machine *machine, machine__map_t fn, void *priv) -- cgit v1.2.3 From 1d6eff930595eef83cfb8486c122249afb66145d Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 9 Apr 2024 23:42:06 -0700 Subject: perf dso: Move dso functions out of dsos.c Move dso and dso_id functions to dso.c to match the struct declarations. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Anne Macedo Cc: Athira Rajeev Cc: Ben Gainey Cc: Changbin Du Cc: Chengen Du Cc: Colin Ian King Cc: Ilkka Koskinen Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kan Liang Cc: Leo Yan Cc: Li Dong Cc: Mark Rutland Cc: Markus Elfring Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Song Liu Cc: Sun Haiyong Cc: Thomas Richter Cc: Yang Jihong Cc: Yanteng Si Cc: Yicong Yang Cc: zhaimingbing Link: https://lore.kernel.org/r/20240410064214.2755936-5-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dso.c | 61 ++++++++++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/dso.h | 4 ++++ tools/perf/util/dsos.c | 61 -------------------------------------------------- 3 files changed, 65 insertions(+), 61 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c index 6e2a7198b382..ad562743d769 100644 --- a/tools/perf/util/dso.c +++ b/tools/perf/util/dso.c @@ -1269,6 +1269,67 @@ static void dso__set_long_name_id(struct dso *dso, const char *name, struct dso_ __dsos__findnew_link_by_longname_id(root, dso, NULL, id); } +static int __dso_id__cmp(struct dso_id *a, struct dso_id *b) +{ + if (a->maj > b->maj) return -1; + if (a->maj < b->maj) return 1; + + if (a->min > b->min) return -1; + if (a->min < b->min) return 1; + + if (a->ino > b->ino) return -1; + if (a->ino < b->ino) return 1; + + /* + * Synthesized MMAP events have zero ino_generation, avoid comparing + * them with MMAP events with actual ino_generation. + * + * I found it harmful because the mismatch resulted in a new + * dso that did not have a build ID whereas the original dso did have a + * build ID. The build ID was essential because the object was not found + * otherwise. - Adrian + */ + if (a->ino_generation && b->ino_generation) { + if (a->ino_generation > b->ino_generation) return -1; + if (a->ino_generation < b->ino_generation) return 1; + } + + return 0; +} + +bool dso_id__empty(struct dso_id *id) +{ + if (!id) + return true; + + return !id->maj && !id->min && !id->ino && !id->ino_generation; +} + +void dso__inject_id(struct dso *dso, struct dso_id *id) +{ + dso->id.maj = id->maj; + dso->id.min = id->min; + dso->id.ino = id->ino; + dso->id.ino_generation = id->ino_generation; +} + +int dso_id__cmp(struct dso_id *a, struct dso_id *b) +{ + /* + * The second is always dso->id, so zeroes if not set, assume passing + * NULL for a means a zeroed id + */ + if (dso_id__empty(a) || dso_id__empty(b)) + return 0; + + return __dso_id__cmp(a, b); +} + +int dso__cmp_id(struct dso *a, struct dso *b) +{ + return __dso_id__cmp(&a->id, &b->id); +} + void dso__set_long_name(struct dso *dso, const char *name, bool name_allocated) { dso__set_long_name_id(dso, name, NULL, name_allocated); diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h index 3d4faad8d5dc..2c295438226d 100644 --- a/tools/perf/util/dso.h +++ b/tools/perf/util/dso.h @@ -238,6 +238,9 @@ static inline void dso__set_loaded(struct dso *dso) dso->loaded = true; } +int dso_id__cmp(struct dso_id *a, struct dso_id *b); +bool dso_id__empty(struct dso_id *id); + struct dso *dso__new_id(const char *name, struct dso_id *id); struct dso *dso__new(const char *name); void dso__delete(struct dso *dso); @@ -245,6 +248,7 @@ void dso__delete(struct dso *dso); int dso__cmp_id(struct dso *a, struct dso *b); void dso__set_short_name(struct dso *dso, const char *name, bool name_allocated); void dso__set_long_name(struct dso *dso, const char *name, bool name_allocated); +void dso__inject_id(struct dso *dso, struct dso_id *id); int dso__name_len(const struct dso *dso); diff --git a/tools/perf/util/dsos.c b/tools/perf/util/dsos.c index d43f64939b12..f816927a21ff 100644 --- a/tools/perf/util/dsos.c +++ b/tools/perf/util/dsos.c @@ -41,67 +41,6 @@ void dsos__exit(struct dsos *dsos) exit_rwsem(&dsos->lock); } -static int __dso_id__cmp(struct dso_id *a, struct dso_id *b) -{ - if (a->maj > b->maj) return -1; - if (a->maj < b->maj) return 1; - - if (a->min > b->min) return -1; - if (a->min < b->min) return 1; - - if (a->ino > b->ino) return -1; - if (a->ino < b->ino) return 1; - - /* - * Synthesized MMAP events have zero ino_generation, avoid comparing - * them with MMAP events with actual ino_generation. - * - * I found it harmful because the mismatch resulted in a new - * dso that did not have a build ID whereas the original dso did have a - * build ID. The build ID was essential because the object was not found - * otherwise. - Adrian - */ - if (a->ino_generation && b->ino_generation) { - if (a->ino_generation > b->ino_generation) return -1; - if (a->ino_generation < b->ino_generation) return 1; - } - - return 0; -} - -static bool dso_id__empty(struct dso_id *id) -{ - if (!id) - return true; - - return !id->maj && !id->min && !id->ino && !id->ino_generation; -} - -static void dso__inject_id(struct dso *dso, struct dso_id *id) -{ - dso->id.maj = id->maj; - dso->id.min = id->min; - dso->id.ino = id->ino; - dso->id.ino_generation = id->ino_generation; -} - -static int dso_id__cmp(struct dso_id *a, struct dso_id *b) -{ - /* - * The second is always dso->id, so zeroes if not set, assume passing - * NULL for a means a zeroed id - */ - if (dso_id__empty(a) || dso_id__empty(b)) - return 0; - - return __dso_id__cmp(a, b); -} - -int dso__cmp_id(struct dso *a, struct dso *b) -{ - return __dso_id__cmp(&a->id, &b->id); -} - bool __dsos__read_build_ids(struct dsos *dsos, bool with_hits) { struct list_head *head = &dsos->head; -- cgit v1.2.3 From 0ffc8fca5c15a70f32c8aff12c566bbd3991bd0a Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 9 Apr 2024 23:42:07 -0700 Subject: perf dsos: Switch more loops to dsos__for_each_dso() Switch loops within dsos.c, add a version that isn't locked. Switch some unlocked loops to hold the read lock. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Anne Macedo Cc: Athira Rajeev Cc: Ben Gainey Cc: Changbin Du Cc: Chengen Du Cc: Colin Ian King Cc: Ilkka Koskinen Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kan Liang Cc: Leo Yan Cc: Li Dong Cc: Mark Rutland Cc: Markus Elfring Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Song Liu Cc: Sun Haiyong Cc: Thomas Richter Cc: Yang Jihong Cc: Yanteng Si Cc: Yicong Yang Cc: zhaimingbing Link: https://lore.kernel.org/r/20240410064214.2755936-6-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/build-id.c | 2 +- tools/perf/util/dsos.c | 258 +++++++++++++++++++++++++++++---------------- tools/perf/util/dsos.h | 8 +- tools/perf/util/machine.c | 8 +- 4 files changed, 174 insertions(+), 102 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/build-id.c b/tools/perf/util/build-id.c index a6d3c253f19f..864bc26b6b46 100644 --- a/tools/perf/util/build-id.c +++ b/tools/perf/util/build-id.c @@ -964,7 +964,7 @@ int perf_session__cache_build_ids(struct perf_session *session) static bool machine__read_build_ids(struct machine *machine, bool with_hits) { - return __dsos__read_build_ids(&machine->dsos, with_hits); + return dsos__read_build_ids(&machine->dsos, with_hits); } bool perf_session__read_build_ids(struct perf_session *session, bool with_hits) diff --git a/tools/perf/util/dsos.c b/tools/perf/util/dsos.c index f816927a21ff..b7fbfb877ae3 100644 --- a/tools/perf/util/dsos.c +++ b/tools/perf/util/dsos.c @@ -41,38 +41,65 @@ void dsos__exit(struct dsos *dsos) exit_rwsem(&dsos->lock); } -bool __dsos__read_build_ids(struct dsos *dsos, bool with_hits) + +static int __dsos__for_each_dso(struct dsos *dsos, + int (*cb)(struct dso *dso, void *data), + void *data) +{ + struct dso *dso; + + list_for_each_entry(dso, &dsos->head, node) { + int err; + + err = cb(dso, data); + if (err) + return err; + } + return 0; +} + +struct dsos__read_build_ids_cb_args { + bool with_hits; + bool have_build_id; +}; + +static int dsos__read_build_ids_cb(struct dso *dso, void *data) { - struct list_head *head = &dsos->head; - bool have_build_id = false; - struct dso *pos; + struct dsos__read_build_ids_cb_args *args = data; struct nscookie nsc; - list_for_each_entry(pos, head, node) { - if (with_hits && !pos->hit && !dso__is_vdso(pos)) - continue; - if (pos->has_build_id) { - have_build_id = true; - continue; - } - nsinfo__mountns_enter(pos->nsinfo, &nsc); - if (filename__read_build_id(pos->long_name, &pos->bid) > 0) { - have_build_id = true; - pos->has_build_id = true; - } else if (errno == ENOENT && pos->nsinfo) { - char *new_name = dso__filename_with_chroot(pos, pos->long_name); - - if (new_name && filename__read_build_id(new_name, - &pos->bid) > 0) { - have_build_id = true; - pos->has_build_id = true; - } - free(new_name); + if (args->with_hits && !dso->hit && !dso__is_vdso(dso)) + return 0; + if (dso->has_build_id) { + args->have_build_id = true; + return 0; + } + nsinfo__mountns_enter(dso->nsinfo, &nsc); + if (filename__read_build_id(dso->long_name, &dso->bid) > 0) { + args->have_build_id = true; + dso->has_build_id = true; + } else if (errno == ENOENT && dso->nsinfo) { + char *new_name = dso__filename_with_chroot(dso, dso->long_name); + + if (new_name && filename__read_build_id(new_name, &dso->bid) > 0) { + args->have_build_id = true; + dso->has_build_id = true; } - nsinfo__mountns_exit(&nsc); + free(new_name); } + nsinfo__mountns_exit(&nsc); + return 0; +} - return have_build_id; +bool dsos__read_build_ids(struct dsos *dsos, bool with_hits) +{ + struct dsos__read_build_ids_cb_args args = { + .with_hits = with_hits, + .have_build_id = false, + }; + + dsos__for_each_dso(dsos, dsos__read_build_ids_cb, &args); + return args.have_build_id; } static int __dso__cmp_long_name(const char *long_name, struct dso_id *id, struct dso *b) @@ -105,6 +132,7 @@ struct dso *__dsos__findnew_link_by_longname_id(struct rb_root *root, struct dso if (!name) name = dso->long_name; + /* * Find node with the matching name */ @@ -185,17 +213,40 @@ static struct dso *__dsos__findnew_by_longname_id(struct rb_root *root, const ch return __dsos__findnew_link_by_longname_id(root, NULL, name, id); } +struct dsos__find_id_cb_args { + const char *name; + struct dso_id *id; + struct dso *res; +}; + +static int dsos__find_id_cb(struct dso *dso, void *data) +{ + struct dsos__find_id_cb_args *args = data; + + if (__dso__cmp_short_name(args->name, args->id, dso) == 0) { + args->res = dso__get(dso); + return 1; + } + return 0; + +} + static struct dso *__dsos__find_id(struct dsos *dsos, const char *name, struct dso_id *id, bool cmp_short) { - struct dso *pos; + struct dso *res; if (cmp_short) { - list_for_each_entry(pos, &dsos->head, node) - if (__dso__cmp_short_name(name, id, pos) == 0) - return dso__get(pos); - return NULL; + struct dsos__find_id_cb_args args = { + .name = name, + .id = id, + .res = NULL, + }; + + __dsos__for_each_dso(dsos, dsos__find_id_cb, &args); + return args.res; } - return __dsos__findnew_by_longname_id(&dsos->root, name, id); + res = __dsos__findnew_by_longname_id(&dsos->root, name, id); + return res; } struct dso *dsos__find(struct dsos *dsos, const char *name, bool cmp_short) @@ -275,48 +326,74 @@ struct dso *dsos__findnew_id(struct dsos *dsos, const char *name, struct dso_id return dso; } -size_t __dsos__fprintf_buildid(struct dsos *dsos, FILE *fp, - bool (skip)(struct dso *dso, int parm), int parm) -{ - struct list_head *head = &dsos->head; - struct dso *pos; - size_t ret = 0; +struct dsos__fprintf_buildid_cb_args { + FILE *fp; + bool (*skip)(struct dso *dso, int parm); + int parm; + size_t ret; +}; - list_for_each_entry(pos, head, node) { - char sbuild_id[SBUILD_ID_SIZE]; +static int dsos__fprintf_buildid_cb(struct dso *dso, void *data) +{ + struct dsos__fprintf_buildid_cb_args *args = data; + char sbuild_id[SBUILD_ID_SIZE]; - if (skip && skip(pos, parm)) - continue; - build_id__sprintf(&pos->bid, sbuild_id); - ret += fprintf(fp, "%-40s %s\n", sbuild_id, pos->long_name); - } - return ret; + if (args->skip && args->skip(dso, args->parm)) + return 0; + build_id__sprintf(&dso->bid, sbuild_id); + args->ret += fprintf(args->fp, "%-40s %s\n", sbuild_id, dso->long_name); + return 0; } -size_t __dsos__fprintf(struct dsos *dsos, FILE *fp) +size_t dsos__fprintf_buildid(struct dsos *dsos, FILE *fp, + bool (*skip)(struct dso *dso, int parm), int parm) { - struct list_head *head = &dsos->head; - struct dso *pos; - size_t ret = 0; + struct dsos__fprintf_buildid_cb_args args = { + .fp = fp, + .skip = skip, + .parm = parm, + .ret = 0, + }; + + dsos__for_each_dso(dsos, dsos__fprintf_buildid_cb, &args); + return args.ret; +} - list_for_each_entry(pos, head, node) { - ret += dso__fprintf(pos, fp); - } +struct dsos__fprintf_cb_args { + FILE *fp; + size_t ret; +}; - return ret; +static int dsos__fprintf_cb(struct dso *dso, void *data) +{ + struct dsos__fprintf_cb_args *args = data; + + args->ret += dso__fprintf(dso, args->fp); + return 0; } -int __dsos__hit_all(struct dsos *dsos) +size_t dsos__fprintf(struct dsos *dsos, FILE *fp) { - struct list_head *head = &dsos->head; - struct dso *pos; + struct dsos__fprintf_cb_args args = { + .fp = fp, + .ret = 0, + }; - list_for_each_entry(pos, head, node) - pos->hit = true; + dsos__for_each_dso(dsos, dsos__fprintf_cb, &args); + return args.ret; +} +static int dsos__hit_all_cb(struct dso *dso, void *data __maybe_unused) +{ + dso->hit = true; return 0; } +int dsos__hit_all(struct dsos *dsos) +{ + return dsos__for_each_dso(dsos, dsos__hit_all_cb, NULL); +} + struct dso *dsos__findnew_module_dso(struct dsos *dsos, struct machine *machine, struct kmod_path *m, @@ -342,49 +419,44 @@ out_unlock: return dso; } -struct dso *dsos__find_kernel_dso(struct dsos *dsos) +static int dsos__find_kernel_dso_cb(struct dso *dso, void *data) { - struct dso *dso, *res = NULL; + struct dso **res = data; + /* + * The cpumode passed to is_kernel_module is not the cpumode of *this* + * event. If we insist on passing correct cpumode to is_kernel_module, + * we should record the cpumode when we adding this dso to the linked + * list. + * + * However we don't really need passing correct cpumode. We know the + * correct cpumode must be kernel mode (if not, we should not link it + * onto kernel_dsos list). + * + * Therefore, we pass PERF_RECORD_MISC_CPUMODE_UNKNOWN. + * is_kernel_module() treats it as a kernel cpumode. + */ + if (!dso->kernel || + is_kernel_module(dso->long_name, PERF_RECORD_MISC_CPUMODE_UNKNOWN)) + return 0; - down_read(&dsos->lock); - list_for_each_entry(dso, &dsos->head, node) { - /* - * The cpumode passed to is_kernel_module is not the cpumode of - * *this* event. If we insist on passing correct cpumode to - * is_kernel_module, we should record the cpumode when we adding - * this dso to the linked list. - * - * However we don't really need passing correct cpumode. We - * know the correct cpumode must be kernel mode (if not, we - * should not link it onto kernel_dsos list). - * - * Therefore, we pass PERF_RECORD_MISC_CPUMODE_UNKNOWN. - * is_kernel_module() treats it as a kernel cpumode. - */ - if (!dso->kernel || - is_kernel_module(dso->long_name, - PERF_RECORD_MISC_CPUMODE_UNKNOWN)) - continue; + *res = dso__get(dso); + return 1; +} - res = dso__get(dso); - break; - } - up_read(&dsos->lock); +struct dso *dsos__find_kernel_dso(struct dsos *dsos) +{ + struct dso *res = NULL; + + dsos__for_each_dso(dsos, dsos__find_kernel_dso_cb, &res); return res; } int dsos__for_each_dso(struct dsos *dsos, int (*cb)(struct dso *dso, void *data), void *data) { - struct dso *dso; + int err; down_read(&dsos->lock); - list_for_each_entry(dso, &dsos->head, node) { - int err; - - err = cb(dso, data); - if (err) - return err; - } + err = __dsos__for_each_dso(dsos, cb, data); up_read(&dsos->lock); - return 0; + return err; } diff --git a/tools/perf/util/dsos.h b/tools/perf/util/dsos.h index 317a263f0e37..50bd51523475 100644 --- a/tools/perf/util/dsos.h +++ b/tools/perf/util/dsos.h @@ -33,16 +33,16 @@ struct dso *dsos__find(struct dsos *dsos, const char *name, bool cmp_short); struct dso *dsos__findnew_id(struct dsos *dsos, const char *name, struct dso_id *id); -bool __dsos__read_build_ids(struct dsos *dsos, bool with_hits); +bool dsos__read_build_ids(struct dsos *dsos, bool with_hits); struct dso *__dsos__findnew_link_by_longname_id(struct rb_root *root, struct dso *dso, const char *name, struct dso_id *id); -size_t __dsos__fprintf_buildid(struct dsos *dsos, FILE *fp, +size_t dsos__fprintf_buildid(struct dsos *dsos, FILE *fp, bool (skip)(struct dso *dso, int parm), int parm); -size_t __dsos__fprintf(struct dsos *dsos, FILE *fp); +size_t dsos__fprintf(struct dsos *dsos, FILE *fp); -int __dsos__hit_all(struct dsos *dsos); +int dsos__hit_all(struct dsos *dsos); struct dso *dsos__findnew_module_dso(struct dsos *dsos, struct machine *machine, struct kmod_path *m, const char *filename); diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index e3deef38405c..c5c895131bca 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -853,11 +853,11 @@ out: size_t machines__fprintf_dsos(struct machines *machines, FILE *fp) { struct rb_node *nd; - size_t ret = __dsos__fprintf(&machines->host.dsos, fp); + size_t ret = dsos__fprintf(&machines->host.dsos, fp); for (nd = rb_first_cached(&machines->guests); nd; nd = rb_next(nd)) { struct machine *pos = rb_entry(nd, struct machine, rb_node); - ret += __dsos__fprintf(&pos->dsos, fp); + ret += dsos__fprintf(&pos->dsos, fp); } return ret; @@ -866,7 +866,7 @@ size_t machines__fprintf_dsos(struct machines *machines, FILE *fp) size_t machine__fprintf_dsos_buildid(struct machine *m, FILE *fp, bool (skip)(struct dso *dso, int parm), int parm) { - return __dsos__fprintf_buildid(&m->dsos, fp, skip, parm); + return dsos__fprintf_buildid(&m->dsos, fp, skip, parm); } size_t machines__fprintf_dsos_buildid(struct machines *machines, FILE *fp, @@ -3232,5 +3232,5 @@ bool machine__is_lock_function(struct machine *machine, u64 addr) int machine__hit_all_dsos(struct machine *machine) { - return __dsos__hit_all(&machine->dsos); + return dsos__hit_all(&machine->dsos); } -- cgit v1.2.3 From 61ff60aab7d6846a5983804d96a316b22aacefa1 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 8 Apr 2024 19:32:15 -0700 Subject: perf util: Add shellcheck to generate-cmdlist.sh Add shellcheck to generate-cmdlist.sh to avoid basic shell script mistakes. Reviewed-by: James Clark Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Athira Rajeev Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kajol Jain Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Oliver Upton Cc: Peter Zijlstra Cc: Ravi Bangoria Link: https://lore.kernel.org/r/20240409023216.2342032-4-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/Build | 14 ++++++++++++++ 1 file changed, 14 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/Build b/tools/perf/util/Build index aec5a590e349..292170a99ab6 100644 --- a/tools/perf/util/Build +++ b/tools/perf/util/Build @@ -389,3 +389,17 @@ $(OUTPUT)util/vsprintf.o: ../lib/vsprintf.c FORCE $(OUTPUT)util/list_sort.o: ../lib/list_sort.c FORCE $(call rule_mkdir) $(call if_changed_dep,cc_o_c) + +ifdef SHELLCHECK + SHELL_TESTS := generate-cmdlist.sh + TEST_LOGS := $(SHELL_TESTS:%=%.shellcheck_log) +else + SHELL_TESTS := + TEST_LOGS := +endif + +$(OUTPUT)%.shellcheck_log: % + $(call rule_mkdir) + $(Q)$(call echo-cmd,test)shellcheck -a -S warning "$<" > $@ || (cat $@ && rm $@ && false) + +perf-y += $(TEST_LOGS) -- cgit v1.2.3 From 988052f4bfcc4ee893ea19e9d9ce888cc8578e5a Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Fri, 5 Apr 2024 21:09:11 -0700 Subject: perf bench uprobe: Add uretprobe variant of uprobe benchmarks Name benchmarks with _ret at the end to avoid creating a new set of benchmarks. Signed-off-by: Ian Rogers Acked-by: Jiri Olsa Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andrei Vagin Cc: Ingo Molnar Cc: Kan Liang Cc: Kees Kook Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240406040911.1603801-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/bench/bench.h | 2 ++ tools/perf/bench/uprobe.c | 20 +++++++++++++++++--- tools/perf/builtin-bench.c | 2 ++ tools/perf/util/bpf_skel/bench_uprobe.bpf.c | 16 ++++++++++++++++ 4 files changed, 37 insertions(+), 3 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/bench/bench.h b/tools/perf/bench/bench.h index faa18e6d2467..9f736423af53 100644 --- a/tools/perf/bench/bench.h +++ b/tools/perf/bench/bench.h @@ -46,6 +46,8 @@ int bench_breakpoint_enable(int argc, const char **argv); int bench_uprobe_baseline(int argc, const char **argv); int bench_uprobe_empty(int argc, const char **argv); int bench_uprobe_trace_printk(int argc, const char **argv); +int bench_uprobe_empty_ret(int argc, const char **argv); +int bench_uprobe_trace_printk_ret(int argc, const char **argv); int bench_pmu_scan(int argc, const char **argv); #define BENCH_FORMAT_DEFAULT_STR "default" diff --git a/tools/perf/bench/uprobe.c b/tools/perf/bench/uprobe.c index b722ff88fe7d..0b90275862e1 100644 --- a/tools/perf/bench/uprobe.c +++ b/tools/perf/bench/uprobe.c @@ -26,9 +26,11 @@ static int loops = LOOPS_DEFAULT; enum bench_uprobe { - BENCH_UPROBE__BASELINE, - BENCH_UPROBE__EMPTY, - BENCH_UPROBE__TRACE_PRINTK, + BENCH_UPROBE__BASELINE, + BENCH_UPROBE__EMPTY, + BENCH_UPROBE__TRACE_PRINTK, + BENCH_UPROBE__EMPTY_RET, + BENCH_UPROBE__TRACE_PRINTK_RET, }; static const struct option options[] = { @@ -81,6 +83,8 @@ static int bench_uprobe__setup_bpf_skel(enum bench_uprobe bench) case BENCH_UPROBE__BASELINE: break; case BENCH_UPROBE__EMPTY: bench_uprobe__attach_uprobe(empty); break; case BENCH_UPROBE__TRACE_PRINTK: bench_uprobe__attach_uprobe(trace_printk); break; + case BENCH_UPROBE__EMPTY_RET: bench_uprobe__attach_uprobe(empty_ret); break; + case BENCH_UPROBE__TRACE_PRINTK_RET: bench_uprobe__attach_uprobe(trace_printk_ret); break; default: fprintf(stderr, "Invalid bench: %d\n", bench); goto cleanup; @@ -197,3 +201,13 @@ int bench_uprobe_trace_printk(int argc, const char **argv) { return bench_uprobe(argc, argv, BENCH_UPROBE__TRACE_PRINTK); } + +int bench_uprobe_empty_ret(int argc, const char **argv) +{ + return bench_uprobe(argc, argv, BENCH_UPROBE__EMPTY_RET); +} + +int bench_uprobe_trace_printk_ret(int argc, const char **argv) +{ + return bench_uprobe(argc, argv, BENCH_UPROBE__TRACE_PRINTK_RET); +} diff --git a/tools/perf/builtin-bench.c b/tools/perf/builtin-bench.c index 1a8898d5b560..2c1a9f3d847a 100644 --- a/tools/perf/builtin-bench.c +++ b/tools/perf/builtin-bench.c @@ -109,6 +109,8 @@ static struct bench uprobe_benchmarks[] = { { "baseline", "Baseline libc usleep(1000) call", bench_uprobe_baseline, }, { "empty", "Attach empty BPF prog to uprobe on usleep, system wide", bench_uprobe_empty, }, { "trace_printk", "Attach trace_printk BPF prog to uprobe on usleep syswide", bench_uprobe_trace_printk, }, + { "empty_ret", "Attach empty BPF prog to uretprobe on usleep, system wide", bench_uprobe_empty_ret, }, + { "trace_printk_ret", "Attach trace_printk BPF prog to uretprobe on usleep syswide", bench_uprobe_trace_printk_ret,}, { NULL, NULL, NULL }, }; diff --git a/tools/perf/util/bpf_skel/bench_uprobe.bpf.c b/tools/perf/util/bpf_skel/bench_uprobe.bpf.c index 2c55896bb33c..a01c7f791fcd 100644 --- a/tools/perf/util/bpf_skel/bench_uprobe.bpf.c +++ b/tools/perf/util/bpf_skel/bench_uprobe.bpf.c @@ -4,6 +4,7 @@ #include unsigned int nr_uprobes; +unsigned int nr_uretprobes; SEC("uprobe") int BPF_UPROBE(empty) @@ -20,4 +21,19 @@ int BPF_UPROBE(trace_printk) return 0; } +SEC("uretprobe") +int BPF_URETPROBE(empty_ret) +{ + return 0; +} + +SEC("uretprobe") +int BPF_URETPROBE(trace_printk_ret) +{ + char fmt[] = "perf bench uretprobe %u"; + + bpf_trace_printk(fmt, sizeof(fmt), ++nr_uretprobes); + return 0; +} + char LICENSE[] SEC("license") = "Dual BSD/GPL"; -- cgit v1.2.3 From 2bc3cf575a162a2ca9c98262a63e95cc2b619de7 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 12 Apr 2024 11:33:07 -0700 Subject: perf annotate-data: Improve debug message with location info To verify it found the correct variable, let's add the location expression to the debug message. $ perf --debug type-profile annotate --data-type ... ----------------------------------------------------------- find data type for 0xaf0(reg15) at schedule+0xeb CU for kernel/sched/core.c (die:0x1180523) frame base: cfa=0 fbreg=6 found "rq" in scope=3/4 (die: 0x11b6a00) type_offset=0xaf0 variable location: reg15 type='struct rq' size=0xfc0 (die:0x11892e2) ----------------------------------------------------------- find data type for 0x7bc(reg3) at tcp_get_info+0x62 CU for net/ipv4/tcp.c (die:0x7b5f516) frame base: cfa=0 fbreg=6 offset: 1980 is bigger than size: 760 check variable "sk" failed (die: 0x7b92b2c) variable location: reg3 type='struct sock' size=0x2f8 (die:0x7b63c3a) ----------------------------------------------------------- ... The first case is fine. It looked up a data type in r15 with offset of 0xaf0 at schedule+0xeb. It found the CU die and the frame base info and the variable "rq" was found in the scope 3/4. Its location is the r15 register and the type size is 0xfc0 which includes 0xaf0. But the second case is not good. It looked up a data type in rbx (reg3) with offset 0x7bc. It found a CU and the frame base which is good so far. And it also found a variable "sk" but the access offset is bigger than the type size (1980 vs. 760 or 0x7bc vs. 0x2f8). The variable has the right location (reg3) but I need to figure out why it accesses beyond what it's supposed to. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240412183310.2518474-2-namhyung@kernel.org [ Fix the build on 32-bit by casting Dwarf_Word to (long) in pr_debug_location() ] Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 99 ++++++++++++++++++++++++++++++++++------- 1 file changed, 82 insertions(+), 17 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 1cd857400038..e53d66c46c54 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -46,6 +46,7 @@ static void pr_debug_type_name(Dwarf_Die *die, enum type_state_kind kind) { struct strbuf sb; char *str; + Dwarf_Word size = 0; if (!debug_type_profile && verbose < 3) return; @@ -72,13 +73,67 @@ static void pr_debug_type_name(Dwarf_Die *die, enum type_state_kind kind) break; } + dwarf_aggregate_size(die, &size); + strbuf_init(&sb, 32); die_get_typename_from_type(die, &sb); str = strbuf_detach(&sb, NULL); - pr_info(" type=%s (die:%lx)\n", str, (long)dwarf_dieoffset(die)); + pr_info(" type='%s' size=%#lx (die:%#lx)\n", + str, (long)size, (long)dwarf_dieoffset(die)); free(str); } +static void pr_debug_location(Dwarf_Die *die, u64 pc, int reg) +{ + ptrdiff_t off = 0; + Dwarf_Attribute attr; + Dwarf_Addr base, start, end; + Dwarf_Op *ops; + size_t nops; + + if (!debug_type_profile && verbose < 3) + return; + + if (dwarf_attr(die, DW_AT_location, &attr) == NULL) + return; + + while ((off = dwarf_getlocations(&attr, off, &base, &start, &end, &ops, &nops)) > 0) { + if (reg != DWARF_REG_PC && end < pc) + continue; + if (reg != DWARF_REG_PC && start > pc) + break; + + pr_info(" variable location: "); + switch (ops->atom) { + case DW_OP_reg0 ...DW_OP_reg31: + pr_info("reg%d\n", ops->atom - DW_OP_reg0); + break; + case DW_OP_breg0 ...DW_OP_breg31: + pr_info("base=reg%d, offset=%#lx\n", + ops->atom - DW_OP_breg0, (long)ops->number); + break; + case DW_OP_regx: + pr_info("reg%ld\n", (long)ops->number); + break; + case DW_OP_bregx: + pr_info("base=reg%ld, offset=%#lx\n", + (long)ops->number, (long)ops->number2); + break; + case DW_OP_fbreg: + pr_info("use frame base, offset=%#lx\n", (long)ops->number); + break; + case DW_OP_addr: + pr_info("address=%#lx\n", (long)ops->number); + break; + default: + pr_info("unknown: code=%#x, number=%#lx\n", + ops->atom, (long)ops->number); + break; + } + break; + } +} + /* * Type information in a register, valid when @ok is true. * The @caller_saved registers are invalidated after a function call. @@ -1404,7 +1459,7 @@ again: found = find_data_type_insn(dloc, reg, &basic_blocks, var_types, cu_die, type_die); if (found > 0) { - pr_debug_dtp("found by insn track: %#x(reg%d) type-offset=%#x", + pr_debug_dtp("found by insn track: %#x(reg%d) type-offset=%#x\n", dloc->op->offset, reg, dloc->type_offset); pr_debug_type_name(type_die, TSR_KIND_TYPE); ret = 0; @@ -1440,16 +1495,16 @@ static int find_data_type_die(struct data_loc_info *dloc, Dwarf_Die *type_die) char buf[64]; if (dloc->op->multi_regs) - snprintf(buf, sizeof(buf), " or reg%d", dloc->op->reg2); + snprintf(buf, sizeof(buf), "reg%d, reg%d", dloc->op->reg1, dloc->op->reg2); else if (dloc->op->reg1 == DWARF_REG_PC) - snprintf(buf, sizeof(buf), " (PC)"); + snprintf(buf, sizeof(buf), "PC"); else - buf[0] = '\0'; + snprintf(buf, sizeof(buf), "reg%d", dloc->op->reg1); pr_debug_dtp("-----------------------------------------------------------\n"); - pr_debug_dtp("%s [%"PRIx64"] for reg%d%s offset=%#x in %s\n", - __func__, dloc->ip - dloc->ms->sym->start, - dloc->op->reg1, buf, dloc->op->offset, dloc->ms->sym->name); + pr_debug_dtp("find data type for %#x(%s) at %s+%#"PRIx64"\n", + dloc->op->offset, buf, dloc->ms->sym->name, + dloc->ip - dloc->ms->sym->start); /* * IP is a relative instruction address from the start of the map, as @@ -1468,14 +1523,15 @@ static int find_data_type_die(struct data_loc_info *dloc, Dwarf_Die *type_die) reg = loc->reg1; offset = loc->offset; - pr_debug_dtp("CU die offset: %#lx\n", (long)dwarf_dieoffset(&cu_die)); + pr_debug_dtp("CU for %s (die:%#lx)\n", + dwarf_diename(&cu_die), (long)dwarf_dieoffset(&cu_die)); if (reg == DWARF_REG_PC) { if (get_global_var_type(&cu_die, dloc, dloc->ip, dloc->var_addr, &offset, type_die)) { dloc->type_offset = offset; - pr_debug_dtp("found PC-rel by addr=%#"PRIx64" offset=%#x", + pr_debug_dtp("found by addr=%#"PRIx64" type_offset=%#x\n", dloc->var_addr, offset); pr_debug_type_name(type_die, TSR_KIND_TYPE); ret = 0; @@ -1537,13 +1593,22 @@ retry: pr_debug_dtp("found \"%s\" in scope=%d/%d (die: %#lx) ", dwarf_diename(&var_die), i+1, nr_scopes, (long)dwarf_dieoffset(&scopes[i])); - if (reg == DWARF_REG_PC) - pr_debug_dtp("%#x(PC) offset=%#x", loc->offset, offset); - else if (reg == DWARF_REG_FB || is_fbreg) - pr_debug_dtp("%#x(reg%d) stack fb_offset=%#x offset=%#x", - loc->offset, reg, fb_offset, offset); - else - pr_debug_dtp("%#x(reg%d)", loc->offset, reg); + if (reg == DWARF_REG_PC) { + pr_debug_dtp("addr=%#"PRIx64" type_offset=%#x\n", + dloc->var_addr, offset); + } else if (reg == DWARF_REG_FB || is_fbreg) { + pr_debug_dtp("stack_offset=%#x type_offset=%#x\n", + fb_offset, offset); + } else { + pr_debug_dtp("type_offset=%#x\n", offset); + } + pr_debug_location(&var_die, pc, reg); + pr_debug_type_name(type_die, TSR_KIND_TYPE); + } else { + pr_debug_dtp("check variable \"%s\" failed (die: %#lx)\n", + dwarf_diename(&var_die), + (long)dwarf_dieoffset(&var_die)); + pr_debug_location(&var_die, pc, reg); pr_debug_type_name(type_die, TSR_KIND_TYPE); } dloc->type_offset = offset; -- cgit v1.2.3 From 645af3fb62bf12911ce1fc79efa676dae9a8289b Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 12 Apr 2024 11:33:08 -0700 Subject: perf dwarf-aux: Check pointer offset when checking variables In match_var_offset(), it checks the offset range with the target type only for non-pointer types. But it also needs to check the pointer types with the target type. This is because there can be more than one pointer variable located in the same register. Let's look at the following example. It's looking up a variable for reg3 at tcp_get_info+0x62. It found "sk" variable but it wasn't the right one since it accesses beyond the target type (struct 'sock' in this case) size. ----------------------------------------------------------- find data type for 0x7bc(reg3) at tcp_get_info+0x62 CU for net/ipv4/tcp.c (die:0x7b5f516) frame base: cfa=0 fbreg=6 offset: 1980 is bigger than size: 760 check variable "sk" failed (die: 0x7b92b2c) variable location: reg3 type='struct sock' size=0x2f8 (die:0x7b63c3a) Actually there was another variable "tp" in the function and it's located at the same (reg3) because it's just type-casted like below. void tcp_get_info(struct sock *sk, struct tcp_info *info) { const struct tcp_sock *tp = tcp_sk(sk); ... The 'struct tcp_sock' contains the 'struct sock' at offset 0 so it can just use the same address as a pointer to tcp_sock. That means it should match variables correctly by checking the offset and size. Actually it cannot distinguish if the offset was smaller than the size of the original struct sock. But I think it's fine as they are the same at that part. So let's check the target type size and retry if it doesn't match. Now it succeeded to find the correct variable. ----------------------------------------------------------- find data type for 0x7bc(reg3) at tcp_get_info+0x62 CU for net/ipv4/tcp.c (die:0x7b5f516) frame base: cfa=0 fbreg=6 found "tp" in scope=1/1 (die: 0x7b92b16) type_offset=0x7bc variable location: reg3 type='struct tcp_sock' size=0xa68 (die:0x7b81380) Fixes: bc10db8eb8955fbc ("perf annotate-data: Support stack variables") Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240412183310.2518474-3-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dwarf-aux.c | 28 +++++++++++++++++++++------- 1 file changed, 21 insertions(+), 7 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c index 7dad99ee3ff3..b361fd7ebd56 100644 --- a/tools/perf/util/dwarf-aux.c +++ b/tools/perf/util/dwarf-aux.c @@ -1361,7 +1361,7 @@ struct find_var_data { #define DWARF_OP_DIRECT_REGS 32 static bool match_var_offset(Dwarf_Die *die_mem, struct find_var_data *data, - u64 addr_offset, u64 addr_type) + u64 addr_offset, u64 addr_type, bool is_pointer) { Dwarf_Die type_die; Dwarf_Word size; @@ -1375,6 +1375,12 @@ static bool match_var_offset(Dwarf_Die *die_mem, struct find_var_data *data, if (die_get_real_type(die_mem, &type_die) == NULL) return false; + if (is_pointer && dwarf_tag(&type_die) == DW_TAG_pointer_type) { + /* Get the target type of the pointer */ + if (die_get_real_type(&type_die, &type_die) == NULL) + return false; + } + if (dwarf_aggregate_size(&type_die, &size) < 0) return false; @@ -1442,31 +1448,38 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg) if (data->is_fbreg && ops->atom == DW_OP_fbreg && data->offset >= (int)ops->number && check_allowed_ops(ops, nops) && - match_var_offset(die_mem, data, data->offset, ops->number)) + match_var_offset(die_mem, data, data->offset, ops->number, + /*is_pointer=*/false)) return DIE_FIND_CB_END; /* Only match with a simple case */ if (data->reg < DWARF_OP_DIRECT_REGS) { /* pointer variables saved in a register 0 to 31 */ if (ops->atom == (DW_OP_reg0 + data->reg) && - check_allowed_ops(ops, nops)) + check_allowed_ops(ops, nops) && + match_var_offset(die_mem, data, data->offset, 0, + /*is_pointer=*/true)) return DIE_FIND_CB_END; /* Local variables accessed by a register + offset */ if (ops->atom == (DW_OP_breg0 + data->reg) && check_allowed_ops(ops, nops) && - match_var_offset(die_mem, data, data->offset, ops->number)) + match_var_offset(die_mem, data, data->offset, ops->number, + /*is_pointer=*/false)) return DIE_FIND_CB_END; } else { /* pointer variables saved in a register 32 or above */ if (ops->atom == DW_OP_regx && ops->number == data->reg && - check_allowed_ops(ops, nops)) + check_allowed_ops(ops, nops) && + match_var_offset(die_mem, data, data->offset, 0, + /*is_pointer=*/true)) return DIE_FIND_CB_END; /* Local variables accessed by a register + offset */ if (ops->atom == DW_OP_bregx && data->reg == ops->number && check_allowed_ops(ops, nops) && - match_var_offset(die_mem, data, data->offset, ops->number2)) + match_var_offset(die_mem, data, data->offset, ops->number2, + /*is_poitner=*/false)) return DIE_FIND_CB_END; } } @@ -1528,7 +1541,8 @@ static int __die_find_var_addr_cb(Dwarf_Die *die_mem, void *arg) continue; if (check_allowed_ops(ops, nops) && - match_var_offset(die_mem, data, data->addr, ops->number)) + match_var_offset(die_mem, data, data->addr, ops->number, + /*is_pointer=*/false)) return DIE_FIND_CB_END; } return DIE_FIND_CB_SIBLING; -- cgit v1.2.3 From 0519fadbbe3b1ed396d944911c1ff3a276701474 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 12 Apr 2024 11:33:09 -0700 Subject: perf dwarf-aux: Check variable address range properly In match_var_offset(), it just checked the end address of the variable with the given offset because it assumed the register holds a pointer to the data type and the offset starts from the base. But I found some cases that the stack pointer (rsp = reg7) register is used to pointer a stack variable while the frame base is maintained by a different register (rbp = reg6). In that case, it cannot simply use the stack pointer as it cannot guarantee that it points to the frame base. So it needs to check both boundaries of the variable location. Before: ----------------------------------------------------------- find data type for 0x7c(reg7) at tcp_getsockopt+0xb62 CU for net/ipv4/tcp.c (die:0x7b5f516) frame base: cfa=0 fbreg=6 no pointer or no type check variable "tss" failed (die: 0x7b95801) variable location: base reg7, offset=0x110 type='struct scm_timestamping_internal' size=0x30 (die:0x7b8c126) So the current code just checks register number for the non-PC and non-FB registers and assuming it has offset 0. But this variable has offset 0x110 so it should not match to this. After: ----------------------------------------------------------- find data type for 0x7c(reg7) at tcp_getsockopt+0xb62 CU for net/ipv4/tcp.c (die:0x7b5f516) frame base: cfa=0 fbreg=6 no pointer or no type check variable "zc" failed (die: 0x7b9580a) variable location: base=reg7, offset=0x40 type='struct tcp_zerocopy_receive' size=0x40 (die:7b947f4) Now it find the correct variable "zc". It was located at reg7 + 0x40 and the size if 0x40 which means it should cover [0x40, 0x80). And the access was for reg7 + 0x7c so it found the right one. But it still failed to use the variable and it would be handled in the next patch. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Masami Hiramatsu Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240412183310.2518474-4-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dwarf-aux.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c index b361fd7ebd56..40cfbdfe2d75 100644 --- a/tools/perf/util/dwarf-aux.c +++ b/tools/perf/util/dwarf-aux.c @@ -1372,6 +1372,9 @@ static bool match_var_offset(Dwarf_Die *die_mem, struct find_var_data *data, return true; } + if (addr_offset < addr_type) + return false; + if (die_get_real_type(die_mem, &type_die) == NULL) return false; @@ -1446,7 +1449,6 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg) /* Local variables accessed using frame base register */ if (data->is_fbreg && ops->atom == DW_OP_fbreg && - data->offset >= (int)ops->number && check_allowed_ops(ops, nops) && match_var_offset(die_mem, data, data->offset, ops->number, /*is_pointer=*/false)) @@ -1537,9 +1539,6 @@ static int __die_find_var_addr_cb(Dwarf_Die *die_mem, void *arg) if (ops->atom != DW_OP_addr) continue; - if (data->addr < ops->number) - continue; - if (check_allowed_ops(ops, nops) && match_var_offset(die_mem, data, data->addr, ops->number, /*is_pointer=*/false)) -- cgit v1.2.3 From a5a00497b9dfefbf6872f387bc7692919e1785d3 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 12 Apr 2024 11:33:10 -0700 Subject: perf annotate-data: Handle RSP if it's not the FB register In some cases, the stack pointer on x86 (rsp = reg7) is used to point variables on stack but it's not the frame base register. Then it should handle the register like normal registers (IOW not to access the other stack variables using offset calculation) but it should not assume it would have a pointer. Before: ----------------------------------------------------------- find data type for 0x7c(reg7) at tcp_getsockopt+0xb62 CU for net/ipv4/tcp.c (die:0x7b5f516) frame base: cfa=0 fbreg=6 no pointer or no type check variable "zc" failed (die: 0x7b9580a) variable location: base=reg7, offset=0x40 type='struct tcp_zerocopy_receive' size=0x40 (die:0x7b947f4) After: ----------------------------------------------------------- find data type for 0x7c(reg7) at tcp_getsockopt+0xb62 CU for net/ipv4/tcp.c (die:0x7b5f516) frame base: cfa=0 fbreg=6 found "zc" in scope=3/3 (die: 0x7b957fc) type_offset=0x3c variable location: base=reg7, offset=0x40 type='struct tcp_zerocopy_receive' size=0x40 (die:0x7b947f4) Note that the type-offset was properly calculated to 0x3c as the variable starts at 0x40. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240412183310.2518474-5-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 27 +++++++++++++++++++-------- 1 file changed, 19 insertions(+), 8 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index e53d66c46c54..12d5faff3b7a 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -25,6 +25,9 @@ #include "symbol_conf.h" #include "thread.h" +/* register number of the stack pointer */ +#define X86_REG_SP 7 + enum type_state_kind { TSR_KIND_INVALID = 0, TSR_KIND_TYPE, @@ -197,7 +200,7 @@ static void init_type_state(struct type_state *state, struct arch *arch) state->regs[10].caller_saved = true; state->regs[11].caller_saved = true; state->ret_reg = 0; - state->stack_reg = 7; + state->stack_reg = X86_REG_SP; } } @@ -382,10 +385,18 @@ static bool find_cu_die(struct debuginfo *di, u64 pc, Dwarf_Die *cu_die) } /* The type info will be saved in @type_die */ -static int check_variable(Dwarf_Die *var_die, Dwarf_Die *type_die, int offset, - bool is_pointer) +static int check_variable(struct data_loc_info *dloc, Dwarf_Die *var_die, + Dwarf_Die *type_die, int reg, int offset, bool is_fbreg) { Dwarf_Word size; + bool is_pointer = true; + + if (reg == DWARF_REG_PC) + is_pointer = false; + else if (reg == dloc->fbreg || is_fbreg) + is_pointer = false; + else if (arch__is(dloc->arch, "x86") && reg == X86_REG_SP) + is_pointer = false; /* Get the type of the variable */ if (die_get_real_type(var_die, type_die) == NULL) { @@ -607,7 +618,6 @@ static bool get_global_var_type(Dwarf_Die *cu_die, struct data_loc_info *dloc, { u64 pc; int offset; - bool is_pointer = false; const char *var_name = NULL; struct global_var_entry *gvar; Dwarf_Die var_die; @@ -623,7 +633,8 @@ static bool get_global_var_type(Dwarf_Die *cu_die, struct data_loc_info *dloc, /* Try to get the variable by address first */ if (die_find_variable_by_addr(cu_die, var_addr, &var_die, &offset) && - check_variable(&var_die, type_die, offset, is_pointer) == 0) { + check_variable(dloc, &var_die, type_die, DWARF_REG_PC, offset, + /*is_fbreg=*/false) == 0) { var_name = dwarf_diename(&var_die); *var_offset = offset; goto ok; @@ -636,7 +647,8 @@ static bool get_global_var_type(Dwarf_Die *cu_die, struct data_loc_info *dloc, /* Try to get the name of global variable */ if (die_find_variable_at(cu_die, var_name, pc, &var_die) && - check_variable(&var_die, type_die, *var_offset, is_pointer) == 0) + check_variable(dloc, &var_die, type_die, DWARF_REG_PC, *var_offset, + /*is_fbreg=*/false) == 0) goto ok; return false; @@ -1587,8 +1599,7 @@ retry: } /* Found a variable, see if it's correct */ - ret = check_variable(&var_die, type_die, offset, - reg != DWARF_REG_PC && !is_fbreg); + ret = check_variable(dloc, &var_die, type_die, reg, offset, is_fbreg); if (ret == 0) { pr_debug_dtp("found \"%s\" in scope=%d/%d (die: %#lx) ", dwarf_diename(&var_die), i+1, nr_scopes, -- cgit v1.2.3 From 0993d724674a9d905ebdc9b9cd0b64e30523c5d6 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 11 Apr 2024 11:17:16 -0700 Subject: perf hist: Move histogram related code to hist.h It's strange that sort.h has the definition of struct hist_entry. As sort.h already includes hist.h, let's move the data structure to hist.h. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Andi Kleen Cc: Athira Rajeev Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240411181718.2367948-1-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/hist.h | 191 +++++++++++++++++++++++++++++++++++++++++++++-- tools/perf/util/sort.h | 188 ---------------------------------------------- tools/perf/util/values.h | 1 + 3 files changed, 184 insertions(+), 196 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h index 4a0aea0c9e00..8f072f3749eb 100644 --- a/tools/perf/util/hist.h +++ b/tools/perf/util/hist.h @@ -4,21 +4,22 @@ #include #include -#include "evsel.h" +#include "callchain.h" #include "color.h" #include "events_stats.h" +#include "evsel.h" +#include "map_symbol.h" #include "mutex.h" +#include "sample.h" +#include "spark.h" +#include "stat.h" -struct hist_entry; -struct hist_entry_ops; struct addr_location; -struct map_symbol; struct mem_info; struct kvm_info; struct branch_info; struct branch_stack; struct block_info; -struct symbol; struct ui_progress; enum hist_filter { @@ -150,6 +151,159 @@ extern const struct hist_iter_ops hist_iter_branch; extern const struct hist_iter_ops hist_iter_mem; extern const struct hist_iter_ops hist_iter_cumulative; +struct res_sample { + u64 time; + int cpu; + int tid; +}; + +struct he_stat { + u64 period; + u64 period_sys; + u64 period_us; + u64 period_guest_sys; + u64 period_guest_us; + u32 nr_events; +}; + +struct namespace_id { + u64 dev; + u64 ino; +}; + +struct hist_entry_diff { + bool computed; + union { + /* PERF_HPP__DELTA */ + double period_ratio_delta; + + /* PERF_HPP__RATIO */ + double period_ratio; + + /* HISTC_WEIGHTED_DIFF */ + s64 wdiff; + + /* PERF_HPP_DIFF__CYCLES */ + s64 cycles; + }; + struct stats stats; + unsigned long svals[NUM_SPARKS]; +}; + +struct hist_entry_ops { + void *(*new)(size_t size); + void (*free)(void *ptr); +}; + +/** + * struct hist_entry - histogram entry + * + * @row_offset - offset from the first callchain expanded to appear on screen + * @nr_rows - rows expanded in callchain, recalculated on folding/unfolding + */ +struct hist_entry { + struct rb_node rb_node_in; + struct rb_node rb_node; + union { + struct list_head node; + struct list_head head; + } pairs; + struct he_stat stat; + struct he_stat *stat_acc; + struct map_symbol ms; + struct thread *thread; + struct comm *comm; + struct namespace_id cgroup_id; + u64 cgroup; + u64 ip; + u64 transaction; + s32 socket; + s32 cpu; + u64 code_page_size; + u64 weight; + u64 ins_lat; + u64 p_stage_cyc; + u8 cpumode; + u8 depth; + int mem_type_off; + struct simd_flags simd_flags; + + /* We are added by hists__add_dummy_entry. */ + bool dummy; + bool leaf; + + char level; + u8 filtered; + + u16 callchain_size; + union { + /* + * Since perf diff only supports the stdio output, TUI + * fields are only accessed from perf report (or perf + * top). So make it a union to reduce memory usage. + */ + struct hist_entry_diff diff; + struct /* for TUI */ { + u16 row_offset; + u16 nr_rows; + bool init_have_children; + bool unfolded; + bool has_children; + bool has_no_entry; + }; + }; + char *srcline; + char *srcfile; + struct symbol *parent; + struct branch_info *branch_info; + long time; + struct hists *hists; + struct mem_info *mem_info; + struct block_info *block_info; + struct kvm_info *kvm_info; + void *raw_data; + u32 raw_size; + int num_res; + struct res_sample *res_samples; + void *trace_output; + struct perf_hpp_list *hpp_list; + struct hist_entry *parent_he; + struct hist_entry_ops *ops; + struct annotated_data_type *mem_type; + union { + /* this is for hierarchical entry structure */ + struct { + struct rb_root_cached hroot_in; + struct rb_root_cached hroot_out; + }; /* non-leaf entries */ + struct rb_root sorted_chain; /* leaf entry has callchains */ + }; + struct callchain_root callchain[0]; /* must be last member */ +}; + +static __pure inline bool hist_entry__has_callchains(struct hist_entry *he) +{ + return he->callchain_size != 0; +} + +static inline bool hist_entry__has_pairs(struct hist_entry *he) +{ + return !list_empty(&he->pairs.node); +} + +static inline struct hist_entry *hist_entry__next_pair(struct hist_entry *he) +{ + if (hist_entry__has_pairs(he)) + return list_entry(he->pairs.node.next, struct hist_entry, pairs.node); + return NULL; +} + +static inline void hist_entry__add_pair(struct hist_entry *pair, + struct hist_entry *he) +{ + list_add_tail(&pair->pairs.node, &he->pairs.head); +} + struct hist_entry *hists__add_entry(struct hists *hists, struct addr_location *al, struct symbol *parent, @@ -186,6 +340,8 @@ int hist_entry__sort_snprintf(struct hist_entry *he, char *bf, size_t size, struct hists *hists); int hist_entry__snprintf_alignment(struct hist_entry *he, struct perf_hpp *hpp, struct perf_hpp_fmt *fmt, int printed); +int hist_entry__sym_snprintf(struct hist_entry *he, char *bf, size_t size, + unsigned int width); void hist_entry__delete(struct hist_entry *he); typedef int (*hists__resort_cb_t)(struct hist_entry *he, void *arg); @@ -238,6 +394,20 @@ void hists__match(struct hists *leader, struct hists *other); int hists__link(struct hists *leader, struct hists *other); int hists__unlink(struct hists *hists); +static inline float hist_entry__get_percent_limit(struct hist_entry *he) +{ + u64 period = he->stat.period; + u64 total_period = hists__total_period(he->hists); + + if (unlikely(total_period == 0)) + return 0; + + if (symbol_conf.cumulate_callchain) + period = he->stat_acc->period; + + return period * 100.0 / total_period; +} + struct hists_evsel { struct evsel evsel; struct hists hists; @@ -460,15 +630,20 @@ struct hist_browser_timer { int refresh; }; -struct res_sample; - enum rstype { A_NORMAL, A_ASM, A_SOURCE }; -struct block_hist; +struct block_hist { + struct hists block_hists; + struct perf_hpp_list block_list; + struct perf_hpp_fmt block_fmt; + int block_idx; + bool valid; + struct hist_entry he; +}; #ifdef HAVE_SLANG_SUPPORT #include "../ui/keysyms.h" diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h index 6f6b4189a389..690892a92cf3 100644 --- a/tools/perf/util/sort.h +++ b/tools/perf/util/sort.h @@ -3,19 +3,9 @@ #define __PERF_SORT_H #include #include -#include -#include -#include "map_symbol.h" -#include "symbol_conf.h" -#include "callchain.h" -#include "values.h" #include "hist.h" -#include "stat.h" -#include "spark.h" struct option; -struct thread; -struct annotated_data_type; extern regex_t parent_regex; extern const char *sort_order; @@ -39,175 +29,6 @@ extern struct sort_entry sort_type; extern const char default_mem_sort_order[]; extern bool chk_double_cl; -struct res_sample { - u64 time; - int cpu; - int tid; -}; - -struct he_stat { - u64 period; - u64 period_sys; - u64 period_us; - u64 period_guest_sys; - u64 period_guest_us; - u32 nr_events; -}; - -struct namespace_id { - u64 dev; - u64 ino; -}; - -struct hist_entry_diff { - bool computed; - union { - /* PERF_HPP__DELTA */ - double period_ratio_delta; - - /* PERF_HPP__RATIO */ - double period_ratio; - - /* HISTC_WEIGHTED_DIFF */ - s64 wdiff; - - /* PERF_HPP_DIFF__CYCLES */ - s64 cycles; - }; - struct stats stats; - unsigned long svals[NUM_SPARKS]; -}; - -struct hist_entry_ops { - void *(*new)(size_t size); - void (*free)(void *ptr); -}; - -/** - * struct hist_entry - histogram entry - * - * @row_offset - offset from the first callchain expanded to appear on screen - * @nr_rows - rows expanded in callchain, recalculated on folding/unfolding - */ -struct hist_entry { - struct rb_node rb_node_in; - struct rb_node rb_node; - union { - struct list_head node; - struct list_head head; - } pairs; - struct he_stat stat; - struct he_stat *stat_acc; - struct map_symbol ms; - struct thread *thread; - struct comm *comm; - struct namespace_id cgroup_id; - u64 cgroup; - u64 ip; - u64 transaction; - s32 socket; - s32 cpu; - u64 code_page_size; - u64 weight; - u64 ins_lat; - u64 p_stage_cyc; - u8 cpumode; - u8 depth; - int mem_type_off; - struct simd_flags simd_flags; - - /* We are added by hists__add_dummy_entry. */ - bool dummy; - bool leaf; - - char level; - u8 filtered; - - u16 callchain_size; - union { - /* - * Since perf diff only supports the stdio output, TUI - * fields are only accessed from perf report (or perf - * top). So make it a union to reduce memory usage. - */ - struct hist_entry_diff diff; - struct /* for TUI */ { - u16 row_offset; - u16 nr_rows; - bool init_have_children; - bool unfolded; - bool has_children; - bool has_no_entry; - }; - }; - char *srcline; - char *srcfile; - struct symbol *parent; - struct branch_info *branch_info; - long time; - struct hists *hists; - struct mem_info *mem_info; - struct block_info *block_info; - struct kvm_info *kvm_info; - void *raw_data; - u32 raw_size; - int num_res; - struct res_sample *res_samples; - void *trace_output; - struct perf_hpp_list *hpp_list; - struct hist_entry *parent_he; - struct hist_entry_ops *ops; - struct annotated_data_type *mem_type; - union { - /* this is for hierarchical entry structure */ - struct { - struct rb_root_cached hroot_in; - struct rb_root_cached hroot_out; - }; /* non-leaf entries */ - struct rb_root sorted_chain; /* leaf entry has callchains */ - }; - struct callchain_root callchain[0]; /* must be last member */ -}; - -static __pure inline bool hist_entry__has_callchains(struct hist_entry *he) -{ - return he->callchain_size != 0; -} - -int hist_entry__sym_snprintf(struct hist_entry *he, char *bf, size_t size, unsigned int width); - -static inline bool hist_entry__has_pairs(struct hist_entry *he) -{ - return !list_empty(&he->pairs.node); -} - -static inline struct hist_entry *hist_entry__next_pair(struct hist_entry *he) -{ - if (hist_entry__has_pairs(he)) - return list_entry(he->pairs.node.next, struct hist_entry, pairs.node); - return NULL; -} - -static inline void hist_entry__add_pair(struct hist_entry *pair, - struct hist_entry *he) -{ - list_add_tail(&pair->pairs.node, &he->pairs.head); -} - -static inline float hist_entry__get_percent_limit(struct hist_entry *he) -{ - u64 period = he->stat.period; - u64 total_period = hists__total_period(he->hists); - - if (unlikely(total_period == 0)) - return 0; - - if (symbol_conf.cumulate_callchain) - period = he->stat_acc->period; - - return period * 100.0 / total_period; -} - enum sort_mode { SORT_MODE__NORMAL, SORT_MODE__BRANCH, @@ -299,15 +120,6 @@ struct sort_entry { u8 se_width_idx; }; -struct block_hist { - struct hists block_hists; - struct perf_hpp_list block_list; - struct perf_hpp_fmt block_fmt; - int block_idx; - bool valid; - struct hist_entry he; -}; - extern struct sort_entry sort_thread; struct evlist; diff --git a/tools/perf/util/values.h b/tools/perf/util/values.h index 8c41f22f42cf..791c1ad606c2 100644 --- a/tools/perf/util/values.h +++ b/tools/perf/util/values.h @@ -2,6 +2,7 @@ #ifndef __PERF_VALUES_H #define __PERF_VALUES_H +#include #include struct perf_read_values { -- cgit v1.2.3 From 6fcf1e65253c52a7584817cbd3393c63d5660467 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 11 Apr 2024 11:17:17 -0700 Subject: perf hist: Add weight fields to hist entry stats Like period and sample numbers, it'd be better to track weight values and display them in the output rather than having them as sort keys. This patch just adds a few more fields to save the weights in a hist entry. It'll be displayed as new output fields in the later patch. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Andi Kleen Cc: Athira Rajeev Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240411181718.2367948-2-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/hist.c | 12 ++++++++++-- tools/perf/util/hist.h | 3 +++ 2 files changed, 13 insertions(+), 2 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c index fa359180ebf8..9d43f8ae412d 100644 --- a/tools/perf/util/hist.c +++ b/tools/perf/util/hist.c @@ -308,6 +308,9 @@ static void he_stat__add_stat(struct he_stat *dest, struct he_stat *src) dest->period_us += src->period_us; dest->period_guest_sys += src->period_guest_sys; dest->period_guest_us += src->period_guest_us; + dest->weight1 += src->weight1; + dest->weight2 += src->weight2; + dest->weight3 += src->weight3; dest->nr_events += src->nr_events; } @@ -315,7 +318,9 @@ static void he_stat__decay(struct he_stat *he_stat) { he_stat->period = (he_stat->period * 7) / 8; he_stat->nr_events = (he_stat->nr_events * 7) / 8; - /* XXX need decay for weight too? */ + he_stat->weight1 = (he_stat->weight1 * 7) / 8; + he_stat->weight2 = (he_stat->weight2 * 7) / 8; + he_stat->weight3 = (he_stat->weight3 * 7) / 8; } static void hists__delete_entry(struct hists *hists, struct hist_entry *he); @@ -614,7 +619,7 @@ static struct hist_entry *hists__findnew_entry(struct hists *hists, cmp = hist_entry__cmp(he, entry); if (!cmp) { if (sample_self) { - he_stat__add_period(&he->stat, period); + he_stat__add_stat(&he->stat, &entry->stat); hist_entry__add_callchain_period(he, period); } if (symbol_conf.cumulate_callchain) @@ -731,6 +736,9 @@ __hists__add_entry(struct hists *hists, .stat = { .nr_events = 1, .period = sample->period, + .weight1 = sample->weight, + .weight2 = sample->ins_lat, + .weight3 = sample->p_stage_cyc, }, .parent = sym_parent, .filtered = symbol__parent_filter(sym_parent) | al->filtered, diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h index 8f072f3749eb..f34f101c36c2 100644 --- a/tools/perf/util/hist.h +++ b/tools/perf/util/hist.h @@ -163,6 +163,9 @@ struct he_stat { u64 period_us; u64 period_guest_sys; u64 period_guest_us; + u64 weight1; + u64 weight2; + u64 weight3; u32 nr_events; }; -- cgit v1.2.3 From 7043dc5286a8c082d449e2257f9f87d7f764e7d4 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 11 Apr 2024 11:17:18 -0700 Subject: perf report: Add weight[123] output fields Add weight1, weight2 and weight3 fields to -F/--fields and their aliases like 'ins_lat', 'p_stage_cyc' and 'retire_lat'. Note that they are in the sort keys too but the difference is that output fields will sum up the weight values and display the average. In the sort key, users can see the distribution of weight value and I think it's confusing we have local vs. global weight for the same weight. For example, I experiment with mem-loads events to get the weights. On my laptop, it seems only weight1 field is supported. $ perf mem record -- perf test -w noploop Let's look at the noploop function only. It has 7 samples. $ perf script -F event,ip,sym,weight | grep noploop # event weight ip sym cpu/mem-loads,ldlat=30/P: 43 55b3c122bffc noploop cpu/mem-loads,ldlat=30/P: 48 55b3c122bffc noploop cpu/mem-loads,ldlat=30/P: 38 55b3c122bffc noploop <--- same weight cpu/mem-loads,ldlat=30/P: 38 55b3c122bffc noploop <--- same weight cpu/mem-loads,ldlat=30/P: 59 55b3c122bffc noploop cpu/mem-loads,ldlat=30/P: 33 55b3c122bffc noploop cpu/mem-loads,ldlat=30/P: 38 55b3c122bffc noploop <--- same weight When you use the 'weight' sort key, it'd show entries with a separate weight value separately. Also note that the first entry has 3 samples with weight value 38, so they are displayed together and the weight value is the sum of 3 samples (114 = 38 * 3). $ perf report -n -s +weight | grep -e Weight -e noploop # Overhead Samples Command Shared Object Symbol Weight 0.53% 3 perf perf [.] noploop 114 0.18% 1 perf perf [.] noploop 59 0.18% 1 perf perf [.] noploop 48 0.18% 1 perf perf [.] noploop 43 0.18% 1 perf perf [.] noploop 33 If you use 'local_weight' sort key, you can see the actual weight. $ perf report -n -s +local_weight | grep -e Weight -e noploop # Overhead Samples Command Shared Object Symbol Local Weight 0.53% 3 perf perf [.] noploop 38 0.18% 1 perf perf [.] noploop 59 0.18% 1 perf perf [.] noploop 48 0.18% 1 perf perf [.] noploop 43 0.18% 1 perf perf [.] noploop 33 But when you use the -F/--field option instead, you can see the average weight for the while noploop function (as it won't group samples by weight value and use the default 'comm,dso,sym' sort keys). $ perf report -n -F +weight | grep -e Weight -e noploop Warning: --fields weight shows the average value unlike in the --sort key. # Overhead Samples Weight1 Command Shared Object Symbol 1.23% 7 42.4 perf perf [.] noploop The weight1 field shows the average value: (38 * 3 + 59 + 48 + 43 + 33) / 7 = 42.4 Also it'd show the warning that 'weight' field has the average value. Using 'weight1' can remove the warning. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Andi Kleen Cc: Athira Rajeev Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Cc: Stephane Eranian Link: https://lore.kernel.org/r/20240411181718.2367948-3-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/Documentation/perf-report.txt | 9 +++- tools/perf/ui/hist.c | 92 +++++++++++++++++++++++++------- tools/perf/util/hist.h | 15 +++++- tools/perf/util/sort.c | 28 ++++++---- tools/perf/util/sort.h | 2 +- 5 files changed, 115 insertions(+), 31 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/Documentation/perf-report.txt b/tools/perf/Documentation/perf-report.txt index d8b863e01fe0..d2b1593ef700 100644 --- a/tools/perf/Documentation/perf-report.txt +++ b/tools/perf/Documentation/perf-report.txt @@ -121,6 +121,9 @@ OPTIONS - type: Data type of sample memory access. - typeoff: Offset in the data type of sample memory access. - symoff: Offset in the symbol. + - weight1: Average value of event specific weight (1st field of weight_struct). + - weight2: Average value of event specific weight (2nd field of weight_struct). + - weight3: Average value of event specific weight (3rd field of weight_struct). By default, comm, dso and symbol keys are used. (i.e. --sort comm,dso,symbol) @@ -198,7 +201,11 @@ OPTIONS --fields=:: Specify output field - multiple keys can be specified in CSV format. Following fields are available: - overhead, overhead_sys, overhead_us, overhead_children, sample and period. + overhead, overhead_sys, overhead_us, overhead_children, sample, period, + weight1, weight2, weight3, ins_lat, p_stage_cyc and retire_lat. The + last 3 names are alias for the corresponding weights. When the weight + fields are used, they will show the average value of the weight. + Also it can contain any sort key(s). By default, every sort keys not specified in -F will be appended diff --git a/tools/perf/ui/hist.c b/tools/perf/ui/hist.c index 2bf959d08354..685ba2a54fd8 100644 --- a/tools/perf/ui/hist.c +++ b/tools/perf/ui/hist.c @@ -25,7 +25,7 @@ static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he, hpp_field_fn get_field, const char *fmt, int len, - hpp_snprint_fn print_fn, bool fmt_percent) + hpp_snprint_fn print_fn, enum perf_hpp_fmt_type fmtype) { int ret; struct hists *hists = he->hists; @@ -33,7 +33,7 @@ static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he, char *buf = hpp->buf; size_t size = hpp->size; - if (fmt_percent) { + if (fmtype == PERF_HPP_FMT_TYPE__PERCENT) { double percent = 0.0; u64 total = hists__total_period(hists); @@ -41,8 +41,16 @@ static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he, percent = 100.0 * get_field(he) / total; ret = hpp__call_print_fn(hpp, print_fn, fmt, len, percent); - } else + } else if (fmtype == PERF_HPP_FMT_TYPE__AVERAGE) { + double average = 0; + + if (he->stat.nr_events) + average = 1.0 * get_field(he) / he->stat.nr_events; + + ret = hpp__call_print_fn(hpp, print_fn, fmt, len, average); + } else { ret = hpp__call_print_fn(hpp, print_fn, fmt, len, get_field(he)); + } if (evsel__is_group_event(evsel)) { int prev_idx, idx_delta; @@ -54,6 +62,7 @@ static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he, list_for_each_entry(pair, &he->pairs.head, pairs.node) { u64 period = get_field(pair); u64 total = hists__total_period(pair->hists); + int nr_samples = pair->stat.nr_events; if (!total) continue; @@ -66,7 +75,7 @@ static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he, * zero-fill group members in the middle which * have no sample */ - if (fmt_percent) { + if (fmtype != PERF_HPP_FMT_TYPE__RAW) { ret += hpp__call_print_fn(hpp, print_fn, fmt, len, 0.0); } else { @@ -75,9 +84,14 @@ static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he, } } - if (fmt_percent) { + if (fmtype == PERF_HPP_FMT_TYPE__PERCENT) { ret += hpp__call_print_fn(hpp, print_fn, fmt, len, 100.0 * period / total); + } else if (fmtype == PERF_HPP_FMT_TYPE__AVERAGE) { + double avg = nr_samples ? (period / nr_samples) : 0; + + ret += hpp__call_print_fn(hpp, print_fn, fmt, + len, avg); } else { ret += hpp__call_print_fn(hpp, print_fn, fmt, len, period); @@ -92,7 +106,7 @@ static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he, /* * zero-fill group members at last which have no sample */ - if (fmt_percent) { + if (fmtype != PERF_HPP_FMT_TYPE__RAW) { ret += hpp__call_print_fn(hpp, print_fn, fmt, len, 0.0); } else { @@ -114,33 +128,35 @@ static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he, int hpp__fmt(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp, struct hist_entry *he, hpp_field_fn get_field, - const char *fmtstr, hpp_snprint_fn print_fn, bool fmt_percent) + const char *fmtstr, hpp_snprint_fn print_fn, + enum perf_hpp_fmt_type fmtype) { int len = fmt->user_len ?: fmt->len; if (symbol_conf.field_sep) { return __hpp__fmt(hpp, he, get_field, fmtstr, 1, - print_fn, fmt_percent); + print_fn, fmtype); } - if (fmt_percent) + if (fmtype == PERF_HPP_FMT_TYPE__PERCENT) len -= 2; /* 2 for a space and a % sign */ else len -= 1; - return __hpp__fmt(hpp, he, get_field, fmtstr, len, print_fn, fmt_percent); + return __hpp__fmt(hpp, he, get_field, fmtstr, len, print_fn, fmtype); } int hpp__fmt_acc(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp, struct hist_entry *he, hpp_field_fn get_field, - const char *fmtstr, hpp_snprint_fn print_fn, bool fmt_percent) + const char *fmtstr, hpp_snprint_fn print_fn, + enum perf_hpp_fmt_type fmtype) { if (!symbol_conf.cumulate_callchain) { int len = fmt->user_len ?: fmt->len; return snprintf(hpp->buf, hpp->size, " %*s", len - 1, "N/A"); } - return hpp__fmt(fmt, hpp, he, get_field, fmtstr, print_fn, fmt_percent); + return hpp__fmt(fmt, hpp, he, get_field, fmtstr, print_fn, fmtype); } static int field_cmp(u64 field_a, u64 field_b) @@ -350,7 +366,7 @@ static int hpp__color_##_type(struct perf_hpp_fmt *fmt, \ struct perf_hpp *hpp, struct hist_entry *he) \ { \ return hpp__fmt(fmt, hpp, he, he_get_##_field, " %*.2f%%", \ - hpp_color_scnprintf, true); \ + hpp_color_scnprintf, PERF_HPP_FMT_TYPE__PERCENT); \ } #define __HPP_ENTRY_PERCENT_FN(_type, _field) \ @@ -358,7 +374,7 @@ static int hpp__entry_##_type(struct perf_hpp_fmt *fmt, \ struct perf_hpp *hpp, struct hist_entry *he) \ { \ return hpp__fmt(fmt, hpp, he, he_get_##_field, " %*.2f%%", \ - hpp_entry_scnprintf, true); \ + hpp_entry_scnprintf, PERF_HPP_FMT_TYPE__PERCENT); \ } #define __HPP_SORT_FN(_type, _field) \ @@ -378,7 +394,7 @@ static int hpp__color_##_type(struct perf_hpp_fmt *fmt, \ struct perf_hpp *hpp, struct hist_entry *he) \ { \ return hpp__fmt_acc(fmt, hpp, he, he_get_acc_##_field, " %*.2f%%", \ - hpp_color_scnprintf, true); \ + hpp_color_scnprintf, PERF_HPP_FMT_TYPE__PERCENT); \ } #define __HPP_ENTRY_ACC_PERCENT_FN(_type, _field) \ @@ -386,7 +402,7 @@ static int hpp__entry_##_type(struct perf_hpp_fmt *fmt, \ struct perf_hpp *hpp, struct hist_entry *he) \ { \ return hpp__fmt_acc(fmt, hpp, he, he_get_acc_##_field, " %*.2f%%", \ - hpp_entry_scnprintf, true); \ + hpp_entry_scnprintf, PERF_HPP_FMT_TYPE__PERCENT); \ } #define __HPP_SORT_ACC_FN(_type, _field) \ @@ -406,7 +422,7 @@ static int hpp__entry_##_type(struct perf_hpp_fmt *fmt, \ struct perf_hpp *hpp, struct hist_entry *he) \ { \ return hpp__fmt(fmt, hpp, he, he_get_raw_##_field, " %*"PRIu64, \ - hpp_entry_scnprintf, false); \ + hpp_entry_scnprintf, PERF_HPP_FMT_TYPE__RAW); \ } #define __HPP_SORT_RAW_FN(_type, _field) \ @@ -416,6 +432,26 @@ static int64_t hpp__sort_##_type(struct perf_hpp_fmt *fmt __maybe_unused, \ return __hpp__sort(a, b, he_get_raw_##_field); \ } +#define __HPP_ENTRY_AVERAGE_FN(_type, _field) \ +static u64 he_get_##_field(struct hist_entry *he) \ +{ \ + return he->stat._field; \ +} \ + \ +static int hpp__entry_##_type(struct perf_hpp_fmt *fmt, \ + struct perf_hpp *hpp, struct hist_entry *he) \ +{ \ + return hpp__fmt(fmt, hpp, he, he_get_##_field, " %*.1f", \ + hpp_entry_scnprintf, PERF_HPP_FMT_TYPE__AVERAGE); \ +} + +#define __HPP_SORT_AVERAGE_FN(_type, _field) \ +static int64_t hpp__sort_##_type(struct perf_hpp_fmt *fmt __maybe_unused, \ + struct hist_entry *a, struct hist_entry *b) \ +{ \ + return __hpp__sort(a, b, he_get_##_field); \ +} + #define HPP_PERCENT_FNS(_type, _field) \ __HPP_COLOR_PERCENT_FN(_type, _field) \ @@ -431,6 +467,10 @@ __HPP_SORT_ACC_FN(_type, _field) __HPP_ENTRY_RAW_FN(_type, _field) \ __HPP_SORT_RAW_FN(_type, _field) +#define HPP_AVERAGE_FNS(_type, _field) \ +__HPP_ENTRY_AVERAGE_FN(_type, _field) \ +__HPP_SORT_AVERAGE_FN(_type, _field) + HPP_PERCENT_FNS(overhead, period) HPP_PERCENT_FNS(overhead_sys, period_sys) HPP_PERCENT_FNS(overhead_us, period_us) @@ -441,6 +481,10 @@ HPP_PERCENT_ACC_FNS(overhead_acc, period) HPP_RAW_FNS(samples, nr_events) HPP_RAW_FNS(period, period) +HPP_AVERAGE_FNS(weight1, weight1) +HPP_AVERAGE_FNS(weight2, weight2) +HPP_AVERAGE_FNS(weight3, weight3) + static int64_t hpp__nop_cmp(struct perf_hpp_fmt *fmt __maybe_unused, struct hist_entry *a __maybe_unused, struct hist_entry *b __maybe_unused) @@ -510,7 +554,10 @@ struct perf_hpp_fmt perf_hpp__format[] = { HPP__COLOR_PRINT_FNS("guest usr", overhead_guest_us, OVERHEAD_GUEST_US), HPP__COLOR_ACC_PRINT_FNS("Children", overhead_acc, OVERHEAD_ACC), HPP__PRINT_FNS("Samples", samples, SAMPLES), - HPP__PRINT_FNS("Period", period, PERIOD) + HPP__PRINT_FNS("Period", period, PERIOD), + HPP__PRINT_FNS("Weight1", weight1, WEIGHT1), + HPP__PRINT_FNS("Weight2", weight2, WEIGHT2), + HPP__PRINT_FNS("Weight3", weight3, WEIGHT3), }; struct perf_hpp_list perf_hpp_list = { @@ -526,6 +573,7 @@ struct perf_hpp_list perf_hpp_list = { #undef HPP_PERCENT_FNS #undef HPP_PERCENT_ACC_FNS #undef HPP_RAW_FNS +#undef HPP_AVERAGE_FNS #undef __HPP_HEADER_FN #undef __HPP_WIDTH_FN @@ -534,9 +582,11 @@ struct perf_hpp_list perf_hpp_list = { #undef __HPP_COLOR_ACC_PERCENT_FN #undef __HPP_ENTRY_ACC_PERCENT_FN #undef __HPP_ENTRY_RAW_FN +#undef __HPP_ENTRY_AVERAGE_FN #undef __HPP_SORT_FN #undef __HPP_SORT_ACC_FN #undef __HPP_SORT_RAW_FN +#undef __HPP_SORT_AVERAGE_FN static void fmt_free(struct perf_hpp_fmt *fmt) { @@ -785,6 +835,12 @@ void perf_hpp__reset_width(struct perf_hpp_fmt *fmt, struct hists *hists) fmt->len = 12; break; + case PERF_HPP__WEIGHT1: + case PERF_HPP__WEIGHT2: + case PERF_HPP__WEIGHT3: + fmt->len = 8; + break; + default: break; } diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h index f34f101c36c2..5260822b9773 100644 --- a/tools/perf/util/hist.h +++ b/tools/perf/util/hist.h @@ -550,6 +550,9 @@ enum { PERF_HPP__OVERHEAD_ACC, PERF_HPP__SAMPLES, PERF_HPP__PERIOD, + PERF_HPP__WEIGHT1, + PERF_HPP__WEIGHT2, + PERF_HPP__WEIGHT3, PERF_HPP__MAX_INDEX }; @@ -596,16 +599,24 @@ void perf_hpp__reset_sort_width(struct perf_hpp_fmt *fmt, struct hists *hists); void perf_hpp__set_user_width(const char *width_list_str); void hists__reset_column_width(struct hists *hists); +enum perf_hpp_fmt_type { + PERF_HPP_FMT_TYPE__RAW, + PERF_HPP_FMT_TYPE__PERCENT, + PERF_HPP_FMT_TYPE__AVERAGE, +}; + typedef u64 (*hpp_field_fn)(struct hist_entry *he); typedef int (*hpp_callback_fn)(struct perf_hpp *hpp, bool front); typedef int (*hpp_snprint_fn)(struct perf_hpp *hpp, const char *fmt, ...); int hpp__fmt(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp, struct hist_entry *he, hpp_field_fn get_field, - const char *fmtstr, hpp_snprint_fn print_fn, bool fmt_percent); + const char *fmtstr, hpp_snprint_fn print_fn, + enum perf_hpp_fmt_type fmtype); int hpp__fmt_acc(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp, struct hist_entry *he, hpp_field_fn get_field, - const char *fmtstr, hpp_snprint_fn print_fn, bool fmt_percent); + const char *fmtstr, hpp_snprint_fn print_fn, + enum perf_hpp_fmt_type fmtype); static inline void advance_hpp(struct perf_hpp *hpp, int inc) { diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c index 92a1bd695e8a..add8601c57fd 100644 --- a/tools/perf/util/sort.c +++ b/tools/perf/util/sort.c @@ -2441,6 +2441,13 @@ static struct hpp_dimension hpp_sort_dimensions[] = { DIM(PERF_HPP__OVERHEAD_ACC, "overhead_children"), DIM(PERF_HPP__SAMPLES, "sample"), DIM(PERF_HPP__PERIOD, "period"), + DIM(PERF_HPP__WEIGHT1, "weight1"), + DIM(PERF_HPP__WEIGHT2, "weight2"), + DIM(PERF_HPP__WEIGHT3, "weight3"), + /* aliases for weight_struct */ + DIM(PERF_HPP__WEIGHT2, "ins_lat"), + DIM(PERF_HPP__WEIGHT3, "retire_lat"), + DIM(PERF_HPP__WEIGHT3, "p_stage_cyc"), }; #undef DIM @@ -3743,26 +3750,29 @@ void sort__setup_elide(FILE *output) } } -int output_field_add(struct perf_hpp_list *list, char *tok) +int output_field_add(struct perf_hpp_list *list, const char *tok) { unsigned int i; - for (i = 0; i < ARRAY_SIZE(common_sort_dimensions); i++) { - struct sort_dimension *sd = &common_sort_dimensions[i]; + for (i = 0; i < ARRAY_SIZE(hpp_sort_dimensions); i++) { + struct hpp_dimension *hd = &hpp_sort_dimensions[i]; - if (!sd->name || strncasecmp(tok, sd->name, strlen(tok))) + if (strncasecmp(tok, hd->name, strlen(tok))) continue; - return __sort_dimension__add_output(list, sd); + if (!strcasecmp(tok, "weight")) + ui__warning("--fields weight shows the average value unlike in the --sort key.\n"); + + return __hpp_dimension__add_output(list, hd); } - for (i = 0; i < ARRAY_SIZE(hpp_sort_dimensions); i++) { - struct hpp_dimension *hd = &hpp_sort_dimensions[i]; + for (i = 0; i < ARRAY_SIZE(common_sort_dimensions); i++) { + struct sort_dimension *sd = &common_sort_dimensions[i]; - if (strncasecmp(tok, hd->name, strlen(tok))) + if (!sd->name || strncasecmp(tok, sd->name, strlen(tok))) continue; - return __hpp_dimension__add_output(list, hd); + return __sort_dimension__add_output(list, sd); } for (i = 0; i < ARRAY_SIZE(bstack_sort_dimensions); i++) { diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h index 690892a92cf3..0bd0ee3ae76b 100644 --- a/tools/perf/util/sort.h +++ b/tools/perf/util/sort.h @@ -141,7 +141,7 @@ void reset_dimensions(void); int sort_dimension__add(struct perf_hpp_list *list, const char *tok, struct evlist *evlist, int level); -int output_field_add(struct perf_hpp_list *list, char *tok); +int output_field_add(struct perf_hpp_list *list, const char *tok); int64_t sort__iaddr_cmp(struct hist_entry *left, struct hist_entry *right); int64_t -- cgit v1.2.3 From b828a23a75a012b9242166b78ff1c3ff2189f436 Mon Sep 17 00:00:00 2001 From: Chen Pei Date: Mon, 15 Apr 2024 17:55:32 +0800 Subject: perf genelf: Fix compiling with libelf on rv32 When cross-compiling perf with libelf, the following error occurred: In file included from tests/genelf.c:14: tests/../util/genelf.h:50:2: error: #error "unsupported architecture" 50 | #error "unsupported architecture" | ^~~~~ tests/../util/genelf.h:59:5: warning: "GEN_ELF_CLASS" is not defined, evaluates to 0 [-Wundef] 59 | #if GEN_ELF_CLASS == ELFCLASS64 Fix this by adding GEN-ELF-ARCH and GEN-ELF-CLASS definitions for rv32. Reviewed-by: Ian Rogers Signed-off-by: Chen Pei Cc: Albert Ou Cc: Palmer Dabbelt Cc: Paul Walmsley Link: https://lore.kernel.org/r/20240415095532.4930-1-cp0613@linux.alibaba.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/genelf.h | 3 +++ 1 file changed, 3 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/genelf.h b/tools/perf/util/genelf.h index 5f18d20ea903..4e2e4f40e134 100644 --- a/tools/perf/util/genelf.h +++ b/tools/perf/util/genelf.h @@ -43,6 +43,9 @@ int jit_add_debug_info(Elf *e, uint64_t code_addr, void *debug, int nr_debug_ent #elif defined(__riscv) && __riscv_xlen == 64 #define GEN_ELF_ARCH EM_RISCV #define GEN_ELF_CLASS ELFCLASS64 +#elif defined(__riscv) && __riscv_xlen == 32 +#define GEN_ELF_ARCH EM_RISCV +#define GEN_ELF_CLASS ELFCLASS32 #elif defined(__loongarch__) #define GEN_ELF_ARCH EM_LOONGARCH #define GEN_ELF_CLASS ELFCLASS64 -- cgit v1.2.3 From 03f2357017c37d68e73d7d8d77abfcb72e12bc86 Mon Sep 17 00:00:00 2001 From: Weilin Wang Date: Fri, 12 Apr 2024 14:07:41 -0700 Subject: perf stat: Add new field in stat_config to enable hardware aware grouping Hardware counter and event information could be used to help creating event groups that better utilize hardware counters and improve multiplexing. Reviewed-by: Ian Rogers Signed-off-by: Weilin Wang Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Caleb Biggers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Namhyung Kim Cc: Perry Taylor Cc: Peter Zijlstra Cc: Samantha Alt Link: https://lore.kernel.org/r/20240412210756.309828-2-weilin.wang@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-stat.c | 5 +++++ tools/perf/util/metricgroup.c | 3 +++ tools/perf/util/metricgroup.h | 1 + tools/perf/util/stat.h | 1 + 4 files changed, 10 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index 65388c57bb5d..65a3dd7ffac3 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -2085,6 +2085,7 @@ static int add_default_attributes(void) stat_config.metric_no_threshold, stat_config.user_requested_cpu_list, stat_config.system_wide, + stat_config.hardware_aware_grouping, &stat_config.metric_events); } @@ -2118,6 +2119,7 @@ static int add_default_attributes(void) stat_config.metric_no_threshold, stat_config.user_requested_cpu_list, stat_config.system_wide, + stat_config.hardware_aware_grouping, &stat_config.metric_events); } @@ -2152,6 +2154,7 @@ static int add_default_attributes(void) /*metric_no_threshold=*/true, stat_config.user_requested_cpu_list, stat_config.system_wide, + stat_config.hardware_aware_grouping, &stat_config.metric_events) < 0) return -1; } @@ -2193,6 +2196,7 @@ static int add_default_attributes(void) /*metric_no_threshold=*/true, stat_config.user_requested_cpu_list, stat_config.system_wide, + stat_config.hardware_aware_grouping, &stat_config.metric_events) < 0) return -1; @@ -2727,6 +2731,7 @@ int cmd_stat(int argc, const char **argv) stat_config.metric_no_threshold, stat_config.user_requested_cpu_list, stat_config.system_wide, + stat_config.hardware_aware_grouping, &stat_config.metric_events); zfree(&metrics); diff --git a/tools/perf/util/metricgroup.c b/tools/perf/util/metricgroup.c index 6ec083af14a1..9be406524617 100644 --- a/tools/perf/util/metricgroup.c +++ b/tools/perf/util/metricgroup.c @@ -1690,12 +1690,15 @@ int metricgroup__parse_groups(struct evlist *perf_evlist, bool metric_no_threshold, const char *user_requested_cpu_list, bool system_wide, + bool hardware_aware_grouping, struct rblist *metric_events) { const struct pmu_metrics_table *table = pmu_metrics_table__find(); if (!table) return -EINVAL; + if (hardware_aware_grouping) + pr_debug("Use hardware aware grouping instead of traditional metric grouping method\n"); return parse_groups(perf_evlist, pmu, str, metric_no_group, metric_no_merge, metric_no_threshold, user_requested_cpu_list, system_wide, diff --git a/tools/perf/util/metricgroup.h b/tools/perf/util/metricgroup.h index d5325c6ec8e1..779f6ede1b51 100644 --- a/tools/perf/util/metricgroup.h +++ b/tools/perf/util/metricgroup.h @@ -77,6 +77,7 @@ int metricgroup__parse_groups(struct evlist *perf_evlist, bool metric_no_threshold, const char *user_requested_cpu_list, bool system_wide, + bool hardware_aware_grouping, struct rblist *metric_events); int metricgroup__parse_groups_test(struct evlist *evlist, const struct pmu_metrics_table *table, diff --git a/tools/perf/util/stat.h b/tools/perf/util/stat.h index d6e5c8787ba2..fd7a187551bd 100644 --- a/tools/perf/util/stat.h +++ b/tools/perf/util/stat.h @@ -87,6 +87,7 @@ struct perf_stat_config { bool metric_no_group; bool metric_no_merge; bool metric_no_threshold; + bool hardware_aware_grouping; bool stop_read_counter; bool iostat_run; char *user_requested_cpu_list; -- cgit v1.2.3 From a529bec023d7d14d9fb7d5b456921630e63edc6b Mon Sep 17 00:00:00 2001 From: Dima Kogan Date: Mon, 15 Apr 2024 21:55:10 -0700 Subject: perf probe-event: Un-hardcode sizeof(buf) In several places we had char buf[64]; ... snprintf(buf, 64, ...); This patch changes it to char buf[64]; ... snprintf(buf, sizeof(buf), ...); so the "64" is only stated once. Signed-off-by: Dima Kogan Acked-by: Masami Hiramatsu Link: https://lore.kernel.org/r/20240416045533.162692-2-dima@secretsauce.net Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/probe-event.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index 5c12459e9765..c1dbad8a83dc 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -236,7 +236,7 @@ static int convert_exec_to_group(const char *exec, char **result) } } - ret = e_snprintf(buf, 64, "%s_%s", PERFPROBE_GROUP, ptr1); + ret = e_snprintf(buf, sizeof(buf), "%s_%s", PERFPROBE_GROUP, ptr1); if (ret < 0) goto out; @@ -2867,7 +2867,7 @@ static int probe_trace_event__set_name(struct probe_trace_event *tev, group = PERFPROBE_GROUP; /* Get an unused new event name */ - ret = get_new_event_name(buf, 64, event, namelist, + ret = get_new_event_name(buf, sizeof(buf), event, namelist, tev->point.retprobe, allow_suffix); if (ret < 0) return ret; -- cgit v1.2.3 From c15ed4442981ccf8e7e0d885f5cb500400dde9d7 Mon Sep 17 00:00:00 2001 From: Dima Kogan Date: Mon, 15 Apr 2024 21:55:11 -0700 Subject: perf probe-event: Better error message for a too-long probe name This is a common failure mode when probing userspace C++ code (where the mangling adds significant length to the symbol names). Prior to this patch, only a very generic error message is produced, making the user guess at what the issue is. Signed-off-by: Dima Kogan Acked-by: Masami Hiramatsu Link: https://lore.kernel.org/r/20240416045533.162692-3-dima@secretsauce.net Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/probe-event.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index c1dbad8a83dc..8a73c9464b70 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -2758,7 +2758,7 @@ static int get_new_event_name(char *buf, size_t len, const char *base, /* Try no suffix number */ ret = e_snprintf(buf, len, "%s%s", nbase, ret_event ? "__return" : ""); if (ret < 0) { - pr_debug("snprintf() failed: %d\n", ret); + pr_warning("snprintf() failed: %d; the event name nbase='%s' is too long\n", ret, nbase); goto out; } if (!strlist__has_entry(namelist, buf)) -- cgit v1.2.3 From 8b734eaa98ac3f22bbaa253273fd16f391b3d5f8 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:17 -0700 Subject: perf parse-events: Factor out '/.../' parsing Factor out the case of an event or PMU name followed by a slash based term list. This is with a view to sharing the code with new legacy hardware parsing. Use early return to reduce indentation in the code. Make parse_events_add_pmu static now it doesn't need sharing with parse-events.y. Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 70 +++++++++++++++++++++++++++++++++++++++- tools/perf/util/parse-events.h | 10 +++--- tools/perf/util/parse-events.y | 73 +++--------------------------------------- 3 files changed, 80 insertions(+), 73 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 6f8b0fa17689..a6f71165ee1a 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1385,7 +1385,7 @@ static bool config_term_percore(struct list_head *config_terms) return false; } -int parse_events_add_pmu(struct parse_events_state *parse_state, +static int parse_events_add_pmu(struct parse_events_state *parse_state, struct list_head *list, const char *name, const struct parse_events_terms *const_parsed_terms, bool auto_merge_stats, void *loc_) @@ -1618,6 +1618,74 @@ out_err: return ok ? 0 : -1; } +int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state, + const char *event_or_pmu, + const struct parse_events_terms *const_parsed_terms, + struct list_head **listp, + void *loc_) +{ + char *pattern = NULL; + YYLTYPE *loc = loc_; + struct perf_pmu *pmu = NULL; + int ok = 0; + char *help; + + *listp = malloc(sizeof(**listp)); + if (!*listp) + return -ENOMEM; + + INIT_LIST_HEAD(*listp); + + /* Attempt to add to list assuming event_or_pmu is a PMU name. */ + if (!parse_events_add_pmu(parse_state, *listp, event_or_pmu, const_parsed_terms, + /*auto_merge_stats=*/false, loc)) + return 0; + + /* Failed to add, try wildcard expansion of event_or_pmu as a PMU name. */ + if (asprintf(&pattern, "%s*", event_or_pmu) < 0) { + zfree(listp); + return -ENOMEM; + } + + while ((pmu = perf_pmus__scan(pmu)) != NULL) { + const char *name = pmu->name; + + if (parse_events__filter_pmu(parse_state, pmu)) + continue; + + if (!strncmp(name, "uncore_", 7) && + strncmp(event_or_pmu, "uncore_", 7)) + name += 7; + if (!perf_pmu__match(pattern, name, event_or_pmu) || + !perf_pmu__match(pattern, pmu->alias_name, event_or_pmu)) { + bool auto_merge_stats = perf_pmu__auto_merge_stats(pmu); + + if (!parse_events_add_pmu(parse_state, *listp, pmu->name, + const_parsed_terms, + auto_merge_stats, loc)) { + ok++; + parse_state->wild_card_pmus = true; + } + } + } + zfree(&pattern); + if (ok) + return 0; + + /* Failure to add, assume event_or_pmu is an event name. */ + zfree(listp); + if (!parse_events_multi_pmu_add(parse_state, event_or_pmu, const_parsed_terms, listp, loc)) + return 0; + + if (asprintf(&help, "Unable to find PMU or event on a PMU of '%s'", event_or_pmu) < 0) + help = NULL; + parse_events_error__handle(parse_state->error, loc->first_column, + strdup("Bad event or PMU"), + help); + zfree(listp); + return -EINVAL; +} + int parse_events__modifier_group(struct list_head *list, char *event_mod) { diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index 809359e8544e..a331b9f0da2b 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -209,10 +209,6 @@ int parse_events_add_breakpoint(struct parse_events_state *parse_state, struct list_head *list, u64 addr, char *type, u64 len, struct parse_events_terms *head_config); -int parse_events_add_pmu(struct parse_events_state *parse_state, - struct list_head *list, const char *name, - const struct parse_events_terms *const_parsed_terms, - bool auto_merge_stats, void *loc); struct evsel *parse_events__add_event(int idx, struct perf_event_attr *attr, const char *name, const char *metric_id, @@ -223,6 +219,12 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, const struct parse_events_terms *const_parsed_terms, struct list_head **listp, void *loc); +int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state, + const char *event_or_pmu, + const struct parse_events_terms *const_parsed_terms, + struct list_head **listp, + void *loc_); + void parse_events__set_leader(char *name, struct list_head *list); void parse_events_update_lists(struct list_head *list_event, struct list_head *list_all); diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y index d70f5d84af92..7764e5895210 100644 --- a/tools/perf/util/parse-events.y +++ b/tools/perf/util/parse-events.y @@ -273,78 +273,15 @@ event_def: event_pmu | event_pmu: PE_NAME opt_pmu_config { - struct parse_events_state *parse_state = _parse_state; /* List of created evsels. */ struct list_head *list = NULL; - char *pattern = NULL; - -#define CLEANUP \ - do { \ - parse_events_terms__delete($2); \ - free(list); \ - free($1); \ - free(pattern); \ - } while(0) + int err = parse_events_multi_pmu_add_or_add_pmu(_parse_state, $1, $2, &list, &@1); - list = alloc_list(); - if (!list) { - CLEANUP; - YYNOMEM; - } - /* Attempt to add to list assuming $1 is a PMU name. */ - if (parse_events_add_pmu(parse_state, list, $1, $2, /*auto_merge_stats=*/false, &@1)) { - struct perf_pmu *pmu = NULL; - int ok = 0; - - /* Failure to add, try wildcard expansion of $1 as a PMU name. */ - if (asprintf(&pattern, "%s*", $1) < 0) { - CLEANUP; - YYNOMEM; - } - - while ((pmu = perf_pmus__scan(pmu)) != NULL) { - const char *name = pmu->name; - - if (parse_events__filter_pmu(parse_state, pmu)) - continue; - - if (!strncmp(name, "uncore_", 7) && - strncmp($1, "uncore_", 7)) - name += 7; - if (!perf_pmu__match(pattern, name, $1) || - !perf_pmu__match(pattern, pmu->alias_name, $1)) { - bool auto_merge_stats = perf_pmu__auto_merge_stats(pmu); - - if (!parse_events_add_pmu(parse_state, list, pmu->name, $2, - auto_merge_stats, &@1)) { - ok++; - parse_state->wild_card_pmus = true; - } - } - } - - if (!ok) { - /* Failure to add, assume $1 is an event name. */ - zfree(&list); - ok = !parse_events_multi_pmu_add(parse_state, $1, $2, &list, &@1); - } - if (!ok) { - struct parse_events_error *error = parse_state->error; - char *help; - - if (asprintf(&help, "Unable to find PMU or event on a PMU of '%s'", $1) < 0) - help = NULL; - parse_events_error__handle(error, @1.first_column, - strdup("Bad event or PMU"), - help); - CLEANUP; - YYABORT; - } - } + parse_events_terms__delete($2); + free($1); + if (err) + PE_ABORT(err); $$ = list; - list = NULL; - CLEANUP; -#undef CLEANUP } | PE_NAME sep_dc -- cgit v1.2.3 From 63dfcde9779bcab5cff909819e8c82d472ea117a Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:18 -0700 Subject: perf parse-events: Directly pass PMU to parse_events_add_pmu() Avoid passing the name of a PMU then finding it again, just directly pass the PMU. parse_events_multi_pmu_add_or_add_pmu() is the only version that needs to find a PMU, so move the find there. Remove the error message as parse_events_multi_pmu_add_or_add_pmu will given an error at the end when a name isn't either a PMU name or event name. Without the error message being created the location in the input parameter (loc) can be removed. Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-3-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 46 ++++++++++++++++-------------------------- 1 file changed, 17 insertions(+), 29 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index a6f71165ee1a..2d5a275dd257 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1386,32 +1386,18 @@ static bool config_term_percore(struct list_head *config_terms) } static int parse_events_add_pmu(struct parse_events_state *parse_state, - struct list_head *list, const char *name, - const struct parse_events_terms *const_parsed_terms, - bool auto_merge_stats, void *loc_) + struct list_head *list, struct perf_pmu *pmu, + const struct parse_events_terms *const_parsed_terms, + bool auto_merge_stats) { struct perf_event_attr attr; struct perf_pmu_info info; - struct perf_pmu *pmu; struct evsel *evsel; struct parse_events_error *err = parse_state->error; - YYLTYPE *loc = loc_; LIST_HEAD(config_terms); struct parse_events_terms parsed_terms; bool alias_rewrote_terms = false; - pmu = parse_state->fake_pmu ?: perf_pmus__find(name); - - if (!pmu) { - char *err_str; - - if (asprintf(&err_str, - "Cannot find PMU `%s'. Missing kernel support?", - name) >= 0) - parse_events_error__handle(err, loc->first_column, err_str, NULL); - return -EINVAL; - } - parse_events_terms__init(&parsed_terms); if (const_parsed_terms) { int ret = parse_events_terms__copy(const_parsed_terms, &parsed_terms); @@ -1425,9 +1411,9 @@ static int parse_events_add_pmu(struct parse_events_state *parse_state, strbuf_init(&sb, /*hint=*/ 0); if (pmu->selectable && list_empty(&parsed_terms.terms)) { - strbuf_addf(&sb, "%s//", name); + strbuf_addf(&sb, "%s//", pmu->name); } else { - strbuf_addf(&sb, "%s/", name); + strbuf_addf(&sb, "%s/", pmu->name); parse_events_terms__to_strbuf(&parsed_terms, &sb); strbuf_addch(&sb, '/'); } @@ -1469,7 +1455,7 @@ static int parse_events_add_pmu(struct parse_events_state *parse_state, strbuf_init(&sb, /*hint=*/ 0); parse_events_terms__to_strbuf(&parsed_terms, &sb); - fprintf(stderr, "..after resolving event: %s/%s/\n", name, sb.buf); + fprintf(stderr, "..after resolving event: %s/%s/\n", pmu->name, sb.buf); strbuf_release(&sb); } @@ -1583,8 +1569,8 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, continue; auto_merge_stats = perf_pmu__auto_merge_stats(pmu); - if (!parse_events_add_pmu(parse_state, list, pmu->name, - &parsed_terms, auto_merge_stats, loc)) { + if (!parse_events_add_pmu(parse_state, list, pmu, + &parsed_terms, auto_merge_stats)) { struct strbuf sb; strbuf_init(&sb, /*hint=*/ 0); @@ -1596,8 +1582,8 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, } if (parse_state->fake_pmu) { - if (!parse_events_add_pmu(parse_state, list, event_name, &parsed_terms, - /*auto_merge_stats=*/true, loc)) { + if (!parse_events_add_pmu(parse_state, list, parse_state->fake_pmu, &parsed_terms, + /*auto_merge_stats=*/true)) { struct strbuf sb; strbuf_init(&sb, /*hint=*/ 0); @@ -1626,7 +1612,7 @@ int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state { char *pattern = NULL; YYLTYPE *loc = loc_; - struct perf_pmu *pmu = NULL; + struct perf_pmu *pmu; int ok = 0; char *help; @@ -1637,10 +1623,12 @@ int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state INIT_LIST_HEAD(*listp); /* Attempt to add to list assuming event_or_pmu is a PMU name. */ - if (!parse_events_add_pmu(parse_state, *listp, event_or_pmu, const_parsed_terms, - /*auto_merge_stats=*/false, loc)) + pmu = parse_state->fake_pmu ?: perf_pmus__find(event_or_pmu); + if (pmu && !parse_events_add_pmu(parse_state, *listp, pmu, const_parsed_terms, + /*auto_merge_stats=*/false)) return 0; + pmu = NULL; /* Failed to add, try wildcard expansion of event_or_pmu as a PMU name. */ if (asprintf(&pattern, "%s*", event_or_pmu) < 0) { zfree(listp); @@ -1660,9 +1648,9 @@ int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state !perf_pmu__match(pattern, pmu->alias_name, event_or_pmu)) { bool auto_merge_stats = perf_pmu__auto_merge_stats(pmu); - if (!parse_events_add_pmu(parse_state, *listp, pmu->name, + if (!parse_events_add_pmu(parse_state, *listp, pmu, const_parsed_terms, - auto_merge_stats, loc)) { + auto_merge_stats)) { ok++; parse_state->wild_card_pmus = true; } -- cgit v1.2.3 From 90b2c210a54e657ea8fcba3d724dba180c716edc Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:19 -0700 Subject: perf parse-events: Avoid copying an empty list In parse_events_add_pmu, delay copying the list of terms until it is known the list contains terms. Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-4-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 25 +++++++++++++------------ 1 file changed, 13 insertions(+), 12 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 2d5a275dd257..3b1f767039fa 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1398,29 +1398,21 @@ static int parse_events_add_pmu(struct parse_events_state *parse_state, struct parse_events_terms parsed_terms; bool alias_rewrote_terms = false; - parse_events_terms__init(&parsed_terms); - if (const_parsed_terms) { - int ret = parse_events_terms__copy(const_parsed_terms, &parsed_terms); - - if (ret) - return ret; - } - if (verbose > 1) { struct strbuf sb; strbuf_init(&sb, /*hint=*/ 0); - if (pmu->selectable && list_empty(&parsed_terms.terms)) { + if (pmu->selectable && const_parsed_terms && + list_empty(&const_parsed_terms->terms)) { strbuf_addf(&sb, "%s//", pmu->name); } else { strbuf_addf(&sb, "%s/", pmu->name); - parse_events_terms__to_strbuf(&parsed_terms, &sb); + parse_events_terms__to_strbuf(const_parsed_terms, &sb); strbuf_addch(&sb, '/'); } fprintf(stderr, "Attempt to add: %s\n", sb.buf); strbuf_release(&sb); } - fix_raw(&parsed_terms, pmu); memset(&attr, 0, sizeof(attr)); if (pmu->perf_event_attr_init_default) @@ -1428,7 +1420,7 @@ static int parse_events_add_pmu(struct parse_events_state *parse_state, attr.type = pmu->type; - if (list_empty(&parsed_terms.terms)) { + if (!const_parsed_terms || list_empty(&const_parsed_terms->terms)) { evsel = __add_event(list, &parse_state->idx, &attr, /*init_attr=*/true, /*name=*/NULL, /*metric_id=*/NULL, pmu, @@ -1437,6 +1429,15 @@ static int parse_events_add_pmu(struct parse_events_state *parse_state, return evsel ? 0 : -ENOMEM; } + parse_events_terms__init(&parsed_terms); + if (const_parsed_terms) { + int ret = parse_events_terms__copy(const_parsed_terms, &parsed_terms); + + if (ret) + return ret; + } + fix_raw(&parsed_terms, pmu); + /* Configure attr/terms with a known PMU, this will set hardcoded terms. */ if (config_attr(&attr, &parsed_terms, parse_state->error, config_term_pmu)) { parse_events_terms__exit(&parsed_terms); -- cgit v1.2.3 From f91fa2ae63604e6fe9eed9ac05eb9ed3b23701cc Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:20 -0700 Subject: perf pmu: Refactor perf_pmu__match() Move all implementation to pmu code. Don't allocate a fnmatch wildcard pattern, matching ignoring the suffix already handles this, and only use fnmatch if the given PMU name has a '*' in it. Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-5-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 19 ++----------------- tools/perf/util/pmu.c | 27 +++++++++++++++++++-------- tools/perf/util/pmu.h | 2 +- 3 files changed, 22 insertions(+), 26 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 3b1f767039fa..39548ec645ec 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1611,7 +1611,6 @@ int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state struct list_head **listp, void *loc_) { - char *pattern = NULL; YYLTYPE *loc = loc_; struct perf_pmu *pmu; int ok = 0; @@ -1631,22 +1630,9 @@ int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state pmu = NULL; /* Failed to add, try wildcard expansion of event_or_pmu as a PMU name. */ - if (asprintf(&pattern, "%s*", event_or_pmu) < 0) { - zfree(listp); - return -ENOMEM; - } - while ((pmu = perf_pmus__scan(pmu)) != NULL) { - const char *name = pmu->name; - - if (parse_events__filter_pmu(parse_state, pmu)) - continue; - - if (!strncmp(name, "uncore_", 7) && - strncmp(event_or_pmu, "uncore_", 7)) - name += 7; - if (!perf_pmu__match(pattern, name, event_or_pmu) || - !perf_pmu__match(pattern, pmu->alias_name, event_or_pmu)) { + if (!parse_events__filter_pmu(parse_state, pmu) && + perf_pmu__match(pmu, event_or_pmu)) { bool auto_merge_stats = perf_pmu__auto_merge_stats(pmu); if (!parse_events_add_pmu(parse_state, *listp, pmu, @@ -1657,7 +1643,6 @@ int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state } } } - zfree(&pattern); if (ok) return 0; diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index ab30f22eaf10..74dd5bd49d9a 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -2064,18 +2064,29 @@ void perf_pmu__warn_invalid_config(struct perf_pmu *pmu, __u64 config, name ?: "N/A", buf, config_name, config); } -int perf_pmu__match(const char *pattern, const char *name, const char *tok) +bool perf_pmu__match(const struct perf_pmu *pmu, const char *tok) { - if (!name) - return -1; + const char *name = pmu->name; + bool need_fnmatch = strchr(tok, '*') != NULL; - if (fnmatch(pattern, name, 0)) - return -1; + if (!strncmp(tok, "uncore_", 7)) + tok += 7; + if (!strncmp(name, "uncore_", 7)) + name += 7; - if (tok && !perf_pmu__match_ignoring_suffix(name, tok)) - return -1; + if (perf_pmu__match_ignoring_suffix(name, tok) || + (need_fnmatch && !fnmatch(tok, name, 0))) + return true; - return 0; + name = pmu->alias_name; + if (!name) + return false; + + if (!strncmp(name, "uncore_", 7)) + name += 7; + + return perf_pmu__match_ignoring_suffix(name, tok) || + (need_fnmatch && !fnmatch(tok, name, 0)); } double __weak perf_pmu__cpu_slots_per_cycle(void) diff --git a/tools/perf/util/pmu.h b/tools/perf/util/pmu.h index 152700f78455..93d03bd3ecbe 100644 --- a/tools/perf/util/pmu.h +++ b/tools/perf/util/pmu.h @@ -263,7 +263,7 @@ void perf_pmu__warn_invalid_config(struct perf_pmu *pmu, __u64 config, const char *config_name); void perf_pmu__warn_invalid_formats(struct perf_pmu *pmu); -int perf_pmu__match(const char *pattern, const char *name, const char *tok); +bool perf_pmu__match(const struct perf_pmu *pmu, const char *tok); double perf_pmu__cpu_slots_per_cycle(void); int perf_pmu__event_source_devices_scnprintf(char *pathname, size_t size); -- cgit v1.2.3 From 62593394f66aaebc8bfc0058bb029cae84bd8748 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:22 -0700 Subject: perf parse-events: Legacy cache names on all PMUs and lower priority Prior behavior is to not look for legacy cache names in sysfs/JSON and to create events on all core PMUs. New behavior is to look for sysfs/JSON events first on all PMUs, for core PMUs add a legacy event if the sysfs/JSON event isn't present. This is done so that there is consistency with how event names in terms are handled and their prioritization of sysfs/JSON over legacy. It may make sense to use a legacy cache event name as an event name on a non-core PMU so we should allow it. Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-7-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 39 +++++++++++++++++++++++++++++++-------- tools/perf/util/parse-events.h | 2 +- 2 files changed, 32 insertions(+), 9 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 39548ec645ec..1440a3b4b674 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -442,17 +442,21 @@ bool parse_events__filter_pmu(const struct parse_events_state *parse_state, return strcmp(parse_state->pmu_filter, pmu->name) != 0; } +static int parse_events_add_pmu(struct parse_events_state *parse_state, + struct list_head *list, struct perf_pmu *pmu, + const struct parse_events_terms *const_parsed_terms, + bool auto_merge_stats); + int parse_events_add_cache(struct list_head *list, int *idx, const char *name, struct parse_events_state *parse_state, - struct parse_events_terms *head_config) + struct parse_events_terms *parsed_terms) { struct perf_pmu *pmu = NULL; bool found_supported = false; - const char *config_name = get_config_name(head_config); - const char *metric_id = get_config_metric_id(head_config); + const char *config_name = get_config_name(parsed_terms); + const char *metric_id = get_config_metric_id(parsed_terms); - /* Legacy cache events are only supported by core PMUs. */ - while ((pmu = perf_pmus__scan_core(pmu)) != NULL) { + while ((pmu = perf_pmus__scan(pmu)) != NULL) { LIST_HEAD(config_terms); struct perf_event_attr attr; int ret; @@ -460,6 +464,24 @@ int parse_events_add_cache(struct list_head *list, int *idx, const char *name, if (parse_events__filter_pmu(parse_state, pmu)) continue; + if (perf_pmu__have_event(pmu, name)) { + /* + * The PMU has the event so add as not a legacy cache + * event. + */ + ret = parse_events_add_pmu(parse_state, list, pmu, + parsed_terms, + perf_pmu__auto_merge_stats(pmu)); + if (ret) + return ret; + continue; + } + + if (!pmu->is_core) { + /* Legacy cache events are only supported by core PMUs. */ + continue; + } + memset(&attr, 0, sizeof(attr)); attr.type = PERF_TYPE_HW_CACHE; @@ -469,11 +491,12 @@ int parse_events_add_cache(struct list_head *list, int *idx, const char *name, found_supported = true; - if (head_config) { - if (config_attr(&attr, head_config, parse_state->error, config_term_common)) + if (parsed_terms) { + if (config_attr(&attr, parsed_terms, parse_state->error, + config_term_common)) return -EINVAL; - if (get_config_terms(head_config, &config_terms)) + if (get_config_terms(parsed_terms, &config_terms)) return -ENOMEM; } diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index a331b9f0da2b..db47913e54bc 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -203,7 +203,7 @@ int parse_events_add_tool(struct parse_events_state *parse_state, int tool_event); int parse_events_add_cache(struct list_head *list, int *idx, const char *name, struct parse_events_state *parse_state, - struct parse_events_terms *head_config); + struct parse_events_terms *parsed_terms); int parse_events__decode_legacy_cache(const char *name, int pmu_type, __u64 *config); int parse_events_add_breakpoint(struct parse_events_state *parse_state, struct list_head *list, -- cgit v1.2.3 From 9d0dba2398ff4bdb2eced7eaa7abd927aadf442b Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:23 -0700 Subject: perf parse-events: Handle PE_TERM_HW in name_or_raw Avoid duplicate logic for name_or_raw and PE_TERM_HW by having a rule to turn PE_TERM_HW into a name_or_raw. Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-8-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.y | 31 +++++-------------------------- 1 file changed, 5 insertions(+), 26 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y index 7764e5895210..254f8aeca461 100644 --- a/tools/perf/util/parse-events.y +++ b/tools/perf/util/parse-events.y @@ -603,6 +603,11 @@ event_term } name_or_raw: PE_RAW | PE_NAME | PE_LEGACY_CACHE +| +PE_TERM_HW +{ + $$ = $1.str; +} event_term: PE_RAW @@ -644,20 +649,6 @@ name_or_raw '=' PE_VALUE $$ = term; } | -name_or_raw '=' PE_TERM_HW -{ - struct parse_events_term *term; - int err = parse_events_term__str(&term, PARSE_EVENTS__TERM_TYPE_USER, - $1, $3.str, &@1, &@3); - - if (err) { - free($1); - free($3.str); - PE_ABORT(err); - } - $$ = term; -} -| PE_LEGACY_CACHE { struct parse_events_term *term; @@ -710,18 +701,6 @@ PE_TERM '=' name_or_raw $$ = term; } | -PE_TERM '=' PE_TERM_HW -{ - struct parse_events_term *term; - int err = parse_events_term__str(&term, $1, /*config=*/NULL, $3.str, &@1, &@3); - - if (err) { - free($3.str); - PE_ABORT(err); - } - $$ = term; -} -| PE_TERM '=' PE_TERM { struct parse_events_term *term; -- cgit v1.2.3 From 5ccc4edfc2a98840d8fb1669963df42d2d433181 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:24 -0700 Subject: perf parse-events: Constify parse_events_add_numeric Allow the term list to be const so that other functions can pass const term lists. Add const as necessary to called functions. Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-9-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 20 +++++++++++--------- tools/perf/util/parse-events.h | 2 +- 2 files changed, 12 insertions(+), 10 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 1440a3b4b674..1b408e3dccc7 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -34,7 +34,8 @@ #ifdef PARSER_DEBUG extern int parse_events_debug; #endif -static int get_config_terms(struct parse_events_terms *head_config, struct list_head *head_terms); +static int get_config_terms(const struct parse_events_terms *head_config, + struct list_head *head_terms); static int parse_events_terms__copy(const struct parse_events_terms *src, struct parse_events_terms *dest); @@ -154,7 +155,7 @@ const char *event_type(int type) return "unknown"; } -static char *get_config_str(struct parse_events_terms *head_terms, +static char *get_config_str(const struct parse_events_terms *head_terms, enum parse_events__term_type type_term) { struct parse_events_term *term; @@ -169,12 +170,12 @@ static char *get_config_str(struct parse_events_terms *head_terms, return NULL; } -static char *get_config_metric_id(struct parse_events_terms *head_terms) +static char *get_config_metric_id(const struct parse_events_terms *head_terms) { return get_config_str(head_terms, PARSE_EVENTS__TERM_TYPE_METRIC_ID); } -static char *get_config_name(struct parse_events_terms *head_terms) +static char *get_config_name(const struct parse_events_terms *head_terms) { return get_config_str(head_terms, PARSE_EVENTS__TERM_TYPE_NAME); } @@ -358,7 +359,7 @@ static int config_term_common(struct perf_event_attr *attr, struct parse_events_term *term, struct parse_events_error *err); static int config_attr(struct perf_event_attr *attr, - struct parse_events_terms *head, + const struct parse_events_terms *head, struct parse_events_error *err, config_term_func_t config_term); @@ -1108,7 +1109,7 @@ static int config_term_tracepoint(struct perf_event_attr *attr, #endif static int config_attr(struct perf_event_attr *attr, - struct parse_events_terms *head, + const struct parse_events_terms *head, struct parse_events_error *err, config_term_func_t config_term) { @@ -1121,7 +1122,8 @@ static int config_attr(struct perf_event_attr *attr, return 0; } -static int get_config_terms(struct parse_events_terms *head_config, struct list_head *head_terms) +static int get_config_terms(const struct parse_events_terms *head_config, + struct list_head *head_terms) { #define ADD_CONFIG_TERM(__type, __weak) \ struct evsel_config_term *__t; \ @@ -1325,7 +1327,7 @@ int parse_events_add_tracepoint(struct list_head *list, int *idx, static int __parse_events_add_numeric(struct parse_events_state *parse_state, struct list_head *list, struct perf_pmu *pmu, u32 type, u32 extended_type, - u64 config, struct parse_events_terms *head_config) + u64 config, const struct parse_events_terms *head_config) { struct perf_event_attr attr; LIST_HEAD(config_terms); @@ -1361,7 +1363,7 @@ static int __parse_events_add_numeric(struct parse_events_state *parse_state, int parse_events_add_numeric(struct parse_events_state *parse_state, struct list_head *list, u32 type, u64 config, - struct parse_events_terms *head_config, + const struct parse_events_terms *head_config, bool wildcard) { struct perf_pmu *pmu = NULL; diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index db47913e54bc..5005782766e9 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -196,7 +196,7 @@ int parse_events_add_tracepoint(struct list_head *list, int *idx, int parse_events_add_numeric(struct parse_events_state *parse_state, struct list_head *list, u32 type, u64 config, - struct parse_events_terms *head_config, + const struct parse_events_terms *head_config, bool wildcard); int parse_events_add_tool(struct parse_events_state *parse_state, struct list_head *list, -- cgit v1.2.3 From 617824a7f0f73e4de325cf8add58e55b28c12493 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:25 -0700 Subject: perf parse-events: Prefer sysfs/JSON hardware events over legacy It was requested that RISC-V be able to add events to the perf tool so the PMU driver didn't need to map legacy events to config encodings: https://lore.kernel.org/lkml/20240217005738.3744121-1-atishp@rivosinc.com/ This change makes the priority of events specified without a PMU the same as those specified with a PMU, namely sysfs and JSON events are checked first before using the legacy encoding. The hw_term is made more generic as a hardware_event that encodes a pair of string and int value, allowing parse_events_multi_pmu_add to fall back on a known encoding when the sysfs/JSON adding fails for core events. As this covers PE_VALUE_SYM_HW, that token is removed and related code simplified. Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-10-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 31 ++++++++++++----- tools/perf/util/parse-events.h | 2 +- tools/perf/util/parse-events.l | 76 +++++++++++++++++++++--------------------- tools/perf/util/parse-events.y | 62 ++++++++++++++++++++++------------ 4 files changed, 103 insertions(+), 68 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 1b408e3dccc7..805872c90a3e 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1543,7 +1543,7 @@ static int parse_events_add_pmu(struct parse_events_state *parse_state, } int parse_events_multi_pmu_add(struct parse_events_state *parse_state, - const char *event_name, + const char *event_name, u64 hw_config, const struct parse_events_terms *const_parsed_terms, struct list_head **listp, void *loc_) { @@ -1551,8 +1551,8 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, struct list_head *list = NULL; struct perf_pmu *pmu = NULL; YYLTYPE *loc = loc_; - int ok = 0; - const char *config; + int ok = 0, core_ok = 0; + const char *tmp; struct parse_events_terms parsed_terms; *listp = NULL; @@ -1565,15 +1565,15 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, return ret; } - config = strdup(event_name); - if (!config) + tmp = strdup(event_name); + if (!tmp) goto out_err; if (parse_events_term__num(&term, PARSE_EVENTS__TERM_TYPE_USER, - config, /*num=*/1, /*novalue=*/true, + tmp, /*num=*/1, /*novalue=*/true, loc, /*loc_val=*/NULL) < 0) { - zfree(&config); + zfree(&tmp); goto out_err; } list_add_tail(&term->list, &parsed_terms.terms); @@ -1604,6 +1604,8 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, pr_debug("%s -> %s/%s/\n", event_name, pmu->name, sb.buf); strbuf_release(&sb); ok++; + if (pmu->is_core) + core_ok++; } } @@ -1620,6 +1622,18 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, } } + if (hw_config != PERF_COUNT_HW_MAX && !core_ok) { + /* + * The event wasn't found on core PMUs but it has a hardware + * config version to try. + */ + if (!parse_events_add_numeric(parse_state, list, + PERF_TYPE_HARDWARE, hw_config, + const_parsed_terms, + /*wildcard=*/true)) + ok++; + } + out_err: parse_events_terms__exit(&parsed_terms); if (ok) @@ -1673,7 +1687,8 @@ int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state /* Failure to add, assume event_or_pmu is an event name. */ zfree(listp); - if (!parse_events_multi_pmu_add(parse_state, event_or_pmu, const_parsed_terms, listp, loc)) + if (!parse_events_multi_pmu_add(parse_state, event_or_pmu, PERF_COUNT_HW_MAX, + const_parsed_terms, listp, loc)) return 0; if (asprintf(&help, "Unable to find PMU or event on a PMU of '%s'", event_or_pmu) < 0) diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index 5005782766e9..7e5afad3feb8 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -215,7 +215,7 @@ struct evsel *parse_events__add_event(int idx, struct perf_event_attr *attr, struct perf_pmu *pmu); int parse_events_multi_pmu_add(struct parse_events_state *parse_state, - const char *event_name, + const char *event_name, u64 hw_config, const struct parse_events_terms *const_parsed_terms, struct list_head **listp, void *loc); diff --git a/tools/perf/util/parse-events.l b/tools/perf/util/parse-events.l index e86c45675e1d..6fe37003ab7b 100644 --- a/tools/perf/util/parse-events.l +++ b/tools/perf/util/parse-events.l @@ -100,12 +100,12 @@ do { \ yyless(0); \ } while (0) -static int sym(yyscan_t scanner, int type, int config) +static int sym(yyscan_t scanner, int config) { YYSTYPE *yylval = parse_events_get_lval(scanner); - yylval->num = (type << 16) + config; - return type == PERF_TYPE_HARDWARE ? PE_VALUE_SYM_HW : PE_VALUE_SYM_SW; + yylval->num = config; + return PE_VALUE_SYM_SW; } static int tool(yyscan_t scanner, enum perf_tool_event event) @@ -124,13 +124,13 @@ static int term(yyscan_t scanner, enum parse_events__term_type type) return PE_TERM; } -static int hw_term(yyscan_t scanner, int config) +static int hw(yyscan_t scanner, int config) { YYSTYPE *yylval = parse_events_get_lval(scanner); char *text = parse_events_get_text(scanner); - yylval->hardware_term.str = strdup(text); - yylval->hardware_term.num = PERF_TYPE_HARDWARE + config; + yylval->hardware_event.str = strdup(text); + yylval->hardware_event.num = config; return PE_TERM_HW; } @@ -246,16 +246,16 @@ percore { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_PERCORE); } aux-output { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_AUX_OUTPUT); } aux-sample-size { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_AUX_SAMPLE_SIZE); } metric-id { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_METRIC_ID); } -cpu-cycles|cycles { return hw_term(yyscanner, PERF_COUNT_HW_CPU_CYCLES); } -stalled-cycles-frontend|idle-cycles-frontend { return hw_term(yyscanner, PERF_COUNT_HW_STALLED_CYCLES_FRONTEND); } -stalled-cycles-backend|idle-cycles-backend { return hw_term(yyscanner, PERF_COUNT_HW_STALLED_CYCLES_BACKEND); } -instructions { return hw_term(yyscanner, PERF_COUNT_HW_INSTRUCTIONS); } -cache-references { return hw_term(yyscanner, PERF_COUNT_HW_CACHE_REFERENCES); } -cache-misses { return hw_term(yyscanner, PERF_COUNT_HW_CACHE_MISSES); } -branch-instructions|branches { return hw_term(yyscanner, PERF_COUNT_HW_BRANCH_INSTRUCTIONS); } -branch-misses { return hw_term(yyscanner, PERF_COUNT_HW_BRANCH_MISSES); } -bus-cycles { return hw_term(yyscanner, PERF_COUNT_HW_BUS_CYCLES); } -ref-cycles { return hw_term(yyscanner, PERF_COUNT_HW_REF_CPU_CYCLES); } +cpu-cycles|cycles { return hw(yyscanner, PERF_COUNT_HW_CPU_CYCLES); } +stalled-cycles-frontend|idle-cycles-frontend { return hw(yyscanner, PERF_COUNT_HW_STALLED_CYCLES_FRONTEND); } +stalled-cycles-backend|idle-cycles-backend { return hw(yyscanner, PERF_COUNT_HW_STALLED_CYCLES_BACKEND); } +instructions { return hw(yyscanner, PERF_COUNT_HW_INSTRUCTIONS); } +cache-references { return hw(yyscanner, PERF_COUNT_HW_CACHE_REFERENCES); } +cache-misses { return hw(yyscanner, PERF_COUNT_HW_CACHE_MISSES); } +branch-instructions|branches { return hw(yyscanner, PERF_COUNT_HW_BRANCH_INSTRUCTIONS); } +branch-misses { return hw(yyscanner, PERF_COUNT_HW_BRANCH_MISSES); } +bus-cycles { return hw(yyscanner, PERF_COUNT_HW_BUS_CYCLES); } +ref-cycles { return hw(yyscanner, PERF_COUNT_HW_REF_CPU_CYCLES); } r{num_raw_hex} { return str(yyscanner, PE_RAW); } r0x{num_raw_hex} { return str(yyscanner, PE_RAW); } , { return ','; } @@ -299,31 +299,31 @@ r0x{num_raw_hex} { return str(yyscanner, PE_RAW); } <> { BEGIN(INITIAL); } } -cpu-cycles|cycles { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_CPU_CYCLES); } -stalled-cycles-frontend|idle-cycles-frontend { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_STALLED_CYCLES_FRONTEND); } -stalled-cycles-backend|idle-cycles-backend { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_STALLED_CYCLES_BACKEND); } -instructions { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_INSTRUCTIONS); } -cache-references { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_CACHE_REFERENCES); } -cache-misses { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_CACHE_MISSES); } -branch-instructions|branches { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_BRANCH_INSTRUCTIONS); } -branch-misses { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_BRANCH_MISSES); } -bus-cycles { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_BUS_CYCLES); } -ref-cycles { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_REF_CPU_CYCLES); } -cpu-clock { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_CPU_CLOCK); } -task-clock { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_TASK_CLOCK); } -page-faults|faults { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_PAGE_FAULTS); } -minor-faults { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_PAGE_FAULTS_MIN); } -major-faults { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_PAGE_FAULTS_MAJ); } -context-switches|cs { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_CONTEXT_SWITCHES); } -cpu-migrations|migrations { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_CPU_MIGRATIONS); } -alignment-faults { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_ALIGNMENT_FAULTS); } -emulation-faults { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_EMULATION_FAULTS); } -dummy { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_DUMMY); } +cpu-cycles|cycles { return hw(yyscanner, PERF_COUNT_HW_CPU_CYCLES); } +stalled-cycles-frontend|idle-cycles-frontend { return hw(yyscanner, PERF_COUNT_HW_STALLED_CYCLES_FRONTEND); } +stalled-cycles-backend|idle-cycles-backend { return hw(yyscanner, PERF_COUNT_HW_STALLED_CYCLES_BACKEND); } +instructions { return hw(yyscanner, PERF_COUNT_HW_INSTRUCTIONS); } +cache-references { return hw(yyscanner, PERF_COUNT_HW_CACHE_REFERENCES); } +cache-misses { return hw(yyscanner, PERF_COUNT_HW_CACHE_MISSES); } +branch-instructions|branches { return hw(yyscanner, PERF_COUNT_HW_BRANCH_INSTRUCTIONS); } +branch-misses { return hw(yyscanner, PERF_COUNT_HW_BRANCH_MISSES); } +bus-cycles { return hw(yyscanner, PERF_COUNT_HW_BUS_CYCLES); } +ref-cycles { return hw(yyscanner, PERF_COUNT_HW_REF_CPU_CYCLES); } +cpu-clock { return sym(yyscanner, PERF_COUNT_SW_CPU_CLOCK); } +task-clock { return sym(yyscanner, PERF_COUNT_SW_TASK_CLOCK); } +page-faults|faults { return sym(yyscanner, PERF_COUNT_SW_PAGE_FAULTS); } +minor-faults { return sym(yyscanner, PERF_COUNT_SW_PAGE_FAULTS_MIN); } +major-faults { return sym(yyscanner, PERF_COUNT_SW_PAGE_FAULTS_MAJ); } +context-switches|cs { return sym(yyscanner, PERF_COUNT_SW_CONTEXT_SWITCHES); } +cpu-migrations|migrations { return sym(yyscanner, PERF_COUNT_SW_CPU_MIGRATIONS); } +alignment-faults { return sym(yyscanner, PERF_COUNT_SW_ALIGNMENT_FAULTS); } +emulation-faults { return sym(yyscanner, PERF_COUNT_SW_EMULATION_FAULTS); } +dummy { return sym(yyscanner, PERF_COUNT_SW_DUMMY); } duration_time { return tool(yyscanner, PERF_TOOL_DURATION_TIME); } user_time { return tool(yyscanner, PERF_TOOL_USER_TIME); } system_time { return tool(yyscanner, PERF_TOOL_SYSTEM_TIME); } -bpf-output { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_BPF_OUTPUT); } -cgroup-switches { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_CGROUP_SWITCHES); } +bpf-output { return sym(yyscanner, PERF_COUNT_SW_BPF_OUTPUT); } +cgroup-switches { return sym(yyscanner, PERF_COUNT_SW_CGROUP_SWITCHES); } {lc_type} { return str(yyscanner, PE_LEGACY_CACHE); } {lc_type}-{lc_op_result} { return str(yyscanner, PE_LEGACY_CACHE); } diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y index 254f8aeca461..31fe8cf428ff 100644 --- a/tools/perf/util/parse-events.y +++ b/tools/perf/util/parse-events.y @@ -55,7 +55,7 @@ static void free_list_evsel(struct list_head* list_evsel) %} %token PE_START_EVENTS PE_START_TERMS -%token PE_VALUE PE_VALUE_SYM_HW PE_VALUE_SYM_SW PE_TERM +%token PE_VALUE PE_VALUE_SYM_SW PE_TERM %token PE_VALUE_SYM_TOOL %token PE_EVENT_NAME %token PE_RAW PE_NAME @@ -66,11 +66,9 @@ static void free_list_evsel(struct list_head* list_evsel) %token PE_DRV_CFG_TERM %token PE_TERM_HW %type PE_VALUE -%type PE_VALUE_SYM_HW %type PE_VALUE_SYM_SW %type PE_VALUE_SYM_TOOL %type PE_TERM -%type value_sym %type PE_RAW %type PE_NAME %type PE_LEGACY_CACHE @@ -87,6 +85,7 @@ static void free_list_evsel(struct list_head* list_evsel) %type opt_pmu_config %destructor { parse_events_terms__delete ($$); } %type event_pmu +%type event_legacy_hardware %type event_legacy_symbol %type event_legacy_cache %type event_legacy_mem @@ -104,8 +103,8 @@ static void free_list_evsel(struct list_head* list_evsel) %destructor { free_list_evsel ($$); } %type tracepoint_name %destructor { free ($$.sys); free ($$.event); } -%type PE_TERM_HW -%destructor { free ($$.str); } +%type PE_TERM_HW +%destructor { free ($$.str); } %union { @@ -119,10 +118,10 @@ static void free_list_evsel(struct list_head* list_evsel) char *sys; char *event; } tracepoint_name; - struct hardware_term { + struct hardware_event { char *str; u64 num; - } hardware_term; + } hardware_event; } %% @@ -263,6 +262,7 @@ PE_EVENT_NAME event_def event_def event_def: event_pmu | + event_legacy_hardware | event_legacy_symbol | event_legacy_cache sep_dc | event_legacy_mem sep_dc | @@ -289,7 +289,7 @@ PE_NAME sep_dc struct list_head *list; int err; - err = parse_events_multi_pmu_add(_parse_state, $1, NULL, &list, &@1); + err = parse_events_multi_pmu_add(_parse_state, $1, PERF_COUNT_HW_MAX, NULL, &list, &@1); if (err < 0) { struct parse_events_state *parse_state = _parse_state; struct parse_events_error *error = parse_state->error; @@ -305,24 +305,45 @@ PE_NAME sep_dc $$ = list; } -value_sym: -PE_VALUE_SYM_HW +event_legacy_hardware: +PE_TERM_HW opt_pmu_config +{ + /* List of created evsels. */ + struct list_head *list = NULL; + int err = parse_events_multi_pmu_add(_parse_state, $1.str, $1.num, $2, &list, &@1); + + free($1.str); + parse_events_terms__delete($2); + if (err) + PE_ABORT(err); + + $$ = list; +} | -PE_VALUE_SYM_SW +PE_TERM_HW sep_dc +{ + struct list_head *list; + int err; + + err = parse_events_multi_pmu_add(_parse_state, $1.str, $1.num, NULL, &list, &@1); + free($1.str); + if (err) + PE_ABORT(err); + $$ = list; +} event_legacy_symbol: -value_sym '/' event_config '/' +PE_VALUE_SYM_SW '/' event_config '/' { struct list_head *list; - int type = $1 >> 16; - int config = $1 & 255; int err; - bool wildcard = (type == PERF_TYPE_HARDWARE || type == PERF_TYPE_HW_CACHE); list = alloc_list(); if (!list) YYNOMEM; - err = parse_events_add_numeric(_parse_state, list, type, config, $3, wildcard); + err = parse_events_add_numeric(_parse_state, list, + /*type=*/PERF_TYPE_SOFTWARE, /*config=*/$1, + $3, /*wildcard=*/false); parse_events_terms__delete($3); if (err) { free_list_evsel(list); @@ -331,18 +352,17 @@ value_sym '/' event_config '/' $$ = list; } | -value_sym sep_slash_slash_dc +PE_VALUE_SYM_SW sep_slash_slash_dc { struct list_head *list; - int type = $1 >> 16; - int config = $1 & 255; - bool wildcard = (type == PERF_TYPE_HARDWARE || type == PERF_TYPE_HW_CACHE); int err; list = alloc_list(); if (!list) YYNOMEM; - err = parse_events_add_numeric(_parse_state, list, type, config, /*head_config=*/NULL, wildcard); + err = parse_events_add_numeric(_parse_state, list, + /*type=*/PERF_TYPE_SOFTWARE, /*config=*/$1, + /*head_config=*/NULL, /*wildcard=*/false); if (err) PE_ABORT(err); $$ = list; -- cgit v1.2.3 From 4e5484b4bfd53d4d12933ab9b19da66909f1598e Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:26 -0700 Subject: perf parse-events: Inline parse_events_update_lists The helper function just wraps a splice and free. Making the free inline removes a comment, so then it just wraps a splice which we can make inline too. Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-11-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 13 ------------- tools/perf/util/parse-events.h | 2 -- tools/perf/util/parse-events.y | 41 +++++++++++++++++++++++++---------------- 3 files changed, 25 insertions(+), 31 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 805872c90a3e..7eba714f0d73 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1720,19 +1720,6 @@ void parse_events__set_leader(char *name, struct list_head *list) leader->group_name = name; } -/* list_event is assumed to point to malloc'ed memory */ -void parse_events_update_lists(struct list_head *list_event, - struct list_head *list_all) -{ - /* - * Called for single event definition. Update the - * 'all event' list, and reinit the 'single event' - * list, for next event definition. - */ - list_splice_tail(list_event, list_all); - free(list_event); -} - struct event_modifier { int eu; int ek; diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index 7e5afad3feb8..e8f2aebea10f 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -226,8 +226,6 @@ int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state void *loc_); void parse_events__set_leader(char *name, struct list_head *list); -void parse_events_update_lists(struct list_head *list_event, - struct list_head *list_all); void parse_events_evlist_error(struct parse_events_state *parse_state, int idx, const char *str); diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y index 31fe8cf428ff..51490f0f8c50 100644 --- a/tools/perf/util/parse-events.y +++ b/tools/perf/util/parse-events.y @@ -125,6 +125,10 @@ static void free_list_evsel(struct list_head* list_evsel) } %% + /* + * Entry points. We are either parsing events or terminals. Just terminal + * parsing is used for parsing events in sysfs. + */ start: PE_START_EVENTS start_events | @@ -132,31 +136,36 @@ PE_START_TERMS start_terms start_events: groups { + /* Take the parsed events, groups.. and place into parse_state. */ + struct list_head *groups = $1; struct parse_events_state *parse_state = _parse_state; - /* frees $1 */ - parse_events_update_lists($1, &parse_state->list); + list_splice_tail(groups, &parse_state->list); + free(groups); } -groups: +groups: /* A list of groups or events. */ groups ',' group { - struct list_head *list = $1; - struct list_head *group = $3; + /* Merge group into the list of events/groups. */ + struct list_head *groups = $1; + struct list_head *group = $3; - /* frees $3 */ - parse_events_update_lists(group, list); - $$ = list; + list_splice_tail(group, groups); + free(group); + $$ = groups; } | groups ',' event { - struct list_head *list = $1; + /* Merge event into the list of events/groups. */ + struct list_head *groups = $1; struct list_head *event = $3; - /* frees $3 */ - parse_events_update_lists(event, list); - $$ = list; + + list_splice_tail(event, groups); + free(event); + $$ = groups; } | group @@ -206,12 +215,12 @@ PE_NAME '{' events '}' events: events ',' event { + struct list_head *events = $1; struct list_head *event = $3; - struct list_head *list = $1; - /* frees $3 */ - parse_events_update_lists(event, list); - $$ = list; + list_splice_tail(event, events); + free(event); + $$ = events; } | event -- cgit v1.2.3 From ba5c371edfd05411522c1fefb00e92b54c08c9db Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:27 -0700 Subject: perf parse-events: Improve error message for bad numbers Use the error handler from the parse_state to give a more informative error message. Before: $ perf stat -e 'cycles/period=99999999999999999999/' true event syntax error: 'cycles/period=99999999999999999999/' \___ parser error Run 'perf list' for a list of valid events Usage: perf stat [] [] -e, --event event selector. use 'perf list' to list available events After: $ perf stat -e 'cycles/period=99999999999999999999/' true event syntax error: 'cycles/period=99999999999999999999/' \___ parser error event syntax error: '..les/period=99999999999999999999/' \___ Bad base 10 number "99999999999999999999" Run 'perf list' for a list of valid events Usage: perf stat [] [] -e, --event event selector. use 'perf list' to list available events Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-12-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.l | 40 ++++++++++++++++++++++++---------------- 1 file changed, 24 insertions(+), 16 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.l b/tools/perf/util/parse-events.l index 6fe37003ab7b..0cd68c9f0d4f 100644 --- a/tools/perf/util/parse-events.l +++ b/tools/perf/util/parse-events.l @@ -18,26 +18,34 @@ char *parse_events_get_text(yyscan_t yyscanner); YYSTYPE *parse_events_get_lval(yyscan_t yyscanner); +int parse_events_get_column(yyscan_t yyscanner); +int parse_events_get_leng(yyscan_t yyscanner); -static int __value(YYSTYPE *yylval, char *str, int base, int token) +static int get_column(yyscan_t scanner) { - u64 num; - - errno = 0; - num = strtoull(str, NULL, base); - if (errno) - return PE_ERROR; - - yylval->num = num; - return token; + return parse_events_get_column(scanner) - parse_events_get_leng(scanner); } -static int value(yyscan_t scanner, int base) +static int value(struct parse_events_state *parse_state, yyscan_t scanner, int base) { YYSTYPE *yylval = parse_events_get_lval(scanner); char *text = parse_events_get_text(scanner); + u64 num; - return __value(yylval, text, base, PE_VALUE); + errno = 0; + num = strtoull(text, NULL, base); + if (errno) { + struct parse_events_error *error = parse_state->error; + char *help = NULL; + + if (asprintf(&help, "Bad base %d number \"%s\"", base, text) > 0) + parse_events_error__handle(error, get_column(scanner), help , NULL); + + return PE_ERROR; + } + + yylval->num = num; + return PE_VALUE; } static int str(yyscan_t scanner, int token) @@ -283,8 +291,8 @@ r0x{num_raw_hex} { return str(yyscanner, PE_RAW); } */ "/"/{digit} { return PE_BP_SLASH; } "/"/{non_digit} { BEGIN(config); return '/'; } -{num_dec} { return value(yyscanner, 10); } -{num_hex} { return value(yyscanner, 16); } +{num_dec} { return value(_parse_state, yyscanner, 10); } +{num_hex} { return value(_parse_state, yyscanner, 16); } /* * We need to separate 'mem:' scanner part, in order to get specific * modifier bits parsed out. Otherwise we would need to handle PE_NAME @@ -330,8 +338,8 @@ cgroup-switches { return sym(yyscanner, PERF_COUNT_SW_CGROUP_SWITCHES); } {lc_type}-{lc_op_result}-{lc_op_result} { return str(yyscanner, PE_LEGACY_CACHE); } mem: { BEGIN(mem); return PE_PREFIX_MEM; } r{num_raw_hex} { return str(yyscanner, PE_RAW); } -{num_dec} { return value(yyscanner, 10); } -{num_hex} { return value(yyscanner, 16); } +{num_dec} { return value(_parse_state, yyscanner, 10); } +{num_hex} { return value(_parse_state, yyscanner, 16); } {modifier_event} { return str(yyscanner, PE_MODIFIER_EVENT); } {name} { return str(yyscanner, PE_NAME); } -- cgit v1.2.3 From e18601d80ce1917f02396b8b0d85b7587a0d3af5 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:28 -0700 Subject: perf parse-events: Inline parse_events_evlist_error Inline parse_events_evlist_error that is only used in parse_events_error. Modify parse_events_error to not report a parser error unless errors haven't already been reported. Make it clearer that the latter case only happens for unrecognized input. Before: $ perf stat -e 'cycles/period=99999999999999999999/' true event syntax error: 'cycles/period=99999999999999999999/' \___ parser error event syntax error: '..les/period=99999999999999999999/' \___ Bad base 10 number "99999999999999999999" Run 'perf list' for a list of valid events Usage: perf stat [] [] -e, --event event selector. use 'perf list' to list available events $ perf stat -e 'cycles:xyz' true event syntax error: 'cycles:xyz' \___ parser error Run 'perf list' for a list of valid events Usage: perf stat [] [] -e, --event event selector. use 'perf list' to list available events After: $ perf stat -e 'cycles/period=99999999999999999999/xyz' true event syntax error: '..les/period=99999999999999999999/xyz' \___ Bad base 10 number "99999999999999999999" Run 'perf list' for a list of valid events Usage: perf stat [] [] -e, --event event selector. use 'perf list' to list available events $ perf stat -e 'cycles:xyz' true event syntax error: 'cycles:xyz' \___ Unrecognized input Run 'perf list' for a list of valid events Usage: perf stat [] [] -e, --event event selector. use 'perf list' to list available events Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-13-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 9 --------- tools/perf/util/parse-events.h | 2 -- tools/perf/util/parse-events.y | 10 ++++++++-- 3 files changed, 8 insertions(+), 13 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 7eba714f0d73..ebada37ef98a 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -2760,15 +2760,6 @@ int parse_events_terms__to_strbuf(const struct parse_events_terms *terms, struct return 0; } -void parse_events_evlist_error(struct parse_events_state *parse_state, - int idx, const char *str) -{ - if (!parse_state->error) - return; - - parse_events_error__handle(parse_state->error, idx, strdup(str), NULL); -} - static void config_terms_list(char *buf, size_t buf_sz) { int i; diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index e8f2aebea10f..290ae6c72ec5 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -226,8 +226,6 @@ int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state void *loc_); void parse_events__set_leader(char *name, struct list_head *list); -void parse_events_evlist_error(struct parse_events_state *parse_state, - int idx, const char *str); struct event_symbol { const char *symbol; diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y index 51490f0f8c50..2c4817e126c1 100644 --- a/tools/perf/util/parse-events.y +++ b/tools/perf/util/parse-events.y @@ -790,9 +790,15 @@ sep_slash_slash_dc: '/' '/' | ':' | %% -void parse_events_error(YYLTYPE *loc, void *parse_state, +void parse_events_error(YYLTYPE *loc, void *_parse_state, void *scanner __maybe_unused, char const *msg __maybe_unused) { - parse_events_evlist_error(parse_state, loc->last_column, "parser error"); + struct parse_events_state *parse_state = _parse_state; + + if (!parse_state->error || !list_empty(&parse_state->error->list)) + return; + + parse_events_error__handle(parse_state->error, loc->last_column, + strdup("Unrecognized input"), NULL); } -- cgit v1.2.3 From e30a7912f498c58062d0b75e59c181dd48f1cc56 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:29 -0700 Subject: perf parse-events: Improvements to modifier parsing Use a struct/bitmap rather than a copied string from lexer. In lexer give improved error message when too many precise flags are given or repeated modifiers. Before: $ perf stat -e 'cycles:kuk' true event syntax error: 'cycles:kuk' \___ Bad modifier ... $ perf stat -e 'cycles:pppp' true event syntax error: 'cycles:pppp' \___ Bad modifier ... $ perf stat -e '{instructions:p,cycles:pp}:pp' -a true event syntax error: '..cycles:pp}:pp' \___ Bad modifier ... After: $ perf stat -e 'cycles:kuk' true event syntax error: 'cycles:kuk' \___ Duplicate modifier 'k' (kernel) ... $ perf stat -e 'cycles:pppp' true event syntax error: 'cycles:pppp' \___ Maximum precise value is 3 ... $ perf stat -e '{instructions:p,cycles:pp}:pp' true event syntax error: '..cycles:pp}:pp' \___ Maximum combined precise value is 3, adding precision to "cycles:pp" ... Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-14-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 250 +++++++++++++++-------------------------- tools/perf/util/parse-events.h | 23 +++- tools/perf/util/parse-events.l | 75 ++++++++++++- tools/perf/util/parse-events.y | 28 ++--- 4 files changed, 194 insertions(+), 182 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index ebada37ef98a..3ab533d0e653 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1700,12 +1700,6 @@ int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state return -EINVAL; } -int parse_events__modifier_group(struct list_head *list, - char *event_mod) -{ - return parse_events__modifier_event(list, event_mod, true); -} - void parse_events__set_leader(char *name, struct list_head *list) { struct evsel *leader; @@ -1720,183 +1714,125 @@ void parse_events__set_leader(char *name, struct list_head *list) leader->group_name = name; } -struct event_modifier { - int eu; - int ek; - int eh; - int eH; - int eG; - int eI; - int precise; - int precise_max; - int exclude_GH; - int sample_read; - int pinned; - int weak; - int exclusive; - int bpf_counter; -}; +static int parse_events__modifier_list(struct parse_events_state *parse_state, + YYLTYPE *loc, + struct list_head *list, + struct parse_events_modifier mod, + bool group) +{ + struct evsel *evsel; -static int get_event_modifier(struct event_modifier *mod, char *str, - struct evsel *evsel) -{ - int eu = evsel ? evsel->core.attr.exclude_user : 0; - int ek = evsel ? evsel->core.attr.exclude_kernel : 0; - int eh = evsel ? evsel->core.attr.exclude_hv : 0; - int eH = evsel ? evsel->core.attr.exclude_host : 0; - int eG = evsel ? evsel->core.attr.exclude_guest : 0; - int eI = evsel ? evsel->core.attr.exclude_idle : 0; - int precise = evsel ? evsel->core.attr.precise_ip : 0; - int precise_max = 0; - int sample_read = 0; - int pinned = evsel ? evsel->core.attr.pinned : 0; - int exclusive = evsel ? evsel->core.attr.exclusive : 0; - - int exclude = eu | ek | eh; - int exclude_GH = evsel ? evsel->exclude_GH : 0; - int weak = 0; - int bpf_counter = 0; - - memset(mod, 0, sizeof(*mod)); - - while (*str) { - if (*str == 'u') { + if (!group && mod.weak) { + parse_events_error__handle(parse_state->error, loc->first_column, + strdup("Weak modifier is for use with groups"), NULL); + return -EINVAL; + } + + __evlist__for_each_entry(list, evsel) { + /* Translate modifiers into the equivalent evsel excludes. */ + int eu = group ? evsel->core.attr.exclude_user : 0; + int ek = group ? evsel->core.attr.exclude_kernel : 0; + int eh = group ? evsel->core.attr.exclude_hv : 0; + int eH = group ? evsel->core.attr.exclude_host : 0; + int eG = group ? evsel->core.attr.exclude_guest : 0; + int exclude = eu | ek | eh; + int exclude_GH = group ? evsel->exclude_GH : 0; + + if (mod.precise) { + /* use of precise requires exclude_guest */ + eG = 1; + } + if (mod.user) { if (!exclude) exclude = eu = ek = eh = 1; if (!exclude_GH && !perf_guest) eG = 1; eu = 0; - } else if (*str == 'k') { + } + if (mod.kernel) { if (!exclude) exclude = eu = ek = eh = 1; ek = 0; - } else if (*str == 'h') { + } + if (mod.hypervisor) { if (!exclude) exclude = eu = ek = eh = 1; eh = 0; - } else if (*str == 'G') { + } + if (mod.guest) { if (!exclude_GH) exclude_GH = eG = eH = 1; eG = 0; - } else if (*str == 'H') { + } + if (mod.host) { if (!exclude_GH) exclude_GH = eG = eH = 1; eH = 0; - } else if (*str == 'I') { - eI = 1; - } else if (*str == 'p') { - precise++; - /* use of precise requires exclude_guest */ - if (!exclude_GH) - eG = 1; - } else if (*str == 'P') { - precise_max = 1; - } else if (*str == 'S') { - sample_read = 1; - } else if (*str == 'D') { - pinned = 1; - } else if (*str == 'e') { - exclusive = 1; - } else if (*str == 'W') { - weak = 1; - } else if (*str == 'b') { - bpf_counter = 1; - } else - break; - - ++str; + } + evsel->core.attr.exclude_user = eu; + evsel->core.attr.exclude_kernel = ek; + evsel->core.attr.exclude_hv = eh; + evsel->core.attr.exclude_host = eH; + evsel->core.attr.exclude_guest = eG; + evsel->exclude_GH = exclude_GH; + + /* Simple modifiers copied to the evsel. */ + if (mod.precise) { + u8 precise = evsel->core.attr.precise_ip + mod.precise; + /* + * precise ip: + * + * 0 - SAMPLE_IP can have arbitrary skid + * 1 - SAMPLE_IP must have constant skid + * 2 - SAMPLE_IP requested to have 0 skid + * 3 - SAMPLE_IP must have 0 skid + * + * See also PERF_RECORD_MISC_EXACT_IP + */ + if (precise > 3) { + char *help; + + if (asprintf(&help, + "Maximum combined precise value is 3, adding precision to \"%s\"", + evsel__name(evsel)) > 0) { + parse_events_error__handle(parse_state->error, + loc->first_column, + help, NULL); + } + return -EINVAL; + } + evsel->core.attr.precise_ip = precise; + } + if (mod.precise_max) + evsel->precise_max = 1; + if (mod.non_idle) + evsel->core.attr.exclude_idle = 1; + if (mod.sample_read) + evsel->sample_read = 1; + if (mod.pinned && evsel__is_group_leader(evsel)) + evsel->core.attr.pinned = 1; + if (mod.exclusive && evsel__is_group_leader(evsel)) + evsel->core.attr.exclusive = 1; + if (mod.weak) + evsel->weak_group = true; + if (mod.bpf) + evsel->bpf_counter = true; } - - /* - * precise ip: - * - * 0 - SAMPLE_IP can have arbitrary skid - * 1 - SAMPLE_IP must have constant skid - * 2 - SAMPLE_IP requested to have 0 skid - * 3 - SAMPLE_IP must have 0 skid - * - * See also PERF_RECORD_MISC_EXACT_IP - */ - if (precise > 3) - return -EINVAL; - - mod->eu = eu; - mod->ek = ek; - mod->eh = eh; - mod->eH = eH; - mod->eG = eG; - mod->eI = eI; - mod->precise = precise; - mod->precise_max = precise_max; - mod->exclude_GH = exclude_GH; - mod->sample_read = sample_read; - mod->pinned = pinned; - mod->weak = weak; - mod->bpf_counter = bpf_counter; - mod->exclusive = exclusive; - return 0; } -/* - * Basic modifier sanity check to validate it contains only one - * instance of any modifier (apart from 'p') present. - */ -static int check_modifier(char *str) +int parse_events__modifier_group(struct parse_events_state *parse_state, void *loc, + struct list_head *list, + struct parse_events_modifier mod) { - char *p = str; - - /* The sizeof includes 0 byte as well. */ - if (strlen(str) > (sizeof("ukhGHpppPSDIWeb") - 1)) - return -1; - - while (*p) { - if (*p != 'p' && strchr(p + 1, *p)) - return -1; - p++; - } - - return 0; + return parse_events__modifier_list(parse_state, loc, list, mod, /*group=*/true); } -int parse_events__modifier_event(struct list_head *list, char *str, bool add) +int parse_events__modifier_event(struct parse_events_state *parse_state, void *loc, + struct list_head *list, + struct parse_events_modifier mod) { - struct evsel *evsel; - struct event_modifier mod; - - if (str == NULL) - return 0; - - if (check_modifier(str)) - return -EINVAL; - - if (!add && get_event_modifier(&mod, str, NULL)) - return -EINVAL; - - __evlist__for_each_entry(list, evsel) { - if (add && get_event_modifier(&mod, str, evsel)) - return -EINVAL; - - evsel->core.attr.exclude_user = mod.eu; - evsel->core.attr.exclude_kernel = mod.ek; - evsel->core.attr.exclude_hv = mod.eh; - evsel->core.attr.precise_ip = mod.precise; - evsel->core.attr.exclude_host = mod.eH; - evsel->core.attr.exclude_guest = mod.eG; - evsel->core.attr.exclude_idle = mod.eI; - evsel->exclude_GH = mod.exclude_GH; - evsel->sample_read = mod.sample_read; - evsel->precise_max = mod.precise_max; - evsel->weak_group = mod.weak; - evsel->bpf_counter = mod.bpf_counter; - - if (evsel__is_group_leader(evsel)) { - evsel->core.attr.pinned = mod.pinned; - evsel->core.attr.exclusive = mod.exclusive; - } - } - - return 0; + return parse_events__modifier_list(parse_state, loc, list, mod, /*group=*/false); } int parse_events_name(struct list_head *list, const char *name) diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index 290ae6c72ec5..f104faef1a78 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -186,8 +186,27 @@ void parse_events_terms__init(struct parse_events_terms *terms); void parse_events_terms__exit(struct parse_events_terms *terms); int parse_events_terms(struct parse_events_terms *terms, const char *str, FILE *input); int parse_events_terms__to_strbuf(const struct parse_events_terms *terms, struct strbuf *sb); -int parse_events__modifier_event(struct list_head *list, char *str, bool add); -int parse_events__modifier_group(struct list_head *list, char *event_mod); + +struct parse_events_modifier { + u8 precise; /* Number of repeated 'p' for precision. */ + bool precise_max : 1; /* 'P' */ + bool non_idle : 1; /* 'I' */ + bool sample_read : 1; /* 'S' */ + bool pinned : 1; /* 'D' */ + bool exclusive : 1; /* 'e' */ + bool weak : 1; /* 'W' */ + bool bpf : 1; /* 'b' */ + bool user : 1; /* 'u' */ + bool kernel : 1; /* 'k' */ + bool hypervisor : 1; /* 'h' */ + bool guest : 1; /* 'G' */ + bool host : 1; /* 'H' */ +}; + +int parse_events__modifier_event(struct parse_events_state *parse_state, void *loc, + struct list_head *list, struct parse_events_modifier mod); +int parse_events__modifier_group(struct parse_events_state *parse_state, void *loc, + struct list_head *list, struct parse_events_modifier mod); int parse_events_name(struct list_head *list, const char *name); int parse_events_add_tracepoint(struct list_head *list, int *idx, const char *sys, const char *event, diff --git a/tools/perf/util/parse-events.l b/tools/perf/util/parse-events.l index 0cd68c9f0d4f..4aaf0c53d9b6 100644 --- a/tools/perf/util/parse-events.l +++ b/tools/perf/util/parse-events.l @@ -142,6 +142,77 @@ static int hw(yyscan_t scanner, int config) return PE_TERM_HW; } +static void modifiers_error(struct parse_events_state *parse_state, yyscan_t scanner, + int pos, char mod_char, const char *mod_name) +{ + struct parse_events_error *error = parse_state->error; + char *help = NULL; + + if (asprintf(&help, "Duplicate modifier '%c' (%s)", mod_char, mod_name) > 0) + parse_events_error__handle(error, get_column(scanner) + pos, help , NULL); +} + +static int modifiers(struct parse_events_state *parse_state, yyscan_t scanner) +{ + YYSTYPE *yylval = parse_events_get_lval(scanner); + char *text = parse_events_get_text(scanner); + struct parse_events_modifier mod = { .precise = 0, }; + + for (size_t i = 0, n = strlen(text); i < n; i++) { +#define CASE(c, field) \ + case c: \ + if (mod.field) { \ + modifiers_error(parse_state, scanner, i, c, #field); \ + return PE_ERROR; \ + } \ + mod.field = true; \ + break + + switch (text[i]) { + CASE('u', user); + CASE('k', kernel); + CASE('h', hypervisor); + CASE('I', non_idle); + CASE('G', guest); + CASE('H', host); + case 'p': + mod.precise++; + /* + * precise ip: + * + * 0 - SAMPLE_IP can have arbitrary skid + * 1 - SAMPLE_IP must have constant skid + * 2 - SAMPLE_IP requested to have 0 skid + * 3 - SAMPLE_IP must have 0 skid + * + * See also PERF_RECORD_MISC_EXACT_IP + */ + if (mod.precise > 3) { + struct parse_events_error *error = parse_state->error; + char *help = strdup("Maximum precise value is 3"); + + if (help) { + parse_events_error__handle(error, get_column(scanner) + i, + help , NULL); + } + return PE_ERROR; + } + break; + CASE('P', precise_max); + CASE('S', sample_read); + CASE('D', pinned); + CASE('W', weak); + CASE('e', exclusive); + CASE('b', bpf); + default: + return PE_ERROR; + } +#undef CASE + } + yylval->mod = mod; + return PE_MODIFIER_EVENT; +} + #define YY_USER_ACTION \ do { \ yylloc->last_column = yylloc->first_column; \ @@ -174,7 +245,7 @@ drv_cfg_term [a-zA-Z0-9_\.]+(=[a-zA-Z0-9_*?\.:]+)? * If you add a modifier you need to update check_modifier(). * Also, the letters in modifier_event must not be in modifier_bp. */ -modifier_event [ukhpPGHSDIWeb]+ +modifier_event [ukhpPGHSDIWeb]{1,15} modifier_bp [rwx]{1,3} lc_type (L1-dcache|l1-d|l1d|L1-data|L1-icache|l1-i|l1i|L1-instruction|LLC|L2|dTLB|d-tlb|Data-TLB|iTLB|i-tlb|Instruction-TLB|branch|branches|bpu|btb|bpc|node) lc_op_result (load|loads|read|store|stores|write|prefetch|prefetches|speculative-read|speculative-load|refs|Reference|ops|access|misses|miss) @@ -341,7 +412,7 @@ r{num_raw_hex} { return str(yyscanner, PE_RAW); } {num_dec} { return value(_parse_state, yyscanner, 10); } {num_hex} { return value(_parse_state, yyscanner, 16); } -{modifier_event} { return str(yyscanner, PE_MODIFIER_EVENT); } +{modifier_event} { return modifiers(_parse_state, yyscanner); } {name} { return str(yyscanner, PE_NAME); } {name_tag} { return str(yyscanner, PE_NAME); } "/" { BEGIN(config); return '/'; } diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y index 2c4817e126c1..79f254189be6 100644 --- a/tools/perf/util/parse-events.y +++ b/tools/perf/util/parse-events.y @@ -68,11 +68,11 @@ static void free_list_evsel(struct list_head* list_evsel) %type PE_VALUE %type PE_VALUE_SYM_SW %type PE_VALUE_SYM_TOOL +%type PE_MODIFIER_EVENT %type PE_TERM %type PE_RAW %type PE_NAME %type PE_LEGACY_CACHE -%type PE_MODIFIER_EVENT %type PE_MODIFIER_BP %type PE_EVENT_NAME %type PE_DRV_CFG_TERM @@ -110,6 +110,7 @@ static void free_list_evsel(struct list_head* list_evsel) { char *str; u64 num; + struct parse_events_modifier mod; enum parse_events__term_type term_type; struct list_head *list_evsel; struct parse_events_terms *list_terms; @@ -175,20 +176,13 @@ event group: group_def ':' PE_MODIFIER_EVENT { + /* Apply the modifier to the events in the group_def. */ struct list_head *list = $1; int err; - err = parse_events__modifier_group(list, $3); - free($3); - if (err) { - struct parse_events_state *parse_state = _parse_state; - struct parse_events_error *error = parse_state->error; - - parse_events_error__handle(error, @3.first_column, - strdup("Bad modifier"), NULL); - free_list_evsel(list); + err = parse_events__modifier_group(_parse_state, &@3, list, $3); + if (err) YYABORT; - } $$ = list; } | @@ -238,17 +232,9 @@ event_name PE_MODIFIER_EVENT * (there could be more events added for multiple tracepoint * definitions via '*?'. */ - err = parse_events__modifier_event(list, $2, false); - free($2); - if (err) { - struct parse_events_state *parse_state = _parse_state; - struct parse_events_error *error = parse_state->error; - - parse_events_error__handle(error, @2.first_column, - strdup("Bad modifier"), NULL); - free_list_evsel(list); + err = parse_events__modifier_event(_parse_state, &@2, list, $2); + if (err) YYABORT; - } $$ = list; } | -- cgit v1.2.3 From 4a20e793652e7742d8c0608a32e6f8adcde9e272 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:30 -0700 Subject: perf parse-event: Constify event_symbol arrays Moves 352 bytes from .data to .data.rel.ro. Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-15-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 4 ++-- tools/perf/util/parse-events.h | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 3ab533d0e653..8d3d692d219d 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -39,7 +39,7 @@ static int get_config_terms(const struct parse_events_terms *head_config, static int parse_events_terms__copy(const struct parse_events_terms *src, struct parse_events_terms *dest); -struct event_symbol event_symbols_hw[PERF_COUNT_HW_MAX] = { +const struct event_symbol event_symbols_hw[PERF_COUNT_HW_MAX] = { [PERF_COUNT_HW_CPU_CYCLES] = { .symbol = "cpu-cycles", .alias = "cycles", @@ -82,7 +82,7 @@ struct event_symbol event_symbols_hw[PERF_COUNT_HW_MAX] = { }, }; -struct event_symbol event_symbols_sw[PERF_COUNT_SW_MAX] = { +const struct event_symbol event_symbols_sw[PERF_COUNT_SW_MAX] = { [PERF_COUNT_SW_CPU_CLOCK] = { .symbol = "cpu-clock", .alias = "", diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index f104faef1a78..0bb5f0c80a5e 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -250,8 +250,8 @@ struct event_symbol { const char *symbol; const char *alias; }; -extern struct event_symbol event_symbols_hw[]; -extern struct event_symbol event_symbols_sw[]; +extern const struct event_symbol event_symbols_hw[]; +extern const struct event_symbol event_symbols_sw[]; char *parse_events_formats_error_string(char *additional_terms); -- cgit v1.2.3 From afd876bbdc97664257523bb5f2dcedbb5cef40d0 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:31 -0700 Subject: perf parse-events: Minor grouping tidy up Add comments. Ensure leader->group_name is freed before overwriting it. Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-16-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 1 + tools/perf/util/parse-events.y | 6 +++++- 2 files changed, 6 insertions(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 8d3d692d219d..1c1b1bcb78e8 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1711,6 +1711,7 @@ void parse_events__set_leader(char *name, struct list_head *list) leader = list_first_entry(list, struct evsel, core.node); __perf_evlist__set_leader(list, &leader->core); + zfree(&leader->group_name); leader->group_name = name; } diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y index 79f254189be6..6f1042272dda 100644 --- a/tools/perf/util/parse-events.y +++ b/tools/perf/util/parse-events.y @@ -193,7 +193,10 @@ PE_NAME '{' events '}' { struct list_head *list = $3; - /* Takes ownership of $1. */ + /* + * Set the first entry of list to be the leader. Set the group name on + * the leader to $1 taking ownership. + */ parse_events__set_leader($1, list); $$ = list; } @@ -202,6 +205,7 @@ PE_NAME '{' events '}' { struct list_head *list = $2; + /* Set the first entry of list to be the leader clearing the group name. */ parse_events__set_leader(NULL, list); $$ = list; } -- cgit v1.2.3 From bb65ff7810b654beb498e7afb4a4688e6ddf0340 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 15 Apr 2024 23:15:32 -0700 Subject: perf parse-events: Tidy the setting of the default event name Add comments. Pass ownership of the event name to save on a strdup. Signed-off-by: Ian Rogers Reviewed-by: Kan Liang Tested-by: Atish Patra Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Beeman Strong Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240416061533.921723-17-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 9 ++++++--- tools/perf/util/parse-events.h | 2 +- tools/perf/util/parse-events.l | 5 +++++ tools/perf/util/parse-events.y | 10 +++++++--- 4 files changed, 19 insertions(+), 7 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 1c1b1bcb78e8..0f308b4db2b9 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1836,18 +1836,21 @@ int parse_events__modifier_event(struct parse_events_state *parse_state, void *l return parse_events__modifier_list(parse_state, loc, list, mod, /*group=*/false); } -int parse_events_name(struct list_head *list, const char *name) +int parse_events__set_default_name(struct list_head *list, char *name) { struct evsel *evsel; + bool used_name = false; __evlist__for_each_entry(list, evsel) { if (!evsel->name) { - evsel->name = strdup(name); + evsel->name = used_name ? strdup(name) : name; + used_name = true; if (!evsel->name) return -ENOMEM; } } - + if (!used_name) + free(name); return 0; } diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index 0bb5f0c80a5e..5695308efab9 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -207,7 +207,7 @@ int parse_events__modifier_event(struct parse_events_state *parse_state, void *l struct list_head *list, struct parse_events_modifier mod); int parse_events__modifier_group(struct parse_events_state *parse_state, void *loc, struct list_head *list, struct parse_events_modifier mod); -int parse_events_name(struct list_head *list, const char *name); +int parse_events__set_default_name(struct list_head *list, char *name); int parse_events_add_tracepoint(struct list_head *list, int *idx, const char *sys, const char *event, struct parse_events_error *error, diff --git a/tools/perf/util/parse-events.l b/tools/perf/util/parse-events.l index 4aaf0c53d9b6..08ea2d845dc3 100644 --- a/tools/perf/util/parse-events.l +++ b/tools/perf/util/parse-events.l @@ -96,6 +96,11 @@ static int drv_str(yyscan_t scanner, int token) return token; } +/* + * Use yyless to return all the characaters to the input. Update the column for + * location debugging. If __alloc is non-zero set yylval to the text for the + * returned token's value. + */ #define REWIND(__alloc) \ do { \ YYSTYPE *__yylval = parse_events_get_lval(yyscanner); \ diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y index 6f1042272dda..68b3b06c7ff0 100644 --- a/tools/perf/util/parse-events.y +++ b/tools/perf/util/parse-events.y @@ -247,10 +247,14 @@ event_name event_name: PE_EVENT_NAME event_def { - int err; + /* + * When an event is parsed the text is rewound and the entire text of + * the event is set to the str of PE_EVENT_NAME token matched here. If + * no name was on an event via a term, set the name to the entire text + * taking ownership of the allocation. + */ + int err = parse_events__set_default_name($2, $1); - err = parse_events_name($2, $1); - free($1); if (err) { free_list_evsel($2); YYNOMEM; -- cgit v1.2.3 From 47557db99a5decba2192303ed34ec1add5177b51 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 24 Apr 2024 16:00:15 -0700 Subject: perf annotate-data: Check if 'struct annotation_source' was allocated on 'perf report' TUI As it removed the sample accounting for code when no symbol sort key is given for 'perf report' TUI, it might not have allocated the 'struct annotated_source' yet. Let's check if it's NULL first. Fixes: 6cdd977ec24e1538 ("perf report: Do not collect sample histogram unnecessarily") Reviewed-by: Ian Rogers Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240424230015.1054013-1-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index dca2c08ab8c5..f5b6b5e5e757 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -906,7 +906,7 @@ int symbol__annotate(struct map_symbol *ms, struct evsel *evsel, if (parch) *parch = arch; - if (!list_empty(¬es->src->source)) + if (notes->src && !list_empty(¬es->src->source)) return 0; args.arch = arch; -- cgit v1.2.3 From f35847de2a65137e011e559f38a3de5902a5463f Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 24 Apr 2024 17:51:56 -0700 Subject: perf annotate: Fallback disassemble to objdump when capstone fails I found some cases that capstone failed to disassemble. Probably my capstone is an old version but anyway there's a chance it can fail. And then it silently stopped in the middle. In my case, it didn't understand "RDPKRU" instruction. Let's check if the capstone disassemble reached the end of the function and fallback to objdump if not. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240425005157.1104789-1-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/disasm.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/disasm.c b/tools/perf/util/disasm.c index 92937809be85..412101f2cf2a 100644 --- a/tools/perf/util/disasm.c +++ b/tools/perf/util/disasm.c @@ -1542,6 +1542,20 @@ static int symbol__disassemble_capstone(char *filename, struct symbol *sym, offset += insn[i].size; } + /* It failed in the middle: probably due to unknown instructions */ + if (offset != len) { + struct list_head *list = ¬es->src->source; + + /* Discard all lines and fallback to objdump */ + while (!list_empty(list)) { + dl = list_first_entry(list, struct disasm_line, al.node); + + list_del_init(&dl->al.node); + disasm_line__free(dl); + } + count = -1; + } + out: if (needs_cs_close) cs_close(&handle); -- cgit v1.2.3 From 8f3ec810bb668d421f9d260e3de3bccca954b56f Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 24 Apr 2024 17:51:57 -0700 Subject: perf annotate: Update DSO binary type when trying build-id dso__disassemble_filename() tries to get the filename for objdump (or capstone) using build-id. But I found sometimes it didn't disassemble some functions. It turned out that those functions belong to a DSO which has no binary type set. It seems it sets the binary type for some special files only - like kernel (kallsyms or kcore) or BPF images. And there's a logic to skip dso with DSO_BINARY_TYPE__NOT_FOUND. As it's checked the build-id cache link, it should set the binary type as DSO_BINARY_TYPE__BUILD_ID_CACHE. Fixes: 873a83731f1cc85c ("perf annotate: Skip DSOs not found") Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240425005157.1104789-2-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/disasm.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/disasm.c b/tools/perf/util/disasm.c index 412101f2cf2a..6d1125e687b7 100644 --- a/tools/perf/util/disasm.c +++ b/tools/perf/util/disasm.c @@ -1156,6 +1156,8 @@ fallback: } } mutex_unlock(&dso->lock); + } else if (dso->binary_type == DSO_BINARY_TYPE__NOT_FOUND) { + dso->binary_type = DSO_BINARY_TYPE__BUILD_ID_CACHE; } free(build_id_path); -- cgit v1.2.3 From e101a05f79fd4ee3e89d2f3fb716493c33a33708 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 26 Mar 2024 10:32:23 +0200 Subject: perf intel-pt: Fix unassigned instruction op (discovered by MemorySanitizer) MemorySanitizer discovered instances where the instruction op value was not assigned.: WARNING: MemorySanitizer: use-of-uninitialized-value #0 0x5581c00a76b3 in intel_pt_sample_flags tools/perf/util/intel-pt.c:1527:17 Uninitialized value was stored to memory at #0 0x5581c005ddf8 in intel_pt_walk_insn tools/perf/util/intel-pt-decoder/intel-pt-decoder.c:1256:25 The op value is used to set branch flags for branch instructions encountered when walking the code, so fix by setting op to INTEL_PT_OP_OTHER in other cases. Fixes: 4c761d805bb2d2ea ("perf intel-pt: Fix intel_pt_fup_event() assumptions about setting state type") Reported-by: Ian Rogers Signed-off-by: Adrian Hunter Tested-by: Ian Rogers Cc: Jiri Olsa Cc: Namhyung Kim Closes: https://lore.kernel.org/linux-perf-users/20240320162619.1272015-1-irogers@google.com/ Link: https://lore.kernel.org/r/20240326083223.10883-1-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/intel-pt-decoder/intel-pt-decoder.c | 2 ++ tools/perf/util/intel-pt.c | 2 ++ 2 files changed, 4 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/intel-pt-decoder/intel-pt-decoder.c b/tools/perf/util/intel-pt-decoder/intel-pt-decoder.c index b450178e3420..e733f6b1f7ac 100644 --- a/tools/perf/util/intel-pt-decoder/intel-pt-decoder.c +++ b/tools/perf/util/intel-pt-decoder/intel-pt-decoder.c @@ -1319,6 +1319,8 @@ static bool intel_pt_fup_event(struct intel_pt_decoder *decoder, bool no_tip) bool ret = false; decoder->state.type &= ~INTEL_PT_BRANCH; + decoder->state.insn_op = INTEL_PT_OP_OTHER; + decoder->state.insn_len = 0; if (decoder->set_fup_cfe_ip || decoder->set_fup_cfe) { bool ip = decoder->set_fup_cfe_ip; diff --git a/tools/perf/util/intel-pt.c b/tools/perf/util/intel-pt.c index f38893e0b036..4db9a098f592 100644 --- a/tools/perf/util/intel-pt.c +++ b/tools/perf/util/intel-pt.c @@ -764,6 +764,7 @@ static int intel_pt_walk_next_insn(struct intel_pt_insn *intel_pt_insn, addr_location__init(&al); intel_pt_insn->length = 0; + intel_pt_insn->op = INTEL_PT_OP_OTHER; if (to_ip && *ip == to_ip) goto out_no_cache; @@ -898,6 +899,7 @@ static int intel_pt_walk_next_insn(struct intel_pt_insn *intel_pt_insn, if (to_ip && *ip == to_ip) { intel_pt_insn->length = 0; + intel_pt_insn->op = INTEL_PT_OP_OTHER; goto out_no_cache; } -- cgit v1.2.3 From d7b60803a790b716dc930bde2155c6ecb7537d51 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 1 May 2024 23:00:06 -0700 Subject: perf dwarf-aux: Add die_collect_global_vars() This function is to search all global variables in the CU. We want to have the list of global variables at once and match them later. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Masami Hiramatsu Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240502060011.1838090-2-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dwarf-aux.c | 62 +++++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/dwarf-aux.h | 8 ++++++ 2 files changed, 70 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c index 40cfbdfe2d75..c0a492e65388 100644 --- a/tools/perf/util/dwarf-aux.c +++ b/tools/perf/util/dwarf-aux.c @@ -1630,6 +1630,68 @@ void die_collect_vars(Dwarf_Die *sc_die, struct die_var_type **var_types) die_find_child(sc_die, __die_collect_vars_cb, (void *)var_types, &die_mem); } + +static int __die_collect_global_vars_cb(Dwarf_Die *die_mem, void *arg) +{ + struct die_var_type **var_types = arg; + Dwarf_Die type_die; + int tag = dwarf_tag(die_mem); + Dwarf_Attribute attr; + Dwarf_Addr base, start, end; + Dwarf_Op *ops; + size_t nops; + struct die_var_type *vt; + + if (tag != DW_TAG_variable) + return DIE_FIND_CB_SIBLING; + + if (dwarf_attr(die_mem, DW_AT_location, &attr) == NULL) + return DIE_FIND_CB_SIBLING; + + /* Only collect the location with an absolute address. */ + if (dwarf_getlocations(&attr, 0, &base, &start, &end, &ops, &nops) <= 0) + return DIE_FIND_CB_SIBLING; + + if (ops->atom != DW_OP_addr) + return DIE_FIND_CB_SIBLING; + + if (!check_allowed_ops(ops, nops)) + return DIE_FIND_CB_SIBLING; + + if (die_get_real_type(die_mem, &type_die) == NULL) + return DIE_FIND_CB_SIBLING; + + vt = malloc(sizeof(*vt)); + if (vt == NULL) + return DIE_FIND_CB_END; + + vt->die_off = dwarf_dieoffset(&type_die); + vt->addr = ops->number; + vt->reg = -1; + vt->offset = 0; + vt->next = *var_types; + *var_types = vt; + + return DIE_FIND_CB_SIBLING; +} + +/** + * die_collect_global_vars - Save all global variables + * @cu_die: a CU DIE + * @var_types: a pointer to save the resulting list + * + * Save all global variables in the @cu_die and save them to @var_types. + * The @var_types is a singly-linked list containing type and location info. + * Actual type can be retrieved using dwarf_offdie() with 'die_off' later. + * + * Callers should free @var_types. + */ +void die_collect_global_vars(Dwarf_Die *cu_die, struct die_var_type **var_types) +{ + Dwarf_Die die_mem; + + die_find_child(cu_die, __die_collect_global_vars_cb, (void *)var_types, &die_mem); +} #endif /* HAVE_DWARF_GETLOCATIONS_SUPPORT */ #ifdef HAVE_DWARF_CFI_SUPPORT diff --git a/tools/perf/util/dwarf-aux.h b/tools/perf/util/dwarf-aux.h index b0f25fbf9668..24446412b869 100644 --- a/tools/perf/util/dwarf-aux.h +++ b/tools/perf/util/dwarf-aux.h @@ -171,6 +171,9 @@ Dwarf_Die *die_find_variable_by_addr(Dwarf_Die *sc_die, Dwarf_Addr addr, /* Save all variables and parameters in this scope */ void die_collect_vars(Dwarf_Die *sc_die, struct die_var_type **var_types); +/* Save all global variables in this CU */ +void die_collect_global_vars(Dwarf_Die *cu_die, struct die_var_type **var_types); + #else /* HAVE_DWARF_GETLOCATIONS_SUPPORT */ static inline int die_get_var_range(Dwarf_Die *sp_die __maybe_unused, @@ -203,6 +206,11 @@ static inline void die_collect_vars(Dwarf_Die *sc_die __maybe_unused, { } +static inline void die_collect_global_vars(Dwarf_Die *cu_die __maybe_unused, + struct die_var_type **var_types __maybe_unused) +{ +} + #endif /* HAVE_DWARF_GETLOCATIONS_SUPPORT */ #ifdef HAVE_DWARF_CFI_SUPPORT -- cgit v1.2.3 From c1da8411e4be2a96a448979baede9a0e86c5baf8 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 1 May 2024 23:00:07 -0700 Subject: perf annotate-data: Collect global variables in advance Currently it looks up global variables from the current CU using address and name. But it sometimes fails to find a variable as the variable can come from a different CU - but it's still strange it failed to find a declaration for some reason. Anyway, it can collect all global variables from all CU once and then lookup them later on. This slightly improves the success rate of my test data set. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240502060011.1838090-3-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 57 +++++++++++++++++++++++++++++++++++++++-- 1 file changed, 55 insertions(+), 2 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 12d5faff3b7a..4dd0911904f2 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -28,6 +28,8 @@ /* register number of the stack pointer */ #define X86_REG_SP 7 +static void delete_var_types(struct die_var_type *var_types); + enum type_state_kind { TSR_KIND_INVALID = 0, TSR_KIND_TYPE, @@ -557,8 +559,8 @@ static bool global_var__add(struct data_loc_info *dloc, u64 addr, if (gvar == NULL) return false; - gvar->name = strdup(name); - if (gvar->name == NULL) { + gvar->name = name ? strdup(name) : NULL; + if (name && gvar->name == NULL) { free(gvar); return false; } @@ -612,6 +614,53 @@ static bool get_global_var_info(struct data_loc_info *dloc, u64 addr, return true; } +static void global_var__collect(struct data_loc_info *dloc) +{ + Dwarf *dwarf = dloc->di->dbg; + Dwarf_Off off, next_off; + Dwarf_Die cu_die, type_die; + size_t header_size; + + /* Iterate all CU and collect global variables that have no location in a register. */ + off = 0; + while (dwarf_nextcu(dwarf, off, &next_off, &header_size, + NULL, NULL, NULL) == 0) { + struct die_var_type *var_types = NULL; + struct die_var_type *pos; + + if (dwarf_offdie(dwarf, off + header_size, &cu_die) == NULL) { + off = next_off; + continue; + } + + die_collect_global_vars(&cu_die, &var_types); + + for (pos = var_types; pos; pos = pos->next) { + const char *var_name = NULL; + int var_offset = 0; + + if (pos->reg != -1) + continue; + + if (!dwarf_offdie(dwarf, pos->die_off, &type_die)) + continue; + + if (!get_global_var_info(dloc, pos->addr, &var_name, + &var_offset)) + continue; + + if (var_offset != 0) + continue; + + global_var__add(dloc, pos->addr, var_name, &type_die); + } + + delete_var_types(var_types); + + off = next_off; + } +} + static bool get_global_var_type(Dwarf_Die *cu_die, struct data_loc_info *dloc, u64 ip, u64 var_addr, int *var_offset, Dwarf_Die *type_die) @@ -620,8 +669,12 @@ static bool get_global_var_type(Dwarf_Die *cu_die, struct data_loc_info *dloc, int offset; const char *var_name = NULL; struct global_var_entry *gvar; + struct dso *dso = map__dso(dloc->ms->map); Dwarf_Die var_die; + if (RB_EMPTY_ROOT(&dso->global_vars)) + global_var__collect(dloc); + gvar = global_var__find(dloc, var_addr); if (gvar) { if (!dwarf_offdie(dloc->di->dbg, gvar->die_offset, type_die)) -- cgit v1.2.3 From 4449c9047dc6f9f68333a720958cd7b58225910d Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 1 May 2024 23:00:08 -0700 Subject: perf annotate-data: Handle direct global variable access Like per-cpu base offset array, sometimes it accesses the global variable directly using the offset. Allow this type of instructions as long as it finds a global variable for the address. movslq %edi, %rcx mov -0x7dc94ae0(,%rcx,8), %rcx <<<--- here As %rcx has a valid type (i.e. array index) from the first instruction, it will be checked by the first case in check_matching_type(). But as it's not a pointer type, the match will fail. But in this case, it should check if it accesses the kernel global array variable. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240502060011.1838090-4-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 27 +++++++++++++++------------ 1 file changed, 15 insertions(+), 12 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 4dd0911904f2..f1e52a531563 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -1256,14 +1256,19 @@ static int check_matching_type(struct type_state *state, if (state->regs[reg].ok && state->regs[reg].kind == TSR_KIND_TYPE) { int tag = dwarf_tag(&state->regs[reg].type); - pr_debug_dtp("\n"); - /* * Normal registers should hold a pointer (or array) to * dereference a memory location. */ - if (tag != DW_TAG_pointer_type && tag != DW_TAG_array_type) + if (tag != DW_TAG_pointer_type && tag != DW_TAG_array_type) { + if (dloc->op->offset < 0 && reg != state->stack_reg) + goto check_kernel; + + pr_debug_dtp("\n"); return -1; + } + + pr_debug_dtp("\n"); /* Remove the pointer and get the target type */ if (die_get_real_type(&state->regs[reg].type, type_die) == NULL) @@ -1376,12 +1381,14 @@ static int check_matching_type(struct type_state *state, return -1; } - if (map__dso(dloc->ms->map)->kernel && arch__is(dloc->arch, "x86")) { +check_kernel: + if (map__dso(dloc->ms->map)->kernel) { u64 addr; int offset; /* Direct this-cpu access like "%gs:0x34740" */ - if (dloc->op->segment == INSN_SEG_X86_GS && dloc->op->imm) { + if (dloc->op->segment == INSN_SEG_X86_GS && dloc->op->imm && + arch__is(dloc->arch, "x86")) { pr_debug_dtp(" this-cpu var\n"); addr = dloc->op->offset; @@ -1394,17 +1401,13 @@ static int check_matching_type(struct type_state *state, return -1; } - /* Access to per-cpu base like "-0x7dcf0500(,%rdx,8)" */ + /* Access to global variable like "-0x7dcf0500(,%rdx,8)" */ if (dloc->op->offset < 0 && reg != state->stack_reg) { - const char *var_name = NULL; - addr = (s64) dloc->op->offset; - if (get_global_var_info(dloc, addr, &var_name, &offset) && - !strcmp(var_name, "__per_cpu_offset") && offset == 0 && - get_global_var_type(cu_die, dloc, dloc->ip, addr, + if (get_global_var_type(cu_die, dloc, dloc->ip, addr, &offset, type_die)) { - pr_debug_dtp(" percpu base\n"); + pr_debug_dtp(" global var\n"); dloc->type_offset = offset; return 1; -- cgit v1.2.3 From eba1f853edf794ec259ec7b5e5a6efee5ede989f Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 1 May 2024 23:00:09 -0700 Subject: perf annotate-data: Check memory access with two registers The following instruction pattern is used to access a global variable. mov $0x231c0, %rax movsql %edi, %rcx mov -0x7dc94ae0(,%rcx,8), %rcx cmpl $0x0, 0xa60(%rcx,%rax,1) <<<--- here The first instruction set the address of the per-cpu variable (here, it is 'runqueues' of type 'struct rq'). The second instruction seems like a cpu number of the per-cpu base. The third instruction get the base offset of per-cpu area for that cpu. The last instruction compares the value of the per-cpu variable at the offset of 0xa60. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240502060011.1838090-5-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 44 ++++++++++++++++++++++++++++++++++++----- 1 file changed, 39 insertions(+), 5 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index f1e52a531563..245e3ef3e2ff 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -1031,22 +1031,37 @@ retry: else if (has_reg_type(state, sreg) && state->regs[sreg].kind == TSR_KIND_PERCPU_BASE) { u64 ip = dloc->ms->sym->start + dl->al.offset; + u64 var_addr = src->offset; int offset; + if (src->multi_regs) { + int reg2 = (sreg == src->reg1) ? src->reg2 : src->reg1; + + if (has_reg_type(state, reg2) && state->regs[reg2].ok && + state->regs[reg2].kind == TSR_KIND_CONST) + var_addr += state->regs[reg2].imm_value; + } + /* * In kernel, %gs points to a per-cpu region for the * current CPU. Access with a constant offset should * be treated as a global variable access. */ - if (get_global_var_type(cu_die, dloc, ip, src->offset, + if (get_global_var_type(cu_die, dloc, ip, var_addr, &offset, &type_die) && die_get_member_type(&type_die, offset, &type_die)) { tsr->type = type_die; tsr->kind = TSR_KIND_TYPE; tsr->ok = true; - pr_debug_dtp("mov [%x] percpu %#x(reg%d) -> reg%d", - insn_offset, src->offset, sreg, dst->reg1); + if (src->multi_regs) { + pr_debug_dtp("mov [%x] percpu %#x(reg%d,reg%d) -> reg%d", + insn_offset, src->offset, src->reg1, + src->reg2, dst->reg1); + } else { + pr_debug_dtp("mov [%x] percpu %#x(reg%d) -> reg%d", + insn_offset, src->offset, sreg, dst->reg1); + } pr_debug_type_name(&tsr->type, tsr->kind); } else { tsr->ok = false; @@ -1340,6 +1355,17 @@ static int check_matching_type(struct type_state *state, pr_debug_dtp(" percpu var\n"); + if (dloc->op->multi_regs) { + int reg2 = dloc->op->reg2; + + if (dloc->op->reg2 == reg) + reg2 = dloc->op->reg1; + + if (has_reg_type(state, reg2) && state->regs[reg2].ok && + state->regs[reg2].kind == TSR_KIND_CONST) + var_addr += state->regs[reg2].imm_value; + } + if (get_global_var_type(cu_die, dloc, dloc->ip, var_addr, &var_offset, type_die)) { dloc->type_offset = var_offset; @@ -1527,8 +1553,16 @@ again: found = find_data_type_insn(dloc, reg, &basic_blocks, var_types, cu_die, type_die); if (found > 0) { - pr_debug_dtp("found by insn track: %#x(reg%d) type-offset=%#x\n", - dloc->op->offset, reg, dloc->type_offset); + char buf[64]; + + if (dloc->op->multi_regs) + snprintf(buf, sizeof(buf), "reg%d, reg%d", + dloc->op->reg1, dloc->op->reg2); + else + snprintf(buf, sizeof(buf), "reg%d", dloc->op->reg1); + + pr_debug_dtp("found by insn track: %#x(%s) type-offset=%#x\n", + dloc->op->offset, buf, dloc->type_offset); pr_debug_type_name(type_die, TSR_KIND_TYPE); ret = 0; break; -- cgit v1.2.3 From af89e8f2bdb2ff9252317307a755f97dd02f6cd7 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 1 May 2024 23:00:10 -0700 Subject: perf annotate-data: Handle multi regs in find_data_type_block() The instruction tracking should be the same for the both registers. Just do it once and compare the result with multi regs as with the previous patches. Then we don't need to call find_data_type_block() separately for each reg. Let's remove the 'reg' argument from the relevant functions. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240502060011.1838090-6-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 23 ++++++++++++----------- 1 file changed, 12 insertions(+), 11 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 245e3ef3e2ff..68fe7999f033 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -1258,11 +1258,12 @@ static void setup_stack_canary(struct data_loc_info *dloc) * are similar to global variables and no additional info is needed. */ static int check_matching_type(struct type_state *state, - struct data_loc_info *dloc, int reg, + struct data_loc_info *dloc, Dwarf_Die *cu_die, Dwarf_Die *type_die) { Dwarf_Word size; u32 insn_offset = dloc->ip - dloc->ms->sym->start; + int reg = dloc->op->reg1; pr_debug_dtp("chk [%x] reg%d offset=%#x ok=%d kind=%d", insn_offset, reg, dloc->op->offset, @@ -1448,7 +1449,7 @@ check_kernel: } /* Iterate instructions in basic blocks and update type table */ -static int find_data_type_insn(struct data_loc_info *dloc, int reg, +static int find_data_type_insn(struct data_loc_info *dloc, struct list_head *basic_blocks, struct die_var_type *var_types, Dwarf_Die *cu_die, Dwarf_Die *type_die) @@ -1481,7 +1482,7 @@ static int find_data_type_insn(struct data_loc_info *dloc, int reg, update_var_state(&state, dloc, addr, dl->al.offset, var_types); if (this_ip == dloc->ip) { - ret = check_matching_type(&state, dloc, reg, + ret = check_matching_type(&state, dloc, cu_die, type_die); goto out; } @@ -1502,7 +1503,7 @@ out: * Construct a list of basic blocks for each scope with variables and try to find * the data type by updating a type state table through instructions. */ -static int find_data_type_block(struct data_loc_info *dloc, int reg, +static int find_data_type_block(struct data_loc_info *dloc, Dwarf_Die *cu_die, Dwarf_Die *scopes, int nr_scopes, Dwarf_Die *type_die) { @@ -1550,7 +1551,7 @@ again: fixup_var_address(var_types, start); /* Find from start of this scope to the target instruction */ - found = find_data_type_insn(dloc, reg, &basic_blocks, var_types, + found = find_data_type_insn(dloc, &basic_blocks, var_types, cu_die, type_die); if (found > 0) { char buf[64]; @@ -1716,8 +1717,13 @@ retry: goto out; } + if (loc->multi_regs && reg == loc->reg1 && loc->reg1 != loc->reg2) { + reg = loc->reg2; + goto retry; + } + if (reg != DWARF_REG_PC) { - ret = find_data_type_block(dloc, reg, &cu_die, scopes, + ret = find_data_type_block(dloc, &cu_die, scopes, nr_scopes, type_die); if (ret == 0) { ann_data_stat.insn_track++; @@ -1725,11 +1731,6 @@ retry: } } - if (loc->multi_regs && reg == loc->reg1 && loc->reg1 != loc->reg2) { - reg = loc->reg2; - goto retry; - } - if (ret < 0) { pr_debug_dtp("no variable found\n"); ann_data_stat.no_var++; -- cgit v1.2.3 From b7d4aacfc894ca2d86b11ef738f94e6c8cf2536b Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Wed, 1 May 2024 23:00:11 -0700 Subject: perf annotate-data: Check kind of stack variables I sometimes see ("unknown type") in the result and it was because it didn't check the type of stack variables properly during the instruction tracking. The stack can carry constant values (without type info) and if the target instruction is accessing the stack location, it resulted in the "unknown type". Maybe we could pick one of integer types for the constant, but it doesn't really mean anything useful. Let's just drop the stack slot if it doesn't have a valid type info. Here's an example how it got the unknown type. Note that 0xffffff48 = -0xb8. ----------------------------------------------------------- find data type for 0xffffff48(reg6) at ... CU for ... frame base: cfa=0 fbreg=6 scope: [2/2] (die:11cb97f) bb: [37 - 3a] var [37] reg15 type='int' size=0x4 (die:0x1180633) bb: [40 - 4b] mov [40] imm=0x1 -> reg13 var [45] reg8 type='sigset_t*' size=0x8 (die:0x11a39ee) mov [45] imm=0x1 -> reg2 <--- here reg2 has a constant bb: [215 - 237] mov [218] reg2 -> -0xb8(stack) constant <--- and save it to the stack mov [225] reg13 -> -0xc4(stack) constant call [22f] find_task_by_vgpid call [22f] return -> reg0 type='struct task_struct*' size=0x8 (die:0x11881e8) bb: [5c8 - 5cf] bb: [2fb - 302] mov [2fb] -0xc4(stack) -> reg13 constant bb: [13b - 14d] mov [143] 0xd50(reg3) -> reg5 type='struct task_struct*' size=0x8 (die:0xa31f3c) bb: [153 - 153] chk [153] reg6 offset=0xffffff48 ok=0 kind=0 fbreg <--- access here found by insn track: 0xffffff48(reg6) type-offset=0 type='G^KU' size=0 (die:0xffffffffffffffff) Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240502060011.1838090-7-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 68fe7999f033..2c98813f95cd 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -1314,6 +1314,9 @@ static int check_matching_type(struct type_state *state, return -1; } + if (stack->kind != TSR_KIND_TYPE) + return 0; + *type_die = stack->type; /* Update the type offset from the start of slot */ dloc->type_offset -= stack->offset; @@ -1343,6 +1346,9 @@ static int check_matching_type(struct type_state *state, return -1; } + if (stack->kind != TSR_KIND_TYPE) + return 0; + *type_die = stack->type; /* Update the type offset from the start of slot */ dloc->type_offset -= fboff + stack->offset; -- cgit v1.2.3 From 3cdd98b42d212160d7aae746a97960a4595cbfc2 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 29 Apr 2024 15:57:38 -0700 Subject: perf maps: Remove check_invariants() from maps__lock() I found that the debug build was a slowed down a lot by the maps lock code since it checks the invariants whenever it gets the pointer to the lock. This means it checks twice the invariants before and after the access. Instead, let's move the checking code within the lock area but after any modification and remove it from the read paths. This would remove (more than) half of the maps lock overhead. The time for perf report with a huge data file (200k+ of MMAP2 events). Non-debug Before After --------- -------- -------- 2m 43s 6m 45s 4m 21s Reviewed-by: Ian Rogers Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240429225738.1491791-1-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/maps.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/maps.c b/tools/perf/util/maps.c index ce13145a9f8e..ac9fb880ddc7 100644 --- a/tools/perf/util/maps.c +++ b/tools/perf/util/maps.c @@ -211,11 +211,6 @@ void maps__set_unwind_libunwind_ops(struct maps *maps, const struct unwind_libun static struct rw_semaphore *maps__lock(struct maps *maps) { - /* - * When the lock is acquired or released the maps invariants should - * hold. - */ - check_invariants(maps); return &RC_CHK_ACCESS(maps)->lock; } @@ -358,6 +353,7 @@ static int map__strcmp(const void *a, const void *b) static int maps__sort_by_name(struct maps *maps) { int err = 0; + down_write(maps__lock(maps)); if (!maps__maps_by_name_sorted(maps)) { struct map **maps_by_name = maps__maps_by_name(maps); @@ -384,6 +380,7 @@ static int maps__sort_by_name(struct maps *maps) maps__set_maps_by_name_sorted(maps, true); } } + check_invariants(maps); up_write(maps__lock(maps)); return err; } @@ -502,6 +499,7 @@ int maps__insert(struct maps *maps, struct map *map) down_write(maps__lock(maps)); ret = __maps__insert(maps, map); + check_invariants(maps); up_write(maps__lock(maps)); return ret; } @@ -536,6 +534,7 @@ void maps__remove(struct maps *maps, struct map *map) { down_write(maps__lock(maps)); __maps__remove(maps, map); + check_invariants(maps); up_write(maps__lock(maps)); } @@ -602,6 +601,7 @@ void maps__remove_maps(struct maps *maps, bool (*cb)(struct map *map, void *data else i++; } + check_invariants(maps); up_write(maps__lock(maps)); } @@ -942,6 +942,8 @@ int maps__copy_from(struct maps *dest, struct maps *parent) map__put(new); } } + check_invariants(dest); + up_read(maps__lock(parent)); up_write(maps__lock(dest)); return err; @@ -1097,6 +1099,7 @@ void maps__fixup_end(struct maps *maps) map__set_end(maps_by_address[n - 1], ~0ULL); RC_CHK_ACCESS(maps)->ends_broken = false; + check_invariants(maps); up_write(maps__lock(maps)); } @@ -1147,6 +1150,8 @@ int maps__merge_in(struct maps *kmaps, struct map *new_map) map__start(kmaps_maps_by_address[first_after_]) >= map__end(new_map)) { /* No overlap so regular insert suffices. */ int ret = __maps__insert(kmaps, new_map); + + check_invariants(kmaps); up_write(maps__lock(kmaps)); return ret; } @@ -1184,6 +1189,7 @@ int maps__merge_in(struct maps *kmaps, struct map *new_map) map__zput(kmaps_maps_by_address[i]); free(kmaps_maps_by_address); + check_invariants(kmaps); up_write(maps__lock(kmaps)); return 0; } -- cgit v1.2.3 From aa1551f299ba414c07edb08ebb4f46eb46006345 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 2 May 2024 14:35:04 -0700 Subject: perf test pmu: Refactor format test and exposed test APIs In tests/pmu.c, make a common utility that creates a PMU in a mkdtemp directory and uses regular PMU parsing logic to load that PMU. Formats must still be eagerly loaded as by default the PMU code assumes devices are going to be in sysfs. In util/pmu.[ch], hide perf_pmu__format_parse but add the eager argument to perf_pmu__lookup called by perf_pmus__add_test_pmu. Later patches will eagerly load other non-sysfs files when eager loading is enabled. In tests/pmu.c, rather than manually constructing a list of term arguments, just use the term parsing code from a string. Add more comments and debug logging. Reviewed-by: Kan Liang Signed-off-by: Ian Rogers Cc: Alexander Shishkin Cc: Bjorn Helgaas Cc: Ingo Molnar Cc: James Clark Cc: Jing Zhang Cc: Jiri Olsa Cc: Jonathan Corbet Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Randy Dunlap Cc: Ravi Bangoria Cc: Thomas Richter Link: https://lore.kernel.org/r/20240502213507.2339733-4-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/tests/pmu.c | 319 +++++++++++++++++++---------------------- tools/perf/util/parse-events.c | 2 +- tools/perf/util/parse-events.h | 2 +- tools/perf/util/pmu.c | 11 +- tools/perf/util/pmu.h | 4 +- tools/perf/util/pmus.c | 16 ++- tools/perf/util/pmus.h | 2 + 7 files changed, 177 insertions(+), 179 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/tests/pmu.c b/tools/perf/tests/pmu.c index 8f18127d876a..424ebdb0f09d 100644 --- a/tools/perf/tests/pmu.c +++ b/tools/perf/tests/pmu.c @@ -2,203 +2,186 @@ #include "parse-events.h" #include "pmu.h" #include "tests.h" +#include "debug.h" #include #include #include -#include -#include -#include - -/* Simulated format definitions. */ -static struct test_format { - const char *name; - const char *value; -} test_formats[] = { - { "krava01", "config:0-1,62-63\n", }, - { "krava02", "config:10-17\n", }, - { "krava03", "config:5\n", }, - { "krava11", "config1:0,2,4,6,8,20-28\n", }, - { "krava12", "config1:63\n", }, - { "krava13", "config1:45-47\n", }, - { "krava21", "config2:0-3,10-13,20-23,30-33,40-43,50-53,60-63\n", }, - { "krava22", "config2:8,18,48,58\n", }, - { "krava23", "config2:28-29,38\n", }, -}; - -/* Simulated users input. */ -static struct parse_events_term test_terms[] = { - { - .config = "krava01", - .val.num = 15, - .type_val = PARSE_EVENTS__TERM_TYPE_NUM, - .type_term = PARSE_EVENTS__TERM_TYPE_USER, - }, - { - .config = "krava02", - .val.num = 170, - .type_val = PARSE_EVENTS__TERM_TYPE_NUM, - .type_term = PARSE_EVENTS__TERM_TYPE_USER, - }, - { - .config = "krava03", - .val.num = 1, - .type_val = PARSE_EVENTS__TERM_TYPE_NUM, - .type_term = PARSE_EVENTS__TERM_TYPE_USER, - }, - { - .config = "krava11", - .val.num = 27, - .type_val = PARSE_EVENTS__TERM_TYPE_NUM, - .type_term = PARSE_EVENTS__TERM_TYPE_USER, - }, - { - .config = "krava12", - .val.num = 1, - .type_val = PARSE_EVENTS__TERM_TYPE_NUM, - .type_term = PARSE_EVENTS__TERM_TYPE_USER, - }, - { - .config = "krava13", - .val.num = 2, - .type_val = PARSE_EVENTS__TERM_TYPE_NUM, - .type_term = PARSE_EVENTS__TERM_TYPE_USER, - }, - { - .config = "krava21", - .val.num = 119, - .type_val = PARSE_EVENTS__TERM_TYPE_NUM, - .type_term = PARSE_EVENTS__TERM_TYPE_USER, - }, - { - .config = "krava22", - .val.num = 11, - .type_val = PARSE_EVENTS__TERM_TYPE_NUM, - .type_term = PARSE_EVENTS__TERM_TYPE_USER, - }, - { - .config = "krava23", - .val.num = 2, - .type_val = PARSE_EVENTS__TERM_TYPE_NUM, - .type_term = PARSE_EVENTS__TERM_TYPE_USER, - }, -}; - -/* - * Prepare format directory data, exported by kernel - * at /sys/bus/event_source/devices//format. - */ -static char *test_format_dir_get(char *dir, size_t sz) -{ - unsigned int i; - - snprintf(dir, sz, "/tmp/perf-pmu-test-format-XXXXXX"); - if (!mkdtemp(dir)) - return NULL; - - for (i = 0; i < ARRAY_SIZE(test_formats); i++) { - char name[PATH_MAX]; - struct test_format *format = &test_formats[i]; - FILE *file; - - scnprintf(name, PATH_MAX, "%s/%s", dir, format->name); - - file = fopen(name, "w"); - if (!file) - return NULL; - - if (1 != fwrite(format->value, strlen(format->value), 1, file)) - break; +#include +#include +#include - fclose(file); - } - - return dir; -} +/* Fake PMUs created in temp directory. */ +static LIST_HEAD(test_pmus); -/* Cleanup format directory. */ -static int test_format_dir_put(char *dir) +/* Cleanup test PMU directory. */ +static int test_pmu_put(const char *dir, struct perf_pmu *pmu) { char buf[PATH_MAX + 20]; + int ret; - snprintf(buf, sizeof(buf), "rm -f %s/*\n", dir); - if (system(buf)) - return -1; + if (scnprintf(buf, sizeof(buf), "rm -fr %s", dir) < 0) { + pr_err("Failure to set up buffer for \"%s\"\n", dir); + return -EINVAL; + } + ret = system(buf); + if (ret) + pr_err("Failure to \"%s\"\n", buf); - snprintf(buf, sizeof(buf), "rmdir %s\n", dir); - return system(buf); + list_del(&pmu->list); + perf_pmu__delete(pmu); + return ret; } -static void add_test_terms(struct parse_events_terms *terms) +/* + * Prepare test PMU directory data, normally exported by kernel at + * /sys/bus/event_source/devices//. Give as input a buffer to hold the file + * path, the result is PMU loaded using that directory. + */ +static struct perf_pmu *test_pmu_get(char *dir, size_t sz) { - unsigned int i; + /* Simulated format definitions. */ + const struct test_format { + const char *name; + const char *value; + } test_formats[] = { + { "krava01", "config:0-1,62-63\n", }, + { "krava02", "config:10-17\n", }, + { "krava03", "config:5\n", }, + { "krava11", "config1:0,2,4,6,8,20-28\n", }, + { "krava12", "config1:63\n", }, + { "krava13", "config1:45-47\n", }, + { "krava21", "config2:0-3,10-13,20-23,30-33,40-43,50-53,60-63\n", }, + { "krava22", "config2:8,18,48,58\n", }, + { "krava23", "config2:28-29,38\n", }, + }; + char name[PATH_MAX]; + int dirfd, file; + struct perf_pmu *pmu = NULL; + ssize_t len; + + /* Create equivalent of sysfs mount point. */ + scnprintf(dir, sz, "/tmp/perf-pmu-test-XXXXXX"); + if (!mkdtemp(dir)) { + pr_err("mkdtemp failed\n"); + dir[0] = '\0'; + return NULL; + } + dirfd = open(dir, O_DIRECTORY); + if (dirfd < 0) { + pr_err("Failed to open test directory \"%s\"\n", dir); + goto err_out; + } - for (i = 0; i < ARRAY_SIZE(test_terms); i++) { - struct parse_events_term *clone; + /* Create the test PMU directory and give it a perf_event_attr type number. */ + if (mkdirat(dirfd, "perf-pmu-test", 0755) < 0) { + pr_err("Failed to mkdir PMU directory\n"); + goto err_out; + } + file = openat(dirfd, "perf-pmu-test/type", O_WRONLY | O_CREAT, 0600); + if (!file) { + pr_err("Failed to open for writing file \"type\"\n"); + goto err_out; + } + len = strlen("9999"); + if (write(file, "9999\n", len) < len) { + close(file); + pr_err("Failed to write to 'type' file\n"); + goto err_out; + } + close(file); - parse_events_term__clone(&clone, &test_terms[i]); - list_add_tail(&clone->list, &terms->terms); + /* Create format directory and files. */ + if (mkdirat(dirfd, "perf-pmu-test/format", 0755) < 0) { + pr_err("Failed to mkdir PMU format directory\n)"); + goto err_out; } + for (size_t i = 0; i < ARRAY_SIZE(test_formats); i++) { + const struct test_format *format = &test_formats[i]; + + if (scnprintf(name, PATH_MAX, "perf-pmu-test/format/%s", format->name) < 0) { + pr_err("Failure to set up path for \"%s\"\n", format->name); + goto err_out; + } + file = openat(dirfd, name, O_WRONLY | O_CREAT, 0600); + if (!file) { + pr_err("Failed to open for writing file \"%s\"\n", name); + goto err_out; + } + + if (write(file, format->value, strlen(format->value)) < 0) { + pr_err("Failed to write to file \"%s\"\n", name); + close(file); + goto err_out; + } + close(file); + } + + /* Make the PMU reading the files created above. */ + pmu = perf_pmus__add_test_pmu(dirfd, "perf-pmu-test"); + if (!pmu) + pr_err("Test PMU creation failed\n"); + +err_out: + if (!pmu) + test_pmu_put(dir, pmu); + if (dirfd >= 0) + close(dirfd); + return pmu; } -static int test__pmu(struct test_suite *test __maybe_unused, int subtest __maybe_unused) +static int test__pmu_format(struct test_suite *test __maybe_unused, int subtest __maybe_unused) { char dir[PATH_MAX]; - char *format; - struct parse_events_terms terms; struct perf_event_attr attr; - struct perf_pmu *pmu; - int fd; - int ret; + struct parse_events_terms terms; + int ret = TEST_FAIL; + struct perf_pmu *pmu = test_pmu_get(dir, sizeof(dir)); - parse_events_terms__init(&terms); - add_test_terms(&terms); - pmu = zalloc(sizeof(*pmu)); - if (!pmu) { - parse_events_terms__exit(&terms); - return -ENOMEM; - } + if (!pmu) + return TEST_FAIL; - INIT_LIST_HEAD(&pmu->format); - INIT_LIST_HEAD(&pmu->aliases); - INIT_LIST_HEAD(&pmu->caps); - format = test_format_dir_get(dir, sizeof(dir)); - if (!format) { - free(pmu); - parse_events_terms__exit(&terms); - return -EINVAL; + parse_events_terms__init(&terms); + if (parse_events_terms(&terms, + "krava01=15,krava02=170,krava03=1,krava11=27,krava12=1," + "krava13=2,krava21=119,krava22=11,krava23=2", + NULL)) { + pr_err("Term parsing failed\n"); + goto err_out; } memset(&attr, 0, sizeof(attr)); - - fd = open(format, O_DIRECTORY); - if (fd < 0) { - ret = fd; - goto out; + ret = perf_pmu__config_terms(pmu, &attr, &terms, /*zero=*/false, /*err=*/NULL); + if (ret) { + pr_err("perf_pmu__config_terms failed"); + goto err_out; } - pmu->name = strdup("perf-pmu-test"); - ret = perf_pmu__format_parse(pmu, fd, /*eager_load=*/true); - if (ret) - goto out; + if (attr.config != 0xc00000000002a823) { + pr_err("Unexpected config value %llx\n", attr.config); + goto err_out; + } + if (attr.config1 != 0x8000400000000145) { + pr_err("Unexpected config1 value %llx\n", attr.config1); + goto err_out; + } + if (attr.config2 != 0x0400000020041d07) { + pr_err("Unexpected config2 value %llx\n", attr.config2); + goto err_out; + } - ret = perf_pmu__config_terms(pmu, &attr, &terms, /*zero=*/false, /*err=*/NULL); - if (ret) - goto out; - - ret = -EINVAL; - if (attr.config != 0xc00000000002a823) - goto out; - if (attr.config1 != 0x8000400000000145) - goto out; - if (attr.config2 != 0x0400000020041d07) - goto out; - - ret = 0; -out: - test_format_dir_put(format); - perf_pmu__delete(pmu); + ret = TEST_OK; +err_out: parse_events_terms__exit(&terms); + test_pmu_put(dir, pmu); return ret; } -DEFINE_SUITE("Parse perf pmu format", pmu); +static struct test_case tests__pmu[] = { + TEST_CASE("Parsing with PMU format directory", pmu_format), + { .name = NULL, } +}; + +struct test_suite suite__pmu = { + .desc = "Sysfs PMU tests", + .test_cases = tests__pmu, +}; diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 0f308b4db2b9..2b9ede311c31 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -2585,7 +2585,7 @@ int parse_events_term__term(struct parse_events_term **term, } int parse_events_term__clone(struct parse_events_term **new, - struct parse_events_term *term) + const struct parse_events_term *term) { char *str; struct parse_events_term temp = *term; diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index 5695308efab9..e7ac1f13376d 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -178,7 +178,7 @@ int parse_events_term__term(struct parse_events_term **term, enum parse_events__term_type term_rhs, void *loc_term, void *loc_val); int parse_events_term__clone(struct parse_events_term **new, - struct parse_events_term *term); + const struct parse_events_term *term); void parse_events_term__delete(struct parse_events_term *term); void parse_events_terms__delete(struct parse_events_terms *terms); diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index 74dd5bd49d9a..fbbc535ed93f 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -182,7 +182,7 @@ static void perf_pmu_format__load(const struct perf_pmu *pmu, struct perf_pmu_fo * Parse & process all the sysfs attributes located under * the directory specified in 'dir' parameter. */ -int perf_pmu__format_parse(struct perf_pmu *pmu, int dirfd, bool eager_load) +static int perf_pmu__format_parse(struct perf_pmu *pmu, int dirfd, bool eager_load) { struct dirent *evt_ent; DIR *format_dir; @@ -232,7 +232,7 @@ int perf_pmu__format_parse(struct perf_pmu *pmu, int dirfd, bool eager_load) * located at: * /sys/bus/event_source/devices//format as sysfs group attributes. */ -static int pmu_format(struct perf_pmu *pmu, int dirfd, const char *name) +static int pmu_format(struct perf_pmu *pmu, int dirfd, const char *name, bool eager_load) { int fd; @@ -241,7 +241,7 @@ static int pmu_format(struct perf_pmu *pmu, int dirfd, const char *name) return 0; /* it'll close the fd */ - if (perf_pmu__format_parse(pmu, fd, /*eager_load=*/false)) + if (perf_pmu__format_parse(pmu, fd, eager_load)) return -1; return 0; @@ -994,7 +994,8 @@ perf_pmu__arch_init(struct perf_pmu *pmu) pmu->mem_events = perf_mem_events; } -struct perf_pmu *perf_pmu__lookup(struct list_head *pmus, int dirfd, const char *name) +struct perf_pmu *perf_pmu__lookup(struct list_head *pmus, int dirfd, const char *name, + bool eager_load) { struct perf_pmu *pmu; __u32 type; @@ -1023,7 +1024,7 @@ struct perf_pmu *perf_pmu__lookup(struct list_head *pmus, int dirfd, const char * type value and format definitions. Load both right * now. */ - if (pmu_format(pmu, dirfd, name)) + if (pmu_format(pmu, dirfd, name, eager_load)) goto err; pmu->is_core = is_pmu_core(name); diff --git a/tools/perf/util/pmu.h b/tools/perf/util/pmu.h index 93d03bd3ecbe..561716aa2b25 100644 --- a/tools/perf/util/pmu.h +++ b/tools/perf/util/pmu.h @@ -214,7 +214,6 @@ int perf_pmu__check_alias(struct perf_pmu *pmu, struct parse_events_terms *head_ struct parse_events_error *err); int perf_pmu__find_event(struct perf_pmu *pmu, const char *event, void *state, pmu_event_callback cb); -int perf_pmu__format_parse(struct perf_pmu *pmu, int dirfd, bool eager_load); void perf_pmu_format__set_value(void *format, int config, unsigned long *bits); bool perf_pmu__has_format(const struct perf_pmu *pmu, const char *name); int perf_pmu__for_each_format(struct perf_pmu *pmu, void *state, pmu_format_callback cb); @@ -272,7 +271,8 @@ int perf_pmu__pathname_scnprintf(char *buf, size_t size, int perf_pmu__event_source_devices_fd(void); int perf_pmu__pathname_fd(int dirfd, const char *pmu_name, const char *filename, int flags); -struct perf_pmu *perf_pmu__lookup(struct list_head *pmus, int dirfd, const char *lookup_name); +struct perf_pmu *perf_pmu__lookup(struct list_head *pmus, int dirfd, const char *lookup_name, + bool eager_load); struct perf_pmu *perf_pmu__create_placeholder_core_pmu(struct list_head *core_pmus); void perf_pmu__delete(struct perf_pmu *pmu); struct perf_pmu *perf_pmus__find_core_pmu(void); diff --git a/tools/perf/util/pmus.c b/tools/perf/util/pmus.c index 2fd369e45832..b9b4c5eb5002 100644 --- a/tools/perf/util/pmus.c +++ b/tools/perf/util/pmus.c @@ -124,7 +124,8 @@ struct perf_pmu *perf_pmus__find(const char *name) return NULL; dirfd = perf_pmu__event_source_devices_fd(); - pmu = perf_pmu__lookup(core_pmu ? &core_pmus : &other_pmus, dirfd, name); + pmu = perf_pmu__lookup(core_pmu ? &core_pmus : &other_pmus, dirfd, name, + /*eager_load=*/false); close(dirfd); if (!pmu) { @@ -159,7 +160,8 @@ static struct perf_pmu *perf_pmu__find2(int dirfd, const char *name) if (core_pmu && read_sysfs_core_pmus) return NULL; - return perf_pmu__lookup(core_pmu ? &core_pmus : &other_pmus, dirfd, name); + return perf_pmu__lookup(core_pmu ? &core_pmus : &other_pmus, dirfd, name, + /*eager_load=*/false); } static int pmus_cmp(void *priv __maybe_unused, @@ -696,3 +698,13 @@ struct perf_pmu *perf_pmus__find_core_pmu(void) { return perf_pmus__scan_core(NULL); } + +struct perf_pmu *perf_pmus__add_test_pmu(int test_sysfs_dirfd, const char *name) +{ + /* + * Some PMU functions read from the sysfs mount point, so care is + * needed, hence passing the eager_load flag to load things like the + * format files. + */ + return perf_pmu__lookup(&other_pmus, test_sysfs_dirfd, name, /*eager_load=*/true); +} diff --git a/tools/perf/util/pmus.h b/tools/perf/util/pmus.h index eec599d8aebd..9d4ded80b8e9 100644 --- a/tools/perf/util/pmus.h +++ b/tools/perf/util/pmus.h @@ -24,4 +24,6 @@ int perf_pmus__num_core_pmus(void); bool perf_pmus__supports_extended_type(void); char *perf_pmus__default_pmu_name(void); +struct perf_pmu *perf_pmus__add_test_pmu(int test_sysfs_dirfd, const char *name); + #endif /* __PMUS_H */ -- cgit v1.2.3 From 18eb2ca8c18f0612c15aa12375e7cba29e97ab1a Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 2 May 2024 14:35:05 -0700 Subject: perf test pmu: Add an eagerly loaded event test Allow events/aliases to be eagerly loaded for a PMU. Factor out the pmu_aliases_parse to allow this. Parse a test event and check it configures the attribute as expected. There is overlap with the parse-events tests, but this test is done with a PMU created in a temp directory and doesn't rely on PMUs in sysfs. Reviewed-by: Kan Liang Signed-off-by: Ian Rogers Cc: Alexander Shishkin Cc: Bjorn Helgaas Cc: Ingo Molnar Cc: James Clark Cc: Jing Zhang Cc: Jiri Olsa Cc: Jonathan Corbet Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Randy Dunlap Cc: Ravi Bangoria Cc: Thomas Richter Link: https://lore.kernel.org/r/20240502213507.2339733-5-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/tests/pmu.c | 76 ++++++++++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/pmu.c | 69 +++++++++++++++++++++++++++++++-------------- 2 files changed, 124 insertions(+), 21 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/tests/pmu.c b/tools/perf/tests/pmu.c index 424ebdb0f09d..071cfc51b52c 100644 --- a/tools/perf/tests/pmu.c +++ b/tools/perf/tests/pmu.c @@ -1,4 +1,6 @@ // SPDX-License-Identifier: GPL-2.0 +#include "evlist.h" +#include "evsel.h" #include "parse-events.h" #include "pmu.h" #include "tests.h" @@ -54,6 +56,9 @@ static struct perf_pmu *test_pmu_get(char *dir, size_t sz) { "krava22", "config2:8,18,48,58\n", }, { "krava23", "config2:28-29,38\n", }, }; + const char *test_event = "krava01=15,krava02=170,krava03=1,krava11=27,krava12=1," + "krava13=2,krava21=119,krava22=11,krava23=2\n"; + char name[PATH_MAX]; int dirfd, file; struct perf_pmu *pmu = NULL; @@ -116,6 +121,24 @@ static struct perf_pmu *test_pmu_get(char *dir, size_t sz) close(file); } + /* Create test event. */ + if (mkdirat(dirfd, "perf-pmu-test/events", 0755) < 0) { + pr_err("Failed to mkdir PMU events directory\n"); + goto err_out; + } + file = openat(dirfd, "perf-pmu-test/events/test-event", O_WRONLY | O_CREAT, 0600); + if (!file) { + pr_err("Failed to open for writing file \"type\"\n"); + goto err_out; + } + len = strlen(test_event); + if (write(file, test_event, len) < len) { + close(file); + pr_err("Failed to write to 'test-event' file\n"); + goto err_out; + } + close(file); + /* Make the PMU reading the files created above. */ pmu = perf_pmus__add_test_pmu(dirfd, "perf-pmu-test"); if (!pmu) @@ -176,8 +199,61 @@ err_out: return ret; } +static int test__pmu_events(struct test_suite *test __maybe_unused, int subtest __maybe_unused) +{ + char dir[PATH_MAX]; + struct parse_events_error err; + struct evlist *evlist; + struct evsel *evsel; + struct perf_event_attr *attr; + int ret = TEST_FAIL; + struct perf_pmu *pmu = test_pmu_get(dir, sizeof(dir)); + const char *event = "perf-pmu-test/test-event/"; + + + if (!pmu) + return TEST_FAIL; + + evlist = evlist__new(); + if (evlist == NULL) { + pr_err("Failed allocation"); + goto err_out; + } + parse_events_error__init(&err); + ret = parse_events(evlist, event, &err); + if (ret) { + pr_debug("failed to parse event '%s', err %d\n", event, ret); + parse_events_error__print(&err, event); + if (parse_events_error__contains(&err, "can't access trace events")) + ret = TEST_SKIP; + goto err_out; + } + evsel = evlist__first(evlist); + attr = &evsel->core.attr; + if (attr->config != 0xc00000000002a823) { + pr_err("Unexpected config value %llx\n", attr->config); + goto err_out; + } + if (attr->config1 != 0x8000400000000145) { + pr_err("Unexpected config1 value %llx\n", attr->config1); + goto err_out; + } + if (attr->config2 != 0x0400000020041d07) { + pr_err("Unexpected config2 value %llx\n", attr->config2); + goto err_out; + } + + ret = TEST_OK; +err_out: + parse_events_error__exit(&err); + evlist__delete(evlist); + test_pmu_put(dir, pmu); + return ret; +} + static struct test_case tests__pmu[] = { TEST_CASE("Parsing with PMU format directory", pmu_format), + TEST_CASE("Parsing with PMU event", pmu_events), { .name = NULL, } }; diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index fbbc535ed93f..7849be4bfea1 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -597,33 +597,18 @@ static inline bool pmu_alias_info_file(const char *name) * Reading the pmu event aliases definition, which should be located at: * /sys/bus/event_source/devices//events as sysfs group attributes. */ -static int pmu_aliases_parse(struct perf_pmu *pmu) +static int __pmu_aliases_parse(struct perf_pmu *pmu, int events_dir_fd) { - char path[PATH_MAX]; struct dirent *evt_ent; DIR *event_dir; - size_t len; - int fd, dir_fd; - len = perf_pmu__event_source_devices_scnprintf(path, sizeof(path)); - if (!len) - return 0; - scnprintf(path + len, sizeof(path) - len, "%s/events", pmu->name); - - dir_fd = open(path, O_DIRECTORY); - if (dir_fd == -1) { - pmu->sysfs_aliases_loaded = true; - return 0; - } - - event_dir = fdopendir(dir_fd); - if (!event_dir){ - close (dir_fd); + event_dir = fdopendir(events_dir_fd); + if (!event_dir) return -EINVAL; - } while ((evt_ent = readdir(event_dir))) { char *name = evt_ent->d_name; + int fd; FILE *file; if (!strcmp(name, ".") || !strcmp(name, "..")) @@ -635,7 +620,7 @@ static int pmu_aliases_parse(struct perf_pmu *pmu) if (pmu_alias_info_file(name)) continue; - fd = openat(dir_fd, name, O_RDONLY); + fd = openat(events_dir_fd, name, O_RDONLY); if (fd == -1) { pr_debug("Cannot open %s\n", name); continue; @@ -653,11 +638,50 @@ static int pmu_aliases_parse(struct perf_pmu *pmu) } closedir(event_dir); - close (dir_fd); pmu->sysfs_aliases_loaded = true; return 0; } +static int pmu_aliases_parse(struct perf_pmu *pmu) +{ + char path[PATH_MAX]; + size_t len; + int events_dir_fd, ret; + + if (pmu->sysfs_aliases_loaded) + return 0; + + len = perf_pmu__event_source_devices_scnprintf(path, sizeof(path)); + if (!len) + return 0; + scnprintf(path + len, sizeof(path) - len, "%s/events", pmu->name); + + events_dir_fd = open(path, O_DIRECTORY); + if (events_dir_fd == -1) { + pmu->sysfs_aliases_loaded = true; + return 0; + } + ret = __pmu_aliases_parse(pmu, events_dir_fd); + close(events_dir_fd); + return ret; +} + +static int pmu_aliases_parse_eager(struct perf_pmu *pmu, int sysfs_fd) +{ + char path[FILENAME_MAX + 7]; + int ret, events_dir_fd; + + scnprintf(path, sizeof(path), "%s/events", pmu->name); + events_dir_fd = openat(sysfs_fd, path, O_DIRECTORY, 0); + if (events_dir_fd == -1) { + pmu->sysfs_aliases_loaded = true; + return 0; + } + ret = __pmu_aliases_parse(pmu, events_dir_fd); + close(events_dir_fd); + return ret; +} + static int pmu_alias_terms(struct perf_pmu_alias *alias, int err_loc, struct list_head *terms) { struct parse_events_term *term, *cloned; @@ -1042,6 +1066,9 @@ struct perf_pmu *perf_pmu__lookup(struct list_head *pmus, int dirfd, const char perf_pmu__arch_init(pmu); + if (eager_load) + pmu_aliases_parse_eager(pmu, dirfd); + return pmu; err: zfree(&pmu->name); -- cgit v1.2.3 From 7b6dd7a923281a7ccb980a0f768d6926721eb3cc Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 2 May 2024 14:35:07 -0700 Subject: perf pmu: Assume sysfs events are always the same case Perf event names aren't case sensitive. For sysfs events the entire directory of events is read then iterated comparing names in a case insensitive way, most often to see if an event is present. Consider: $ perf stat -e inst_retired.any true The event inst_retired.any may be present in any PMU, so every PMU's sysfs events are loaded and then searched with strcasecmp to see if any match. This event is only present on the cpu PMU as a JSON event so a lot of events were loaded from sysfs unnecessarily just to prove an event didn't exist there. This change avoids loading all the events by assuming sysfs event names are always either lower or uppercase. It uses file exists and only loads the events when the desired event is present. For the example above, the number of openat calls measured by 'perf trace' on a tigerlake laptop goes from 325 down to 255. The reduction will be larger for machines with many PMUs, particularly replicated uncore PMUs. Ensure pmu_aliases_parse() is called before all uses of the aliases list, but remove some "pmu->sysfs_aliases_loaded" tests as they are now part of the function. Reviewed-by: Kan Liang Signed-off-by: Ian Rogers Cc: Alexander Shishkin Cc: Bjorn Helgaas Cc: Ingo Molnar Cc: James Clark Cc: Jing Zhang Cc: Jiri Olsa Cc: Jonathan Corbet Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Randy Dunlap Cc: Ravi Bangoria Cc: Thomas Richter Link: https://lore.kernel.org/r/20240502213507.2339733-7-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/pmu.c | 31 ++++++++++++++++++++++++++----- 1 file changed, 26 insertions(+), 5 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index 7849be4bfea1..b3b072feef02 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -425,9 +425,30 @@ static struct perf_pmu_alias *perf_pmu__find_alias(struct perf_pmu *pmu, { struct perf_pmu_alias *alias; - if (load && !pmu->sysfs_aliases_loaded) - pmu_aliases_parse(pmu); + if (load && !pmu->sysfs_aliases_loaded) { + bool has_sysfs_event; + char event_file_name[FILENAME_MAX + 8]; + /* + * Test if alias/event 'name' exists in the PMU's sysfs/events + * directory. If not skip parsing the sysfs aliases. Sysfs event + * name must be all lower or all upper case. + */ + scnprintf(event_file_name, sizeof(event_file_name), "events/%s", name); + for (size_t i = 7, n = 7 + strlen(name); i < n; i++) + event_file_name[i] = tolower(event_file_name[i]); + + has_sysfs_event = perf_pmu__file_exists(pmu, event_file_name); + if (!has_sysfs_event) { + for (size_t i = 7, n = 7 + strlen(name); i < n; i++) + event_file_name[i] = toupper(event_file_name[i]); + + has_sysfs_event = perf_pmu__file_exists(pmu, event_file_name); + } + if (has_sysfs_event) + pmu_aliases_parse(pmu); + + } list_for_each_entry(alias, &pmu->aliases, list) { if (!strcasecmp(alias->name, name)) return alias; @@ -1717,9 +1738,7 @@ size_t perf_pmu__num_events(struct perf_pmu *pmu) { size_t nr; - if (!pmu->sysfs_aliases_loaded) - pmu_aliases_parse(pmu); - + pmu_aliases_parse(pmu); nr = pmu->sysfs_aliases; if (pmu->cpu_aliases_added) @@ -1778,6 +1797,7 @@ int perf_pmu__for_each_event(struct perf_pmu *pmu, bool skip_duplicate_pmus, struct strbuf sb; strbuf_init(&sb, /*hint=*/ 0); + pmu_aliases_parse(pmu); pmu_add_cpu_aliases(pmu); list_for_each_entry(event, &pmu->aliases, list) { size_t buf_used; @@ -2193,6 +2213,7 @@ const char *perf_pmu__name_from_config(struct perf_pmu *pmu, u64 config) if (!pmu) return NULL; + pmu_aliases_parse(pmu); pmu_add_cpu_aliases(pmu); list_for_each_entry(event, &pmu->aliases, list) { struct perf_event_attr attr = {.config = 0,}; -- cgit v1.2.3 From 3f4ac23a990853ab5012037767281dfd4beb4b15 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Sat, 4 May 2024 14:37:57 -0700 Subject: perf dsos: Switch backing storage to array from rbtree/list MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit DSOs were held on a list for fast iteration and in an rbtree for fast finds. Switch to using a lazily sorted array where iteration is just iterating through the array and binary searches are the same complexity as searching the rbtree. The find may need to sort the array first which does increase the complexity, but add operations have lower complexity and overall the complexity should remain about the same. The set name operations on the dso just records that the array is no longer sorted, avoiding complexity in rebalancing the rbtree. Tighter locking discipline is enforced to avoid the array being resorted while long and short names or ids are changed. The array is smaller in size, replacing 6 pointers with 2, and so even with extra allocated space in the array, the array may be 50% unoccupied, the memory saving should be at least 2x. Committer testing: On a previous version of this patchset we were getting a lot of warnings about deleting a DSO still on a list, now it is ok: root@x1:~# perf probe -l root@x1:~# perf probe finish_task_switch Added new event: probe:finish_task_switch (on finish_task_switch) You can now use it in all perf tools, such as: perf record -e probe:finish_task_switch -aR sleep 1 root@x1:~# perf probe -l probe:finish_task_switch (on finish_task_switch@kernel/sched/core.c) root@x1:~# perf trace -e probe:finish_task_switch/max-stack=8/ --max-events=1 0.000 migration/0/19 probe:finish_task_switch(__probe_ip: -1894408688) finish_task_switch.isra.0 ([kernel.kallsyms]) __schedule ([kernel.kallsyms]) schedule ([kernel.kallsyms]) smpboot_thread_fn ([kernel.kallsyms]) kthread ([kernel.kallsyms]) ret_from_fork ([kernel.kallsyms]) ret_from_fork_asm ([kernel.kallsyms]) root@x1:~# root@x1:~# perf probe -d probe:* Removed event: probe:finish_task_switch root@x1:~# perf probe -l root@x1:~# I also ran the full 'perf test' suite after applying this one, no regressions. Signed-off-by: Ian Rogers Tested-by: Arnaldo Carvalho de Melo Cc: Adrian Hunter Cc: Ahelenia ZiemiaÅ„ska Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Ben Gainey Cc: Changbin Du Cc: Chengen Du Cc: Colin Ian King Cc: Dima Kogan Cc: Ilkka Koskinen Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kan Liang Cc: Leo Yan Cc: Li Dong Cc: Mark Rutland Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Paran Lee Cc: Peter Zijlstra Cc: Song Liu Cc: Sun Haiyong Cc: Thomas Richter Cc: Tiezhu Yang Cc: Yanteng Si Cc: zhaimingbing Link: https://lore.kernel.org/r/20240504213803.218974-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dso.c | 67 +++++++++++------- tools/perf/util/dso.h | 10 ++- tools/perf/util/dsos.c | 188 +++++++++++++++++++++++++++++++------------------ tools/perf/util/dsos.h | 21 +++--- 4 files changed, 177 insertions(+), 109 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c index ad562743d769..3caca60a6ce3 100644 --- a/tools/perf/util/dso.c +++ b/tools/perf/util/dso.c @@ -1241,35 +1241,35 @@ struct dso *machine__findnew_kernel(struct machine *machine, const char *name, return dso; } -static void dso__set_long_name_id(struct dso *dso, const char *name, struct dso_id *id, bool name_allocated) +static void dso__set_long_name_id(struct dso *dso, const char *name, bool name_allocated) { - struct rb_root *root = dso->root; + struct dsos *dsos = dso->dsos; if (name == NULL) return; - if (dso->long_name_allocated) - free((char *)dso->long_name); - - if (root) { - rb_erase(&dso->rb_node, root); + if (dsos) { /* - * __dsos__findnew_link_by_longname_id() isn't guaranteed to - * add it back, so a clean removal is required here. + * Need to avoid re-sorting the dsos breaking by non-atomically + * renaming the dso. */ - RB_CLEAR_NODE(&dso->rb_node); - dso->root = NULL; + down_write(&dsos->lock); } + if (dso->long_name_allocated) + free((char *)dso->long_name); + dso->long_name = name; dso->long_name_len = strlen(name); dso->long_name_allocated = name_allocated; - if (root) - __dsos__findnew_link_by_longname_id(root, dso, NULL, id); + if (dsos) { + dsos->sorted = false; + up_write(&dsos->lock); + } } -static int __dso_id__cmp(struct dso_id *a, struct dso_id *b) +static int __dso_id__cmp(const struct dso_id *a, const struct dso_id *b) { if (a->maj > b->maj) return -1; if (a->maj < b->maj) return 1; @@ -1297,7 +1297,7 @@ static int __dso_id__cmp(struct dso_id *a, struct dso_id *b) return 0; } -bool dso_id__empty(struct dso_id *id) +bool dso_id__empty(const struct dso_id *id) { if (!id) return true; @@ -1305,15 +1305,22 @@ bool dso_id__empty(struct dso_id *id) return !id->maj && !id->min && !id->ino && !id->ino_generation; } -void dso__inject_id(struct dso *dso, struct dso_id *id) +void __dso__inject_id(struct dso *dso, struct dso_id *id) { + struct dsos *dsos = dso->dsos; + + /* dsos write lock held by caller. */ + dso->id.maj = id->maj; dso->id.min = id->min; dso->id.ino = id->ino; dso->id.ino_generation = id->ino_generation; + + if (dsos) + dsos->sorted = false; } -int dso_id__cmp(struct dso_id *a, struct dso_id *b) +int dso_id__cmp(const struct dso_id *a, const struct dso_id *b) { /* * The second is always dso->id, so zeroes if not set, assume passing @@ -1332,20 +1339,34 @@ int dso__cmp_id(struct dso *a, struct dso *b) void dso__set_long_name(struct dso *dso, const char *name, bool name_allocated) { - dso__set_long_name_id(dso, name, NULL, name_allocated); + dso__set_long_name_id(dso, name, name_allocated); } void dso__set_short_name(struct dso *dso, const char *name, bool name_allocated) { + struct dsos *dsos = dso->dsos; + if (name == NULL) return; + if (dsos) { + /* + * Need to avoid re-sorting the dsos breaking by non-atomically + * renaming the dso. + */ + down_write(&dsos->lock); + } if (dso->short_name_allocated) free((char *)dso->short_name); dso->short_name = name; dso->short_name_len = strlen(name); dso->short_name_allocated = name_allocated; + + if (dsos) { + dsos->sorted = false; + up_write(&dsos->lock); + } } int dso__name_len(const struct dso *dso) @@ -1381,7 +1402,7 @@ struct dso *dso__new_id(const char *name, struct dso_id *id) strcpy(dso->name, name); if (id) dso->id = *id; - dso__set_long_name_id(dso, dso->name, id, false); + dso__set_long_name_id(dso, dso->name, false); dso__set_short_name(dso, dso->name, false); dso->symbols = RB_ROOT_CACHED; dso->symbol_names = NULL; @@ -1406,9 +1427,6 @@ struct dso *dso__new_id(const char *name, struct dso_id *id) dso->is_kmod = 0; dso->needs_swap = DSO_SWAP__UNSET; dso->comp = COMP_ID__NONE; - RB_CLEAR_NODE(&dso->rb_node); - dso->root = NULL; - INIT_LIST_HEAD(&dso->node); INIT_LIST_HEAD(&dso->data.open_entry); mutex_init(&dso->lock); refcount_set(&dso->refcnt, 1); @@ -1424,9 +1442,8 @@ struct dso *dso__new(const char *name) void dso__delete(struct dso *dso) { - if (!RB_EMPTY_NODE(&dso->rb_node)) - pr_err("DSO %s is still in rbtree when being deleted!\n", - dso->long_name); + if (dso->dsos) + pr_err("DSO %s is still in rbtree when being deleted!\n", dso->long_name); /* free inlines first, as they reference symbols */ inlines__tree_delete(&dso->inlined_nodes); diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h index 2c295438226d..b22dec8b3f3a 100644 --- a/tools/perf/util/dso.h +++ b/tools/perf/util/dso.h @@ -146,9 +146,7 @@ struct auxtrace_cache; struct dso { struct mutex lock; - struct list_head node; - struct rb_node rb_node; /* rbtree node sorted by long name */ - struct rb_root *root; /* root of rbtree that rb_node is in */ + struct dsos *dsos; struct rb_root_cached symbols; struct symbol **symbol_names; size_t symbol_names_len; @@ -238,8 +236,8 @@ static inline void dso__set_loaded(struct dso *dso) dso->loaded = true; } -int dso_id__cmp(struct dso_id *a, struct dso_id *b); -bool dso_id__empty(struct dso_id *id); +int dso_id__cmp(const struct dso_id *a, const struct dso_id *b); +bool dso_id__empty(const struct dso_id *id); struct dso *dso__new_id(const char *name, struct dso_id *id); struct dso *dso__new(const char *name); @@ -248,7 +246,7 @@ void dso__delete(struct dso *dso); int dso__cmp_id(struct dso *a, struct dso *b); void dso__set_short_name(struct dso *dso, const char *name, bool name_allocated); void dso__set_long_name(struct dso *dso, const char *name, bool name_allocated); -void dso__inject_id(struct dso *dso, struct dso_id *id); +void __dso__inject_id(struct dso *dso, struct dso_id *id); int dso__name_len(const struct dso *dso); diff --git a/tools/perf/util/dsos.c b/tools/perf/util/dsos.c index b7fbfb877ae3..c01f0b538e65 100644 --- a/tools/perf/util/dsos.c +++ b/tools/perf/util/dsos.c @@ -14,24 +14,30 @@ void dsos__init(struct dsos *dsos) { - INIT_LIST_HEAD(&dsos->head); - dsos->root = RB_ROOT; init_rwsem(&dsos->lock); + + dsos->cnt = 0; + dsos->allocated = 0; + dsos->dsos = NULL; + dsos->sorted = true; } static void dsos__purge(struct dsos *dsos) { - struct dso *pos, *n; - down_write(&dsos->lock); - list_for_each_entry_safe(pos, n, &dsos->head, node) { - RB_CLEAR_NODE(&pos->rb_node); - pos->root = NULL; - list_del_init(&pos->node); - dso__put(pos); + for (unsigned int i = 0; i < dsos->cnt; i++) { + struct dso *dso = dsos->dsos[i]; + + dso->dsos = NULL; + dso__put(dso); } + zfree(&dsos->dsos); + dsos->cnt = 0; + dsos->allocated = 0; + dsos->sorted = true; + up_write(&dsos->lock); } @@ -46,9 +52,8 @@ static int __dsos__for_each_dso(struct dsos *dsos, int (*cb)(struct dso *dso, void *data), void *data) { - struct dso *dso; - - list_for_each_entry(dso, &dsos->head, node) { + for (unsigned int i = 0; i < dsos->cnt; i++) { + struct dso *dso = dsos->dsos[i]; int err; err = cb(dso, data); @@ -119,16 +124,47 @@ static int dso__cmp_short_name(struct dso *a, struct dso *b) return __dso__cmp_short_name(a->short_name, &a->id, b); } +static int dsos__cmp_long_name_id_short_name(const void *va, const void *vb) +{ + const struct dso *a = *((const struct dso **)va); + const struct dso *b = *((const struct dso **)vb); + int rc = strcmp(a->long_name, b->long_name); + + if (!rc) { + rc = dso_id__cmp(&a->id, &b->id); + if (!rc) + rc = strcmp(a->short_name, b->short_name); + } + return rc; +} + /* * Find a matching entry and/or link current entry to RB tree. * Either one of the dso or name parameter must be non-NULL or the * function will not work. */ -struct dso *__dsos__findnew_link_by_longname_id(struct rb_root *root, struct dso *dso, - const char *name, struct dso_id *id) +struct dso *__dsos__findnew_link_by_longname_id(struct dsos *dsos, + struct dso *dso, + const char *name, + struct dso_id *id, + bool write_locked) { - struct rb_node **p = &root->rb_node; - struct rb_node *parent = NULL; + int low = 0, high = dsos->cnt - 1; + + if (!dsos->sorted) { + if (!write_locked) { + up_read(&dsos->lock); + down_write(&dsos->lock); + dso = __dsos__findnew_link_by_longname_id(dsos, dso, name, id, + /*write_locked=*/true); + up_write(&dsos->lock); + down_read(&dsos->lock); + return dso; + } + qsort(dsos->dsos, dsos->cnt, sizeof(struct dso *), + dsos__cmp_long_name_id_short_name); + dsos->sorted = true; + } if (!name) name = dso->long_name; @@ -136,11 +172,11 @@ struct dso *__dsos__findnew_link_by_longname_id(struct rb_root *root, struct dso /* * Find node with the matching name */ - while (*p) { - struct dso *this = rb_entry(*p, struct dso, rb_node); + while (low <= high) { + int mid = (low + high) / 2; + struct dso *this = dsos->dsos[mid]; int rc = __dso__cmp_long_name(name, id, this); - parent = *p; if (rc == 0) { /* * In case the new DSO is a duplicate of an existing @@ -161,56 +197,53 @@ struct dso *__dsos__findnew_link_by_longname_id(struct rb_root *root, struct dso } } if (rc < 0) - p = &parent->rb_left; + high = mid - 1; else - p = &parent->rb_right; - } - if (dso) { - /* Add new node and rebalance tree */ - rb_link_node(&dso->rb_node, parent, p); - rb_insert_color(&dso->rb_node, root); - dso->root = root; + low = mid + 1; } + if (dso) + __dsos__add(dsos, dso); return NULL; } -void __dsos__add(struct dsos *dsos, struct dso *dso) +int __dsos__add(struct dsos *dsos, struct dso *dso) { - list_add_tail(&dso->node, &dsos->head); - __dsos__findnew_link_by_longname_id(&dsos->root, dso, NULL, &dso->id); - /* - * It is now in the linked list, grab a reference, then garbage collect - * this when needing memory, by looking at LRU dso instances in the - * list with atomic_read(&dso->refcnt) == 1, i.e. no references - * anywhere besides the one for the list, do, under a lock for the - * list: remove it from the list, then a dso__put(), that probably will - * be the last and will then call dso__delete(), end of life. - * - * That, or at the end of the 'struct machine' lifetime, when all - * 'struct dso' instances will be removed from the list, in - * dsos__exit(), if they have no other reference from some other data - * structure. - * - * E.g.: after processing a 'perf.data' file and storing references - * to objects instantiated while processing events, we will have - * references to the 'thread', 'map', 'dso' structs all from 'struct - * hist_entry' instances, but we may not need anything not referenced, - * so we might as well call machines__exit()/machines__delete() and - * garbage collect it. - */ - dso__get(dso); + if (dsos->cnt == dsos->allocated) { + unsigned int to_allocate = 2; + struct dso **temp; + + if (dsos->allocated > 0) + to_allocate = dsos->allocated * 2; + temp = realloc(dsos->dsos, sizeof(struct dso *) * to_allocate); + if (!temp) + return -ENOMEM; + dsos->dsos = temp; + dsos->allocated = to_allocate; + } + dsos->dsos[dsos->cnt++] = dso__get(dso); + if (dsos->cnt >= 2 && dsos->sorted) { + dsos->sorted = dsos__cmp_long_name_id_short_name(&dsos->dsos[dsos->cnt - 2], + &dsos->dsos[dsos->cnt - 1]) + <= 0; + } + dso->dsos = dsos; + return 0; } -void dsos__add(struct dsos *dsos, struct dso *dso) +int dsos__add(struct dsos *dsos, struct dso *dso) { + int ret; + down_write(&dsos->lock); - __dsos__add(dsos, dso); + ret = __dsos__add(dsos, dso); up_write(&dsos->lock); + return ret; } -static struct dso *__dsos__findnew_by_longname_id(struct rb_root *root, const char *name, struct dso_id *id) +static struct dso *__dsos__findnew_by_longname_id(struct dsos *dsos, const char *name, + struct dso_id *id, bool write_locked) { - return __dsos__findnew_link_by_longname_id(root, NULL, name, id); + return __dsos__findnew_link_by_longname_id(dsos, NULL, name, id, write_locked); } struct dsos__find_id_cb_args { @@ -231,7 +264,8 @@ static int dsos__find_id_cb(struct dso *dso, void *data) } -static struct dso *__dsos__find_id(struct dsos *dsos, const char *name, struct dso_id *id, bool cmp_short) +static struct dso *__dsos__find_id(struct dsos *dsos, const char *name, struct dso_id *id, + bool cmp_short, bool write_locked) { struct dso *res; @@ -245,7 +279,7 @@ static struct dso *__dsos__find_id(struct dsos *dsos, const char *name, struct d __dsos__for_each_dso(dsos, dsos__find_id_cb, &args); return args.res; } - res = __dsos__findnew_by_longname_id(&dsos->root, name, id); + res = __dsos__findnew_by_longname_id(dsos, name, id, write_locked); return res; } @@ -254,7 +288,7 @@ struct dso *dsos__find(struct dsos *dsos, const char *name, bool cmp_short) struct dso *res; down_read(&dsos->lock); - res = __dsos__find_id(dsos, name, NULL, cmp_short); + res = __dsos__find_id(dsos, name, NULL, cmp_short, /*write_locked=*/false); up_read(&dsos->lock); return res; } @@ -296,8 +330,13 @@ static struct dso *__dsos__addnew_id(struct dsos *dsos, const char *name, struct struct dso *dso = dso__new_id(name, id); if (dso != NULL) { - __dsos__add(dsos, dso); + /* + * The dsos lock is held on entry, so rename the dso before + * adding it to avoid needing to take the dsos lock again to say + * the array isn't sorted. + */ dso__set_basename(dso); + __dsos__add(dsos, dso); } return dso; } @@ -309,10 +348,10 @@ struct dso *__dsos__addnew(struct dsos *dsos, const char *name) static struct dso *__dsos__findnew_id(struct dsos *dsos, const char *name, struct dso_id *id) { - struct dso *dso = __dsos__find_id(dsos, name, id, false); + struct dso *dso = __dsos__find_id(dsos, name, id, false, /*write_locked=*/true); if (dso && dso_id__empty(&dso->id) && !dso_id__empty(id)) - dso__inject_id(dso, id); + __dso__inject_id(dso, id); return dso ? dso : __dsos__addnew_id(dsos, name, id); } @@ -403,18 +442,27 @@ struct dso *dsos__findnew_module_dso(struct dsos *dsos, down_write(&dsos->lock); - dso = __dsos__find_id(dsos, m->name, NULL, /*cmp_short=*/true); + dso = __dsos__find_id(dsos, m->name, NULL, /*cmp_short=*/true, /*write_locked=*/true); + if (dso) { + up_write(&dsos->lock); + return dso; + } + /* + * Failed to find the dso so create it. Change the name before adding it + * to the array, to avoid unnecessary sorts and potential locking + * issues. + */ + dso = dso__new_id(m->name, /*id=*/NULL); if (!dso) { - dso = __dsos__addnew(dsos, m->name); - if (dso == NULL) - goto out_unlock; - - dso__set_module_info(dso, m, machine); - dso__set_long_name(dso, strdup(filename), true); - dso->kernel = DSO_SPACE__KERNEL; + up_write(&dsos->lock); + return NULL; } + dso__set_basename(dso); + dso__set_module_info(dso, m, machine); + dso__set_long_name(dso, strdup(filename), true); + dso->kernel = DSO_SPACE__KERNEL; + __dsos__add(dsos, dso); -out_unlock: up_write(&dsos->lock); return dso; } diff --git a/tools/perf/util/dsos.h b/tools/perf/util/dsos.h index 50bd51523475..c1b3979ad4bd 100644 --- a/tools/perf/util/dsos.h +++ b/tools/perf/util/dsos.h @@ -14,20 +14,22 @@ struct kmod_path; struct machine; /* - * DSOs are put into both a list for fast iteration and rbtree for fast - * long name lookup. + * Collection of DSOs as an array for iteration speed, but sorted for O(n) + * lookup. */ struct dsos { - struct list_head head; - struct rb_root root; /* rbtree root sorted by long name */ struct rw_semaphore lock; + struct dso **dsos; + unsigned int cnt; + unsigned int allocated; + bool sorted; }; void dsos__init(struct dsos *dsos); void dsos__exit(struct dsos *dsos); -void __dsos__add(struct dsos *dsos, struct dso *dso); -void dsos__add(struct dsos *dsos, struct dso *dso); +int __dsos__add(struct dsos *dsos, struct dso *dso); +int dsos__add(struct dsos *dsos, struct dso *dso); struct dso *__dsos__addnew(struct dsos *dsos, const char *name); struct dso *dsos__find(struct dsos *dsos, const char *name, bool cmp_short); @@ -35,8 +37,11 @@ struct dso *dsos__findnew_id(struct dsos *dsos, const char *name, struct dso_id bool dsos__read_build_ids(struct dsos *dsos, bool with_hits); -struct dso *__dsos__findnew_link_by_longname_id(struct rb_root *root, struct dso *dso, - const char *name, struct dso_id *id); +struct dso *__dsos__findnew_link_by_longname_id(struct dsos *dsos, + struct dso *dso, + const char *name, + struct dso_id *id, + bool write_locked); size_t dsos__fprintf_buildid(struct dsos *dsos, FILE *fp, bool (skip)(struct dso *dso, int parm), int parm); -- cgit v1.2.3 From dfd48165bbf752e41dd51e77b92db3f848a4a99a Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Sat, 4 May 2024 14:37:58 -0700 Subject: perf dsos: Remove __dsos__addnew() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Function no longer used so remove. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Ahelenia ZiemiaÅ„ska Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Ben Gainey Cc: Changbin Du Cc: Chengen Du Cc: Colin Ian King Cc: Dima Kogan Cc: Ilkka Koskinen Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kan Liang Cc: Leo Yan Cc: Li Dong Cc: Mark Rutland Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Paran Lee Cc: Peter Zijlstra Cc: Song Liu Cc: Sun Haiyong Cc: Thomas Richter Cc: Tiezhu Yang Cc: Yanteng Si Cc: zhaimingbing Link: https://lore.kernel.org/r/20240504213803.218974-3-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dsos.c | 5 ----- tools/perf/util/dsos.h | 1 - 2 files changed, 6 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dsos.c b/tools/perf/util/dsos.c index c01f0b538e65..89f87a197a44 100644 --- a/tools/perf/util/dsos.c +++ b/tools/perf/util/dsos.c @@ -341,11 +341,6 @@ static struct dso *__dsos__addnew_id(struct dsos *dsos, const char *name, struct return dso; } -struct dso *__dsos__addnew(struct dsos *dsos, const char *name) -{ - return __dsos__addnew_id(dsos, name, NULL); -} - static struct dso *__dsos__findnew_id(struct dsos *dsos, const char *name, struct dso_id *id) { struct dso *dso = __dsos__find_id(dsos, name, id, false, /*write_locked=*/true); diff --git a/tools/perf/util/dsos.h b/tools/perf/util/dsos.h index c1b3979ad4bd..d1497b11d64c 100644 --- a/tools/perf/util/dsos.h +++ b/tools/perf/util/dsos.h @@ -30,7 +30,6 @@ void dsos__exit(struct dsos *dsos); int __dsos__add(struct dsos *dsos, struct dso *dso); int dsos__add(struct dsos *dsos, struct dso *dso); -struct dso *__dsos__addnew(struct dsos *dsos, const char *name); struct dso *dsos__find(struct dsos *dsos, const char *name, bool cmp_short); struct dso *dsos__findnew_id(struct dsos *dsos, const char *name, struct dso_id *id); -- cgit v1.2.3 From 7410d6008d28d65bead6aa85909e6915f2c8fc61 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Sat, 4 May 2024 14:37:59 -0700 Subject: perf dsos: Remove __dsos__findnew_link_by_longname_id() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Function was only called in dsos.c with the dso parameter as NULL. Remove the function and specialize for the dso being NULL case removing other unused functions along the way. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Ahelenia ZiemiaÅ„ska Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Ben Gainey Cc: Changbin Du Cc: Chengen Du Cc: Colin Ian King Cc: Dima Kogan Cc: Ilkka Koskinen Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kan Liang Cc: Leo Yan Cc: Li Dong Cc: Mark Rutland Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Paran Lee Cc: Peter Zijlstra Cc: Song Liu Cc: Sun Haiyong Cc: Thomas Richter Cc: Tiezhu Yang Cc: Yanteng Si Cc: zhaimingbing Link: https://lore.kernel.org/r/20240504213803.218974-4-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dsos.c | 51 ++++++++++---------------------------------------- tools/perf/util/dsos.h | 6 ------ 2 files changed, 10 insertions(+), 47 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dsos.c b/tools/perf/util/dsos.c index 89f87a197a44..e2ccceac7619 100644 --- a/tools/perf/util/dsos.c +++ b/tools/perf/util/dsos.c @@ -119,11 +119,6 @@ static int __dso__cmp_short_name(const char *short_name, struct dso_id *id, stru return rc ?: dso_id__cmp(id, &b->id); } -static int dso__cmp_short_name(struct dso *a, struct dso *b) -{ - return __dso__cmp_short_name(a->short_name, &a->id, b); -} - static int dsos__cmp_long_name_id_short_name(const void *va, const void *vb) { const struct dso *a = *((const struct dso **)va); @@ -143,20 +138,21 @@ static int dsos__cmp_long_name_id_short_name(const void *va, const void *vb) * Either one of the dso or name parameter must be non-NULL or the * function will not work. */ -struct dso *__dsos__findnew_link_by_longname_id(struct dsos *dsos, - struct dso *dso, - const char *name, - struct dso_id *id, - bool write_locked) +static struct dso *__dsos__find_by_longname_id(struct dsos *dsos, + const char *name, + struct dso_id *id, + bool write_locked) { int low = 0, high = dsos->cnt - 1; if (!dsos->sorted) { if (!write_locked) { + struct dso *dso; + up_read(&dsos->lock); down_write(&dsos->lock); - dso = __dsos__findnew_link_by_longname_id(dsos, dso, name, id, - /*write_locked=*/true); + dso = __dsos__find_by_longname_id(dsos, name, id, + /*write_locked=*/true); up_write(&dsos->lock); down_read(&dsos->lock); return dso; @@ -166,9 +162,6 @@ struct dso *__dsos__findnew_link_by_longname_id(struct dsos *dsos, dsos->sorted = true; } - if (!name) - name = dso->long_name; - /* * Find node with the matching name */ @@ -178,31 +171,13 @@ struct dso *__dsos__findnew_link_by_longname_id(struct dsos *dsos, int rc = __dso__cmp_long_name(name, id, this); if (rc == 0) { - /* - * In case the new DSO is a duplicate of an existing - * one, print a one-time warning & put the new entry - * at the end of the list of duplicates. - */ - if (!dso || (dso == this)) - return dso__get(this); /* Find matching dso */ - /* - * The core kernel DSOs may have duplicated long name. - * In this case, the short name should be different. - * Comparing the short names to differentiate the DSOs. - */ - rc = dso__cmp_short_name(dso, this); - if (rc == 0) { - pr_err("Duplicated dso name: %s\n", name); - return NULL; - } + return dso__get(this); /* Find matching dso */ } if (rc < 0) high = mid - 1; else low = mid + 1; } - if (dso) - __dsos__add(dsos, dso); return NULL; } @@ -240,12 +215,6 @@ int dsos__add(struct dsos *dsos, struct dso *dso) return ret; } -static struct dso *__dsos__findnew_by_longname_id(struct dsos *dsos, const char *name, - struct dso_id *id, bool write_locked) -{ - return __dsos__findnew_link_by_longname_id(dsos, NULL, name, id, write_locked); -} - struct dsos__find_id_cb_args { const char *name; struct dso_id *id; @@ -279,7 +248,7 @@ static struct dso *__dsos__find_id(struct dsos *dsos, const char *name, struct d __dsos__for_each_dso(dsos, dsos__find_id_cb, &args); return args.res; } - res = __dsos__findnew_by_longname_id(dsos, name, id, write_locked); + res = __dsos__find_by_longname_id(dsos, name, id, write_locked); return res; } diff --git a/tools/perf/util/dsos.h b/tools/perf/util/dsos.h index d1497b11d64c..6c13b65648bc 100644 --- a/tools/perf/util/dsos.h +++ b/tools/perf/util/dsos.h @@ -36,12 +36,6 @@ struct dso *dsos__findnew_id(struct dsos *dsos, const char *name, struct dso_id bool dsos__read_build_ids(struct dsos *dsos, bool with_hits); -struct dso *__dsos__findnew_link_by_longname_id(struct dsos *dsos, - struct dso *dso, - const char *name, - struct dso_id *id, - bool write_locked); - size_t dsos__fprintf_buildid(struct dsos *dsos, FILE *fp, bool (skip)(struct dso *dso, int parm), int parm); size_t dsos__fprintf(struct dsos *dsos, FILE *fp); -- cgit v1.2.3 From 7a9418cf7f05a74cbc9d4c750ee1bfddaa11f121 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Sat, 4 May 2024 14:38:00 -0700 Subject: perf dsos: Switch hand crafted code to bsearch() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Switch to using the bsearch library function rather than having a hand written binary search. Const-ify some static functions to avoid compiler warnings. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Ahelenia ZiemiaÅ„ska Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Ben Gainey Cc: Changbin Du Cc: Chengen Du Cc: Colin Ian King Cc: Dima Kogan Cc: Ilkka Koskinen Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kan Liang Cc: Leo Yan Cc: Li Dong Cc: Mark Rutland Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Paran Lee Cc: Peter Zijlstra Cc: Song Liu Cc: Sun Haiyong Cc: Thomas Richter Cc: Tiezhu Yang Cc: Yanteng Si Cc: zhaimingbing Link: https://lore.kernel.org/r/20240504213803.218974-5-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dsos.c | 46 +++++++++++++++++++++++++++------------------- 1 file changed, 27 insertions(+), 19 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dsos.c b/tools/perf/util/dsos.c index e2ccceac7619..2e4e86dc6c17 100644 --- a/tools/perf/util/dsos.c +++ b/tools/perf/util/dsos.c @@ -107,13 +107,15 @@ bool dsos__read_build_ids(struct dsos *dsos, bool with_hits) return args.have_build_id; } -static int __dso__cmp_long_name(const char *long_name, struct dso_id *id, struct dso *b) +static int __dso__cmp_long_name(const char *long_name, const struct dso_id *id, + const struct dso *b) { int rc = strcmp(long_name, b->long_name); return rc ?: dso_id__cmp(id, &b->id); } -static int __dso__cmp_short_name(const char *short_name, struct dso_id *id, struct dso *b) +static int __dso__cmp_short_name(const char *short_name, const struct dso_id *id, + const struct dso *b) { int rc = strcmp(short_name, b->short_name); return rc ?: dso_id__cmp(id, &b->id); @@ -133,6 +135,19 @@ static int dsos__cmp_long_name_id_short_name(const void *va, const void *vb) return rc; } +struct dsos__key { + const char *long_name; + const struct dso_id *id; +}; + +static int dsos__cmp_key_long_name_id(const void *vkey, const void *vdso) +{ + const struct dsos__key *key = vkey; + const struct dso *dso = *((const struct dso **)vdso); + + return __dso__cmp_long_name(key->long_name, key->id, dso); +} + /* * Find a matching entry and/or link current entry to RB tree. * Either one of the dso or name parameter must be non-NULL or the @@ -143,7 +158,11 @@ static struct dso *__dsos__find_by_longname_id(struct dsos *dsos, struct dso_id *id, bool write_locked) { - int low = 0, high = dsos->cnt - 1; + struct dsos__key key = { + .long_name = name, + .id = id, + }; + struct dso **res; if (!dsos->sorted) { if (!write_locked) { @@ -162,23 +181,12 @@ static struct dso *__dsos__find_by_longname_id(struct dsos *dsos, dsos->sorted = true; } - /* - * Find node with the matching name - */ - while (low <= high) { - int mid = (low + high) / 2; - struct dso *this = dsos->dsos[mid]; - int rc = __dso__cmp_long_name(name, id, this); + res = bsearch(&key, dsos->dsos, dsos->cnt, sizeof(struct dso *), + dsos__cmp_key_long_name_id); + if (!res) + return NULL; - if (rc == 0) { - return dso__get(this); /* Find matching dso */ - } - if (rc < 0) - high = mid - 1; - else - low = mid + 1; - } - return NULL; + return dso__get(*res); } int __dsos__add(struct dsos *dsos, struct dso *dso) -- cgit v1.2.3 From ee756ef7491eafd70f390343a1d90930af125a51 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Sat, 4 May 2024 14:38:01 -0700 Subject: perf dso: Add reference count checking and accessor functions MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Add reference count checking to struct dso, this can help with implementing correct reference counting discipline. To avoid RC_CHK_ACCESS everywhere, add accessor functions for the variables in struct dso. The majority of the change is mechanical in nature and not easy to split up. Committer testing: 'perf test' up to this patch shows no regressions. But: util/symbol.c: In function ‘dso__load_bfd_symbols’: util/symbol.c:1683:9: error: too few arguments to function ‘dso__set_adjust_symbols’ 1683 | dso__set_adjust_symbols(dso); | ^~~~~~~~~~~~~~~~~~~~~~~ In file included from util/symbol.c:21: util/dso.h:268:20: note: declared here 268 | static inline void dso__set_adjust_symbols(struct dso *dso, bool val) | ^~~~~~~~~~~~~~~~~~~~~~~ make[6]: *** [/home/acme/git/perf-tools-next/tools/build/Makefile.build:106: /tmp/tmp.ZWHbQftdN6/util/symbol.o] Error 1 MKDIR /tmp/tmp.ZWHbQftdN6/tests/workloads/ make[6]: *** Waiting for unfinished jobs.... This was updated: - symbols__fixup_end(&dso->symbols, false); - symbols__fixup_duplicate(&dso->symbols); - dso->adjust_symbols = 1; + symbols__fixup_end(dso__symbols(dso), false); + symbols__fixup_duplicate(dso__symbols(dso)); + dso__set_adjust_symbols(dso); But not build tested with BUILD_NONDISTRO and libbfd devel files installed (binutils-devel on fedora). Add the missing argument: symbols__fixup_end(dso__symbols(dso), false); symbols__fixup_duplicate(dso__symbols(dso)); - dso__set_adjust_symbols(dso); + dso__set_adjust_symbols(dso, true); Signed-off-by: Ian Rogers Tested-by: Arnaldo Carvalho de Melo Cc: Adrian Hunter Cc: Ahelenia ZiemiaÅ„ska Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Ben Gainey Cc: Changbin Du Cc: Chengen Du Cc: Colin Ian King Cc: Dima Kogan Cc: Ilkka Koskinen Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kan Liang Cc: Leo Yan Cc: Li Dong Cc: Mark Rutland Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Paran Lee Cc: Peter Zijlstra Cc: Song Liu Cc: Sun Haiyong Cc: Thomas Richter Cc: Tiezhu Yang Cc: Yanteng Si Cc: zhaimingbing Link: https://lore.kernel.org/r/20240504213803.218974-6-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-annotate.c | 6 +- tools/perf/builtin-buildid-cache.c | 2 +- tools/perf/builtin-buildid-list.c | 18 +- tools/perf/builtin-inject.c | 71 ++- tools/perf/builtin-kallsyms.c | 2 +- tools/perf/builtin-mem.c | 4 +- tools/perf/builtin-report.c | 6 +- tools/perf/builtin-script.c | 8 +- tools/perf/builtin-top.c | 4 +- tools/perf/builtin-trace.c | 2 +- tools/perf/tests/code-reading.c | 8 +- tools/perf/tests/dso-data.c | 11 +- tools/perf/tests/hists_common.c | 6 +- tools/perf/tests/hists_cumulate.c | 4 +- tools/perf/tests/hists_output.c | 2 +- tools/perf/tests/maps.c | 4 +- tools/perf/tests/symbols.c | 8 +- tools/perf/tests/vmlinux-kallsyms.c | 6 +- tools/perf/ui/browsers/annotate.c | 6 +- tools/perf/ui/browsers/hists.c | 8 +- tools/perf/ui/browsers/map.c | 4 +- tools/perf/util/annotate-data.c | 18 +- tools/perf/util/annotate.c | 17 +- tools/perf/util/auxtrace.c | 2 +- tools/perf/util/block-info.c | 2 +- tools/perf/util/bpf-event.c | 8 +- tools/perf/util/build-id.c | 38 +- tools/perf/util/callchain.c | 2 +- tools/perf/util/data-convert-json.c | 2 +- tools/perf/util/db-export.c | 6 +- tools/perf/util/disasm.c | 40 +- tools/perf/util/dlfilter.c | 12 +- tools/perf/util/dso.c | 368 ++++++++-------- tools/perf/util/dso.h | 488 +++++++++++++++++++-- tools/perf/util/dsos.c | 54 +-- tools/perf/util/event.c | 8 +- tools/perf/util/header.c | 8 +- tools/perf/util/hist.c | 4 +- tools/perf/util/intel-pt.c | 22 +- tools/perf/util/machine.c | 46 +- tools/perf/util/map.c | 77 ++-- tools/perf/util/maps.c | 14 +- tools/perf/util/print_insn.c | 2 +- tools/perf/util/probe-event.c | 25 +- .../perf/util/scripting-engines/trace-event-perl.c | 6 +- .../util/scripting-engines/trace-event-python.c | 21 +- tools/perf/util/sort.c | 19 +- tools/perf/util/srcline.c | 65 +-- tools/perf/util/symbol-elf.c | 94 ++-- tools/perf/util/symbol-minimal.c | 4 +- tools/perf/util/symbol.c | 186 ++++---- tools/perf/util/symbol_fprintf.c | 4 +- tools/perf/util/synthetic-events.c | 24 +- tools/perf/util/thread.c | 4 +- tools/perf/util/unwind-libunwind-local.c | 18 +- tools/perf/util/unwind-libunwind.c | 2 +- tools/perf/util/vdso.c | 8 +- 57 files changed, 1169 insertions(+), 739 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 83812b9d5363..50d2fb222d48 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -219,7 +219,7 @@ static int process_branch_callback(struct evsel *evsel, } if (a.map != NULL) - map__dso(a.map)->hit = 1; + dso__set_hit(map__dso(a.map)); hist__account_cycles(sample->branch_stack, al, sample, false, NULL); @@ -254,7 +254,7 @@ static int evsel__add_sample(struct evsel *evsel, struct perf_sample *sample, if (al->sym != NULL) { struct dso *dso = map__dso(al->map); - rb_erase_cached(&al->sym->rb_node, &dso->symbols); + rb_erase_cached(&al->sym->rb_node, dso__symbols(dso)); symbol__delete(al->sym); dso__reset_find_symbol_cache(dso); } @@ -419,7 +419,7 @@ static void hists__find_annotations(struct hists *hists, struct hist_entry *he = rb_entry(nd, struct hist_entry, rb_node); struct annotation *notes; - if (he->ms.sym == NULL || map__dso(he->ms.map)->annotate_warned) + if (he->ms.sym == NULL || dso__annotate_warned(map__dso(he->ms.map))) goto find_next; if (ann->sym_hist_filter && diff --git a/tools/perf/builtin-buildid-cache.c b/tools/perf/builtin-buildid-cache.c index e2a40f1d9225..b0511d16aeb6 100644 --- a/tools/perf/builtin-buildid-cache.c +++ b/tools/perf/builtin-buildid-cache.c @@ -286,7 +286,7 @@ static bool dso__missing_buildid_cache(struct dso *dso, int parm __maybe_unused) pr_warning("Problems with %s file, consider removing it from the cache\n", filename); - } else if (memcmp(dso->bid.data, bid.data, bid.size)) { + } else if (memcmp(dso__bid(dso)->data, bid.data, bid.size)) { pr_warning("Problems with %s file, consider removing it from the cache\n", filename); } diff --git a/tools/perf/builtin-buildid-list.c b/tools/perf/builtin-buildid-list.c index c9037477865a..383d5de36ce4 100644 --- a/tools/perf/builtin-buildid-list.c +++ b/tools/perf/builtin-buildid-list.c @@ -26,16 +26,18 @@ static int buildid__map_cb(struct map *map, void *arg __maybe_unused) { const struct dso *dso = map__dso(map); char bid_buf[SBUILD_ID_SIZE]; + const char *dso_long_name = dso__long_name(dso); + const char *dso_short_name = dso__short_name(dso); memset(bid_buf, 0, sizeof(bid_buf)); - if (dso->has_build_id) - build_id__sprintf(&dso->bid, bid_buf); + if (dso__has_build_id(dso)) + build_id__sprintf(dso__bid_const(dso), bid_buf); printf("%s %16" PRIx64 " %16" PRIx64, bid_buf, map__start(map), map__end(map)); - if (dso->long_name != NULL) { - printf(" %s", dso->long_name); - } else if (dso->short_name != NULL) { - printf(" %s", dso->short_name); - } + if (dso_long_name != NULL) + printf(" %s", dso_long_name); + else if (dso_short_name != NULL) + printf(" %s", dso_short_name); + printf("\n"); return 0; @@ -76,7 +78,7 @@ static int filename__fprintf_build_id(const char *name, FILE *fp) static bool dso__skip_buildid(struct dso *dso, int with_hits) { - return with_hits && !dso->hit; + return with_hits && !dso__hit(dso); } static int perf_session__list_build_ids(bool force, bool with_hits) diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c index ce5e28eaad90..a212678d47be 100644 --- a/tools/perf/builtin-inject.c +++ b/tools/perf/builtin-inject.c @@ -445,10 +445,9 @@ static struct dso *findnew_dso(int pid, int tid, const char *filename, } if (dso) { - mutex_lock(&dso->lock); - nsinfo__put(dso->nsinfo); - dso->nsinfo = nsi; - mutex_unlock(&dso->lock); + mutex_lock(dso__lock(dso)); + dso__set_nsinfo(dso, nsi); + mutex_unlock(dso__lock(dso)); } else nsinfo__put(nsi); @@ -466,8 +465,8 @@ static int perf_event__repipe_buildid_mmap(struct perf_tool *tool, dso = findnew_dso(event->mmap.pid, event->mmap.tid, event->mmap.filename, NULL, machine); - if (dso && !dso->hit) { - dso->hit = 1; + if (dso && !dso__hit(dso)) { + dso__set_hit(dso); dso__inject_build_id(dso, tool, machine, sample->cpumode, 0); } dso__put(dso); @@ -492,7 +491,7 @@ static int perf_event__repipe_mmap2(struct perf_tool *tool, event->mmap2.filename, NULL, machine); if (dso) { /* mark it not to inject build-id */ - dso->hit = 1; + dso__set_hit(dso); } dso__put(dso); } @@ -544,7 +543,7 @@ static int perf_event__repipe_buildid_mmap2(struct perf_tool *tool, event->mmap2.filename, NULL, machine); if (dso) { /* mark it not to inject build-id */ - dso->hit = 1; + dso__set_hit(dso); } dso__put(dso); perf_event__repipe(tool, event, sample, machine); @@ -554,8 +553,8 @@ static int perf_event__repipe_buildid_mmap2(struct perf_tool *tool, dso = findnew_dso(event->mmap2.pid, event->mmap2.tid, event->mmap2.filename, &dso_id, machine); - if (dso && !dso->hit) { - dso->hit = 1; + if (dso && !dso__hit(dso)) { + dso__set_hit(dso); dso__inject_build_id(dso, tool, machine, sample->cpumode, event->mmap2.flags); } @@ -631,24 +630,24 @@ static int dso__read_build_id(struct dso *dso) { struct nscookie nsc; - if (dso->has_build_id) + if (dso__has_build_id(dso)) return 0; - mutex_lock(&dso->lock); - nsinfo__mountns_enter(dso->nsinfo, &nsc); - if (filename__read_build_id(dso->long_name, &dso->bid) > 0) - dso->has_build_id = true; - else if (dso->nsinfo) { - char *new_name = dso__filename_with_chroot(dso, dso->long_name); + mutex_lock(dso__lock(dso)); + nsinfo__mountns_enter(dso__nsinfo(dso), &nsc); + if (filename__read_build_id(dso__long_name(dso), dso__bid(dso)) > 0) + dso__set_has_build_id(dso); + else if (dso__nsinfo(dso)) { + char *new_name = dso__filename_with_chroot(dso, dso__long_name(dso)); - if (new_name && filename__read_build_id(new_name, &dso->bid) > 0) - dso->has_build_id = true; + if (new_name && filename__read_build_id(new_name, dso__bid(dso)) > 0) + dso__set_has_build_id(dso); free(new_name); } nsinfo__mountns_exit(&nsc); - mutex_unlock(&dso->lock); + mutex_unlock(dso__lock(dso)); - return dso->has_build_id ? 0 : -1; + return dso__has_build_id(dso) ? 0 : -1; } static struct strlist *perf_inject__parse_known_build_ids( @@ -700,14 +699,14 @@ static bool perf_inject__lookup_known_build_id(struct perf_inject *inject, dso_name = strchr(build_id, ' '); bid_len = dso_name - pos->s; dso_name = skip_spaces(dso_name); - if (strcmp(dso->long_name, dso_name)) + if (strcmp(dso__long_name(dso), dso_name)) continue; for (int ix = 0; 2 * ix + 1 < bid_len; ++ix) { - dso->bid.data[ix] = (hex(build_id[2 * ix]) << 4 | - hex(build_id[2 * ix + 1])); + dso__bid(dso)->data[ix] = (hex(build_id[2 * ix]) << 4 | + hex(build_id[2 * ix + 1])); } - dso->bid.size = bid_len / 2; - dso->has_build_id = 1; + dso__bid(dso)->size = bid_len / 2; + dso__set_has_build_id(dso); return true; } return false; @@ -720,9 +719,9 @@ static int dso__inject_build_id(struct dso *dso, struct perf_tool *tool, tool); int err; - if (is_anon_memory(dso->long_name) || flags & MAP_HUGETLB) + if (is_anon_memory(dso__long_name(dso)) || flags & MAP_HUGETLB) return 0; - if (is_no_dso_memory(dso->long_name)) + if (is_no_dso_memory(dso__long_name(dso))) return 0; if (inject->known_build_ids != NULL && @@ -730,14 +729,14 @@ static int dso__inject_build_id(struct dso *dso, struct perf_tool *tool, return 1; if (dso__read_build_id(dso) < 0) { - pr_debug("no build_id found for %s\n", dso->long_name); + pr_debug("no build_id found for %s\n", dso__long_name(dso)); return -1; } err = perf_event__synthesize_build_id(tool, dso, cpumode, perf_event__repipe, machine); if (err) { - pr_err("Can't synthesize build_id event for %s\n", dso->long_name); + pr_err("Can't synthesize build_id event for %s\n", dso__long_name(dso)); return -1; } @@ -763,8 +762,8 @@ int perf_event__inject_buildid(struct perf_tool *tool, union perf_event *event, if (thread__find_map(thread, sample->cpumode, sample->ip, &al)) { struct dso *dso = map__dso(al.map); - if (!dso->hit) { - dso->hit = 1; + if (!dso__hit(dso)) { + dso__set_hit(dso); dso__inject_build_id(dso, tool, machine, sample->cpumode, map__flags(al.map)); } @@ -1146,8 +1145,8 @@ static bool dso__is_in_kernel_space(struct dso *dso) return false; return dso__is_kcore(dso) || - dso->kernel || - is_kernel_module(dso->long_name, PERF_RECORD_MISC_CPUMODE_UNKNOWN); + dso__kernel(dso) || + is_kernel_module(dso__long_name(dso), PERF_RECORD_MISC_CPUMODE_UNKNOWN); } static u64 evlist__first_id(struct evlist *evlist) @@ -1181,7 +1180,7 @@ static int synthesize_build_id(struct perf_inject *inject, struct dso *dso, pid_ if (!machine) return -ENOMEM; - dso->hit = 1; + dso__set_hit(dso); return perf_event__synthesize_build_id(&inject->tool, dso, cpumode, process_build_id, machine); @@ -1192,7 +1191,7 @@ static int guest_session__add_build_ids_cb(struct dso *dso, void *data) struct guest_session *gs = data; struct perf_inject *inject = container_of(gs, struct perf_inject, guest_session); - if (!dso->has_build_id) + if (!dso__has_build_id(dso)) return 0; return synthesize_build_id(inject, dso, gs->machine_pid); diff --git a/tools/perf/builtin-kallsyms.c b/tools/perf/builtin-kallsyms.c index 7f75c5b73f26..a3c2ffdc1af8 100644 --- a/tools/perf/builtin-kallsyms.c +++ b/tools/perf/builtin-kallsyms.c @@ -38,7 +38,7 @@ static int __cmd_kallsyms(int argc, const char **argv) dso = map__dso(map); printf("%s: %s %s %#" PRIx64 "-%#" PRIx64 " (%#" PRIx64 "-%#" PRIx64")\n", - symbol->name, dso->short_name, dso->long_name, + symbol->name, dso__short_name(dso), dso__long_name(dso), map__unmap_ip(map, symbol->start), map__unmap_ip(map, symbol->end), symbol->start, symbol->end); } diff --git a/tools/perf/builtin-mem.c b/tools/perf/builtin-mem.c index 5b851e64e4a1..863fcd735dae 100644 --- a/tools/perf/builtin-mem.c +++ b/tools/perf/builtin-mem.c @@ -213,7 +213,7 @@ dump_raw_samples(struct perf_tool *tool, if (al.map != NULL) { dso = map__dso(al.map); if (dso) - dso->hit = 1; + dso__set_hit(dso); } field_sep = symbol_conf.field_sep; @@ -255,7 +255,7 @@ dump_raw_samples(struct perf_tool *tool, symbol_conf.field_sep, sample->data_src, symbol_conf.field_sep, - dso ? dso->long_name : "???", + dso ? dso__long_name(dso) : "???", al.sym ? al.sym->name : "???"); out_put: addr_location__exit(&al); diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index dafba6e030ef..b5525f4f7090 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -322,7 +322,7 @@ static int process_sample_event(struct perf_tool *tool, } if (al.map != NULL) - map__dso(al.map)->hit = 1; + dso__set_hit(map__dso(al.map)); if (ui__has_annotation() || rep->symbol_ipc || rep->total_cycles_mode) { hist__account_cycles(sample->branch_stack, &al, sample, @@ -609,7 +609,7 @@ static void report__warn_kptr_restrict(const struct report *rep) return; if (kernel_map == NULL || - (map__dso(kernel_map)->hit && + (dso__hit(map__dso(kernel_map)) && (kernel_kmap->ref_reloc_sym == NULL || kernel_kmap->ref_reloc_sym->addr == 0))) { const char *desc = @@ -850,7 +850,7 @@ static int maps__fprintf_task_cb(struct map *map, void *data) prot & PROT_EXEC ? 'x' : '-', map__flags(map) ? 's' : 'p', map__pgoff(map), - dso->id.ino, dso->name); + dso__id_const(dso)->ino, dso__name(dso)); if (ret < 0) return ret; diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index 647cb31a47c8..f7c3c3868c3c 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -1017,11 +1017,11 @@ static int perf_sample__fprintf_brstackoff(struct perf_sample *sample, to = entries[i].to; if (thread__find_map_fb(thread, sample->cpumode, from, &alf) && - !map__dso(alf.map)->adjust_symbols) + !dso__adjust_symbols(map__dso(alf.map))) from = map__dso_map_ip(alf.map, from); if (thread__find_map_fb(thread, sample->cpumode, to, &alt) && - !map__dso(alt.map)->adjust_symbols) + !dso__adjust_symbols(map__dso(alt.map))) to = map__dso_map_ip(alt.map, to); printed += fprintf(fp, " 0x%"PRIx64, from); @@ -1082,7 +1082,7 @@ static int grab_bb(u8 *buffer, u64 start, u64 end, pr_debug("\tcannot resolve %" PRIx64 "-%" PRIx64 "\n", start, end); goto out; } - if (dso->data.status == DSO_DATA_STATUS_ERROR) { + if (dso__data(dso)->status == DSO_DATA_STATUS_ERROR) { pr_debug("\tcannot resolve %" PRIx64 "-%" PRIx64 "\n", start, end); goto out; } @@ -1094,7 +1094,7 @@ static int grab_bb(u8 *buffer, u64 start, u64 end, len = dso__data_read_offset(dso, machine, offset, (u8 *)buffer, end - start + MAXINSN); - *is64bit = dso->is_64_bit; + *is64bit = dso__is_64_bit(dso); if (len <= 0) pr_debug("\tcannot fetch code for block at %" PRIx64 "-%" PRIx64 "\n", start, end); diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 5ac6dcc64cef..1d6aef51c122 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -129,7 +129,7 @@ static int perf_top__parse_source(struct perf_top *top, struct hist_entry *he) /* * We can't annotate with just /proc/kallsyms */ - if (dso->symtab_type == DSO_BINARY_TYPE__KALLSYMS && !dso__is_kcore(dso)) { + if (dso__symtab_type(dso) == DSO_BINARY_TYPE__KALLSYMS && !dso__is_kcore(dso)) { pr_err("Can't annotate %s: No vmlinux file was found in the " "path\n", sym->name); sleep(1); @@ -182,7 +182,7 @@ static void ui__warn_map_erange(struct map *map, struct symbol *sym, u64 ip) "Tools: %s\n\n" "Not all samples will be on the annotation output.\n\n" "Please report to linux-kernel@vger.kernel.org\n", - ip, dso->long_name, dso__symtab_origin(dso), + ip, dso__long_name(dso), dso__symtab_origin(dso), map__start(map), map__end(map), sym->start, sym->end, sym->binding == STB_GLOBAL ? 'g' : sym->binding == STB_LOCAL ? 'l' : 'w', sym->name, diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 3818d3a62779..51eca671c797 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -2917,7 +2917,7 @@ static void print_location(FILE *f, struct perf_sample *sample, { if ((verbose > 0 || print_dso) && al->map) - fprintf(f, "%s@", map__dso(al->map)->long_name); + fprintf(f, "%s@", dso__long_name(map__dso(al->map))); if ((verbose > 0 || print_sym) && al->sym) fprintf(f, "%s+0x%" PRIx64, al->sym->name, diff --git a/tools/perf/tests/code-reading.c b/tools/perf/tests/code-reading.c index 29d2f3ee4e10..27c82cfb7e7d 100644 --- a/tools/perf/tests/code-reading.c +++ b/tools/perf/tests/code-reading.c @@ -253,9 +253,9 @@ static int read_object_code(u64 addr, size_t len, u8 cpumode, goto out; } dso = map__dso(al.map); - pr_debug("File is: %s\n", dso->long_name); + pr_debug("File is: %s\n", dso__long_name(dso)); - if (dso->symtab_type == DSO_BINARY_TYPE__KALLSYMS && !dso__is_kcore(dso)) { + if (dso__symtab_type(dso) == DSO_BINARY_TYPE__KALLSYMS && !dso__is_kcore(dso)) { pr_debug("Unexpected kernel address - skipping\n"); goto out; } @@ -274,7 +274,7 @@ static int read_object_code(u64 addr, size_t len, u8 cpumode, * modules to manage long jumps. Check if the ip offset falls in stubs * sections for kernel modules. And skip module address after text end */ - if (dso->is_kmod && al.addr > dso->text_end) { + if (dso__is_kmod(dso) && al.addr > dso__text_end(dso)) { pr_debug("skipping the module address %#"PRIx64" after text end\n", al.addr); goto out; } @@ -315,7 +315,7 @@ static int read_object_code(u64 addr, size_t len, u8 cpumode, state->done[state->done_cnt++] = map__start(al.map); } - objdump_name = dso->long_name; + objdump_name = dso__long_name(dso); if (dso__needs_decompress(dso)) { if (dso__decompress_kmodule_path(dso, objdump_name, decomp_name, diff --git a/tools/perf/tests/dso-data.c b/tools/perf/tests/dso-data.c index 2d67422c1222..fde4eca84b6f 100644 --- a/tools/perf/tests/dso-data.c +++ b/tools/perf/tests/dso-data.c @@ -228,7 +228,8 @@ static void dsos__delete(int cnt) for (i = 0; i < cnt; i++) { struct dso *dso = dsos[i]; - unlink(dso->name); + dso__data_close(dso); + unlink(dso__name(dso)); dso__put(dso); } @@ -289,14 +290,14 @@ static int test__dso_data_cache(struct test_suite *test __maybe_unused, int subt } /* verify the first one is already open */ - TEST_ASSERT_VAL("dsos[0] is not open", dsos[0]->data.fd != -1); + TEST_ASSERT_VAL("dsos[0] is not open", dso__data(dsos[0])->fd != -1); /* open +1 dso to reach the allowed limit */ fd = dso__data_fd(dsos[i], &machine); TEST_ASSERT_VAL("failed to get fd", fd > 0); /* should force the first one to be closed */ - TEST_ASSERT_VAL("failed to close dsos[0]", dsos[0]->data.fd == -1); + TEST_ASSERT_VAL("failed to close dsos[0]", dso__data(dsos[0])->fd == -1); /* cleanup everything */ dsos__delete(dso_cnt); @@ -371,7 +372,7 @@ static int test__dso_data_reopen(struct test_suite *test __maybe_unused, int sub * dso_0 should get closed, because we reached * the file descriptor limit */ - TEST_ASSERT_VAL("failed to close dso_0", dso_0->data.fd == -1); + TEST_ASSERT_VAL("failed to close dso_0", dso__data(dso_0)->fd == -1); /* open dso_0 */ fd = dso__data_fd(dso_0, &machine); @@ -381,7 +382,7 @@ static int test__dso_data_reopen(struct test_suite *test __maybe_unused, int sub * dso_1 should get closed, because we reached * the file descriptor limit */ - TEST_ASSERT_VAL("failed to close dso_1", dso_1->data.fd == -1); + TEST_ASSERT_VAL("failed to close dso_1", dso__data(dso_1)->fd == -1); /* cleanup everything */ close(fd_extra); diff --git a/tools/perf/tests/hists_common.c b/tools/perf/tests/hists_common.c index d08add0f4da6..187f12f5bc21 100644 --- a/tools/perf/tests/hists_common.c +++ b/tools/perf/tests/hists_common.c @@ -146,7 +146,7 @@ struct machine *setup_fake_machine(struct machines *machines) goto out; } - symbols__insert(&dso->symbols, sym); + symbols__insert(dso__symbols(dso), sym); } dso__put(dso); @@ -183,7 +183,7 @@ void print_hists_in(struct hists *hists) pr_info("%2d: entry: %-8s [%-8s] %20s: period = %"PRIu64"\n", i, thread__comm_str(he->thread), - dso->short_name, + dso__short_name(dso), he->ms.sym->name, he->stat.period); } @@ -212,7 +212,7 @@ void print_hists_out(struct hists *hists) pr_info("%2d: entry: %8s:%5d [%-8s] %20s: period = %"PRIu64"/%"PRIu64"\n", i, thread__comm_str(he->thread), thread__tid(he->thread), - dso->short_name, + dso__short_name(dso), he->ms.sym->name, he->stat.period, he->stat_acc ? he->stat_acc->period : 0); } diff --git a/tools/perf/tests/hists_cumulate.c b/tools/perf/tests/hists_cumulate.c index 71dacb0fec4d..1e0f5a310fd5 100644 --- a/tools/perf/tests/hists_cumulate.c +++ b/tools/perf/tests/hists_cumulate.c @@ -164,11 +164,11 @@ static void put_fake_samples(void) typedef int (*test_fn_t)(struct evsel *, struct machine *); #define COMM(he) (thread__comm_str(he->thread)) -#define DSO(he) (map__dso(he->ms.map)->short_name) +#define DSO(he) (dso__short_name(map__dso(he->ms.map))) #define SYM(he) (he->ms.sym->name) #define CPU(he) (he->cpu) #define DEPTH(he) (he->callchain->max_depth) -#define CDSO(cl) (map__dso(cl->ms.map)->short_name) +#define CDSO(cl) (dso__short_name(map__dso(cl->ms.map))) #define CSYM(cl) (cl->ms.sym->name) struct result { diff --git a/tools/perf/tests/hists_output.c b/tools/perf/tests/hists_output.c index ba1cccf57049..33b5cc8352a7 100644 --- a/tools/perf/tests/hists_output.c +++ b/tools/perf/tests/hists_output.c @@ -129,7 +129,7 @@ static void put_fake_samples(void) typedef int (*test_fn_t)(struct evsel *, struct machine *); #define COMM(he) (thread__comm_str(he->thread)) -#define DSO(he) (map__dso(he->ms.map)->short_name) +#define DSO(he) (dso__short_name(map__dso(he->ms.map))) #define SYM(he) (he->ms.sym->name) #define CPU(he) (he->cpu) #define PID(he) (thread__tid(he->thread)) diff --git a/tools/perf/tests/maps.c b/tools/perf/tests/maps.c index b15417a0d617..4f1f9385ea9c 100644 --- a/tools/perf/tests/maps.c +++ b/tools/perf/tests/maps.c @@ -26,7 +26,7 @@ static int check_maps_cb(struct map *map, void *data) if (map__start(map) != merged->start || map__end(map) != merged->end || - strcmp(map__dso(map)->name, merged->name) || + strcmp(dso__name(map__dso(map)), merged->name) || refcount_read(map__refcnt(map)) != 1) { return 1; } @@ -39,7 +39,7 @@ static int failed_cb(struct map *map, void *data __maybe_unused) pr_debug("\tstart: %" PRIu64 " end: %" PRIu64 " name: '%s' refcnt: %d\n", map__start(map), map__end(map), - map__dso(map)->name, + dso__name(map__dso(map)), refcount_read(map__refcnt(map))); return 0; diff --git a/tools/perf/tests/symbols.c b/tools/perf/tests/symbols.c index d208105919ed..ee20a366f32f 100644 --- a/tools/perf/tests/symbols.c +++ b/tools/perf/tests/symbols.c @@ -81,7 +81,7 @@ static int create_map(struct test_info *ti, char *filename, struct map **map_p) * If 'filename' matches a current kernel module, must use a kernel * map. Find the one that already exists. */ - if (dso && dso->kernel) { + if (dso && dso__kernel(dso) != DSO_SPACE__USER) { *map_p = find_module_map(ti->machine, dso); dso__put(dso); if (!*map_p) { @@ -116,7 +116,7 @@ static int test_dso(struct dso *dso) if (verbose > 1) dso__fprintf(dso, stderr); - for (nd = rb_first_cached(&dso->symbols); nd; nd = rb_next(nd)) { + for (nd = rb_first_cached(dso__symbols(dso)); nd; nd = rb_next(nd)) { struct symbol *sym = rb_entry(nd, struct symbol, rb_node); if (sym->type != STT_FUNC && sym->type != STT_GNU_IFUNC) @@ -145,7 +145,7 @@ static int subdivided_dso_cb(struct dso *dso, struct machine *machine __maybe_un { struct dso *text_dso = d; - if (dso != text_dso && strstarts(dso->short_name, text_dso->short_name)) + if (dso != text_dso && strstarts(dso__short_name(dso), dso__short_name(text_dso))) if (test_dso(dso) != TEST_OK) return -1; @@ -190,7 +190,7 @@ static int test_file(struct test_info *ti, char *filename) ret = test_dso(dso); /* Module dso is split into many dsos by section */ - if (ret == TEST_OK && dso->kernel) + if (ret == TEST_OK && dso__kernel(dso) != DSO_SPACE__USER) ret = process_subdivided_dso(ti->machine, dso); out_put: map__put(map); diff --git a/tools/perf/tests/vmlinux-kallsyms.c b/tools/perf/tests/vmlinux-kallsyms.c index fecbf851bb2e..e30fd55f8e51 100644 --- a/tools/perf/tests/vmlinux-kallsyms.c +++ b/tools/perf/tests/vmlinux-kallsyms.c @@ -129,7 +129,7 @@ static int test__vmlinux_matches_kallsyms_cb1(struct map *map, void *data) * cases. */ struct map *pair = maps__find_by_name(args->kallsyms.kmaps, - (dso->kernel ? dso->short_name : dso->name)); + (dso__kernel(dso) ? dso__short_name(dso) : dso__name(dso))); if (pair) { map__set_priv(pair, 1); @@ -162,11 +162,11 @@ static int test__vmlinux_matches_kallsyms_cb2(struct map *map, void *data) } pr_info("WARN: %" PRIx64 "-%" PRIx64 " %" PRIx64 " %s in kallsyms as", - map__start(map), map__end(map), map__pgoff(map), dso->name); + map__start(map), map__end(map), map__pgoff(map), dso__name(dso)); if (mem_end != map__end(pair)) pr_info(":\nWARN: *%" PRIx64 "-%" PRIx64 " %" PRIx64, map__start(pair), map__end(pair), map__pgoff(pair)); - pr_info(" %s\n", dso->name); + pr_info(" %s\n", dso__name(dso)); map__set_priv(pair, 1); } map__put(pair); diff --git a/tools/perf/ui/browsers/annotate.c b/tools/perf/ui/browsers/annotate.c index 0dd429cf612c..ea986430241e 100644 --- a/tools/perf/ui/browsers/annotate.c +++ b/tools/perf/ui/browsers/annotate.c @@ -438,7 +438,7 @@ static int sym_title(struct symbol *sym, struct map *map, char *title, size_t sz, int percent_type) { return snprintf(title, sz, "%s %s [Percent: %s]", sym->name, - map__dso(map)->long_name, + dso__long_name(map__dso(map)), percent_type_str(percent_type)); } @@ -967,14 +967,14 @@ int symbol__tui_annotate(struct map_symbol *ms, struct evsel *evsel, return -1; dso = map__dso(ms->map); - if (dso->annotate_warned) + if (dso__annotate_warned(dso)) return -1; if (not_annotated || !sym->annotate2) { err = symbol__annotate2(ms, evsel, &browser.arch); if (err) { char msg[BUFSIZ]; - dso->annotate_warned = true; + dso__set_annotate_warned(dso); symbol__strerror_disassemble(ms, err, msg, sizeof(msg)); ui__error("Couldn't annotate %s:\n%s", sym->name, msg); return -1; diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c index 71b32591d61a..b7219df51236 100644 --- a/tools/perf/ui/browsers/hists.c +++ b/tools/perf/ui/browsers/hists.c @@ -2489,7 +2489,7 @@ add_annotate_opt(struct hist_browser *browser __maybe_unused, { struct dso *dso; - if (!ms->map || (dso = map__dso(ms->map)) == NULL || dso->annotate_warned) + if (!ms->map || (dso = map__dso(ms->map)) == NULL || dso__annotate_warned(dso)) return 0; if (!ms->sym) @@ -2608,7 +2608,7 @@ static int hists_browser__zoom_map(struct hist_browser *browser, struct map *map } else { struct dso *dso = map__dso(map); ui_helpline__fpush("To zoom out press ESC or ENTER + \"Zoom out of %s DSO\"", - __map__is_kernel(map) ? "the Kernel" : dso->short_name); + __map__is_kernel(map) ? "the Kernel" : dso__short_name(dso)); browser->hists->dso_filter = dso; perf_hpp__set_elide(HISTC_DSO, true); pstack__push(browser->pstack, &browser->hists->dso_filter); @@ -2634,7 +2634,7 @@ add_dso_opt(struct hist_browser *browser, struct popup_action *act, if (asprintf(optstr, "Zoom %s %s DSO (use the 'k' hotkey to zoom directly into the kernel)", browser->hists->dso_filter ? "out of" : "into", - __map__is_kernel(map) ? "the Kernel" : map__dso(map)->short_name) < 0) + __map__is_kernel(map) ? "the Kernel" : dso__short_name(map__dso(map))) < 0) return 0; act->ms.map = map; @@ -3110,7 +3110,7 @@ do_hotkey: // key came straight from options ui__popup_menu() if (!browser->selection || !browser->selection->map || !map__dso(browser->selection->map) || - map__dso(browser->selection->map)->annotate_warned) { + dso__annotate_warned(map__dso(browser->selection->map))) { continue; } diff --git a/tools/perf/ui/browsers/map.c b/tools/perf/ui/browsers/map.c index 3d1b958d8832..fba55175a935 100644 --- a/tools/perf/ui/browsers/map.c +++ b/tools/perf/ui/browsers/map.c @@ -76,7 +76,7 @@ static int map_browser__run(struct map_browser *browser) { int key; - if (ui_browser__show(&browser->b, map__dso(browser->map)->long_name, + if (ui_browser__show(&browser->b, dso__long_name(map__dso(browser->map)), "Press ESC to exit, %s / to search", verbose > 0 ? "" : "restart with -v to use") < 0) return -1; @@ -106,7 +106,7 @@ int map__browse(struct map *map) { struct map_browser mb = { .b = { - .entries = &map__dso(map)->symbols, + .entries = dso__symbols(map__dso(map)), .refresh = ui_browser__rb_tree_refresh, .seek = ui_browser__rb_tree_seek, .write = map_browser__write, diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 2c98813f95cd..faefa444af1e 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -336,7 +336,7 @@ static struct annotated_data_type *dso__findnew_data_type(struct dso *dso, /* Check existing nodes in dso->data_types tree */ key.self.type_name = type_name; key.self.size = size; - node = rb_find(&key, &dso->data_types, data_type_cmp); + node = rb_find(&key, dso__data_types(dso), data_type_cmp); if (node) { result = rb_entry(node, struct annotated_data_type, node); free(type_name); @@ -357,7 +357,7 @@ static struct annotated_data_type *dso__findnew_data_type(struct dso *dso, if (symbol_conf.annotate_data_member) add_member_types(result, type_die); - rb_add(&result->node, &dso->data_types, data_type_less); + rb_add(&result->node, dso__data_types(dso), data_type_less); return result; } @@ -538,7 +538,7 @@ static struct global_var_entry *global_var__find(struct data_loc_info *dloc, u64 struct dso *dso = map__dso(dloc->ms->map); struct rb_node *node; - node = rb_find((void *)(uintptr_t)addr, &dso->global_vars, global_var_cmp); + node = rb_find((void *)(uintptr_t)addr, dso__global_vars(dso), global_var_cmp); if (node == NULL) return NULL; @@ -569,7 +569,7 @@ static bool global_var__add(struct data_loc_info *dloc, u64 addr, gvar->end = addr + size; gvar->die_offset = dwarf_dieoffset(type_die); - rb_add(&gvar->node, &dso->global_vars, global_var_less); + rb_add(&gvar->node, dso__global_vars(dso), global_var_less); return true; } @@ -672,7 +672,7 @@ static bool get_global_var_type(Dwarf_Die *cu_die, struct data_loc_info *dloc, struct dso *dso = map__dso(dloc->ms->map); Dwarf_Die var_die; - if (RB_EMPTY_ROOT(&dso->global_vars)) + if (RB_EMPTY_ROOT(dso__global_vars(dso))) global_var__collect(dloc); gvar = global_var__find(dloc, var_addr); @@ -889,7 +889,7 @@ static void update_insn_state_x86(struct type_state *state, return; tsr = &state->regs[dst->reg1]; - if (map__dso(dloc->ms->map)->kernel && + if (dso__kernel(map__dso(dloc->ms->map)) && src->segment == INSN_SEG_X86_GS && src->imm) { u64 ip = dloc->ms->sym->start + dl->al.offset; u64 var_addr; @@ -1415,7 +1415,7 @@ static int check_matching_type(struct type_state *state, } check_kernel: - if (map__dso(dloc->ms->map)->kernel) { + if (dso__kernel(map__dso(dloc->ms->map))) { u64 addr; int offset; @@ -1767,7 +1767,7 @@ struct annotated_data_type *find_data_type(struct data_loc_info *dloc) struct dso *dso = map__dso(dloc->ms->map); Dwarf_Die type_die; - dloc->di = debuginfo__new(dso->long_name); + dloc->di = debuginfo__new(dso__long_name(dso)); if (dloc->di == NULL) { pr_debug_dtp("cannot get the debug info\n"); return NULL; @@ -1901,7 +1901,7 @@ static void print_annotated_data_header(struct hist_entry *he, struct evsel *evs } printf("Annotate type: '%s' in %s (%d samples):\n", - he->mem_type->self.type_name, dso->name, nr_samples); + he->mem_type->self.type_name, dso__name(dso), nr_samples); if (evsel__is_group_event(evsel)) { struct evsel *pos; diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index f5b6b5e5e757..d7d55263fc91 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -1103,7 +1103,7 @@ int symbol__annotate_printf(struct map_symbol *ms, struct evsel *evsel) int graph_dotted_len; char buf[512]; - filename = strdup(dso->long_name); + filename = strdup(dso__long_name(dso)); if (!filename) return -ENOMEM; @@ -1268,7 +1268,7 @@ int map_symbol__annotation_dump(struct map_symbol *ms, struct evsel *evsel) } fprintf(fp, "%s() %s\nEvent: %s\n\n", - ms->sym->name, map__dso(ms->map)->long_name, ev_name); + ms->sym->name, dso__long_name(map__dso(ms->map)), ev_name); symbol__annotate_fprintf2(ms->sym, fp); fclose(fp); @@ -1526,7 +1526,7 @@ int symbol__tty_annotate2(struct map_symbol *ms, struct evsel *evsel) if (err) { char msg[BUFSIZ]; - dso->annotate_warned = true; + dso__set_annotate_warned(dso); symbol__strerror_disassemble(ms, err, msg, sizeof(msg)); ui__error("Couldn't annotate %s:\n%s", sym->name, msg); return -1; @@ -1535,13 +1535,12 @@ int symbol__tty_annotate2(struct map_symbol *ms, struct evsel *evsel) if (annotate_opts.print_lines) { srcline_full_filename = annotate_opts.full_path; symbol__calc_lines(ms, &source_line); - print_summary(&source_line, dso->long_name); + print_summary(&source_line, dso__long_name(dso)); } hists__scnprintf_title(hists, buf, sizeof(buf)); fprintf(stdout, "%s, [percent: %s]\n%s() %s\n", - buf, percent_type_str(annotate_opts.percent_type), sym->name, - dso->long_name); + buf, percent_type_str(annotate_opts.percent_type), sym->name, dso__long_name(dso)); symbol__annotate_fprintf2(sym, stdout); annotated_source__purge(symbol__annotation(sym)->src); @@ -1560,7 +1559,7 @@ int symbol__tty_annotate(struct map_symbol *ms, struct evsel *evsel) if (err) { char msg[BUFSIZ]; - dso->annotate_warned = true; + dso__set_annotate_warned(dso); symbol__strerror_disassemble(ms, err, msg, sizeof(msg)); ui__error("Couldn't annotate %s:\n%s", sym->name, msg); return -1; @@ -1571,7 +1570,7 @@ int symbol__tty_annotate(struct map_symbol *ms, struct evsel *evsel) if (annotate_opts.print_lines) { srcline_full_filename = annotate_opts.full_path; symbol__calc_lines(ms, &source_line); - print_summary(&source_line, dso->long_name); + print_summary(&source_line, dso__long_name(dso)); } symbol__annotate_printf(ms, evsel); @@ -2400,7 +2399,7 @@ retry: } /* This CPU access in kernel - pretend PC-relative addressing */ - if (map__dso(ms->map)->kernel && arch__is(arch, "x86") && + if (dso__kernel(map__dso(ms->map)) && arch__is(arch, "x86") && op_loc->segment == INSN_SEG_X86_GS && op_loc->imm) { dloc.var_addr = op_loc->offset; op_loc->reg1 = DWARF_REG_PC; diff --git a/tools/perf/util/auxtrace.c b/tools/perf/util/auxtrace.c index cfa2153d4611..3fd350de0051 100644 --- a/tools/perf/util/auxtrace.c +++ b/tools/perf/util/auxtrace.c @@ -2654,7 +2654,7 @@ static int addr_filter__entire_dso(struct addr_filter *filt, struct dso *dso) } filt->addr = 0; - filt->size = dso->data.file_size; + filt->size = dso__data(dso)->file_size; return 0; } diff --git a/tools/perf/util/block-info.c b/tools/perf/util/block-info.c index dec910989701..895ee8adf3b3 100644 --- a/tools/perf/util/block-info.c +++ b/tools/perf/util/block-info.c @@ -319,7 +319,7 @@ static int block_dso_entry(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp, if (map && map__dso(map)) { return scnprintf(hpp->buf, hpp->size, "%*s", block_fmt->width, - map__dso(map)->short_name); + dso__short_name(map__dso(map))); } return scnprintf(hpp->buf, hpp->size, "%*s", block_fmt->width, diff --git a/tools/perf/util/bpf-event.c b/tools/perf/util/bpf-event.c index 83709146a48a..827695cd0408 100644 --- a/tools/perf/util/bpf-event.c +++ b/tools/perf/util/bpf-event.c @@ -59,10 +59,10 @@ static int machine__process_bpf_event_load(struct machine *machine, if (map) { struct dso *dso = map__dso(map); - dso->binary_type = DSO_BINARY_TYPE__BPF_PROG_INFO; - dso->bpf_prog.id = id; - dso->bpf_prog.sub_id = i; - dso->bpf_prog.env = env; + dso__set_binary_type(dso, DSO_BINARY_TYPE__BPF_PROG_INFO); + dso__bpf_prog(dso)->id = id; + dso__bpf_prog(dso)->sub_id = i; + dso__bpf_prog(dso)->env = env; map__put(map); } } diff --git a/tools/perf/util/build-id.c b/tools/perf/util/build-id.c index 864bc26b6b46..83a1581e8cf1 100644 --- a/tools/perf/util/build-id.c +++ b/tools/perf/util/build-id.c @@ -60,7 +60,7 @@ int build_id__mark_dso_hit(struct perf_tool *tool __maybe_unused, addr_location__init(&al); if (thread__find_map(thread, sample->cpumode, sample->ip, &al)) - map__dso(al.map)->hit = 1; + dso__set_hit(map__dso(al.map)); addr_location__exit(&al); thread__put(thread); @@ -272,10 +272,10 @@ char *__dso__build_id_filename(const struct dso *dso, char *bf, size_t size, bool alloc = (bf == NULL); int ret; - if (!dso->has_build_id) + if (!dso__has_build_id(dso)) return NULL; - build_id__sprintf(&dso->bid, sbuild_id); + build_id__sprintf(dso__bid_const(dso), sbuild_id); linkname = build_id_cache__linkname(sbuild_id, NULL, 0); if (!linkname) return NULL; @@ -340,25 +340,25 @@ static int machine__write_buildid_table_cb(struct dso *dso, void *data) size_t name_len; bool in_kernel = false; - if (!dso->has_build_id) + if (!dso__has_build_id(dso)) return 0; - if (!dso->hit && !dso__is_vdso(dso)) + if (!dso__hit(dso) && !dso__is_vdso(dso)) return 0; if (dso__is_vdso(dso)) { - name = dso->short_name; - name_len = dso->short_name_len; + name = dso__short_name(dso); + name_len = dso__short_name_len(dso); } else if (dso__is_kcore(dso)) { name = args->machine->mmap_name; name_len = strlen(name); } else { - name = dso->long_name; - name_len = dso->long_name_len; + name = dso__long_name(dso); + name_len = dso__long_name_len(dso); } - in_kernel = dso->kernel || is_kernel_module(name, PERF_RECORD_MISC_CPUMODE_UNKNOWN); - return write_buildid(name, name_len, &dso->bid, args->machine->pid, + in_kernel = dso__kernel(dso) || is_kernel_module(name, PERF_RECORD_MISC_CPUMODE_UNKNOWN); + return write_buildid(name, name_len, dso__bid(dso), args->machine->pid, in_kernel ? args->kmisc : args->umisc, args->fd); } @@ -876,11 +876,11 @@ static bool dso__build_id_mismatch(struct dso *dso, const char *name) struct build_id bid; bool ret = false; - mutex_lock(&dso->lock); - if (filename__read_build_id_ns(name, &bid, dso->nsinfo) >= 0) + mutex_lock(dso__lock(dso)); + if (filename__read_build_id_ns(name, &bid, dso__nsinfo(dso)) >= 0) ret = !dso__build_id_equal(dso, &bid); - mutex_unlock(&dso->lock); + mutex_unlock(dso__lock(dso)); return ret; } @@ -890,13 +890,13 @@ static int dso__cache_build_id(struct dso *dso, struct machine *machine, { bool is_kallsyms = dso__is_kallsyms(dso); bool is_vdso = dso__is_vdso(dso); - const char *name = dso->long_name; + const char *name = dso__long_name(dso); const char *proper_name = NULL; const char *root_dir = NULL; char *allocated_name = NULL; int ret = 0; - if (!dso->has_build_id) + if (!dso__has_build_id(dso)) return 0; if (dso__is_kcore(dso)) { @@ -921,10 +921,10 @@ static int dso__cache_build_id(struct dso *dso, struct machine *machine, if (!is_kallsyms && dso__build_id_mismatch(dso, name)) goto out_free; - mutex_lock(&dso->lock); - ret = build_id_cache__add_b(&dso->bid, name, dso->nsinfo, + mutex_lock(dso__lock(dso)); + ret = build_id_cache__add_b(dso__bid(dso), name, dso__nsinfo(dso), is_kallsyms, is_vdso, proper_name, root_dir); - mutex_unlock(&dso->lock); + mutex_unlock(dso__lock(dso)); out_free: free(allocated_name); return ret; diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index 7517d16c02ec..68feed871809 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -1205,7 +1205,7 @@ char *callchain_list__sym_name(struct callchain_list *cl, if (show_dso) scnprintf(bf + printed, bfsize - printed, " %s", cl->ms.map ? - map__dso(cl->ms.map)->short_name : + dso__short_name(map__dso(cl->ms.map)) : "unknown"); return bf; diff --git a/tools/perf/util/data-convert-json.c b/tools/perf/util/data-convert-json.c index 09d57efd2d9d..3cf64f5b23ee 100644 --- a/tools/perf/util/data-convert-json.c +++ b/tools/perf/util/data-convert-json.c @@ -134,7 +134,7 @@ static void output_sample_callchain_entry(struct perf_tool *tool, output_json_key_string(out, false, 5, "symbol", al->sym->name); if (dso) { - const char *dso_name = dso->short_name; + const char *dso_name = dso__short_name(dso); if (dso_name && strlen(dso_name) > 0) { fputc(',', out); diff --git a/tools/perf/util/db-export.c b/tools/perf/util/db-export.c index 106429155c2e..50f916374d87 100644 --- a/tools/perf/util/db-export.c +++ b/tools/perf/util/db-export.c @@ -146,10 +146,10 @@ int db_export__comm_thread(struct db_export *dbe, struct comm *comm, int db_export__dso(struct db_export *dbe, struct dso *dso, struct machine *machine) { - if (dso->db_id) + if (dso__db_id(dso)) return 0; - dso->db_id = ++dbe->dso_last_db_id; + dso__set_db_id(dso, ++dbe->dso_last_db_id); if (dbe->export_dso) return dbe->export_dso(dbe, dso, machine); @@ -184,7 +184,7 @@ static int db_ids_from_al(struct db_export *dbe, struct addr_location *al, err = db_export__dso(dbe, dso, maps__machine(al->maps)); if (err) return err; - *dso_db_id = dso->db_id; + *dso_db_id = dso__db_id(dso); if (!al->sym) { al->sym = symbol__new(al->addr, 0, 0, 0, "unknown"); diff --git a/tools/perf/util/disasm.c b/tools/perf/util/disasm.c index 6d1125e687b7..72aec8f61b94 100644 --- a/tools/perf/util/disasm.c +++ b/tools/perf/util/disasm.c @@ -1054,8 +1054,8 @@ int symbol__strerror_disassemble(struct map_symbol *ms, int errnum, char *buf, s char bf[SBUILD_ID_SIZE + 15] = " with build id "; char *build_id_msg = NULL; - if (dso->has_build_id) { - build_id__sprintf(&dso->bid, bf + 15); + if (dso__has_build_id(dso)) { + build_id__sprintf(dso__bid(dso), bf + 15); build_id_msg = bf; } scnprintf(buf, buflen, @@ -1077,11 +1077,11 @@ int symbol__strerror_disassemble(struct map_symbol *ms, int errnum, char *buf, s scnprintf(buf, buflen, "Problems while parsing the CPUID in the arch specific initialization."); break; case SYMBOL_ANNOTATE_ERRNO__BPF_INVALID_FILE: - scnprintf(buf, buflen, "Invalid BPF file: %s.", dso->long_name); + scnprintf(buf, buflen, "Invalid BPF file: %s.", dso__long_name(dso)); break; case SYMBOL_ANNOTATE_ERRNO__BPF_MISSING_BTF: scnprintf(buf, buflen, "The %s BPF file has no BTF section, compile with -g or use pahole -J.", - dso->long_name); + dso__long_name(dso)); break; default: scnprintf(buf, buflen, "Internal error: Invalid %d error code\n", errnum); @@ -1099,7 +1099,7 @@ static int dso__disassemble_filename(struct dso *dso, char *filename, size_t fil char *pos; int len; - if (dso->symtab_type == DSO_BINARY_TYPE__KALLSYMS && + if (dso__symtab_type(dso) == DSO_BINARY_TYPE__KALLSYMS && !dso__is_kcore(dso)) return SYMBOL_ANNOTATE_ERRNO__NO_VMLINUX; @@ -1108,7 +1108,7 @@ static int dso__disassemble_filename(struct dso *dso, char *filename, size_t fil __symbol__join_symfs(filename, filename_size, build_id_filename); free(build_id_filename); } else { - if (dso->has_build_id) + if (dso__has_build_id(dso)) return ENOMEM; goto fallback; } @@ -1142,22 +1142,22 @@ fallback: * cache, or is just a kallsyms file, well, lets hope that this * DSO is the same as when 'perf record' ran. */ - if (dso->kernel && dso->long_name[0] == '/') - snprintf(filename, filename_size, "%s", dso->long_name); + if (dso__kernel(dso) && dso__long_name(dso)[0] == '/') + snprintf(filename, filename_size, "%s", dso__long_name(dso)); else - __symbol__join_symfs(filename, filename_size, dso->long_name); + __symbol__join_symfs(filename, filename_size, dso__long_name(dso)); - mutex_lock(&dso->lock); - if (access(filename, R_OK) && errno == ENOENT && dso->nsinfo) { + mutex_lock(dso__lock(dso)); + if (access(filename, R_OK) && errno == ENOENT && dso__nsinfo(dso)) { char *new_name = dso__filename_with_chroot(dso, filename); if (new_name) { strlcpy(filename, new_name, filename_size); free(new_name); } } - mutex_unlock(&dso->lock); - } else if (dso->binary_type == DSO_BINARY_TYPE__NOT_FOUND) { - dso->binary_type = DSO_BINARY_TYPE__BUILD_ID_CACHE; + mutex_unlock(dso__lock(dso)); + } else if (dso__binary_type(dso) == DSO_BINARY_TYPE__NOT_FOUND) { + dso__set_binary_type(dso, DSO_BINARY_TYPE__BUILD_ID_CACHE); } free(build_id_path); @@ -1425,7 +1425,7 @@ static void print_capstone_detail(cs_insn *insn, char *buf, size_t len, orig_addr = addr + insn->size + op->mem.disp; addr = map__objdump_2mem(map, orig_addr); - if (map__dso(map)->kernel) { + if (dso__kernel(map__dso(map))) { /* * The kernel maps can be splitted into sections, * let's find the map first and the search the symbol. @@ -1479,7 +1479,7 @@ static int symbol__disassemble_capstone(char *filename, struct symbol *sym, if (args->options->objdump_path) return -1; - nsinfo__mountns_enter(dso->nsinfo, &nsc); + nsinfo__mountns_enter(dso__nsinfo(dso), &nsc); fd = open(filename, O_RDONLY); nsinfo__mountns_exit(&nsc); if (fd < 0) @@ -1679,13 +1679,13 @@ int symbol__disassemble(struct symbol *sym, struct annotate_args *args) map__unmap_ip(map, sym->end)); pr_debug("annotating [%p] %30s : [%p] %30s\n", - dso, dso->long_name, sym, sym->name); + dso, dso__long_name(dso), sym, sym->name); - if (dso->binary_type == DSO_BINARY_TYPE__BPF_PROG_INFO) { + if (dso__binary_type(dso) == DSO_BINARY_TYPE__BPF_PROG_INFO) { return symbol__disassemble_bpf(sym, args); - } else if (dso->binary_type == DSO_BINARY_TYPE__BPF_IMAGE) { + } else if (dso__binary_type(dso) == DSO_BINARY_TYPE__BPF_IMAGE) { return symbol__disassemble_bpf_image(sym, args); - } else if (dso->binary_type == DSO_BINARY_TYPE__NOT_FOUND) { + } else if (dso__binary_type(dso) == DSO_BINARY_TYPE__NOT_FOUND) { return -1; } else if (dso__is_kcore(dso)) { kce.kcore_filename = symfs_filename; diff --git a/tools/perf/util/dlfilter.c b/tools/perf/util/dlfilter.c index 908e16813722..7d180bdaedbc 100644 --- a/tools/perf/util/dlfilter.c +++ b/tools/perf/util/dlfilter.c @@ -33,13 +33,13 @@ static void al_to_d_al(struct addr_location *al, struct perf_dlfilter_al *d_al) if (al->map) { struct dso *dso = map__dso(al->map); - if (symbol_conf.show_kernel_path && dso->long_name) - d_al->dso = dso->long_name; + if (symbol_conf.show_kernel_path && dso__long_name(dso)) + d_al->dso = dso__long_name(dso); else - d_al->dso = dso->name; - d_al->is_64_bit = dso->is_64_bit; - d_al->buildid_size = dso->bid.size; - d_al->buildid = dso->bid.data; + d_al->dso = dso__name(dso); + d_al->is_64_bit = dso__is_64_bit(dso); + d_al->buildid_size = dso__bid(dso)->size; + d_al->buildid = dso__bid(dso)->data; } else { d_al->dso = NULL; d_al->is_64_bit = 0; diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c index 3caca60a6ce3..27db65e96e04 100644 --- a/tools/perf/util/dso.c +++ b/tools/perf/util/dso.c @@ -40,6 +40,12 @@ static const char * const debuglink_paths[] = { "/usr/lib/debug%s/%s" }; +void dso__set_nsinfo(struct dso *dso, struct nsinfo *nsi) +{ + nsinfo__put(RC_CHK_ACCESS(dso)->nsinfo); + RC_CHK_ACCESS(dso)->nsinfo = nsi; +} + char dso__symtab_origin(const struct dso *dso) { static const char origin[] = { @@ -63,14 +69,14 @@ char dso__symtab_origin(const struct dso *dso) [DSO_BINARY_TYPE__GUEST_VMLINUX] = 'V', }; - if (dso == NULL || dso->symtab_type == DSO_BINARY_TYPE__NOT_FOUND) + if (dso == NULL || dso__symtab_type(dso) == DSO_BINARY_TYPE__NOT_FOUND) return '!'; - return origin[dso->symtab_type]; + return origin[dso__symtab_type(dso)]; } bool dso__is_object_file(const struct dso *dso) { - switch (dso->binary_type) { + switch (dso__binary_type(dso)) { case DSO_BINARY_TYPE__KALLSYMS: case DSO_BINARY_TYPE__GUEST_KALLSYMS: case DSO_BINARY_TYPE__JAVA_JIT: @@ -117,7 +123,7 @@ int dso__read_binary_type_filename(const struct dso *dso, char symfile[PATH_MAX]; unsigned int i; - len = __symbol__join_symfs(filename, size, dso->long_name); + len = __symbol__join_symfs(filename, size, dso__long_name(dso)); last_slash = filename + len; while (last_slash != filename && *last_slash != '/') last_slash--; @@ -159,12 +165,12 @@ int dso__read_binary_type_filename(const struct dso *dso, case DSO_BINARY_TYPE__FEDORA_DEBUGINFO: len = __symbol__join_symfs(filename, size, "/usr/lib/debug"); - snprintf(filename + len, size - len, "%s.debug", dso->long_name); + snprintf(filename + len, size - len, "%s.debug", dso__long_name(dso)); break; case DSO_BINARY_TYPE__UBUNTU_DEBUGINFO: len = __symbol__join_symfs(filename, size, "/usr/lib/debug"); - snprintf(filename + len, size - len, "%s", dso->long_name); + snprintf(filename + len, size - len, "%s", dso__long_name(dso)); break; case DSO_BINARY_TYPE__MIXEDUP_UBUNTU_DEBUGINFO: @@ -173,13 +179,13 @@ int dso__read_binary_type_filename(const struct dso *dso, * /usr/lib/debug/lib when it is expected to be in * /usr/lib/debug/usr/lib */ - if (strlen(dso->long_name) < 9 || - strncmp(dso->long_name, "/usr/lib/", 9)) { + if (strlen(dso__long_name(dso)) < 9 || + strncmp(dso__long_name(dso), "/usr/lib/", 9)) { ret = -1; break; } len = __symbol__join_symfs(filename, size, "/usr/lib/debug"); - snprintf(filename + len, size - len, "%s", dso->long_name + 4); + snprintf(filename + len, size - len, "%s", dso__long_name(dso) + 4); break; case DSO_BINARY_TYPE__OPENEMBEDDED_DEBUGINFO: @@ -187,29 +193,29 @@ int dso__read_binary_type_filename(const struct dso *dso, const char *last_slash; size_t dir_size; - last_slash = dso->long_name + dso->long_name_len; - while (last_slash != dso->long_name && *last_slash != '/') + last_slash = dso__long_name(dso) + dso__long_name_len(dso); + while (last_slash != dso__long_name(dso) && *last_slash != '/') last_slash--; len = __symbol__join_symfs(filename, size, ""); - dir_size = last_slash - dso->long_name + 2; + dir_size = last_slash - dso__long_name(dso) + 2; if (dir_size > (size - len)) { ret = -1; break; } - len += scnprintf(filename + len, dir_size, "%s", dso->long_name); + len += scnprintf(filename + len, dir_size, "%s", dso__long_name(dso)); len += scnprintf(filename + len , size - len, ".debug%s", last_slash); break; } case DSO_BINARY_TYPE__BUILDID_DEBUGINFO: - if (!dso->has_build_id) { + if (!dso__has_build_id(dso)) { ret = -1; break; } - build_id__sprintf(&dso->bid, build_id_hex); + build_id__sprintf(dso__bid_const(dso), build_id_hex); len = __symbol__join_symfs(filename, size, "/usr/lib/debug/.build-id/"); snprintf(filename + len, size - len, "%.2s/%s.debug", build_id_hex, build_id_hex + 2); @@ -218,23 +224,23 @@ int dso__read_binary_type_filename(const struct dso *dso, case DSO_BINARY_TYPE__VMLINUX: case DSO_BINARY_TYPE__GUEST_VMLINUX: case DSO_BINARY_TYPE__SYSTEM_PATH_DSO: - __symbol__join_symfs(filename, size, dso->long_name); + __symbol__join_symfs(filename, size, dso__long_name(dso)); break; case DSO_BINARY_TYPE__GUEST_KMODULE: case DSO_BINARY_TYPE__GUEST_KMODULE_COMP: path__join3(filename, size, symbol_conf.symfs, - root_dir, dso->long_name); + root_dir, dso__long_name(dso)); break; case DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE: case DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE_COMP: - __symbol__join_symfs(filename, size, dso->long_name); + __symbol__join_symfs(filename, size, dso__long_name(dso)); break; case DSO_BINARY_TYPE__KCORE: case DSO_BINARY_TYPE__GUEST_KCORE: - snprintf(filename, size, "%s", dso->long_name); + snprintf(filename, size, "%s", dso__long_name(dso)); break; default: @@ -310,8 +316,8 @@ bool is_kernel_module(const char *pathname, int cpumode) bool dso__needs_decompress(struct dso *dso) { - return dso->symtab_type == DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE_COMP || - dso->symtab_type == DSO_BINARY_TYPE__GUEST_KMODULE_COMP; + return dso__symtab_type(dso) == DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE_COMP || + dso__symtab_type(dso) == DSO_BINARY_TYPE__GUEST_KMODULE_COMP; } int filename__decompress(const char *name, char *pathname, @@ -363,11 +369,10 @@ static int decompress_kmodule(struct dso *dso, const char *name, if (!dso__needs_decompress(dso)) return -1; - if (dso->comp == COMP_ID__NONE) + if (dso__comp(dso) == COMP_ID__NONE) return -1; - return filename__decompress(name, pathname, len, dso->comp, - &dso->load_errno); + return filename__decompress(name, pathname, len, dso__comp(dso), dso__load_errno(dso)); } int dso__decompress_kmodule_fd(struct dso *dso, const char *name) @@ -468,17 +473,17 @@ void dso__set_module_info(struct dso *dso, struct kmod_path *m, struct machine *machine) { if (machine__is_host(machine)) - dso->symtab_type = DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE; + dso__set_symtab_type(dso, DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE); else - dso->symtab_type = DSO_BINARY_TYPE__GUEST_KMODULE; + dso__set_symtab_type(dso, DSO_BINARY_TYPE__GUEST_KMODULE); /* _KMODULE_COMP should be next to _KMODULE */ if (m->kmod && m->comp) { - dso->symtab_type++; - dso->comp = m->comp; + dso__set_symtab_type(dso, dso__symtab_type(dso) + 1); + dso__set_comp(dso, m->comp); } - dso->is_kmod = 1; + dso__set_is_kmod(dso); dso__set_short_name(dso, strdup(m->name), true); } @@ -491,13 +496,15 @@ static pthread_mutex_t dso__data_open_lock = PTHREAD_MUTEX_INITIALIZER; static void dso__list_add(struct dso *dso) { - list_add_tail(&dso->data.open_entry, &dso__data_open); + list_add_tail(&dso__data(dso)->open_entry, &dso__data_open); + dso__data(dso)->dso = dso__get(dso); dso__data_open_cnt++; } static void dso__list_del(struct dso *dso) { - list_del_init(&dso->data.open_entry); + list_del_init(&dso__data(dso)->open_entry); + dso__put(dso__data(dso)->dso); WARN_ONCE(dso__data_open_cnt <= 0, "DSO data fd counter out of bounds."); dso__data_open_cnt--; @@ -528,7 +535,7 @@ static int do_open(char *name) char *dso__filename_with_chroot(const struct dso *dso, const char *filename) { - return filename_with_chroot(nsinfo__pid(dso->nsinfo), filename); + return filename_with_chroot(nsinfo__pid(dso__nsinfo_const(dso)), filename); } static int __open_dso(struct dso *dso, struct machine *machine) @@ -541,18 +548,18 @@ static int __open_dso(struct dso *dso, struct machine *machine) if (!name) return -ENOMEM; - mutex_lock(&dso->lock); + mutex_lock(dso__lock(dso)); if (machine) root_dir = machine->root_dir; - if (dso__read_binary_type_filename(dso, dso->binary_type, + if (dso__read_binary_type_filename(dso, dso__binary_type(dso), root_dir, name, PATH_MAX)) goto out; if (!is_regular_file(name)) { char *new_name; - if (errno != ENOENT || dso->nsinfo == NULL) + if (errno != ENOENT || dso__nsinfo(dso) == NULL) goto out; new_name = dso__filename_with_chroot(dso, name); @@ -568,7 +575,7 @@ static int __open_dso(struct dso *dso, struct machine *machine) size_t len = sizeof(newpath); if (dso__decompress_kmodule_path(dso, name, newpath, len) < 0) { - fd = -dso->load_errno; + fd = -(*dso__load_errno(dso)); goto out; } @@ -582,7 +589,7 @@ static int __open_dso(struct dso *dso, struct machine *machine) unlink(name); out: - mutex_unlock(&dso->lock); + mutex_unlock(dso__lock(dso)); free(name); return fd; } @@ -601,13 +608,13 @@ static int open_dso(struct dso *dso, struct machine *machine) int fd; struct nscookie nsc; - if (dso->binary_type != DSO_BINARY_TYPE__BUILD_ID_CACHE) { - mutex_lock(&dso->lock); - nsinfo__mountns_enter(dso->nsinfo, &nsc); - mutex_unlock(&dso->lock); + if (dso__binary_type(dso) != DSO_BINARY_TYPE__BUILD_ID_CACHE) { + mutex_lock(dso__lock(dso)); + nsinfo__mountns_enter(dso__nsinfo(dso), &nsc); + mutex_unlock(dso__lock(dso)); } fd = __open_dso(dso, machine); - if (dso->binary_type != DSO_BINARY_TYPE__BUILD_ID_CACHE) + if (dso__binary_type(dso) != DSO_BINARY_TYPE__BUILD_ID_CACHE) nsinfo__mountns_exit(&nsc); if (fd >= 0) { @@ -624,10 +631,10 @@ static int open_dso(struct dso *dso, struct machine *machine) static void close_data_fd(struct dso *dso) { - if (dso->data.fd >= 0) { - close(dso->data.fd); - dso->data.fd = -1; - dso->data.file_size = 0; + if (dso__data(dso)->fd >= 0) { + close(dso__data(dso)->fd); + dso__data(dso)->fd = -1; + dso__data(dso)->file_size = 0; dso__list_del(dso); } } @@ -646,10 +653,10 @@ static void close_dso(struct dso *dso) static void close_first_dso(void) { - struct dso *dso; + struct dso_data *dso_data; - dso = list_first_entry(&dso__data_open, struct dso, data.open_entry); - close_dso(dso); + dso_data = list_first_entry(&dso__data_open, struct dso_data, open_entry); + close_dso(dso_data->dso); } static rlim_t get_fd_limit(void) @@ -728,28 +735,29 @@ static void try_to_open_dso(struct dso *dso, struct machine *machine) DSO_BINARY_TYPE__NOT_FOUND, }; int i = 0; + struct dso_data *dso_data = dso__data(dso); - if (dso->data.fd >= 0) + if (dso_data->fd >= 0) return; - if (dso->binary_type != DSO_BINARY_TYPE__NOT_FOUND) { - dso->data.fd = open_dso(dso, machine); + if (dso__binary_type(dso) != DSO_BINARY_TYPE__NOT_FOUND) { + dso_data->fd = open_dso(dso, machine); goto out; } do { - dso->binary_type = binary_type_data[i++]; + dso__set_binary_type(dso, binary_type_data[i++]); - dso->data.fd = open_dso(dso, machine); - if (dso->data.fd >= 0) + dso_data->fd = open_dso(dso, machine); + if (dso_data->fd >= 0) goto out; - } while (dso->binary_type != DSO_BINARY_TYPE__NOT_FOUND); + } while (dso__binary_type(dso) != DSO_BINARY_TYPE__NOT_FOUND); out: - if (dso->data.fd >= 0) - dso->data.status = DSO_DATA_STATUS_OK; + if (dso_data->fd >= 0) + dso_data->status = DSO_DATA_STATUS_OK; else - dso->data.status = DSO_DATA_STATUS_ERROR; + dso_data->status = DSO_DATA_STATUS_ERROR; } /** @@ -763,7 +771,7 @@ out: */ int dso__data_get_fd(struct dso *dso, struct machine *machine) { - if (dso->data.status == DSO_DATA_STATUS_ERROR) + if (dso__data(dso)->status == DSO_DATA_STATUS_ERROR) return -1; if (pthread_mutex_lock(&dso__data_open_lock) < 0) @@ -771,10 +779,10 @@ int dso__data_get_fd(struct dso *dso, struct machine *machine) try_to_open_dso(dso, machine); - if (dso->data.fd < 0) + if (dso__data(dso)->fd < 0) pthread_mutex_unlock(&dso__data_open_lock); - return dso->data.fd; + return dso__data(dso)->fd; } void dso__data_put_fd(struct dso *dso __maybe_unused) @@ -786,10 +794,10 @@ bool dso__data_status_seen(struct dso *dso, enum dso_data_status_seen by) { u32 flag = 1 << by; - if (dso->data.status_seen & flag) + if (dso__data(dso)->status_seen & flag) return true; - dso->data.status_seen |= flag; + dso__data(dso)->status_seen |= flag; return false; } @@ -799,12 +807,13 @@ static ssize_t bpf_read(struct dso *dso, u64 offset, char *data) { struct bpf_prog_info_node *node; ssize_t size = DSO__DATA_CACHE_SIZE; + struct dso_bpf_prog *dso_bpf_prog = dso__bpf_prog(dso); u64 len; u8 *buf; - node = perf_env__find_bpf_prog_info(dso->bpf_prog.env, dso->bpf_prog.id); + node = perf_env__find_bpf_prog_info(dso_bpf_prog->env, dso_bpf_prog->id); if (!node || !node->info_linear) { - dso->data.status = DSO_DATA_STATUS_ERROR; + dso__data(dso)->status = DSO_DATA_STATUS_ERROR; return -1; } @@ -822,14 +831,15 @@ static ssize_t bpf_read(struct dso *dso, u64 offset, char *data) static int bpf_size(struct dso *dso) { struct bpf_prog_info_node *node; + struct dso_bpf_prog *dso_bpf_prog = dso__bpf_prog(dso); - node = perf_env__find_bpf_prog_info(dso->bpf_prog.env, dso->bpf_prog.id); + node = perf_env__find_bpf_prog_info(dso_bpf_prog->env, dso_bpf_prog->id); if (!node || !node->info_linear) { - dso->data.status = DSO_DATA_STATUS_ERROR; + dso__data(dso)->status = DSO_DATA_STATUS_ERROR; return -1; } - dso->data.file_size = node->info_linear->info.jited_prog_len; + dso__data(dso)->file_size = node->info_linear->info.jited_prog_len; return 0; } #endif // HAVE_LIBBPF_SUPPORT @@ -837,10 +847,10 @@ static int bpf_size(struct dso *dso) static void dso_cache__free(struct dso *dso) { - struct rb_root *root = &dso->data.cache; + struct rb_root *root = &dso__data(dso)->cache; struct rb_node *next = rb_first(root); - mutex_lock(&dso->lock); + mutex_lock(dso__lock(dso)); while (next) { struct dso_cache *cache; @@ -849,12 +859,12 @@ dso_cache__free(struct dso *dso) rb_erase(&cache->rb_node, root); free(cache); } - mutex_unlock(&dso->lock); + mutex_unlock(dso__lock(dso)); } static struct dso_cache *__dso_cache__find(struct dso *dso, u64 offset) { - const struct rb_root *root = &dso->data.cache; + const struct rb_root *root = &dso__data(dso)->cache; struct rb_node * const *p = &root->rb_node; const struct rb_node *parent = NULL; struct dso_cache *cache; @@ -880,13 +890,13 @@ static struct dso_cache *__dso_cache__find(struct dso *dso, u64 offset) static struct dso_cache * dso_cache__insert(struct dso *dso, struct dso_cache *new) { - struct rb_root *root = &dso->data.cache; + struct rb_root *root = &dso__data(dso)->cache; struct rb_node **p = &root->rb_node; struct rb_node *parent = NULL; struct dso_cache *cache; u64 offset = new->offset; - mutex_lock(&dso->lock); + mutex_lock(dso__lock(dso)); while (*p != NULL) { u64 end; @@ -907,7 +917,7 @@ dso_cache__insert(struct dso *dso, struct dso_cache *new) cache = NULL; out: - mutex_unlock(&dso->lock); + mutex_unlock(dso__lock(dso)); return cache; } @@ -932,18 +942,18 @@ static ssize_t file_read(struct dso *dso, struct machine *machine, pthread_mutex_lock(&dso__data_open_lock); /* - * dso->data.fd might be closed if other thread opened another + * dso__data(dso)->fd might be closed if other thread opened another * file (dso) due to open file limit (RLIMIT_NOFILE). */ try_to_open_dso(dso, machine); - if (dso->data.fd < 0) { - dso->data.status = DSO_DATA_STATUS_ERROR; + if (dso__data(dso)->fd < 0) { + dso__data(dso)->status = DSO_DATA_STATUS_ERROR; ret = -errno; goto out; } - ret = pread(dso->data.fd, data, DSO__DATA_CACHE_SIZE, offset); + ret = pread(dso__data(dso)->fd, data, DSO__DATA_CACHE_SIZE, offset); out: pthread_mutex_unlock(&dso__data_open_lock); return ret; @@ -963,11 +973,11 @@ static struct dso_cache *dso_cache__populate(struct dso *dso, return NULL; } #ifdef HAVE_LIBBPF_SUPPORT - if (dso->binary_type == DSO_BINARY_TYPE__BPF_PROG_INFO) + if (dso__binary_type(dso) == DSO_BINARY_TYPE__BPF_PROG_INFO) *ret = bpf_read(dso, cache_offset, cache->data); else #endif - if (dso->binary_type == DSO_BINARY_TYPE__OOL) + if (dso__binary_type(dso) == DSO_BINARY_TYPE__OOL) *ret = DSO__DATA_CACHE_SIZE; else *ret = file_read(dso, machine, cache_offset, cache->data); @@ -1056,25 +1066,25 @@ static int file_size(struct dso *dso, struct machine *machine) pthread_mutex_lock(&dso__data_open_lock); /* - * dso->data.fd might be closed if other thread opened another + * dso__data(dso)->fd might be closed if other thread opened another * file (dso) due to open file limit (RLIMIT_NOFILE). */ try_to_open_dso(dso, machine); - if (dso->data.fd < 0) { + if (dso__data(dso)->fd < 0) { ret = -errno; - dso->data.status = DSO_DATA_STATUS_ERROR; + dso__data(dso)->status = DSO_DATA_STATUS_ERROR; goto out; } - if (fstat(dso->data.fd, &st) < 0) { + if (fstat(dso__data(dso)->fd, &st) < 0) { ret = -errno; pr_err("dso cache fstat failed: %s\n", str_error_r(errno, sbuf, sizeof(sbuf))); - dso->data.status = DSO_DATA_STATUS_ERROR; + dso__data(dso)->status = DSO_DATA_STATUS_ERROR; goto out; } - dso->data.file_size = st.st_size; + dso__data(dso)->file_size = st.st_size; out: pthread_mutex_unlock(&dso__data_open_lock); @@ -1083,13 +1093,13 @@ out: int dso__data_file_size(struct dso *dso, struct machine *machine) { - if (dso->data.file_size) + if (dso__data(dso)->file_size) return 0; - if (dso->data.status == DSO_DATA_STATUS_ERROR) + if (dso__data(dso)->status == DSO_DATA_STATUS_ERROR) return -1; #ifdef HAVE_LIBBPF_SUPPORT - if (dso->binary_type == DSO_BINARY_TYPE__BPF_PROG_INFO) + if (dso__binary_type(dso) == DSO_BINARY_TYPE__BPF_PROG_INFO) return bpf_size(dso); #endif return file_size(dso, machine); @@ -1108,7 +1118,7 @@ off_t dso__data_size(struct dso *dso, struct machine *machine) return -1; /* For now just estimate dso data size is close to file size */ - return dso->data.file_size; + return dso__data(dso)->file_size; } static ssize_t data_read_write_offset(struct dso *dso, struct machine *machine, @@ -1119,7 +1129,7 @@ static ssize_t data_read_write_offset(struct dso *dso, struct machine *machine, return -1; /* Check the offset sanity. */ - if (offset > dso->data.file_size) + if (offset > dso__data(dso)->file_size) return -1; if (offset + size < offset) @@ -1142,7 +1152,7 @@ static ssize_t data_read_write_offset(struct dso *dso, struct machine *machine, ssize_t dso__data_read_offset(struct dso *dso, struct machine *machine, u64 offset, u8 *data, ssize_t size) { - if (dso->data.status == DSO_DATA_STATUS_ERROR) + if (dso__data(dso)->status == DSO_DATA_STATUS_ERROR) return -1; return data_read_write_offset(dso, machine, offset, data, size, true); @@ -1182,7 +1192,7 @@ ssize_t dso__data_write_cache_offs(struct dso *dso, struct machine *machine, { u8 *data = (u8 *)data_in; /* cast away const to use same fns for r/w */ - if (dso->data.status == DSO_DATA_STATUS_ERROR) + if (dso__data(dso)->status == DSO_DATA_STATUS_ERROR) return -1; return data_read_write_offset(dso, machine, offset, data, size, false); @@ -1235,7 +1245,7 @@ struct dso *machine__findnew_kernel(struct machine *machine, const char *name, */ if (dso != NULL) { dso__set_short_name(dso, short_name, false); - dso->kernel = dso_type; + dso__set_kernel(dso, dso_type); } return dso; @@ -1243,7 +1253,7 @@ struct dso *machine__findnew_kernel(struct machine *machine, const char *name, static void dso__set_long_name_id(struct dso *dso, const char *name, bool name_allocated) { - struct dsos *dsos = dso->dsos; + struct dsos *dsos = dso__dsos(dso); if (name == NULL) return; @@ -1256,12 +1266,12 @@ static void dso__set_long_name_id(struct dso *dso, const char *name, bool name_a down_write(&dsos->lock); } - if (dso->long_name_allocated) - free((char *)dso->long_name); + if (dso__long_name_allocated(dso)) + free((char *)dso__long_name(dso)); - dso->long_name = name; - dso->long_name_len = strlen(name); - dso->long_name_allocated = name_allocated; + RC_CHK_ACCESS(dso)->long_name = name; + RC_CHK_ACCESS(dso)->long_name_len = strlen(name); + dso__set_long_name_allocated(dso, name_allocated); if (dsos) { dsos->sorted = false; @@ -1307,14 +1317,15 @@ bool dso_id__empty(const struct dso_id *id) void __dso__inject_id(struct dso *dso, struct dso_id *id) { - struct dsos *dsos = dso->dsos; + struct dsos *dsos = dso__dsos(dso); + struct dso_id *dso_id = dso__id(dso); /* dsos write lock held by caller. */ - dso->id.maj = id->maj; - dso->id.min = id->min; - dso->id.ino = id->ino; - dso->id.ino_generation = id->ino_generation; + dso_id->maj = id->maj; + dso_id->min = id->min; + dso_id->ino = id->ino; + dso_id->ino_generation = id->ino_generation; if (dsos) dsos->sorted = false; @@ -1334,7 +1345,7 @@ int dso_id__cmp(const struct dso_id *a, const struct dso_id *b) int dso__cmp_id(struct dso *a, struct dso *b) { - return __dso_id__cmp(&a->id, &b->id); + return __dso_id__cmp(dso__id(a), dso__id(b)); } void dso__set_long_name(struct dso *dso, const char *name, bool name_allocated) @@ -1344,7 +1355,7 @@ void dso__set_long_name(struct dso *dso, const char *name, bool name_allocated) void dso__set_short_name(struct dso *dso, const char *name, bool name_allocated) { - struct dsos *dsos = dso->dsos; + struct dsos *dsos = dso__dsos(dso); if (name == NULL) return; @@ -1356,12 +1367,12 @@ void dso__set_short_name(struct dso *dso, const char *name, bool name_allocated) */ down_write(&dsos->lock); } - if (dso->short_name_allocated) - free((char *)dso->short_name); + if (dso__short_name_allocated(dso)) + free((char *)dso__short_name(dso)); - dso->short_name = name; - dso->short_name_len = strlen(name); - dso->short_name_allocated = name_allocated; + RC_CHK_ACCESS(dso)->short_name = name; + RC_CHK_ACCESS(dso)->short_name_len = strlen(name); + dso__set_short_name_allocated(dso, name_allocated); if (dsos) { dsos->sorted = false; @@ -1374,40 +1385,44 @@ int dso__name_len(const struct dso *dso) if (!dso) return strlen("[unknown]"); if (verbose > 0) - return dso->long_name_len; + return dso__long_name_len(dso); - return dso->short_name_len; + return dso__short_name_len(dso); } bool dso__loaded(const struct dso *dso) { - return dso->loaded; + return RC_CHK_ACCESS(dso)->loaded; } bool dso__sorted_by_name(const struct dso *dso) { - return dso->sorted_by_name; + return RC_CHK_ACCESS(dso)->sorted_by_name; } void dso__set_sorted_by_name(struct dso *dso) { - dso->sorted_by_name = true; + RC_CHK_ACCESS(dso)->sorted_by_name = true; } struct dso *dso__new_id(const char *name, struct dso_id *id) { - struct dso *dso = calloc(1, sizeof(*dso) + strlen(name) + 1); + RC_STRUCT(dso) *dso = zalloc(sizeof(*dso) + strlen(name) + 1); + struct dso *res; + struct dso_data *data; - if (dso != NULL) { + if (!dso) + return NULL; + + if (ADD_RC_CHK(res, dso)) { strcpy(dso->name, name); if (id) dso->id = *id; - dso__set_long_name_id(dso, dso->name, false); - dso__set_short_name(dso, dso->name, false); + dso__set_long_name_id(res, dso->name, false); + dso__set_short_name(res, dso->name, false); dso->symbols = RB_ROOT_CACHED; dso->symbol_names = NULL; dso->symbol_names_len = 0; - dso->data.cache = RB_ROOT; dso->inlined_nodes = RB_ROOT_CACHED; dso->srclines = RB_ROOT_CACHED; dso->data_types = RB_ROOT; @@ -1427,12 +1442,16 @@ struct dso *dso__new_id(const char *name, struct dso_id *id) dso->is_kmod = 0; dso->needs_swap = DSO_SWAP__UNSET; dso->comp = COMP_ID__NONE; - INIT_LIST_HEAD(&dso->data.open_entry); mutex_init(&dso->lock); refcount_set(&dso->refcnt, 1); + data = &dso->data; + data->cache = RB_ROOT; + data->fd = -1; + data->status = DSO_DATA_STATUS_UNKNOWN; + INIT_LIST_HEAD(&data->open_entry); + data->dso = NULL; /* Set when on the open_entry list. */ } - - return dso; + return res; } struct dso *dso__new(const char *name) @@ -1442,71 +1461,78 @@ struct dso *dso__new(const char *name) void dso__delete(struct dso *dso) { - if (dso->dsos) - pr_err("DSO %s is still in rbtree when being deleted!\n", dso->long_name); + if (dso__dsos(dso)) + pr_err("DSO %s is still in rbtree when being deleted!\n", dso__long_name(dso)); /* free inlines first, as they reference symbols */ - inlines__tree_delete(&dso->inlined_nodes); - srcline__tree_delete(&dso->srclines); - symbols__delete(&dso->symbols); - dso->symbol_names_len = 0; - zfree(&dso->symbol_names); - annotated_data_type__tree_delete(&dso->data_types); - global_var_type__tree_delete(&dso->global_vars); - - if (dso->short_name_allocated) { - zfree((char **)&dso->short_name); - dso->short_name_allocated = false; + inlines__tree_delete(&RC_CHK_ACCESS(dso)->inlined_nodes); + srcline__tree_delete(&RC_CHK_ACCESS(dso)->srclines); + symbols__delete(&RC_CHK_ACCESS(dso)->symbols); + RC_CHK_ACCESS(dso)->symbol_names_len = 0; + zfree(&RC_CHK_ACCESS(dso)->symbol_names); + annotated_data_type__tree_delete(dso__data_types(dso)); + global_var_type__tree_delete(dso__global_vars(dso)); + + if (RC_CHK_ACCESS(dso)->short_name_allocated) { + zfree((char **)&RC_CHK_ACCESS(dso)->short_name); + RC_CHK_ACCESS(dso)->short_name_allocated = false; } - if (dso->long_name_allocated) { - zfree((char **)&dso->long_name); - dso->long_name_allocated = false; + if (RC_CHK_ACCESS(dso)->long_name_allocated) { + zfree((char **)&RC_CHK_ACCESS(dso)->long_name); + RC_CHK_ACCESS(dso)->long_name_allocated = false; } dso__data_close(dso); - auxtrace_cache__free(dso->auxtrace_cache); + auxtrace_cache__free(RC_CHK_ACCESS(dso)->auxtrace_cache); dso_cache__free(dso); dso__free_a2l(dso); - zfree(&dso->symsrc_filename); - nsinfo__zput(dso->nsinfo); - mutex_destroy(&dso->lock); - free(dso); + zfree(&RC_CHK_ACCESS(dso)->symsrc_filename); + nsinfo__zput(RC_CHK_ACCESS(dso)->nsinfo); + mutex_destroy(dso__lock(dso)); + RC_CHK_FREE(dso); } struct dso *dso__get(struct dso *dso) { - if (dso) - refcount_inc(&dso->refcnt); - return dso; + struct dso *result; + + if (RC_CHK_GET(result, dso)) + refcount_inc(&RC_CHK_ACCESS(dso)->refcnt); + + return result; } void dso__put(struct dso *dso) { - if (dso && refcount_dec_and_test(&dso->refcnt)) + if (dso && refcount_dec_and_test(&RC_CHK_ACCESS(dso)->refcnt)) dso__delete(dso); + else + RC_CHK_PUT(dso); } void dso__set_build_id(struct dso *dso, struct build_id *bid) { - dso->bid = *bid; - dso->has_build_id = 1; + RC_CHK_ACCESS(dso)->bid = *bid; + RC_CHK_ACCESS(dso)->has_build_id = 1; } bool dso__build_id_equal(const struct dso *dso, struct build_id *bid) { - if (dso->bid.size > bid->size && dso->bid.size == BUILD_ID_SIZE) { + const struct build_id *dso_bid = dso__bid_const(dso); + + if (dso_bid->size > bid->size && dso_bid->size == BUILD_ID_SIZE) { /* * For the backward compatibility, it allows a build-id has * trailing zeros. */ - return !memcmp(dso->bid.data, bid->data, bid->size) && - !memchr_inv(&dso->bid.data[bid->size], 0, - dso->bid.size - bid->size); + return !memcmp(dso_bid->data, bid->data, bid->size) && + !memchr_inv(&dso_bid->data[bid->size], 0, + dso_bid->size - bid->size); } - return dso->bid.size == bid->size && - memcmp(dso->bid.data, bid->data, dso->bid.size) == 0; + return dso_bid->size == bid->size && + memcmp(dso_bid->data, bid->data, dso_bid->size) == 0; } void dso__read_running_kernel_build_id(struct dso *dso, struct machine *machine) @@ -1516,8 +1542,8 @@ void dso__read_running_kernel_build_id(struct dso *dso, struct machine *machine) if (machine__is_default_guest(machine)) return; sprintf(path, "%s/sys/kernel/notes", machine->root_dir); - if (sysfs__read_build_id(path, &dso->bid) == 0) - dso->has_build_id = true; + if (sysfs__read_build_id(path, dso__bid(dso)) == 0) + dso__set_has_build_id(dso); } int dso__kernel_module_get_build_id(struct dso *dso, @@ -1528,14 +1554,14 @@ int dso__kernel_module_get_build_id(struct dso *dso, * kernel module short names are of the form "[module]" and * we need just "module" here. */ - const char *name = dso->short_name + 1; + const char *name = dso__short_name(dso) + 1; snprintf(filename, sizeof(filename), "%s/sys/module/%.*s/notes/.note.gnu.build-id", root_dir, (int)strlen(name) - 1, name); - if (sysfs__read_build_id(filename, &dso->bid) == 0) - dso->has_build_id = true; + if (sysfs__read_build_id(filename, dso__bid(dso)) == 0) + dso__set_has_build_id(dso); return 0; } @@ -1544,21 +1570,21 @@ static size_t dso__fprintf_buildid(struct dso *dso, FILE *fp) { char sbuild_id[SBUILD_ID_SIZE]; - build_id__sprintf(&dso->bid, sbuild_id); + build_id__sprintf(dso__bid(dso), sbuild_id); return fprintf(fp, "%s", sbuild_id); } size_t dso__fprintf(struct dso *dso, FILE *fp) { struct rb_node *nd; - size_t ret = fprintf(fp, "dso: %s (", dso->short_name); + size_t ret = fprintf(fp, "dso: %s (", dso__short_name(dso)); - if (dso->short_name != dso->long_name) - ret += fprintf(fp, "%s, ", dso->long_name); + if (dso__short_name(dso) != dso__long_name(dso)) + ret += fprintf(fp, "%s, ", dso__long_name(dso)); ret += fprintf(fp, "%sloaded, ", dso__loaded(dso) ? "" : "NOT "); ret += dso__fprintf_buildid(dso, fp); ret += fprintf(fp, ")\n"); - for (nd = rb_first_cached(&dso->symbols); nd; nd = rb_next(nd)) { + for (nd = rb_first_cached(dso__symbols(dso)); nd; nd = rb_next(nd)) { struct symbol *pos = rb_entry(nd, struct symbol, rb_node); ret += symbol__fprintf(pos, fp); } @@ -1582,7 +1608,7 @@ enum dso_type dso__type(struct dso *dso, struct machine *machine) int dso__strerror_load(struct dso *dso, char *buf, size_t buflen) { - int idx, errnum = dso->load_errno; + int idx, errnum = *dso__load_errno(dso); /* * This must have a same ordering as the enum dso_load_errno. */ diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h index b22dec8b3f3a..f9689dd60de3 100644 --- a/tools/perf/util/dso.h +++ b/tools/perf/util/dso.h @@ -11,6 +11,7 @@ #include #include "build-id.h" #include "mutex.h" +#include struct machine; struct map; @@ -100,26 +101,27 @@ enum dso_load_errno { __DSO_LOAD_ERRNO__END, }; -#define DSO__SWAP(dso, type, val) \ -({ \ - type ____r = val; \ - BUG_ON(dso->needs_swap == DSO_SWAP__UNSET); \ - if (dso->needs_swap == DSO_SWAP__YES) { \ - switch (sizeof(____r)) { \ - case 2: \ - ____r = bswap_16(val); \ - break; \ - case 4: \ - ____r = bswap_32(val); \ - break; \ - case 8: \ - ____r = bswap_64(val); \ - break; \ - default: \ - BUG_ON(1); \ - } \ - } \ - ____r; \ +#define DSO__SWAP(dso, type, val) \ +({ \ + type ____r = val; \ + enum dso_swap_type ___dst = dso__needs_swap(dso); \ + BUG_ON(___dst == DSO_SWAP__UNSET); \ + if (___dst == DSO_SWAP__YES) { \ + switch (sizeof(____r)) { \ + case 2: \ + ____r = bswap_16(val); \ + break; \ + case 4: \ + ____r = bswap_32(val); \ + break; \ + case 8: \ + ____r = bswap_64(val); \ + break; \ + default: \ + BUG_ON(1); \ + } \ + } \ + ____r; \ }) #define DSO__DATA_CACHE_SIZE 4096 @@ -142,9 +144,29 @@ struct dso_cache { char data[]; }; +struct dso_data { + struct rb_root cache; + struct list_head open_entry; + struct dso *dso; + int fd; + int status; + u32 status_seen; + u64 file_size; + u64 elf_base_addr; + u64 debug_frame_offset; + u64 eh_frame_hdr_addr; + u64 eh_frame_hdr_offset; +}; + +struct dso_bpf_prog { + u32 id; + u32 sub_id; + struct perf_env *env; +}; + struct auxtrace_cache; -struct dso { +DECLARE_RC_STRUCT(dso) { struct mutex lock; struct dsos *dsos; struct rb_root_cached symbols; @@ -176,24 +198,9 @@ struct dso { u64 db_id; }; /* bpf prog information */ - struct { - struct perf_env *env; - u32 id; - u32 sub_id; - } bpf_prog; + struct dso_bpf_prog bpf_prog; /* dso data file */ - struct { - struct rb_root cache; - struct list_head open_entry; - u64 file_size; - u64 elf_base_addr; - u64 debug_frame_offset; - u64 eh_frame_hdr_addr; - u64 eh_frame_hdr_offset; - int fd; - int status; - u32 status_seen; - } data; + struct dso_data data; struct dso_id id; unsigned int a2l_fails; int comp; @@ -229,11 +236,388 @@ struct dso { * @n: the 'struct rb_node *' to use as a temporary storage */ #define dso__for_each_symbol(dso, pos, n) \ - symbols__for_each_entry(&(dso)->symbols, pos, n) + symbols__for_each_entry(dso__symbols(dso), pos, n) + +static inline void *dso__a2l(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->a2l; +} + +static inline void dso__set_a2l(struct dso *dso, void *val) +{ + RC_CHK_ACCESS(dso)->a2l = val; +} + +static inline unsigned int dso__a2l_fails(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->a2l_fails; +} + +static inline void dso__set_a2l_fails(struct dso *dso, unsigned int val) +{ + RC_CHK_ACCESS(dso)->a2l_fails = val; +} + +static inline bool dso__adjust_symbols(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->adjust_symbols; +} + +static inline void dso__set_adjust_symbols(struct dso *dso, bool val) +{ + RC_CHK_ACCESS(dso)->adjust_symbols = val; +} + +static inline bool dso__annotate_warned(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->annotate_warned; +} + +static inline void dso__set_annotate_warned(struct dso *dso) +{ + RC_CHK_ACCESS(dso)->annotate_warned = 1; +} + +static inline struct auxtrace_cache *dso__auxtrace_cache(struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->auxtrace_cache; +} + +static inline void dso__set_auxtrace_cache(struct dso *dso, struct auxtrace_cache *cache) +{ + RC_CHK_ACCESS(dso)->auxtrace_cache = cache; +} + +static inline struct build_id *dso__bid(struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->bid; +} + +static inline const struct build_id *dso__bid_const(const struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->bid; +} + +static inline struct dso_bpf_prog *dso__bpf_prog(struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->bpf_prog; +} + +static inline bool dso__has_build_id(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->has_build_id; +} + +static inline void dso__set_has_build_id(struct dso *dso) +{ + RC_CHK_ACCESS(dso)->has_build_id = true; +} + +static inline bool dso__has_srcline(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->has_srcline; +} + +static inline void dso__set_has_srcline(struct dso *dso, bool val) +{ + RC_CHK_ACCESS(dso)->has_srcline = val; +} + +static inline int dso__comp(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->comp; +} + +static inline void dso__set_comp(struct dso *dso, int comp) +{ + RC_CHK_ACCESS(dso)->comp = comp; +} + +static inline struct dso_data *dso__data(struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->data; +} + +static inline u64 dso__db_id(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->db_id; +} + +static inline void dso__set_db_id(struct dso *dso, u64 db_id) +{ + RC_CHK_ACCESS(dso)->db_id = db_id; +} + +static inline struct dsos *dso__dsos(struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->dsos; +} + +static inline void dso__set_dsos(struct dso *dso, struct dsos *dsos) +{ + RC_CHK_ACCESS(dso)->dsos = dsos; +} + +static inline bool dso__header_build_id(struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->header_build_id; +} + +static inline void dso__set_header_build_id(struct dso *dso, bool val) +{ + RC_CHK_ACCESS(dso)->header_build_id = val; +} + +static inline bool dso__hit(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->hit; +} + +static inline void dso__set_hit(struct dso *dso) +{ + RC_CHK_ACCESS(dso)->hit = 1; +} + +static inline struct dso_id *dso__id(struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->id; +} + +static inline const struct dso_id *dso__id_const(const struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->id; +} + +static inline struct rb_root_cached *dso__inlined_nodes(struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->inlined_nodes; +} + +static inline bool dso__is_64_bit(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->is_64_bit; +} + +static inline void dso__set_is_64_bit(struct dso *dso, bool is) +{ + RC_CHK_ACCESS(dso)->is_64_bit = is; +} + +static inline bool dso__is_kmod(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->is_kmod; +} + +static inline void dso__set_is_kmod(struct dso *dso) +{ + RC_CHK_ACCESS(dso)->is_kmod = 1; +} + +static inline enum dso_space_type dso__kernel(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->kernel; +} + +static inline void dso__set_kernel(struct dso *dso, enum dso_space_type kernel) +{ + RC_CHK_ACCESS(dso)->kernel = kernel; +} + +static inline u64 dso__last_find_result_addr(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->last_find_result.addr; +} + +static inline void dso__set_last_find_result_addr(struct dso *dso, u64 addr) +{ + RC_CHK_ACCESS(dso)->last_find_result.addr = addr; +} + +static inline struct symbol *dso__last_find_result_symbol(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->last_find_result.symbol; +} + +static inline void dso__set_last_find_result_symbol(struct dso *dso, struct symbol *symbol) +{ + RC_CHK_ACCESS(dso)->last_find_result.symbol = symbol; +} + +static inline enum dso_load_errno *dso__load_errno(struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->load_errno; +} static inline void dso__set_loaded(struct dso *dso) { - dso->loaded = true; + RC_CHK_ACCESS(dso)->loaded = true; +} + +static inline struct mutex *dso__lock(struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->lock; +} + +static inline const char *dso__long_name(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->long_name; +} + +static inline bool dso__long_name_allocated(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->long_name_allocated; +} + +static inline void dso__set_long_name_allocated(struct dso *dso, bool allocated) +{ + RC_CHK_ACCESS(dso)->long_name_allocated = allocated; +} + +static inline u16 dso__long_name_len(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->long_name_len; +} + +static inline const char *dso__name(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->name; +} + +static inline enum dso_swap_type dso__needs_swap(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->needs_swap; +} + +static inline void dso__set_needs_swap(struct dso *dso, enum dso_swap_type type) +{ + RC_CHK_ACCESS(dso)->needs_swap = type; +} + +static inline struct nsinfo *dso__nsinfo(struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->nsinfo; +} + +static inline const struct nsinfo *dso__nsinfo_const(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->nsinfo; +} + +static inline struct nsinfo **dso__nsinfo_ptr(struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->nsinfo; +} + +void dso__set_nsinfo(struct dso *dso, struct nsinfo *nsi); + +static inline u8 dso__rel(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->rel; +} + +static inline void dso__set_rel(struct dso *dso, u8 rel) +{ + RC_CHK_ACCESS(dso)->rel = rel; +} + +static inline const char *dso__short_name(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->short_name; +} + +static inline bool dso__short_name_allocated(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->short_name_allocated; +} + +static inline void dso__set_short_name_allocated(struct dso *dso, bool allocated) +{ + RC_CHK_ACCESS(dso)->short_name_allocated = allocated; +} + +static inline u16 dso__short_name_len(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->short_name_len; +} + +static inline struct rb_root_cached *dso__srclines(struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->srclines; +} + +static inline struct rb_root *dso__data_types(struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->data_types; +} + +static inline struct rb_root *dso__global_vars(struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->global_vars; +} + +static inline struct rb_root_cached *dso__symbols(struct dso *dso) +{ + return &RC_CHK_ACCESS(dso)->symbols; +} + +static inline struct symbol **dso__symbol_names(struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->symbol_names; +} + +static inline void dso__set_symbol_names(struct dso *dso, struct symbol **names) +{ + RC_CHK_ACCESS(dso)->symbol_names = names; +} + +static inline size_t dso__symbol_names_len(struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->symbol_names_len; +} + +static inline void dso__set_symbol_names_len(struct dso *dso, size_t len) +{ + RC_CHK_ACCESS(dso)->symbol_names_len = len; +} + +static inline const char *dso__symsrc_filename(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->symsrc_filename; +} + +static inline void dso__set_symsrc_filename(struct dso *dso, char *val) +{ + RC_CHK_ACCESS(dso)->symsrc_filename = val; +} + +static inline enum dso_binary_type dso__symtab_type(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->symtab_type; +} + +static inline void dso__set_symtab_type(struct dso *dso, enum dso_binary_type bt) +{ + RC_CHK_ACCESS(dso)->symtab_type = bt; +} + +static inline u64 dso__text_end(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->text_end; +} + +static inline void dso__set_text_end(struct dso *dso, u64 val) +{ + RC_CHK_ACCESS(dso)->text_end = val; +} + +static inline u64 dso__text_offset(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->text_offset; +} + +static inline void dso__set_text_offset(struct dso *dso, u64 val) +{ + RC_CHK_ACCESS(dso)->text_offset = val; } int dso_id__cmp(const struct dso_id *a, const struct dso_id *b); @@ -265,7 +649,7 @@ bool dso__loaded(const struct dso *dso); static inline bool dso__has_symbols(const struct dso *dso) { - return !RB_EMPTY_ROOT(&dso->symbols.rb_root); + return !RB_EMPTY_ROOT(&RC_CHK_ACCESS(dso)->symbols.rb_root); } char *dso__filename_with_chroot(const struct dso *dso, const char *filename); @@ -381,21 +765,33 @@ void dso__reset_find_symbol_cache(struct dso *dso); size_t dso__fprintf_symbols_by_name(struct dso *dso, FILE *fp); size_t dso__fprintf(struct dso *dso, FILE *fp); +static inline enum dso_binary_type dso__binary_type(const struct dso *dso) +{ + return RC_CHK_ACCESS(dso)->binary_type; +} + +static inline void dso__set_binary_type(struct dso *dso, enum dso_binary_type bt) +{ + RC_CHK_ACCESS(dso)->binary_type = bt; +} + static inline bool dso__is_vmlinux(const struct dso *dso) { - return dso->binary_type == DSO_BINARY_TYPE__VMLINUX || - dso->binary_type == DSO_BINARY_TYPE__GUEST_VMLINUX; + enum dso_binary_type bt = dso__binary_type(dso); + + return bt == DSO_BINARY_TYPE__VMLINUX || bt == DSO_BINARY_TYPE__GUEST_VMLINUX; } static inline bool dso__is_kcore(const struct dso *dso) { - return dso->binary_type == DSO_BINARY_TYPE__KCORE || - dso->binary_type == DSO_BINARY_TYPE__GUEST_KCORE; + enum dso_binary_type bt = dso__binary_type(dso); + + return bt == DSO_BINARY_TYPE__KCORE || bt == DSO_BINARY_TYPE__GUEST_KCORE; } static inline bool dso__is_kallsyms(const struct dso *dso) { - return dso->kernel && dso->long_name[0] != '/'; + return RC_CHK_ACCESS(dso)->kernel && RC_CHK_ACCESS(dso)->long_name[0] != '/'; } bool dso__is_object_file(const struct dso *dso); diff --git a/tools/perf/util/dsos.c b/tools/perf/util/dsos.c index 2e4e86dc6c17..ab3d0c01dd63 100644 --- a/tools/perf/util/dsos.c +++ b/tools/perf/util/dsos.c @@ -29,7 +29,7 @@ static void dsos__purge(struct dsos *dsos) for (unsigned int i = 0; i < dsos->cnt; i++) { struct dso *dso = dsos->dsos[i]; - dso->dsos = NULL; + dso__set_dsos(dso, NULL); dso__put(dso); } @@ -73,22 +73,22 @@ static int dsos__read_build_ids_cb(struct dso *dso, void *data) struct dsos__read_build_ids_cb_args *args = data; struct nscookie nsc; - if (args->with_hits && !dso->hit && !dso__is_vdso(dso)) + if (args->with_hits && !dso__hit(dso) && !dso__is_vdso(dso)) return 0; - if (dso->has_build_id) { + if (dso__has_build_id(dso)) { args->have_build_id = true; return 0; } - nsinfo__mountns_enter(dso->nsinfo, &nsc); - if (filename__read_build_id(dso->long_name, &dso->bid) > 0) { + nsinfo__mountns_enter(dso__nsinfo(dso), &nsc); + if (filename__read_build_id(dso__long_name(dso), dso__bid(dso)) > 0) { args->have_build_id = true; - dso->has_build_id = true; - } else if (errno == ENOENT && dso->nsinfo) { - char *new_name = dso__filename_with_chroot(dso, dso->long_name); + dso__set_has_build_id(dso); + } else if (errno == ENOENT && dso__nsinfo(dso)) { + char *new_name = dso__filename_with_chroot(dso, dso__long_name(dso)); - if (new_name && filename__read_build_id(new_name, &dso->bid) > 0) { + if (new_name && filename__read_build_id(new_name, dso__bid(dso)) > 0) { args->have_build_id = true; - dso->has_build_id = true; + dso__set_has_build_id(dso); } free(new_name); } @@ -110,27 +110,27 @@ bool dsos__read_build_ids(struct dsos *dsos, bool with_hits) static int __dso__cmp_long_name(const char *long_name, const struct dso_id *id, const struct dso *b) { - int rc = strcmp(long_name, b->long_name); - return rc ?: dso_id__cmp(id, &b->id); + int rc = strcmp(long_name, dso__long_name(b)); + return rc ?: dso_id__cmp(id, dso__id_const(b)); } static int __dso__cmp_short_name(const char *short_name, const struct dso_id *id, const struct dso *b) { - int rc = strcmp(short_name, b->short_name); - return rc ?: dso_id__cmp(id, &b->id); + int rc = strcmp(short_name, dso__short_name(b)); + return rc ?: dso_id__cmp(id, dso__id_const(b)); } static int dsos__cmp_long_name_id_short_name(const void *va, const void *vb) { const struct dso *a = *((const struct dso **)va); const struct dso *b = *((const struct dso **)vb); - int rc = strcmp(a->long_name, b->long_name); + int rc = strcmp(dso__long_name(a), dso__long_name(b)); if (!rc) { - rc = dso_id__cmp(&a->id, &b->id); + rc = dso_id__cmp(dso__id_const(a), dso__id_const(b)); if (!rc) - rc = strcmp(a->short_name, b->short_name); + rc = strcmp(dso__short_name(a), dso__short_name(b)); } return rc; } @@ -209,7 +209,7 @@ int __dsos__add(struct dsos *dsos, struct dso *dso) &dsos->dsos[dsos->cnt - 1]) <= 0; } - dso->dsos = dsos; + dso__set_dsos(dso, dsos); return 0; } @@ -275,7 +275,7 @@ static void dso__set_basename(struct dso *dso) char *base, *lname; int tid; - if (sscanf(dso->long_name, "/tmp/perf-%d.map", &tid) == 1) { + if (sscanf(dso__long_name(dso), "/tmp/perf-%d.map", &tid) == 1) { if (asprintf(&base, "[JIT] tid %d", tid) < 0) return; } else { @@ -283,7 +283,7 @@ static void dso__set_basename(struct dso *dso) * basename() may modify path buffer, so we must pass * a copy. */ - lname = strdup(dso->long_name); + lname = strdup(dso__long_name(dso)); if (!lname) return; @@ -322,7 +322,7 @@ static struct dso *__dsos__findnew_id(struct dsos *dsos, const char *name, struc { struct dso *dso = __dsos__find_id(dsos, name, id, false, /*write_locked=*/true); - if (dso && dso_id__empty(&dso->id) && !dso_id__empty(id)) + if (dso && dso_id__empty(dso__id(dso)) && !dso_id__empty(id)) __dso__inject_id(dso, id); return dso ? dso : __dsos__addnew_id(dsos, name, id); @@ -351,8 +351,8 @@ static int dsos__fprintf_buildid_cb(struct dso *dso, void *data) if (args->skip && args->skip(dso, args->parm)) return 0; - build_id__sprintf(&dso->bid, sbuild_id); - args->ret += fprintf(args->fp, "%-40s %s\n", sbuild_id, dso->long_name); + build_id__sprintf(dso__bid(dso), sbuild_id); + args->ret += fprintf(args->fp, "%-40s %s\n", sbuild_id, dso__long_name(dso)); return 0; } @@ -396,7 +396,7 @@ size_t dsos__fprintf(struct dsos *dsos, FILE *fp) static int dsos__hit_all_cb(struct dso *dso, void *data __maybe_unused) { - dso->hit = true; + dso__set_hit(dso); return 0; } @@ -432,7 +432,7 @@ struct dso *dsos__findnew_module_dso(struct dsos *dsos, dso__set_basename(dso); dso__set_module_info(dso, m, machine); dso__set_long_name(dso, strdup(filename), true); - dso->kernel = DSO_SPACE__KERNEL; + dso__set_kernel(dso, DSO_SPACE__KERNEL); __dsos__add(dsos, dso); up_write(&dsos->lock); @@ -455,8 +455,8 @@ static int dsos__find_kernel_dso_cb(struct dso *dso, void *data) * Therefore, we pass PERF_RECORD_MISC_CPUMODE_UNKNOWN. * is_kernel_module() treats it as a kernel cpumode. */ - if (!dso->kernel || - is_kernel_module(dso->long_name, PERF_RECORD_MISC_CPUMODE_UNKNOWN)) + if (!dso__kernel(dso) || + is_kernel_module(dso__long_name(dso), PERF_RECORD_MISC_CPUMODE_UNKNOWN)) return 0; *res = dso__get(dso); diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 198903157f9e..f32f9abf6344 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -726,7 +726,7 @@ int machine__resolve(struct machine *machine, struct addr_location *al, dso = al->map ? map__dso(al->map) : NULL; dump_printf(" ...... dso: %s\n", dso - ? dso->long_name + ? dso__long_name(dso) : (al->level == 'H' ? "[hypervisor]" : "")); if (thread__is_filtered(thread)) @@ -750,10 +750,10 @@ int machine__resolve(struct machine *machine, struct addr_location *al, if (al->map) { if (symbol_conf.dso_list && (!dso || !(strlist__has_entry(symbol_conf.dso_list, - dso->short_name) || - (dso->short_name != dso->long_name && + dso__short_name(dso)) || + (dso__short_name(dso) != dso__long_name(dso) && strlist__has_entry(symbol_conf.dso_list, - dso->long_name))))) { + dso__long_name(dso)))))) { al->filtered |= (1 << HIST_FILTER__DSO); } diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 3fe28edc3d01..55e9553861d0 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -2308,7 +2308,7 @@ static int __event_process_build_id(struct perf_record_header_build_id *bev, build_id__init(&bid, bev->data, size); dso__set_build_id(dso, &bid); - dso->header_build_id = 1; + dso__set_header_build_id(dso, true); if (dso_space != DSO_SPACE__USER) { struct kmod_path m = { .name = NULL, }; @@ -2316,13 +2316,13 @@ static int __event_process_build_id(struct perf_record_header_build_id *bev, if (!kmod_path__parse_name(&m, filename) && m.kmod) dso__set_module_info(dso, &m, machine); - dso->kernel = dso_space; + dso__set_kernel(dso, dso_space); free(m.name); } - build_id__sprintf(&dso->bid, sbuild_id); + build_id__sprintf(dso__bid(dso), sbuild_id); pr_debug("build id event received for %s: %s [%zu]\n", - dso->long_name, sbuild_id, size); + dso__long_name(dso), sbuild_id, size); dso__put(dso); } diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c index 9d43f8ae412d..55ea6afcc437 100644 --- a/tools/perf/util/hist.c +++ b/tools/perf/util/hist.c @@ -2136,7 +2136,7 @@ static bool hists__filter_entry_by_dso(struct hists *hists, struct hist_entry *he) { if (hists->dso_filter != NULL && - (he->ms.map == NULL || map__dso(he->ms.map) != hists->dso_filter)) { + (he->ms.map == NULL || !RC_CHK_EQUAL(map__dso(he->ms.map), hists->dso_filter))) { he->filtered |= (1 << HIST_FILTER__DSO); return true; } @@ -2816,7 +2816,7 @@ int __hists__scnprintf_title(struct hists *hists, char *bf, size_t size, bool sh } if (dso) printed += scnprintf(bf + printed, size - printed, - ", DSO: %s", dso->short_name); + ", DSO: %s", dso__short_name(dso)); if (socket_id > -1) printed += scnprintf(bf + printed, size - printed, ", Processor Socket: %d", socket_id); diff --git a/tools/perf/util/intel-pt.c b/tools/perf/util/intel-pt.c index 4db9a098f592..d6d7b7512505 100644 --- a/tools/perf/util/intel-pt.c +++ b/tools/perf/util/intel-pt.c @@ -598,15 +598,15 @@ static struct auxtrace_cache *intel_pt_cache(struct dso *dso, struct auxtrace_cache *c; unsigned int bits; - if (dso->auxtrace_cache) - return dso->auxtrace_cache; + if (dso__auxtrace_cache(dso)) + return dso__auxtrace_cache(dso); bits = intel_pt_cache_size(dso, machine); /* Ignoring cache creation failure */ c = auxtrace_cache__new(bits, sizeof(struct intel_pt_cache_entry), 200); - dso->auxtrace_cache = c; + dso__set_auxtrace_cache(dso, c); return c; } @@ -650,7 +650,7 @@ intel_pt_cache_lookup(struct dso *dso, struct machine *machine, u64 offset) if (!c) return NULL; - return auxtrace_cache__lookup(dso->auxtrace_cache, offset); + return auxtrace_cache__lookup(dso__auxtrace_cache(dso), offset); } static void intel_pt_cache_invalidate(struct dso *dso, struct machine *machine, @@ -661,7 +661,7 @@ static void intel_pt_cache_invalidate(struct dso *dso, struct machine *machine, if (!c) return; - auxtrace_cache__remove(dso->auxtrace_cache, offset); + auxtrace_cache__remove(dso__auxtrace_cache(dso), offset); } static inline bool intel_pt_guest_kernel_ip(uint64_t ip) @@ -821,8 +821,8 @@ static int intel_pt_walk_next_insn(struct intel_pt_insn *intel_pt_insn, } dso = map__dso(al.map); - if (dso->data.status == DSO_DATA_STATUS_ERROR && - dso__data_status_seen(dso, DSO_DATA_STATUS_SEEN_ITRACE)) { + if (dso__data(dso)->status == DSO_DATA_STATUS_ERROR && + dso__data_status_seen(dso, DSO_DATA_STATUS_SEEN_ITRACE)) { ret = -ENOENT; goto out_ret; } @@ -855,7 +855,7 @@ static int intel_pt_walk_next_insn(struct intel_pt_insn *intel_pt_insn, /* Load maps to ensure dso->is_64_bit has been updated */ map__load(al.map); - x86_64 = dso->is_64_bit; + x86_64 = dso__is_64_bit(dso); while (1) { len = dso__data_read_offset(dso, machine, @@ -1010,7 +1010,7 @@ static int __intel_pt_pgd_ip(uint64_t ip, void *data) offset = map__map_ip(al.map, ip); - res = intel_pt_match_pgd_ip(ptq->pt, ip, offset, map__dso(al.map)->long_name); + res = intel_pt_match_pgd_ip(ptq->pt, ip, offset, dso__long_name(map__dso(al.map))); addr_location__exit(&al); return res; } @@ -3418,7 +3418,7 @@ static int intel_pt_text_poke(struct intel_pt *pt, union perf_event *event) } dso = map__dso(al.map); - if (!dso || !dso->auxtrace_cache) + if (!dso || !dso__auxtrace_cache(dso)) continue; offset = map__map_ip(al.map, addr); @@ -3438,7 +3438,7 @@ static int intel_pt_text_poke(struct intel_pt *pt, union perf_event *event) } else { intel_pt_cache_invalidate(dso, machine, offset); intel_pt_log("Invalidated instruction cache for %s at %#"PRIx64"\n", - dso->long_name, addr); + dso__long_name(dso), addr); } } out: diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index c5c895131bca..0b8fb14f5ff6 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -694,7 +694,7 @@ static int machine__process_ksymbol_register(struct machine *machine, err = -ENOMEM; goto out; } - dso->kernel = DSO_SPACE__KERNEL; + dso__set_kernel(dso, DSO_SPACE__KERNEL); map = map__new2(0, dso); dso__put(dso); if (!map) { @@ -702,8 +702,8 @@ static int machine__process_ksymbol_register(struct machine *machine, goto out; } if (event->ksymbol.ksym_type == PERF_RECORD_KSYMBOL_TYPE_OOL) { - dso->binary_type = DSO_BINARY_TYPE__OOL; - dso->data.file_size = event->ksymbol.len; + dso__set_binary_type(dso, DSO_BINARY_TYPE__OOL); + dso__data(dso)->file_size = event->ksymbol.len; dso__set_loaded(dso); } @@ -718,7 +718,7 @@ static int machine__process_ksymbol_register(struct machine *machine, dso__set_loaded(dso); if (is_bpf_image(event->ksymbol.name)) { - dso->binary_type = DSO_BINARY_TYPE__BPF_IMAGE; + dso__set_binary_type(dso, DSO_BINARY_TYPE__BPF_IMAGE); dso__set_long_name(dso, "", false); } } else { @@ -888,17 +888,17 @@ size_t machine__fprintf_vmlinux_path(struct machine *machine, FILE *fp) size_t printed = 0; struct dso *kdso = machine__kernel_dso(machine); - if (kdso->has_build_id) { + if (dso__has_build_id(kdso)) { char filename[PATH_MAX]; - if (dso__build_id_filename(kdso, filename, sizeof(filename), - false)) + + if (dso__build_id_filename(kdso, filename, sizeof(filename), false)) printed += fprintf(fp, "[0] %s\n", filename); } - for (i = 0; i < vmlinux_path__nr_entries; ++i) - printed += fprintf(fp, "[%d] %s\n", - i + kdso->has_build_id, vmlinux_path[i]); - + for (i = 0; i < vmlinux_path__nr_entries; ++i) { + printed += fprintf(fp, "[%d] %s\n", i + dso__has_build_id(kdso), + vmlinux_path[i]); + } return printed; } @@ -948,7 +948,7 @@ static struct dso *machine__get_kernel(struct machine *machine) DSO_SPACE__KERNEL_GUEST); } - if (kernel != NULL && (!kernel->has_build_id)) + if (kernel != NULL && (!dso__has_build_id(kernel))) dso__read_running_kernel_build_id(kernel, machine); return kernel; @@ -1313,8 +1313,8 @@ static char *get_kernel_version(const char *root_dir) static bool is_kmod_dso(struct dso *dso) { - return dso->symtab_type == DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE || - dso->symtab_type == DSO_BINARY_TYPE__GUEST_KMODULE; + return dso__symtab_type(dso) == DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE || + dso__symtab_type(dso) == DSO_BINARY_TYPE__GUEST_KMODULE; } static int maps__set_module_path(struct maps *maps, const char *path, struct kmod_path *m) @@ -1341,8 +1341,8 @@ static int maps__set_module_path(struct maps *maps, const char *path, struct kmo * we need to update the symtab_type if needed. */ if (m->comp && is_kmod_dso(dso)) { - dso->symtab_type++; - dso->comp = m->comp; + dso__set_symtab_type(dso, dso__symtab_type(dso)); + dso__set_comp(dso, m->comp); } map__put(map); return 0; @@ -1643,13 +1643,13 @@ static int machine__process_kernel_mmap_event(struct machine *machine, if (kernel == NULL) goto out_problem; - kernel->kernel = dso_space; + dso__set_kernel(kernel, dso_space); if (__machine__create_kernel_maps(machine, kernel) < 0) { dso__put(kernel); goto out_problem; } - if (strstr(kernel->long_name, "vmlinux")) + if (strstr(dso__long_name(kernel), "vmlinux")) dso__set_short_name(kernel, "[kernel.vmlinux]", false); if (machine__update_kernel_mmap(machine, xm->start, xm->end) < 0) { @@ -2031,14 +2031,14 @@ static char *callchain_srcline(struct map_symbol *ms, u64 ip) return srcline; dso = map__dso(map); - srcline = srcline__tree_find(&dso->srclines, ip); + srcline = srcline__tree_find(dso__srclines(dso), ip); if (!srcline) { bool show_sym = false; bool show_addr = callchain_param.key == CCKEY_ADDRESS; srcline = get_srcline(dso, map__rip_2objdump(map, ip), ms->sym, show_sym, show_addr, ip); - srcline__tree_insert(&dso->srclines, ip, srcline); + srcline__tree_insert(dso__srclines(dso), ip, srcline); } return srcline; @@ -2836,12 +2836,12 @@ static int append_inlines(struct callchain_cursor *cursor, struct map_symbol *ms addr = map__rip_2objdump(map, addr); dso = map__dso(map); - inline_node = inlines__tree_find(&dso->inlined_nodes, addr); + inline_node = inlines__tree_find(dso__inlined_nodes(dso), addr); if (!inline_node) { inline_node = dso__parse_addr_inlines(dso, addr, sym); if (!inline_node) return ret; - inlines__tree_insert(&dso->inlined_nodes, inline_node); + inlines__tree_insert(dso__inlined_nodes(dso), inline_node); } ilist_ms = (struct map_symbol) { @@ -3130,7 +3130,7 @@ char *machine__resolve_kernel_addr(void *vmachine, unsigned long long *addrp, ch if (sym == NULL) return NULL; - *modp = __map__is_kmodule(map) ? (char *)map__dso(map)->short_name : NULL; + *modp = __map__is_kmodule(map) ? (char *)dso__short_name(map__dso(map)) : NULL; *addrp = map__unmap_ip(map, sym->start); return sym->name; } diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index cca871959f87..117c4bb78b35 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -168,7 +168,7 @@ struct map *map__new(struct machine *machine, u64 start, u64 len, if (dso == NULL) goto out_delete; - assert(!dso->kernel); + assert(!dso__kernel(dso)); map__init(result, start, start + len, pgoff, dso); if (anon || no_dso) { @@ -182,10 +182,9 @@ struct map *map__new(struct machine *machine, u64 start, u64 len, if (!(prot & PROT_EXEC)) dso__set_loaded(dso); } - mutex_lock(&dso->lock); - nsinfo__put(dso->nsinfo); - dso->nsinfo = nsi; - mutex_unlock(&dso->lock); + mutex_lock(dso__lock(dso)); + dso__set_nsinfo(dso, nsi); + mutex_unlock(dso__lock(dso)); if (build_id__is_defined(bid)) { dso__set_build_id(dso, bid); @@ -197,9 +196,9 @@ struct map *map__new(struct machine *machine, u64 start, u64 len, * have it missing. */ header_bid_dso = dsos__find(&machine->dsos, filename, false); - if (header_bid_dso && header_bid_dso->header_build_id) { - dso__set_build_id(dso, &header_bid_dso->bid); - dso->header_build_id = 1; + if (header_bid_dso && dso__header_build_id(header_bid_dso)) { + dso__set_build_id(dso, dso__bid(header_bid_dso)); + dso__set_header_build_id(dso, 1); } } dso__put(dso); @@ -221,7 +220,7 @@ struct map *map__new2(u64 start, struct dso *dso) struct map *result; RC_STRUCT(map) *map; - map = calloc(1, sizeof(*map) + (dso->kernel ? sizeof(struct kmap) : 0)); + map = calloc(1, sizeof(*map) + (dso__kernel(dso) ? sizeof(struct kmap) : 0)); if (ADD_RC_CHK(result, map)) { /* * ->end will be filled after we load all the symbols @@ -234,7 +233,7 @@ struct map *map__new2(u64 start, struct dso *dso) bool __map__is_kernel(const struct map *map) { - if (!map__dso(map)->kernel) + if (!dso__kernel(map__dso(map))) return false; return machine__kernel_map(maps__machine(map__kmaps((struct map *)map))) == map; } @@ -251,7 +250,7 @@ bool __map__is_bpf_prog(const struct map *map) const char *name; struct dso *dso = map__dso(map); - if (dso->binary_type == DSO_BINARY_TYPE__BPF_PROG_INFO) + if (dso__binary_type(dso) == DSO_BINARY_TYPE__BPF_PROG_INFO) return true; /* @@ -259,7 +258,7 @@ bool __map__is_bpf_prog(const struct map *map) * type of DSO_BINARY_TYPE__BPF_PROG_INFO. In such cases, we can * guess the type based on name. */ - name = dso->short_name; + name = dso__short_name(dso); return name && (strstr(name, "bpf_prog_") == name); } @@ -268,7 +267,7 @@ bool __map__is_bpf_image(const struct map *map) const char *name; struct dso *dso = map__dso(map); - if (dso->binary_type == DSO_BINARY_TYPE__BPF_IMAGE) + if (dso__binary_type(dso) == DSO_BINARY_TYPE__BPF_IMAGE) return true; /* @@ -276,7 +275,7 @@ bool __map__is_bpf_image(const struct map *map) * type of DSO_BINARY_TYPE__BPF_IMAGE. In such cases, we can * guess the type based on name. */ - name = dso->short_name; + name = dso__short_name(dso); return name && is_bpf_image(name); } @@ -284,7 +283,7 @@ bool __map__is_ool(const struct map *map) { const struct dso *dso = map__dso(map); - return dso && dso->binary_type == DSO_BINARY_TYPE__OOL; + return dso && dso__binary_type(dso) == DSO_BINARY_TYPE__OOL; } bool map__has_symbols(const struct map *map) @@ -315,7 +314,7 @@ void map__put(struct map *map) void map__fixup_start(struct map *map) { struct dso *dso = map__dso(map); - struct rb_root_cached *symbols = &dso->symbols; + struct rb_root_cached *symbols = dso__symbols(dso); struct rb_node *nd = rb_first_cached(symbols); if (nd != NULL) { @@ -328,7 +327,7 @@ void map__fixup_start(struct map *map) void map__fixup_end(struct map *map) { struct dso *dso = map__dso(map); - struct rb_root_cached *symbols = &dso->symbols; + struct rb_root_cached *symbols = dso__symbols(dso); struct rb_node *nd = rb_last(&symbols->rb_root); if (nd != NULL) { @@ -342,7 +341,7 @@ void map__fixup_end(struct map *map) int map__load(struct map *map) { struct dso *dso = map__dso(map); - const char *name = dso->long_name; + const char *name = dso__long_name(dso); int nr; if (dso__loaded(dso)) @@ -350,10 +349,10 @@ int map__load(struct map *map) nr = dso__load(dso, map); if (nr < 0) { - if (dso->has_build_id) { + if (dso__has_build_id(dso)) { char sbuild_id[SBUILD_ID_SIZE]; - build_id__sprintf(&dso->bid, sbuild_id); + build_id__sprintf(dso__bid(dso), sbuild_id); pr_debug("%s with build id %s not found", name, sbuild_id); } else pr_debug("Failed to open %s", name); @@ -415,7 +414,7 @@ struct map *map__clone(struct map *from) size_t size = sizeof(RC_STRUCT(map)); struct dso *dso = map__dso(from); - if (dso && dso->kernel) + if (dso && dso__kernel(dso)) size += sizeof(struct kmap); map = memdup(RC_CHK_ACCESS(from), size); @@ -432,14 +431,14 @@ size_t map__fprintf(struct map *map, FILE *fp) const struct dso *dso = map__dso(map); return fprintf(fp, " %" PRIx64 "-%" PRIx64 " %" PRIx64 " %s\n", - map__start(map), map__end(map), map__pgoff(map), dso->name); + map__start(map), map__end(map), map__pgoff(map), dso__name(dso)); } static bool prefer_dso_long_name(const struct dso *dso, bool print_off) { - return dso->long_name && + return dso__long_name(dso) && (symbol_conf.show_kernel_path || - (print_off && (dso->name[0] == '[' || dso__is_kcore(dso)))); + (print_off && (dso__name(dso)[0] == '[' || dso__is_kcore(dso)))); } static size_t __map__fprintf_dsoname(struct map *map, bool print_off, FILE *fp) @@ -450,9 +449,9 @@ static size_t __map__fprintf_dsoname(struct map *map, bool print_off, FILE *fp) if (dso) { if (prefer_dso_long_name(dso, print_off)) - dsoname = dso->long_name; + dsoname = dso__long_name(dso); else - dsoname = dso->name; + dsoname = dso__name(dso); } if (symbol_conf.pad_output_len_dso) { @@ -545,14 +544,14 @@ u64 map__rip_2objdump(struct map *map, u64 rip) } } - if (!dso->adjust_symbols) + if (!dso__adjust_symbols(dso)) return rip; - if (dso->rel) + if (dso__rel(dso)) return rip - map__pgoff(map); - if (dso->kernel == DSO_SPACE__USER) - return rip + dso->text_offset; + if (dso__kernel(dso) == DSO_SPACE__USER) + return rip + dso__text_offset(dso); return map__unmap_ip(map, rip) - map__reloc(map); } @@ -573,14 +572,14 @@ u64 map__objdump_2mem(struct map *map, u64 ip) { const struct dso *dso = map__dso(map); - if (!dso->adjust_symbols) + if (!dso__adjust_symbols(dso)) return map__unmap_ip(map, ip); - if (dso->rel) + if (dso__rel(dso)) return map__unmap_ip(map, ip + map__pgoff(map)); - if (dso->kernel == DSO_SPACE__USER) - return map__unmap_ip(map, ip - dso->text_offset); + if (dso__kernel(dso) == DSO_SPACE__USER) + return map__unmap_ip(map, ip - dso__text_offset(dso)); return ip + map__reloc(map); } @@ -590,14 +589,14 @@ u64 map__objdump_2rip(struct map *map, u64 ip) { const struct dso *dso = map__dso(map); - if (!dso->adjust_symbols) + if (!dso__adjust_symbols(dso)) return ip; - if (dso->rel) + if (dso__rel(dso)) return ip + map__pgoff(map); - if (dso->kernel == DSO_SPACE__USER) - return ip - dso->text_offset; + if (dso__kernel(dso) == DSO_SPACE__USER) + return ip - dso__text_offset(dso); return map__map_ip(map, ip + map__reloc(map)); } @@ -613,7 +612,7 @@ struct kmap *__map__kmap(struct map *map) { const struct dso *dso = map__dso(map); - if (!dso || !dso->kernel) + if (!dso || !dso__kernel(dso)) return NULL; return (struct kmap *)(&RC_CHK_ACCESS(map)[1]); } diff --git a/tools/perf/util/maps.c b/tools/perf/util/maps.c index ac9fb880ddc7..61eb742d91e3 100644 --- a/tools/perf/util/maps.c +++ b/tools/perf/util/maps.c @@ -76,7 +76,7 @@ static void check_invariants(const struct maps *maps __maybe_unused) /* Expect at least 1 reference count. */ assert(refcount_read(map__refcnt(map)) > 0); - if (map__dso(map) && map__dso(map)->kernel) + if (map__dso(map) && dso__kernel(map__dso(map))) assert(RC_CHK_EQUAL(map__kmap(map)->kmaps, maps)); if (i > 0) { @@ -341,7 +341,7 @@ static int map__strcmp(const void *a, const void *b) const struct map *map_b = *(const struct map * const *)b; const struct dso *dso_a = map__dso(map_a); const struct dso *dso_b = map__dso(map_b); - int ret = strcmp(dso_a->short_name, dso_b->short_name); + int ret = strcmp(dso__short_name(dso_a), dso__short_name(dso_b)); if (ret == 0 && RC_CHK_ACCESS(map_a) != RC_CHK_ACCESS(map_b)) { /* Ensure distinct but name equal maps have an order. */ @@ -482,7 +482,7 @@ static int __maps__insert(struct maps *maps, struct map *new) } if (map__end(new) < map__start(new)) RC_CHK_ACCESS(maps)->ends_broken = true; - if (dso && dso->kernel) { + if (dso && dso__kernel(dso)) { struct kmap *kmap = map__kmap(new); if (kmap) @@ -766,7 +766,7 @@ sort_again: if (use_browser) { pr_debug("overlapping maps in %s (disable tui for more info)\n", - map__dso(new)->name); + dso__name(map__dso(new))); } else if (verbose >= 2) { pr_debug("overlapping maps:\n"); map__fprintf(new, fp); @@ -989,7 +989,7 @@ static int map__strcmp_name(const void *name, const void *b) { const struct dso *dso = map__dso(*(const struct map **)b); - return strcmp(name, dso->short_name); + return strcmp(name, dso__short_name(dso)); } struct map *maps__find_by_name(struct maps *maps, const char *name) @@ -1008,7 +1008,7 @@ struct map *maps__find_by_name(struct maps *maps, const char *name) if (i < maps__nr_maps(maps) && maps__maps_by_name(maps)) { struct dso *dso = map__dso(maps__maps_by_name(maps)[i]); - if (dso && strcmp(dso->short_name, name) == 0) { + if (dso && strcmp(dso__short_name(dso), name) == 0) { result = map__get(maps__maps_by_name(maps)[i]); done = true; } @@ -1045,7 +1045,7 @@ struct map *maps__find_by_name(struct maps *maps, const char *name) struct map *pos = maps_by_address[i]; struct dso *dso = map__dso(pos); - if (dso && strcmp(dso->short_name, name) == 0) { + if (dso && strcmp(dso__short_name(dso), name) == 0) { result = map__get(pos); break; } diff --git a/tools/perf/util/print_insn.c b/tools/perf/util/print_insn.c index aab11d8e1b1d..a950e9157d2d 100644 --- a/tools/perf/util/print_insn.c +++ b/tools/perf/util/print_insn.c @@ -104,7 +104,7 @@ static bool is64bitip(struct machine *machine, struct addr_location *al) const struct dso *dso = al->map ? map__dso(al->map) : NULL; if (dso) - return dso->is_64_bit; + return dso__is_64_bit(dso); return machine__is(machine, "x86_64") || machine__normalized_is(machine, "arm64") || diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index 8a73c9464b70..a17c9b8a7a79 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -159,8 +159,8 @@ static int kernel_get_module_map_cb(struct map *map, void *data) { struct kernel_get_module_map_cb_args *args = data; struct dso *dso = map__dso(map); - const char *short_name = dso->short_name; /* short_name is "[module]" */ - u16 short_name_len = dso->short_name_len; + const char *short_name = dso__short_name(dso); + u16 short_name_len = dso__short_name_len(dso); if (strncmp(short_name + 1, args->module, short_name_len - 2) == 0 && args->module[short_name_len - 2] == '\0') { @@ -202,10 +202,9 @@ struct map *get_target_map(const char *target, struct nsinfo *nsi, bool user) map = dso__new_map(target); dso = map ? map__dso(map) : NULL; if (dso) { - mutex_lock(&dso->lock); - nsinfo__put(dso->nsinfo); - dso->nsinfo = nsinfo__get(nsi); - mutex_unlock(&dso->lock); + mutex_lock(dso__lock(dso)); + dso__set_nsinfo(dso, nsinfo__get(nsi)); + mutex_unlock(dso__lock(dso)); } return map; } else { @@ -368,11 +367,11 @@ static int kernel_get_module_dso(const char *module, struct dso **pdso) map = machine__kernel_map(host_machine); dso = map__dso(map); - if (!dso->has_build_id) + if (!dso__has_build_id(dso)) dso__read_running_kernel_build_id(dso, host_machine); vmlinux_name = symbol_conf.vmlinux_name; - dso->load_errno = 0; + *dso__load_errno(dso) = 0; if (vmlinux_name) ret = dso__load_vmlinux(dso, map, vmlinux_name, false); else @@ -499,7 +498,7 @@ static struct debuginfo *open_from_debuginfod(struct dso *dso, struct nsinfo *ns if (!c) return NULL; - build_id__sprintf(&dso->bid, sbuild_id); + build_id__sprintf(dso__bid(dso), sbuild_id); fd = debuginfod_find_debuginfo(c, (const unsigned char *)sbuild_id, 0, &path); if (fd >= 0) @@ -542,7 +541,7 @@ static struct debuginfo *open_debuginfo(const char *module, struct nsinfo *nsi, if (!module || !strchr(module, '/')) { err = kernel_get_module_dso(module, &dso); if (err < 0) { - if (!dso || dso->load_errno == 0) { + if (!dso || *dso__load_errno(dso) == 0) { if (!str_error_r(-err, reason, STRERR_BUFSIZE)) strcpy(reason, "(unknown)"); } else @@ -559,7 +558,7 @@ static struct debuginfo *open_debuginfo(const char *module, struct nsinfo *nsi, } return NULL; } - path = dso->long_name; + path = dso__long_name(dso); } nsinfo__mountns_enter(nsi, &nsc); ret = debuginfo__new(path); @@ -3795,8 +3794,8 @@ int show_available_funcs(const char *target, struct nsinfo *nsi, /* Show all (filtered) symbols */ setup_pager(); - for (size_t i = 0; i < dso->symbol_names_len; i++) { - struct symbol *pos = dso->symbol_names[i]; + for (size_t i = 0; i < dso__symbol_names_len(dso); i++) { + struct symbol *pos = dso__symbol_names(dso)[i]; if (strfilter__compare(_filter, pos->name)) printf("%s\n", pos->name); diff --git a/tools/perf/util/scripting-engines/trace-event-perl.c b/tools/perf/util/scripting-engines/trace-event-perl.c index b072ac5d3bc2..e16257d5ab2c 100644 --- a/tools/perf/util/scripting-engines/trace-event-perl.c +++ b/tools/perf/util/scripting-engines/trace-event-perl.c @@ -320,10 +320,10 @@ static SV *perl_process_callchain(struct perf_sample *sample, const char *dsoname = "[unknown]"; if (dso) { - if (symbol_conf.show_kernel_path && dso->long_name) - dsoname = dso->long_name; + if (symbol_conf.show_kernel_path && dso__long_name(dso)) + dsoname = dso__long_name(dso); else - dsoname = dso->name; + dsoname = dso__name(dso); } if (!hv_stores(elem, "dso", newSVpv(dsoname,0))) { hv_undef(elem); diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c index 8aa301948de5..c2caa5720299 100644 --- a/tools/perf/util/scripting-engines/trace-event-python.c +++ b/tools/perf/util/scripting-engines/trace-event-python.c @@ -393,10 +393,10 @@ static const char *get_dsoname(struct map *map) struct dso *dso = map ? map__dso(map) : NULL; if (dso) { - if (symbol_conf.show_kernel_path && dso->long_name) - dsoname = dso->long_name; + if (symbol_conf.show_kernel_path && dso__long_name(dso)) + dsoname = dso__long_name(dso); else - dsoname = dso->name; + dsoname = dso__name(dso); } return dsoname; @@ -799,8 +799,9 @@ static void set_sym_in_dict(PyObject *dict, struct addr_location *al, if (al->map) { struct dso *dso = map__dso(al->map); - pydict_set_item_string_decref(dict, dso_field, _PyUnicode_FromString(dso->name)); - build_id__sprintf(&dso->bid, sbuild_id); + pydict_set_item_string_decref(dict, dso_field, + _PyUnicode_FromString(dso__name(dso))); + build_id__sprintf(dso__bid(dso), sbuild_id); pydict_set_item_string_decref(dict, dso_bid_field, _PyUnicode_FromString(sbuild_id)); pydict_set_item_string_decref(dict, dso_map_start, @@ -1246,14 +1247,14 @@ static int python_export_dso(struct db_export *dbe, struct dso *dso, char sbuild_id[SBUILD_ID_SIZE]; PyObject *t; - build_id__sprintf(&dso->bid, sbuild_id); + build_id__sprintf(dso__bid(dso), sbuild_id); t = tuple_new(5); - tuple_set_d64(t, 0, dso->db_id); + tuple_set_d64(t, 0, dso__db_id(dso)); tuple_set_d64(t, 1, machine->db_id); - tuple_set_string(t, 2, dso->short_name); - tuple_set_string(t, 3, dso->long_name); + tuple_set_string(t, 2, dso__short_name(dso)); + tuple_set_string(t, 3, dso__long_name(dso)); tuple_set_string(t, 4, sbuild_id); call_object(tables->dso_handler, t, "dso_table"); @@ -1273,7 +1274,7 @@ static int python_export_symbol(struct db_export *dbe, struct symbol *sym, t = tuple_new(6); tuple_set_d64(t, 0, *sym_db_id); - tuple_set_d64(t, 1, dso->db_id); + tuple_set_d64(t, 1, dso__db_id(dso)); tuple_set_d64(t, 2, sym->start); tuple_set_d64(t, 3, sym->end); tuple_set_s32(t, 4, sym->binding); diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c index add8601c57fd..704664e5b4ea 100644 --- a/tools/perf/util/sort.c +++ b/tools/perf/util/sort.c @@ -239,11 +239,11 @@ static int64_t _sort__dso_cmp(struct map *map_l, struct map *map_r) return cmp_null(dso_r, dso_l); if (verbose > 0) { - dso_name_l = dso_l->long_name; - dso_name_r = dso_r->long_name; + dso_name_l = dso__long_name(dso_l); + dso_name_r = dso__long_name(dso_r); } else { - dso_name_l = dso_l->short_name; - dso_name_r = dso_r->short_name; + dso_name_l = dso__short_name(dso_l); + dso_name_r = dso__short_name(dso_r); } return strcmp(dso_name_l, dso_name_r); @@ -262,7 +262,7 @@ static int _hist_entry__dso_snprintf(struct map *map, char *bf, const char *dso_name = "[unknown]"; if (dso) - dso_name = verbose > 0 ? dso->long_name : dso->short_name; + dso_name = verbose > 0 ? dso__long_name(dso) : dso__short_name(dso); return repsep_snprintf(bf, size, "%-*.*s", width, width, dso_name); } @@ -364,7 +364,7 @@ static int _hist_entry__sym_snprintf(struct map_symbol *ms, char o = dso ? dso__symtab_origin(dso) : '!'; u64 rip = ip; - if (dso && dso->kernel && dso->adjust_symbols) + if (dso && dso__kernel(dso) && dso__adjust_symbols(dso)) rip = map__unmap_ip(map, ip); ret += repsep_snprintf(bf, size, "%-#*llx %c ", @@ -1586,8 +1586,8 @@ sort__dcacheline_cmp(struct hist_entry *left, struct hist_entry *right) */ if ((left->cpumode != PERF_RECORD_MISC_KERNEL) && - (!(map__flags(l_map) & MAP_SHARED)) && !l_dso->id.maj && !l_dso->id.min && - !l_dso->id.ino && !l_dso->id.ino_generation) { + (!(map__flags(l_map) & MAP_SHARED)) && !dso__id(l_dso)->maj && !dso__id(l_dso)->min && + !dso__id(l_dso)->ino && !dso__id(l_dso)->ino_generation) { /* userspace anonymous */ if (thread__pid(left->thread) > thread__pid(right->thread)) @@ -1626,7 +1626,8 @@ static int hist_entry__dcacheline_snprintf(struct hist_entry *he, char *bf, if ((he->cpumode != PERF_RECORD_MISC_KERNEL) && map && !(map__prot(map) & PROT_EXEC) && (map__flags(map) & MAP_SHARED) && - (dso->id.maj || dso->id.min || dso->id.ino || dso->id.ino_generation)) + (dso__id(dso)->maj || dso__id(dso)->min || dso__id(dso)->ino || + dso__id(dso)->ino_generation)) level = 's'; else if (!map) level = 'X'; diff --git a/tools/perf/util/srcline.c b/tools/perf/util/srcline.c index 7addc34afcf5..9d670d8c1c08 100644 --- a/tools/perf/util/srcline.c +++ b/tools/perf/util/srcline.c @@ -27,14 +27,14 @@ bool srcline_full_filename; char *srcline__unknown = (char *)"??:0"; -static const char *dso__name(struct dso *dso) +static const char *srcline_dso_name(struct dso *dso) { const char *dso_name; - if (dso->symsrc_filename) - dso_name = dso->symsrc_filename; + if (dso__symsrc_filename(dso)) + dso_name = dso__symsrc_filename(dso); else - dso_name = dso->long_name; + dso_name = dso__long_name(dso); if (dso_name[0] == '[') return NULL; @@ -638,7 +638,7 @@ static int addr2line(const char *dso_name, u64 addr, struct inline_node *node, struct symbol *sym __maybe_unused) { - struct child_process *a2l = dso->a2l; + struct child_process *a2l = dso__a2l(dso); char *record_function = NULL; char *record_filename = NULL; unsigned int record_line_nr = 0; @@ -655,8 +655,9 @@ static int addr2line(const char *dso_name, u64 addr, if (!filename__has_section(dso_name, ".debug_line")) goto out; - dso->a2l = addr2line_subprocess_init(symbol_conf.addr2line_path, dso_name); - a2l = dso->a2l; + dso__set_a2l(dso, + addr2line_subprocess_init(symbol_conf.addr2line_path, dso_name)); + a2l = dso__a2l(dso); } if (a2l == NULL) { @@ -770,7 +771,7 @@ out: free(record_function); free(record_filename); if (io.eof) { - dso->a2l = NULL; + dso__set_a2l(dso, NULL); addr2line_subprocess_cleanup(a2l); } return ret; @@ -778,14 +779,14 @@ out: void dso__free_a2l(struct dso *dso) { - struct child_process *a2l = dso->a2l; + struct child_process *a2l = dso__a2l(dso); if (!a2l) return; addr2line_subprocess_cleanup(a2l); - dso->a2l = NULL; + dso__set_a2l(dso, NULL); } #endif /* HAVE_LIBBFD_SUPPORT */ @@ -823,33 +824,34 @@ char *__get_srcline(struct dso *dso, u64 addr, struct symbol *sym, char *srcline; const char *dso_name; - if (!dso->has_srcline) + if (!dso__has_srcline(dso)) goto out; - dso_name = dso__name(dso); + dso_name = srcline_dso_name(dso); if (dso_name == NULL) - goto out; + goto out_err; if (!addr2line(dso_name, addr, &file, &line, dso, unwind_inlines, NULL, sym)) - goto out; + goto out_err; srcline = srcline_from_fileline(file, line); free(file); if (!srcline) - goto out; + goto out_err; - dso->a2l_fails = 0; + dso__set_a2l_fails(dso, 0); return srcline; -out: - if (dso->a2l_fails && ++dso->a2l_fails > A2L_FAIL_LIMIT) { - dso->has_srcline = 0; +out_err: + dso__set_a2l_fails(dso, dso__a2l_fails(dso) + 1); + if (dso__a2l_fails(dso) > A2L_FAIL_LIMIT) { + dso__set_has_srcline(dso, false); dso__free_a2l(dso); } - +out: if (!show_addr) return (show_sym && sym) ? strndup(sym->name, sym->namelen) : SRCLINE_UNKNOWN; @@ -858,7 +860,7 @@ out: if (asprintf(&srcline, "%s+%" PRIu64, show_sym ? sym->name : "", ip - sym->start) < 0) return SRCLINE_UNKNOWN; - } else if (asprintf(&srcline, "%s[%" PRIx64 "]", dso->short_name, addr) < 0) + } else if (asprintf(&srcline, "%s[%" PRIx64 "]", dso__short_name(dso), addr) < 0) return SRCLINE_UNKNOWN; return srcline; } @@ -869,22 +871,23 @@ char *get_srcline_split(struct dso *dso, u64 addr, unsigned *line) char *file = NULL; const char *dso_name; - if (!dso->has_srcline) - goto out; + if (!dso__has_srcline(dso)) + return NULL; - dso_name = dso__name(dso); + dso_name = srcline_dso_name(dso); if (dso_name == NULL) - goto out; + goto out_err; if (!addr2line(dso_name, addr, &file, line, dso, true, NULL, NULL)) - goto out; + goto out_err; - dso->a2l_fails = 0; + dso__set_a2l_fails(dso, 0); return file; -out: - if (dso->a2l_fails && ++dso->a2l_fails > A2L_FAIL_LIMIT) { - dso->has_srcline = 0; +out_err: + dso__set_a2l_fails(dso, dso__a2l_fails(dso) + 1); + if (dso__a2l_fails(dso) > A2L_FAIL_LIMIT) { + dso__set_has_srcline(dso, false); dso__free_a2l(dso); } @@ -982,7 +985,7 @@ struct inline_node *dso__parse_addr_inlines(struct dso *dso, u64 addr, { const char *dso_name; - dso_name = dso__name(dso); + dso_name = srcline_dso_name(dso); if (dso_name == NULL) return NULL; diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index 0b91f813c4fa..3be5e8d1e278 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -174,7 +174,7 @@ static inline bool elf_sec__is_data(const GElf_Shdr *shdr, static bool elf_sec__filter(GElf_Shdr *shdr, Elf_Data *secstrs) { - return elf_sec__is_text(shdr, secstrs) || + return elf_sec__is_text(shdr, secstrs) || elf_sec__is_data(shdr, secstrs); } @@ -312,8 +312,8 @@ static char *demangle_sym(struct dso *dso, int kmodule, const char *elf_name) * DWARF DW_compile_unit has this, but we don't always have access * to it... */ - if (!want_demangle(dso->kernel || kmodule)) - return demangled; + if (!want_demangle(dso__kernel(dso) || kmodule)) + return demangled; demangled = cxx_demangle_sym(elf_name, verbose > 0, verbose > 0); if (demangled == NULL) { @@ -470,7 +470,7 @@ static bool get_plt_sizes(struct dso *dso, GElf_Ehdr *ehdr, GElf_Shdr *shdr_plt, } if (*plt_entry_size) return true; - pr_debug("Missing PLT entry size for %s\n", dso->long_name); + pr_debug("Missing PLT entry size for %s\n", dso__long_name(dso)); return false; } @@ -654,7 +654,7 @@ static int dso__synthesize_plt_got_symbols(struct dso *dso, Elf *elf, sym = symbol__new(shdr.sh_offset + i, shdr.sh_entsize, STB_GLOBAL, STT_FUNC, buf); if (!sym) goto out; - symbols__insert(&dso->symbols, sym); + symbols__insert(dso__symbols(dso), sym); } err = 0; out: @@ -708,7 +708,7 @@ int dso__synthesize_plt_symbols(struct dso *dso, struct symsrc *ss) plt_sym = symbol__new(shdr_plt.sh_offset, plt_header_size, STB_GLOBAL, STT_FUNC, ".plt"); if (!plt_sym) goto out_elf_end; - symbols__insert(&dso->symbols, plt_sym); + symbols__insert(dso__symbols(dso), plt_sym); /* Only x86 has .plt.got */ if (machine_is_x86(ehdr.e_machine) && @@ -830,7 +830,7 @@ int dso__synthesize_plt_symbols(struct dso *dso, struct symsrc *ss) goto out_elf_end; plt_offset += plt_entry_size; - symbols__insert(&dso->symbols, f); + symbols__insert(dso__symbols(dso), f); ++nr; } @@ -840,7 +840,7 @@ out_elf_end: if (err == 0) return nr; pr_debug("%s: problems reading %s PLT info.\n", - __func__, dso->long_name); + __func__, dso__long_name(dso)); return 0; } @@ -1175,19 +1175,19 @@ static int dso__swap_init(struct dso *dso, unsigned char eidata) { static unsigned int const endian = 1; - dso->needs_swap = DSO_SWAP__NO; + dso__set_needs_swap(dso, DSO_SWAP__NO); switch (eidata) { case ELFDATA2LSB: /* We are big endian, DSO is little endian. */ if (*(unsigned char const *)&endian != 1) - dso->needs_swap = DSO_SWAP__YES; + dso__set_needs_swap(dso, DSO_SWAP__YES); break; case ELFDATA2MSB: /* We are little endian, DSO is big endian. */ if (*(unsigned char const *)&endian != 0) - dso->needs_swap = DSO_SWAP__YES; + dso__set_needs_swap(dso, DSO_SWAP__YES); break; default: @@ -1238,11 +1238,11 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name, if (fd < 0) return -1; - type = dso->symtab_type; + type = dso__symtab_type(dso); } else { fd = open(name, O_RDONLY); if (fd < 0) { - dso->load_errno = errno; + *dso__load_errno(dso) = errno; return -1; } } @@ -1250,37 +1250,37 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name, elf = elf_begin(fd, PERF_ELF_C_READ_MMAP, NULL); if (elf == NULL) { pr_debug("%s: cannot read %s ELF file.\n", __func__, name); - dso->load_errno = DSO_LOAD_ERRNO__INVALID_ELF; + *dso__load_errno(dso) = DSO_LOAD_ERRNO__INVALID_ELF; goto out_close; } if (gelf_getehdr(elf, &ehdr) == NULL) { - dso->load_errno = DSO_LOAD_ERRNO__INVALID_ELF; + *dso__load_errno(dso) = DSO_LOAD_ERRNO__INVALID_ELF; pr_debug("%s: cannot get elf header.\n", __func__); goto out_elf_end; } if (dso__swap_init(dso, ehdr.e_ident[EI_DATA])) { - dso->load_errno = DSO_LOAD_ERRNO__INTERNAL_ERROR; + *dso__load_errno(dso) = DSO_LOAD_ERRNO__INTERNAL_ERROR; goto out_elf_end; } /* Always reject images with a mismatched build-id: */ - if (dso->has_build_id && !symbol_conf.ignore_vmlinux_buildid) { + if (dso__has_build_id(dso) && !symbol_conf.ignore_vmlinux_buildid) { u8 build_id[BUILD_ID_SIZE]; struct build_id bid; int size; size = elf_read_build_id(elf, build_id, BUILD_ID_SIZE); if (size <= 0) { - dso->load_errno = DSO_LOAD_ERRNO__CANNOT_READ_BUILDID; + *dso__load_errno(dso) = DSO_LOAD_ERRNO__CANNOT_READ_BUILDID; goto out_elf_end; } build_id__init(&bid, build_id, size); if (!dso__build_id_equal(dso, &bid)) { pr_debug("%s: build id mismatch for %s.\n", __func__, name); - dso->load_errno = DSO_LOAD_ERRNO__MISMATCHING_BUILDID; + *dso__load_errno(dso) = DSO_LOAD_ERRNO__MISMATCHING_BUILDID; goto out_elf_end; } } @@ -1305,14 +1305,14 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name, if (ss->opdshdr.sh_type != SHT_PROGBITS) ss->opdsec = NULL; - if (dso->kernel == DSO_SPACE__USER) + if (dso__kernel(dso) == DSO_SPACE__USER) ss->adjust_symbols = true; else ss->adjust_symbols = elf__needs_adjust_symbols(ehdr); ss->name = strdup(name); if (!ss->name) { - dso->load_errno = errno; + *dso__load_errno(dso) = errno; goto out_elf_end; } @@ -1432,7 +1432,7 @@ static int dso__process_kernel_symbol(struct dso *dso, struct map *map, if (adjust_kernel_syms) sym->st_value -= shdr->sh_addr - shdr->sh_offset; - if (strcmp(section_name, (curr_dso->short_name + dso->short_name_len)) == 0) + if (strcmp(section_name, (dso__short_name(curr_dso) + dso__short_name_len(dso))) == 0) return 0; if (strcmp(section_name, ".text") == 0) { @@ -1441,7 +1441,7 @@ static int dso__process_kernel_symbol(struct dso *dso, struct map *map, * kallsyms and identity maps. Overwrite it to * map to the kernel dso. */ - if (*remap_kernel && dso->kernel && !kmodule) { + if (*remap_kernel && dso__kernel(dso) && !kmodule) { *remap_kernel = false; map__set_start(map, shdr->sh_addr + ref_reloc(kmap)); map__set_end(map, map__start(map) + shdr->sh_size); @@ -1489,7 +1489,7 @@ static int dso__process_kernel_symbol(struct dso *dso, struct map *map, return 0; } - snprintf(dso_name, sizeof(dso_name), "%s%s", dso->short_name, section_name); + snprintf(dso_name, sizeof(dso_name), "%s%s", dso__short_name(dso), section_name); curr_map = maps__find_by_name(kmaps, dso_name); if (curr_map == NULL) { @@ -1501,17 +1501,17 @@ static int dso__process_kernel_symbol(struct dso *dso, struct map *map, curr_dso = dso__new(dso_name); if (curr_dso == NULL) return -1; - curr_dso->kernel = dso->kernel; - curr_dso->long_name = dso->long_name; - curr_dso->long_name_len = dso->long_name_len; - curr_dso->binary_type = dso->binary_type; - curr_dso->adjust_symbols = dso->adjust_symbols; + dso__set_kernel(curr_dso, dso__kernel(dso)); + RC_CHK_ACCESS(curr_dso)->long_name = dso__long_name(dso); + RC_CHK_ACCESS(curr_dso)->long_name_len = dso__long_name_len(dso); + dso__set_binary_type(curr_dso, dso__binary_type(dso)); + dso__set_adjust_symbols(curr_dso, dso__adjust_symbols(dso)); curr_map = map__new2(start, curr_dso); dso__put(curr_dso); if (curr_map == NULL) return -1; - if (curr_dso->kernel) + if (dso__kernel(curr_dso)) map__kmap(curr_map)->kmaps = kmaps; if (adjust_kernel_syms) { @@ -1521,7 +1521,7 @@ static int dso__process_kernel_symbol(struct dso *dso, struct map *map, } else { map__set_mapping_type(curr_map, MAPPING_TYPE__IDENTITY); } - curr_dso->symtab_type = dso->symtab_type; + dso__set_symtab_type(curr_dso, dso__symtab_type(dso)); if (maps__insert(kmaps, curr_map)) return -1; /* @@ -1547,7 +1547,7 @@ static int dso__load_sym_internal(struct dso *dso, struct map *map, struct symsrc *syms_ss, struct symsrc *runtime_ss, int kmodule, int dynsym) { - struct kmap *kmap = dso->kernel ? map__kmap(map) : NULL; + struct kmap *kmap = dso__kernel(dso) ? map__kmap(map) : NULL; struct maps *kmaps = kmap ? map__kmaps(map) : NULL; struct map *curr_map = map; struct dso *curr_dso = dso; @@ -1581,8 +1581,8 @@ dso__load_sym_internal(struct dso *dso, struct map *map, struct symsrc *syms_ss, if (elf_section_by_name(runtime_ss->elf, &runtime_ss->ehdr, &tshdr, ".text", NULL)) { - dso->text_offset = tshdr.sh_addr - tshdr.sh_offset; - dso->text_end = tshdr.sh_offset + tshdr.sh_size; + dso__set_text_offset(dso, tshdr.sh_addr - tshdr.sh_offset); + dso__set_text_end(dso, tshdr.sh_offset + tshdr.sh_size); } if (runtime_ss->opdsec) @@ -1641,16 +1641,16 @@ dso__load_sym_internal(struct dso *dso, struct map *map, struct symsrc *syms_ss, * attempted to prelink vdso to its virtual address. */ if (dso__is_vdso(dso)) - map__set_reloc(map, map__start(map) - dso->text_offset); + map__set_reloc(map, map__start(map) - dso__text_offset(dso)); - dso->adjust_symbols = runtime_ss->adjust_symbols || ref_reloc(kmap); + dso__set_adjust_symbols(dso, runtime_ss->adjust_symbols || ref_reloc(kmap)); /* * Initial kernel and module mappings do not map to the dso. * Flag the fixups. */ - if (dso->kernel) { + if (dso__kernel(dso)) { remap_kernel = true; - adjust_kernel_syms = dso->adjust_symbols; + adjust_kernel_syms = dso__adjust_symbols(dso); } if (kmodule && adjust_kernel_syms) @@ -1743,7 +1743,7 @@ dso__load_sym_internal(struct dso *dso, struct map *map, struct symsrc *syms_ss, (sym.st_value & 1)) --sym.st_value; - if (dso->kernel) { + if (dso__kernel(dso)) { if (dso__process_kernel_symbol(dso, map, &sym, &shdr, kmaps, kmap, &curr_dso, &curr_map, section_name, adjust_kernel_syms, kmodule, &remap_kernel, max_text_sh_offset)) @@ -1792,7 +1792,7 @@ dso__load_sym_internal(struct dso *dso, struct map *map, struct symsrc *syms_ss, arch__sym_update(f, &sym); - __symbols__insert(&curr_dso->symbols, f, dso->kernel); + __symbols__insert(dso__symbols(curr_dso), f, dso__kernel(dso)); nr++; } @@ -1800,8 +1800,8 @@ dso__load_sym_internal(struct dso *dso, struct map *map, struct symsrc *syms_ss, * For misannotated, zeroed, ASM function sizes. */ if (nr > 0) { - symbols__fixup_end(&dso->symbols, false); - symbols__fixup_duplicate(&dso->symbols); + symbols__fixup_end(dso__symbols(dso), false); + symbols__fixup_duplicate(dso__symbols(dso)); if (kmap) { /* * We need to fixup this here too because we create new @@ -1821,16 +1821,16 @@ int dso__load_sym(struct dso *dso, struct map *map, struct symsrc *syms_ss, int nr = 0; int err = -1; - dso->symtab_type = syms_ss->type; - dso->is_64_bit = syms_ss->is_64_bit; - dso->rel = syms_ss->ehdr.e_type == ET_REL; + dso__set_symtab_type(dso, syms_ss->type); + dso__set_is_64_bit(dso, syms_ss->is_64_bit); + dso__set_rel(dso, syms_ss->ehdr.e_type == ET_REL); /* * Modules may already have symbols from kallsyms, but those symbols * have the wrong values for the dso maps, so remove them. */ if (kmodule && syms_ss->symtab) - symbols__delete(&dso->symbols); + symbols__delete(dso__symbols(dso)); if (!syms_ss->symtab) { /* @@ -1838,7 +1838,7 @@ int dso__load_sym(struct dso *dso, struct map *map, struct symsrc *syms_ss, * to using kallsyms. The vmlinux runtime symbols aren't * of much use. */ - if (dso->kernel) + if (dso__kernel(dso)) return err; } else { err = dso__load_sym_internal(dso, map, syms_ss, runtime_ss, diff --git a/tools/perf/util/symbol-minimal.c b/tools/perf/util/symbol-minimal.c index 1da8b713509c..c6f369b5d893 100644 --- a/tools/perf/util/symbol-minimal.c +++ b/tools/perf/util/symbol-minimal.c @@ -273,7 +273,7 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name, out_close: close(fd); out_errno: - dso->load_errno = errno; + RC_CHK_ACCESS(dso)->load_errno = errno; return -1; } @@ -348,7 +348,7 @@ int dso__load_sym(struct dso *dso, struct map *map __maybe_unused, ret = fd__is_64_bit(ss->fd); if (ret >= 0) - dso->is_64_bit = ret; + RC_CHK_ACCESS(dso)->is_64_bit = ret; if (filename__read_build_id(ss->name, &bid) > 0) dso__set_build_id(dso, &bid); diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 9ebdb8e13c0b..7772a4d3e66c 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -532,52 +532,52 @@ static struct symbol *symbols__find_by_name(struct symbol *symbols[], void dso__reset_find_symbol_cache(struct dso *dso) { - dso->last_find_result.addr = 0; - dso->last_find_result.symbol = NULL; + dso__set_last_find_result_addr(dso, 0); + dso__set_last_find_result_symbol(dso, NULL); } void dso__insert_symbol(struct dso *dso, struct symbol *sym) { - __symbols__insert(&dso->symbols, sym, dso->kernel); + __symbols__insert(dso__symbols(dso), sym, dso__kernel(dso)); /* update the symbol cache if necessary */ - if (dso->last_find_result.addr >= sym->start && - (dso->last_find_result.addr < sym->end || + if (dso__last_find_result_addr(dso) >= sym->start && + (dso__last_find_result_addr(dso) < sym->end || sym->start == sym->end)) { - dso->last_find_result.symbol = sym; + dso__set_last_find_result_symbol(dso, sym); } } void dso__delete_symbol(struct dso *dso, struct symbol *sym) { - rb_erase_cached(&sym->rb_node, &dso->symbols); + rb_erase_cached(&sym->rb_node, dso__symbols(dso)); symbol__delete(sym); dso__reset_find_symbol_cache(dso); } struct symbol *dso__find_symbol(struct dso *dso, u64 addr) { - if (dso->last_find_result.addr != addr || dso->last_find_result.symbol == NULL) { - dso->last_find_result.addr = addr; - dso->last_find_result.symbol = symbols__find(&dso->symbols, addr); + if (dso__last_find_result_addr(dso) != addr || dso__last_find_result_symbol(dso) == NULL) { + dso__set_last_find_result_addr(dso, addr); + dso__set_last_find_result_symbol(dso, symbols__find(dso__symbols(dso), addr)); } - return dso->last_find_result.symbol; + return dso__last_find_result_symbol(dso); } struct symbol *dso__find_symbol_nocache(struct dso *dso, u64 addr) { - return symbols__find(&dso->symbols, addr); + return symbols__find(dso__symbols(dso), addr); } struct symbol *dso__first_symbol(struct dso *dso) { - return symbols__first(&dso->symbols); + return symbols__first(dso__symbols(dso)); } struct symbol *dso__last_symbol(struct dso *dso) { - return symbols__last(&dso->symbols); + return symbols__last(dso__symbols(dso)); } struct symbol *dso__next_symbol(struct symbol *sym) @@ -587,11 +587,11 @@ struct symbol *dso__next_symbol(struct symbol *sym) struct symbol *dso__next_symbol_by_name(struct dso *dso, size_t *idx) { - if (*idx + 1 >= dso->symbol_names_len) + if (*idx + 1 >= dso__symbol_names_len(dso)) return NULL; ++*idx; - return dso->symbol_names[*idx]; + return dso__symbol_names(dso)[*idx]; } /* @@ -599,27 +599,29 @@ struct symbol *dso__next_symbol_by_name(struct dso *dso, size_t *idx) */ struct symbol *dso__find_symbol_by_name(struct dso *dso, const char *name, size_t *idx) { - struct symbol *s = symbols__find_by_name(dso->symbol_names, dso->symbol_names_len, - name, SYMBOL_TAG_INCLUDE__NONE, idx); - if (!s) - s = symbols__find_by_name(dso->symbol_names, dso->symbol_names_len, - name, SYMBOL_TAG_INCLUDE__DEFAULT_ONLY, idx); + struct symbol *s = symbols__find_by_name(dso__symbol_names(dso), + dso__symbol_names_len(dso), + name, SYMBOL_TAG_INCLUDE__NONE, idx); + if (!s) { + s = symbols__find_by_name(dso__symbol_names(dso), dso__symbol_names_len(dso), + name, SYMBOL_TAG_INCLUDE__DEFAULT_ONLY, idx); + } return s; } void dso__sort_by_name(struct dso *dso) { - mutex_lock(&dso->lock); + mutex_lock(dso__lock(dso)); if (!dso__sorted_by_name(dso)) { size_t len; - dso->symbol_names = symbols__sort_by_name(&dso->symbols, &len); - if (dso->symbol_names) { - dso->symbol_names_len = len; + dso__set_symbol_names(dso, symbols__sort_by_name(dso__symbols(dso), &len)); + if (dso__symbol_names(dso)) { + dso__set_symbol_names_len(dso, len); dso__set_sorted_by_name(dso); } } - mutex_unlock(&dso->lock); + mutex_unlock(dso__lock(dso)); } /* @@ -746,7 +748,7 @@ static int map__process_kallsym_symbol(void *arg, const char *name, { struct symbol *sym; struct dso *dso = arg; - struct rb_root_cached *root = &dso->symbols; + struct rb_root_cached *root = dso__symbols(dso); if (!symbol_type__filter(type)) return 0; @@ -786,8 +788,8 @@ static int maps__split_kallsyms_for_kcore(struct maps *kmaps, struct dso *dso) { struct symbol *pos; int count = 0; - struct rb_root_cached old_root = dso->symbols; - struct rb_root_cached *root = &dso->symbols; + struct rb_root_cached *root = dso__symbols(dso); + struct rb_root_cached old_root = *root; struct rb_node *next = rb_first_cached(root); if (!kmaps) @@ -821,13 +823,13 @@ static int maps__split_kallsyms_for_kcore(struct maps *kmaps, struct dso *dso) pos->end = map__end(curr_map); if (pos->end) pos->end -= map__start(curr_map) - map__pgoff(curr_map); - symbols__insert(&curr_map_dso->symbols, pos); + symbols__insert(dso__symbols(curr_map_dso), pos); ++count; map__put(curr_map); } /* Symbols have been adjusted */ - dso->adjust_symbols = 1; + dso__set_adjust_symbols(dso, true); return count; } @@ -844,7 +846,7 @@ static int maps__split_kallsyms(struct maps *kmaps, struct dso *dso, u64 delta, struct map *curr_map = map__get(initial_map); struct symbol *pos; int count = 0, moved = 0; - struct rb_root_cached *root = &dso->symbols; + struct rb_root_cached *root = dso__symbols(dso); struct rb_node *next = rb_first_cached(root); int kernel_range = 0; bool x86_64; @@ -871,9 +873,9 @@ static int maps__split_kallsyms(struct maps *kmaps, struct dso *dso, u64 delta, *module++ = '\0'; curr_map_dso = map__dso(curr_map); - if (strcmp(curr_map_dso->short_name, module)) { + if (strcmp(dso__short_name(curr_map_dso), module)) { if (!RC_CHK_EQUAL(curr_map, initial_map) && - dso->kernel == DSO_SPACE__KERNEL_GUEST && + dso__kernel(dso) == DSO_SPACE__KERNEL_GUEST && machine__is_default_guest(machine)) { /* * We assume all symbols of a module are @@ -896,7 +898,7 @@ static int maps__split_kallsyms(struct maps *kmaps, struct dso *dso, u64 delta, goto discard_symbol; } curr_map_dso = map__dso(curr_map); - if (curr_map_dso->loaded && + if (dso__loaded(curr_map_dso) && !machine__is_default_guest(machine)) goto discard_symbol; } @@ -932,7 +934,7 @@ static int maps__split_kallsyms(struct maps *kmaps, struct dso *dso, u64 delta, goto add_symbol; } - if (dso->kernel == DSO_SPACE__KERNEL_GUEST) + if (dso__kernel(dso) == DSO_SPACE__KERNEL_GUEST) snprintf(dso_name, sizeof(dso_name), "[guest.kernel].%d", kernel_range++); @@ -946,7 +948,7 @@ static int maps__split_kallsyms(struct maps *kmaps, struct dso *dso, u64 delta, if (ndso == NULL) return -1; - ndso->kernel = dso->kernel; + dso__set_kernel(ndso, dso__kernel(dso)); curr_map = map__new2(pos->start, ndso); if (curr_map == NULL) { @@ -971,7 +973,7 @@ add_symbol: struct dso *curr_map_dso = map__dso(curr_map); rb_erase_cached(&pos->rb_node, root); - symbols__insert(&curr_map_dso->symbols, pos); + symbols__insert(dso__symbols(curr_map_dso), pos); ++moved; } else ++count; @@ -983,7 +985,7 @@ discard_symbol: } if (!RC_CHK_EQUAL(curr_map, initial_map) && - dso->kernel == DSO_SPACE__KERNEL_GUEST && + dso__kernel(dso) == DSO_SPACE__KERNEL_GUEST && machine__is_default_guest(maps__machine(kmaps))) { dso__set_loaded(map__dso(curr_map)); } @@ -1157,7 +1159,7 @@ static int do_validate_kcore_modules_cb(struct map *old_map, void *data) dso = map__dso(old_map); /* Module must be in memory at the same address */ - mi = find_module(dso->short_name, modules); + mi = find_module(dso__short_name(dso), modules); if (!mi || mi->start != map__start(old_map)) return -EINVAL; @@ -1326,7 +1328,7 @@ static int dso__load_kcore(struct dso *dso, struct map *map, &is_64_bit); if (err) goto out_err; - dso->is_64_bit = is_64_bit; + dso__set_is_64_bit(dso, is_64_bit); if (list_empty(&md.maps)) { err = -EINVAL; @@ -1418,10 +1420,10 @@ static int dso__load_kcore(struct dso *dso, struct map *map, * Set the data type and long name so that kcore can be read via * dso__data_read_addr(). */ - if (dso->kernel == DSO_SPACE__KERNEL_GUEST) - dso->binary_type = DSO_BINARY_TYPE__GUEST_KCORE; + if (dso__kernel(dso) == DSO_SPACE__KERNEL_GUEST) + dso__set_binary_type(dso, DSO_BINARY_TYPE__GUEST_KCORE); else - dso->binary_type = DSO_BINARY_TYPE__KCORE; + dso__set_binary_type(dso, DSO_BINARY_TYPE__KCORE); dso__set_long_name(dso, strdup(kcore_filename), true); close(fd); @@ -1482,13 +1484,13 @@ int __dso__load_kallsyms(struct dso *dso, const char *filename, if (kallsyms__delta(kmap, filename, &delta)) return -1; - symbols__fixup_end(&dso->symbols, true); - symbols__fixup_duplicate(&dso->symbols); + symbols__fixup_end(dso__symbols(dso), true); + symbols__fixup_duplicate(dso__symbols(dso)); - if (dso->kernel == DSO_SPACE__KERNEL_GUEST) - dso->symtab_type = DSO_BINARY_TYPE__GUEST_KALLSYMS; + if (dso__kernel(dso) == DSO_SPACE__KERNEL_GUEST) + dso__set_symtab_type(dso, DSO_BINARY_TYPE__GUEST_KALLSYMS); else - dso->symtab_type = DSO_BINARY_TYPE__KALLSYMS; + dso__set_symtab_type(dso, DSO_BINARY_TYPE__KALLSYMS); if (!no_kcore && !dso__load_kcore(dso, map, filename)) return maps__split_kallsyms_for_kcore(kmap->kmaps, dso); @@ -1544,7 +1546,7 @@ static int dso__load_perf_map(const char *map_path, struct dso *dso) if (sym == NULL) goto out_delete_line; - symbols__insert(&dso->symbols, sym); + symbols__insert(dso__symbols(dso), sym); nr_syms++; } @@ -1670,15 +1672,15 @@ int dso__load_bfd_symbols(struct dso *dso, const char *debugfile) if (!symbol) goto out_free; - symbols__insert(&dso->symbols, symbol); + symbols__insert(dso__symbols(dso), symbol); } #ifdef bfd_get_section #undef bfd_asymbol_section #endif - symbols__fixup_end(&dso->symbols, false); - symbols__fixup_duplicate(&dso->symbols); - dso->adjust_symbols = 1; + symbols__fixup_end(dso__symbols(dso), false); + symbols__fixup_duplicate(dso__symbols(dso)); + dso__set_adjust_symbols(dso, true); err = 0; out_free: @@ -1701,17 +1703,17 @@ static bool dso__is_compatible_symtab_type(struct dso *dso, bool kmod, case DSO_BINARY_TYPE__MIXEDUP_UBUNTU_DEBUGINFO: case DSO_BINARY_TYPE__BUILDID_DEBUGINFO: case DSO_BINARY_TYPE__OPENEMBEDDED_DEBUGINFO: - return !kmod && dso->kernel == DSO_SPACE__USER; + return !kmod && dso__kernel(dso) == DSO_SPACE__USER; case DSO_BINARY_TYPE__KALLSYMS: case DSO_BINARY_TYPE__VMLINUX: case DSO_BINARY_TYPE__KCORE: - return dso->kernel == DSO_SPACE__KERNEL; + return dso__kernel(dso) == DSO_SPACE__KERNEL; case DSO_BINARY_TYPE__GUEST_KALLSYMS: case DSO_BINARY_TYPE__GUEST_VMLINUX: case DSO_BINARY_TYPE__GUEST_KCORE: - return dso->kernel == DSO_SPACE__KERNEL_GUEST; + return dso__kernel(dso) == DSO_SPACE__KERNEL_GUEST; case DSO_BINARY_TYPE__GUEST_KMODULE: case DSO_BINARY_TYPE__GUEST_KMODULE_COMP: @@ -1721,7 +1723,7 @@ static bool dso__is_compatible_symtab_type(struct dso *dso, bool kmod, * kernel modules know their symtab type - it's set when * creating a module dso in machine__addnew_module_map(). */ - return kmod && dso->symtab_type == type; + return kmod && dso__symtab_type(dso) == type; case DSO_BINARY_TYPE__BUILD_ID_CACHE: case DSO_BINARY_TYPE__BUILD_ID_CACHE_DEBUGINFO: @@ -1789,18 +1791,19 @@ int dso__load(struct dso *dso, struct map *map) struct build_id bid; struct nscookie nsc; char newmapname[PATH_MAX]; - const char *map_path = dso->long_name; + const char *map_path = dso__long_name(dso); - mutex_lock(&dso->lock); - perfmap = strncmp(dso->name, "/tmp/perf-", 10) == 0; + mutex_lock(dso__lock(dso)); + perfmap = strncmp(dso__name(dso), "/tmp/perf-", 10) == 0; if (perfmap) { - if (dso->nsinfo && (dso__find_perf_map(newmapname, - sizeof(newmapname), &dso->nsinfo) == 0)) { + if (dso__nsinfo(dso) && + (dso__find_perf_map(newmapname, sizeof(newmapname), + dso__nsinfo_ptr(dso)) == 0)) { map_path = newmapname; } } - nsinfo__mountns_enter(dso->nsinfo, &nsc); + nsinfo__mountns_enter(dso__nsinfo(dso), &nsc); /* check again under the dso->lock */ if (dso__loaded(dso)) { @@ -1808,15 +1811,15 @@ int dso__load(struct dso *dso, struct map *map) goto out; } - kmod = dso->symtab_type == DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE || - dso->symtab_type == DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE_COMP || - dso->symtab_type == DSO_BINARY_TYPE__GUEST_KMODULE || - dso->symtab_type == DSO_BINARY_TYPE__GUEST_KMODULE_COMP; + kmod = dso__symtab_type(dso) == DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE || + dso__symtab_type(dso) == DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE_COMP || + dso__symtab_type(dso) == DSO_BINARY_TYPE__GUEST_KMODULE || + dso__symtab_type(dso) == DSO_BINARY_TYPE__GUEST_KMODULE_COMP; - if (dso->kernel && !kmod) { - if (dso->kernel == DSO_SPACE__KERNEL) + if (dso__kernel(dso) && !kmod) { + if (dso__kernel(dso) == DSO_SPACE__KERNEL) ret = dso__load_kernel_sym(dso, map); - else if (dso->kernel == DSO_SPACE__KERNEL_GUEST) + else if (dso__kernel(dso) == DSO_SPACE__KERNEL_GUEST) ret = dso__load_guest_kernel_sym(dso, map); machine = maps__machine(map__kmaps(map)); @@ -1825,12 +1828,13 @@ int dso__load(struct dso *dso, struct map *map) goto out; } - dso->adjust_symbols = 0; + dso__set_adjust_symbols(dso, false); if (perfmap) { ret = dso__load_perf_map(map_path, dso); - dso->symtab_type = ret > 0 ? DSO_BINARY_TYPE__JAVA_JIT : - DSO_BINARY_TYPE__NOT_FOUND; + dso__set_symtab_type(dso, ret > 0 + ? DSO_BINARY_TYPE__JAVA_JIT + : DSO_BINARY_TYPE__NOT_FOUND); goto out; } @@ -1845,9 +1849,9 @@ int dso__load(struct dso *dso, struct map *map) * Read the build id if possible. This is required for * DSO_BINARY_TYPE__BUILDID_DEBUGINFO to work */ - if (!dso->has_build_id && - is_regular_file(dso->long_name)) { - __symbol__join_symfs(name, PATH_MAX, dso->long_name); + if (!dso__has_build_id(dso) && + is_regular_file(dso__long_name(dso))) { + __symbol__join_symfs(name, PATH_MAX, dso__long_name(dso)); if (filename__read_build_id(name, &bid) > 0) dso__set_build_id(dso, &bid); } @@ -1881,7 +1885,7 @@ int dso__load(struct dso *dso, struct map *map) nsinfo__mountns_exit(&nsc); is_reg = is_regular_file(name); - if (!is_reg && errno == ENOENT && dso->nsinfo) { + if (!is_reg && errno == ENOENT && dso__nsinfo(dso)) { char *new_name = dso__filename_with_chroot(dso, name); if (new_name) { is_reg = is_regular_file(new_name); @@ -1898,7 +1902,7 @@ int dso__load(struct dso *dso, struct map *map) sirc = symsrc__init(ss, dso, name, symtab_type); if (nsexit) - nsinfo__mountns_enter(dso->nsinfo, &nsc); + nsinfo__mountns_enter(dso__nsinfo(dso), &nsc); if (bfdrc == 0) { ret = 0; @@ -1911,8 +1915,8 @@ int dso__load(struct dso *dso, struct map *map) if (!syms_ss && symsrc__has_symtab(ss)) { syms_ss = ss; next_slot = true; - if (!dso->symsrc_filename) - dso->symsrc_filename = strdup(name); + if (!dso__symsrc_filename(dso)) + dso__set_symsrc_filename(dso, strdup(name)); } if (!runtime_ss && symsrc__possibly_runtime(ss)) { @@ -1959,11 +1963,11 @@ int dso__load(struct dso *dso, struct map *map) symsrc__destroy(&ss_[ss_pos - 1]); out_free: free(name); - if (ret < 0 && strstr(dso->name, " (deleted)") != NULL) + if (ret < 0 && strstr(dso__name(dso), " (deleted)") != NULL) ret = 0; out: dso__set_loaded(dso); - mutex_unlock(&dso->lock); + mutex_unlock(dso__lock(dso)); nsinfo__mountns_exit(&nsc); return ret; @@ -1982,7 +1986,7 @@ int dso__load_vmlinux(struct dso *dso, struct map *map, else symbol__join_symfs(symfs_vmlinux, vmlinux); - if (dso->kernel == DSO_SPACE__KERNEL_GUEST) + if (dso__kernel(dso) == DSO_SPACE__KERNEL_GUEST) symtab_type = DSO_BINARY_TYPE__GUEST_VMLINUX; else symtab_type = DSO_BINARY_TYPE__VMLINUX; @@ -1995,10 +1999,10 @@ int dso__load_vmlinux(struct dso *dso, struct map *map, * an incorrect long name unless we set it here first. */ dso__set_long_name(dso, vmlinux, vmlinux_allocated); - if (dso->kernel == DSO_SPACE__KERNEL_GUEST) - dso->binary_type = DSO_BINARY_TYPE__GUEST_VMLINUX; + if (dso__kernel(dso) == DSO_SPACE__KERNEL_GUEST) + dso__set_binary_type(dso, DSO_BINARY_TYPE__GUEST_VMLINUX); else - dso->binary_type = DSO_BINARY_TYPE__VMLINUX; + dso__set_binary_type(dso, DSO_BINARY_TYPE__VMLINUX); err = dso__load_sym(dso, map, &ss, &ss, 0); symsrc__destroy(&ss); @@ -2091,7 +2095,7 @@ static char *dso__find_kallsyms(struct dso *dso, struct map *map) bool is_host = false; char path[PATH_MAX]; - if (!dso->has_build_id) { + if (!dso__has_build_id(dso)) { /* * Last resort, if we don't have a build-id and couldn't find * any vmlinux file, try the running kernel kallsyms table. @@ -2116,7 +2120,7 @@ static char *dso__find_kallsyms(struct dso *dso, struct map *map) goto proc_kallsyms; } - build_id__sprintf(&dso->bid, sbuild_id); + build_id__sprintf(dso__bid(dso), sbuild_id); /* Find kallsyms in build-id cache with kcore */ scnprintf(path, sizeof(path), "%s/%s/%s", @@ -2209,7 +2213,7 @@ do_kallsyms: free(kallsyms_allocated_filename); if (err > 0 && !dso__is_kcore(dso)) { - dso->binary_type = DSO_BINARY_TYPE__KALLSYMS; + dso__set_binary_type(dso, DSO_BINARY_TYPE__KALLSYMS); dso__set_long_name(dso, DSO__NAME_KALLSYMS, false); map__fixup_start(map); map__fixup_end(map); @@ -2252,7 +2256,7 @@ static int dso__load_guest_kernel_sym(struct dso *dso, struct map *map) if (err > 0) pr_debug("Using %s for symbols\n", kallsyms_filename); if (err > 0 && !dso__is_kcore(dso)) { - dso->binary_type = DSO_BINARY_TYPE__GUEST_KALLSYMS; + dso__set_binary_type(dso, DSO_BINARY_TYPE__GUEST_KALLSYMS); dso__set_long_name(dso, machine->mmap_name, false); map__fixup_start(map); map__fixup_end(map); diff --git a/tools/perf/util/symbol_fprintf.c b/tools/perf/util/symbol_fprintf.c index 088f4abf230f..53e1af4ed9ac 100644 --- a/tools/perf/util/symbol_fprintf.c +++ b/tools/perf/util/symbol_fprintf.c @@ -64,8 +64,8 @@ size_t dso__fprintf_symbols_by_name(struct dso *dso, { size_t ret = 0; - for (size_t i = 0; i < dso->symbol_names_len; i++) { - struct symbol *pos = dso->symbol_names[i]; + for (size_t i = 0; i < dso__symbol_names_len(dso); i++) { + struct symbol *pos = dso__symbol_names(dso)[i]; ret += fprintf(fp, "%s\n", pos->name); } diff --git a/tools/perf/util/synthetic-events.c b/tools/perf/util/synthetic-events.c index 2a0289c14959..5498048f56ea 100644 --- a/tools/perf/util/synthetic-events.c +++ b/tools/perf/util/synthetic-events.c @@ -385,8 +385,8 @@ static void perf_record_mmap2__read_build_id(struct perf_record_mmap2 *event, id.ino_generation = event->ino_generation; dso = dsos__findnew_id(&machine->dsos, event->filename, &id); - if (dso && dso->has_build_id) { - bid = dso->bid; + if (dso && dso__has_build_id(dso)) { + bid = *dso__bid(dso); rc = 0; goto out; } @@ -407,7 +407,7 @@ out: event->__reserved_1 = 0; event->__reserved_2 = 0; - if (dso && !dso->has_build_id) + if (dso && !dso__has_build_id(dso)) dso__set_build_id(dso, &bid); } else { if (event->filename[0] == '/') { @@ -684,7 +684,7 @@ static int perf_event__synthesize_modules_maps_cb(struct map *map, void *data) dso = map__dso(map); if (symbol_conf.buildid_mmap2) { - size = PERF_ALIGN(dso->long_name_len + 1, sizeof(u64)); + size = PERF_ALIGN(dso__long_name_len(dso) + 1, sizeof(u64)); event->mmap2.header.type = PERF_RECORD_MMAP2; event->mmap2.header.size = (sizeof(event->mmap2) - (sizeof(event->mmap2.filename) - size)); @@ -694,11 +694,11 @@ static int perf_event__synthesize_modules_maps_cb(struct map *map, void *data) event->mmap2.len = map__size(map); event->mmap2.pid = args->machine->pid; - memcpy(event->mmap2.filename, dso->long_name, dso->long_name_len + 1); + memcpy(event->mmap2.filename, dso__long_name(dso), dso__long_name_len(dso) + 1); perf_record_mmap2__read_build_id(&event->mmap2, args->machine, false); } else { - size = PERF_ALIGN(dso->long_name_len + 1, sizeof(u64)); + size = PERF_ALIGN(dso__long_name_len(dso) + 1, sizeof(u64)); event->mmap.header.type = PERF_RECORD_MMAP; event->mmap.header.size = (sizeof(event->mmap) - (sizeof(event->mmap.filename) - size)); @@ -708,7 +708,7 @@ static int perf_event__synthesize_modules_maps_cb(struct map *map, void *data) event->mmap.len = map__size(map); event->mmap.pid = args->machine->pid; - memcpy(event->mmap.filename, dso->long_name, dso->long_name_len + 1); + memcpy(event->mmap.filename, dso__long_name(dso), dso__long_name_len(dso) + 1); } if (perf_tool__process_synth_event(args->tool, event, args->machine, args->process) != 0) @@ -2231,20 +2231,20 @@ int perf_event__synthesize_build_id(struct perf_tool *tool, struct dso *pos, u16 union perf_event ev; size_t len; - if (!pos->hit) + if (!dso__hit(pos)) return 0; memset(&ev, 0, sizeof(ev)); - len = pos->long_name_len + 1; + len = dso__long_name_len(pos) + 1; len = PERF_ALIGN(len, NAME_ALIGN); - ev.build_id.size = min(pos->bid.size, sizeof(pos->bid.data)); - memcpy(&ev.build_id.build_id, pos->bid.data, ev.build_id.size); + ev.build_id.size = min(dso__bid(pos)->size, sizeof(dso__bid(pos)->data)); + memcpy(&ev.build_id.build_id, dso__bid(pos)->data, ev.build_id.size); ev.build_id.header.type = PERF_RECORD_HEADER_BUILD_ID; ev.build_id.header.misc = misc | PERF_RECORD_MISC_BUILD_ID_SIZE; ev.build_id.pid = machine->pid; ev.build_id.header.size = sizeof(ev.build_id) + len; - memcpy(&ev.build_id.filename, pos->long_name, pos->long_name_len); + memcpy(&ev.build_id.filename, dso__long_name(pos), dso__long_name_len(pos)); return process(tool, &ev, NULL, machine); } diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index 1aa8962dcf52..0a473112f881 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -457,14 +457,14 @@ int thread__memcpy(struct thread *thread, struct machine *machine, dso = map__dso(al.map); - if (!dso || dso->data.status == DSO_DATA_STATUS_ERROR || map__load(al.map) < 0) { + if (!dso || dso__data(dso)->status == DSO_DATA_STATUS_ERROR || map__load(al.map) < 0) { addr_location__exit(&al); return -1; } offset = map__map_ip(al.map, ip); if (is64bit) - *is64bit = dso->is_64_bit; + *is64bit = dso__is_64_bit(dso); addr_location__exit(&al); diff --git a/tools/perf/util/unwind-libunwind-local.c b/tools/perf/util/unwind-libunwind-local.c index 6a5ac0faa6f4..cde267ea3e99 100644 --- a/tools/perf/util/unwind-libunwind-local.c +++ b/tools/perf/util/unwind-libunwind-local.c @@ -329,27 +329,27 @@ static int read_unwind_spec_eh_frame(struct dso *dso, struct unwind_info *ui, }; int ret, fd; - if (dso->data.eh_frame_hdr_offset == 0) { + if (dso__data(dso)->eh_frame_hdr_offset == 0) { fd = dso__data_get_fd(dso, ui->machine); if (fd < 0) return -EINVAL; /* Check the .eh_frame section for unwinding info */ ret = elf_section_address_and_offset(fd, ".eh_frame_hdr", - &dso->data.eh_frame_hdr_addr, - &dso->data.eh_frame_hdr_offset); - dso->data.elf_base_addr = elf_base_address(fd); + &dso__data(dso)->eh_frame_hdr_addr, + &dso__data(dso)->eh_frame_hdr_offset); + dso__data(dso)->elf_base_addr = elf_base_address(fd); dso__data_put_fd(dso); - if (ret || dso->data.eh_frame_hdr_offset == 0) + if (ret || dso__data(dso)->eh_frame_hdr_offset == 0) return -EINVAL; } maps__for_each_map(thread__maps(ui->thread), read_unwind_spec_eh_frame_maps_cb, &args); - args.base_addr -= dso->data.elf_base_addr; + args.base_addr -= dso__data(dso)->elf_base_addr; /* Address of .eh_frame_hdr */ - *segbase = args.base_addr + dso->data.eh_frame_hdr_addr; - ret = unwind_spec_ehframe(dso, ui->machine, dso->data.eh_frame_hdr_offset, + *segbase = args.base_addr + dso__data(dso)->eh_frame_hdr_addr; + ret = unwind_spec_ehframe(dso, ui->machine, dso__data(dso)->eh_frame_hdr_offset, table_data, fde_count); if (ret) return ret; @@ -460,7 +460,7 @@ find_proc_info(unw_addr_space_t as, unw_word_t ip, unw_proc_info_t *pi, return -EINVAL; } - pr_debug("unwind: find_proc_info dso %s\n", dso->name); + pr_debug("unwind: find_proc_info dso %s\n", dso__name(dso)); /* Check the .eh_frame section for unwinding info */ if (!read_unwind_spec_eh_frame(dso, ui, &table_data, &segbase, &fde_count)) { diff --git a/tools/perf/util/unwind-libunwind.c b/tools/perf/util/unwind-libunwind.c index 2728eb4f13ea..cb8be6acfb6f 100644 --- a/tools/perf/util/unwind-libunwind.c +++ b/tools/perf/util/unwind-libunwind.c @@ -25,7 +25,7 @@ int unwind__prepare_access(struct maps *maps, struct map *map, bool *initialized return 0; if (maps__addr_space(maps)) { - pr_debug("unwind: thread map already set, dso=%s\n", dso->name); + pr_debug("unwind: thread map already set, dso=%s\n", dso__name(dso)); if (initialized) *initialized = true; return 0; diff --git a/tools/perf/util/vdso.c b/tools/perf/util/vdso.c index 35532dcbff74..1b6f8f6db7aa 100644 --- a/tools/perf/util/vdso.c +++ b/tools/perf/util/vdso.c @@ -148,7 +148,7 @@ static int machine__thread_dso_type_maps_cb(struct map *map, void *data) struct machine__thread_dso_type_maps_cb_args *args = data; struct dso *dso = map__dso(map); - if (!dso || dso->long_name[0] != '/') + if (!dso || dso__long_name(dso)[0] != '/') return 0; args->dso_type = dso__type(dso, args->machine); @@ -361,7 +361,7 @@ struct dso *machine__findnew_vdso(struct machine *machine, bool dso__is_vdso(struct dso *dso) { - return !strcmp(dso->short_name, DSO__NAME_VDSO) || - !strcmp(dso->short_name, DSO__NAME_VDSO32) || - !strcmp(dso->short_name, DSO__NAME_VDSOX32); + return !strcmp(dso__short_name(dso), DSO__NAME_VDSO) || + !strcmp(dso__short_name(dso), DSO__NAME_VDSO32) || + !strcmp(dso__short_name(dso), DSO__NAME_VDSOX32); } -- cgit v1.2.3 From 7fdc33f84261466591fc00e66168d9c809e7e4c0 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 6 May 2024 11:01:01 -0700 Subject: perf map: Add missing dso__put() in map__new() A dso__put() is needed for the dsos__find() when the map is created and a buildid is sought. Fixes: f649ed80f3cabbf1 ("perf dsos: Tidy reference counting and locking") Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Athira Rajeev Cc: Changbin Du Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Tiezhu Yang Link: https://lore.kernel.org/r/20240506180104.485674-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/map.c | 1 + 1 file changed, 1 insertion(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index 117c4bb78b35..e1d14936a60d 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -200,6 +200,7 @@ struct map *map__new(struct machine *machine, u64 start, u64 len, dso__set_build_id(dso, dso__bid(header_bid_dso)); dso__set_header_build_id(dso, 1); } + dso__put(header_bid_dso); } dso__put(dso); } -- cgit v1.2.3 From ee5061f82449f7ac867a39daae64c9f9681156e8 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 6 May 2024 11:01:02 -0700 Subject: perf symbol-elf: Ensure dso__put() in machine__process_ksymbol_register() The dso__put() after the map creation causes a use after put in dso__set_loaded(). To ensure there is a +1 reference count on both sides of the if-else, do a dso__get() on the found map's dso. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Athira Rajeev Cc: Changbin Du Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Tiezhu Yang Link: https://lore.kernel.org/r/20240506180104.485674-3-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index 0b8fb14f5ff6..a3ff2ab154bd 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -683,7 +683,7 @@ static int machine__process_ksymbol_register(struct machine *machine, struct perf_sample *sample __maybe_unused) { struct symbol *sym; - struct dso *dso; + struct dso *dso = NULL; struct map *map = maps__find(machine__kernel_maps(machine), event->ksymbol.addr); int err = 0; @@ -696,7 +696,6 @@ static int machine__process_ksymbol_register(struct machine *machine, } dso__set_kernel(dso, DSO_SPACE__KERNEL); map = map__new2(0, dso); - dso__put(dso); if (!map) { err = -ENOMEM; goto out; @@ -722,7 +721,7 @@ static int machine__process_ksymbol_register(struct machine *machine, dso__set_long_name(dso, "", false); } } else { - dso = map__dso(map); + dso = dso__get(map__dso(map)); } sym = symbol__new(map__map_ip(map, map__start(map)), @@ -735,6 +734,7 @@ static int machine__process_ksymbol_register(struct machine *machine, dso__insert_symbol(dso, sym); out: map__put(map); + dso__put(dso); return err; } -- cgit v1.2.3 From 23106e318888849e95babe4c2bdc8a1a948ac36d Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 6 May 2024 11:01:03 -0700 Subject: perf symbol-elf: dso__load_sym_internal() reference count fixes dso__load_sym_internal() passed curr_mapp as an out argument to dso__process_kernel_symbol(). The out argument was never used so remove it to simplify the reference counting logic. Simplify reference counting issues with curr_dso by ensuring the value it points to has a +1 reference count, and then putting as necessary. This avoids some reference counting games when the dso is created making the code more obviously correct with some possible introduced overhead due to the reference counting get/puts. This, however, silences reference count checking and we can always optimize from a seemingly correct point. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Athira Rajeev Cc: Changbin Du Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Tiezhu Yang Link: https://lore.kernel.org/r/20240506180104.485674-4-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol-elf.c | 51 ++++++++++++++++++++++---------------------- 1 file changed, 25 insertions(+), 26 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index 3be5e8d1e278..e398abfd13a0 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -1419,7 +1419,7 @@ void __weak arch__sym_update(struct symbol *s __maybe_unused, static int dso__process_kernel_symbol(struct dso *dso, struct map *map, GElf_Sym *sym, GElf_Shdr *shdr, struct maps *kmaps, struct kmap *kmap, - struct dso **curr_dsop, struct map **curr_mapp, + struct dso **curr_dsop, const char *section_name, bool adjust_kernel_syms, bool kmodule, bool *remap_kernel, u64 max_text_sh_offset) @@ -1470,8 +1470,8 @@ static int dso__process_kernel_symbol(struct dso *dso, struct map *map, map__set_pgoff(map, shdr->sh_offset); } - *curr_mapp = map; - *curr_dsop = dso; + dso__put(*curr_dsop); + *curr_dsop = dso__get(dso); return 0; } @@ -1484,8 +1484,8 @@ static int dso__process_kernel_symbol(struct dso *dso, struct map *map, */ if (kmodule && adjust_kernel_syms && is_exe_text(shdr->sh_flags) && shdr->sh_offset <= max_text_sh_offset) { - *curr_mapp = map; - *curr_dsop = dso; + dso__put(*curr_dsop); + *curr_dsop = dso__get(dso); return 0; } @@ -1507,10 +1507,10 @@ static int dso__process_kernel_symbol(struct dso *dso, struct map *map, dso__set_binary_type(curr_dso, dso__binary_type(dso)); dso__set_adjust_symbols(curr_dso, dso__adjust_symbols(dso)); curr_map = map__new2(start, curr_dso); - dso__put(curr_dso); - if (curr_map == NULL) + if (curr_map == NULL) { + dso__put(curr_dso); return -1; - + } if (dso__kernel(curr_dso)) map__kmap(curr_map)->kmaps = kmaps; @@ -1524,21 +1524,15 @@ static int dso__process_kernel_symbol(struct dso *dso, struct map *map, dso__set_symtab_type(curr_dso, dso__symtab_type(dso)); if (maps__insert(kmaps, curr_map)) return -1; - /* - * Add it before we drop the reference to curr_map, i.e. while - * we still are sure to have a reference to this DSO via - * *curr_map->dso. - */ dsos__add(&maps__machine(kmaps)->dsos, curr_dso); - /* kmaps already got it */ - map__put(curr_map); dso__set_loaded(curr_dso); - *curr_mapp = curr_map; + dso__put(*curr_dsop); *curr_dsop = curr_dso; } else { - *curr_dsop = map__dso(curr_map); - map__put(curr_map); + dso__put(*curr_dsop); + *curr_dsop = dso__get(map__dso(curr_map)); } + map__put(curr_map); return 0; } @@ -1549,11 +1543,9 @@ dso__load_sym_internal(struct dso *dso, struct map *map, struct symsrc *syms_ss, { struct kmap *kmap = dso__kernel(dso) ? map__kmap(map) : NULL; struct maps *kmaps = kmap ? map__kmaps(map) : NULL; - struct map *curr_map = map; - struct dso *curr_dso = dso; + struct dso *curr_dso = NULL; Elf_Data *symstrs, *secstrs, *secstrs_run, *secstrs_sym; uint32_t nr_syms; - int err = -1; uint32_t idx; GElf_Ehdr ehdr; GElf_Shdr shdr; @@ -1656,6 +1648,7 @@ dso__load_sym_internal(struct dso *dso, struct map *map, struct symsrc *syms_ss, if (kmodule && adjust_kernel_syms) max_text_sh_offset = max_text_section(runtime_ss->elf, &runtime_ss->ehdr); + curr_dso = dso__get(dso); elf_symtab__for_each_symbol(syms, nr_syms, idx, sym) { struct symbol *f; const char *elf_name = elf_sym__name(&sym, symstrs); @@ -1744,9 +1737,13 @@ dso__load_sym_internal(struct dso *dso, struct map *map, struct symsrc *syms_ss, --sym.st_value; if (dso__kernel(dso)) { - if (dso__process_kernel_symbol(dso, map, &sym, &shdr, kmaps, kmap, &curr_dso, &curr_map, - section_name, adjust_kernel_syms, kmodule, - &remap_kernel, max_text_sh_offset)) + if (dso__process_kernel_symbol(dso, map, &sym, &shdr, + kmaps, kmap, &curr_dso, + section_name, + adjust_kernel_syms, + kmodule, + &remap_kernel, + max_text_sh_offset)) goto out_elf_end; } else if ((used_opd && runtime_ss->adjust_symbols) || (!used_opd && syms_ss->adjust_symbols)) { @@ -1795,6 +1792,7 @@ dso__load_sym_internal(struct dso *dso, struct map *map, struct symsrc *syms_ss, __symbols__insert(dso__symbols(curr_dso), f, dso__kernel(dso)); nr++; } + dso__put(curr_dso); /* * For misannotated, zeroed, ASM function sizes. @@ -1810,9 +1808,10 @@ dso__load_sym_internal(struct dso *dso, struct map *map, struct symsrc *syms_ss, maps__fixup_end(kmaps); } } - err = nr; + return nr; out_elf_end: - return err; + dso__put(curr_dso); + return -1; } int dso__load_sym(struct dso *dso, struct map *map, struct symsrc *syms_ss, -- cgit v1.2.3 From 37862d6fdced70a72b5a06d8f3440f7c567d5272 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 6 May 2024 11:01:04 -0700 Subject: perf dso: Use container_of() to avoid a pointer in 'struct dso_data' The dso pointer in 'struct dso_data' is necessary for reference count checking to account for the dso_data forming a global list of open dso's with references to the dso. The dso pointer also allows for the indirection that reference count checking needs. Outside of reference count checking the indirection isn't needed and container_of() is more efficient and saves space. The reference count won't be increased by placing items onto the global list, matching how things were before the reference count checking change, but we assert the dso is in dsos holding it live (and that the set of open dsos is a subset of all dsos for the machine). Update the DSO data tests so that they use a dsos struct to make the invariant true. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Athira Rajeev Cc: Changbin Du Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Tiezhu Yang Link: https://lore.kernel.org/r/20240506180104.485674-5-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/tests/dso-data.c | 60 ++++++++++++++++++++++----------------------- tools/perf/util/dso.c | 16 +++++++++++- tools/perf/util/dso.h | 2 ++ 3 files changed, 46 insertions(+), 32 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/tests/dso-data.c b/tools/perf/tests/dso-data.c index fde4eca84b6f..5286ae8bd2d7 100644 --- a/tools/perf/tests/dso-data.c +++ b/tools/perf/tests/dso-data.c @@ -10,6 +10,7 @@ #include #include #include "dso.h" +#include "dsos.h" #include "machine.h" #include "symbol.h" #include "tests.h" @@ -123,9 +124,10 @@ static int test__dso_data(struct test_suite *test __maybe_unused, int subtest __ TEST_ASSERT_VAL("No test file", file); memset(&machine, 0, sizeof(machine)); + dsos__init(&machine.dsos); - dso = dso__new((const char *)file); - + dso = dso__new(file); + TEST_ASSERT_VAL("Failed to add dso", !dsos__add(&machine.dsos, dso)); TEST_ASSERT_VAL("Failed to access to dso", dso__data_fd(dso, &machine) >= 0); @@ -170,6 +172,7 @@ static int test__dso_data(struct test_suite *test __maybe_unused, int subtest __ } dso__put(dso); + dsos__exit(&machine.dsos); unlink(file); return 0; } @@ -199,41 +202,35 @@ static long open_files_cnt(void) return nr - 1; } -static struct dso **dsos; - -static int dsos__create(int cnt, int size) +static int dsos__create(int cnt, int size, struct dsos *dsos) { int i; - dsos = malloc(sizeof(*dsos) * cnt); - TEST_ASSERT_VAL("failed to alloc dsos array", dsos); + dsos__init(dsos); for (i = 0; i < cnt; i++) { - char *file; + struct dso *dso; + char *file = test_file(size); - file = test_file(size); TEST_ASSERT_VAL("failed to get dso file", file); - - dsos[i] = dso__new(file); - TEST_ASSERT_VAL("failed to get dso", dsos[i]); + dso = dso__new(file); + TEST_ASSERT_VAL("failed to get dso", dso); + TEST_ASSERT_VAL("failed to add dso", !dsos__add(dsos, dso)); + dso__put(dso); } return 0; } -static void dsos__delete(int cnt) +static void dsos__delete(struct dsos *dsos) { - int i; - - for (i = 0; i < cnt; i++) { - struct dso *dso = dsos[i]; + for (unsigned int i = 0; i < dsos->cnt; i++) { + struct dso *dso = dsos->dsos[i]; dso__data_close(dso); unlink(dso__name(dso)); - dso__put(dso); } - - free(dsos); + dsos__exit(dsos); } static int set_fd_limit(int n) @@ -267,10 +264,10 @@ static int test__dso_data_cache(struct test_suite *test __maybe_unused, int subt /* and this is now our dso open FDs limit */ dso_cnt = limit / 2; TEST_ASSERT_VAL("failed to create dsos\n", - !dsos__create(dso_cnt, TEST_FILE_SIZE)); + !dsos__create(dso_cnt, TEST_FILE_SIZE, &machine.dsos)); for (i = 0; i < (dso_cnt - 1); i++) { - struct dso *dso = dsos[i]; + struct dso *dso = machine.dsos.dsos[i]; /* * Open dsos via dso__data_fd(), it opens the data @@ -290,17 +287,17 @@ static int test__dso_data_cache(struct test_suite *test __maybe_unused, int subt } /* verify the first one is already open */ - TEST_ASSERT_VAL("dsos[0] is not open", dso__data(dsos[0])->fd != -1); + TEST_ASSERT_VAL("dsos[0] is not open", dso__data(machine.dsos.dsos[0])->fd != -1); /* open +1 dso to reach the allowed limit */ - fd = dso__data_fd(dsos[i], &machine); + fd = dso__data_fd(machine.dsos.dsos[i], &machine); TEST_ASSERT_VAL("failed to get fd", fd > 0); /* should force the first one to be closed */ - TEST_ASSERT_VAL("failed to close dsos[0]", dso__data(dsos[0])->fd == -1); + TEST_ASSERT_VAL("failed to close dsos[0]", dso__data(machine.dsos.dsos[0])->fd == -1); /* cleanup everything */ - dsos__delete(dso_cnt); + dsos__delete(&machine.dsos); /* Make sure we did not leak any file descriptor. */ nr_end = open_files_cnt(); @@ -325,9 +322,9 @@ static int test__dso_data_reopen(struct test_suite *test __maybe_unused, int sub long nr_end, nr = open_files_cnt(), lim = new_limit(3); int fd, fd_extra; -#define dso_0 (dsos[0]) -#define dso_1 (dsos[1]) -#define dso_2 (dsos[2]) +#define dso_0 (machine.dsos.dsos[0]) +#define dso_1 (machine.dsos.dsos[1]) +#define dso_2 (machine.dsos.dsos[2]) /* Rest the internal dso open counter limit. */ reset_fd_limit(); @@ -347,7 +344,8 @@ static int test__dso_data_reopen(struct test_suite *test __maybe_unused, int sub TEST_ASSERT_VAL("failed to set file limit", !set_fd_limit((lim))); - TEST_ASSERT_VAL("failed to create dsos\n", !dsos__create(3, TEST_FILE_SIZE)); + TEST_ASSERT_VAL("failed to create dsos\n", + !dsos__create(3, TEST_FILE_SIZE, &machine.dsos)); /* open dso_0 */ fd = dso__data_fd(dso_0, &machine); @@ -386,7 +384,7 @@ static int test__dso_data_reopen(struct test_suite *test __maybe_unused, int sub /* cleanup everything */ close(fd_extra); - dsos__delete(3); + dsos__delete(&machine.dsos); /* Make sure we did not leak any file descriptor. */ nr_end = open_files_cnt(); diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c index 27db65e96e04..dde706b71da7 100644 --- a/tools/perf/util/dso.c +++ b/tools/perf/util/dso.c @@ -497,14 +497,20 @@ static pthread_mutex_t dso__data_open_lock = PTHREAD_MUTEX_INITIALIZER; static void dso__list_add(struct dso *dso) { list_add_tail(&dso__data(dso)->open_entry, &dso__data_open); +#ifdef REFCNT_CHECKING dso__data(dso)->dso = dso__get(dso); +#endif + /* Assume the dso is part of dsos, hence the optional reference count above. */ + assert(dso__dsos(dso)); dso__data_open_cnt++; } static void dso__list_del(struct dso *dso) { list_del_init(&dso__data(dso)->open_entry); +#ifdef REFCNT_CHECKING dso__put(dso__data(dso)->dso); +#endif WARN_ONCE(dso__data_open_cnt <= 0, "DSO data fd counter out of bounds."); dso__data_open_cnt--; @@ -654,9 +660,15 @@ static void close_dso(struct dso *dso) static void close_first_dso(void) { struct dso_data *dso_data; + struct dso *dso; dso_data = list_first_entry(&dso__data_open, struct dso_data, open_entry); - close_dso(dso_data->dso); +#ifdef REFCNT_CHECKING + dso = dso_data->dso; +#else + dso = container_of(dso_data, struct dso, data); +#endif + close_dso(dso); } static rlim_t get_fd_limit(void) @@ -1449,7 +1461,9 @@ struct dso *dso__new_id(const char *name, struct dso_id *id) data->fd = -1; data->status = DSO_DATA_STATUS_UNKNOWN; INIT_LIST_HEAD(&data->open_entry); +#ifdef REFCNT_CHECKING data->dso = NULL; /* Set when on the open_entry list. */ +#endif } return res; } diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h index f9689dd60de3..df2c98402af3 100644 --- a/tools/perf/util/dso.h +++ b/tools/perf/util/dso.h @@ -147,7 +147,9 @@ struct dso_cache { struct dso_data { struct rb_root cache; struct list_head open_entry; +#ifdef REFCNT_CHECKING struct dso *dso; +#endif int fd; int status; u32 status_seen; -- cgit v1.2.3 From 69fb6eab1969d09187feff14f370e01032054f1f Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 7 May 2024 00:04:06 -0300 Subject: perf annotate: Use zfree() to avoid possibly accessing dangling pointers When freeing a->b it is good practice to set a->b to NULL using zfree(&a->b) so that when we have a bug where a reference to a freed 'a' pointer is kept somewhere, we can more quickly cause a segfault if some code tries to use a->b. This is mostly done but some new cases were introduced recently, convert them to zfree(). Cc: Adrian Hunter Cc: Ian Rogers Cc: Jiri Olsa Cc: Kan Liang Cc: Namhyung Kim Link: https://lore.kernel.org/lkml/ZjmbHHrjIm5YRIBv@x1 Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browsers/annotate-data.c | 3 ++- tools/perf/util/annotate-data.c | 17 +++++++++-------- tools/perf/util/annotate.c | 4 ++-- 3 files changed, 13 insertions(+), 11 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/ui/browsers/annotate-data.c b/tools/perf/ui/browsers/annotate-data.c index a4a0f042f201..8d6bf08d371d 100644 --- a/tools/perf/ui/browsers/annotate-data.c +++ b/tools/perf/ui/browsers/annotate-data.c @@ -1,6 +1,7 @@ // SPDX-License-Identifier: GPL-2.0 #include #include +#include #include #include "ui/browser.h" @@ -130,7 +131,7 @@ static void annotated_data_browser__delete_entries(struct annotated_data_browser list_for_each_entry_safe(pos, tmp, &browser->entries, node) { list_del_init(&pos->node); - free(pos->hists); + zfree(&pos->hists); free(pos); } } diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index faefa444af1e..57e7d4b3550b 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -8,6 +8,7 @@ #include #include #include +#include #include "annotate.h" #include "annotate-data.h" @@ -311,8 +312,8 @@ static void delete_members(struct annotated_member *member) list_for_each_entry_safe(child, tmp, &member->children, node) { list_del(&child->node); delete_members(child); - free(child->type_name); - free(child->var_name); + zfree(&child->type_name); + zfree(&child->var_name); free(child); } } @@ -582,7 +583,7 @@ void global_var_type__tree_delete(struct rb_root *root) rb_erase(node, root); gvar = rb_entry(node, struct global_var_entry, node); - free(gvar->name); + zfree(&gvar->name); free(gvar); } } @@ -1817,16 +1818,16 @@ static int alloc_data_type_histograms(struct annotated_data_type *adt, int nr_en err: while (--i >= 0) - free(adt->histograms[i]); - free(adt->histograms); + zfree(&(adt->histograms[i])); + zfree(&adt->histograms); return -ENOMEM; } static void delete_data_type_histograms(struct annotated_data_type *adt) { for (int i = 0; i < adt->nr_histograms; i++) - free(adt->histograms[i]); - free(adt->histograms); + zfree(&(adt->histograms[i])); + zfree(&adt->histograms); } void annotated_data_type__tree_delete(struct rb_root *root) @@ -1840,7 +1841,7 @@ void annotated_data_type__tree_delete(struct rb_root *root) pos = rb_entry(node, struct annotated_data_type, node); delete_members(&pos->self); delete_data_type_histograms(pos); - free(pos->self.type_name); + zfree(&pos->self.type_name); free(pos); } } diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index d7d55263fc91..2b178835c1f3 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -2618,13 +2618,13 @@ static void delete_basic_blocks(struct basic_block_data *bb_data) list_for_each_entry_safe(link, tmp, &bb_data->queue, node) { list_del(&link->node); - free(link->bb); + zfree(&link->bb); free(link); } list_for_each_entry_safe(link, tmp, &bb_data->visited, node) { list_del(&link->node); - free(link->bb); + zfree(&link->bb); free(link); } } -- cgit v1.2.3 From 54ef362e4daa4a4ecfa2abdc251b21564d27784e Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 7 May 2024 00:04:06 -0300 Subject: perf callchain: Use zfree() to avoid possibly accessing dangling pointers When freeing a->b it is good practice to set a->b to NULL using zfree(&a->b) so that when we have a bug where a reference to a freed 'a' pointer is kept somewhere, we can more quickly cause a segfault if some code tries to use a->b. Convert one such case in the callchain code. Cc: Adrian Hunter Cc: Ian Rogers Cc: Jiri Olsa Cc: Kan Liang Cc: Namhyung Kim Link: https://lore.kernel.org/lkml/ZjmcGobQ8E52EyjJ@x1 Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/callchain.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index 68feed871809..1730b852a947 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -606,7 +606,7 @@ fill_node(struct callchain_node *node, struct callchain_cursor *cursor) call->brtype_stat = zalloc(sizeof(*call->brtype_stat)); if (!call->brtype_stat) { perror("not enough memory for the code path branch statistics"); - free(call->brtype_stat); + zfree(&call->brtype_stat); return -ENOMEM; } } -- cgit v1.2.3 From 36e8aa90fd6c577f783d5d8b02fbc205bb8e7f86 Mon Sep 17 00:00:00 2001 From: Athira Rajeev Date: Mon, 6 May 2024 17:49:00 +0530 Subject: perf annotate: Fix a comment about multi_regs in extract_reg_offset function Fix a comment in function which explains how multi_regs field gets set for an instruction. In the example, "mov %rsi, 8(%rbx,%rcx,4)", the comment mistakenly referred to "dst_multi_regs = 0". Correct it to use "src_multi_regs = 0" Signed-off-by: Athira Rajeev Acked-by: Namhyung Kim Cc: Adrian Hunter Cc: Akanksha J N Cc: Christophe Leroy Cc: Disha Goel Cc: Ian Rogers Cc: Jiri Olsa Cc: Kajol Jain Cc: Madhavan Srinivasan Cc: Segher Boessenkool Link: https://lore.kernel.org/r/20240506121906.76639-4-atrajeev@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 2b178835c1f3..0ffae315ca08 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -2092,7 +2092,7 @@ static int extract_reg_offset(struct arch *arch, const char *str, * mov 0x18, %r8 # src_reg1 = -1, src_mem = 0 * # dst_reg1 = r8, dst_mem = 0 * - * mov %rsi, 8(%rbx,%rcx,4) # src_reg1 = rsi, src_mem = 0, dst_multi_regs = 0 + * mov %rsi, 8(%rbx,%rcx,4) # src_reg1 = rsi, src_mem = 0, src_multi_regs = 0 * # dst_reg1 = rbx, dst_reg2 = rcx, dst_mem = 1 * # dst_multi_regs = 1, dst_offset = 8 */ -- cgit v1.2.3 From 0d2e3f251149b758458586df61578827eef7dc8d Mon Sep 17 00:00:00 2001 From: James Clark Date: Mon, 29 Apr 2024 16:21:46 +0100 Subject: perf cs-etm: Print error for new PERF_RECORD_AUX_OUTPUT_HW_ID versions The likely fix for this is to update perf so print a helpful message. Signed-off-by: James Clark Tested-by: Ganapatrao Kulkarni Acked-by: Anshuman Khandual Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Alexandre Torgue Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: John Garry Cc: Leo Yan Cc: Mark Rutland Cc: Maxime Coquelin Cc: Mike Leach Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Steve Clevenger Cc: Suzuki Poulouse Cc: Will Deacon Link: https://lore.kernel.org/r/20240429152207.479221-2-james.clark@arm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/cs-etm.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/cs-etm.c b/tools/perf/util/cs-etm.c index d65d7485886c..32818bd7cd17 100644 --- a/tools/perf/util/cs-etm.c +++ b/tools/perf/util/cs-etm.c @@ -335,8 +335,11 @@ static int cs_etm__process_aux_output_hw_id(struct perf_session *session, trace_chan_id = FIELD_GET(CS_AUX_HW_ID_TRACE_ID_MASK, hw_id); /* check that we can handle this version */ - if (version > CS_AUX_HW_ID_CURR_VERSION) + if (version > CS_AUX_HW_ID_CURR_VERSION) { + pr_err("CS ETM Trace: PERF_RECORD_AUX_OUTPUT_HW_ID version %d not supported. Please update Perf.\n", + version); return -EINVAL; + } /* get access to the etm metadata */ etm = container_of(session->auxtrace, struct cs_etm_auxtrace, auxtrace); -- cgit v1.2.3 From ee73fe99f77b066afc5035727130a6b8016e64b6 Mon Sep 17 00:00:00 2001 From: James Clark Date: Mon, 29 Apr 2024 16:21:47 +0100 Subject: perf auxtrace: Allow number of queues to be specified Currently it's only possible to initialize with the default number of queues and then use auxtrace_queues__add_event() to grow the array. But that's problematic if you don't have a real event to pass into that function yet. The queues hold a void *priv member to store custom state, and for Coresight we want to create decoders upfront before receiving data, so add a new function that allows pre-allocating queues. One reason to do this is because we might need to store metadata (HW_ID events) that effects other queues, but never actually receive auxtrace data on that queue. Reviewed-by: Anshuman Khandual Signed-off-by: James Clark Tested-by: Ganapatrao Kulkarni Acked-by: Adrian Hunter Cc: Alexander Shishkin Cc: Alexandre Torgue Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: John Garry Cc: Leo Yan Cc: Mark Rutland Cc: Maxime Coquelin Cc: Mike Leach Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Steve Clevenger Cc: Suzuki Poulouse Cc: Will Deacon Link: https://lore.kernel.org/r/20240429152207.479221-3-james.clark@arm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/auxtrace.c | 9 +++++++-- tools/perf/util/auxtrace.h | 1 + 2 files changed, 8 insertions(+), 2 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/auxtrace.c b/tools/perf/util/auxtrace.c index 3fd350de0051..e2f317063eec 100644 --- a/tools/perf/util/auxtrace.c +++ b/tools/perf/util/auxtrace.c @@ -218,15 +218,20 @@ static struct auxtrace_queue *auxtrace_alloc_queue_array(unsigned int nr_queues) return queue_array; } -int auxtrace_queues__init(struct auxtrace_queues *queues) +int auxtrace_queues__init_nr(struct auxtrace_queues *queues, int nr_queues) { - queues->nr_queues = AUXTRACE_INIT_NR_QUEUES; + queues->nr_queues = nr_queues; queues->queue_array = auxtrace_alloc_queue_array(queues->nr_queues); if (!queues->queue_array) return -ENOMEM; return 0; } +int auxtrace_queues__init(struct auxtrace_queues *queues) +{ + return auxtrace_queues__init_nr(queues, AUXTRACE_INIT_NR_QUEUES); +} + static int auxtrace_queues__grow(struct auxtrace_queues *queues, unsigned int new_nr_queues) { diff --git a/tools/perf/util/auxtrace.h b/tools/perf/util/auxtrace.h index 55702215a82d..8a6ec9565835 100644 --- a/tools/perf/util/auxtrace.h +++ b/tools/perf/util/auxtrace.h @@ -521,6 +521,7 @@ int auxtrace_mmap__read_snapshot(struct mmap *map, struct perf_tool *tool, process_auxtrace_t fn, size_t snapshot_size); +int auxtrace_queues__init_nr(struct auxtrace_queues *queues, int nr_queues); int auxtrace_queues__init(struct auxtrace_queues *queues); int auxtrace_queues__add_event(struct auxtrace_queues *queues, struct perf_session *session, -- cgit v1.2.3 From a3f7768bcf48281df14d98715f076c5656571527 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 7 May 2024 11:35:39 -0700 Subject: perf annotate: Fix memory leak in annotated_source Freeing hash map doesn't free the entries added to the hashmap, add the missing free(). Fixes: d3e7cad6f36d9e80 ("perf annotate: Add a hashmap for symbol histogram") Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Ben Gainey Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Kajol Jain Cc: Kan Liang Cc: K Prateek Nayak Cc: Li Dong Cc: Mark Rutland Cc: Namhyung Kim Cc: Oliver Upton Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Sun Haiyong Cc: Tim Chen Cc: Yanteng Si Cc: Yicong Yang Link: https://lore.kernel.org/r/20240507183545.1236093-3-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 0ffae315ca08..541988cf6e19 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -107,9 +107,15 @@ static struct annotated_source *annotated_source__new(void) static __maybe_unused void annotated_source__delete(struct annotated_source *src) { + struct hashmap_entry *cur; + size_t bkt; + if (src == NULL) return; + hashmap__for_each_entry(src->samples, cur, bkt) + zfree(&cur->pvalue); + hashmap__free(src->samples); zfree(&src->histograms); free(src); -- cgit v1.2.3 From 557b32c343925dfac480ef331494437ea223932d Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 7 May 2024 11:35:40 -0700 Subject: perf block-info: Remove unused refcount block_info__get() has no callers so the refcount is only ever one. As such remove the reference counting logic and turn puts to deletes. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Ben Gainey Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kajol Jain Cc: Kan Liang Cc: Li Dong Cc: Mark Rutland Cc: Namhyung Kim Cc: Oliver Upton Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Sun Haiyong Cc: Tim Chen Cc: Yanteng Si Cc: Yicong Yang Link: https://lore.kernel.org/r/20240507183545.1236093-4-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/block-info.c | 22 +++++----------------- tools/perf/util/block-info.h | 15 +-------------- tools/perf/util/hist.c | 4 ++-- 3 files changed, 8 insertions(+), 33 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/block-info.c b/tools/perf/util/block-info.c index 895ee8adf3b3..04068d48683f 100644 --- a/tools/perf/util/block-info.c +++ b/tools/perf/util/block-info.c @@ -43,26 +43,14 @@ static struct block_header_column { } }; -struct block_info *block_info__get(struct block_info *bi) -{ - if (bi) - refcount_inc(&bi->refcnt); - return bi; -} - -void block_info__put(struct block_info *bi) +struct block_info *block_info__new(void) { - if (bi && refcount_dec_and_test(&bi->refcnt)) - free(bi); + return zalloc(sizeof(struct block_info)); } -struct block_info *block_info__new(void) +void block_info__delete(struct block_info *bi) { - struct block_info *bi = zalloc(sizeof(*bi)); - - if (bi) - refcount_set(&bi->refcnt, 1); - return bi; + free(bi); } int64_t __block_info__cmp(struct hist_entry *left, struct hist_entry *right) @@ -148,7 +136,7 @@ int block_info__process_sym(struct hist_entry *he, struct block_hist *bh, he_block = hists__add_entry_block(&bh->block_hists, &al, bi); if (!he_block) { - block_info__put(bi); + block_info__delete(bi); return -1; } } diff --git a/tools/perf/util/block-info.h b/tools/perf/util/block-info.h index 96f53e89795e..0b9e1aad4c55 100644 --- a/tools/perf/util/block-info.h +++ b/tools/perf/util/block-info.h @@ -3,7 +3,6 @@ #define __PERF_BLOCK_H #include -#include #include "hist.h" #include "symbol.h" #include "sort.h" @@ -19,7 +18,6 @@ struct block_info { u64 total_cycles; int num; int num_aggr; - refcount_t refcnt; }; struct block_fmt { @@ -48,19 +46,8 @@ struct block_report { int nr_fmts; }; -struct block_hist; - struct block_info *block_info__new(void); -struct block_info *block_info__get(struct block_info *bi); -void block_info__put(struct block_info *bi); - -static inline void __block_info__zput(struct block_info **bi) -{ - block_info__put(*bi); - *bi = NULL; -} - -#define block_info__zput(bi) __block_info__zput(&bi) +void block_info__delete(struct block_info *bi); int64_t __block_info__cmp(struct hist_entry *left, struct hist_entry *right); diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c index 55ea6afcc437..b8a508cd0b14 100644 --- a/tools/perf/util/hist.c +++ b/tools/perf/util/hist.c @@ -631,7 +631,7 @@ static struct hist_entry *hists__findnew_entry(struct hists *hists, */ mem_info__zput(entry->mem_info); - block_info__zput(entry->block_info); + block_info__delete(entry->block_info); kvm_info__zput(entry->kvm_info); @@ -1341,7 +1341,7 @@ void hist_entry__delete(struct hist_entry *he) } if (he->block_info) - block_info__zput(he->block_info); + block_info__delete(he->block_info); if (he->kvm_info) kvm_info__zput(he->kvm_info); -- cgit v1.2.3 From a8cd4766d9128b897af6d4e0d22604f3bdaf2f82 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 7 May 2024 11:35:41 -0700 Subject: perf cpumap: Remove refcnt from 'struct cpu_aggr_map' It is assigned a value of 1 and never incremented. Remove and replace puts with delete. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Ben Gainey Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kajol Jain Cc: Kan Liang Cc: Li Dong Cc: Mark Rutland Cc: Namhyung Kim Cc: Oliver Upton Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Sun Haiyong Cc: Tim Chen Cc: Yanteng Si Cc: Yicong Yang Link: https://lore.kernel.org/r/20240507183545.1236093-5-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-stat.c | 16 +++------------- tools/perf/util/cpumap.c | 2 -- tools/perf/util/cpumap.h | 2 -- 3 files changed, 3 insertions(+), 17 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index 65a3dd7ffac3..35f79b48e8dc 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -1631,23 +1631,13 @@ static int perf_stat_init_aggr_mode(void) static void cpu_aggr_map__delete(struct cpu_aggr_map *map) { - if (map) { - WARN_ONCE(refcount_read(&map->refcnt) != 0, - "cpu_aggr_map refcnt unbalanced\n"); - free(map); - } -} - -static void cpu_aggr_map__put(struct cpu_aggr_map *map) -{ - if (map && refcount_dec_and_test(&map->refcnt)) - cpu_aggr_map__delete(map); + free(map); } static void perf_stat__exit_aggr_mode(void) { - cpu_aggr_map__put(stat_config.aggr_map); - cpu_aggr_map__put(stat_config.cpus_aggr_map); + cpu_aggr_map__delete(stat_config.aggr_map); + cpu_aggr_map__delete(stat_config.cpus_aggr_map); stat_config.aggr_map = NULL; stat_config.cpus_aggr_map = NULL; } diff --git a/tools/perf/util/cpumap.c b/tools/perf/util/cpumap.c index 6a270d640acb..27094211edd8 100644 --- a/tools/perf/util/cpumap.c +++ b/tools/perf/util/cpumap.c @@ -180,8 +180,6 @@ struct cpu_aggr_map *cpu_aggr_map__empty_new(int nr) cpus->nr = nr; for (i = 0; i < nr; i++) cpus->map[i] = aggr_cpu_id__empty(); - - refcount_set(&cpus->refcnt, 1); } return cpus; diff --git a/tools/perf/util/cpumap.h b/tools/perf/util/cpumap.h index 26cf76c693f5..ee0f6139b04a 100644 --- a/tools/perf/util/cpumap.h +++ b/tools/perf/util/cpumap.h @@ -5,7 +5,6 @@ #include #include #include -#include /** Identify where counts are aggregated, -1 implies not to aggregate. */ struct aggr_cpu_id { @@ -37,7 +36,6 @@ struct aggr_cpu_id { /** A collection of aggr_cpu_id values, the "built" version is sorted and uniqued. */ struct cpu_aggr_map { - refcount_t refcnt; /** Number of valid entries. */ int nr; /** The entries. */ -- cgit v1.2.3 From 13ca628716c6f2c3c1c477b007582ad9cfb8eba5 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 7 May 2024 11:35:42 -0700 Subject: perf comm: Add reference count checking to 'struct comm_str' Reference count checking of an rbtree is troublesome as each pointer should have a reference, switch to using a sorted array. Remove an indirection by embedding the reference count with the string. Use pthread_once to safely initialize the comm_strs and reader writer mutex. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Ben Gainey Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kajol Jain Cc: Kan Liang Cc: Li Dong Cc: Mark Rutland Cc: Namhyung Kim Cc: Oliver Upton Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Sun Haiyong Cc: Tim Chen Cc: Yanteng Si Cc: Yicong Yang Link: https://lore.kernel.org/r/20240507183545.1236093-6-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/comm.c | 196 +++++++++++++++++++++++++++++++------------------ 1 file changed, 126 insertions(+), 70 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/comm.c b/tools/perf/util/comm.c index afb8d4fd2644..1aa9a08e5b03 100644 --- a/tools/perf/util/comm.c +++ b/tools/perf/util/comm.c @@ -1,108 +1,164 @@ // SPDX-License-Identifier: GPL-2.0 #include "comm.h" #include -#include -#include #include +#include #include -#include #include #include "rwsem.h" -struct comm_str { - char *str; - struct rb_node rb_node; +DECLARE_RC_STRUCT(comm_str) { refcount_t refcnt; + char str[]; }; -/* Should perhaps be moved to struct machine */ -static struct rb_root comm_str_root; -static struct rw_semaphore comm_str_lock = {.lock = PTHREAD_RWLOCK_INITIALIZER,}; +static struct comm_strs { + struct rw_semaphore lock; + struct comm_str **strs; + int num_strs; + int capacity; +} _comm_strs; + +static void comm_strs__init(void) +{ + init_rwsem(&_comm_strs.lock); + _comm_strs.capacity = 16; + _comm_strs.num_strs = 0; + _comm_strs.strs = calloc(16, sizeof(*_comm_strs.strs)); +} + +static struct comm_strs *comm_strs__get(void) +{ + static pthread_once_t comm_strs_type_once = PTHREAD_ONCE_INIT; + + pthread_once(&comm_strs_type_once, comm_strs__init); + + return &_comm_strs; +} + +static refcount_t *comm_str__refcnt(struct comm_str *cs) +{ + return &RC_CHK_ACCESS(cs)->refcnt; +} + +static const char *comm_str__str(const struct comm_str *cs) +{ + return &RC_CHK_ACCESS(cs)->str[0]; +} static struct comm_str *comm_str__get(struct comm_str *cs) { - if (cs && refcount_inc_not_zero(&cs->refcnt)) - return cs; + struct comm_str *result; + + if (RC_CHK_GET(result, cs)) + refcount_inc_not_zero(comm_str__refcnt(cs)); - return NULL; + return result; } static void comm_str__put(struct comm_str *cs) { - if (cs && refcount_dec_and_test(&cs->refcnt)) { - down_write(&comm_str_lock); - rb_erase(&cs->rb_node, &comm_str_root); - up_write(&comm_str_lock); - zfree(&cs->str); - free(cs); + if (cs && refcount_dec_and_test(comm_str__refcnt(cs))) { + struct comm_strs *comm_strs = comm_strs__get(); + int i; + + down_write(&comm_strs->lock); + for (i = 0; i < comm_strs->num_strs; i++) { + if (comm_strs->strs[i] == cs) + break; + } + for (; i < comm_strs->num_strs - 1; i++) + comm_strs->strs[i] = comm_strs->strs[i + 1]; + + comm_strs->num_strs--; + up_write(&comm_strs->lock); + RC_CHK_FREE(cs); + } else { + RC_CHK_PUT(cs); } } -static struct comm_str *comm_str__alloc(const char *str) +static struct comm_str *comm_str__new(const char *str) { - struct comm_str *cs; + struct comm_str *result = NULL; + RC_STRUCT(comm_str) *cs; - cs = zalloc(sizeof(*cs)); - if (!cs) - return NULL; - - cs->str = strdup(str); - if (!cs->str) { - free(cs); - return NULL; + cs = malloc(sizeof(*cs) + strlen(str) + 1); + if (ADD_RC_CHK(result, cs)) { + refcount_set(comm_str__refcnt(result), 1); + strcpy(&cs->str[0], str); } + return result; +} - refcount_set(&cs->refcnt, 1); +static int comm_str__cmp(const void *_lhs, const void *_rhs) +{ + const struct comm_str *lhs = *(const struct comm_str * const *)_lhs; + const struct comm_str *rhs = *(const struct comm_str * const *)_rhs; - return cs; + return strcmp(comm_str__str(lhs), comm_str__str(rhs)); } -static -struct comm_str *__comm_str__findnew(const char *str, struct rb_root *root) +static int comm_str__search(const void *_key, const void *_member) { - struct rb_node **p = &root->rb_node; - struct rb_node *parent = NULL; - struct comm_str *iter, *new; - int cmp; - - while (*p != NULL) { - parent = *p; - iter = rb_entry(parent, struct comm_str, rb_node); - - /* - * If we race with comm_str__put, iter->refcnt is 0 - * and it will be removed within comm_str__put call - * shortly, ignore it in this search. - */ - cmp = strcmp(str, iter->str); - if (!cmp && comm_str__get(iter)) - return iter; - - if (cmp < 0) - p = &(*p)->rb_left; - else - p = &(*p)->rb_right; - } + const char *key = _key; + const struct comm_str *member = *(const struct comm_str * const *)_member; - new = comm_str__alloc(str); - if (!new) - return NULL; + return strcmp(key, comm_str__str(member)); +} + +static struct comm_str *__comm_strs__find(struct comm_strs *comm_strs, const char *str) +{ + struct comm_str **result; - rb_link_node(&new->rb_node, parent, p); - rb_insert_color(&new->rb_node, root); + result = bsearch(str, comm_strs->strs, comm_strs->num_strs, sizeof(struct comm_str *), + comm_str__search); - return new; + if (!result) + return NULL; + + return comm_str__get(*result); } -static struct comm_str *comm_str__findnew(const char *str, struct rb_root *root) +static struct comm_str *comm_strs__findnew(const char *str) { - struct comm_str *cs; + struct comm_strs *comm_strs = comm_strs__get(); + struct comm_str *result; - down_write(&comm_str_lock); - cs = __comm_str__findnew(str, root); - up_write(&comm_str_lock); + if (!comm_strs) + return NULL; - return cs; + down_read(&comm_strs->lock); + result = __comm_strs__find(comm_strs, str); + up_read(&comm_strs->lock); + if (result) + return result; + + down_write(&comm_strs->lock); + result = __comm_strs__find(comm_strs, str); + if (!result) { + if (comm_strs->num_strs == comm_strs->capacity) { + struct comm_str **tmp; + + tmp = reallocarray(comm_strs->strs, + comm_strs->capacity + 16, + sizeof(*comm_strs->strs)); + if (!tmp) { + up_write(&comm_strs->lock); + return NULL; + } + comm_strs->strs = tmp; + comm_strs->capacity += 16; + } + result = comm_str__new(str); + if (result) { + comm_strs->strs[comm_strs->num_strs++] = result; + qsort(comm_strs->strs, comm_strs->num_strs, sizeof(struct comm_str *), + comm_str__cmp); + } + } + up_write(&comm_strs->lock); + return result; } struct comm *comm__new(const char *str, u64 timestamp, bool exec) @@ -115,7 +171,7 @@ struct comm *comm__new(const char *str, u64 timestamp, bool exec) comm->start = timestamp; comm->exec = exec; - comm->comm_str = comm_str__findnew(str, &comm_str_root); + comm->comm_str = comm_strs__findnew(str); if (!comm->comm_str) { free(comm); return NULL; @@ -128,7 +184,7 @@ int comm__override(struct comm *comm, const char *str, u64 timestamp, bool exec) { struct comm_str *new, *old = comm->comm_str; - new = comm_str__findnew(str, &comm_str_root); + new = comm_strs__findnew(str); if (!new) return -ENOMEM; @@ -149,5 +205,5 @@ void comm__free(struct comm *comm) const char *comm__str(const struct comm *comm) { - return comm->comm_str->str; + return comm_str__str(comm->comm_str); } -- cgit v1.2.3 From ad3003a65a3ce1abf3b30af265bb36e23224a7aa Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 7 May 2024 11:35:43 -0700 Subject: perf mem-info: Move mem-info out of mem-events and symbol Move mem-info to its own header rather than having it split between mem-events and symbol. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Ben Gainey Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kajol Jain Cc: Kan Liang Cc: Li Dong Cc: Mark Rutland Cc: Namhyung Kim Cc: Oliver Upton Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Sun Haiyong Cc: Tim Chen Cc: Yanteng Si Cc: Yicong Yang Link: https://lore.kernel.org/r/20240507183545.1236093-7-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-c2c.c | 1 + tools/perf/builtin-report.c | 1 + tools/perf/builtin-script.c | 1 + tools/perf/tests/mem.c | 1 + tools/perf/util/Build | 1 + tools/perf/util/hist.c | 1 + tools/perf/util/machine.c | 1 + tools/perf/util/mem-events.c | 16 ++++++------ tools/perf/util/mem-events.h | 29 ++++++++-------------- tools/perf/util/mem-info.c | 28 +++++++++++++++++++++ tools/perf/util/mem-info.h | 28 +++++++++++++++++++++ .../util/scripting-engines/trace-event-python.c | 1 + tools/perf/util/sort.c | 1 + tools/perf/util/symbol.c | 26 +------------------ tools/perf/util/symbol.h | 12 --------- 15 files changed, 85 insertions(+), 63 deletions(-) create mode 100644 tools/perf/util/mem-info.c create mode 100644 tools/perf/util/mem-info.h (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-c2c.c b/tools/perf/builtin-c2c.c index 1f1d17df9b9a..c624414b2285 100644 --- a/tools/perf/builtin-c2c.c +++ b/tools/perf/builtin-c2c.c @@ -38,6 +38,7 @@ #include "ui/browsers/hists.h" #include "thread.h" #include "mem2node.h" +#include "mem-info.h" #include "symbol.h" #include "ui/ui.h" #include "ui/progress.h" diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index b5525f4f7090..6eb1d90b589f 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -31,6 +31,7 @@ #include "util/evsel.h" #include "util/evswitch.h" #include "util/header.h" +#include "util/mem-info.h" #include "util/session.h" #include "util/srcline.h" #include "util/tool.h" diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index f7c3c3868c3c..58af4f3aa592 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -32,6 +32,7 @@ #include "util/time-utils.h" #include "util/path.h" #include "util/event.h" +#include "util/mem-info.h" #include "ui/ui.h" #include "print_binary.h" #include "print_insn.h" diff --git a/tools/perf/tests/mem.c b/tools/perf/tests/mem.c index 56014ec7d49d..f694a60d8a73 100644 --- a/tools/perf/tests/mem.c +++ b/tools/perf/tests/mem.c @@ -1,6 +1,7 @@ // SPDX-License-Identifier: GPL-2.0 #include "util/map_symbol.h" #include "util/mem-events.h" +#include "util/mem-info.h" #include "util/symbol.h" #include "linux/perf_event.h" #include "util/debug.h" diff --git a/tools/perf/util/Build b/tools/perf/util/Build index 292170a99ab6..da64efd8718f 100644 --- a/tools/perf/util/Build +++ b/tools/perf/util/Build @@ -141,6 +141,7 @@ perf-y += term.o perf-y += help-unknown-cmd.o perf-y += dlfilter.o perf-y += mem-events.o +perf-y += mem-info.o perf-y += vsprintf.o perf-y += units.o perf-y += time-utils.o diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c index b8a508cd0b14..56453a02cdf4 100644 --- a/tools/perf/util/hist.c +++ b/tools/perf/util/hist.c @@ -9,6 +9,7 @@ #include "map_symbol.h" #include "branch.h" #include "mem-events.h" +#include "mem-info.h" #include "session.h" #include "namespaces.h" #include "cgroup.h" diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index a3ff2ab154bd..d5a01ef92668 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -16,6 +16,7 @@ #include "map_symbol.h" #include "branch.h" #include "mem-events.h" +#include "mem-info.h" #include "path.h" #include "srcline.h" #include "symbol.h" diff --git a/tools/perf/util/mem-events.c b/tools/perf/util/mem-events.c index 637cbd4a7bfb..f618a5ccc7af 100644 --- a/tools/perf/util/mem-events.c +++ b/tools/perf/util/mem-events.c @@ -10,7 +10,9 @@ #include #include "map_symbol.h" #include "mem-events.h" +#include "mem-info.h" #include "debug.h" +#include "evsel.h" #include "symbol.h" #include "pmu.h" #include "pmus.h" @@ -281,7 +283,7 @@ static const char * const tlb_access[] = { "Fault", }; -int perf_mem__tlb_scnprintf(char *out, size_t sz, struct mem_info *mem_info) +int perf_mem__tlb_scnprintf(char *out, size_t sz, const struct mem_info *mem_info) { size_t l = 0, i; u64 m = PERF_MEM_TLB_NA; @@ -359,7 +361,7 @@ static const char * const mem_hops[] = { "board", }; -static int perf_mem__op_scnprintf(char *out, size_t sz, struct mem_info *mem_info) +static int perf_mem__op_scnprintf(char *out, size_t sz, const struct mem_info *mem_info) { u64 op = PERF_MEM_LOCK_NA; int l; @@ -383,7 +385,7 @@ static int perf_mem__op_scnprintf(char *out, size_t sz, struct mem_info *mem_inf return l; } -int perf_mem__lvl_scnprintf(char *out, size_t sz, struct mem_info *mem_info) +int perf_mem__lvl_scnprintf(char *out, size_t sz, const struct mem_info *mem_info) { union perf_mem_data_src data_src; int printed = 0; @@ -465,7 +467,7 @@ static const char * const snoopx_access[] = { "Peer", }; -int perf_mem__snp_scnprintf(char *out, size_t sz, struct mem_info *mem_info) +int perf_mem__snp_scnprintf(char *out, size_t sz, const struct mem_info *mem_info) { size_t i, l = 0; u64 m = PERF_MEM_SNOOP_NA; @@ -507,7 +509,7 @@ int perf_mem__snp_scnprintf(char *out, size_t sz, struct mem_info *mem_info) return l; } -int perf_mem__lck_scnprintf(char *out, size_t sz, struct mem_info *mem_info) +int perf_mem__lck_scnprintf(char *out, size_t sz, const struct mem_info *mem_info) { u64 mask = PERF_MEM_LOCK_NA; int l; @@ -525,7 +527,7 @@ int perf_mem__lck_scnprintf(char *out, size_t sz, struct mem_info *mem_info) return l; } -int perf_mem__blk_scnprintf(char *out, size_t sz, struct mem_info *mem_info) +int perf_mem__blk_scnprintf(char *out, size_t sz, const struct mem_info *mem_info) { size_t l = 0; u64 mask = PERF_MEM_BLK_NA; @@ -548,7 +550,7 @@ int perf_mem__blk_scnprintf(char *out, size_t sz, struct mem_info *mem_info) return l; } -int perf_script__meminfo_scnprintf(char *out, size_t sz, struct mem_info *mem_info) +int perf_script__meminfo_scnprintf(char *out, size_t sz, const struct mem_info *mem_info) { int i = 0; diff --git a/tools/perf/util/mem-events.h b/tools/perf/util/mem-events.h index 15d5f0320d27..ca31014d7934 100644 --- a/tools/perf/util/mem-events.h +++ b/tools/perf/util/mem-events.h @@ -3,13 +3,7 @@ #define __PERF_MEM_EVENTS_H #include -#include -#include #include -#include -#include -#include "stat.h" -#include "evsel.h" struct perf_mem_event { bool record; @@ -21,13 +15,6 @@ struct perf_mem_event { const char *event_name; }; -struct mem_info { - struct addr_map_symbol iaddr; - struct addr_map_symbol daddr; - union perf_mem_data_src data_src; - refcount_t refcnt; -}; - enum { PERF_MEM_EVENTS__LOAD, PERF_MEM_EVENTS__STORE, @@ -35,6 +22,10 @@ enum { PERF_MEM_EVENTS__MAX, }; +struct evsel; +struct mem_info; +struct perf_pmu; + extern unsigned int perf_mem_events__loads_ldlat; extern struct perf_mem_event perf_mem_events[PERF_MEM_EVENTS__MAX]; @@ -49,13 +40,13 @@ bool is_mem_loads_aux_event(struct evsel *leader); void perf_pmu__mem_events_list(struct perf_pmu *pmu); int perf_mem_events__record_args(const char **rec_argv, int *argv_nr); -int perf_mem__tlb_scnprintf(char *out, size_t sz, struct mem_info *mem_info); -int perf_mem__lvl_scnprintf(char *out, size_t sz, struct mem_info *mem_info); -int perf_mem__snp_scnprintf(char *out, size_t sz, struct mem_info *mem_info); -int perf_mem__lck_scnprintf(char *out, size_t sz, struct mem_info *mem_info); -int perf_mem__blk_scnprintf(char *out, size_t sz, struct mem_info *mem_info); +int perf_mem__tlb_scnprintf(char *out, size_t sz, const struct mem_info *mem_info); +int perf_mem__lvl_scnprintf(char *out, size_t sz, const struct mem_info *mem_info); +int perf_mem__snp_scnprintf(char *out, size_t sz, const struct mem_info *mem_info); +int perf_mem__lck_scnprintf(char *out, size_t sz, const struct mem_info *mem_info); +int perf_mem__blk_scnprintf(char *out, size_t sz, const struct mem_info *mem_info); -int perf_script__meminfo_scnprintf(char *bf, size_t size, struct mem_info *mem_info); +int perf_script__meminfo_scnprintf(char *bf, size_t size, const struct mem_info *mem_info); struct c2c_stats { u32 nr_entries; diff --git a/tools/perf/util/mem-info.c b/tools/perf/util/mem-info.c new file mode 100644 index 000000000000..ff0dfdb5369a --- /dev/null +++ b/tools/perf/util/mem-info.c @@ -0,0 +1,28 @@ +// SPDX-License-Identifier: GPL-2.0 +#include +#include "mem-info.h" + +struct mem_info *mem_info__get(struct mem_info *mi) +{ + if (mi) + refcount_inc(&mi->refcnt); + return mi; +} + +void mem_info__put(struct mem_info *mi) +{ + if (mi && refcount_dec_and_test(&mi->refcnt)) { + addr_map_symbol__exit(&mi->iaddr); + addr_map_symbol__exit(&mi->daddr); + free(mi); + } +} + +struct mem_info *mem_info__new(void) +{ + struct mem_info *mi = zalloc(sizeof(*mi)); + + if (mi) + refcount_set(&mi->refcnt, 1); + return mi; +} diff --git a/tools/perf/util/mem-info.h b/tools/perf/util/mem-info.h new file mode 100644 index 000000000000..bb7d8375d73c --- /dev/null +++ b/tools/perf/util/mem-info.h @@ -0,0 +1,28 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef __PERF_MEM_INFO_H +#define __PERF_MEM_INFO_H + +#include +#include +#include "map_symbol.h" + +struct mem_info { + struct addr_map_symbol iaddr; + struct addr_map_symbol daddr; + union perf_mem_data_src data_src; + refcount_t refcnt; +}; + +struct mem_info *mem_info__new(void); +struct mem_info *mem_info__get(struct mem_info *mi); +void mem_info__put(struct mem_info *mi); + +static inline void __mem_info__zput(struct mem_info **mi) +{ + mem_info__put(*mi); + *mi = NULL; +} + +#define mem_info__zput(mi) __mem_info__zput(&mi) + +#endif /* __PERF_MEM_INFO_H */ diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c index c2caa5720299..c77fe08a0545 100644 --- a/tools/perf/util/scripting-engines/trace-event-python.c +++ b/tools/perf/util/scripting-engines/trace-event-python.c @@ -45,6 +45,7 @@ #include "../thread.h" #include "../comm.h" #include "../machine.h" +#include "../mem-info.h" #include "../db-export.h" #include "../thread-stack.h" #include "../trace-event.h" diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c index 704664e5b4ea..711ef69306f3 100644 --- a/tools/perf/util/sort.c +++ b/tools/perf/util/sort.c @@ -23,6 +23,7 @@ #include "strlist.h" #include "strbuf.h" #include "mem-events.h" +#include "mem-info.h" #include "annotate.h" #include "annotate-data.h" #include "event.h" diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 7772a4d3e66c..eb3319baa1b5 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -27,6 +27,7 @@ #include "symbol.h" #include "map_symbol.h" #include "mem-events.h" +#include "mem-info.h" #include "symsrc.h" #include "strlist.h" #include "intlist.h" @@ -2570,31 +2571,6 @@ int symbol__config_symfs(const struct option *opt __maybe_unused, return 0; } -struct mem_info *mem_info__get(struct mem_info *mi) -{ - if (mi) - refcount_inc(&mi->refcnt); - return mi; -} - -void mem_info__put(struct mem_info *mi) -{ - if (mi && refcount_dec_and_test(&mi->refcnt)) { - addr_map_symbol__exit(&mi->iaddr); - addr_map_symbol__exit(&mi->daddr); - free(mi); - } -} - -struct mem_info *mem_info__new(void) -{ - struct mem_info *mi = zalloc(sizeof(*mi)); - - if (mi) - refcount_set(&mi->refcnt, 1); - return mi; -} - /* * Checks that user supplied symbol kernel files are accessible because * the default mechanism for accessing elf files fails silently. i.e. if diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 071837ddce2a..3fb5d146d9b1 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -268,18 +268,6 @@ enum { SDT_NOTE_IDX_REFCTR, }; -struct mem_info *mem_info__new(void); -struct mem_info *mem_info__get(struct mem_info *mi); -void mem_info__put(struct mem_info *mi); - -static inline void __mem_info__zput(struct mem_info **mi) -{ - mem_info__put(*mi); - *mi = NULL; -} - -#define mem_info__zput(mi) __mem_info__zput(&mi) - int symbol__validate_sym_arguments(void); #endif /* __PERF_SYMBOL */ -- cgit v1.2.3 From 1a8c2e0177df250093b482b0c0034b53fdc5409f Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 7 May 2024 11:35:44 -0700 Subject: perf mem-info: Add reference count checking Add reference count checking and switch 'struct mem_info' usage to use accessor functions. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Ben Gainey Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kajol Jain Cc: Kan Liang Cc: Li Dong Cc: Mark Rutland Cc: Namhyung Kim Cc: Oliver Upton Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Sun Haiyong Cc: Tim Chen Cc: Yanteng Si Cc: Yicong Yang Link: https://lore.kernel.org/r/20240507183545.1236093-8-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-c2c.c | 12 ++-- tools/perf/builtin-report.c | 2 +- tools/perf/builtin-script.c | 11 +++- tools/perf/tests/mem.c | 10 ++-- tools/perf/util/hist.c | 26 +++++---- tools/perf/util/machine.c | 6 +- tools/perf/util/mem-events.c | 20 +++---- tools/perf/util/mem-info.c | 29 +++++---- tools/perf/util/mem-info.h | 28 ++++++++- .../util/scripting-engines/trace-event-python.c | 11 +++- tools/perf/util/sort.c | 68 +++++++++++----------- 11 files changed, 135 insertions(+), 88 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/builtin-c2c.c b/tools/perf/builtin-c2c.c index c624414b2285..c157bd31f2e5 100644 --- a/tools/perf/builtin-c2c.c +++ b/tools/perf/builtin-c2c.c @@ -530,7 +530,7 @@ static int dcacheline_entry(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp, char buf[20]; if (he->mem_info) - addr = cl_address(he->mem_info->daddr.addr, chk_double_cl); + addr = cl_address(mem_info__daddr(he->mem_info)->addr, chk_double_cl); return scnprintf(hpp->buf, hpp->size, "%*s", width, HEX_STR(buf, addr)); } @@ -568,7 +568,7 @@ static int offset_entry(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp, char buf[20]; if (he->mem_info) - addr = cl_offset(he->mem_info->daddr.al_addr, chk_double_cl); + addr = cl_offset(mem_info__daddr(he->mem_info)->al_addr, chk_double_cl); return scnprintf(hpp->buf, hpp->size, "%*s", width, HEX_STR(buf, addr)); } @@ -580,10 +580,10 @@ offset_cmp(struct perf_hpp_fmt *fmt __maybe_unused, uint64_t l = 0, r = 0; if (left->mem_info) - l = cl_offset(left->mem_info->daddr.addr, chk_double_cl); + l = cl_offset(mem_info__daddr(left->mem_info)->addr, chk_double_cl); if (right->mem_info) - r = cl_offset(right->mem_info->daddr.addr, chk_double_cl); + r = cl_offset(mem_info__daddr(right->mem_info)->addr, chk_double_cl); return (int64_t)(r - l); } @@ -597,7 +597,7 @@ iaddr_entry(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp, char buf[20]; if (he->mem_info) - addr = he->mem_info->iaddr.addr; + addr = mem_info__iaddr(he->mem_info)->addr; return scnprintf(hpp->buf, hpp->size, "%*s", width, HEX_STR(buf, addr)); } @@ -2593,7 +2593,7 @@ perf_c2c_cacheline_browser__title(struct hist_browser *browser, he = cl_browser->he; if (he->mem_info) - addr = cl_address(he->mem_info->daddr.addr, chk_double_cl); + addr = cl_address(mem_info__daddr(he->mem_info)->addr, chk_double_cl); scnprintf(bf, size, "Cacheline 0x%lx", addr); return 0; diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 6eb1d90b589f..0892b6e3e5e7 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -186,7 +186,7 @@ static int hist_iter__report_callback(struct hist_entry_iter *iter, } else if (rep->mem_mode) { mi = he->mem_info; - err = addr_map_symbol__inc_samples(&mi->daddr, sample, evsel); + err = addr_map_symbol__inc_samples(mem_info__daddr(mi), sample, evsel); if (err) goto out; diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index 58af4f3aa592..c16224b1fef3 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -2051,13 +2051,18 @@ static int evlist__max_name_len(struct evlist *evlist) static int data_src__fprintf(u64 data_src, FILE *fp) { - struct mem_info mi = { .data_src.val = data_src }; + struct mem_info *mi = mem_info__new(); char decode[100]; char out[100]; static int maxlen; int len; - perf_script__meminfo_scnprintf(decode, 100, &mi); + if (!mi) + return -ENOMEM; + + mem_info__data_src(mi)->val = data_src; + perf_script__meminfo_scnprintf(decode, 100, mi); + mem_info__put(mi); len = scnprintf(out, 100, "%16" PRIx64 " %s", data_src, decode); if (maxlen < len) @@ -2498,7 +2503,7 @@ static int process_attr(struct perf_tool *tool, union perf_event *event, evsel = evlist__last(*pevlist); if (!evsel->priv) { - if (scr->per_event_dump) { + if (scr->per_event_dump) { evsel->priv = evsel_script__new(evsel, scr->session->data); if (!evsel->priv) return -ENOMEM; diff --git a/tools/perf/tests/mem.c b/tools/perf/tests/mem.c index f694a60d8a73..cb3d749e157b 100644 --- a/tools/perf/tests/mem.c +++ b/tools/perf/tests/mem.c @@ -13,12 +13,14 @@ static int check(union perf_mem_data_src data_src, { char out[100]; char failure[100]; - struct mem_info mi = { .data_src = data_src }; - + struct mem_info *mi = mem_info__new(); int n; - n = perf_mem__snp_scnprintf(out, sizeof out, &mi); - n += perf_mem__lvl_scnprintf(out + n, sizeof out - n, &mi); + TEST_ASSERT_VAL("Memory allocation failed", mi); + *mem_info__data_src(mi) = data_src; + n = perf_mem__snp_scnprintf(out, sizeof out, mi); + n += perf_mem__lvl_scnprintf(out + n, sizeof out - n, mi); + mem_info__put(mi); scnprintf(failure, sizeof failure, "unexpected %s", out); TEST_ASSERT_VAL(failure, !strcmp(string, out)); return 0; diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c index 56453a02cdf4..00814d42d5f1 100644 --- a/tools/perf/util/hist.c +++ b/tools/perf/util/hist.c @@ -154,8 +154,8 @@ void hists__calc_col_len(struct hists *hists, struct hist_entry *h) } if (h->mem_info) { - if (h->mem_info->daddr.ms.sym) { - symlen = (int)h->mem_info->daddr.ms.sym->namelen + 4 + if (mem_info__daddr(h->mem_info)->ms.sym) { + symlen = (int)mem_info__daddr(h->mem_info)->ms.sym->namelen + 4 + unresolved_col_width + 2; hists__new_col_len(hists, HISTC_MEM_DADDR_SYMBOL, symlen); @@ -169,8 +169,8 @@ void hists__calc_col_len(struct hists *hists, struct hist_entry *h) symlen); } - if (h->mem_info->iaddr.ms.sym) { - symlen = (int)h->mem_info->iaddr.ms.sym->namelen + 4 + if (mem_info__iaddr(h->mem_info)->ms.sym) { + symlen = (int)mem_info__iaddr(h->mem_info)->ms.sym->namelen + 4 + unresolved_col_width + 2; hists__new_col_len(hists, HISTC_MEM_IADDR_SYMBOL, symlen); @@ -180,8 +180,8 @@ void hists__calc_col_len(struct hists *hists, struct hist_entry *h) symlen); } - if (h->mem_info->daddr.ms.map) { - symlen = dso__name_len(map__dso(h->mem_info->daddr.ms.map)); + if (mem_info__daddr(h->mem_info)->ms.map) { + symlen = dso__name_len(map__dso(mem_info__daddr(h->mem_info)->ms.map)); hists__new_col_len(hists, HISTC_MEM_DADDR_DSO, symlen); } else { @@ -477,8 +477,10 @@ static int hist_entry__init(struct hist_entry *he, } if (he->mem_info) { - he->mem_info->iaddr.ms.map = map__get(he->mem_info->iaddr.ms.map); - he->mem_info->daddr.ms.map = map__get(he->mem_info->daddr.ms.map); + mem_info__iaddr(he->mem_info)->ms.map = + map__get(mem_info__iaddr(he->mem_info)->ms.map); + mem_info__daddr(he->mem_info)->ms.map = + map__get(mem_info__daddr(he->mem_info)->ms.map); } if (hist_entry__has_callchains(he) && symbol_conf.use_callchain) @@ -526,8 +528,8 @@ err_infos: zfree(&he->branch_info); } if (he->mem_info) { - map_symbol__exit(&he->mem_info->iaddr.ms); - map_symbol__exit(&he->mem_info->daddr.ms); + map_symbol__exit(&mem_info__iaddr(he->mem_info)->ms); + map_symbol__exit(&mem_info__daddr(he->mem_info)->ms); } err: map_symbol__exit(&he->ms); @@ -1336,8 +1338,8 @@ void hist_entry__delete(struct hist_entry *he) } if (he->mem_info) { - map_symbol__exit(&he->mem_info->iaddr.ms); - map_symbol__exit(&he->mem_info->daddr.ms); + map_symbol__exit(&mem_info__iaddr(he->mem_info)->ms); + map_symbol__exit(&mem_info__daddr(he->mem_info)->ms); mem_info__zput(he->mem_info); } diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index d5a01ef92668..8477edefc299 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -2013,11 +2013,11 @@ struct mem_info *sample__resolve_mem(struct perf_sample *sample, if (!mi) return NULL; - ip__resolve_ams(al->thread, &mi->iaddr, sample->ip); - ip__resolve_data(al->thread, al->cpumode, &mi->daddr, + ip__resolve_ams(al->thread, mem_info__iaddr(mi), sample->ip); + ip__resolve_data(al->thread, al->cpumode, mem_info__daddr(mi), sample->addr, sample->phys_addr, sample->data_page_size); - mi->data_src.val = sample->data_src; + mem_info__data_src(mi)->val = sample->data_src; return mi; } diff --git a/tools/perf/util/mem-events.c b/tools/perf/util/mem-events.c index f618a5ccc7af..6dda47bb774f 100644 --- a/tools/perf/util/mem-events.c +++ b/tools/perf/util/mem-events.c @@ -293,7 +293,7 @@ int perf_mem__tlb_scnprintf(char *out, size_t sz, const struct mem_info *mem_inf out[0] = '\0'; if (mem_info) - m = mem_info->data_src.mem_dtlb; + m = mem_info__const_data_src(mem_info)->mem_dtlb; hit = m & PERF_MEM_TLB_HIT; miss = m & PERF_MEM_TLB_MISS; @@ -367,7 +367,7 @@ static int perf_mem__op_scnprintf(char *out, size_t sz, const struct mem_info *m int l; if (mem_info) - op = mem_info->data_src.mem_op; + op = mem_info__const_data_src(mem_info)->mem_op; if (op & PERF_MEM_OP_NA) l = scnprintf(out, sz, "N/A"); @@ -400,7 +400,7 @@ int perf_mem__lvl_scnprintf(char *out, size_t sz, const struct mem_info *mem_inf if (!mem_info) goto na; - data_src = mem_info->data_src; + data_src = *mem_info__const_data_src(mem_info); if (data_src.mem_lvl & PERF_MEM_LVL_HIT) memcpy(hit_miss, "hit", 3); @@ -476,7 +476,7 @@ int perf_mem__snp_scnprintf(char *out, size_t sz, const struct mem_info *mem_inf out[0] = '\0'; if (mem_info) - m = mem_info->data_src.mem_snoop; + m = mem_info__const_data_src(mem_info)->mem_snoop; for (i = 0; m && i < ARRAY_SIZE(snoop_access); i++, m >>= 1) { if (!(m & 0x1)) @@ -490,7 +490,7 @@ int perf_mem__snp_scnprintf(char *out, size_t sz, const struct mem_info *mem_inf m = 0; if (mem_info) - m = mem_info->data_src.mem_snoopx; + m = mem_info__const_data_src(mem_info)->mem_snoopx; for (i = 0; m && i < ARRAY_SIZE(snoopx_access); i++, m >>= 1) { if (!(m & 0x1)) @@ -515,7 +515,7 @@ int perf_mem__lck_scnprintf(char *out, size_t sz, const struct mem_info *mem_inf int l; if (mem_info) - mask = mem_info->data_src.mem_lock; + mask = mem_info__const_data_src(mem_info)->mem_lock; if (mask & PERF_MEM_LOCK_NA) l = scnprintf(out, sz, "N/A"); @@ -536,7 +536,7 @@ int perf_mem__blk_scnprintf(char *out, size_t sz, const struct mem_info *mem_inf out[0] = '\0'; if (mem_info) - mask = mem_info->data_src.mem_blk; + mask = mem_info__const_data_src(mem_info)->mem_blk; if (!mask || (mask & PERF_MEM_BLK_NA)) { l += scnprintf(out + l, sz - l, " N/A"); @@ -572,8 +572,8 @@ int perf_script__meminfo_scnprintf(char *out, size_t sz, const struct mem_info * int c2c_decode_stats(struct c2c_stats *stats, struct mem_info *mi) { - union perf_mem_data_src *data_src = &mi->data_src; - u64 daddr = mi->daddr.addr; + union perf_mem_data_src *data_src = mem_info__data_src(mi); + u64 daddr = mem_info__daddr(mi)->addr; u64 op = data_src->mem_op; u64 lvl = data_src->mem_lvl; u64 snoop = data_src->mem_snoop; @@ -700,7 +700,7 @@ do { \ return -1; } - if (!mi->daddr.ms.map || !mi->iaddr.ms.map) { + if (!mem_info__daddr(mi)->ms.map || !mem_info__iaddr(mi)->ms.map) { stats->nomap++; return -1; } diff --git a/tools/perf/util/mem-info.c b/tools/perf/util/mem-info.c index ff0dfdb5369a..27d67721a695 100644 --- a/tools/perf/util/mem-info.c +++ b/tools/perf/util/mem-info.c @@ -4,25 +4,32 @@ struct mem_info *mem_info__get(struct mem_info *mi) { - if (mi) - refcount_inc(&mi->refcnt); - return mi; + struct mem_info *result; + + if (RC_CHK_GET(result, mi)) + refcount_inc(mem_info__refcnt(mi)); + + return result; } void mem_info__put(struct mem_info *mi) { - if (mi && refcount_dec_and_test(&mi->refcnt)) { - addr_map_symbol__exit(&mi->iaddr); - addr_map_symbol__exit(&mi->daddr); - free(mi); + if (mi && refcount_dec_and_test(mem_info__refcnt(mi))) { + addr_map_symbol__exit(mem_info__iaddr(mi)); + addr_map_symbol__exit(mem_info__daddr(mi)); + RC_CHK_FREE(mi); + } else { + RC_CHK_PUT(mi); } } struct mem_info *mem_info__new(void) { - struct mem_info *mi = zalloc(sizeof(*mi)); + struct mem_info *result = NULL; + RC_STRUCT(mem_info) *mi = zalloc(sizeof(*mi)); + + if (ADD_RC_CHK(result, mi)) + refcount_set(mem_info__refcnt(result), 1); - if (mi) - refcount_set(&mi->refcnt, 1); - return mi; + return result; } diff --git a/tools/perf/util/mem-info.h b/tools/perf/util/mem-info.h index bb7d8375d73c..0f68e29f311b 100644 --- a/tools/perf/util/mem-info.h +++ b/tools/perf/util/mem-info.h @@ -4,9 +4,10 @@ #include #include +#include #include "map_symbol.h" -struct mem_info { +DECLARE_RC_STRUCT(mem_info) { struct addr_map_symbol iaddr; struct addr_map_symbol daddr; union perf_mem_data_src data_src; @@ -25,4 +26,29 @@ static inline void __mem_info__zput(struct mem_info **mi) #define mem_info__zput(mi) __mem_info__zput(&mi) +static inline struct addr_map_symbol *mem_info__iaddr(struct mem_info *mi) +{ + return &RC_CHK_ACCESS(mi)->iaddr; +} + +static inline struct addr_map_symbol *mem_info__daddr(struct mem_info *mi) +{ + return &RC_CHK_ACCESS(mi)->daddr; +} + +static inline union perf_mem_data_src *mem_info__data_src(struct mem_info *mi) +{ + return &RC_CHK_ACCESS(mi)->data_src; +} + +static inline const union perf_mem_data_src *mem_info__const_data_src(const struct mem_info *mi) +{ + return &RC_CHK_ACCESS(mi)->data_src; +} + +static inline refcount_t *mem_info__refcnt(struct mem_info *mi) +{ + return &RC_CHK_ACCESS(mi)->refcnt; +} + #endif /* __PERF_MEM_INFO_H */ diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c index c77fe08a0545..fb00f3ad6815 100644 --- a/tools/perf/util/scripting-engines/trace-event-python.c +++ b/tools/perf/util/scripting-engines/trace-event-python.c @@ -721,15 +721,20 @@ static void set_sample_read_in_dict(PyObject *dict_sample, } static void set_sample_datasrc_in_dict(PyObject *dict, - struct perf_sample *sample) + struct perf_sample *sample) { - struct mem_info mi = { .data_src.val = sample->data_src }; + struct mem_info *mi = mem_info__new(); char decode[100]; + if (!mi) + Py_FatalError("couldn't create mem-info"); + pydict_set_item_string_decref(dict, "datasrc", PyLong_FromUnsignedLongLong(sample->data_src)); - perf_script__meminfo_scnprintf(decode, 100, &mi); + mem_info__data_src(mi)->val = sample->data_src; + perf_script__meminfo_scnprintf(decode, 100, mi); + mem_info__put(mi); pydict_set_item_string_decref(dict, "datasrc_decode", _PyUnicode_FromString(decode)); diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c index 711ef69306f3..cd39ea972193 100644 --- a/tools/perf/util/sort.c +++ b/tools/perf/util/sort.c @@ -1365,9 +1365,9 @@ sort__daddr_cmp(struct hist_entry *left, struct hist_entry *right) uint64_t l = 0, r = 0; if (left->mem_info) - l = left->mem_info->daddr.addr; + l = mem_info__daddr(left->mem_info)->addr; if (right->mem_info) - r = right->mem_info->daddr.addr; + r = mem_info__daddr(right->mem_info)->addr; return (int64_t)(r - l); } @@ -1379,8 +1379,8 @@ static int hist_entry__daddr_snprintf(struct hist_entry *he, char *bf, struct map_symbol *ms = NULL; if (he->mem_info) { - addr = he->mem_info->daddr.addr; - ms = &he->mem_info->daddr.ms; + addr = mem_info__daddr(he->mem_info)->addr; + ms = &mem_info__daddr(he->mem_info)->ms; } return _hist_entry__sym_snprintf(ms, addr, he->level, bf, size, width); } @@ -1391,9 +1391,9 @@ sort__iaddr_cmp(struct hist_entry *left, struct hist_entry *right) uint64_t l = 0, r = 0; if (left->mem_info) - l = left->mem_info->iaddr.addr; + l = mem_info__iaddr(left->mem_info)->addr; if (right->mem_info) - r = right->mem_info->iaddr.addr; + r = mem_info__iaddr(right->mem_info)->addr; return (int64_t)(r - l); } @@ -1405,8 +1405,8 @@ static int hist_entry__iaddr_snprintf(struct hist_entry *he, char *bf, struct map_symbol *ms = NULL; if (he->mem_info) { - addr = he->mem_info->iaddr.addr; - ms = &he->mem_info->iaddr.ms; + addr = mem_info__iaddr(he->mem_info)->addr; + ms = &mem_info__iaddr(he->mem_info)->ms; } return _hist_entry__sym_snprintf(ms, addr, he->level, bf, size, width); } @@ -1418,9 +1418,9 @@ sort__dso_daddr_cmp(struct hist_entry *left, struct hist_entry *right) struct map *map_r = NULL; if (left->mem_info) - map_l = left->mem_info->daddr.ms.map; + map_l = mem_info__daddr(left->mem_info)->ms.map; if (right->mem_info) - map_r = right->mem_info->daddr.ms.map; + map_r = mem_info__daddr(right->mem_info)->ms.map; return _sort__dso_cmp(map_l, map_r); } @@ -1431,7 +1431,7 @@ static int hist_entry__dso_daddr_snprintf(struct hist_entry *he, char *bf, struct map *map = NULL; if (he->mem_info) - map = he->mem_info->daddr.ms.map; + map = mem_info__daddr(he->mem_info)->ms.map; return _hist_entry__dso_snprintf(map, bf, size, width); } @@ -1443,12 +1443,12 @@ sort__locked_cmp(struct hist_entry *left, struct hist_entry *right) union perf_mem_data_src data_src_r; if (left->mem_info) - data_src_l = left->mem_info->data_src; + data_src_l = *mem_info__data_src(left->mem_info); else data_src_l.mem_lock = PERF_MEM_LOCK_NA; if (right->mem_info) - data_src_r = right->mem_info->data_src; + data_src_r = *mem_info__data_src(right->mem_info); else data_src_r.mem_lock = PERF_MEM_LOCK_NA; @@ -1471,12 +1471,12 @@ sort__tlb_cmp(struct hist_entry *left, struct hist_entry *right) union perf_mem_data_src data_src_r; if (left->mem_info) - data_src_l = left->mem_info->data_src; + data_src_l = *mem_info__data_src(left->mem_info); else data_src_l.mem_dtlb = PERF_MEM_TLB_NA; if (right->mem_info) - data_src_r = right->mem_info->data_src; + data_src_r = *mem_info__data_src(right->mem_info); else data_src_r.mem_dtlb = PERF_MEM_TLB_NA; @@ -1499,12 +1499,12 @@ sort__lvl_cmp(struct hist_entry *left, struct hist_entry *right) union perf_mem_data_src data_src_r; if (left->mem_info) - data_src_l = left->mem_info->data_src; + data_src_l = *mem_info__data_src(left->mem_info); else data_src_l.mem_lvl = PERF_MEM_LVL_NA; if (right->mem_info) - data_src_r = right->mem_info->data_src; + data_src_r = *mem_info__data_src(right->mem_info); else data_src_r.mem_lvl = PERF_MEM_LVL_NA; @@ -1527,12 +1527,12 @@ sort__snoop_cmp(struct hist_entry *left, struct hist_entry *right) union perf_mem_data_src data_src_r; if (left->mem_info) - data_src_l = left->mem_info->data_src; + data_src_l = *mem_info__data_src(left->mem_info); else data_src_l.mem_snoop = PERF_MEM_SNOOP_NA; if (right->mem_info) - data_src_r = right->mem_info->data_src; + data_src_r = *mem_info__data_src(right->mem_info); else data_src_r.mem_snoop = PERF_MEM_SNOOP_NA; @@ -1563,8 +1563,8 @@ sort__dcacheline_cmp(struct hist_entry *left, struct hist_entry *right) if (left->cpumode > right->cpumode) return -1; if (left->cpumode < right->cpumode) return 1; - l_map = left->mem_info->daddr.ms.map; - r_map = right->mem_info->daddr.ms.map; + l_map = mem_info__daddr(left->mem_info)->ms.map; + r_map = mem_info__daddr(right->mem_info)->ms.map; /* if both are NULL, jump to sort on al_addr instead */ if (!l_map && !r_map) @@ -1599,8 +1599,8 @@ sort__dcacheline_cmp(struct hist_entry *left, struct hist_entry *right) addr: /* al_addr does all the right addr - start + offset calculations */ - l = cl_address(left->mem_info->daddr.al_addr, chk_double_cl); - r = cl_address(right->mem_info->daddr.al_addr, chk_double_cl); + l = cl_address(mem_info__daddr(left->mem_info)->al_addr, chk_double_cl); + r = cl_address(mem_info__daddr(right->mem_info)->al_addr, chk_double_cl); if (l > r) return -1; if (l < r) return 1; @@ -1617,11 +1617,11 @@ static int hist_entry__dcacheline_snprintf(struct hist_entry *he, char *bf, char level = he->level; if (he->mem_info) { - struct map *map = he->mem_info->daddr.ms.map; + struct map *map = mem_info__daddr(he->mem_info)->ms.map; struct dso *dso = map ? map__dso(map) : NULL; - addr = cl_address(he->mem_info->daddr.al_addr, chk_double_cl); - ms = &he->mem_info->daddr.ms; + addr = cl_address(mem_info__daddr(he->mem_info)->al_addr, chk_double_cl); + ms = &mem_info__daddr(he->mem_info)->ms; /* print [s] for shared data mmaps */ if ((he->cpumode != PERF_RECORD_MISC_KERNEL) && @@ -1806,12 +1806,12 @@ sort__blocked_cmp(struct hist_entry *left, struct hist_entry *right) union perf_mem_data_src data_src_r; if (left->mem_info) - data_src_l = left->mem_info->data_src; + data_src_l = *mem_info__data_src(left->mem_info); else data_src_l.mem_blk = PERF_MEM_BLK_NA; if (right->mem_info) - data_src_r = right->mem_info->data_src; + data_src_r = *mem_info__data_src(right->mem_info); else data_src_r.mem_blk = PERF_MEM_BLK_NA; @@ -1840,9 +1840,9 @@ sort__phys_daddr_cmp(struct hist_entry *left, struct hist_entry *right) uint64_t l = 0, r = 0; if (left->mem_info) - l = left->mem_info->daddr.phys_addr; + l = mem_info__daddr(left->mem_info)->phys_addr; if (right->mem_info) - r = right->mem_info->daddr.phys_addr; + r = mem_info__daddr(right->mem_info)->phys_addr; return (int64_t)(r - l); } @@ -1854,7 +1854,7 @@ static int hist_entry__phys_daddr_snprintf(struct hist_entry *he, char *bf, size_t ret = 0; size_t len = BITS_PER_LONG / 4; - addr = he->mem_info->daddr.phys_addr; + addr = mem_info__daddr(he->mem_info)->phys_addr; ret += repsep_snprintf(bf + ret, size - ret, "[%c] ", he->level); @@ -1881,9 +1881,9 @@ sort__data_page_size_cmp(struct hist_entry *left, struct hist_entry *right) uint64_t l = 0, r = 0; if (left->mem_info) - l = left->mem_info->daddr.data_page_size; + l = mem_info__daddr(left->mem_info)->data_page_size; if (right->mem_info) - r = right->mem_info->daddr.data_page_size; + r = mem_info__daddr(right->mem_info)->data_page_size; return (int64_t)(r - l); } @@ -1894,7 +1894,7 @@ static int hist_entry__data_page_size_snprintf(struct hist_entry *he, char *bf, char str[PAGE_SIZE_NAME_LEN]; return repsep_snprintf(bf, size, "%-*s", width, - get_page_size_name(he->mem_info->daddr.data_page_size, str)); + get_page_size_name(mem_info__daddr(he->mem_info)->data_page_size, str)); } struct sort_entry sort_mem_data_page_size = { -- cgit v1.2.3 From d561e170bd07db62c9b8dbe3d0f59085447ca77a Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 7 May 2024 11:35:45 -0700 Subject: perf hist: Avoid 'struct hist_entry_iter' mem_info memory leak 'struct mem_info' is reference counted while 'struct branch_info' and he_cache (struct hist_entry **) are not. Break apart the priv field in 'struct hist_entry_iter' so that we can know which values are owned by the iter and do the appropriate free or put. Move hide_unresolved to marginally shrink the size of the now grown struct. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Ben Gainey Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: K Prateek Nayak Cc: Kajol Jain Cc: Kan Liang Cc: Li Dong Cc: Mark Rutland Cc: Namhyung Kim Cc: Oliver Upton Cc: Paran Lee Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Sun Haiyong Cc: Tim Chen Cc: Yanteng Si Cc: Yicong Yang Link: https://lore.kernel.org/r/20240507183545.1236093-9-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/hist.c | 39 ++++++++++++++------------------------- tools/perf/util/hist.h | 8 +++++--- 2 files changed, 19 insertions(+), 28 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c index 00814d42d5f1..2e9e193179dd 100644 --- a/tools/perf/util/hist.c +++ b/tools/perf/util/hist.c @@ -476,13 +476,6 @@ static int hist_entry__init(struct hist_entry *he, he->branch_info->to.ms.map = map__get(he->branch_info->to.ms.map); } - if (he->mem_info) { - mem_info__iaddr(he->mem_info)->ms.map = - map__get(mem_info__iaddr(he->mem_info)->ms.map); - mem_info__daddr(he->mem_info)->ms.map = - map__get(mem_info__daddr(he->mem_info)->ms.map); - } - if (hist_entry__has_callchains(he) && symbol_conf.use_callchain) callchain_init(he->callchain); @@ -574,7 +567,6 @@ static struct hist_entry *hist_entry__new(struct hist_entry *template, he = NULL; } } - return he; } @@ -747,7 +739,7 @@ __hists__add_entry(struct hists *hists, .filtered = symbol__parent_filter(sym_parent) | al->filtered, .hists = hists, .branch_info = bi, - .mem_info = mi, + .mem_info = mem_info__get(mi), .kvm_info = ki, .block_info = block_info, .transaction = sample->transaction, @@ -836,7 +828,7 @@ iter_prepare_mem_entry(struct hist_entry_iter *iter, struct addr_location *al) if (mi == NULL) return -ENOMEM; - iter->priv = mi; + iter->mi = mi; return 0; } @@ -844,7 +836,7 @@ static int iter_add_single_mem_entry(struct hist_entry_iter *iter, struct addr_location *al) { u64 cost; - struct mem_info *mi = iter->priv; + struct mem_info *mi = iter->mi; struct hists *hists = evsel__hists(iter->evsel); struct perf_sample *sample = iter->sample; struct hist_entry *he; @@ -891,12 +883,7 @@ iter_finish_mem_entry(struct hist_entry_iter *iter, err = hist_entry__append_callchain(he, iter->sample); out: - /* - * We don't need to free iter->priv (mem_info) here since the mem info - * was either already freed in hists__findnew_entry() or passed to a - * new hist entry by hist_entry__new(). - */ - iter->priv = NULL; + mem_info__zput(iter->mi); iter->he = NULL; return err; @@ -915,7 +902,7 @@ iter_prepare_branch_entry(struct hist_entry_iter *iter, struct addr_location *al iter->curr = 0; iter->total = sample->branch_stack->nr; - iter->priv = bi; + iter->bi = bi; return 0; } @@ -929,7 +916,7 @@ iter_add_single_branch_entry(struct hist_entry_iter *iter __maybe_unused, static int iter_next_branch_entry(struct hist_entry_iter *iter, struct addr_location *al) { - struct branch_info *bi = iter->priv; + struct branch_info *bi = iter->bi; int i = iter->curr; if (bi == NULL) @@ -958,7 +945,7 @@ iter_add_next_branch_entry(struct hist_entry_iter *iter, struct addr_location *a int i = iter->curr; int err = 0; - bi = iter->priv; + bi = iter->bi; if (iter->hide_unresolved && !(bi[i].from.ms.sym && bi[i].to.ms.sym)) goto out; @@ -987,7 +974,7 @@ static int iter_finish_branch_entry(struct hist_entry_iter *iter, struct addr_location *al __maybe_unused) { - zfree(&iter->priv); + zfree(&iter->bi); iter->he = NULL; return iter->curr >= iter->total ? 0 : -1; @@ -1055,7 +1042,7 @@ iter_prepare_cumulative_entry(struct hist_entry_iter *iter, if (he_cache == NULL) return -ENOMEM; - iter->priv = he_cache; + iter->he_cache = he_cache; iter->curr = 0; return 0; @@ -1068,7 +1055,7 @@ iter_add_single_cumulative_entry(struct hist_entry_iter *iter, struct evsel *evsel = iter->evsel; struct hists *hists = evsel__hists(evsel); struct perf_sample *sample = iter->sample; - struct hist_entry **he_cache = iter->priv; + struct hist_entry **he_cache = iter->he_cache; struct hist_entry *he; int err = 0; @@ -1126,7 +1113,7 @@ iter_add_next_cumulative_entry(struct hist_entry_iter *iter, { struct evsel *evsel = iter->evsel; struct perf_sample *sample = iter->sample; - struct hist_entry **he_cache = iter->priv; + struct hist_entry **he_cache = iter->he_cache; struct hist_entry *he; struct hist_entry he_tmp = { .hists = evsel__hists(evsel), @@ -1192,7 +1179,9 @@ static int iter_finish_cumulative_entry(struct hist_entry_iter *iter, struct addr_location *al __maybe_unused) { - zfree(&iter->priv); + mem_info__zput(iter->mi); + zfree(&iter->bi); + zfree(&iter->he_cache); iter->he = NULL; return 0; diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h index 5260822b9773..8fb3bdd29188 100644 --- a/tools/perf/util/hist.h +++ b/tools/perf/util/hist.h @@ -132,18 +132,20 @@ struct hist_entry_iter { int total; int curr; - bool hide_unresolved; - struct evsel *evsel; struct perf_sample *sample; struct hist_entry *he; struct symbol *parent; - void *priv; + + struct mem_info *mi; + struct branch_info *bi; + struct hist_entry **he_cache; const struct hist_iter_ops *ops; /* user-defined callback function (optional) */ int (*add_entry_cb)(struct hist_entry_iter *iter, struct addr_location *al, bool single, void *arg); + bool hide_unresolved; }; extern const struct hist_iter_ops hist_iter_normal; -- cgit v1.2.3 From 187c219b57eaf3e1b7a3cab2c6a8b7909bdbf4a9 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 6 May 2024 21:13:38 -0700 Subject: perf dwarf-aux: Print array type name with "[]" It's confusing both pointers and arrays are printed as *. Let's print array types with [] so that we can identify them easily. Although it's interchangable, sometimes it can cause confusion with size like in the below example. Note that it is not the same with C syntax where it goes to the variable names, but we want to have it in the type names (like in Go language). Before: mov [20] 0x68(reg5) -> reg0 type='struct page**' size=0x80 (die:0x4e61d32) After: mov [20] 0x68(reg5) -> reg0 type='struct page*[]' size=0x80 (die:0x4e61d32) Signed-off-by: Namhyung Kim Acked-by: Masami Hiramatsu Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240507041338.2081775-1-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dwarf-aux.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c index c0a492e65388..ec988f294497 100644 --- a/tools/perf/util/dwarf-aux.c +++ b/tools/perf/util/dwarf-aux.c @@ -1110,8 +1110,10 @@ int die_get_typename_from_type(Dwarf_Die *type_die, struct strbuf *buf) const char *tmp = ""; tag = dwarf_tag(type_die); - if (tag == DW_TAG_array_type || tag == DW_TAG_pointer_type) + if (tag == DW_TAG_pointer_type) tmp = "*"; + else if (tag == DW_TAG_array_type) + tmp = "[]"; else if (tag == DW_TAG_subroutine_type) { /* Function pointer */ return strbuf_add(buf, "(function_type)", 15); -- cgit v1.2.3 From de6a908384fb1a8327ba46a2baec67d1dfe9a3e1 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 7 May 2024 20:52:59 -0700 Subject: perf comm: Fix comm_str__put() for reference count checking Searching for the entry in the array needs to avoid the intermediate pointer with reference count checking. Refactor the array removal to binary search for the entry. Change the array to hold an entry with a reference count (so the intermediate pointer can work) and remove from the array when the reference count on a comm_str falls to 1. Fixes: 13ca628716c6f2c3 ("perf comm: Add reference count checking to 'struct comm_str'") Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Kan Liang Cc: Leo Yan Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240508035301.1554434-3-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/comm.c | 45 +++++++++++++++++++++++++++++++-------------- 1 file changed, 31 insertions(+), 14 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/comm.c b/tools/perf/util/comm.c index 1aa9a08e5b03..233f2b6edf52 100644 --- a/tools/perf/util/comm.c +++ b/tools/perf/util/comm.c @@ -19,6 +19,8 @@ static struct comm_strs { int capacity; } _comm_strs; +static void comm_strs__remove_if_last(struct comm_str *cs); + static void comm_strs__init(void) { init_rwsem(&_comm_strs.lock); @@ -58,22 +60,15 @@ static struct comm_str *comm_str__get(struct comm_str *cs) static void comm_str__put(struct comm_str *cs) { - if (cs && refcount_dec_and_test(comm_str__refcnt(cs))) { - struct comm_strs *comm_strs = comm_strs__get(); - int i; + if (!cs) + return; - down_write(&comm_strs->lock); - for (i = 0; i < comm_strs->num_strs; i++) { - if (comm_strs->strs[i] == cs) - break; - } - for (; i < comm_strs->num_strs - 1; i++) - comm_strs->strs[i] = comm_strs->strs[i + 1]; - - comm_strs->num_strs--; - up_write(&comm_strs->lock); + if (refcount_dec_and_test(comm_str__refcnt(cs))) { RC_CHK_FREE(cs); } else { + if (refcount_read(comm_str__refcnt(cs)) == 1) + comm_strs__remove_if_last(cs); + RC_CHK_PUT(cs); } } @@ -107,6 +102,28 @@ static int comm_str__search(const void *_key, const void *_member) return strcmp(key, comm_str__str(member)); } +static void comm_strs__remove_if_last(struct comm_str *cs) +{ + struct comm_strs *comm_strs = comm_strs__get(); + + down_write(&comm_strs->lock); + /* + * Are there only references from the array, if so remove the array + * reference under the write lock so that we don't race with findnew. + */ + if (refcount_read(comm_str__refcnt(cs)) == 1) { + struct comm_str **entry; + + entry = bsearch(comm_str__str(cs), comm_strs->strs, comm_strs->num_strs, + sizeof(struct comm_str *), comm_str__search); + comm_str__put(*entry); + for (int i = entry - comm_strs->strs; i < comm_strs->num_strs - 1; i++) + comm_strs->strs[i] = comm_strs->strs[i + 1]; + comm_strs->num_strs--; + } + up_write(&comm_strs->lock); +} + static struct comm_str *__comm_strs__find(struct comm_strs *comm_strs, const char *str) { struct comm_str **result; @@ -158,7 +175,7 @@ static struct comm_str *comm_strs__findnew(const char *str) } } up_write(&comm_strs->lock); - return result; + return comm_str__get(result); } struct comm *comm__new(const char *str, u64 timestamp, bool exec) -- cgit v1.2.3 From 3536c2575e88a890cf696b4ccd3da36bc937853b Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Tue, 7 May 2024 20:53:01 -0700 Subject: perf thread: Fixes to thread__new() related to initializing comm Freeing the thread on failure won't work with reference count checking, use thread__delete(). Don't allocate the comm_str, use a stack allocation instead. Fixes: f6005cafebab72f8 ("perf thread: Add reference count checking") Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Kan Liang Cc: Leo Yan Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240508035301.1554434-5-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/thread.c | 14 +++++--------- 1 file changed, 5 insertions(+), 9 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index 0a473112f881..87c59aa9fe38 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -39,12 +39,13 @@ int thread__init_maps(struct thread *thread, struct machine *machine) struct thread *thread__new(pid_t pid, pid_t tid) { - char *comm_str; - struct comm *comm; RC_STRUCT(thread) *_thread = zalloc(sizeof(*_thread)); struct thread *thread; if (ADD_RC_CHK(thread, _thread) != NULL) { + struct comm *comm; + char comm_str[32]; + thread__set_pid(thread, pid); thread__set_tid(thread, tid); thread__set_ppid(thread, -1); @@ -56,13 +57,8 @@ struct thread *thread__new(pid_t pid, pid_t tid) init_rwsem(thread__namespaces_lock(thread)); init_rwsem(thread__comm_lock(thread)); - comm_str = malloc(32); - if (!comm_str) - goto err_thread; - - snprintf(comm_str, 32, ":%d", tid); + snprintf(comm_str, sizeof(comm_str), ":%d", tid); comm = comm__new(comm_str, 0, false); - free(comm_str); if (!comm) goto err_thread; @@ -76,7 +72,7 @@ struct thread *thread__new(pid_t pid, pid_t tid) return thread; err_thread: - free(thread); + thread__delete(thread); return NULL; } -- cgit v1.2.3 From c9d492378faec5c1fb8ea1534c620f7320bbb23a Mon Sep 17 00:00:00 2001 From: James Clark Date: Wed, 8 May 2024 15:14:57 +0100 Subject: perf dwarf-aux: Fix build with HAVE_DWARF_CFI_SUPPORT check_allowed_ops() is used from both HAVE_DWARF_GETLOCATIONS_SUPPORT and HAVE_DWARF_CFI_SUPPORT sections, so move it into the right place so that it's available when either are defined. This shows up when doing a static cross compile for arm64: $ make ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu- LDFLAGS="-static" \ EXTRA_PERFLIBS="-lexpat" util/dwarf-aux.c:1723:6: error: implicit declaration of function 'check_allowed_ops' Fixes: 55442cc2f22d0727 ("perf dwarf-aux: Check allowed DWARF Ops") Reviewed-by: Ian Rogers Signed-off-by: James Clark Acked-by: Masami Hiramatsu Acked-by: Namhyung Kim Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: Jiri Olsa Cc: Leo Yan Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240508141458.439017-1-james.clark@arm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dwarf-aux.c | 56 ++++++++++++++++++++++----------------------- 1 file changed, 28 insertions(+), 28 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c index ec988f294497..44ef968a7ad3 100644 --- a/tools/perf/util/dwarf-aux.c +++ b/tools/perf/util/dwarf-aux.c @@ -1217,6 +1217,34 @@ static int offset_from_dwarf_op(Dwarf_Op *op) } return -1; } + +static bool check_allowed_ops(Dwarf_Op *ops, size_t nops) +{ + /* The first op is checked separately */ + ops++; + nops--; + + /* + * It needs to make sure if the location expression matches to the given + * register and offset exactly. Thus it rejects any complex expressions + * and only allows a few of selected operators that doesn't change the + * location. + */ + while (nops) { + switch (ops->atom) { + case DW_OP_stack_value: + case DW_OP_deref_size: + case DW_OP_deref: + case DW_OP_piece: + break; + default: + return false; + } + ops++; + nops--; + } + return true; +} #endif /* HAVE_DWARF_GETLOCATIONS_SUPPORT || HAVE_DWARF_CFI_SUPPORT */ #ifdef HAVE_DWARF_GETLOCATIONS_SUPPORT @@ -1397,34 +1425,6 @@ static bool match_var_offset(Dwarf_Die *die_mem, struct find_var_data *data, return true; } -static bool check_allowed_ops(Dwarf_Op *ops, size_t nops) -{ - /* The first op is checked separately */ - ops++; - nops--; - - /* - * It needs to make sure if the location expression matches to the given - * register and offset exactly. Thus it rejects any complex expressions - * and only allows a few of selected operators that doesn't change the - * location. - */ - while (nops) { - switch (ops->atom) { - case DW_OP_stack_value: - case DW_OP_deref_size: - case DW_OP_deref: - case DW_OP_piece: - break; - default: - return false; - } - ops++; - nops--; - } - return true; -} - /* Only checks direct child DIEs in the given scope. */ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg) { -- cgit v1.2.3 From d790ead8a60c0f687446f7d8faa96943dc158912 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 9 May 2024 08:32:45 -0700 Subject: perf tracepoint: Don't scan all tracepoints to test if one exists In is_valid_tracepoint, rather than scanning "/sys/kernel/tracing/events/*/*" skipping any path where "/sys/kernel/tracing/events/*/*/id" doesn't exist, and then testing if "*:*" matches the tracepoint name, just use the given tracepoint name replace the ':' with '/' and see if the id file exists. This turns a nested directory search into a single file available test. Rather than return 1 for valid and 0 for invalid, return true and false. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240509153245.1990426-1-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/tracepoint.c | 56 +++++++++++++++++--------------------------- tools/perf/util/tracepoint.h | 3 ++- 2 files changed, 24 insertions(+), 35 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/tracepoint.c b/tools/perf/util/tracepoint.c index 92dd8b455b90..95377ed5d87b 100644 --- a/tools/perf/util/tracepoint.c +++ b/tools/perf/util/tracepoint.c @@ -4,10 +4,12 @@ #include #include #include +#include #include #include #include +#include "fncache.h" int tp_event_has_id(const char *dir_path, struct dirent *evt_dir) { @@ -26,39 +28,25 @@ int tp_event_has_id(const char *dir_path, struct dirent *evt_dir) /* * Check whether event is in /tracing/events */ -int is_valid_tracepoint(const char *event_string) +bool is_valid_tracepoint(const char *event_string) { - DIR *sys_dir, *evt_dir; - struct dirent *sys_dirent, *evt_dirent; - char evt_path[MAXPATHLEN]; - char *dir_path; - - sys_dir = tracing_events__opendir(); - if (!sys_dir) - return 0; - - for_each_subsystem(sys_dir, sys_dirent) { - dir_path = get_events_file(sys_dirent->d_name); - if (!dir_path) - continue; - evt_dir = opendir(dir_path); - if (!evt_dir) - goto next; - - for_each_event(dir_path, evt_dir, evt_dirent) { - snprintf(evt_path, MAXPATHLEN, "%s:%s", - sys_dirent->d_name, evt_dirent->d_name); - if (!strcmp(evt_path, event_string)) { - closedir(evt_dir); - put_events_file(dir_path); - closedir(sys_dir); - return 1; - } - } - closedir(evt_dir); -next: - put_events_file(dir_path); - } - closedir(sys_dir); - return 0; + char *dst, *path = malloc(strlen(event_string) + 4); /* Space for "/id\0". */ + bool have_file = false; /* Conservatively return false if memory allocation failed. */ + const char *src; + + if (!path) + return false; + + /* Copy event_string replacing the ':' with '/'. */ + for (src = event_string, dst = path; *src; src++, dst++) + *dst = (*src == ':') ? '/' : *src; + /* Add "/id\0". */ + memcpy(dst, "/id", 4); + + dst = get_events_file(path); + if (dst) + have_file = file_available(dst); + free(dst); + free(path); + return have_file; } diff --git a/tools/perf/util/tracepoint.h b/tools/perf/util/tracepoint.h index c4a110fe87d7..65ccb01fc312 100644 --- a/tools/perf/util/tracepoint.h +++ b/tools/perf/util/tracepoint.h @@ -4,6 +4,7 @@ #include #include +#include int tp_event_has_id(const char *dir_path, struct dirent *evt_dir); @@ -20,6 +21,6 @@ int tp_event_has_id(const char *dir_path, struct dirent *evt_dir); (strcmp(sys_dirent->d_name, ".")) && \ (strcmp(sys_dirent->d_name, ".."))) -int is_valid_tracepoint(const char *event_string); +bool is_valid_tracepoint(const char *event_string); #endif /* __PERF_TRACEPOINT_H */ -- cgit v1.2.3 From 9fe410a7ef483a9aca08bf620d8ddfd35ac99bc7 Mon Sep 17 00:00:00 2001 From: James Clark Date: Tue, 7 May 2024 15:12:05 +0100 Subject: perf symbols: Remove map from list before updating addresses Make the order of operations remove, update, add. Updating addresses before the map is removed causes the ordering check to fail when the map is removed. This can be reproduced when running Perf on an Arm system with a static kernel and Perf uses kcore rather than other sources: $ perf record -- ls $ perf report util/maps.c:96: check_invariants: Assertion `map__end(prev) <= map__start(map) || map__start(prev) == map__start(map)' failed Fixes: 659ad3492b913c90 ("perf maps: Switch from rbtree to lazily sorted array for addresses") Signed-off-by: James Clark Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Athira Rajeev Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240507141210.195939-2-james.clark@arm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index eb3319baa1b5..0d4de786358d 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1378,13 +1378,15 @@ static int dso__load_kcore(struct dso *dso, struct map *map, if (RC_CHK_EQUAL(new_map, replacement_map)) { struct map *map_ref; - map__set_start(map, map__start(new_map)); - map__set_end(map, map__end(new_map)); - map__set_pgoff(map, map__pgoff(new_map)); - map__set_mapping_type(map, map__mapping_type(new_map)); /* Ensure maps are correctly ordered */ map_ref = map__get(map); maps__remove(kmaps, map_ref); + + map__set_start(map_ref, map__start(new_map)); + map__set_end(map_ref, map__end(new_map)); + map__set_pgoff(map_ref, map__pgoff(new_map)); + map__set_mapping_type(map_ref, map__mapping_type(new_map)); + err = maps__insert(kmaps, map_ref); map__put(map_ref); map__put(new_map); -- cgit v1.2.3 From fd81f52e311f11f3eba842439948f989ffa8ccc2 Mon Sep 17 00:00:00 2001 From: James Clark Date: Tue, 7 May 2024 15:12:06 +0100 Subject: perf maps: Re-use __maps__free_maps_by_name() maps__merge_in() hard codes the steps to free the maps_by_name list. It seems to not map__put() each element before freeing, and it sets maps_by_name_sorted to true after freeing, which may be harmless but is inconsistent with maps__init() and other functions. maps__maps_by_name_addr() is also quite hard to read because we already have maps__maps_by_name() and maps__maps_by_address(), but the function is only used in that place so delete it. Reviewed-by: Ian Rogers Signed-off-by: James Clark Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Athira Rajeev Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240507141210.195939-3-james.clark@arm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/maps.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/maps.c b/tools/perf/util/maps.c index 61eb742d91e3..16b39db594f4 100644 --- a/tools/perf/util/maps.c +++ b/tools/perf/util/maps.c @@ -124,11 +124,6 @@ static void maps__set_maps_by_address(struct maps *maps, struct map **new) } -static struct map ***maps__maps_by_name_addr(struct maps *maps) -{ - return &RC_CHK_ACCESS(maps)->maps_by_name; -} - static void maps__set_nr_maps_allocated(struct maps *maps, unsigned int nr_maps_allocated) { RC_CHK_ACCESS(maps)->nr_maps_allocated = nr_maps_allocated; @@ -284,6 +279,9 @@ void maps__put(struct maps *maps) static void __maps__free_maps_by_name(struct maps *maps) { + if (!maps__maps_by_name(maps)) + return; + /* * Free everything to try to do it from the rbtree in the next search */ @@ -291,6 +289,9 @@ static void __maps__free_maps_by_name(struct maps *maps) map__put(maps__maps_by_name(maps)[i]); zfree(&RC_CHK_ACCESS(maps)->maps_by_name); + + /* Consistent with maps__init(). When maps_by_name == NULL, maps_by_name_sorted == false */ + maps__set_maps_by_name_sorted(maps, false); } static int map__start_cmp(const void *a, const void *b) @@ -1167,8 +1168,7 @@ int maps__merge_in(struct maps *kmaps, struct map *new_map) } maps__set_maps_by_address(kmaps, merged_maps_by_address); maps__set_maps_by_address_sorted(kmaps, true); - zfree(maps__maps_by_name_addr(kmaps)); - maps__set_maps_by_name_sorted(kmaps, true); + __maps__free_maps_by_name(kmaps); maps__set_nr_maps_allocated(kmaps, merged_nr_maps_allocated); /* Copy entries before the new_map that can't overlap. */ -- cgit v1.2.3 From f30232b20fadea8c0f2f43f764bc06e51e8cfcdf Mon Sep 17 00:00:00 2001 From: James Clark Date: Tue, 7 May 2024 15:12:07 +0100 Subject: perf symbols: Update kcore map before merging in remaining symbols When loading kcore, the main vmlinux map is updated in the same loop that merges the remaining maps. If a map that overlaps is merged in before kcore, the list can become unsortable when the main map addresses are updated. This will later trigger the check_invariants() assert: $ perf record $ perf report util/maps.c:96: check_invariants: Assertion `map__end(prev) <= map__start(map) || map__start(prev) == map__start(map)' failed. Aborted Fix it by moving the main map update prior to the loop so that maps__merge_in() can split it if necessary. Fixes: 659ad3492b913c90 ("perf maps: Switch from rbtree to lazily sorted array for addresses") Signed-off-by: James Clark Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Athira Rajeev Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240507141210.195939-4-james.clark@arm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol.c | 40 +++++++++++++++++++++------------------- 1 file changed, 21 insertions(+), 19 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 0d4de786358d..f21fe5772bd8 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1290,7 +1290,7 @@ static int dso__load_kcore(struct dso *dso, struct map *map, { struct maps *kmaps = map__kmaps(map); struct kcore_mapfn_data md; - struct map *replacement_map = NULL; + struct map *map_ref, *replacement_map = NULL; struct machine *machine; bool is_64_bit; int err, fd; @@ -1368,6 +1368,24 @@ static int dso__load_kcore(struct dso *dso, struct map *map, if (!replacement_map) replacement_map = list_entry(md.maps.next, struct map_list_node, node)->map; + /* + * Update addresses of vmlinux map. Re-insert it to ensure maps are + * correctly ordered. Do this before using maps__merge_in() for the + * remaining maps so vmlinux gets split if necessary. + */ + map_ref = map__get(map); + maps__remove(kmaps, map_ref); + + map__set_start(map_ref, map__start(replacement_map)); + map__set_end(map_ref, map__end(replacement_map)); + map__set_pgoff(map_ref, map__pgoff(replacement_map)); + map__set_mapping_type(map_ref, map__mapping_type(replacement_map)); + + err = maps__insert(kmaps, map_ref); + map__put(map_ref); + if (err) + goto out_err; + /* Add new maps */ while (!list_empty(&md.maps)) { struct map_list_node *new_node = list_entry(md.maps.next, struct map_list_node, node); @@ -1375,24 +1393,8 @@ static int dso__load_kcore(struct dso *dso, struct map *map, list_del_init(&new_node->node); - if (RC_CHK_EQUAL(new_map, replacement_map)) { - struct map *map_ref; - - /* Ensure maps are correctly ordered */ - map_ref = map__get(map); - maps__remove(kmaps, map_ref); - - map__set_start(map_ref, map__start(new_map)); - map__set_end(map_ref, map__end(new_map)); - map__set_pgoff(map_ref, map__pgoff(new_map)); - map__set_mapping_type(map_ref, map__mapping_type(new_map)); - - err = maps__insert(kmaps, map_ref); - map__put(map_ref); - map__put(new_map); - if (err) - goto out_err; - } else { + /* skip if replacement_map, already inserted above */ + if (!RC_CHK_EQUAL(new_map, replacement_map)) { /* * Merge kcore map into existing maps, * and ensure that current maps (eBPF) -- cgit v1.2.3 From 25626e19ae6df34f336f235b6b3dbd1b566d2738 Mon Sep 17 00:00:00 2001 From: James Clark Date: Tue, 7 May 2024 15:12:08 +0100 Subject: perf symbols: Fix ownership of string in dso__load_vmlinux() The linked commit updated dso__load_vmlinux() to call dso__set_long_name() before loading the symbols. Loading the symbols may not succeed but dso__set_long_name() takes ownership of the string. The two callers of this function free the string themselves on failure cases, resulting in the following error: $ perf record -- ls $ perf report free(): double free detected in tcache 2 Fix it by always taking ownership of the string, even on failure. This means the string is either freed at the very first early exit condition, or later when the dso is deleted or the long name is replaced. Now no special return value is needed to signify that the caller needs to free the string. Fixes: e59fea47f83e8a9a ("perf symbols: Fix DSO kernel load and symbol process to correctly map DSO to its long_name, type and adjust_symbols") Reviewed-by: Ian Rogers Signed-off-by: James Clark Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Athira Rajeev Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240507141210.195939-5-james.clark@arm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index f21fe5772bd8..9e5940b5bc59 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1978,6 +1978,10 @@ out: return ret; } +/* + * Always takes ownership of vmlinux when vmlinux_allocated == true, even if + * it returns an error. + */ int dso__load_vmlinux(struct dso *dso, struct map *map, const char *vmlinux, bool vmlinux_allocated) { @@ -1996,8 +2000,11 @@ int dso__load_vmlinux(struct dso *dso, struct map *map, else symtab_type = DSO_BINARY_TYPE__VMLINUX; - if (symsrc__init(&ss, dso, symfs_vmlinux, symtab_type)) + if (symsrc__init(&ss, dso, symfs_vmlinux, symtab_type)) { + if (vmlinux_allocated) + free((char *) vmlinux); return -1; + } /* * dso__load_sym() may copy 'dso' which will result in the copies having @@ -2040,7 +2047,6 @@ int dso__load_vmlinux_path(struct dso *dso, struct map *map) err = dso__load_vmlinux(dso, map, filename, true); if (err > 0) goto out; - free(filename); } out: return err; @@ -2192,7 +2198,6 @@ static int dso__load_kernel_sym(struct dso *dso, struct map *map) err = dso__load_vmlinux(dso, map, filename, true); if (err > 0) return err; - free(filename); } if (!symbol_conf.ignore_vmlinux && vmlinux_path != NULL) { -- cgit v1.2.3 From 11a42964850b5b6f866f64b0157e6e99b4d7ab9d Mon Sep 17 00:00:00 2001 From: Dominique Martinet Date: Fri, 10 May 2024 07:24:30 +0900 Subject: perf parse-events: pass parse_state to add_tracepoint The next patch will add another flag to parse_state that we will want to pass to evsel__newtp_idx(), so pass the whole parse_state all the way down instead of giving only the index Originally-by: Jiri Olsa Reviewed-by: Ian Rogers Signed-off-by: Dominique Martinet Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240510-perf_digit-v4-1-db1553f3233b@codewreck.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 31 ++++++++++++++++++------------- tools/perf/util/parse-events.h | 3 ++- tools/perf/util/parse-events.y | 2 +- 3 files changed, 21 insertions(+), 15 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 2b9ede311c31..e6a2a80b02df 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -543,13 +543,14 @@ static void tracepoint_error(struct parse_events_error *e, int err, parse_events_error__handle(e, column, strdup(str), strdup(help)); } -static int add_tracepoint(struct list_head *list, int *idx, +static int add_tracepoint(struct parse_events_state *parse_state, + struct list_head *list, const char *sys_name, const char *evt_name, struct parse_events_error *err, struct parse_events_terms *head_config, void *loc_) { YYLTYPE *loc = loc_; - struct evsel *evsel = evsel__newtp_idx(sys_name, evt_name, (*idx)++); + struct evsel *evsel = evsel__newtp_idx(sys_name, evt_name, parse_state->idx++); if (IS_ERR(evsel)) { tracepoint_error(err, PTR_ERR(evsel), sys_name, evt_name, loc->first_column); @@ -568,7 +569,8 @@ static int add_tracepoint(struct list_head *list, int *idx, return 0; } -static int add_tracepoint_multi_event(struct list_head *list, int *idx, +static int add_tracepoint_multi_event(struct parse_events_state *parse_state, + struct list_head *list, const char *sys_name, const char *evt_name, struct parse_events_error *err, struct parse_events_terms *head_config, YYLTYPE *loc) @@ -602,7 +604,7 @@ static int add_tracepoint_multi_event(struct list_head *list, int *idx, found++; - ret = add_tracepoint(list, idx, sys_name, evt_ent->d_name, + ret = add_tracepoint(parse_state, list, sys_name, evt_ent->d_name, err, head_config, loc); } @@ -616,19 +618,21 @@ static int add_tracepoint_multi_event(struct list_head *list, int *idx, return ret; } -static int add_tracepoint_event(struct list_head *list, int *idx, +static int add_tracepoint_event(struct parse_events_state *parse_state, + struct list_head *list, const char *sys_name, const char *evt_name, struct parse_events_error *err, struct parse_events_terms *head_config, YYLTYPE *loc) { return strpbrk(evt_name, "*?") ? - add_tracepoint_multi_event(list, idx, sys_name, evt_name, + add_tracepoint_multi_event(parse_state, list, sys_name, evt_name, err, head_config, loc) : - add_tracepoint(list, idx, sys_name, evt_name, + add_tracepoint(parse_state, list, sys_name, evt_name, err, head_config, loc); } -static int add_tracepoint_multi_sys(struct list_head *list, int *idx, +static int add_tracepoint_multi_sys(struct parse_events_state *parse_state, + struct list_head *list, const char *sys_name, const char *evt_name, struct parse_events_error *err, struct parse_events_terms *head_config, YYLTYPE *loc) @@ -654,7 +658,7 @@ static int add_tracepoint_multi_sys(struct list_head *list, int *idx, if (!strglobmatch(events_ent->d_name, sys_name)) continue; - ret = add_tracepoint_event(list, idx, events_ent->d_name, + ret = add_tracepoint_event(parse_state, list, events_ent->d_name, evt_name, err, head_config, loc); } @@ -1291,7 +1295,8 @@ static int get_config_chgs(struct perf_pmu *pmu, struct parse_events_terms *head return 0; } -int parse_events_add_tracepoint(struct list_head *list, int *idx, +int parse_events_add_tracepoint(struct parse_events_state *parse_state, + struct list_head *list, const char *sys, const char *event, struct parse_events_error *err, struct parse_events_terms *head_config, void *loc_) @@ -1307,14 +1312,14 @@ int parse_events_add_tracepoint(struct list_head *list, int *idx, } if (strpbrk(sys, "*?")) - return add_tracepoint_multi_sys(list, idx, sys, event, + return add_tracepoint_multi_sys(parse_state, list, sys, event, err, head_config, loc); else - return add_tracepoint_event(list, idx, sys, event, + return add_tracepoint_event(parse_state, list, sys, event, err, head_config, loc); #else + (void)parse_state; (void)list; - (void)idx; (void)sys; (void)event; (void)head_config; diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index e7ac1f13376d..c06984bd3af8 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -208,7 +208,8 @@ int parse_events__modifier_event(struct parse_events_state *parse_state, void *l int parse_events__modifier_group(struct parse_events_state *parse_state, void *loc, struct list_head *list, struct parse_events_modifier mod); int parse_events__set_default_name(struct list_head *list, char *name); -int parse_events_add_tracepoint(struct list_head *list, int *idx, +int parse_events_add_tracepoint(struct parse_events_state *parse_state, + struct list_head *list, const char *sys, const char *event, struct parse_events_error *error, struct parse_events_terms *head_config, void *loc); diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y index 68b3b06c7ff0..c94a3994177e 100644 --- a/tools/perf/util/parse-events.y +++ b/tools/perf/util/parse-events.y @@ -497,7 +497,7 @@ tracepoint_name opt_event_config if (!list) YYNOMEM; - err = parse_events_add_tracepoint(list, &parse_state->idx, $1.sys, $1.event, + err = parse_events_add_tracepoint(parse_state, list, $1.sys, $1.event, error, $2, &@1); parse_events_terms__delete($2); -- cgit v1.2.3 From a2a6604e1c5836ff4121f170af1328c6a141cca4 Mon Sep 17 00:00:00 2001 From: Dominique Martinet Date: Fri, 10 May 2024 07:24:31 +0900 Subject: perf parse-events: Add new 'fake_tp' parameter for tests The next commit will allow tracepoints starting with digits, but most systems do not have any available by default so tests should skip the actual "check if it exists in /sys/kernel/debug/tracing" step. In order to do that, add a new boolean flag specifying if we should actually "format" the probe or not. Originally-by: Jiri Olsa Reviewed-by: Ian Rogers Signed-off-by: Dominique Martinet Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240510-perf_digit-v4-2-db1553f3233b@codewreck.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/tests/parse-events.c | 6 ++++-- tools/perf/tests/pmu-events.c | 2 +- tools/perf/util/evlist.c | 3 ++- tools/perf/util/evsel.c | 20 +++++++++++++------- tools/perf/util/evsel.h | 4 ++-- tools/perf/util/metricgroup.c | 3 ++- tools/perf/util/parse-events.c | 9 ++++++--- tools/perf/util/parse-events.h | 6 ++++-- 8 files changed, 34 insertions(+), 19 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/tests/parse-events.c b/tools/perf/tests/parse-events.c index 993e482f094c..c3b77570bb57 100644 --- a/tools/perf/tests/parse-events.c +++ b/tools/perf/tests/parse-events.c @@ -2493,7 +2493,8 @@ static int test_event(const struct evlist_test *e) return TEST_FAIL; } parse_events_error__init(&err); - ret = parse_events(evlist, e->name, &err); + ret = __parse_events(evlist, e->name, /*pmu_filter=*/NULL, &err, /*fake_pmu=*/NULL, + /*warn_if_reordered=*/true, /*fake_tp=*/true); if (ret) { pr_debug("failed to parse event '%s', err %d\n", e->name, ret); parse_events_error__print(&err, e->name); @@ -2521,7 +2522,8 @@ static int test_event_fake_pmu(const char *str) parse_events_error__init(&err); ret = __parse_events(evlist, str, /*pmu_filter=*/NULL, &err, - &perf_pmu__fake, /*warn_if_reordered=*/true); + &perf_pmu__fake, /*warn_if_reordered=*/true, + /*fake_tp=*/true); if (ret) { pr_debug("failed to parse event '%s', err %d\n", str, ret); diff --git a/tools/perf/tests/pmu-events.c b/tools/perf/tests/pmu-events.c index f39aadacd7a6..ff3e7bc0a77f 100644 --- a/tools/perf/tests/pmu-events.c +++ b/tools/perf/tests/pmu-events.c @@ -842,7 +842,7 @@ static int check_parse_id(const char *id, struct parse_events_error *error, *cur = '/'; ret = __parse_events(evlist, dup, /*pmu_filter=*/NULL, error, fake_pmu, - /*warn_if_reordered=*/true); + /*warn_if_reordered=*/true, /*fake_tp=*/false); free(dup); evlist__delete(evlist); diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c index 55a300a0977b..3a719edafc7a 100644 --- a/tools/perf/util/evlist.c +++ b/tools/perf/util/evlist.c @@ -298,7 +298,8 @@ struct evsel *evlist__add_aux_dummy(struct evlist *evlist, bool system_wide) #ifdef HAVE_LIBTRACEEVENT struct evsel *evlist__add_sched_switch(struct evlist *evlist, bool system_wide) { - struct evsel *evsel = evsel__newtp_idx("sched", "sched_switch", 0); + struct evsel *evsel = evsel__newtp_idx("sched", "sched_switch", 0, + /*format=*/true); if (IS_ERR(evsel)) return evsel; diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c index 3536404e9447..4f818ab6b662 100644 --- a/tools/perf/util/evsel.c +++ b/tools/perf/util/evsel.c @@ -452,7 +452,7 @@ out_err: * Returns pointer with encoded error via interface. */ #ifdef HAVE_LIBTRACEEVENT -struct evsel *evsel__newtp_idx(const char *sys, const char *name, int idx) +struct evsel *evsel__newtp_idx(const char *sys, const char *name, int idx, bool format) { struct evsel *evsel = zalloc(perf_evsel__object.size); int err = -ENOMEM; @@ -469,14 +469,20 @@ struct evsel *evsel__newtp_idx(const char *sys, const char *name, int idx) if (asprintf(&evsel->name, "%s:%s", sys, name) < 0) goto out_free; - evsel->tp_format = trace_event__tp_format(sys, name); - if (IS_ERR(evsel->tp_format)) { - err = PTR_ERR(evsel->tp_format); - goto out_free; + event_attr_init(&attr); + + if (format) { + evsel->tp_format = trace_event__tp_format(sys, name); + if (IS_ERR(evsel->tp_format)) { + err = PTR_ERR(evsel->tp_format); + goto out_free; + } + attr.config = evsel->tp_format->id; + } else { + attr.config = (__u64) -1; } - event_attr_init(&attr); - attr.config = evsel->tp_format->id; + attr.sample_period = 1; evsel__init(evsel, &attr, idx); } diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h index 517cff431de2..375a38e15cd9 100644 --- a/tools/perf/util/evsel.h +++ b/tools/perf/util/evsel.h @@ -234,14 +234,14 @@ void free_config_terms(struct list_head *config_terms); #ifdef HAVE_LIBTRACEEVENT -struct evsel *evsel__newtp_idx(const char *sys, const char *name, int idx); +struct evsel *evsel__newtp_idx(const char *sys, const char *name, int idx, bool format); /* * Returns pointer with encoded error via interface. */ static inline struct evsel *evsel__newtp(const char *sys, const char *name) { - return evsel__newtp_idx(sys, name, 0); + return evsel__newtp_idx(sys, name, 0, true); } #endif diff --git a/tools/perf/util/metricgroup.c b/tools/perf/util/metricgroup.c index 9be406524617..69f6a46402c3 100644 --- a/tools/perf/util/metricgroup.c +++ b/tools/perf/util/metricgroup.c @@ -1502,7 +1502,8 @@ static int parse_ids(bool metric_no_merge, struct perf_pmu *fake_pmu, pr_debug("Parsing metric events '%s'\n", events.buf); parse_events_error__init(&parse_error); ret = __parse_events(parsed_evlist, events.buf, /*pmu_filter=*/NULL, - &parse_error, fake_pmu, /*warn_if_reordered=*/false); + &parse_error, fake_pmu, /*warn_if_reordered=*/false, + /*fake_tp=*/false); if (ret) { parse_events_error__print(&parse_error, events.buf); goto err_out; diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index e6a2a80b02df..30f958069076 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -550,7 +550,8 @@ static int add_tracepoint(struct parse_events_state *parse_state, struct parse_events_terms *head_config, void *loc_) { YYLTYPE *loc = loc_; - struct evsel *evsel = evsel__newtp_idx(sys_name, evt_name, parse_state->idx++); + struct evsel *evsel = evsel__newtp_idx(sys_name, evt_name, parse_state->idx++, + !parse_state->fake_tp); if (IS_ERR(evsel)) { tracepoint_error(err, PTR_ERR(evsel), sys_name, evt_name, loc->first_column); @@ -2135,7 +2136,7 @@ static int parse_events__sort_events_and_fix_groups(struct list_head *list) int __parse_events(struct evlist *evlist, const char *str, const char *pmu_filter, struct parse_events_error *err, struct perf_pmu *fake_pmu, - bool warn_if_reordered) + bool warn_if_reordered, bool fake_tp) { struct parse_events_state parse_state = { .list = LIST_HEAD_INIT(parse_state.list), @@ -2143,6 +2144,7 @@ int __parse_events(struct evlist *evlist, const char *str, const char *pmu_filte .error = err, .stoken = PE_START_EVENTS, .fake_pmu = fake_pmu, + .fake_tp = fake_tp, .pmu_filter = pmu_filter, .match_legacy_cache_terms = true, }; @@ -2352,7 +2354,8 @@ int parse_events_option(const struct option *opt, const char *str, parse_events_error__init(&err); ret = __parse_events(*args->evlistp, str, args->pmu_filter, &err, - /*fake_pmu=*/NULL, /*warn_if_reordered=*/true); + /*fake_pmu=*/NULL, /*warn_if_reordered=*/true, + /*fake_tp=*/false); if (ret) { parse_events_error__print(&err, str); diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index c06984bd3af8..f2baa69fff98 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -32,14 +32,14 @@ int parse_events_option_new_evlist(const struct option *opt, const char *str, in __attribute__((nonnull(1, 2, 4))) int __parse_events(struct evlist *evlist, const char *str, const char *pmu_filter, struct parse_events_error *error, struct perf_pmu *fake_pmu, - bool warn_if_reordered); + bool warn_if_reordered, bool fake_tp); __attribute__((nonnull(1, 2, 3))) static inline int parse_events(struct evlist *evlist, const char *str, struct parse_events_error *err) { return __parse_events(evlist, str, /*pmu_filter=*/NULL, err, /*fake_pmu=*/NULL, - /*warn_if_reordered=*/true); + /*warn_if_reordered=*/true, /*fake_tp=*/false); } int parse_event(struct evlist *evlist, const char *str); @@ -152,6 +152,8 @@ struct parse_events_state { int stoken; /* Special fake PMU marker for testing. */ struct perf_pmu *fake_pmu; + /* Skip actual tracepoint processing for testing. */ + bool fake_tp; /* If non-null, when wildcard matching only match the given PMU. */ const char *pmu_filter; /* Should PE_LEGACY_NAME tokens be generated for config terms? */ -- cgit v1.2.3 From 5ceb57990bf41684e9bc186128a07025adb896bd Mon Sep 17 00:00:00 2001 From: Dominique Martinet Date: Fri, 10 May 2024 07:24:32 +0900 Subject: perf parse: Allow tracepoint names to start with digits Tracepoints can start with digits, although we don't have many of these: $ rg -g '*.h' '\bTRACE_EVENT\([0-9]' net/mac802154/trace.h 53:TRACE_EVENT(802154_drv_return_int, ... net/ieee802154/trace.h 66:TRACE_EVENT(802154_rdev_add_virtual_intf, ... include/trace/events/9p.h 124:TRACE_EVENT(9p_client_req, ... Just allow names to start with digits too so e.g. "perf trace -e '9p:*'" works Reviewed-by: Ian Rogers Signed-off-by: Dominique Martinet Tested-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240510-perf_digit-v4-3-db1553f3233b@codewreck.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/tests/parse-events.c | 7 +++++++ tools/perf/util/parse-events.l | 4 ++-- 2 files changed, 9 insertions(+), 2 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/tests/parse-events.c b/tools/perf/tests/parse-events.c index c3b77570bb57..edc2adcf1bae 100644 --- a/tools/perf/tests/parse-events.c +++ b/tools/perf/tests/parse-events.c @@ -2269,6 +2269,13 @@ static const struct evlist_test test__events[] = { .check = test__checkevent_breakpoint_2_events, /* 3 */ }, +#ifdef HAVE_LIBTRACEEVENT + { + .name = "9p:9p_client_req", + .check = test__checkevent_tracepoint, + /* 4 */ + }, +#endif }; static const struct evlist_test test__events_pmu[] = { diff --git a/tools/perf/util/parse-events.l b/tools/perf/util/parse-events.l index 08ea2d845dc3..99d585d272e0 100644 --- a/tools/perf/util/parse-events.l +++ b/tools/perf/util/parse-events.l @@ -242,8 +242,8 @@ event [^,{}/]+ num_dec [0-9]+ num_hex 0x[a-fA-F0-9]{1,16} num_raw_hex [a-fA-F0-9]{1,16} -name [a-zA-Z_*?\[\]][a-zA-Z0-9_*?.\[\]!\-]* -name_tag [\'][a-zA-Z_*?\[\]][a-zA-Z0-9_*?\-,\.\[\]:=]*[\'] +name [a-zA-Z0-9_*?\[\]][a-zA-Z0-9_*?.\[\]!\-]* +name_tag [\'][a-zA-Z0-9_*?\[\]][a-zA-Z0-9_*?\-,\.\[\]:=]*[\'] name_minus [a-zA-Z_*?][a-zA-Z0-9\-_*?.:]* drv_cfg_term [a-zA-Z0-9_\.]+(=[a-zA-Z0-9_*?\.:]+)? /* -- cgit v1.2.3 From e2eeef290c4adad7a0f95c4a41e1a992326a7829 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 9 May 2024 11:22:35 -0700 Subject: perf tools: Ignore deleted cgroups On large systems, cgroups can be created and deleted often. That means there's a race between perf tools and cgroups when it gets the cgroup name and opens the cgroup. I got a report that 'perf stat' with many cgroups failed quite often due to the missing cgroups on such a large machine. I think we can ignore such cgroups when expanding events and use id 0 if it fails to read the cgroup id. IIUC 0 is not a vaild cgroup id so it won't update event counts for the failed cgroups. Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240509182235.2319599-1-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/bpf_counter_cgroup.c | 5 ++--- tools/perf/util/cgroup.c | 4 +++- 2 files changed, 5 insertions(+), 4 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/bpf_counter_cgroup.c b/tools/perf/util/bpf_counter_cgroup.c index 1c82377ed78b..ea29c372f339 100644 --- a/tools/perf/util/bpf_counter_cgroup.c +++ b/tools/perf/util/bpf_counter_cgroup.c @@ -136,9 +136,8 @@ static int bperf_load_program(struct evlist *evlist) cgrp = evsel->cgrp; if (read_cgroup_id(cgrp) < 0) { - pr_err("Failed to get cgroup id\n"); - err = -1; - goto out; + pr_debug("Failed to get cgroup id for %s\n", cgrp->name); + cgrp->id = 0; } map_fd = bpf_map__fd(skel->maps.cgrp_idx); diff --git a/tools/perf/util/cgroup.c b/tools/perf/util/cgroup.c index fcb509058499..0f759dd96db7 100644 --- a/tools/perf/util/cgroup.c +++ b/tools/perf/util/cgroup.c @@ -465,9 +465,11 @@ int evlist__expand_cgroup(struct evlist *evlist, const char *str, name = cn->name + prefix_len; if (name[0] == '/' && name[1]) name++; + + /* the cgroup can go away in the meantime */ cgrp = cgroup__new(name, open_cgroup); if (cgrp == NULL) - goto out_err; + continue; leader = NULL; evlist__for_each_entry(orig_list, pos) { -- cgit v1.2.3 From 9ef30265a483f0405e4f7b3f15cda251b9a2c7da Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 10 May 2024 14:04:51 -0700 Subject: perf annotate: Fix segfault on sample histogram A symbol can have no samples, then accessing the annotated_source->samples hashmap will result in a segfault. Fixes: a3f7768bcf48281d ("perf annotate: Fix memory leak in annotated_source") Reviewed-by: Ian Rogers Signed-off-by: Namhyung Kim Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240510210452.2449944-1-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 541988cf6e19..1451caf25e77 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -113,10 +113,11 @@ static __maybe_unused void annotated_source__delete(struct annotated_source *src if (src == NULL) return; - hashmap__for_each_entry(src->samples, cur, bkt) - zfree(&cur->pvalue); - - hashmap__free(src->samples); + if (src->samples) { + hashmap__for_each_entry(src->samples, cur, bkt) + zfree(&cur->pvalue); + hashmap__free(src->samples); + } zfree(&src->histograms); free(src); } -- cgit v1.2.3 From 2af1280b190c408bd590704806dd0d2d1cf52db5 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 10 May 2024 14:04:52 -0700 Subject: perf annotate-data: Ensure the number of type histograms Arnaldo reported that there is a case where nr_histograms and histograms don't agree each other. It ended up in a segfault trying to access a NULL histograms array. Let's make sure to update the nr_histograms when the histograms array is changed. Reported-by: Arnaldo Carvalho de Melo Reviewed-by: Ian Rogers Signed-off-by: Namhyung Kim Tested-by: Arnaldo Carvalho de Melo Cc: Adrian Hunter Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240510210452.2449944-2-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/annotate-data.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c index 57e7d4b3550b..965da6c0b542 100644 --- a/tools/perf/util/annotate-data.c +++ b/tools/perf/util/annotate-data.c @@ -1800,7 +1800,6 @@ static int alloc_data_type_histograms(struct annotated_data_type *adt, int nr_en sz += sizeof(struct type_hist_entry) * adt->self.size; /* Allocate a table of pointers for each event */ - adt->nr_histograms = nr_entries; adt->histograms = calloc(nr_entries, sizeof(*adt->histograms)); if (adt->histograms == NULL) return -ENOMEM; @@ -1814,6 +1813,8 @@ static int alloc_data_type_histograms(struct annotated_data_type *adt, int nr_en if (adt->histograms[i] == NULL) goto err; } + + adt->nr_histograms = nr_entries; return 0; err: @@ -1827,7 +1828,9 @@ static void delete_data_type_histograms(struct annotated_data_type *adt) { for (int i = 0; i < adt->nr_histograms; i++) zfree(&(adt->histograms[i])); + zfree(&adt->histograms); + adt->nr_histograms = 0; } void annotated_data_type__tree_delete(struct rb_root *root) -- cgit v1.2.3 From 193a9e30207f54777ff42d0d8be8389edc522277 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 9 May 2024 22:13:09 -0700 Subject: perf stat: Don't display metric header for non-leader uncore events On an Intel tigerlake laptop a metric like: { "BriefDescription": "Test", "MetricExpr": "imc_free_running@data_read@ + imc_free_running@data_write@", "MetricGroup": "Test", "MetricName": "Test", "ScaleUnit": "6.103515625e-5MiB" }, Will have 4 events: uncore_imc_free_running_0/data_read/ uncore_imc_free_running_0/data_write/ uncore_imc_free_running_1/data_read/ uncore_imc_free_running_1/data_write/ If aggregration is disabled with metric-only 2 column headers are needed: $ perf stat -M test --metric-only -A -a sleep 1 Performance counter stats for 'system wide': MiB Test MiB Test CPU0 1821.0 1820.5 But when not, the counts aggregated in the metric leader and only 1 column should be shown: $ perf stat -M test --metric-only -a sleep 1 Performance counter stats for 'system wide': MiB Test 5909.4 1.001258915 seconds time elapsed Achieve this by skipping events that aren't metric leaders when printing column headers and aggregation isn't disabled. The bug is long standing, the fixes tag is set to a refactor as that is as far back as is reasonable to backport. Fixes: 088519f318be3a41 ("perf stat: Move the display functions to stat-display.c") Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kaige Ye Cc: Kan Liang Cc: K Prateek Nayak Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Yicong Yang Link: https://lore.kernel.org/r/20240510051309.2452468-1-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/stat-display.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'tools/perf/util') diff --git a/tools/perf/util/stat-display.c b/tools/perf/util/stat-display.c index bfc1d705f437..91d2f7f65df7 100644 --- a/tools/perf/util/stat-display.c +++ b/tools/perf/util/stat-display.c @@ -1223,6 +1223,9 @@ static void print_metric_headers(struct perf_stat_config *config, /* Print metrics headers only */ evlist__for_each_entry(evlist, counter) { + if (config->aggr_mode != AGGR_NONE && counter->metric_leader != counter) + continue; + os.evsel = counter; perf_stat__print_shadow_stats(config, counter, 0, -- cgit v1.2.3 From d9c5f5f94c2d356fdf3503f7fcaf254512bc032d Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Fri, 10 May 2024 17:36:01 -0700 Subject: perf pmu: Count sys and cpuid JSON events separately Sys events are eagerly loaded as each event has a compat option that may mean the event is or isn't associated with the PMU. These shouldn't be counted as loaded_json_events as that is used for JSON events matching the CPUID that may or may not have been loaded. The mismatch causes issues on ARM64 that uses sys events. Fixes: e6ff1eed3584362d ("perf pmu: Lazily add JSON events") Closes: https://lore.kernel.org/lkml/20240510024729.1075732-1-justin.he@arm.com/ Reported-by: Jia He Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: John Garry Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20240511003601.2666907-1-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/pmu.c | 70 +++++++++++++++++++++++++++++++++++---------------- tools/perf/util/pmu.h | 6 +++-- 2 files changed, 53 insertions(+), 23 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index b3b072feef02..888ce9912275 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -36,6 +36,18 @@ struct perf_pmu perf_pmu__fake = { #define UNIT_MAX_LEN 31 /* max length for event unit name */ +enum event_source { + /* An event loaded from /sys/devices//events. */ + EVENT_SRC_SYSFS, + /* An event loaded from a CPUID matched json file. */ + EVENT_SRC_CPU_JSON, + /* + * An event loaded from a /sys/devices//identifier matched json + * file. + */ + EVENT_SRC_SYS_JSON, +}; + /** * struct perf_pmu_alias - An event either read from sysfs or builtin in * pmu-events.c, created by parsing the pmu-events json files. @@ -521,7 +533,7 @@ static int update_alias(const struct pmu_event *pe, static int perf_pmu__new_alias(struct perf_pmu *pmu, const char *name, const char *desc, const char *val, FILE *val_fd, - const struct pmu_event *pe) + const struct pmu_event *pe, enum event_source src) { struct perf_pmu_alias *alias; int ret; @@ -574,25 +586,30 @@ static int perf_pmu__new_alias(struct perf_pmu *pmu, const char *name, } snprintf(alias->unit, sizeof(alias->unit), "%s", unit); } - if (!pe) { - /* Update an event from sysfs with json data. */ - struct update_alias_data data = { - .pmu = pmu, - .alias = alias, - }; - + switch (src) { + default: + case EVENT_SRC_SYSFS: alias->from_sysfs = true; if (pmu->events_table) { + /* Update an event from sysfs with json data. */ + struct update_alias_data data = { + .pmu = pmu, + .alias = alias, + }; if (pmu_events_table__find_event(pmu->events_table, pmu, name, update_alias, &data) == 0) - pmu->loaded_json_aliases++; + pmu->cpu_json_aliases++; } - } - - if (!pe) pmu->sysfs_aliases++; - else - pmu->loaded_json_aliases++; + break; + case EVENT_SRC_CPU_JSON: + pmu->cpu_json_aliases++; + break; + case EVENT_SRC_SYS_JSON: + pmu->sys_json_aliases++; + break; + + } list_add_tail(&alias->list, &pmu->aliases); return 0; } @@ -653,7 +670,8 @@ static int __pmu_aliases_parse(struct perf_pmu *pmu, int events_dir_fd) } if (perf_pmu__new_alias(pmu, name, /*desc=*/ NULL, - /*val=*/ NULL, file, /*pe=*/ NULL) < 0) + /*val=*/ NULL, file, /*pe=*/ NULL, + EVENT_SRC_SYSFS) < 0) pr_debug("Cannot set up %s\n", name); fclose(file); } @@ -946,7 +964,8 @@ static int pmu_add_cpu_aliases_map_callback(const struct pmu_event *pe, { struct perf_pmu *pmu = vdata; - perf_pmu__new_alias(pmu, pe->name, pe->desc, pe->event, /*val_fd=*/ NULL, pe); + perf_pmu__new_alias(pmu, pe->name, pe->desc, pe->event, /*val_fd=*/ NULL, + pe, EVENT_SRC_CPU_JSON); return 0; } @@ -981,13 +1000,14 @@ static int pmu_add_sys_aliases_iter_fn(const struct pmu_event *pe, return 0; if (pmu_uncore_alias_match(pe->pmu, pmu->name) && - pmu_uncore_identifier_match(pe->compat, pmu->id)) { + pmu_uncore_identifier_match(pe->compat, pmu->id)) { perf_pmu__new_alias(pmu, pe->name, pe->desc, pe->event, /*val_fd=*/ NULL, - pe); + pe, + EVENT_SRC_SYS_JSON); } return 0; @@ -1082,6 +1102,12 @@ struct perf_pmu *perf_pmu__lookup(struct list_head *pmus, int dirfd, const char pmu->max_precise = pmu_max_precise(dirfd, pmu); pmu->alias_name = pmu_find_alias_name(pmu, dirfd); pmu->events_table = perf_pmu__find_events_table(pmu); + /* + * Load the sys json events/aliases when loading the PMU as each event + * may have a different compat regular expression. We therefore can't + * know the number of sys json events/aliases without computing the + * regular expressions for them all. + */ pmu_add_sys_aliases(pmu); list_add_tail(&pmu->list, pmus); @@ -1739,12 +1765,14 @@ size_t perf_pmu__num_events(struct perf_pmu *pmu) size_t nr; pmu_aliases_parse(pmu); - nr = pmu->sysfs_aliases; + nr = pmu->sysfs_aliases + pmu->sys_json_aliases;; if (pmu->cpu_aliases_added) - nr += pmu->loaded_json_aliases; + nr += pmu->cpu_json_aliases; else if (pmu->events_table) - nr += pmu_events_table__num_events(pmu->events_table, pmu) - pmu->loaded_json_aliases; + nr += pmu_events_table__num_events(pmu->events_table, pmu) - pmu->cpu_json_aliases; + else + assert(pmu->cpu_json_aliases == 0); return pmu->selectable ? nr + 1 : nr; } diff --git a/tools/perf/util/pmu.h b/tools/perf/util/pmu.h index 561716aa2b25..b2d3fd291f02 100644 --- a/tools/perf/util/pmu.h +++ b/tools/perf/util/pmu.h @@ -123,8 +123,10 @@ struct perf_pmu { const struct pmu_events_table *events_table; /** @sysfs_aliases: Number of sysfs aliases loaded. */ uint32_t sysfs_aliases; - /** @sysfs_aliases: Number of json event aliases loaded. */ - uint32_t loaded_json_aliases; + /** @cpu_json_aliases: Number of json event aliases loaded specific to the CPUID. */ + uint32_t cpu_json_aliases; + /** @sys_json_aliases: Number of json event aliases loaded matching the PMU's identifier. */ + uint32_t sys_json_aliases; /** @sysfs_aliases_loaded: Are sysfs aliases loaded from disk? */ bool sysfs_aliases_loaded; /** -- cgit v1.2.3 From 4f1b067359ac8364cdb7f9fda41085fa85789d0f Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 26 May 2024 08:13:21 -0300 Subject: Revert "perf parse-events: Prefer sysfs/JSON hardware events over legacy" This reverts commit 617824a7f0f73e4de325cf8add58e55b28c12493. This made a simple 'perf record -e cycles:pp make -j199' stop working on the Ampere ARM64 system Linus uses to test ARM64 kernels, as discussed at length in the threads in the Link tags below. The fix provided by Ian wasn't acceptable and work to fix this will take time we don't have at this point, so lets revert this and work on it on the next devel cycle. Reported-by: Linus Torvalds Cc: Adrian Hunter Cc: Bhaskar Chowdhury Cc: Ethan Adams Cc: Ian Rogers Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Kan Liang Cc: Namhyung Kim Cc: Thomas Gleixner Cc: Thomas Richter Cc: Tycho Andersen Cc: Yang Jihong Link: https://lore.kernel.org/lkml/CAHk-=wi5Ri=yR2jBVk-4HzTzpoAWOgstr1LEvg_-OXtJvXXJOA@mail.gmail.com Link: https://lore.kernel.org/lkml/CAHk-=wiWvtFyedDNpoV7a8Fq_FpbB+F5KmWK2xPY3QoYseOf_A@mail.gmail.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 31 +++++------------ tools/perf/util/parse-events.h | 2 +- tools/perf/util/parse-events.l | 76 +++++++++++++++++++++--------------------- tools/perf/util/parse-events.y | 62 ++++++++++++---------------------- 4 files changed, 68 insertions(+), 103 deletions(-) (limited to 'tools/perf/util') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 30f958069076..6ed0f9c5581d 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1549,7 +1549,7 @@ static int parse_events_add_pmu(struct parse_events_state *parse_state, } int parse_events_multi_pmu_add(struct parse_events_state *parse_state, - const char *event_name, u64 hw_config, + const char *event_name, const struct parse_events_terms *const_parsed_terms, struct list_head **listp, void *loc_) { @@ -1557,8 +1557,8 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, struct list_head *list = NULL; struct perf_pmu *pmu = NULL; YYLTYPE *loc = loc_; - int ok = 0, core_ok = 0; - const char *tmp; + int ok = 0; + const char *config; struct parse_events_terms parsed_terms; *listp = NULL; @@ -1571,15 +1571,15 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, return ret; } - tmp = strdup(event_name); - if (!tmp) + config = strdup(event_name); + if (!config) goto out_err; if (parse_events_term__num(&term, PARSE_EVENTS__TERM_TYPE_USER, - tmp, /*num=*/1, /*novalue=*/true, + config, /*num=*/1, /*novalue=*/true, loc, /*loc_val=*/NULL) < 0) { - zfree(&tmp); + zfree(&config); goto out_err; } list_add_tail(&term->list, &parsed_terms.terms); @@ -1610,8 +1610,6 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, pr_debug("%s -> %s/%s/\n", event_name, pmu->name, sb.buf); strbuf_release(&sb); ok++; - if (pmu->is_core) - core_ok++; } } @@ -1628,18 +1626,6 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, } } - if (hw_config != PERF_COUNT_HW_MAX && !core_ok) { - /* - * The event wasn't found on core PMUs but it has a hardware - * config version to try. - */ - if (!parse_events_add_numeric(parse_state, list, - PERF_TYPE_HARDWARE, hw_config, - const_parsed_terms, - /*wildcard=*/true)) - ok++; - } - out_err: parse_events_terms__exit(&parsed_terms); if (ok) @@ -1693,8 +1679,7 @@ int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state /* Failure to add, assume event_or_pmu is an event name. */ zfree(listp); - if (!parse_events_multi_pmu_add(parse_state, event_or_pmu, PERF_COUNT_HW_MAX, - const_parsed_terms, listp, loc)) + if (!parse_events_multi_pmu_add(parse_state, event_or_pmu, const_parsed_terms, listp, loc)) return 0; if (asprintf(&help, "Unable to find PMU or event on a PMU of '%s'", event_or_pmu) < 0) diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index f2baa69fff98..e13de2c8b706 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -237,7 +237,7 @@ struct evsel *parse_events__add_event(int idx, struct perf_event_attr *attr, struct perf_pmu *pmu); int parse_events_multi_pmu_add(struct parse_events_state *parse_state, - const char *event_name, u64 hw_config, + const char *event_name, const struct parse_events_terms *const_parsed_terms, struct list_head **listp, void *loc); diff --git a/tools/perf/util/parse-events.l b/tools/perf/util/parse-events.l index 99d585d272e0..16045c383ada 100644 --- a/tools/perf/util/parse-events.l +++ b/tools/perf/util/parse-events.l @@ -113,12 +113,12 @@ do { \ yyless(0); \ } while (0) -static int sym(yyscan_t scanner, int config) +static int sym(yyscan_t scanner, int type, int config) { YYSTYPE *yylval = parse_events_get_lval(scanner); - yylval->num = config; - return PE_VALUE_SYM_SW; + yylval->num = (type << 16) + config; + return type == PERF_TYPE_HARDWARE ? PE_VALUE_SYM_HW : PE_VALUE_SYM_SW; } static int tool(yyscan_t scanner, enum perf_tool_event event) @@ -137,13 +137,13 @@ static int term(yyscan_t scanner, enum parse_events__term_type type) return PE_TERM; } -static int hw(yyscan_t scanner, int config) +static int hw_term(yyscan_t scanner, int config) { YYSTYPE *yylval = parse_events_get_lval(scanner); char *text = parse_events_get_text(scanner); - yylval->hardware_event.str = strdup(text); - yylval->hardware_event.num = config; + yylval->hardware_term.str = strdup(text); + yylval->hardware_term.num = PERF_TYPE_HARDWARE + config; return PE_TERM_HW; } @@ -330,16 +330,16 @@ percore { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_PERCORE); } aux-output { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_AUX_OUTPUT); } aux-sample-size { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_AUX_SAMPLE_SIZE); } metric-id { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_METRIC_ID); } -cpu-cycles|cycles { return hw(yyscanner, PERF_COUNT_HW_CPU_CYCLES); } -stalled-cycles-frontend|idle-cycles-frontend { return hw(yyscanner, PERF_COUNT_HW_STALLED_CYCLES_FRONTEND); } -stalled-cycles-backend|idle-cycles-backend { return hw(yyscanner, PERF_COUNT_HW_STALLED_CYCLES_BACKEND); } -instructions { return hw(yyscanner, PERF_COUNT_HW_INSTRUCTIONS); } -cache-references { return hw(yyscanner, PERF_COUNT_HW_CACHE_REFERENCES); } -cache-misses { return hw(yyscanner, PERF_COUNT_HW_CACHE_MISSES); } -branch-instructions|branches { return hw(yyscanner, PERF_COUNT_HW_BRANCH_INSTRUCTIONS); } -branch-misses { return hw(yyscanner, PERF_COUNT_HW_BRANCH_MISSES); } -bus-cycles { return hw(yyscanner, PERF_COUNT_HW_BUS_CYCLES); } -ref-cycles { return hw(yyscanner, PERF_COUNT_HW_REF_CPU_CYCLES); } +cpu-cycles|cycles { return hw_term(yyscanner, PERF_COUNT_HW_CPU_CYCLES); } +stalled-cycles-frontend|idle-cycles-frontend { return hw_term(yyscanner, PERF_COUNT_HW_STALLED_CYCLES_FRONTEND); } +stalled-cycles-backend|idle-cycles-backend { return hw_term(yyscanner, PERF_COUNT_HW_STALLED_CYCLES_BACKEND); } +instructions { return hw_term(yyscanner, PERF_COUNT_HW_INSTRUCTIONS); } +cache-references { return hw_term(yyscanner, PERF_COUNT_HW_CACHE_REFERENCES); } +cache-misses { return hw_term(yyscanner, PERF_COUNT_HW_CACHE_MISSES); } +branch-instructions|branches { return hw_term(yyscanner, PERF_COUNT_HW_BRANCH_INSTRUCTIONS); } +branch-misses { return hw_term(yyscanner, PERF_COUNT_HW_BRANCH_MISSES); } +bus-cycles { return hw_term(yyscanner, PERF_COUNT_HW_BUS_CYCLES); } +ref-cycles { return hw_term(yyscanner, PERF_COUNT_HW_REF_CPU_CYCLES); } r{num_raw_hex} { return str(yyscanner, PE_RAW); } r0x{num_raw_hex} { return str(yyscanner, PE_RAW); } , { return ','; } @@ -383,31 +383,31 @@ r0x{num_raw_hex} { return str(yyscanner, PE_RAW); } <> { BEGIN(INITIAL); } } -cpu-cycles|cycles { return hw(yyscanner, PERF_COUNT_HW_CPU_CYCLES); } -stalled-cycles-frontend|idle-cycles-frontend { return hw(yyscanner, PERF_COUNT_HW_STALLED_CYCLES_FRONTEND); } -stalled-cycles-backend|idle-cycles-backend { return hw(yyscanner, PERF_COUNT_HW_STALLED_CYCLES_BACKEND); } -instructions { return hw(yyscanner, PERF_COUNT_HW_INSTRUCTIONS); } -cache-references { return hw(yyscanner, PERF_COUNT_HW_CACHE_REFERENCES); } -cache-misses { return hw(yyscanner, PERF_COUNT_HW_CACHE_MISSES); } -branch-instructions|branches { return hw(yyscanner, PERF_COUNT_HW_BRANCH_INSTRUCTIONS); } -branch-misses { return hw(yyscanner, PERF_COUNT_HW_BRANCH_MISSES); } -bus-cycles { return hw(yyscanner, PERF_COUNT_HW_BUS_CYCLES); } -ref-cycles { return hw(yyscanner, PERF_COUNT_HW_REF_CPU_CYCLES); } -cpu-clock { return sym(yyscanner, PERF_COUNT_SW_CPU_CLOCK); } -task-clock { return sym(yyscanner, PERF_COUNT_SW_TASK_CLOCK); } -page-faults|faults { return sym(yyscanner, PERF_COUNT_SW_PAGE_FAULTS); } -minor-faults { return sym(yyscanner, PERF_COUNT_SW_PAGE_FAULTS_MIN); } -major-faults { return sym(yyscanner, PERF_COUNT_SW_PAGE_FAULTS_MAJ); } -context-switches|cs { return sym(yyscanner, PERF_COUNT_SW_CONTEXT_SWITCHES); } -cpu-migrations|migrations { return sym(yyscanner, PERF_COUNT_SW_CPU_MIGRATIONS); } -alignment-faults { return sym(yyscanner, PERF_COUNT_SW_ALIGNMENT_FAULTS); } -emulation-faults { return sym(yyscanner, PERF_COUNT_SW_EMULATION_FAULTS); } -dummy { return sym(yyscanner, PERF_COUNT_SW_DUMMY); } +cpu-cycles|cycles { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_CPU_CYCLES); } +stalled-cycles-frontend|idle-cycles-frontend { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_STALLED_CYCLES_FRONTEND); } +stalled-cycles-backend|idle-cycles-backend { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_STALLED_CYCLES_BACKEND); } +instructions { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_INSTRUCTIONS); } +cache-references { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_CACHE_REFERENCES); } +cache-misses { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_CACHE_MISSES); } +branch-instructions|branches { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_BRANCH_INSTRUCTIONS); } +branch-misses { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_BRANCH_MISSES); } +bus-cycles { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_BUS_CYCLES); } +ref-cycles { return sym(yyscanner, PERF_TYPE_HARDWARE, PERF_COUNT_HW_REF_CPU_CYCLES); } +cpu-clock { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_CPU_CLOCK); } +task-clock { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_TASK_CLOCK); } +page-faults|faults { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_PAGE_FAULTS); } +minor-faults { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_PAGE_FAULTS_MIN); } +major-faults { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_PAGE_FAULTS_MAJ); } +context-switches|cs { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_CONTEXT_SWITCHES); } +cpu-migrations|migrations { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_CPU_MIGRATIONS); } +alignment-faults { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_ALIGNMENT_FAULTS); } +emulation-faults { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_EMULATION_FAULTS); } +dummy { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_DUMMY); } duration_time { return tool(yyscanner, PERF_TOOL_DURATION_TIME); } user_time { return tool(yyscanner, PERF_TOOL_USER_TIME); } system_time { return tool(yyscanner, PERF_TOOL_SYSTEM_TIME); } -bpf-output { return sym(yyscanner, PERF_COUNT_SW_BPF_OUTPUT); } -cgroup-switches { return sym(yyscanner, PERF_COUNT_SW_CGROUP_SWITCHES); } +bpf-output { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_BPF_OUTPUT); } +cgroup-switches { return sym(yyscanner, PERF_TYPE_SOFTWARE, PERF_COUNT_SW_CGROUP_SWITCHES); } {lc_type} { return str(yyscanner, PE_LEGACY_CACHE); } {lc_type}-{lc_op_result} { return str(yyscanner, PE_LEGACY_CACHE); } diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y index c94a3994177e..b3c51f06cbdc 100644 --- a/tools/perf/util/parse-events.y +++ b/tools/perf/util/parse-events.y @@ -55,7 +55,7 @@ static void free_list_evsel(struct list_head* list_evsel) %} %token PE_START_EVENTS PE_START_TERMS -%token PE_VALUE PE_VALUE_SYM_SW PE_TERM +%token PE_VALUE PE_VALUE_SYM_HW PE_VALUE_SYM_SW PE_TERM %token PE_VALUE_SYM_TOOL %token PE_EVENT_NAME %token PE_RAW PE_NAME @@ -66,10 +66,12 @@ static void free_list_evsel(struct list_head* list_evsel) %token PE_DRV_CFG_TERM %token PE_TERM_HW %type PE_VALUE +%type PE_VALUE_SYM_HW %type PE_VALUE_SYM_SW %type PE_VALUE_SYM_TOOL %type PE_MODIFIER_EVENT %type PE_TERM +%type value_sym %type PE_RAW %type PE_NAME %type PE_LEGACY_CACHE @@ -85,7 +87,6 @@ static void free_list_evsel(struct list_head* list_evsel) %type opt_pmu_config %destructor { parse_events_terms__delete ($$); } %type event_pmu -%type event_legacy_hardware %type event_legacy_symbol %type event_legacy_cache %type event_legacy_mem @@ -103,8 +104,8 @@ static void free_list_evsel(struct list_head* list_evsel) %destructor { free_list_evsel ($$); } %type tracepoint_name %destructor { free ($$.sys); free ($$.event); } -%type PE_TERM_HW -%destructor { free ($$.str); } +%type PE_TERM_HW +%destructor { free ($$.str); } %union { @@ -119,10 +120,10 @@ static void free_list_evsel(struct list_head* list_evsel) char *sys; char *event; } tracepoint_name; - struct hardware_event { + struct hardware_term { char *str; u64 num; - } hardware_event; + } hardware_term; } %% @@ -265,7 +266,6 @@ PE_EVENT_NAME event_def event_def event_def: event_pmu | - event_legacy_hardware | event_legacy_symbol | event_legacy_cache sep_dc | event_legacy_mem sep_dc | @@ -292,7 +292,7 @@ PE_NAME sep_dc struct list_head *list; int err; - err = parse_events_multi_pmu_add(_parse_state, $1, PERF_COUNT_HW_MAX, NULL, &list, &@1); + err = parse_events_multi_pmu_add(_parse_state, $1, NULL, &list, &@1); if (err < 0) { struct parse_events_state *parse_state = _parse_state; struct parse_events_error *error = parse_state->error; @@ -308,45 +308,24 @@ PE_NAME sep_dc $$ = list; } -event_legacy_hardware: -PE_TERM_HW opt_pmu_config -{ - /* List of created evsels. */ - struct list_head *list = NULL; - int err = parse_events_multi_pmu_add(_parse_state, $1.str, $1.num, $2, &list, &@1); - - free($1.str); - parse_events_terms__delete($2); - if (err) - PE_ABORT(err); - - $$ = list; -} +value_sym: +PE_VALUE_SYM_HW | -PE_TERM_HW sep_dc -{ - struct list_head *list; - int err; - - err = parse_events_multi_pmu_add(_parse_state, $1.str, $1.num, NULL, &list, &@1); - free($1.str); - if (err) - PE_ABORT(err); - $$ = list; -} +PE_VALUE_SYM_SW event_legacy_symbol: -PE_VALUE_SYM_SW '/' event_config '/' +value_sym '/' event_config '/' { struct list_head *list; + int type = $1 >> 16; + int config = $1 & 255; int err; + bool wildcard = (type == PERF_TYPE_HARDWARE || type == PERF_TYPE_HW_CACHE); list = alloc_list(); if (!list) YYNOMEM; - err = parse_events_add_numeric(_parse_state, list, - /*type=*/PERF_TYPE_SOFTWARE, /*config=*/$1, - $3, /*wildcard=*/false); + err = parse_events_add_numeric(_parse_state, list, type, config, $3, wildcard); parse_events_terms__delete($3); if (err) { free_list_evsel(list); @@ -355,17 +334,18 @@ PE_VALUE_SYM_SW '/' event_config '/' $$ = list; } | -PE_VALUE_SYM_SW sep_slash_slash_dc +value_sym sep_slash_slash_dc { struct list_head *list; + int type = $1 >> 16; + int config = $1 & 255; + bool wildcard = (type == PERF_TYPE_HARDWARE || type == PERF_TYPE_HW_CACHE); int err; list = alloc_list(); if (!list) YYNOMEM; - err = parse_events_add_numeric(_parse_state, list, - /*type=*/PERF_TYPE_SOFTWARE, /*config=*/$1, - /*head_config=*/NULL, /*wildcard=*/false); + err = parse_events_add_numeric(_parse_state, list, type, config, /*head_config=*/NULL, wildcard); if (err) PE_ABORT(err); $$ = list; -- cgit v1.2.3