From be8ee18152b0523752f3a44900363838bd1573bb Mon Sep 17 00:00:00 2001 From: Atul Kumar Pant Date: Sat, 18 Jan 2025 14:09:27 +0530 Subject: sched_ext: Fixes typos in comments Fixes some spelling errors in the comments. Signed-off-by: Atul Kumar Pant Signed-off-by: Tejun Heo --- kernel/sched/ext.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'kernel/sched') diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 8857c0709bdd..283d7f1addc5 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -416,7 +416,7 @@ struct sched_ext_ops { /** * @update_idle: Update the idle state of a CPU - * @cpu: CPU to udpate the idle state for + * @cpu: CPU to update the idle state for * @idle: whether entering or exiting the idle state * * This operation is called when @rq's CPU goes or leaves the idle @@ -1214,7 +1214,7 @@ static bool scx_kf_allowed_if_unlocked(void) /** * nldsq_next_task - Iterate to the next task in a non-local DSQ - * @dsq: user dsq being interated + * @dsq: user dsq being iterated * @cur: current position, %NULL to start iteration * @rev: walk backwards * @@ -2078,7 +2078,7 @@ static void set_task_runnable(struct rq *rq, struct task_struct *p) /* * list_add_tail() must be used. scx_ops_bypass() depends on tasks being - * appened to the runnable_list. + * appended to the runnable_list. */ list_add_tail(&p->scx.runnable_node, &rq->scx.runnable_list); } @@ -2480,7 +2480,7 @@ static struct rq *move_task_between_dsqs(struct task_struct *p, u64 enq_flags, /* * A poorly behaving BPF scheduler can live-lock the system by e.g. incessantly * banging on the same DSQ on a large NUMA system to the point where switching - * to the bypass mode can take a long time. Inject artifical delays while the + * to the bypass mode can take a long time. Inject artificial delays while the * bypass mode is switching to guarantee timely completion. */ static void scx_ops_breather(struct rq *rq) @@ -3144,7 +3144,7 @@ static struct task_struct *pick_task_scx(struct rq *rq) * * Unless overridden by ops.core_sched_before(), @p->scx.core_sched_at is used * to implement the default task ordering. The older the timestamp, the higher - * prority the task - the global FIFO ordering matching the default scheduling + * priority the task - the global FIFO ordering matching the default scheduling * behavior. * * When ops.core_sched_before() is enabled, @p->scx.core_sched_at is used to @@ -4590,7 +4590,7 @@ static int scx_cgroup_init(void) cgroup_warned_missing_idle = false; /* - * scx_tg_on/offline() are excluded thorugh scx_cgroup_rwsem. If we walk + * scx_tg_on/offline() are excluded through scx_cgroup_rwsem. If we walk * cgroups and init, all online cgroups are initialized. */ rcu_read_lock(); -- cgit v1.2.3 From 2279563e3a8cac367b267b09c15cf1e39c06c5cc Mon Sep 17 00:00:00 2001 From: Andrea Righi Date: Wed, 22 Jan 2025 10:05:25 +0100 Subject: sched_ext: Include task weight in the error state dump Report the task weight when dumping the task state during an error exit. Moreover, adjust the output format to display dsq_vtime, slice, and weight on the same line. This can help identify whether certain tasks were excessively prioritized or de-prioritized due to large niceness gaps. Signed-off-by: Andrea Righi Signed-off-by: Tejun Heo --- kernel/sched/ext.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'kernel/sched') diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 283d7f1addc5..7081c7be5f62 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -5277,9 +5277,10 @@ static void scx_dump_task(struct seq_buf *s, struct scx_dump_ctx *dctx, scx_get_task_state(p), p->scx.flags & ~SCX_TASK_STATE_MASK, p->scx.dsq_flags, ops_state & SCX_OPSS_STATE_MASK, ops_state >> SCX_OPSS_QSEQ_SHIFT); - dump_line(s, " sticky/holding_cpu=%d/%d dsq_id=%s dsq_vtime=%llu slice=%llu", - p->scx.sticky_cpu, p->scx.holding_cpu, dsq_id_buf, - p->scx.dsq_vtime, p->scx.slice); + dump_line(s, " sticky/holding_cpu=%d/%d dsq_id=%s", + p->scx.sticky_cpu, p->scx.holding_cpu, dsq_id_buf); + dump_line(s, " dsq_vtime=%llu slice=%llu weight=%u", + p->scx.dsq_vtime, p->scx.slice, p->scx.weight); dump_line(s, " cpus=%*pb", cpumask_pr_args(p->cpus_ptr)); if (SCX_HAS_OP(dump_task)) { -- cgit v1.2.3 From d6f3e7d564b2309e1f17e709a70eca78d7ca2bb8 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 24 Jan 2025 12:22:12 -1000 Subject: sched_ext: Fix incorrect autogroup migration detection scx_move_task() is called from sched_move_task() and tells the BPF scheduler that cgroup migration is being committed. sched_move_task() is used by both cgroup and autogroup migrations and scx_move_task() tried to filter out autogroup migrations by testing the destination cgroup and PF_EXITING but this is not enough. In fact, without explicitly tagging the thread which is doing the cgroup migration, there is no good way to tell apart scx_move_task() invocations for racing migration to the root cgroup and an autogroup migration. This led to scx_move_task() incorrectly ignoring a migration from non-root cgroup to an autogroup of the root cgroup triggering the following warning: WARNING: CPU: 7 PID: 1 at kernel/sched/ext.c:3725 scx_cgroup_can_attach+0x196/0x340 ... Call Trace: cgroup_migrate_execute+0x5b1/0x700 cgroup_attach_task+0x296/0x400 __cgroup_procs_write+0x128/0x140 cgroup_procs_write+0x17/0x30 kernfs_fop_write_iter+0x141/0x1f0 vfs_write+0x31d/0x4a0 __x64_sys_write+0x72/0xf0 do_syscall_64+0x82/0x160 entry_SYSCALL_64_after_hwframe+0x76/0x7e Fix it by adding an argument to sched_move_task() that indicates whether the moving is for a cgroup or autogroup migration. After the change, scx_move_task() is called only for cgroup migrations and renamed to scx_cgroup_move_task(). Link: https://github.com/sched-ext/scx/issues/370 Fixes: 819513666966 ("sched_ext: Add cgroup support") Cc: stable@vger.kernel.org # v6.12+ Acked-by: Peter Zijlstra (Intel) Signed-off-by: Tejun Heo --- kernel/sched/autogroup.c | 4 ++-- kernel/sched/core.c | 7 ++++--- kernel/sched/ext.c | 15 +-------------- kernel/sched/ext.h | 4 ++-- kernel/sched/sched.h | 2 +- 5 files changed, 10 insertions(+), 22 deletions(-) (limited to 'kernel/sched') diff --git a/kernel/sched/autogroup.c b/kernel/sched/autogroup.c index db68a964e34e..c4a3ccf6a8ac 100644 --- a/kernel/sched/autogroup.c +++ b/kernel/sched/autogroup.c @@ -150,7 +150,7 @@ void sched_autogroup_exit_task(struct task_struct *p) * see this thread after that: we can no longer use signal->autogroup. * See the PF_EXITING check in task_wants_autogroup(). */ - sched_move_task(p); + sched_move_task(p, true); } static void @@ -182,7 +182,7 @@ autogroup_move_group(struct task_struct *p, struct autogroup *ag) * sched_autogroup_exit_task(). */ for_each_thread(p, t) - sched_move_task(t); + sched_move_task(t, true); unlock_task_sighand(p, &flags); autogroup_kref_put(prev); diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 901170708e2a..e77897a62442 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -9042,7 +9042,7 @@ static void sched_change_group(struct task_struct *tsk, struct task_group *group * now. This function just updates tsk->se.cfs_rq and tsk->se.parent to reflect * its new group. */ -void sched_move_task(struct task_struct *tsk) +void sched_move_task(struct task_struct *tsk, bool for_autogroup) { int queued, running, queue_flags = DEQUEUE_SAVE | DEQUEUE_MOVE | DEQUEUE_NOCLOCK; @@ -9071,7 +9071,8 @@ void sched_move_task(struct task_struct *tsk) put_prev_task(rq, tsk); sched_change_group(tsk, group); - scx_move_task(tsk); + if (!for_autogroup) + scx_cgroup_move_task(tsk); if (queued) enqueue_task(rq, tsk, queue_flags); @@ -9172,7 +9173,7 @@ static void cpu_cgroup_attach(struct cgroup_taskset *tset) struct cgroup_subsys_state *css; cgroup_taskset_for_each(task, css, tset) - sched_move_task(task); + sched_move_task(task, false); scx_cgroup_finish_attach(); } diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 7081c7be5f62..c7b159f48834 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -4323,24 +4323,11 @@ err: return ops_sanitize_err("cgroup_prep_move", ret); } -void scx_move_task(struct task_struct *p) +void scx_cgroup_move_task(struct task_struct *p) { if (!scx_cgroup_enabled) return; - /* - * We're called from sched_move_task() which handles both cgroup and - * autogroup moves. Ignore the latter. - * - * Also ignore exiting tasks, because in the exit path tasks transition - * from the autogroup to the root group, so task_group_is_autogroup() - * alone isn't able to catch exiting autogroup tasks. This is safe for - * cgroup_move(), because cgroup migrations never happen for PF_EXITING - * tasks. - */ - if (task_group_is_autogroup(task_group(p)) || (p->flags & PF_EXITING)) - return; - /* * @p must have ops.cgroup_prep_move() called on it and thus * cgrp_moving_from set. diff --git a/kernel/sched/ext.h b/kernel/sched/ext.h index 4d022d17ac7d..1079b56b0f7a 100644 --- a/kernel/sched/ext.h +++ b/kernel/sched/ext.h @@ -73,7 +73,7 @@ static inline void scx_update_idle(struct rq *rq, bool idle, bool do_notify) {} int scx_tg_online(struct task_group *tg); void scx_tg_offline(struct task_group *tg); int scx_cgroup_can_attach(struct cgroup_taskset *tset); -void scx_move_task(struct task_struct *p); +void scx_cgroup_move_task(struct task_struct *p); void scx_cgroup_finish_attach(void); void scx_cgroup_cancel_attach(struct cgroup_taskset *tset); void scx_group_set_weight(struct task_group *tg, unsigned long cgrp_weight); @@ -82,7 +82,7 @@ void scx_group_set_idle(struct task_group *tg, bool idle); static inline int scx_tg_online(struct task_group *tg) { return 0; } static inline void scx_tg_offline(struct task_group *tg) {} static inline int scx_cgroup_can_attach(struct cgroup_taskset *tset) { return 0; } -static inline void scx_move_task(struct task_struct *p) {} +static inline void scx_cgroup_move_task(struct task_struct *p) {} static inline void scx_cgroup_finish_attach(void) {} static inline void scx_cgroup_cancel_attach(struct cgroup_taskset *tset) {} static inline void scx_group_set_weight(struct task_group *tg, unsigned long cgrp_weight) {} diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 38e0e323dda2..b93c8c3dc05a 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -572,7 +572,7 @@ extern void sched_online_group(struct task_group *tg, extern void sched_destroy_group(struct task_group *tg); extern void sched_release_group(struct task_group *tg); -extern void sched_move_task(struct task_struct *tsk); +extern void sched_move_task(struct task_struct *tsk, bool for_autogroup); #ifdef CONFIG_FAIR_GROUP_SCHED extern int sched_group_set_shares(struct task_group *tg, unsigned long shares); -- cgit v1.2.3 From 1626e5ef0b00386a4fd083fa7c46c8edbd75f9b4 Mon Sep 17 00:00:00 2001 From: Andrea Righi Date: Mon, 27 Jan 2025 23:06:16 +0100 Subject: sched_ext: Fix lock imbalance in dispatch_to_local_dsq() While performing the rq locking dance in dispatch_to_local_dsq(), we may trigger the following lock imbalance condition, in particular when multiple tasks are rapidly changing CPU affinity (i.e., running a `stress-ng --race-sched 0`): [ 13.413579] ===================================== [ 13.413660] WARNING: bad unlock balance detected! [ 13.413729] 6.13.0-virtme #15 Not tainted [ 13.413792] ------------------------------------- [ 13.413859] kworker/1:1/80 is trying to release lock (&rq->__lock) at: [ 13.413954] [] dispatch_to_local_dsq+0x108/0x1a0 [ 13.414111] but there are no more locks to release! [ 13.414176] [ 13.414176] other info that might help us debug this: [ 13.414258] 1 lock held by kworker/1:1/80: [ 13.414318] #0: ffff8b66feb41698 (&rq->__lock){-.-.}-{2:2}, at: raw_spin_rq_lock_nested+0x20/0x90 [ 13.414612] [ 13.414612] stack backtrace: [ 13.415255] CPU: 1 UID: 0 PID: 80 Comm: kworker/1:1 Not tainted 6.13.0-virtme #15 [ 13.415505] Workqueue: 0x0 (events) [ 13.415567] Sched_ext: dsp_local_on (enabled+all), task: runnable_at=-2ms [ 13.415570] Call Trace: [ 13.415700] [ 13.415744] dump_stack_lvl+0x78/0xe0 [ 13.415806] ? dispatch_to_local_dsq+0x108/0x1a0 [ 13.415884] print_unlock_imbalance_bug+0x11b/0x130 [ 13.415965] ? dispatch_to_local_dsq+0x108/0x1a0 [ 13.416226] lock_release+0x231/0x2c0 [ 13.416326] _raw_spin_unlock+0x1b/0x40 [ 13.416422] dispatch_to_local_dsq+0x108/0x1a0 [ 13.416554] flush_dispatch_buf+0x199/0x1d0 [ 13.416652] balance_one+0x194/0x370 [ 13.416751] balance_scx+0x61/0x1e0 [ 13.416848] prev_balance+0x43/0xb0 [ 13.416947] __pick_next_task+0x6b/0x1b0 [ 13.417052] __schedule+0x20d/0x1740 This happens because dispatch_to_local_dsq() is racing with dispatch_dequeue() and, when the latter wins, we incorrectly assume that the task has been moved to dst_rq. Fix by properly tracking the currently locked rq. Fixes: 4d3ca89bdd31 ("sched_ext: Refactor consume_remote_task()") Signed-off-by: Andrea Righi Signed-off-by: Tejun Heo --- kernel/sched/ext.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) (limited to 'kernel/sched') diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index c7b159f48834..a6d6d6dadde5 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -2575,6 +2575,9 @@ static void dispatch_to_local_dsq(struct rq *rq, struct scx_dispatch_q *dst_dsq, { struct rq *src_rq = task_rq(p); struct rq *dst_rq = container_of(dst_dsq, struct rq, scx.local_dsq); +#ifdef CONFIG_SMP + struct rq *locked_rq = rq; +#endif /* * We're synchronized against dequeue through DISPATCHING. As @p can't @@ -2611,8 +2614,9 @@ static void dispatch_to_local_dsq(struct rq *rq, struct scx_dispatch_q *dst_dsq, atomic_long_set_release(&p->scx.ops_state, SCX_OPSS_NONE); /* switch to @src_rq lock */ - if (rq != src_rq) { - raw_spin_rq_unlock(rq); + if (locked_rq != src_rq) { + raw_spin_rq_unlock(locked_rq); + locked_rq = src_rq; raw_spin_rq_lock(src_rq); } @@ -2630,6 +2634,8 @@ static void dispatch_to_local_dsq(struct rq *rq, struct scx_dispatch_q *dst_dsq, } else { move_remote_task_to_local_dsq(p, enq_flags, src_rq, dst_rq); + /* task has been moved to dst_rq, which is now locked */ + locked_rq = dst_rq; } /* if the destination CPU is idle, wake it up */ @@ -2638,8 +2644,8 @@ static void dispatch_to_local_dsq(struct rq *rq, struct scx_dispatch_q *dst_dsq, } /* switch back to @rq lock */ - if (rq != dst_rq) { - raw_spin_rq_unlock(dst_rq); + if (locked_rq != rq) { + raw_spin_rq_unlock(locked_rq); raw_spin_rq_lock(rq); } #else /* CONFIG_SMP */ -- cgit v1.2.3 From bcc6244e13b4d4903511a1ea84368abf925031c0 Mon Sep 17 00:00:00 2001 From: Jann Horn Date: Wed, 29 Jan 2025 20:53:03 +0100 Subject: sched: Clarify wake_up_q()'s write to task->wake_q.next Clarify that wake_up_q() does an atomic write to task->wake_q.next, after which a concurrent __wake_q_add() can immediately overwrite task->wake_q.next again. Signed-off-by: Jann Horn Signed-off-by: Peter Zijlstra (Intel) Link: https://lkml.kernel.org/r/20250129-sched-wakeup-prettier-v1-1-2f51f5f663fa@google.com --- kernel/sched/core.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel/sched') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 3e5a6bf587f9..8931d9b1e895 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -1055,9 +1055,10 @@ void wake_up_q(struct wake_q_head *head) struct task_struct *task; task = container_of(node, struct task_struct, wake_q); - /* Task can safely be re-inserted now: */ node = node->next; - task->wake_q.next = NULL; + /* pairs with cmpxchg_relaxed() in __wake_q_add() */ + WRITE_ONCE(task->wake_q.next, NULL); + /* Task can safely be re-inserted now. */ /* * wake_up_process() executes a full barrier, which pairs with -- cgit v1.2.3 From 2fa0fbeb69edd367b7c44f484e8dc5a5a1a311ef Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 7 Feb 2025 10:58:23 -1000 Subject: sched_ext: Implement auto local dispatching of migration disabled tasks Migration disabled tasks are special and pinned to their previous CPUs. They tripped up some unsuspecting BPF schedulers as their ->nr_cpus_allowed may not agree with the bits set in ->cpus_ptr. Make it easier for BPF schedulers by automatically dispatching them to the pinned local DSQs by default. If a BPF scheduler wants to handle migration disabled tasks explicitly, it can set SCX_OPS_ENQ_MIGRATION_DISABLED. Signed-off-by: Tejun Heo Acked-by: Andrea Righi --- kernel/sched/ext.c | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) (limited to 'kernel/sched') diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index a6d6d6dadde5..efdbf4d85a21 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -122,6 +122,19 @@ enum scx_ops_flags { */ SCX_OPS_SWITCH_PARTIAL = 1LLU << 3, + /* + * A migration disabled task can only execute on its current CPU. By + * default, such tasks are automatically put on the CPU's local DSQ with + * the default slice on enqueue. If this ops flag is set, they also go + * through ops.enqueue(). + * + * A migration disabled task never invokes ops.select_cpu() as it can + * only select the current CPU. Also, p->cpus_ptr will only contain its + * current CPU while p->nr_cpus_allowed keeps tracking p->user_cpus_ptr + * and thus may disagree with cpumask_weight(p->cpus_ptr). + */ + SCX_OPS_ENQ_MIGRATION_DISABLED = 1LLU << 4, + /* * CPU cgroup support flags */ @@ -130,6 +143,7 @@ enum scx_ops_flags { SCX_OPS_ALL_FLAGS = SCX_OPS_KEEP_BUILTIN_IDLE | SCX_OPS_ENQ_LAST | SCX_OPS_ENQ_EXITING | + SCX_OPS_ENQ_MIGRATION_DISABLED | SCX_OPS_SWITCH_PARTIAL | SCX_OPS_HAS_CGROUP_WEIGHT, }; @@ -882,6 +896,7 @@ static bool scx_warned_zero_slice; static DEFINE_STATIC_KEY_FALSE(scx_ops_enq_last); static DEFINE_STATIC_KEY_FALSE(scx_ops_enq_exiting); +static DEFINE_STATIC_KEY_FALSE(scx_ops_enq_migration_disabled); static DEFINE_STATIC_KEY_FALSE(scx_ops_cpu_preempt); static DEFINE_STATIC_KEY_FALSE(scx_builtin_idle_enabled); @@ -2014,6 +2029,11 @@ static void do_enqueue_task(struct rq *rq, struct task_struct *p, u64 enq_flags, unlikely(p->flags & PF_EXITING)) goto local; + /* see %SCX_OPS_ENQ_MIGRATION_DISABLED */ + if (!static_branch_unlikely(&scx_ops_enq_migration_disabled) && + is_migration_disabled(p)) + goto local; + if (!SCX_HAS_OP(enqueue)) goto global; @@ -5052,6 +5072,7 @@ static void scx_ops_disable_workfn(struct kthread_work *work) static_branch_disable(&scx_has_op[i]); static_branch_disable(&scx_ops_enq_last); static_branch_disable(&scx_ops_enq_exiting); + static_branch_disable(&scx_ops_enq_migration_disabled); static_branch_disable(&scx_ops_cpu_preempt); static_branch_disable(&scx_builtin_idle_enabled); synchronize_rcu(); @@ -5661,6 +5682,8 @@ static int scx_ops_enable(struct sched_ext_ops *ops, struct bpf_link *link) if (ops->flags & SCX_OPS_ENQ_EXITING) static_branch_enable(&scx_ops_enq_exiting); + if (ops->flags & SCX_OPS_ENQ_MIGRATION_DISABLED) + static_branch_enable(&scx_ops_enq_migration_disabled); if (scx_ops.cpu_acquire || scx_ops.cpu_release) static_branch_enable(&scx_ops_cpu_preempt); -- cgit v1.2.3 From 32966821574cd2917bd60f2554f435fe527f4702 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 7 Feb 2025 10:59:06 -1000 Subject: sched_ext: Fix migration disabled handling in targeted dispatches A dispatch operation that can target a specific local DSQ - scx_bpf_dsq_move_to_local() or scx_bpf_dsq_move() - checks whether the task can be migrated to the target CPU using task_can_run_on_remote_rq(). If the task can't be migrated to the targeted CPU, it is bounced through a global DSQ. task_can_run_on_remote_rq() assumes that the task is on a CPU that's different from the targeted CPU but the callers doesn't uphold the assumption and may call the function when the task is already on the target CPU. When such task has migration disabled, task_can_run_on_remote_rq() ends up returning %false incorrectly unnecessarily bouncing the task to a global DSQ. Fix it by updating the callers to only call task_can_run_on_remote_rq() when the task is on a different CPU than the target CPU. As this is a bit subtle, for clarity and documentation: - Make task_can_run_on_remote_rq() trigger SCHED_WARN_ON() if the task is on the same CPU as the target CPU. - is_migration_disabled() test in task_can_run_on_remote_rq() cannot trigger if the task is on a different CPU than the target CPU as the preceding task_allowed_on_cpu() test should fail beforehand. Convert the test into SCHED_WARN_ON(). Signed-off-by: Tejun Heo Fixes: 4c30f5ce4f7a ("sched_ext: Implement scx_bpf_dispatch[_vtime]_from_dsq()") Fixes: 0366017e0973 ("sched_ext: Use task_can_run_on_remote_rq() test in dispatch_to_local_dsq()") Cc: stable@vger.kernel.org # v6.12+ --- kernel/sched/ext.c | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) (limited to 'kernel/sched') diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index efdbf4d85a21..e01144340d67 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -2333,12 +2333,16 @@ static void move_remote_task_to_local_dsq(struct task_struct *p, u64 enq_flags, * * - The BPF scheduler is bypassed while the rq is offline and we can always say * no to the BPF scheduler initiated migrations while offline. + * + * The caller must ensure that @p and @rq are on different CPUs. */ static bool task_can_run_on_remote_rq(struct task_struct *p, struct rq *rq, bool trigger_error) { int cpu = cpu_of(rq); + SCHED_WARN_ON(task_cpu(p) == cpu); + /* * We don't require the BPF scheduler to avoid dispatching to offline * CPUs mostly for convenience but also because CPUs can go offline @@ -2352,8 +2356,11 @@ static bool task_can_run_on_remote_rq(struct task_struct *p, struct rq *rq, return false; } - if (unlikely(is_migration_disabled(p))) - return false; + /* + * If @p has migration disabled, @p->cpus_ptr only contains its current + * CPU and the above task_allowed_on_cpu() test should have failed. + */ + SCHED_WARN_ON(is_migration_disabled(p)); if (!scx_rq_online(rq)) return false; @@ -2457,7 +2464,8 @@ static struct rq *move_task_between_dsqs(struct task_struct *p, u64 enq_flags, if (dst_dsq->id == SCX_DSQ_LOCAL) { dst_rq = container_of(dst_dsq, struct rq, scx.local_dsq); - if (!task_can_run_on_remote_rq(p, dst_rq, true)) { + if (src_rq != dst_rq && + unlikely(!task_can_run_on_remote_rq(p, dst_rq, true))) { dst_dsq = find_global_dsq(p); dst_rq = src_rq; } @@ -2611,7 +2619,8 @@ static void dispatch_to_local_dsq(struct rq *rq, struct scx_dispatch_q *dst_dsq, } #ifdef CONFIG_SMP - if (unlikely(!task_can_run_on_remote_rq(p, dst_rq, true))) { + if (src_rq != dst_rq && + unlikely(!task_can_run_on_remote_rq(p, dst_rq, true))) { dispatch_enqueue(find_global_dsq(p), p, enq_flags | SCX_ENQ_CLEAR_OPSS); return; -- cgit v1.2.3 From f3f08c3acfb8860e07a22814a344e83c99ad7398 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 10 Feb 2025 09:27:09 -1000 Subject: sched_ext: Fix incorrect assumption about migration disabled tasks in task_can_run_on_remote_rq() While fixing migration disabled task handling, 32966821574c ("sched_ext: Fix migration disabled handling in targeted dispatches") assumed that a migration disabled task's ->cpus_ptr would only have the pinned CPU. While this is eventually true for migration disabled tasks that are switched out, ->cpus_ptr update is performed by migrate_disable_switch() which is called right before context_switch() in __scheduler(). However, the task is enqueued earlier during pick_next_task() via put_prev_task_scx(), so there is a race window where another CPU can see the task on a DSQ. If the CPU tries to dispatch the migration disabled task while in that window, task_allowed_on_cpu() will succeed and task_can_run_on_remote_rq() will subsequently trigger SCHED_WARN(is_migration_disabled()). WARNING: CPU: 8 PID: 1837 at kernel/sched/ext.c:2466 task_can_run_on_remote_rq+0x12e/0x140 Sched_ext: layered (enabled+all), task: runnable_at=-10ms RIP: 0010:task_can_run_on_remote_rq+0x12e/0x140 ... consume_dispatch_q+0xab/0x220 scx_bpf_dsq_move_to_local+0x58/0xd0 bpf_prog_84dd17b0654b6cf0_layered_dispatch+0x290/0x1cfa bpf__sched_ext_ops_dispatch+0x4b/0xab balance_one+0x1fe/0x3b0 balance_scx+0x61/0x1d0 prev_balance+0x46/0xc0 __pick_next_task+0x73/0x1c0 __schedule+0x206/0x1730 schedule+0x3a/0x160 __do_sys_sched_yield+0xe/0x20 do_syscall_64+0xbb/0x1e0 entry_SYSCALL_64_after_hwframe+0x77/0x7f Fix it by converting the SCHED_WARN() back to a regular failure path. Also, perform the migration disabled test before task_allowed_on_cpu() test so that BPF schedulers which fail to handle migration disabled tasks can be noticed easily. While at it, adjust scx_ops_error() message for !task_allowed_on_cpu() case for brevity and consistency. Signed-off-by: Tejun Heo Fixes: 32966821574c ("sched_ext: Fix migration disabled handling in targeted dispatches") Acked-by: Andrea Righi Reported-by: Jake Hillion --- kernel/sched/ext.c | 29 +++++++++++++++++++++-------- 1 file changed, 21 insertions(+), 8 deletions(-) (limited to 'kernel/sched') diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index e01144340d67..54edd0e2132a 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -2343,6 +2343,25 @@ static bool task_can_run_on_remote_rq(struct task_struct *p, struct rq *rq, SCHED_WARN_ON(task_cpu(p) == cpu); + /* + * If @p has migration disabled, @p->cpus_ptr is updated to contain only + * the pinned CPU in migrate_disable_switch() while @p is being switched + * out. However, put_prev_task_scx() is called before @p->cpus_ptr is + * updated and thus another CPU may see @p on a DSQ inbetween leading to + * @p passing the below task_allowed_on_cpu() check while migration is + * disabled. + * + * Test the migration disabled state first as the race window is narrow + * and the BPF scheduler failing to check migration disabled state can + * easily be masked if task_allowed_on_cpu() is done first. + */ + if (unlikely(is_migration_disabled(p))) { + if (trigger_error) + scx_ops_error("SCX_DSQ_LOCAL[_ON] cannot move migration disabled %s[%d] from CPU %d to %d", + p->comm, p->pid, task_cpu(p), cpu); + return false; + } + /* * We don't require the BPF scheduler to avoid dispatching to offline * CPUs mostly for convenience but also because CPUs can go offline @@ -2351,17 +2370,11 @@ static bool task_can_run_on_remote_rq(struct task_struct *p, struct rq *rq, */ if (!task_allowed_on_cpu(p, cpu)) { if (trigger_error) - scx_ops_error("SCX_DSQ_LOCAL[_ON] verdict target cpu %d not allowed for %s[%d]", - cpu_of(rq), p->comm, p->pid); + scx_ops_error("SCX_DSQ_LOCAL[_ON] target CPU %d not allowed for %s[%d]", + cpu, p->comm, p->pid); return false; } - /* - * If @p has migration disabled, @p->cpus_ptr only contains its current - * CPU and the above task_allowed_on_cpu() test should have failed. - */ - SCHED_WARN_ON(is_migration_disabled(p)); - if (!scx_rq_online(rq)) return false; -- cgit v1.2.3 From f5717c93a1b999970f3a64d771a1a9ee68cc37d0 Mon Sep 17 00:00:00 2001 From: Chuyi Zhou Date: Wed, 12 Feb 2025 21:09:35 +0800 Subject: sched_ext: Use SCX_CALL_OP_TASK in task_tick_scx Now when we use scx_bpf_task_cgroup() in ops.tick() to get the cgroup of the current task, the following error will occur: scx_foo[3795244] triggered exit kind 1024: runtime error (called on a task not being operated on) The reason is that we are using SCX_CALL_OP() instead of SCX_CALL_OP_TASK() when calling ops.tick(), which triggers the error during the subsequent scx_kf_allowed_on_arg_tasks() check. SCX_CALL_OP_TASK() was first introduced in commit 36454023f50b ("sched_ext: Track tasks that are subjects of the in-flight SCX operation") to ensure task's rq lock is held when accessing task's sched_group. Since ops.tick() is marked as SCX_KF_TERMINAL and task_tick_scx() is protected by the rq lock, we can use SCX_CALL_OP_TASK() to avoid the above issue. Similarly, the same changes should be made for ops.disable() and ops.exit_task(), as they are also protected by task_rq_lock() and it's safe to access the task's task_group. Fixes: 36454023f50b ("sched_ext: Track tasks that are subjects of the in-flight SCX operation") Signed-off-by: Chuyi Zhou Signed-off-by: Tejun Heo --- kernel/sched/ext.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel/sched') diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 54edd0e2132a..5a81d9a1e31f 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -3899,7 +3899,7 @@ static void task_tick_scx(struct rq *rq, struct task_struct *curr, int queued) curr->scx.slice = 0; touch_core_sched(rq, curr); } else if (SCX_HAS_OP(tick)) { - SCX_CALL_OP(SCX_KF_REST, tick, curr); + SCX_CALL_OP_TASK(SCX_KF_REST, tick, curr); } if (!curr->scx.slice) @@ -4046,7 +4046,7 @@ static void scx_ops_disable_task(struct task_struct *p) WARN_ON_ONCE(scx_get_task_state(p) != SCX_TASK_ENABLED); if (SCX_HAS_OP(disable)) - SCX_CALL_OP(SCX_KF_REST, disable, p); + SCX_CALL_OP_TASK(SCX_KF_REST, disable, p); scx_set_task_state(p, SCX_TASK_READY); } @@ -4075,7 +4075,7 @@ static void scx_ops_exit_task(struct task_struct *p) } if (SCX_HAS_OP(exit_task)) - SCX_CALL_OP(SCX_KF_REST, exit_task, p, &args); + SCX_CALL_OP_TASK(SCX_KF_REST, exit_task, p, &args); scx_set_task_state(p, SCX_TASK_NONE); } -- cgit v1.2.3 From 02d954c0fdf91845169cdacc7405b120f90afe01 Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Mon, 10 Feb 2025 16:32:50 +0100 Subject: sched: Compact RSEQ concurrency IDs with reduced threads and affinity When a process reduces its number of threads or clears bits in its CPU affinity mask, the mm_cid allocation should eventually converge towards smaller values. However, the change introduced by: commit 7e019dcc470f ("sched: Improve cache locality of RSEQ concurrency IDs for intermittent workloads") adds a per-mm/CPU recent_cid which is never unset unless a thread migrates. This is a tradeoff between: A) Preserving cache locality after a transition from many threads to few threads, or after reducing the hamming weight of the allowed CPU mask. B) Making the mm_cid upper bounds wrt nr threads and allowed CPU mask easy to document and understand. C) Allowing applications to eventually react to mm_cid compaction after reduction of the nr threads or allowed CPU mask, making the tracking of mm_cid compaction easier by shrinking it back towards 0 or not. D) Making sure applications that periodically reduce and then increase again the nr threads or allowed CPU mask still benefit from good cache locality with mm_cid. Introduce the following changes: * After shrinking the number of threads or reducing the number of allowed CPUs, reduce the value of max_nr_cid so expansion of CID allocation will preserve cache locality if the number of threads or allowed CPUs increase again. * Only re-use a recent_cid if it is within the max_nr_cid upper bound, else find the first available CID. Fixes: 7e019dcc470f ("sched: Improve cache locality of RSEQ concurrency IDs for intermittent workloads") Signed-off-by: Mathieu Desnoyers Signed-off-by: Gabriele Monaco Signed-off-by: Peter Zijlstra (Intel) Tested-by: Gabriele Monaco Link: https://lkml.kernel.org/r/20250210153253.460471-2-gmonaco@redhat.com --- include/linux/mm_types.h | 7 ++++--- kernel/sched/sched.h | 25 ++++++++++++++++++++++--- 2 files changed, 26 insertions(+), 6 deletions(-) (limited to 'kernel/sched') diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 6b27db7f9496..0234f14f2aa6 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -875,10 +875,11 @@ struct mm_struct { */ unsigned int nr_cpus_allowed; /** - * @max_nr_cid: Maximum number of concurrency IDs allocated. + * @max_nr_cid: Maximum number of allowed concurrency + * IDs allocated. * - * Track the highest number of concurrency IDs allocated for the - * mm. + * Track the highest number of allowed concurrency IDs + * allocated for the mm. */ atomic_t max_nr_cid; /** diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index b93c8c3dc05a..c8512a9fb022 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -3698,10 +3698,28 @@ static inline int __mm_cid_try_get(struct task_struct *t, struct mm_struct *mm) { struct cpumask *cidmask = mm_cidmask(mm); struct mm_cid __percpu *pcpu_cid = mm->pcpu_cid; - int cid = __this_cpu_read(pcpu_cid->recent_cid); + int cid, max_nr_cid, allowed_max_nr_cid; + /* + * After shrinking the number of threads or reducing the number + * of allowed cpus, reduce the value of max_nr_cid so expansion + * of cid allocation will preserve cache locality if the number + * of threads or allowed cpus increase again. + */ + max_nr_cid = atomic_read(&mm->max_nr_cid); + while ((allowed_max_nr_cid = min_t(int, READ_ONCE(mm->nr_cpus_allowed), + atomic_read(&mm->mm_users))), + max_nr_cid > allowed_max_nr_cid) { + /* atomic_try_cmpxchg loads previous mm->max_nr_cid into max_nr_cid. */ + if (atomic_try_cmpxchg(&mm->max_nr_cid, &max_nr_cid, allowed_max_nr_cid)) { + max_nr_cid = allowed_max_nr_cid; + break; + } + } /* Try to re-use recent cid. This improves cache locality. */ - if (!mm_cid_is_unset(cid) && !cpumask_test_and_set_cpu(cid, cidmask)) + cid = __this_cpu_read(pcpu_cid->recent_cid); + if (!mm_cid_is_unset(cid) && cid < max_nr_cid && + !cpumask_test_and_set_cpu(cid, cidmask)) return cid; /* * Expand cid allocation if the maximum number of concurrency @@ -3709,8 +3727,9 @@ static inline int __mm_cid_try_get(struct task_struct *t, struct mm_struct *mm) * and number of threads. Expanding cid allocation as much as * possible improves cache locality. */ - cid = atomic_read(&mm->max_nr_cid); + cid = max_nr_cid; while (cid < READ_ONCE(mm->nr_cpus_allowed) && cid < atomic_read(&mm->mm_users)) { + /* atomic_try_cmpxchg loads previous mm->max_nr_cid into cid. */ if (!atomic_try_cmpxchg(&mm->max_nr_cid, &cid, cid + 1)) continue; if (!cpumask_test_and_set_cpu(cid, cidmask)) -- cgit v1.2.3