user/sven/linux.git/drivers/perf, branch v5.4.3

Merge branches 'for-next/52-bit-kva', 'for-next/cpu-topology', 'for-next/error-injection', 'for-next/perf', 'for-next/psci-cpuidle', 'for-next/rng', 'for-next/smpboot', 'for-next/tbi' and 'for-next/tlbi' into for-next/core

2019-08-30T11:46:12Z

* for-next/52-bit-kva: (25 commits) Support for 52-bit virtual addressing in kernel space * for-next/cpu-topology: (9 commits) Move CPU topology parsing into core code and add support for ACPI 6.3 * for-next/error-injection: (2 commits) Support for function error injection via kprobes * for-next/perf: (8 commits) Support for i.MX8 DDR PMU and proper SMMUv3 group validation * for-next/psci-cpuidle: (7 commits) Move PSCI idle code into a new CPUidle driver * for-next/rng: (4 commits) Support for 'rng-seed' property being passed in the devicetree * for-next/smpboot: (3 commits) Reduce fragility of secondary CPU bringup in debug configurations * for-next/tbi: (10 commits) Introduce new syscall ABI with relaxed requirements for pointer tags * for-next/tlbi: (6 commits) Handle spurious page faults arising from kernel space

perf/imx_ddr: Add support for AXI ID filtering

2019-08-28T13:16:55Z

AXI filtering is used by events 0x41 and 0x42 to count reads or writes with an ARID or AWID matching a specified filter. The filter is exposed to userspace as an (ID, MASK) pair, where each set bit in the mask causes the corresponding bit in the ID to be ignored when matching against the ID of memory transactions for the purposes of incrementing the counter. For example: # perf stat -a -e imx8_ddr0/axid-read,axi_mask=0xff,axi_id=0x800/ cmd will count all read transactions from AXI IDs 0x800 - 0x8ff. If the 'axi_mask' is omitted, then it is treated as 0x0 which means that the 'axi_id' will be matched exactly. Signed-off-by: Joakim Zhang Signed-off-by: Will Deacon

perf/smmuv3: Validate groups for global filtering

2019-08-27T18:37:04Z

With global filtering, it becomes possible for users to construct self-contradictory groups with conflicting filters. Make sure we cover that when initially validating events. Signed-off-by: Robin Murphy Signed-off-by: Will Deacon

perf/smmuv3: Validate group size

2019-08-27T18:37:04Z

Ensure that a group will actually fit into the available counters. Signed-off-by: Robin Murphy Signed-off-by: Will Deacon

perf: Remove dev_err() usage after platform_get_irq()

2019-07-31T16:15:20Z

We don't need dev_err() messages when platform_get_irq() fails now that platform_get_irq() prints an error message itself when something goes wrong. Let's remove these prints with a simple semantic patch. // @@ expression ret; struct platform_device *E; @@ ret = ( platform_get_irq(E, ...) | platform_get_irq_byname(E, ...) ); if ( \( ret < 0 \| ret <= 0 \) ) { ( -if (ret != -EPROBE_DEFER) -{ ... -dev_err(...); -... } | ... -dev_err(...); ) ... } // While we're here, remove braces on if statements that only have one statement (manually). Cc: Will Deacon Cc: Mark Rutland Cc: linux-arm-kernel@lists.infradead.org Cc: Greg Kroah-Hartman Signed-off-by: Stephen Boyd Signed-off-by: Will Deacon

perf/imx_ddr: Add MODULE_DEVICE_TABLE

2019-07-31T16:15:16Z

This is required for automatic probing when driver is built as a module. Fixes: 9a66d36cc7ac ("drivers/perf: imx_ddr: Add DDR performance counter support to perf") Acked-by: Frank Li Signed-off-by: Leonard Crestez Signed-off-by: Will Deacon

drivers/perf: arm_pmu: Fix failure path in PM notifier

2019-07-29T10:43:48Z

Handling of the CPU_PM_ENTER_FAILED transition in the Arm PMU PM notifier code incorrectly skips restoration of the counters. Fix the logic so that CPU_PM_ENTER_FAILED follows the same path as CPU_PM_EXIT. Cc: Fixes: da4e4f18afe0f372 ("drivers/perf: arm_pmu: implement CPU_PM notifier") Reported-by: Anders Roxell Acked-by: Lorenzo Pieralisi Signed-off-by: Will Deacon

docs: perf: move to the admin-guide

2019-07-15T12:20:27Z

The perf infrastructure is used for userspace to track issues. At least a good part of what's described here is related to it. So, add it to the admin-guide. Signed-off-by: Mauro Carvalho Chehab

docs: perf: convert to ReST

2019-07-15T12:20:26Z

Rename the perf documentation files to ReST, add an index for them and adjust in order to produce a nice html output via the Sphinx build system. At its new index.rst, let's add a :orphan: while this is not linked to the main index.rst file, in order to avoid build warnings. Signed-off-by: Mauro Carvalho Chehab

Merge tag 'arm64-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux

2019-07-08T16:54:55Z

Pull arm64 updates from Catalin Marinas: - arm64 support for syscall emulation via PTRACE_SYSEMU{,_SINGLESTEP} - Wire up VM_FLUSH_RESET_PERMS for arm64, allowing the core code to manage the permissions of executable vmalloc regions more strictly - Slight performance improvement by keeping softirqs enabled while touching the FPSIMD/SVE state (kernel_neon_begin/end) - Expose a couple of ARMv8.5 features to user (HWCAP): CondM (new XAFLAG and AXFLAG instructions for floating point comparison flags manipulation) and FRINT (rounding floating point numbers to integers) - Re-instate ARM64_PSEUDO_NMI support which was previously marked as BROKEN due to some bugs (now fixed) - Improve parking of stopped CPUs and implement an arm64-specific panic_smp_self_stop() to avoid warning on not being able to stop secondary CPUs during panic - perf: enable the ARM Statistical Profiling Extensions (SPE) on ACPI platforms - perf: DDR performance monitor support for iMX8QXP - cache_line_size() can now be set from DT or ACPI/PPTT if provided to cope with a system cache info not exposed via the CPUID registers - Avoid warning on hardware cache line size greater than ARCH_DMA_MINALIGN if the system is fully coherent - arm64 do_page_fault() and hugetlb cleanups - Refactor set_pte_at() to avoid redundant READ_ONCE(*ptep) - Ignore ACPI 5.1 FADTs reported as 5.0 (infer from the 'arm_boot_flags' introduced in 5.1) - CONFIG_RANDOMIZE_BASE now enabled in defconfig - Allow the selection of ARM64_MODULE_PLTS, currently only done via RANDOMIZE_BASE (and an erratum workaround), allowing modules to spill over into the vmalloc area - Make ZONE_DMA32 configurable * tag 'arm64-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux: (54 commits) perf: arm_spe: Enable ACPI/Platform automatic module loading arm_pmu: acpi: spe: Add initial MADT/SPE probing ACPI/PPTT: Add function to return ACPI 6.3 Identical tokens ACPI/PPTT: Modify node flag detection to find last IDENTICAL x86/entry: Simplify _TIF_SYSCALL_EMU handling arm64: rename dump_instr as dump_kernel_instr arm64/mm: Drop [PTE|PMD]_TYPE_FAULT arm64: Implement panic_smp_self_stop() arm64: Improve parking of stopped CPUs arm64: Expose FRINT capabilities to userspace arm64: Expose ARMv8.5 CondM capability to userspace arm64: defconfig: enable CONFIG_RANDOMIZE_BASE arm64: ARM64_MODULES_PLTS must depend on MODULES arm64: bpf: do not allocate executable memory arm64/kprobes: set VM_FLUSH_RESET_PERMS on kprobe instruction pages arm64/mm: wire up CONFIG_ARCH_HAS_SET_DIRECT_MAP arm64: module: create module allocations without exec permissions arm64: Allow user selection of ARM64_MODULE_PLTS acpi/arm64: ignore 5.1 FADTs that are reported as 5.0 arm64: Allow selecting Pseudo-NMI again ...