From 7266f0a6d3bb73f42ea06656d3cc48c7d0386f71 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 9 Oct 2024 10:00:09 +0200 Subject: [PATCH 01/16] fs/bcachefs: Fix __wait_on_freeing_inode() definition of waitqueue entry The following recent commit made DEFINE_WAIT_BIT() type requirements stricter: 2382d68d7d43 ("sched: change wake_up_bit() and related function to expect unsigned long *") .. which results in a build failure: > fs/bcachefs/fs.c: In function '__wait_on_freeing_inode': > fs/bcachefs/fs.c:281:31: error: initialization of 'long unsigned int *' from incompatible pointer type 'u32 *' {aka 'unsigned int *'} [-Wincompatible-pointer-types] > 281 | DEFINE_WAIT_BIT(wait, &inode->v.i_state, __I_NEW); Since this code relies on the waitqueue initialization within inode_bit_waitqueue() anyway, the DEFINE_WAIT_BIT() initialization is unnecessary - we can just declare a waitqueue entry. Reported-by: Stephen Rothwell Suggested-by: NeilBrown Signed-off-by: Ingo Molnar --- fs/bcachefs/fs.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/bcachefs/fs.c b/fs/bcachefs/fs.c index 5bfc26d58270..c410133541ba 100644 --- a/fs/bcachefs/fs.c +++ b/fs/bcachefs/fs.c @@ -183,8 +183,9 @@ static void __wait_on_freeing_inode(struct bch_fs *c, struct bch_inode_info *inode, subvol_inum inum) { + struct wait_bit_queue_entry wait; wait_queue_head_t *wq; - DEFINE_WAIT_BIT(wait, &inode->v.i_state, __I_NEW); + wq = inode_bit_waitqueue(&wait, &inode->v, __I_NEW); prepare_to_wait(wq, &wait.wq_entry, TASK_UNINTERRUPTIBLE); spin_unlock(&inode->v.i_lock); -- 2.51.0 From d16b7eb6f523eeac3cff13001ef2a59cd462aa73 Mon Sep 17 00:00:00 2001 From: Phil Auld Date: Fri, 4 Oct 2024 08:37:29 -0400 Subject: [PATCH 02/16] sched/deadline: Use hrtick_enabled_dl() before start_hrtick_dl() The deadline server code moved one of the start_hrtick_dl() calls but dropped the dl specific hrtick_enabled check. This causes hrticks to get armed even when sched_feat(HRTICK_DL) is false. Fix it. Fixes: 63ba8422f876 ("sched/deadline: Introduce deadline servers") Signed-off-by: Phil Auld Signed-off-by: Peter Zijlstra (Intel) Acked-by: Juri Lelli Link: https://lore.kernel.org/r/20241004123729.460668-1-pauld@redhat.com --- kernel/sched/deadline.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c index 9ce93d0bf452..be1b917dc8ce 100644 --- a/kernel/sched/deadline.c +++ b/kernel/sched/deadline.c @@ -2385,7 +2385,7 @@ static void set_next_task_dl(struct rq *rq, struct task_struct *p, bool first) deadline_queue_push_tasks(rq); - if (hrtick_enabled(rq)) + if (hrtick_enabled_dl(rq)) start_hrtick_dl(rq, &p->dl); } -- 2.51.0 From 73ab05aa46b02d96509cb029a8d04fca7bbde8c7 Mon Sep 17 00:00:00 2001 From: Waiman Long Date: Wed, 9 Oct 2024 21:44:32 -0400 Subject: [PATCH 03/16] sched/core: Disable page allocation in task_tick_mm_cid() With KASAN and PREEMPT_RT enabled, calling task_work_add() in task_tick_mm_cid() may cause the following splat. [ 63.696416] BUG: sleeping function called from invalid context at kernel/locking/spinlock_rt.c:48 [ 63.696416] in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 610, name: modprobe [ 63.696416] preempt_count: 10001, expected: 0 [ 63.696416] RCU nest depth: 1, expected: 1 This problem is caused by the following call trace. sched_tick() [ acquire rq->__lock ] -> task_tick_mm_cid() -> task_work_add() -> __kasan_record_aux_stack() -> kasan_save_stack() -> stack_depot_save_flags() -> alloc_pages_mpol_noprof() -> __alloc_pages_noprof() -> get_page_from_freelist() -> rmqueue() -> rmqueue_pcplist() -> __rmqueue_pcplist() -> rmqueue_bulk() -> rt_spin_lock() The rq lock is a raw_spinlock_t. We can't sleep while holding it. IOW, we can't call alloc_pages() in stack_depot_save_flags(). The task_tick_mm_cid() function with its task_work_add() call was introduced by commit 223baf9d17f2 ("sched: Fix performance regression introduced by mm_cid") in v6.4 kernel. Fortunately, there is a kasan_record_aux_stack_noalloc() variant that calls stack_depot_save_flags() while not allowing it to allocate new pages. To allow task_tick_mm_cid() to use task_work without page allocation, a new TWAF_NO_ALLOC flag is added to enable calling kasan_record_aux_stack_noalloc() instead of kasan_record_aux_stack() if set. The task_tick_mm_cid() function is modified to add this new flag. The possible downside is the missing stack trace in a KASAN report due to new page allocation required when task_work_add_noallloc() is called which should be rare. Fixes: 223baf9d17f2 ("sched: Fix performance regression introduced by mm_cid") Signed-off-by: Waiman Long Signed-off-by: Peter Zijlstra (Intel) Link: https://lkml.kernel.org/r/20241010014432.194742-1-longman@redhat.com --- include/linux/task_work.h | 5 ++++- kernel/sched/core.c | 4 +++- kernel/task_work.c | 15 +++++++++++++-- 3 files changed, 20 insertions(+), 4 deletions(-) diff --git a/include/linux/task_work.h b/include/linux/task_work.h index cf5e7e891a77..2964171856e0 100644 --- a/include/linux/task_work.h +++ b/include/linux/task_work.h @@ -14,11 +14,14 @@ init_task_work(struct callback_head *twork, task_work_func_t func) } enum task_work_notify_mode { - TWA_NONE, + TWA_NONE = 0, TWA_RESUME, TWA_SIGNAL, TWA_SIGNAL_NO_IPI, TWA_NMI_CURRENT, + + TWA_FLAGS = 0xff00, + TWAF_NO_ALLOC = 0x0100, }; static inline bool task_work_pending(struct task_struct *task) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 43e453ab7e20..0259301e572e 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -10458,7 +10458,9 @@ void task_tick_mm_cid(struct rq *rq, struct task_struct *curr) return; if (time_before(now, READ_ONCE(curr->mm->mm_cid_next_scan))) return; - task_work_add(curr, work, TWA_RESUME); + + /* No page allocation under rq lock */ + task_work_add(curr, work, TWA_RESUME | TWAF_NO_ALLOC); } void sched_mm_cid_exit_signals(struct task_struct *t) diff --git a/kernel/task_work.c b/kernel/task_work.c index 5d14d639ac71..c969f1f26be5 100644 --- a/kernel/task_work.c +++ b/kernel/task_work.c @@ -55,15 +55,26 @@ int task_work_add(struct task_struct *task, struct callback_head *work, enum task_work_notify_mode notify) { struct callback_head *head; + int flags = notify & TWA_FLAGS; + notify &= ~TWA_FLAGS; if (notify == TWA_NMI_CURRENT) { if (WARN_ON_ONCE(task != current)) return -EINVAL; if (!IS_ENABLED(CONFIG_IRQ_WORK)) return -EINVAL; } else { - /* record the work call stack in order to print it in KASAN reports */ - kasan_record_aux_stack(work); + /* + * Record the work call stack in order to print it in KASAN + * reports. + * + * Note that stack allocation can fail if TWAF_NO_ALLOC flag + * is set and new page is needed to expand the stack buffer. + */ + if (flags & TWAF_NO_ALLOC) + kasan_record_aux_stack_noalloc(work); + else + kasan_record_aux_stack(work); } head = READ_ONCE(task->task_works); -- 2.51.0 From 98442f0ccd828ac42e89281a815e9e7a97533822 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 10 Oct 2024 11:54:38 +0200 Subject: [PATCH 04/16] sched: Fix delayed_dequeue vs switched_from_fair() Commit 2e0199df252a ("sched/fair: Prepare exit/cleanup paths for delayed_dequeue") and its follow up fixes try to deal with a rather unfortunate situation where is task is enqueued in a new class, even though it shouldn't have been. Mostly because the existing ->switched_to/from() hooks are in the wrong place for this case. This all led to Paul being able to trigger failures at something like once per 10k CPU hours of RCU torture. For now, do the ugly thing and move the code to the right place by ignoring the switch hooks. Note: Clean up the whole sched_class::switch*_{to,from}() thing. Fixes: 2e0199df252a ("sched/fair: Prepare exit/cleanup paths for delayed_dequeue") Reported-by: Paul E. McKenney Signed-off-by: Peter Zijlstra (Intel) Link: https://lkml.kernel.org/r/20241003185037.GA5594@noisy.programming.kicks-ass.net --- kernel/sched/core.c | 29 ++++++++++++++++++----------- kernel/sched/ext.c | 4 ++-- kernel/sched/fair.c | 16 ---------------- kernel/sched/sched.h | 2 +- kernel/sched/syscalls.c | 13 +++++++++---- 5 files changed, 30 insertions(+), 34 deletions(-) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 0259301e572e..a860996622a6 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -7010,20 +7010,20 @@ int default_wake_function(wait_queue_entry_t *curr, unsigned mode, int wake_flag } EXPORT_SYMBOL(default_wake_function); -void __setscheduler_prio(struct task_struct *p, int prio) +const struct sched_class *__setscheduler_class(struct task_struct *p, int prio) { if (dl_prio(prio)) - p->sched_class = &dl_sched_class; - else if (rt_prio(prio)) - p->sched_class = &rt_sched_class; + return &dl_sched_class; + + if (rt_prio(prio)) + return &rt_sched_class; + #ifdef CONFIG_SCHED_CLASS_EXT - else if (task_should_scx(p)) - p->sched_class = &ext_sched_class; + if (task_should_scx(p)) + return &ext_sched_class; #endif - else - p->sched_class = &fair_sched_class; - p->prio = prio; + return &fair_sched_class; } #ifdef CONFIG_RT_MUTEXES @@ -7069,7 +7069,7 @@ void rt_mutex_setprio(struct task_struct *p, struct task_struct *pi_task) { int prio, oldprio, queued, running, queue_flag = DEQUEUE_SAVE | DEQUEUE_MOVE | DEQUEUE_NOCLOCK; - const struct sched_class *prev_class; + const struct sched_class *prev_class, *next_class; struct rq_flags rf; struct rq *rq; @@ -7127,6 +7127,11 @@ void rt_mutex_setprio(struct task_struct *p, struct task_struct *pi_task) queue_flag &= ~DEQUEUE_MOVE; prev_class = p->sched_class; + next_class = __setscheduler_class(p, prio); + + if (prev_class != next_class && p->se.sched_delayed) + dequeue_task(rq, p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK); + queued = task_on_rq_queued(p); running = task_current(rq, p); if (queued) @@ -7164,7 +7169,9 @@ void rt_mutex_setprio(struct task_struct *p, struct task_struct *pi_task) p->rt.timeout = 0; } - __setscheduler_prio(p, prio); + p->sched_class = next_class; + p->prio = prio; + check_class_changing(rq, p, prev_class); if (queued) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 3cd7c50a51c5..6f9de573ee93 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -4471,7 +4471,7 @@ static void scx_ops_disable_workfn(struct kthread_work *work) sched_deq_and_put_task(p, DEQUEUE_SAVE | DEQUEUE_MOVE, &ctx); p->scx.slice = min_t(u64, p->scx.slice, SCX_SLICE_DFL); - __setscheduler_prio(p, p->prio); + p->sched_class = __setscheduler_class(p, p->prio); check_class_changing(task_rq(p), p, old_class); sched_enq_and_set_task(&ctx); @@ -5186,7 +5186,7 @@ static int scx_ops_enable(struct sched_ext_ops *ops, struct bpf_link *link) sched_deq_and_put_task(p, DEQUEUE_SAVE | DEQUEUE_MOVE, &ctx); - __setscheduler_prio(p, p->prio); + p->sched_class = __setscheduler_class(p, p->prio); check_class_changing(task_rq(p), p, old_class); sched_enq_and_set_task(&ctx); diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index ab497fafa7be..c157d4860a3b 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -13177,22 +13177,6 @@ static void attach_task_cfs_rq(struct task_struct *p) static void switched_from_fair(struct rq *rq, struct task_struct *p) { detach_task_cfs_rq(p); - /* - * Since this is called after changing class, this is a little weird - * and we cannot use DEQUEUE_DELAYED. - */ - if (p->se.sched_delayed) { - /* First, dequeue it from its new class' structures */ - dequeue_task(rq, p, DEQUEUE_NOCLOCK | DEQUEUE_SLEEP); - /* - * Now, clean up the fair_sched_class side of things - * related to sched_delayed being true and that wasn't done - * due to the generic dequeue not using DEQUEUE_DELAYED. - */ - finish_delayed_dequeue_entity(&p->se); - p->se.rel_deadline = 0; - __block_task(rq, p); - } } static void switched_to_fair(struct rq *rq, struct task_struct *p) diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index b1c3588a8f00..fba524c81c63 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -3797,7 +3797,7 @@ static inline int rt_effective_prio(struct task_struct *p, int prio) extern int __sched_setscheduler(struct task_struct *p, const struct sched_attr *attr, bool user, bool pi); extern int __sched_setaffinity(struct task_struct *p, struct affinity_context *ctx); -extern void __setscheduler_prio(struct task_struct *p, int prio); +extern const struct sched_class *__setscheduler_class(struct task_struct *p, int prio); extern void set_load_weight(struct task_struct *p, bool update_load); extern void enqueue_task(struct rq *rq, struct task_struct *p, int flags); extern bool dequeue_task(struct rq *rq, struct task_struct *p, int flags); diff --git a/kernel/sched/syscalls.c b/kernel/sched/syscalls.c index aa70beee9895..0470bcc3d204 100644 --- a/kernel/sched/syscalls.c +++ b/kernel/sched/syscalls.c @@ -529,7 +529,7 @@ int __sched_setscheduler(struct task_struct *p, { int oldpolicy = -1, policy = attr->sched_policy; int retval, oldprio, newprio, queued, running; - const struct sched_class *prev_class; + const struct sched_class *prev_class, *next_class; struct balance_callback *head; struct rq_flags rf; int reset_on_fork; @@ -706,6 +706,12 @@ change: queue_flags &= ~DEQUEUE_MOVE; } + prev_class = p->sched_class; + next_class = __setscheduler_class(p, newprio); + + if (prev_class != next_class && p->se.sched_delayed) + dequeue_task(rq, p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK); + queued = task_on_rq_queued(p); running = task_current(rq, p); if (queued) @@ -713,11 +719,10 @@ change: if (running) put_prev_task(rq, p); - prev_class = p->sched_class; - if (!(attr->sched_flags & SCHED_FLAG_KEEP_PARAMS)) { __setscheduler_params(p, attr); - __setscheduler_prio(p, newprio); + p->sched_class = next_class; + p->prio = newprio; } __setscheduler_uclamp(p, attr); check_class_changing(rq, p, prev_class); -- 2.51.0 From f5aaff7bfa11fb0b2ee6b8fd7bbc16cfceea2ad3 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 10 Oct 2024 08:28:36 +0000 Subject: [PATCH 05/16] sched/core: Dequeue PSI signals for blocked tasks that are delayed psi_dequeue() in for blocked task expects psi_sched_switch() to clear the TSK_.*RUNNING PSI flags and set the TSK_IOWAIT flags however psi_sched_switch() uses "!task_on_rq_queued(prev)" to detect if the task is blocked or still runnable which is no longer true with DELAY_DEQUEUE since a blocking task can be left queued on the runqueue. This can lead to PSI splats similar to: psi: inconsistent task state! task=... cpu=... psi_flags=4 clear=0 set=4 when the task is requeued since the TSK_RUNNING flag was not cleared when the task was blocked. Explicitly communicate that the task was blocked to psi_sched_switch() even if it was delayed and is still on the runqueue. [ prateek: Broke off the relevant part from [1], commit message ] Fixes: 152e11f6df29 ("sched/fair: Implement delayed dequeue") Closes: https://lore.kernel.org/lkml/20240830123458.3557-1-spasswolf@web.de/ Closes: https://lore.kernel.org/all/cd67fbcd-d659-4822-bb90-7e8fbb40a856@molgen.mpg.de/ Signed-off-by: Peter Zijlstra (Intel) Not-yet-signed-off-by: Peter Zijlstra Signed-off-by: K Prateek Nayak Signed-off-by: Peter Zijlstra (Intel) Tested-by: Johannes Weiner Link: https://lore.kernel.org/lkml/20241004123506.GR18071@noisy.programming.kicks-ass.net/ [1] --- kernel/sched/core.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index a860996622a6..9e09140ccb34 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -6537,6 +6537,7 @@ static void __sched notrace __schedule(int sched_mode) * as a preemption by schedule_debug() and RCU. */ bool preempt = sched_mode > SM_NONE; + bool block = false; unsigned long *switch_count; unsigned long prev_state; struct rq_flags rf; @@ -6622,6 +6623,7 @@ static void __sched notrace __schedule(int sched_mode) * After this, schedule() must not care about p->state any more. */ block_task(rq, prev, flags); + block = true; } switch_count = &prev->nvcsw; } @@ -6667,7 +6669,7 @@ picked: migrate_disable_switch(rq, prev); psi_account_irqtime(rq, prev, next); - psi_sched_switch(prev, next, !task_on_rq_queued(prev)); + psi_sched_switch(prev, next, block); trace_sched_switch(preempt, prev, next, prev_state); -- 2.51.0 From c6508124193d42bbc3224571eb75bfa4c1821fbb Mon Sep 17 00:00:00 2001 From: Johannes Weiner Date: Fri, 11 Oct 2024 10:49:33 +0200 Subject: [PATCH 06/16] sched/psi: Fix mistaken CPU pressure indication after corrupted task state bug Since sched_delayed tasks remain queued even after blocking, the load balancer can migrate them between runqueues while PSI considers them to be asleep. As a result, it misreads the migration requeue followed by a wakeup as a double queue: psi: inconsistent task state! task=... cpu=... psi_flags=4 clear=. set=4 First, call psi_enqueue() after p->sched_class->enqueue_task(). A wakeup will clear p->se.sched_delayed while a migration will not, so psi can use that flag to tell them apart. Then teach psi to migrate any "sleep" state when delayed-dequeue tasks are being migrated. Delayed-dequeue tasks can be revived by ttwu_runnable(), which will call down with a new ENQUEUE_DELAYED. Instead of further complicating the wakeup conditional in enqueue_task(), identify migration contexts instead and default to wakeup handling for all other cases. It's not just the warning in dmesg, the task state corruption causes a permanent CPU pressure indication, which messes with workload/machine health monitoring. Debugged-by-and-original-fix-by: K Prateek Nayak Fixes: 152e11f6df29 ("sched/fair: Implement delayed dequeue") Closes: https://lore.kernel.org/lkml/20240830123458.3557-1-spasswolf@web.de/ Closes: https://lore.kernel.org/all/cd67fbcd-d659-4822-bb90-7e8fbb40a856@molgen.mpg.de/ Signed-off-by: Johannes Weiner Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Ingo Molnar Tested-by: K Prateek Nayak Link: https://lkml.kernel.org/r/20241010193712.GC181795@cmpxchg.org --- kernel/sched/core.c | 12 +++++------ kernel/sched/stats.h | 48 ++++++++++++++++++++++++++++++-------------- 2 files changed, 39 insertions(+), 21 deletions(-) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 9e09140ccb34..71232f8f9b96 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2012,11 +2012,6 @@ void enqueue_task(struct rq *rq, struct task_struct *p, int flags) if (!(flags & ENQUEUE_NOCLOCK)) update_rq_clock(rq); - if (!(flags & ENQUEUE_RESTORE)) { - sched_info_enqueue(rq, p); - psi_enqueue(p, (flags & ENQUEUE_WAKEUP) && !(flags & ENQUEUE_MIGRATED)); - } - p->sched_class->enqueue_task(rq, p, flags); /* * Must be after ->enqueue_task() because ENQUEUE_DELAYED can clear @@ -2024,6 +2019,11 @@ void enqueue_task(struct rq *rq, struct task_struct *p, int flags) */ uclamp_rq_inc(rq, p); + if (!(flags & ENQUEUE_RESTORE)) { + sched_info_enqueue(rq, p); + psi_enqueue(p, flags & ENQUEUE_MIGRATED); + } + if (sched_core_enabled(rq)) sched_core_enqueue(rq, p); } @@ -2041,7 +2041,7 @@ inline bool dequeue_task(struct rq *rq, struct task_struct *p, int flags) if (!(flags & DEQUEUE_SAVE)) { sched_info_dequeue(rq, p); - psi_dequeue(p, flags & DEQUEUE_SLEEP); + psi_dequeue(p, !(flags & DEQUEUE_SLEEP)); } /* diff --git a/kernel/sched/stats.h b/kernel/sched/stats.h index 237780aa3c53..767e098a3bd1 100644 --- a/kernel/sched/stats.h +++ b/kernel/sched/stats.h @@ -119,45 +119,63 @@ static inline void psi_account_irqtime(struct rq *rq, struct task_struct *curr, /* * PSI tracks state that persists across sleeps, such as iowaits and * memory stalls. As a result, it has to distinguish between sleeps, - * where a task's runnable state changes, and requeues, where a task - * and its state are being moved between CPUs and runqueues. + * where a task's runnable state changes, and migrations, where a task + * and its runnable state are being moved between CPUs and runqueues. + * + * A notable case is a task whose dequeue is delayed. PSI considers + * those sleeping, but because they are still on the runqueue they can + * go through migration requeues. In this case, *sleeping* states need + * to be transferred. */ -static inline void psi_enqueue(struct task_struct *p, bool wakeup) +static inline void psi_enqueue(struct task_struct *p, bool migrate) { - int clear = 0, set = TSK_RUNNING; + int clear = 0, set = 0; if (static_branch_likely(&psi_disabled)) return; - if (p->in_memstall) - set |= TSK_MEMSTALL_RUNNING; - - if (!wakeup) { + if (p->se.sched_delayed) { + /* CPU migration of "sleeping" task */ + SCHED_WARN_ON(!migrate); if (p->in_memstall) set |= TSK_MEMSTALL; + if (p->in_iowait) + set |= TSK_IOWAIT; + } else if (migrate) { + /* CPU migration of runnable task */ + set = TSK_RUNNING; + if (p->in_memstall) + set |= TSK_MEMSTALL | TSK_MEMSTALL_RUNNING; } else { + /* Wakeup of new or sleeping task */ if (p->in_iowait) clear |= TSK_IOWAIT; + set = TSK_RUNNING; + if (p->in_memstall) + set |= TSK_MEMSTALL_RUNNING; } psi_task_change(p, clear, set); } -static inline void psi_dequeue(struct task_struct *p, bool sleep) +static inline void psi_dequeue(struct task_struct *p, bool migrate) { if (static_branch_likely(&psi_disabled)) return; + /* + * When migrating a task to another CPU, clear all psi + * state. The enqueue callback above will work it out. + */ + if (migrate) + psi_task_change(p, p->psi_flags, 0); + /* * A voluntary sleep is a dequeue followed by a task switch. To * avoid walking all ancestors twice, psi_task_switch() handles * TSK_RUNNING and TSK_IOWAIT for us when it moves TSK_ONCPU. * Do nothing here. */ - if (sleep) - return; - - psi_task_change(p, p->psi_flags, 0); } static inline void psi_ttwu_dequeue(struct task_struct *p) @@ -190,8 +208,8 @@ static inline void psi_sched_switch(struct task_struct *prev, } #else /* CONFIG_PSI */ -static inline void psi_enqueue(struct task_struct *p, bool wakeup) {} -static inline void psi_dequeue(struct task_struct *p, bool sleep) {} +static inline void psi_enqueue(struct task_struct *p, bool migrate) {} +static inline void psi_dequeue(struct task_struct *p, bool migrate) {} static inline void psi_ttwu_dequeue(struct task_struct *p) {} static inline void psi_sched_switch(struct task_struct *prev, struct task_struct *next, -- 2.51.0 From cd9626e9ebc77edec33023fe95dab4b04ffc819d Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 10 Oct 2024 11:38:10 +0200 Subject: [PATCH 07/16] sched/fair: Fix external p->on_rq users Sean noted that ever since commit 152e11f6df29 ("sched/fair: Implement delayed dequeue") KVM's preemption notifiers have started mis-classifying preemption vs blocking. Notably p->on_rq is no longer sufficient to determine if a task is runnable or blocked -- the aforementioned commit introduces tasks that remain on the runqueue even through they will not run again, and should be considered blocked for many cases. Add the task_is_runnable() helper to classify things and audit all external users of the p->on_rq state. Also add a few comments. Fixes: 152e11f6df29 ("sched/fair: Implement delayed dequeue") Reported-by: Sean Christopherson Tested-by: Sean Christopherson Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Ingo Molnar Link: https://lkml.kernel.org/r/20241010091843.GK33184@noisy.programming.kicks-ass.net --- include/linux/sched.h | 5 +++++ kernel/events/core.c | 2 +- kernel/freezer.c | 7 ++++++- kernel/rcu/tasks.h | 9 +++++++++ kernel/sched/core.c | 12 +++++++++--- kernel/time/tick-sched.c | 6 ++++++ kernel/trace/trace_selftest.c | 2 +- virt/kvm/kvm_main.c | 2 +- 8 files changed, 38 insertions(+), 7 deletions(-) diff --git a/include/linux/sched.h b/include/linux/sched.h index e6ee4258169a..8a9517e6640c 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -2133,6 +2133,11 @@ static inline void set_task_cpu(struct task_struct *p, unsigned int cpu) #endif /* CONFIG_SMP */ +static inline bool task_is_runnable(struct task_struct *p) +{ + return p->on_rq && !p->se.sched_delayed; +} + extern bool sched_task_on_rq(struct task_struct *p); extern unsigned long get_wchan(struct task_struct *p); extern struct task_struct *cpu_curr_snapshot(int cpu); diff --git a/kernel/events/core.c b/kernel/events/core.c index e3589c4287cb..cdd09769e6c5 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -9251,7 +9251,7 @@ static void perf_event_switch(struct task_struct *task, }, }; - if (!sched_in && task->on_rq) { + if (!sched_in && task_is_runnable(task)) { switch_event.event_id.header.misc |= PERF_RECORD_MISC_SWITCH_OUT_PREEMPT; } diff --git a/kernel/freezer.c b/kernel/freezer.c index 44bbd7dbd2c8..8d530d0949ff 100644 --- a/kernel/freezer.c +++ b/kernel/freezer.c @@ -109,7 +109,12 @@ static int __set_task_frozen(struct task_struct *p, void *arg) { unsigned int state = READ_ONCE(p->__state); - if (p->on_rq) + /* + * Allow freezing the sched_delayed tasks; they will not execute until + * ttwu() fixes them up, so it is safe to swap their state now, instead + * of waiting for them to get fully dequeued. + */ + if (task_is_runnable(p)) return 0; if (p != current && task_curr(p)) diff --git a/kernel/rcu/tasks.h b/kernel/rcu/tasks.h index 6333f4ccf024..4d7ee95df06e 100644 --- a/kernel/rcu/tasks.h +++ b/kernel/rcu/tasks.h @@ -985,6 +985,15 @@ static bool rcu_tasks_is_holdout(struct task_struct *t) if (!READ_ONCE(t->on_rq)) return false; + /* + * t->on_rq && !t->se.sched_delayed *could* be considered sleeping but + * since it is a spurious state (it will transition into the + * traditional blocked state or get woken up without outside + * dependencies), not considering it such should only affect timing. + * + * Be conservative for now and not include it. + */ + /* * Idle tasks (or idle injection) within the idle loop are RCU-tasks * quiescent states. But CPU boot code performed by the idle task diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 71232f8f9b96..7db711ba6d12 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -548,6 +548,11 @@ sched_core_dequeue(struct rq *rq, struct task_struct *p, int flags) { } * ON_RQ_MIGRATING state is used for migration without holding both * rq->locks. It indicates task_cpu() is not stable, see task_rq_lock(). * + * Additionally it is possible to be ->on_rq but still be considered not + * runnable when p->se.sched_delayed is true. These tasks are on the runqueue + * but will be dequeued as soon as they get picked again. See the + * task_is_runnable() helper. + * * p->on_cpu <- { 0, 1 }: * * is set by prepare_task() and cleared by finish_task() such that it will be @@ -4317,9 +4322,10 @@ static bool __task_needs_rq_lock(struct task_struct *p) * @arg: Argument to function. * * Fix the task in it's current state by avoiding wakeups and or rq operations - * and call @func(@arg) on it. This function can use ->on_rq and task_curr() - * to work out what the state is, if required. Given that @func can be invoked - * with a runqueue lock held, it had better be quite lightweight. + * and call @func(@arg) on it. This function can use task_is_runnable() and + * task_curr() to work out what the state is, if required. Given that @func + * can be invoked with a runqueue lock held, it had better be quite + * lightweight. * * Returns: * Whatever @func returns diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 753a184c7090..f203f000da1a 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -434,6 +434,12 @@ static void tick_nohz_kick_task(struct task_struct *tsk) * smp_mb__after_spin_lock() * tick_nohz_task_switch() * LOAD p->tick_dep_mask + * + * XXX given a task picks up the dependency on schedule(), should we + * only care about tasks that are currently on the CPU instead of all + * that are on the runqueue? + * + * That is, does this want to be: task_on_cpu() / task_curr()? */ if (!sched_task_on_rq(tsk)) return; diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index c4ad7cd7e778..1469dd8075fa 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -1485,7 +1485,7 @@ trace_selftest_startup_wakeup(struct tracer *trace, struct trace_array *tr) /* reset the max latency */ tr->max_latency = 0; - while (p->on_rq) { + while (task_is_runnable(p)) { /* * Sleep to make sure the -deadline thread is asleep too. * On virtual machines we can't rely on timings, diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 05cbb2548d99..0c666f1870af 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -6387,7 +6387,7 @@ static void kvm_sched_out(struct preempt_notifier *pn, WRITE_ONCE(vcpu->scheduled_out, true); - if (current->on_rq && vcpu->wants_to_run) { + if (task_is_runnable(current) && vcpu->wants_to_run) { WRITE_ONCE(vcpu->preempted, true); WRITE_ONCE(vcpu->ready, true); } -- 2.51.0 From 8e113df990c9df70fc6d83ebd53ee1b2867c23c4 Mon Sep 17 00:00:00 2001 From: Zhongqiu Han Date: Wed, 9 Oct 2024 17:37:45 +0800 Subject: [PATCH 08/16] sched: idle: Optimize the generic idle loop by removing needless memory barrier The memory barrier rmb() in generic idle loop do_idle() function is not needed, it doesn't order any load instruction, just remove it as needless rmb() can cause performance impact. The rmb() was introduced by the tglx/history.git commit f2f1b44c75c4 ("[PATCH] Remove RCU abuse in cpu_idle()") to order the loads between cpu_idle_map and pm_idle. It pairs with wmb() in function cpu_idle_wait(). And then with the removal of cpu_idle_state in function cpu_idle() and wmb() in function cpu_idle_wait() in commit 783e391b7b5b ("x86: Simplify cpu_idle_wait"), rmb() no longer has a reason to exist. After that, commit d16699123434 ("idle: Implement generic idle function") implemented a generic idle function cpu_idle_loop() which resembles the functionality found in arch/. And it retained the rmb() in generic idle loop in file kernel/cpu/idle.c. And at last, commit cf37b6b48428 ("sched/idle: Move cpu/idle.c to sched/idle.c") moved cpu/idle.c to sched/idle.c. And commit c1de45ca831a ("sched/idle: Add support for tasks that inject idle") renamed function cpu_idle_loop() to do_idle(). History Tree: https://git.kernel.org/pub/scm/linux/kernel/git/tglx/history.git Signed-off-by: Zhongqiu Han Signed-off-by: Peter Zijlstra (Intel) Link: https://lkml.kernel.org/r/20241009093745.9504-1-quic_zhonhan@quicinc.com --- kernel/sched/idle.c | 1 - 1 file changed, 1 deletion(-) diff --git a/kernel/sched/idle.c b/kernel/sched/idle.c index d2f096bb274c..ab911d1335ba 100644 --- a/kernel/sched/idle.c +++ b/kernel/sched/idle.c @@ -271,7 +271,6 @@ static void do_idle(void) tick_nohz_idle_enter(); while (!need_resched()) { - rmb(); /* * Interrupts shouldn't be re-enabled from that point on until -- 2.51.0 From 7e019dcc470f27066c98697e43d930df8d54bd9c Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Wed, 9 Oct 2024 09:50:07 -0400 Subject: [PATCH 09/16] sched: Improve cache locality of RSEQ concurrency IDs for intermittent workloads commit 223baf9d17f25 ("sched: Fix performance regression introduced by mm_cid") introduced a per-mm/cpu current concurrency id (mm_cid), which keeps a reference to the concurrency id allocated for each CPU. This reference expires shortly after a 100ms delay. These per-CPU references keep the per-mm-cid data cache-local in situations where threads are running at least once on each CPU within each 100ms window, thus keeping the per-cpu reference alive. However, intermittent workloads behaving in bursts spaced by more than 100ms on each CPU exhibit bad cache locality and degraded performance compared to purely per-cpu data indexing, because concurrency IDs are allocated over various CPUs and cores, therefore losing cache locality of the associated data. Introduce the following changes to improve per-mm-cid cache locality: - Add a "recent_cid" field to the per-mm/cpu mm_cid structure to keep track of which mm_cid value was last used, and use it as a hint to attempt re-allocating the same concurrency ID the next time this mm/cpu needs to allocate a concurrency ID, - Add a per-mm CPUs allowed mask, which keeps track of the union of CPUs allowed for all threads belonging to this mm. This cpumask is only set during the lifetime of the mm, never cleared, so it represents the union of all the CPUs allowed since the beginning of the mm lifetime (note that the mm_cpumask() is really arch-specific and tailored to the TLB flush needs, and is thus _not_ a viable approach for this), - Add a per-mm nr_cpus_allowed to keep track of the weight of the per-mm CPUs allowed mask (for fast access), - Add a per-mm max_nr_cid to keep track of the highest number of concurrency IDs allocated for the mm. This is used for expanding the concurrency ID allocation within the upper bound defined by: min(mm->nr_cpus_allowed, mm->mm_users) When the next unused CID value reaches this threshold, stop trying to expand the cid allocation and use the first available cid value instead. Spreading allocation to use all the cid values within the range [ 0, min(mm->nr_cpus_allowed, mm->mm_users) - 1 ] improves cache locality while preserving mm_cid compactness within the expected user limits, - In __mm_cid_try_get, only return cid values within the range [ 0, mm->nr_cpus_allowed ] rather than [ 0, nr_cpu_ids ]. This prevents allocating cids above the number of allowed cpus in rare scenarios where cid allocation races with a concurrent remote-clear of the per-mm/cpu cid. This improvement is made possible by the addition of the per-mm CPUs allowed mask, - In sched_mm_cid_migrate_to, use mm->nr_cpus_allowed rather than t->nr_cpus_allowed. This criterion was really meant to compare the number of mm->mm_users to the number of CPUs allowed for the entire mm. Therefore, the prior comparison worked fine when all threads shared the same CPUs allowed mask, but not so much in scenarios where those threads have different masks (e.g. each thread pinned to a single CPU). This improvement is made possible by the addition of the per-mm CPUs allowed mask. * Benchmarks Each thread increments 16kB worth of 8-bit integers in bursts, with a configurable delay between each thread's execution. Each thread run one after the other (no threads run concurrently). The order of thread execution in the sequence is random. The thread execution sequence begins again after all threads have executed. The 16kB areas are allocated with rseq_mempool and indexed by either cpu_id, mm_cid (not cache-local), or cache-local mm_cid. Each thread is pinned to its own core. Testing configurations: 8-core/1-L3: Use 8 cores within a single L3 24-core/24-L3: Use 24 cores, 1 core per L3 192-core/24-L3: Use 192 cores (all cores in the system) 384-thread/24-L3: Use 384 HW threads (all HW threads in the system) Intermittent workload delays between threads: 200ms, 10ms. Hardware: CPU(s): 384 On-line CPU(s) list: 0-383 Vendor ID: AuthenticAMD Model name: AMD EPYC 9654 96-Core Processor Thread(s) per core: 2 Core(s) per socket: 96 Socket(s): 2 Caches (sum of all): L1d: 6 MiB (192 instances) L1i: 6 MiB (192 instances) L2: 192 MiB (192 instances) L3: 768 MiB (24 instances) Each result is an average of 5 test runs. The cache-local speedup is calculated as: (cache-local mm_cid) / (mm_cid). Intermittent workload delay: 200ms per-cpu mm_cid cache-local mm_cid cache-local speedup (ns) (ns) (ns) 8-core/1-L3 1374 19289 1336 14.4x 24-core/24-L3 2423 26721 1594 16.7x 192-core/24-L3 2291 15826 2153 7.3x 384-thread/24-L3 1874 13234 1907 6.9x Intermittent workload delay: 10ms per-cpu mm_cid cache-local mm_cid cache-local speedup (ns) (ns) (ns) 8-core/1-L3 662 756 686 1.1x 24-core/24-L3 1378 3648 1035 3.5x 192-core/24-L3 1439 10833 1482 7.3x 384-thread/24-L3 1503 10570 1556 6.8x [ This deprecates the prior "sched: NUMA-aware per-memory-map concurrency IDs" patch series with a simpler and more general approach. ] [ This patch applies on top of v6.12-rc1. ] Signed-off-by: Mathieu Desnoyers Signed-off-by: Peter Zijlstra (Intel) Acked-by: Marco Elver Link: https://lore.kernel.org/lkml/20240823185946.418340-1-mathieu.desnoyers@efficios.com/ --- fs/exec.c | 2 +- include/linux/mm_types.h | 72 +++++++++++++++++++++++++++++++++++----- kernel/fork.c | 2 +- kernel/sched/core.c | 22 +++++++----- kernel/sched/sched.h | 48 +++++++++++++++++++-------- 5 files changed, 112 insertions(+), 34 deletions(-) diff --git a/fs/exec.c b/fs/exec.c index 6c53920795c2..aaa605529a75 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -990,7 +990,7 @@ static int exec_mmap(struct mm_struct *mm) active_mm = tsk->active_mm; tsk->active_mm = mm; tsk->mm = mm; - mm_init_cid(mm); + mm_init_cid(mm, tsk); /* * This prevents preemption while active_mm is being loaded and * it and mm are being updated, which could cause problems for diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 6e3bdf8e38bc..381d22eba088 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -782,6 +782,7 @@ struct vm_area_struct { struct mm_cid { u64 time; int cid; + int recent_cid; }; #endif @@ -852,6 +853,27 @@ struct mm_struct { * When the next mm_cid scan is due (in jiffies). */ unsigned long mm_cid_next_scan; + /** + * @nr_cpus_allowed: Number of CPUs allowed for mm. + * + * Number of CPUs allowed in the union of all mm's + * threads allowed CPUs. + */ + unsigned int nr_cpus_allowed; + /** + * @max_nr_cid: Maximum number of concurrency IDs allocated. + * + * Track the highest number of concurrency IDs allocated for the + * mm. + */ + atomic_t max_nr_cid; + /** + * @cpus_allowed_lock: Lock protecting mm cpus_allowed. + * + * Provide mutual exclusion for mm cpus_allowed and + * mm nr_cpus_allowed updates. + */ + raw_spinlock_t cpus_allowed_lock; #endif #ifdef CONFIG_MMU atomic_long_t pgtables_bytes; /* size of all page tables */ @@ -1170,18 +1192,30 @@ static inline int mm_cid_clear_lazy_put(int cid) return cid & ~MM_CID_LAZY_PUT; } +/* + * mm_cpus_allowed: Union of all mm's threads allowed CPUs. + */ +static inline cpumask_t *mm_cpus_allowed(struct mm_struct *mm) +{ + unsigned long bitmap = (unsigned long)mm; + + bitmap += offsetof(struct mm_struct, cpu_bitmap); + /* Skip cpu_bitmap */ + bitmap += cpumask_size(); + return (struct cpumask *)bitmap; +} + /* Accessor for struct mm_struct's cidmask. */ static inline cpumask_t *mm_cidmask(struct mm_struct *mm) { - unsigned long cid_bitmap = (unsigned long)mm; + unsigned long cid_bitmap = (unsigned long)mm_cpus_allowed(mm); - cid_bitmap += offsetof(struct mm_struct, cpu_bitmap); - /* Skip cpu_bitmap */ + /* Skip mm_cpus_allowed */ cid_bitmap += cpumask_size(); return (struct cpumask *)cid_bitmap; } -static inline void mm_init_cid(struct mm_struct *mm) +static inline void mm_init_cid(struct mm_struct *mm, struct task_struct *p) { int i; @@ -1189,17 +1223,22 @@ static inline void mm_init_cid(struct mm_struct *mm) struct mm_cid *pcpu_cid = per_cpu_ptr(mm->pcpu_cid, i); pcpu_cid->cid = MM_CID_UNSET; + pcpu_cid->recent_cid = MM_CID_UNSET; pcpu_cid->time = 0; } + mm->nr_cpus_allowed = p->nr_cpus_allowed; + atomic_set(&mm->max_nr_cid, 0); + raw_spin_lock_init(&mm->cpus_allowed_lock); + cpumask_copy(mm_cpus_allowed(mm), &p->cpus_mask); cpumask_clear(mm_cidmask(mm)); } -static inline int mm_alloc_cid_noprof(struct mm_struct *mm) +static inline int mm_alloc_cid_noprof(struct mm_struct *mm, struct task_struct *p) { mm->pcpu_cid = alloc_percpu_noprof(struct mm_cid); if (!mm->pcpu_cid) return -ENOMEM; - mm_init_cid(mm); + mm_init_cid(mm, p); return 0; } #define mm_alloc_cid(...) alloc_hooks(mm_alloc_cid_noprof(__VA_ARGS__)) @@ -1212,16 +1251,31 @@ static inline void mm_destroy_cid(struct mm_struct *mm) static inline unsigned int mm_cid_size(void) { - return cpumask_size(); + return 2 * cpumask_size(); /* mm_cpus_allowed(), mm_cidmask(). */ +} + +static inline void mm_set_cpus_allowed(struct mm_struct *mm, const struct cpumask *cpumask) +{ + struct cpumask *mm_allowed = mm_cpus_allowed(mm); + + if (!mm) + return; + /* The mm_cpus_allowed is the union of each thread allowed CPUs masks. */ + raw_spin_lock(&mm->cpus_allowed_lock); + cpumask_or(mm_allowed, mm_allowed, cpumask); + WRITE_ONCE(mm->nr_cpus_allowed, cpumask_weight(mm_allowed)); + raw_spin_unlock(&mm->cpus_allowed_lock); } #else /* CONFIG_SCHED_MM_CID */ -static inline void mm_init_cid(struct mm_struct *mm) { } -static inline int mm_alloc_cid(struct mm_struct *mm) { return 0; } +static inline void mm_init_cid(struct mm_struct *mm, struct task_struct *p) { } +static inline int mm_alloc_cid(struct mm_struct *mm, struct task_struct *p) { return 0; } static inline void mm_destroy_cid(struct mm_struct *mm) { } + static inline unsigned int mm_cid_size(void) { return 0; } +static inline void mm_set_cpus_allowed(struct mm_struct *mm, const struct cpumask *cpumask) { } #endif /* CONFIG_SCHED_MM_CID */ struct mmu_gather; diff --git a/kernel/fork.c b/kernel/fork.c index 89ceb4a68af2..7d950e93f080 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1298,7 +1298,7 @@ static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p, if (init_new_context(p, mm)) goto fail_nocontext; - if (mm_alloc_cid(mm)) + if (mm_alloc_cid(mm, p)) goto fail_cid; if (percpu_counter_init_many(mm->rss_stat, 0, GFP_KERNEL_ACCOUNT, diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 7db711ba6d12..f5ec452e2c5e 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2696,6 +2696,7 @@ __do_set_cpus_allowed(struct task_struct *p, struct affinity_context *ctx) put_prev_task(rq, p); p->sched_class->set_cpus_allowed(p, ctx); + mm_set_cpus_allowed(p->mm, ctx->new_mask); if (queued) enqueue_task(rq, p, ENQUEUE_RESTORE | ENQUEUE_NOCLOCK); @@ -10243,6 +10244,7 @@ int __sched_mm_cid_migrate_from_try_steal_cid(struct rq *src_rq, */ if (!try_cmpxchg(&src_pcpu_cid->cid, &lazy_cid, MM_CID_UNSET)) return -1; + WRITE_ONCE(src_pcpu_cid->recent_cid, MM_CID_UNSET); return src_cid; } @@ -10255,7 +10257,8 @@ void sched_mm_cid_migrate_to(struct rq *dst_rq, struct task_struct *t) { struct mm_cid *src_pcpu_cid, *dst_pcpu_cid; struct mm_struct *mm = t->mm; - int src_cid, dst_cid, src_cpu; + int src_cid, src_cpu; + bool dst_cid_is_set; struct rq *src_rq; lockdep_assert_rq_held(dst_rq); @@ -10272,9 +10275,9 @@ void sched_mm_cid_migrate_to(struct rq *dst_rq, struct task_struct *t) * allocation closest to 0 in cases where few threads migrate around * many CPUs. * - * If destination cid is already set, we may have to just clear - * the src cid to ensure compactness in frequent migrations - * scenarios. + * If destination cid or recent cid is already set, we may have + * to just clear the src cid to ensure compactness in frequent + * migrations scenarios. * * It is not useful to clear the src cid when the number of threads is * greater or equal to the number of allowed CPUs, because user-space @@ -10282,9 +10285,9 @@ void sched_mm_cid_migrate_to(struct rq *dst_rq, struct task_struct *t) * allowed CPUs. */ dst_pcpu_cid = per_cpu_ptr(mm->pcpu_cid, cpu_of(dst_rq)); - dst_cid = READ_ONCE(dst_pcpu_cid->cid); - if (!mm_cid_is_unset(dst_cid) && - atomic_read(&mm->mm_users) >= t->nr_cpus_allowed) + dst_cid_is_set = !mm_cid_is_unset(READ_ONCE(dst_pcpu_cid->cid)) || + !mm_cid_is_unset(READ_ONCE(dst_pcpu_cid->recent_cid)); + if (dst_cid_is_set && atomic_read(&mm->mm_users) >= READ_ONCE(mm->nr_cpus_allowed)) return; src_pcpu_cid = per_cpu_ptr(mm->pcpu_cid, src_cpu); src_rq = cpu_rq(src_cpu); @@ -10295,13 +10298,14 @@ void sched_mm_cid_migrate_to(struct rq *dst_rq, struct task_struct *t) src_cid); if (src_cid == -1) return; - if (!mm_cid_is_unset(dst_cid)) { + if (dst_cid_is_set) { __mm_cid_put(mm, src_cid); return; } /* Move src_cid to dst cpu. */ mm_cid_snapshot_time(dst_rq, mm); WRITE_ONCE(dst_pcpu_cid->cid, src_cid); + WRITE_ONCE(dst_pcpu_cid->recent_cid, src_cid); } static void sched_mm_cid_remote_clear(struct mm_struct *mm, struct mm_cid *pcpu_cid, @@ -10540,7 +10544,7 @@ void sched_mm_cid_after_execve(struct task_struct *t) * Matches barrier in sched_mm_cid_remote_clear_old(). */ smp_mb(); - t->last_mm_cid = t->mm_cid = mm_cid_get(rq, mm); + t->last_mm_cid = t->mm_cid = mm_cid_get(rq, t, mm); } rseq_set_notify_resume(t); } diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index fba524c81c63..20b6e75604ec 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -3596,24 +3596,41 @@ static inline void mm_cid_put(struct mm_struct *mm) __mm_cid_put(mm, mm_cid_clear_lazy_put(cid)); } -static inline int __mm_cid_try_get(struct mm_struct *mm) +static inline int __mm_cid_try_get(struct task_struct *t, struct mm_struct *mm) { - struct cpumask *cpumask; - int cid; + struct cpumask *cidmask = mm_cidmask(mm); + struct mm_cid __percpu *pcpu_cid = mm->pcpu_cid; + int cid = __this_cpu_read(pcpu_cid->recent_cid); - cpumask = mm_cidmask(mm); + /* Try to re-use recent cid. This improves cache locality. */ + if (!mm_cid_is_unset(cid) && !cpumask_test_and_set_cpu(cid, cidmask)) + return cid; + /* + * Expand cid allocation if the maximum number of concurrency + * IDs allocated (max_nr_cid) is below the number cpus allowed + * and number of threads. Expanding cid allocation as much as + * possible improves cache locality. + */ + cid = atomic_read(&mm->max_nr_cid); + while (cid < READ_ONCE(mm->nr_cpus_allowed) && cid < atomic_read(&mm->mm_users)) { + if (!atomic_try_cmpxchg(&mm->max_nr_cid, &cid, cid + 1)) + continue; + if (!cpumask_test_and_set_cpu(cid, cidmask)) + return cid; + } /* + * Find the first available concurrency id. * Retry finding first zero bit if the mask is temporarily * filled. This only happens during concurrent remote-clear * which owns a cid without holding a rq lock. */ for (;;) { - cid = cpumask_first_zero(cpumask); - if (cid < nr_cpu_ids) + cid = cpumask_first_zero(cidmask); + if (cid < READ_ONCE(mm->nr_cpus_allowed)) break; cpu_relax(); } - if (cpumask_test_and_set_cpu(cid, cpumask)) + if (cpumask_test_and_set_cpu(cid, cidmask)) return -1; return cid; @@ -3631,7 +3648,8 @@ static inline void mm_cid_snapshot_time(struct rq *rq, struct mm_struct *mm) WRITE_ONCE(pcpu_cid->time, rq->clock); } -static inline int __mm_cid_get(struct rq *rq, struct mm_struct *mm) +static inline int __mm_cid_get(struct rq *rq, struct task_struct *t, + struct mm_struct *mm) { int cid; @@ -3641,13 +3659,13 @@ static inline int __mm_cid_get(struct rq *rq, struct mm_struct *mm) * guarantee forward progress. */ if (!READ_ONCE(use_cid_lock)) { - cid = __mm_cid_try_get(mm); + cid = __mm_cid_try_get(t, mm); if (cid >= 0) goto end; raw_spin_lock(&cid_lock); } else { raw_spin_lock(&cid_lock); - cid = __mm_cid_try_get(mm); + cid = __mm_cid_try_get(t, mm); if (cid >= 0) goto unlock; } @@ -3667,7 +3685,7 @@ static inline int __mm_cid_get(struct rq *rq, struct mm_struct *mm) * all newcoming allocations observe the use_cid_lock flag set. */ do { - cid = __mm_cid_try_get(mm); + cid = __mm_cid_try_get(t, mm); cpu_relax(); } while (cid < 0); /* @@ -3684,7 +3702,8 @@ end: return cid; } -static inline int mm_cid_get(struct rq *rq, struct mm_struct *mm) +static inline int mm_cid_get(struct rq *rq, struct task_struct *t, + struct mm_struct *mm) { struct mm_cid __percpu *pcpu_cid = mm->pcpu_cid; struct cpumask *cpumask; @@ -3701,8 +3720,9 @@ static inline int mm_cid_get(struct rq *rq, struct mm_struct *mm) if (try_cmpxchg(&this_cpu_ptr(pcpu_cid)->cid, &cid, MM_CID_UNSET)) __mm_cid_put(mm, mm_cid_clear_lazy_put(cid)); } - cid = __mm_cid_get(rq, mm); + cid = __mm_cid_get(rq, t, mm); __this_cpu_write(pcpu_cid->cid, cid); + __this_cpu_write(pcpu_cid->recent_cid, cid); return cid; } @@ -3755,7 +3775,7 @@ static inline void switch_mm_cid(struct rq *rq, prev->mm_cid = -1; } if (next->mm_cid_active) - next->last_mm_cid = next->mm_cid = mm_cid_get(rq, next->mm); + next->last_mm_cid = next->mm_cid = mm_cid_get(rq, next, next->mm); } #else /* !CONFIG_SCHED_MM_CID: */ -- 2.51.0 From 894d1b3db41cf7e6ae0304429a1747b3c3f390bc Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 9 Oct 2024 16:53:34 -0700 Subject: [PATCH 10/16] locking/mutex: Remove wakeups from under mutex::wait_lock In preparation to nest mutex::wait_lock under rq::lock we need to remove wakeups from under it. Do this by utilizing wake_qs to defer the wakeup until after the lock is dropped. [Heavily changed after 55f036ca7e74 ("locking: WW mutex cleanup") and 08295b3b5bee ("locking: Implement an algorithm choice for Wound-Wait mutexes")] [jstultz: rebased to mainline, added extra wake_up_q & init to avoid hangs, similar to Connor's rework of this patch] Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Juri Lelli Signed-off-by: John Stultz Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Metin Kaya Acked-by: Davidlohr Bueso Tested-by: K Prateek Nayak Tested-by: Metin Kaya Link: https://lore.kernel.org/r/20241009235352.1614323-2-jstultz@google.com --- kernel/futex/pi.c | 6 +++- kernel/locking/mutex.c | 16 ++++++++--- kernel/locking/rtmutex.c | 51 ++++++++++++++++++++++++--------- kernel/locking/rtmutex_api.c | 12 ++++++-- kernel/locking/rtmutex_common.h | 3 +- kernel/locking/rwbase_rt.c | 8 +++++- kernel/locking/rwsem.c | 4 +-- kernel/locking/spinlock_rt.c | 5 ++-- kernel/locking/ww_mutex.h | 30 ++++++++++++------- 9 files changed, 96 insertions(+), 39 deletions(-) diff --git a/kernel/futex/pi.c b/kernel/futex/pi.c index 5722467f2737..d62cca5ed8f4 100644 --- a/kernel/futex/pi.c +++ b/kernel/futex/pi.c @@ -922,6 +922,7 @@ int futex_lock_pi(u32 __user *uaddr, unsigned int flags, ktime_t *time, int tryl struct rt_mutex_waiter rt_waiter; struct futex_hash_bucket *hb; struct futex_q q = futex_q_init; + DEFINE_WAKE_Q(wake_q); int res, ret; if (!IS_ENABLED(CONFIG_FUTEX_PI)) @@ -1018,8 +1019,11 @@ retry_private: * such that futex_unlock_pi() is guaranteed to observe the waiter when * it sees the futex_q::pi_state. */ - ret = __rt_mutex_start_proxy_lock(&q.pi_state->pi_mutex, &rt_waiter, current); + ret = __rt_mutex_start_proxy_lock(&q.pi_state->pi_mutex, &rt_waiter, current, &wake_q); + preempt_disable(); raw_spin_unlock_irq(&q.pi_state->pi_mutex.wait_lock); + wake_up_q(&wake_q); + preempt_enable(); if (ret) { if (ret == 1) diff --git a/kernel/locking/mutex.c b/kernel/locking/mutex.c index cbae8c0b89ab..6c94da061ec2 100644 --- a/kernel/locking/mutex.c +++ b/kernel/locking/mutex.c @@ -575,6 +575,7 @@ __mutex_lock_common(struct mutex *lock, unsigned int state, unsigned int subclas struct lockdep_map *nest_lock, unsigned long ip, struct ww_acquire_ctx *ww_ctx, const bool use_ww_ctx) { + DEFINE_WAKE_Q(wake_q); struct mutex_waiter waiter; struct ww_mutex *ww; int ret; @@ -625,7 +626,7 @@ __mutex_lock_common(struct mutex *lock, unsigned int state, unsigned int subclas */ if (__mutex_trylock(lock)) { if (ww_ctx) - __ww_mutex_check_waiters(lock, ww_ctx); + __ww_mutex_check_waiters(lock, ww_ctx, &wake_q); goto skip_wait; } @@ -645,7 +646,7 @@ __mutex_lock_common(struct mutex *lock, unsigned int state, unsigned int subclas * Add in stamp order, waking up waiters that must kill * themselves. */ - ret = __ww_mutex_add_waiter(&waiter, lock, ww_ctx); + ret = __ww_mutex_add_waiter(&waiter, lock, ww_ctx, &wake_q); if (ret) goto err_early_kill; } @@ -681,6 +682,10 @@ __mutex_lock_common(struct mutex *lock, unsigned int state, unsigned int subclas } raw_spin_unlock(&lock->wait_lock); + /* Make sure we do wakeups before calling schedule */ + wake_up_q(&wake_q); + wake_q_init(&wake_q); + schedule_preempt_disabled(); first = __mutex_waiter_is_first(lock, &waiter); @@ -714,7 +719,7 @@ acquired: */ if (!ww_ctx->is_wait_die && !__mutex_waiter_is_first(lock, &waiter)) - __ww_mutex_check_waiters(lock, ww_ctx); + __ww_mutex_check_waiters(lock, ww_ctx, &wake_q); } __mutex_remove_waiter(lock, &waiter); @@ -730,6 +735,7 @@ skip_wait: ww_mutex_lock_acquired(ww, ww_ctx); raw_spin_unlock(&lock->wait_lock); + wake_up_q(&wake_q); preempt_enable(); return 0; @@ -741,6 +747,7 @@ err_early_kill: raw_spin_unlock(&lock->wait_lock); debug_mutex_free_waiter(&waiter); mutex_release(&lock->dep_map, ip); + wake_up_q(&wake_q); preempt_enable(); return ret; } @@ -951,9 +958,10 @@ static noinline void __sched __mutex_unlock_slowpath(struct mutex *lock, unsigne if (owner & MUTEX_FLAG_HANDOFF) __mutex_handoff(lock, next); + preempt_disable(); raw_spin_unlock(&lock->wait_lock); - wake_up_q(&wake_q); + preempt_enable(); } #ifndef CONFIG_DEBUG_LOCK_ALLOC diff --git a/kernel/locking/rtmutex.c b/kernel/locking/rtmutex.c index ebebd0eec7f6..c7de80ee1f9d 100644 --- a/kernel/locking/rtmutex.c +++ b/kernel/locking/rtmutex.c @@ -34,13 +34,15 @@ static inline int __ww_mutex_add_waiter(struct rt_mutex_waiter *waiter, struct rt_mutex *lock, - struct ww_acquire_ctx *ww_ctx) + struct ww_acquire_ctx *ww_ctx, + struct wake_q_head *wake_q) { return 0; } static inline void __ww_mutex_check_waiters(struct rt_mutex *lock, - struct ww_acquire_ctx *ww_ctx) + struct ww_acquire_ctx *ww_ctx, + struct wake_q_head *wake_q) { } @@ -1201,7 +1203,8 @@ static int __sched task_blocks_on_rt_mutex(struct rt_mutex_base *lock, struct rt_mutex_waiter *waiter, struct task_struct *task, struct ww_acquire_ctx *ww_ctx, - enum rtmutex_chainwalk chwalk) + enum rtmutex_chainwalk chwalk, + struct wake_q_head *wake_q) { struct task_struct *owner = rt_mutex_owner(lock); struct rt_mutex_waiter *top_waiter = waiter; @@ -1245,7 +1248,10 @@ static int __sched task_blocks_on_rt_mutex(struct rt_mutex_base *lock, /* Check whether the waiter should back out immediately */ rtm = container_of(lock, struct rt_mutex, rtmutex); - res = __ww_mutex_add_waiter(waiter, rtm, ww_ctx); + preempt_disable(); + res = __ww_mutex_add_waiter(waiter, rtm, ww_ctx, wake_q); + wake_up_q(wake_q); + preempt_enable(); if (res) { raw_spin_lock(&task->pi_lock); rt_mutex_dequeue(lock, waiter); @@ -1674,12 +1680,14 @@ static void __sched rt_mutex_handle_deadlock(int res, int detect_deadlock, * @state: The task state for sleeping * @chwalk: Indicator whether full or partial chainwalk is requested * @waiter: Initializer waiter for blocking + * @wake_q: The wake_q to wake tasks after we release the wait_lock */ static int __sched __rt_mutex_slowlock(struct rt_mutex_base *lock, struct ww_acquire_ctx *ww_ctx, unsigned int state, enum rtmutex_chainwalk chwalk, - struct rt_mutex_waiter *waiter) + struct rt_mutex_waiter *waiter, + struct wake_q_head *wake_q) { struct rt_mutex *rtm = container_of(lock, struct rt_mutex, rtmutex); struct ww_mutex *ww = ww_container_of(rtm); @@ -1690,7 +1698,7 @@ static int __sched __rt_mutex_slowlock(struct rt_mutex_base *lock, /* Try to acquire the lock again: */ if (try_to_take_rt_mutex(lock, current, NULL)) { if (build_ww_mutex() && ww_ctx) { - __ww_mutex_check_waiters(rtm, ww_ctx); + __ww_mutex_check_waiters(rtm, ww_ctx, wake_q); ww_mutex_lock_acquired(ww, ww_ctx); } return 0; @@ -1700,7 +1708,7 @@ static int __sched __rt_mutex_slowlock(struct rt_mutex_base *lock, trace_contention_begin(lock, LCB_F_RT); - ret = task_blocks_on_rt_mutex(lock, waiter, current, ww_ctx, chwalk); + ret = task_blocks_on_rt_mutex(lock, waiter, current, ww_ctx, chwalk, wake_q); if (likely(!ret)) ret = rt_mutex_slowlock_block(lock, ww_ctx, state, NULL, waiter); @@ -1708,7 +1716,7 @@ static int __sched __rt_mutex_slowlock(struct rt_mutex_base *lock, /* acquired the lock */ if (build_ww_mutex() && ww_ctx) { if (!ww_ctx->is_wait_die) - __ww_mutex_check_waiters(rtm, ww_ctx); + __ww_mutex_check_waiters(rtm, ww_ctx, wake_q); ww_mutex_lock_acquired(ww, ww_ctx); } } else { @@ -1730,7 +1738,8 @@ static int __sched __rt_mutex_slowlock(struct rt_mutex_base *lock, static inline int __rt_mutex_slowlock_locked(struct rt_mutex_base *lock, struct ww_acquire_ctx *ww_ctx, - unsigned int state) + unsigned int state, + struct wake_q_head *wake_q) { struct rt_mutex_waiter waiter; int ret; @@ -1739,7 +1748,7 @@ static inline int __rt_mutex_slowlock_locked(struct rt_mutex_base *lock, waiter.ww_ctx = ww_ctx; ret = __rt_mutex_slowlock(lock, ww_ctx, state, RT_MUTEX_MIN_CHAINWALK, - &waiter); + &waiter, wake_q); debug_rt_mutex_free_waiter(&waiter); return ret; @@ -1755,6 +1764,7 @@ static int __sched rt_mutex_slowlock(struct rt_mutex_base *lock, struct ww_acquire_ctx *ww_ctx, unsigned int state) { + DEFINE_WAKE_Q(wake_q); unsigned long flags; int ret; @@ -1776,8 +1786,11 @@ static int __sched rt_mutex_slowlock(struct rt_mutex_base *lock, * irqsave/restore variants. */ raw_spin_lock_irqsave(&lock->wait_lock, flags); - ret = __rt_mutex_slowlock_locked(lock, ww_ctx, state); + ret = __rt_mutex_slowlock_locked(lock, ww_ctx, state, &wake_q); + preempt_disable(); raw_spin_unlock_irqrestore(&lock->wait_lock, flags); + wake_up_q(&wake_q); + preempt_enable(); rt_mutex_post_schedule(); return ret; @@ -1803,8 +1816,10 @@ static __always_inline int __rt_mutex_lock(struct rt_mutex_base *lock, /** * rtlock_slowlock_locked - Slow path lock acquisition for RT locks * @lock: The underlying RT mutex + * @wake_q: The wake_q to wake tasks after we release the wait_lock */ -static void __sched rtlock_slowlock_locked(struct rt_mutex_base *lock) +static void __sched rtlock_slowlock_locked(struct rt_mutex_base *lock, + struct wake_q_head *wake_q) { struct rt_mutex_waiter waiter; struct task_struct *owner; @@ -1821,7 +1836,7 @@ static void __sched rtlock_slowlock_locked(struct rt_mutex_base *lock) trace_contention_begin(lock, LCB_F_RT); - task_blocks_on_rt_mutex(lock, &waiter, current, NULL, RT_MUTEX_MIN_CHAINWALK); + task_blocks_on_rt_mutex(lock, &waiter, current, NULL, RT_MUTEX_MIN_CHAINWALK, wake_q); for (;;) { /* Try to acquire the lock again */ @@ -1832,7 +1847,11 @@ static void __sched rtlock_slowlock_locked(struct rt_mutex_base *lock) owner = rt_mutex_owner(lock); else owner = NULL; + preempt_disable(); raw_spin_unlock_irq(&lock->wait_lock); + wake_up_q(wake_q); + wake_q_init(wake_q); + preempt_enable(); if (!owner || !rtmutex_spin_on_owner(lock, &waiter, owner)) schedule_rtlock(); @@ -1857,10 +1876,14 @@ static void __sched rtlock_slowlock_locked(struct rt_mutex_base *lock) static __always_inline void __sched rtlock_slowlock(struct rt_mutex_base *lock) { unsigned long flags; + DEFINE_WAKE_Q(wake_q); raw_spin_lock_irqsave(&lock->wait_lock, flags); - rtlock_slowlock_locked(lock); + rtlock_slowlock_locked(lock, &wake_q); + preempt_disable(); raw_spin_unlock_irqrestore(&lock->wait_lock, flags); + wake_up_q(&wake_q); + preempt_enable(); } #endif /* RT_MUTEX_BUILD_SPINLOCKS */ diff --git a/kernel/locking/rtmutex_api.c b/kernel/locking/rtmutex_api.c index a6974d044593..2bc14c049a64 100644 --- a/kernel/locking/rtmutex_api.c +++ b/kernel/locking/rtmutex_api.c @@ -275,6 +275,7 @@ void __sched rt_mutex_proxy_unlock(struct rt_mutex_base *lock) * @lock: the rt_mutex to take * @waiter: the pre-initialized rt_mutex_waiter * @task: the task to prepare + * @wake_q: the wake_q to wake tasks after we release the wait_lock * * Starts the rt_mutex acquire; it enqueues the @waiter and does deadlock * detection. It does not wait, see rt_mutex_wait_proxy_lock() for that. @@ -291,7 +292,8 @@ void __sched rt_mutex_proxy_unlock(struct rt_mutex_base *lock) */ int __sched __rt_mutex_start_proxy_lock(struct rt_mutex_base *lock, struct rt_mutex_waiter *waiter, - struct task_struct *task) + struct task_struct *task, + struct wake_q_head *wake_q) { int ret; @@ -302,7 +304,7 @@ int __sched __rt_mutex_start_proxy_lock(struct rt_mutex_base *lock, /* We enforce deadlock detection for futexes */ ret = task_blocks_on_rt_mutex(lock, waiter, task, NULL, - RT_MUTEX_FULL_CHAINWALK); + RT_MUTEX_FULL_CHAINWALK, wake_q); if (ret && !rt_mutex_owner(lock)) { /* @@ -341,12 +343,16 @@ int __sched rt_mutex_start_proxy_lock(struct rt_mutex_base *lock, struct task_struct *task) { int ret; + DEFINE_WAKE_Q(wake_q); raw_spin_lock_irq(&lock->wait_lock); - ret = __rt_mutex_start_proxy_lock(lock, waiter, task); + ret = __rt_mutex_start_proxy_lock(lock, waiter, task, &wake_q); if (unlikely(ret)) remove_waiter(lock, waiter); + preempt_disable(); raw_spin_unlock_irq(&lock->wait_lock); + wake_up_q(&wake_q); + preempt_enable(); return ret; } diff --git a/kernel/locking/rtmutex_common.h b/kernel/locking/rtmutex_common.h index 1162e07cdaea..c38a2d2d4a7e 100644 --- a/kernel/locking/rtmutex_common.h +++ b/kernel/locking/rtmutex_common.h @@ -83,7 +83,8 @@ extern void rt_mutex_init_proxy_locked(struct rt_mutex_base *lock, extern void rt_mutex_proxy_unlock(struct rt_mutex_base *lock); extern int __rt_mutex_start_proxy_lock(struct rt_mutex_base *lock, struct rt_mutex_waiter *waiter, - struct task_struct *task); + struct task_struct *task, + struct wake_q_head *); extern int rt_mutex_start_proxy_lock(struct rt_mutex_base *lock, struct rt_mutex_waiter *waiter, struct task_struct *task); diff --git a/kernel/locking/rwbase_rt.c b/kernel/locking/rwbase_rt.c index 34a59569db6b..9f4322c07486 100644 --- a/kernel/locking/rwbase_rt.c +++ b/kernel/locking/rwbase_rt.c @@ -69,6 +69,7 @@ static int __sched __rwbase_read_lock(struct rwbase_rt *rwb, unsigned int state) { struct rt_mutex_base *rtm = &rwb->rtmutex; + DEFINE_WAKE_Q(wake_q); int ret; rwbase_pre_schedule(); @@ -110,7 +111,7 @@ static int __sched __rwbase_read_lock(struct rwbase_rt *rwb, * For rwlocks this returns 0 unconditionally, so the below * !ret conditionals are optimized out. */ - ret = rwbase_rtmutex_slowlock_locked(rtm, state); + ret = rwbase_rtmutex_slowlock_locked(rtm, state, &wake_q); /* * On success the rtmutex is held, so there can't be a writer @@ -121,7 +122,12 @@ static int __sched __rwbase_read_lock(struct rwbase_rt *rwb, */ if (!ret) atomic_inc(&rwb->readers); + + preempt_disable(); raw_spin_unlock_irq(&rtm->wait_lock); + wake_up_q(&wake_q); + preempt_enable(); + if (!ret) rwbase_rtmutex_unlock(rtm); diff --git a/kernel/locking/rwsem.c b/kernel/locking/rwsem.c index 2bbb6eca5144..2ddb827e3bea 100644 --- a/kernel/locking/rwsem.c +++ b/kernel/locking/rwsem.c @@ -1413,8 +1413,8 @@ static inline void __downgrade_write(struct rw_semaphore *sem) #define rwbase_rtmutex_lock_state(rtm, state) \ __rt_mutex_lock(rtm, state) -#define rwbase_rtmutex_slowlock_locked(rtm, state) \ - __rt_mutex_slowlock_locked(rtm, NULL, state) +#define rwbase_rtmutex_slowlock_locked(rtm, state, wq) \ + __rt_mutex_slowlock_locked(rtm, NULL, state, wq) #define rwbase_rtmutex_unlock(rtm) \ __rt_mutex_unlock(rtm) diff --git a/kernel/locking/spinlock_rt.c b/kernel/locking/spinlock_rt.c index 38e292454fcc..014143934e00 100644 --- a/kernel/locking/spinlock_rt.c +++ b/kernel/locking/spinlock_rt.c @@ -162,9 +162,10 @@ rwbase_rtmutex_lock_state(struct rt_mutex_base *rtm, unsigned int state) } static __always_inline int -rwbase_rtmutex_slowlock_locked(struct rt_mutex_base *rtm, unsigned int state) +rwbase_rtmutex_slowlock_locked(struct rt_mutex_base *rtm, unsigned int state, + struct wake_q_head *wake_q) { - rtlock_slowlock_locked(rtm); + rtlock_slowlock_locked(rtm, wake_q); return 0; } diff --git a/kernel/locking/ww_mutex.h b/kernel/locking/ww_mutex.h index 76d204b7d29c..a54bd16d0f17 100644 --- a/kernel/locking/ww_mutex.h +++ b/kernel/locking/ww_mutex.h @@ -275,7 +275,7 @@ __ww_ctx_less(struct ww_acquire_ctx *a, struct ww_acquire_ctx *b) */ static bool __ww_mutex_die(struct MUTEX *lock, struct MUTEX_WAITER *waiter, - struct ww_acquire_ctx *ww_ctx) + struct ww_acquire_ctx *ww_ctx, struct wake_q_head *wake_q) { if (!ww_ctx->is_wait_die) return false; @@ -284,7 +284,7 @@ __ww_mutex_die(struct MUTEX *lock, struct MUTEX_WAITER *waiter, #ifndef WW_RT debug_mutex_wake_waiter(lock, waiter); #endif - wake_up_process(waiter->task); + wake_q_add(wake_q, waiter->task); } return true; @@ -299,7 +299,8 @@ __ww_mutex_die(struct MUTEX *lock, struct MUTEX_WAITER *waiter, */ static bool __ww_mutex_wound(struct MUTEX *lock, struct ww_acquire_ctx *ww_ctx, - struct ww_acquire_ctx *hold_ctx) + struct ww_acquire_ctx *hold_ctx, + struct wake_q_head *wake_q) { struct task_struct *owner = __ww_mutex_owner(lock); @@ -331,7 +332,7 @@ static bool __ww_mutex_wound(struct MUTEX *lock, * wakeup pending to re-read the wounded state. */ if (owner != current) - wake_up_process(owner); + wake_q_add(wake_q, owner); return true; } @@ -352,7 +353,8 @@ static bool __ww_mutex_wound(struct MUTEX *lock, * The current task must not be on the wait list. */ static void -__ww_mutex_check_waiters(struct MUTEX *lock, struct ww_acquire_ctx *ww_ctx) +__ww_mutex_check_waiters(struct MUTEX *lock, struct ww_acquire_ctx *ww_ctx, + struct wake_q_head *wake_q) { struct MUTEX_WAITER *cur; @@ -364,8 +366,8 @@ __ww_mutex_check_waiters(struct MUTEX *lock, struct ww_acquire_ctx *ww_ctx) if (!cur->ww_ctx) continue; - if (__ww_mutex_die(lock, cur, ww_ctx) || - __ww_mutex_wound(lock, cur->ww_ctx, ww_ctx)) + if (__ww_mutex_die(lock, cur, ww_ctx, wake_q) || + __ww_mutex_wound(lock, cur->ww_ctx, ww_ctx, wake_q)) break; } } @@ -377,6 +379,8 @@ __ww_mutex_check_waiters(struct MUTEX *lock, struct ww_acquire_ctx *ww_ctx) static __always_inline void ww_mutex_set_context_fastpath(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) { + DEFINE_WAKE_Q(wake_q); + ww_mutex_lock_acquired(lock, ctx); /* @@ -405,8 +409,11 @@ ww_mutex_set_context_fastpath(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) * die or wound us. */ lock_wait_lock(&lock->base); - __ww_mutex_check_waiters(&lock->base, ctx); + __ww_mutex_check_waiters(&lock->base, ctx, &wake_q); + preempt_disable(); unlock_wait_lock(&lock->base); + wake_up_q(&wake_q); + preempt_enable(); } static __always_inline int @@ -488,7 +495,8 @@ __ww_mutex_check_kill(struct MUTEX *lock, struct MUTEX_WAITER *waiter, static inline int __ww_mutex_add_waiter(struct MUTEX_WAITER *waiter, struct MUTEX *lock, - struct ww_acquire_ctx *ww_ctx) + struct ww_acquire_ctx *ww_ctx, + struct wake_q_head *wake_q) { struct MUTEX_WAITER *cur, *pos = NULL; bool is_wait_die; @@ -532,7 +540,7 @@ __ww_mutex_add_waiter(struct MUTEX_WAITER *waiter, pos = cur; /* Wait-Die: ensure younger waiters die. */ - __ww_mutex_die(lock, cur, ww_ctx); + __ww_mutex_die(lock, cur, ww_ctx, wake_q); } __ww_waiter_add(lock, waiter, pos); @@ -550,7 +558,7 @@ __ww_mutex_add_waiter(struct MUTEX_WAITER *waiter, * such that either we or the fastpath will wound @ww->ctx. */ smp_mb(); - __ww_mutex_wound(lock, ww_ctx, ww->ctx); + __ww_mutex_wound(lock, ww_ctx, ww->ctx, wake_q); } return 0; -- 2.51.0 From 5ec58525a1f1bd6ca8ea778e9df55cd82bc02e11 Mon Sep 17 00:00:00 2001 From: Juri Lelli Date: Wed, 9 Oct 2024 16:53:35 -0700 Subject: [PATCH 11/16] locking/mutex: Make mutex::wait_lock irq safe With the proxy-execution series, we traverse the task->mutex->task blocked_on/owner chain in the scheduler core. We do this while holding the rq::lock to keep the structures in place while taking and releasing the alternating lock types. Since the mutex::wait_lock is one of the locks we will take in this way under the rq::lock in the scheduler core, we need to make sure that its usage elsewhere is irq safe. [rebase & fix {un,}lock_wait_lock helpers in ww_mutex.h] Signed-off-by: Juri Lelli Signed-off-by: Connor O'Brien Signed-off-by: John Stultz Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Metin Kaya Reviewed-by: Valentin Schneider Tested-by: K Prateek Nayak Tested-by: Metin Kaya Link: https://lore.kernel.org/r/20241009235352.1614323-3-jstultz@google.com --- kernel/locking/mutex.c | 18 ++++++++++-------- kernel/locking/ww_mutex.h | 21 +++++++++++---------- 2 files changed, 21 insertions(+), 18 deletions(-) diff --git a/kernel/locking/mutex.c b/kernel/locking/mutex.c index 6c94da061ec2..cd248d1767eb 100644 --- a/kernel/locking/mutex.c +++ b/kernel/locking/mutex.c @@ -578,6 +578,7 @@ __mutex_lock_common(struct mutex *lock, unsigned int state, unsigned int subclas DEFINE_WAKE_Q(wake_q); struct mutex_waiter waiter; struct ww_mutex *ww; + unsigned long flags; int ret; if (!use_ww_ctx) @@ -620,7 +621,7 @@ __mutex_lock_common(struct mutex *lock, unsigned int state, unsigned int subclas return 0; } - raw_spin_lock(&lock->wait_lock); + raw_spin_lock_irqsave(&lock->wait_lock, flags); /* * After waiting to acquire the wait_lock, try again. */ @@ -681,7 +682,7 @@ __mutex_lock_common(struct mutex *lock, unsigned int state, unsigned int subclas goto err; } - raw_spin_unlock(&lock->wait_lock); + raw_spin_unlock_irqrestore(&lock->wait_lock, flags); /* Make sure we do wakeups before calling schedule */ wake_up_q(&wake_q); wake_q_init(&wake_q); @@ -706,9 +707,9 @@ __mutex_lock_common(struct mutex *lock, unsigned int state, unsigned int subclas trace_contention_begin(lock, LCB_F_MUTEX); } - raw_spin_lock(&lock->wait_lock); + raw_spin_lock_irqsave(&lock->wait_lock, flags); } - raw_spin_lock(&lock->wait_lock); + raw_spin_lock_irqsave(&lock->wait_lock, flags); acquired: __set_current_state(TASK_RUNNING); @@ -734,7 +735,7 @@ skip_wait: if (ww_ctx) ww_mutex_lock_acquired(ww, ww_ctx); - raw_spin_unlock(&lock->wait_lock); + raw_spin_unlock_irqrestore(&lock->wait_lock, flags); wake_up_q(&wake_q); preempt_enable(); return 0; @@ -744,7 +745,7 @@ err: __mutex_remove_waiter(lock, &waiter); err_early_kill: trace_contention_end(lock, ret); - raw_spin_unlock(&lock->wait_lock); + raw_spin_unlock_irqrestore(&lock->wait_lock, flags); debug_mutex_free_waiter(&waiter); mutex_release(&lock->dep_map, ip); wake_up_q(&wake_q); @@ -915,6 +916,7 @@ static noinline void __sched __mutex_unlock_slowpath(struct mutex *lock, unsigne struct task_struct *next = NULL; DEFINE_WAKE_Q(wake_q); unsigned long owner; + unsigned long flags; mutex_release(&lock->dep_map, ip); @@ -941,7 +943,7 @@ static noinline void __sched __mutex_unlock_slowpath(struct mutex *lock, unsigne } } - raw_spin_lock(&lock->wait_lock); + raw_spin_lock_irqsave(&lock->wait_lock, flags); debug_mutex_unlock(lock); if (!list_empty(&lock->wait_list)) { /* get the first entry from the wait-list: */ @@ -959,7 +961,7 @@ static noinline void __sched __mutex_unlock_slowpath(struct mutex *lock, unsigne __mutex_handoff(lock, next); preempt_disable(); - raw_spin_unlock(&lock->wait_lock); + raw_spin_unlock_irqrestore(&lock->wait_lock, flags); wake_up_q(&wake_q); preempt_enable(); } diff --git a/kernel/locking/ww_mutex.h b/kernel/locking/ww_mutex.h index a54bd16d0f17..37f025a096c9 100644 --- a/kernel/locking/ww_mutex.h +++ b/kernel/locking/ww_mutex.h @@ -70,14 +70,14 @@ __ww_mutex_has_waiters(struct mutex *lock) return atomic_long_read(&lock->owner) & MUTEX_FLAG_WAITERS; } -static inline void lock_wait_lock(struct mutex *lock) +static inline void lock_wait_lock(struct mutex *lock, unsigned long *flags) { - raw_spin_lock(&lock->wait_lock); + raw_spin_lock_irqsave(&lock->wait_lock, *flags); } -static inline void unlock_wait_lock(struct mutex *lock) +static inline void unlock_wait_lock(struct mutex *lock, unsigned long *flags) { - raw_spin_unlock(&lock->wait_lock); + raw_spin_unlock_irqrestore(&lock->wait_lock, *flags); } static inline void lockdep_assert_wait_lock_held(struct mutex *lock) @@ -144,14 +144,14 @@ __ww_mutex_has_waiters(struct rt_mutex *lock) return rt_mutex_has_waiters(&lock->rtmutex); } -static inline void lock_wait_lock(struct rt_mutex *lock) +static inline void lock_wait_lock(struct rt_mutex *lock, unsigned long *flags) { - raw_spin_lock(&lock->rtmutex.wait_lock); + raw_spin_lock_irqsave(&lock->rtmutex.wait_lock, *flags); } -static inline void unlock_wait_lock(struct rt_mutex *lock) +static inline void unlock_wait_lock(struct rt_mutex *lock, unsigned long *flags) { - raw_spin_unlock(&lock->rtmutex.wait_lock); + raw_spin_unlock_irqrestore(&lock->rtmutex.wait_lock, *flags); } static inline void lockdep_assert_wait_lock_held(struct rt_mutex *lock) @@ -380,6 +380,7 @@ static __always_inline void ww_mutex_set_context_fastpath(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) { DEFINE_WAKE_Q(wake_q); + unsigned long flags; ww_mutex_lock_acquired(lock, ctx); @@ -408,10 +409,10 @@ ww_mutex_set_context_fastpath(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) * Uh oh, we raced in fastpath, check if any of the waiters need to * die or wound us. */ - lock_wait_lock(&lock->base); + lock_wait_lock(&lock->base, &flags); __ww_mutex_check_waiters(&lock->base, ctx, &wake_q); preempt_disable(); - unlock_wait_lock(&lock->base); + unlock_wait_lock(&lock->base, &flags); wake_up_q(&wake_q); preempt_enable(); } -- 2.51.0 From 3a9320ecb06c6c5ca5a8a595717e5186b5f20141 Mon Sep 17 00:00:00 2001 From: Juri Lelli Date: Wed, 9 Oct 2024 16:53:36 -0700 Subject: [PATCH 12/16] locking/mutex: Expose __mutex_owner() Implementing proxy execution requires that scheduler code be able to identify the current owner of a mutex. Expose __mutex_owner() for this purpose (alone!). Includes a null mutex check, so that users of the function can be simplified. [Removed the EXPORT_SYMBOL] [jstultz: Reworked per Peter's suggestions] Signed-off-by: Juri Lelli Signed-off-by: Valentin Schneider Signed-off-by: Connor O'Brien Signed-off-by: John Stultz Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Metin Kaya Reviewed-by: Valentin Schneider Tested-by: K Prateek Nayak Tested-by: Metin Kaya Link: https://lore.kernel.org/r/20241009235352.1614323-4-jstultz@google.com --- kernel/locking/mutex.c | 25 ------------------------- kernel/locking/mutex.h | 27 +++++++++++++++++++++++++++ 2 files changed, 27 insertions(+), 25 deletions(-) diff --git a/kernel/locking/mutex.c b/kernel/locking/mutex.c index cd248d1767eb..3302e52f0c96 100644 --- a/kernel/locking/mutex.c +++ b/kernel/locking/mutex.c @@ -56,31 +56,6 @@ __mutex_init(struct mutex *lock, const char *name, struct lock_class_key *key) } EXPORT_SYMBOL(__mutex_init); -/* - * @owner: contains: 'struct task_struct *' to the current lock owner, - * NULL means not owned. Since task_struct pointers are aligned at - * at least L1_CACHE_BYTES, we have low bits to store extra state. - * - * Bit0 indicates a non-empty waiter list; unlock must issue a wakeup. - * Bit1 indicates unlock needs to hand the lock to the top-waiter - * Bit2 indicates handoff has been done and we're waiting for pickup. - */ -#define MUTEX_FLAG_WAITERS 0x01 -#define MUTEX_FLAG_HANDOFF 0x02 -#define MUTEX_FLAG_PICKUP 0x04 - -#define MUTEX_FLAGS 0x07 - -/* - * Internal helper function; C doesn't allow us to hide it :/ - * - * DO NOT USE (outside of mutex code). - */ -static inline struct task_struct *__mutex_owner(struct mutex *lock) -{ - return (struct task_struct *)(atomic_long_read(&lock->owner) & ~MUTEX_FLAGS); -} - static inline struct task_struct *__owner_task(unsigned long owner) { return (struct task_struct *)(owner & ~MUTEX_FLAGS); diff --git a/kernel/locking/mutex.h b/kernel/locking/mutex.h index 0b2a79c4013b..cbff35b9b7ae 100644 --- a/kernel/locking/mutex.h +++ b/kernel/locking/mutex.h @@ -20,6 +20,33 @@ struct mutex_waiter { #endif }; +/* + * @owner: contains: 'struct task_struct *' to the current lock owner, + * NULL means not owned. Since task_struct pointers are aligned at + * at least L1_CACHE_BYTES, we have low bits to store extra state. + * + * Bit0 indicates a non-empty waiter list; unlock must issue a wakeup. + * Bit1 indicates unlock needs to hand the lock to the top-waiter + * Bit2 indicates handoff has been done and we're waiting for pickup. + */ +#define MUTEX_FLAG_WAITERS 0x01 +#define MUTEX_FLAG_HANDOFF 0x02 +#define MUTEX_FLAG_PICKUP 0x04 + +#define MUTEX_FLAGS 0x07 + +/* + * Internal helper function; C doesn't allow us to hide it :/ + * + * DO NOT USE (outside of mutex & scheduler code). + */ +static inline struct task_struct *__mutex_owner(struct mutex *lock) +{ + if (!lock) + return NULL; + return (struct task_struct *)(atomic_long_read(&lock->owner) & ~MUTEX_FLAGS); +} + #ifdef CONFIG_DEBUG_MUTEXES extern void debug_mutex_lock_common(struct mutex *lock, struct mutex_waiter *waiter); -- 2.51.0 From 2b05a0b4c08ffd6dedfbd27af8708742cde39b95 Mon Sep 17 00:00:00 2001 From: Connor O'Brien Date: Wed, 9 Oct 2024 16:53:37 -0700 Subject: [PATCH 13/16] sched: Add move_queued_task_locked helper Switch logic that deactivates, sets the task cpu, and reactivates a task on a different rq to use a helper that will be later extended to push entire blocked task chains. This patch was broken out from a larger chain migration patch originally by Connor O'Brien. [jstultz: split out from larger chain migration patch] Signed-off-by: Connor O'Brien Signed-off-by: John Stultz Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Metin Kaya Reviewed-by: Valentin Schneider Reviewed-by: Qais Yousef Tested-by: K Prateek Nayak Tested-by: Metin Kaya Link: https://lore.kernel.org/r/20241009235352.1614323-5-jstultz@google.com --- kernel/sched/core.c | 13 +++---------- kernel/sched/deadline.c | 8 ++------ kernel/sched/rt.c | 8 ++------ kernel/sched/sched.h | 12 ++++++++++++ 4 files changed, 19 insertions(+), 22 deletions(-) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index f5ec452e2c5e..ab0b775a292d 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2620,9 +2620,7 @@ int push_cpu_stop(void *arg) // XXX validate p is still the highest prio task if (task_rq(p) == rq) { - deactivate_task(rq, p, 0); - set_task_cpu(p, lowest_rq->cpu); - activate_task(lowest_rq, p, 0); + move_queued_task_locked(rq, lowest_rq, p); resched_curr(lowest_rq); } @@ -3309,9 +3307,7 @@ static void __migrate_swap_task(struct task_struct *p, int cpu) rq_pin_lock(src_rq, &srf); rq_pin_lock(dst_rq, &drf); - deactivate_task(src_rq, p, 0); - set_task_cpu(p, cpu); - activate_task(dst_rq, p, 0); + move_queued_task_locked(src_rq, dst_rq, p); wakeup_preempt(dst_rq, p, 0); rq_unpin_lock(dst_rq, &drf); @@ -6300,10 +6296,7 @@ static bool try_steal_cookie(int this, int that) if (sched_task_is_throttled(p, this)) goto next; - deactivate_task(src, p, 0); - set_task_cpu(p, this); - activate_task(dst, p, 0); - + move_queued_task_locked(src, dst, p); resched_curr(dst); success = true; diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c index be1b917dc8ce..4acf5e3a5005 100644 --- a/kernel/sched/deadline.c +++ b/kernel/sched/deadline.c @@ -2751,9 +2751,7 @@ retry: goto retry; } - deactivate_task(rq, next_task, 0); - set_task_cpu(next_task, later_rq->cpu); - activate_task(later_rq, next_task, 0); + move_queued_task_locked(rq, later_rq, next_task); ret = 1; resched_curr(later_rq); @@ -2839,9 +2837,7 @@ static void pull_dl_task(struct rq *this_rq) if (is_migration_disabled(p)) { push_task = get_push_task(src_rq); } else { - deactivate_task(src_rq, p, 0); - set_task_cpu(p, this_cpu); - activate_task(this_rq, p, 0); + move_queued_task_locked(src_rq, this_rq, p); dmin = p->dl.deadline; resched = true; } diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 172c588de542..e2506ab33c97 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -2088,9 +2088,7 @@ retry: goto retry; } - deactivate_task(rq, next_task, 0); - set_task_cpu(next_task, lowest_rq->cpu); - activate_task(lowest_rq, next_task, 0); + move_queued_task_locked(rq, lowest_rq, next_task); resched_curr(lowest_rq); ret = 1; @@ -2361,9 +2359,7 @@ static void pull_rt_task(struct rq *this_rq) if (is_migration_disabled(p)) { push_task = get_push_task(src_rq); } else { - deactivate_task(src_rq, p, 0); - set_task_cpu(p, this_cpu); - activate_task(this_rq, p, 0); + move_queued_task_locked(src_rq, this_rq, p); resched = true; } /* diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 20b6e75604ec..71ce1b05ac94 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -3788,6 +3788,18 @@ static inline void init_sched_mm_cid(struct task_struct *t) { } extern u64 avg_vruntime(struct cfs_rq *cfs_rq); extern int entity_eligible(struct cfs_rq *cfs_rq, struct sched_entity *se); +#ifdef CONFIG_SMP +static inline +void move_queued_task_locked(struct rq *src_rq, struct rq *dst_rq, struct task_struct *task) +{ + lockdep_assert_rq_held(src_rq); + lockdep_assert_rq_held(dst_rq); + + deactivate_task(src_rq, task, 0); + set_task_cpu(task, dst_rq->cpu); + activate_task(dst_rq, task, 0); +} +#endif #ifdef CONFIG_RT_MUTEXES -- 2.51.0 From 18adad1dac3334ed34f60ad4de2960df03058142 Mon Sep 17 00:00:00 2001 From: Connor O'Brien Date: Wed, 9 Oct 2024 16:53:38 -0700 Subject: [PATCH 14/16] sched: Consolidate pick_*_task to task_is_pushable helper This patch consolidates rt and deadline pick_*_task functions to a task_is_pushable() helper This patch was broken out from a larger chain migration patch originally by Connor O'Brien. [jstultz: split out from larger chain migration patch, renamed helper function] Signed-off-by: Connor O'Brien Signed-off-by: John Stultz Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Metin Kaya Reviewed-by: Valentin Schneider Reviewed-by: Christian Loehle Tested-by: K Prateek Nayak Tested-by: Metin Kaya Link: https://lore.kernel.org/r/20241009235352.1614323-6-jstultz@google.com --- kernel/sched/deadline.c | 10 +--------- kernel/sched/rt.c | 11 +---------- kernel/sched/sched.h | 10 ++++++++++ 3 files changed, 12 insertions(+), 19 deletions(-) diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c index 4acf5e3a5005..a4683f8fe685 100644 --- a/kernel/sched/deadline.c +++ b/kernel/sched/deadline.c @@ -2487,14 +2487,6 @@ static void task_fork_dl(struct task_struct *p) /* Only try algorithms three times */ #define DL_MAX_TRIES 3 -static int pick_dl_task(struct rq *rq, struct task_struct *p, int cpu) -{ - if (!task_on_cpu(rq, p) && - cpumask_test_cpu(cpu, &p->cpus_mask)) - return 1; - return 0; -} - /* * Return the earliest pushable rq's task, which is suitable to be executed * on the CPU, NULL otherwise: @@ -2513,7 +2505,7 @@ next_node: if (next_node) { p = __node_2_pdl(next_node); - if (pick_dl_task(rq, p, cpu)) + if (task_is_pushable(rq, p, cpu)) return p; next_node = rb_next(next_node); diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index e2506ab33c97..c5c22fc51824 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -1773,15 +1773,6 @@ static void put_prev_task_rt(struct rq *rq, struct task_struct *p, struct task_s /* Only try algorithms three times */ #define RT_MAX_TRIES 3 -static int pick_rt_task(struct rq *rq, struct task_struct *p, int cpu) -{ - if (!task_on_cpu(rq, p) && - cpumask_test_cpu(cpu, &p->cpus_mask)) - return 1; - - return 0; -} - /* * Return the highest pushable rq's task, which is suitable to be executed * on the CPU, NULL otherwise @@ -1795,7 +1786,7 @@ static struct task_struct *pick_highest_pushable_task(struct rq *rq, int cpu) return NULL; plist_for_each_entry(p, head, pushable_tasks) { - if (pick_rt_task(rq, p, cpu)) + if (task_is_pushable(rq, p, cpu)) return p; } diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 71ce1b05ac94..449335299d2f 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -3799,6 +3799,16 @@ void move_queued_task_locked(struct rq *src_rq, struct rq *dst_rq, struct task_s set_task_cpu(task, dst_rq->cpu); activate_task(dst_rq, task, 0); } + +static inline +bool task_is_pushable(struct rq *rq, struct task_struct *p, int cpu) +{ + if (!task_on_cpu(rq, p) && + cpumask_test_cpu(cpu, &p->cpus_mask)) + return true; + + return false; +} #endif #ifdef CONFIG_RT_MUTEXES -- 2.51.0 From 7b3d61f6578ab06f130ecc13cd2f3010a6c295bb Mon Sep 17 00:00:00 2001 From: John Stultz Date: Wed, 9 Oct 2024 16:53:39 -0700 Subject: [PATCH 15/16] sched: Split out __schedule() deactivate task logic into a helper As we're going to re-use the deactivation logic, split it into a helper. Signed-off-by: John Stultz Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Metin Kaya Reviewed-by: Qais Yousef Tested-by: K Prateek Nayak Tested-by: Metin Kaya Link: https://lore.kernel.org/r/20241009235352.1614323-7-jstultz@google.com --- kernel/sched/core.c | 67 +++++++++++++++++++++++++++------------------ 1 file changed, 40 insertions(+), 27 deletions(-) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index ab0b775a292d..b534de6e543a 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -6490,6 +6490,45 @@ pick_next_task(struct rq *rq, struct task_struct *prev, struct rq_flags *rf) #define SM_PREEMPT 1 #define SM_RTLOCK_WAIT 2 +/* + * Helper function for __schedule() + * + * If a task does not have signals pending, deactivate it + * Otherwise marks the task's __state as RUNNING + */ +static bool try_to_block_task(struct rq *rq, struct task_struct *p, + unsigned long task_state) +{ + int flags = DEQUEUE_NOCLOCK; + + if (signal_pending_state(task_state, p)) { + WRITE_ONCE(p->__state, TASK_RUNNING); + return false; + } + + p->sched_contributes_to_load = + (task_state & TASK_UNINTERRUPTIBLE) && + !(task_state & TASK_NOLOAD) && + !(task_state & TASK_FROZEN); + + if (unlikely(is_special_task_state(task_state))) + flags |= DEQUEUE_SPECIAL; + + /* + * __schedule() ttwu() + * prev_state = prev->state; if (p->on_rq && ...) + * if (prev_state) goto out; + * p->on_rq = 0; smp_acquire__after_ctrl_dep(); + * p->state = TASK_WAKING + * + * Where __schedule() and ttwu() have matching control dependencies. + * + * After this, schedule() must not care about p->state any more. + */ + block_task(rq, p, flags); + return true; +} + /* * __schedule() is the main scheduler function. * @@ -6598,33 +6637,7 @@ static void __sched notrace __schedule(int sched_mode) goto picked; } } else if (!preempt && prev_state) { - if (signal_pending_state(prev_state, prev)) { - WRITE_ONCE(prev->__state, TASK_RUNNING); - } else { - int flags = DEQUEUE_NOCLOCK; - - prev->sched_contributes_to_load = - (prev_state & TASK_UNINTERRUPTIBLE) && - !(prev_state & TASK_NOLOAD) && - !(prev_state & TASK_FROZEN); - - if (unlikely(is_special_task_state(prev_state))) - flags |= DEQUEUE_SPECIAL; - - /* - * __schedule() ttwu() - * prev_state = prev->state; if (p->on_rq && ...) - * if (prev_state) goto out; - * p->on_rq = 0; smp_acquire__after_ctrl_dep(); - * p->state = TASK_WAKING - * - * Where __schedule() and ttwu() have matching control dependencies. - * - * After this, schedule() must not care about p->state any more. - */ - block_task(rq, prev, flags); - block = true; - } + block = try_to_block_task(rq, prev, prev_state); switch_count = &prev->nvcsw; } -- 2.51.0 From af0c8b2bf67b25756f27644936e74fd9a6273bd2 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 9 Oct 2024 16:53:40 -0700 Subject: [PATCH 16/16] sched: Split scheduler and execution contexts Let's define the "scheduling context" as all the scheduler state in task_struct for the task chosen to run, which we'll call the donor task, and the "execution context" as all state required to actually run the task. Currently both are intertwined in task_struct. We want to logically split these such that we can use the scheduling context of the donor task selected to be scheduled, but use the execution context of a different task to actually be run. To this purpose, introduce rq->donor field to point to the task_struct chosen from the runqueue by the scheduler, and will be used for scheduler state, and preserve rq->curr to indicate the execution context of the task that will actually be run. This patch introduces the donor field as a union with curr, so it doesn't cause the contexts to be split yet, but adds the logic to handle everything separately. [add additional comments and update more sched_class code to use rq::proxy] [jstultz: Rebased and resolved minor collisions, reworked to use accessors, tweaked update_curr_common to use rq_proxy fixing rt scheduling issues] Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Juri Lelli Signed-off-by: Connor O'Brien Signed-off-by: John Stultz Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Metin Kaya Tested-by: K Prateek Nayak Tested-by: Metin Kaya Link: https://lore.kernel.org/r/20241009235352.1614323-8-jstultz@google.com --- kernel/sched/core.c | 45 +++++++++++++++++++++++--------------- kernel/sched/deadline.c | 39 +++++++++++++++++---------------- kernel/sched/fair.c | 28 ++++++++++++------------ kernel/sched/pelt.c | 2 +- kernel/sched/rt.c | 48 ++++++++++++++++++++--------------------- kernel/sched/sched.h | 28 +++++++++++++++++++++--- kernel/sched/syscalls.c | 4 ++-- 7 files changed, 114 insertions(+), 80 deletions(-) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index b534de6e543a..dcfc4c473284 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -832,7 +832,7 @@ static enum hrtimer_restart hrtick(struct hrtimer *timer) rq_lock(rq, &rf); update_rq_clock(rq); - rq->curr->sched_class->task_tick(rq, rq->curr, 1); + rq->donor->sched_class->task_tick(rq, rq->curr, 1); rq_unlock(rq, &rf); return HRTIMER_NORESTART; @@ -2135,16 +2135,18 @@ void check_class_changed(struct rq *rq, struct task_struct *p, void wakeup_preempt(struct rq *rq, struct task_struct *p, int flags) { - if (p->sched_class == rq->curr->sched_class) - rq->curr->sched_class->wakeup_preempt(rq, p, flags); - else if (sched_class_above(p->sched_class, rq->curr->sched_class)) + struct task_struct *donor = rq->donor; + + if (p->sched_class == donor->sched_class) + donor->sched_class->wakeup_preempt(rq, p, flags); + else if (sched_class_above(p->sched_class, donor->sched_class)) resched_curr(rq); /* * A queue event has occurred, and we're going to schedule. In * this case, we can save a useless back to back clock update. */ - if (task_on_rq_queued(rq->curr) && test_tsk_need_resched(rq->curr)) + if (task_on_rq_queued(donor) && test_tsk_need_resched(rq->curr)) rq_clock_skip_update(rq); } @@ -2680,7 +2682,7 @@ __do_set_cpus_allowed(struct task_struct *p, struct affinity_context *ctx) lockdep_assert_held(&p->pi_lock); queued = task_on_rq_queued(p); - running = task_current(rq, p); + running = task_current_donor(rq, p); if (queued) { /* @@ -5507,7 +5509,7 @@ unsigned long long task_sched_runtime(struct task_struct *p) * project cycles that may never be accounted to this * thread, breaking clock_gettime(). */ - if (task_current(rq, p) && task_on_rq_queued(p)) { + if (task_current_donor(rq, p) && task_on_rq_queued(p)) { prefetch_curr_exec_start(p); update_rq_clock(rq); p->sched_class->update_curr(rq); @@ -5575,7 +5577,8 @@ void sched_tick(void) { int cpu = smp_processor_id(); struct rq *rq = cpu_rq(cpu); - struct task_struct *curr; + /* accounting goes to the donor task */ + struct task_struct *donor; struct rq_flags rf; unsigned long hw_pressure; u64 resched_latency; @@ -5586,19 +5589,19 @@ void sched_tick(void) sched_clock_tick(); rq_lock(rq, &rf); + donor = rq->donor; - curr = rq->curr; - psi_account_irqtime(rq, curr, NULL); + psi_account_irqtime(rq, donor, NULL); update_rq_clock(rq); hw_pressure = arch_scale_hw_pressure(cpu_of(rq)); update_hw_load_avg(rq_clock_task(rq), rq, hw_pressure); - curr->sched_class->task_tick(rq, curr, 0); + donor->sched_class->task_tick(rq, donor, 0); if (sched_feat(LATENCY_WARN)) resched_latency = cpu_resched_latency(rq); calc_global_load_tick(rq); sched_core_tick(rq); - task_tick_mm_cid(rq, curr); + task_tick_mm_cid(rq, donor); scx_tick(rq); rq_unlock(rq, &rf); @@ -5608,8 +5611,8 @@ void sched_tick(void) perf_event_task_tick(); - if (curr->flags & PF_WQ_WORKER) - wq_worker_tick(curr); + if (donor->flags & PF_WQ_WORKER) + wq_worker_tick(donor); #ifdef CONFIG_SMP if (!scx_switched_all()) { @@ -5676,6 +5679,12 @@ static void sched_tick_remote(struct work_struct *work) struct task_struct *curr = rq->curr; if (cpu_online(cpu)) { + /* + * Since this is a remote tick for full dynticks mode, + * we are always sure that there is no proxy (only a + * single task is running). + */ + SCHED_WARN_ON(rq->curr != rq->donor); update_rq_clock(rq); if (!is_idle_task(curr)) { @@ -6642,6 +6651,7 @@ static void __sched notrace __schedule(int sched_mode) } next = pick_next_task(rq, prev, &rf); + rq_set_donor(rq, next); picked: clear_tsk_need_resched(prev); clear_preempt_need_resched(); @@ -7148,7 +7158,7 @@ void rt_mutex_setprio(struct task_struct *p, struct task_struct *pi_task) dequeue_task(rq, p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK); queued = task_on_rq_queued(p); - running = task_current(rq, p); + running = task_current_donor(rq, p); if (queued) dequeue_task(rq, p, queue_flag); if (running) @@ -7718,6 +7728,7 @@ void __init init_idle(struct task_struct *idle, int cpu) rcu_read_unlock(); rq->idle = idle; + rq_set_donor(rq, idle); rcu_assign_pointer(rq->curr, idle); idle->on_rq = TASK_ON_RQ_QUEUED; #ifdef CONFIG_SMP @@ -7807,7 +7818,7 @@ void sched_setnuma(struct task_struct *p, int nid) rq = task_rq_lock(p, &rf); queued = task_on_rq_queued(p); - running = task_current(rq, p); + running = task_current_donor(rq, p); if (queued) dequeue_task(rq, p, DEQUEUE_SAVE); @@ -8957,7 +8968,7 @@ void sched_move_task(struct task_struct *tsk) update_rq_clock(rq); - running = task_current(rq, tsk); + running = task_current_donor(rq, tsk); queued = task_on_rq_queued(tsk); if (queued) diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c index a4683f8fe685..d9d5a702f1a6 100644 --- a/kernel/sched/deadline.c +++ b/kernel/sched/deadline.c @@ -1339,7 +1339,7 @@ static enum hrtimer_restart dl_task_timer(struct hrtimer *timer) #endif enqueue_task_dl(rq, p, ENQUEUE_REPLENISH); - if (dl_task(rq->curr)) + if (dl_task(rq->donor)) wakeup_preempt_dl(rq, p, 0); else resched_curr(rq); @@ -1736,11 +1736,11 @@ int dl_server_apply_params(struct sched_dl_entity *dl_se, u64 runtime, u64 perio */ static void update_curr_dl(struct rq *rq) { - struct task_struct *curr = rq->curr; - struct sched_dl_entity *dl_se = &curr->dl; + struct task_struct *donor = rq->donor; + struct sched_dl_entity *dl_se = &donor->dl; s64 delta_exec; - if (!dl_task(curr) || !on_dl_rq(dl_se)) + if (!dl_task(donor) || !on_dl_rq(dl_se)) return; /* @@ -2213,7 +2213,7 @@ static int find_later_rq(struct task_struct *task); static int select_task_rq_dl(struct task_struct *p, int cpu, int flags) { - struct task_struct *curr; + struct task_struct *curr, *donor; bool select_rq; struct rq *rq; @@ -2224,6 +2224,7 @@ select_task_rq_dl(struct task_struct *p, int cpu, int flags) rcu_read_lock(); curr = READ_ONCE(rq->curr); /* unlocked access */ + donor = READ_ONCE(rq->donor); /* * If we are dealing with a -deadline task, we must @@ -2234,9 +2235,9 @@ select_task_rq_dl(struct task_struct *p, int cpu, int flags) * other hand, if it has a shorter deadline, we * try to make it stay here, it might be important. */ - select_rq = unlikely(dl_task(curr)) && + select_rq = unlikely(dl_task(donor)) && (curr->nr_cpus_allowed < 2 || - !dl_entity_preempt(&p->dl, &curr->dl)) && + !dl_entity_preempt(&p->dl, &donor->dl)) && p->nr_cpus_allowed > 1; /* @@ -2299,7 +2300,7 @@ static void check_preempt_equal_dl(struct rq *rq, struct task_struct *p) * let's hope p can move out. */ if (rq->curr->nr_cpus_allowed == 1 || - !cpudl_find(&rq->rd->cpudl, rq->curr, NULL)) + !cpudl_find(&rq->rd->cpudl, rq->donor, NULL)) return; /* @@ -2338,7 +2339,7 @@ static int balance_dl(struct rq *rq, struct task_struct *p, struct rq_flags *rf) static void wakeup_preempt_dl(struct rq *rq, struct task_struct *p, int flags) { - if (dl_entity_preempt(&p->dl, &rq->curr->dl)) { + if (dl_entity_preempt(&p->dl, &rq->donor->dl)) { resched_curr(rq); return; } @@ -2348,7 +2349,7 @@ static void wakeup_preempt_dl(struct rq *rq, struct task_struct *p, * In the unlikely case current and p have the same deadline * let us try to decide what's the best thing to do... */ - if ((p->dl.deadline == rq->curr->dl.deadline) && + if ((p->dl.deadline == rq->donor->dl.deadline) && !test_tsk_need_resched(rq->curr)) check_preempt_equal_dl(rq, p); #endif /* CONFIG_SMP */ @@ -2380,7 +2381,7 @@ static void set_next_task_dl(struct rq *rq, struct task_struct *p, bool first) if (!first) return; - if (rq->curr->sched_class != &dl_sched_class) + if (rq->donor->sched_class != &dl_sched_class) update_dl_rq_load_avg(rq_clock_pelt(rq), rq, 0); deadline_queue_push_tasks(rq); @@ -2699,8 +2700,8 @@ retry: * can move away, it makes sense to just reschedule * without going further in pushing next_task. */ - if (dl_task(rq->curr) && - dl_time_before(next_task->dl.deadline, rq->curr->dl.deadline) && + if (dl_task(rq->donor) && + dl_time_before(next_task->dl.deadline, rq->donor->dl.deadline) && rq->curr->nr_cpus_allowed > 1) { resched_curr(rq); return 0; @@ -2823,7 +2824,7 @@ static void pull_dl_task(struct rq *this_rq) * deadline than the current task of its runqueue. */ if (dl_time_before(p->dl.deadline, - src_rq->curr->dl.deadline)) + src_rq->donor->dl.deadline)) goto skip; if (is_migration_disabled(p)) { @@ -2862,9 +2863,9 @@ static void task_woken_dl(struct rq *rq, struct task_struct *p) if (!task_on_cpu(rq, p) && !test_tsk_need_resched(rq->curr) && p->nr_cpus_allowed > 1 && - dl_task(rq->curr) && + dl_task(rq->donor) && (rq->curr->nr_cpus_allowed < 2 || - !dl_entity_preempt(&p->dl, &rq->curr->dl))) { + !dl_entity_preempt(&p->dl, &rq->donor->dl))) { push_dl_tasks(rq); } } @@ -3039,12 +3040,12 @@ static void switched_to_dl(struct rq *rq, struct task_struct *p) return; } - if (rq->curr != p) { + if (rq->donor != p) { #ifdef CONFIG_SMP if (p->nr_cpus_allowed > 1 && rq->dl.overloaded) deadline_queue_push_tasks(rq); #endif - if (dl_task(rq->curr)) + if (dl_task(rq->donor)) wakeup_preempt_dl(rq, p, 0); else resched_curr(rq); @@ -3073,7 +3074,7 @@ static void prio_changed_dl(struct rq *rq, struct task_struct *p, if (!rq->dl.overloaded) deadline_queue_pull_task(rq); - if (task_current(rq, p)) { + if (task_current_donor(rq, p)) { /* * If we now have a earlier deadline task than p, * then reschedule, provided p is still on this diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index d9c33f6c1d54..6512258dc71f 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -1200,12 +1200,12 @@ static inline bool do_preempt_short(struct cfs_rq *cfs_rq, */ s64 update_curr_common(struct rq *rq) { - struct task_struct *curr = rq->curr; + struct task_struct *donor = rq->donor; s64 delta_exec; - delta_exec = update_curr_se(rq, &curr->se); + delta_exec = update_curr_se(rq, &donor->se); if (likely(delta_exec > 0)) - update_curr_task(curr, delta_exec); + update_curr_task(donor, delta_exec); return delta_exec; } @@ -1258,7 +1258,7 @@ static void update_curr(struct cfs_rq *cfs_rq) static void update_curr_fair(struct rq *rq) { - update_curr(cfs_rq_of(&rq->curr->se)); + update_curr(cfs_rq_of(&rq->donor->se)); } static inline void @@ -6815,7 +6815,7 @@ static void hrtick_start_fair(struct rq *rq, struct task_struct *p) s64 delta = slice - ran; if (delta < 0) { - if (task_current(rq, p)) + if (task_current_donor(rq, p)) resched_curr(rq); return; } @@ -6830,12 +6830,12 @@ static void hrtick_start_fair(struct rq *rq, struct task_struct *p) */ static void hrtick_update(struct rq *rq) { - struct task_struct *curr = rq->curr; + struct task_struct *donor = rq->donor; - if (!hrtick_enabled_fair(rq) || curr->sched_class != &fair_sched_class) + if (!hrtick_enabled_fair(rq) || donor->sched_class != &fair_sched_class) return; - hrtick_start_fair(rq, curr); + hrtick_start_fair(rq, donor); } #else /* !CONFIG_SCHED_HRTICK */ static inline void @@ -8750,9 +8750,9 @@ static void set_next_buddy(struct sched_entity *se) */ static void check_preempt_wakeup_fair(struct rq *rq, struct task_struct *p, int wake_flags) { - struct task_struct *curr = rq->curr; - struct sched_entity *se = &curr->se, *pse = &p->se; - struct cfs_rq *cfs_rq = task_cfs_rq(curr); + struct task_struct *donor = rq->donor; + struct sched_entity *se = &donor->se, *pse = &p->se; + struct cfs_rq *cfs_rq = task_cfs_rq(donor); int cse_is_idle, pse_is_idle; if (unlikely(se == pse)) @@ -8781,7 +8781,7 @@ static void check_preempt_wakeup_fair(struct rq *rq, struct task_struct *p, int * prevents us from potentially nominating it as a false LAST_BUDDY * below. */ - if (test_tsk_need_resched(curr)) + if (test_tsk_need_resched(rq->curr)) return; if (!sched_feat(WAKEUP_PREEMPTION)) @@ -13080,7 +13080,7 @@ prio_changed_fair(struct rq *rq, struct task_struct *p, int oldprio) * our priority decreased, or if we are not currently running on * this runqueue and our priority is higher than the current's */ - if (task_current(rq, p)) { + if (task_current_donor(rq, p)) { if (p->prio > oldprio) resched_curr(rq); } else @@ -13187,7 +13187,7 @@ static void switched_to_fair(struct rq *rq, struct task_struct *p) * kick off the schedule if running, otherwise just see * if we can still preempt the current task. */ - if (task_current(rq, p)) + if (task_current_donor(rq, p)) resched_curr(rq); else wakeup_preempt(rq, p, 0); diff --git a/kernel/sched/pelt.c b/kernel/sched/pelt.c index a9c65d97b3ca..fc07382361a8 100644 --- a/kernel/sched/pelt.c +++ b/kernel/sched/pelt.c @@ -476,7 +476,7 @@ int update_irq_load_avg(struct rq *rq, u64 running) bool update_other_load_avgs(struct rq *rq) { u64 now = rq_clock_pelt(rq); - const struct sched_class *curr_class = rq->curr->sched_class; + const struct sched_class *curr_class = rq->donor->sched_class; unsigned long hw_pressure = arch_scale_hw_pressure(cpu_of(rq)); lockdep_assert_rq_held(rq); diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index c5c22fc51824..bd66a46b06ac 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -528,7 +528,7 @@ static void dequeue_rt_entity(struct sched_rt_entity *rt_se, unsigned int flags) static void sched_rt_rq_enqueue(struct rt_rq *rt_rq) { - struct task_struct *curr = rq_of_rt_rq(rt_rq)->curr; + struct task_struct *donor = rq_of_rt_rq(rt_rq)->donor; struct rq *rq = rq_of_rt_rq(rt_rq); struct sched_rt_entity *rt_se; @@ -542,7 +542,7 @@ static void sched_rt_rq_enqueue(struct rt_rq *rt_rq) else if (!on_rt_rq(rt_se)) enqueue_rt_entity(rt_se, 0); - if (rt_rq->highest_prio.curr < curr->prio) + if (rt_rq->highest_prio.curr < donor->prio) resched_curr(rq); } } @@ -988,10 +988,10 @@ static inline int rt_se_prio(struct sched_rt_entity *rt_se) */ static void update_curr_rt(struct rq *rq) { - struct task_struct *curr = rq->curr; + struct task_struct *donor = rq->donor; s64 delta_exec; - if (curr->sched_class != &rt_sched_class) + if (donor->sched_class != &rt_sched_class) return; delta_exec = update_curr_common(rq); @@ -999,7 +999,7 @@ static void update_curr_rt(struct rq *rq) return; #ifdef CONFIG_RT_GROUP_SCHED - struct sched_rt_entity *rt_se = &curr->rt; + struct sched_rt_entity *rt_se = &donor->rt; if (!rt_bandwidth_enabled()) return; @@ -1535,7 +1535,7 @@ static int find_lowest_rq(struct task_struct *task); static int select_task_rq_rt(struct task_struct *p, int cpu, int flags) { - struct task_struct *curr; + struct task_struct *curr, *donor; struct rq *rq; bool test; @@ -1547,6 +1547,7 @@ select_task_rq_rt(struct task_struct *p, int cpu, int flags) rcu_read_lock(); curr = READ_ONCE(rq->curr); /* unlocked access */ + donor = READ_ONCE(rq->donor); /* * If the current task on @p's runqueue is an RT task, then @@ -1575,8 +1576,8 @@ select_task_rq_rt(struct task_struct *p, int cpu, int flags) * systems like big.LITTLE. */ test = curr && - unlikely(rt_task(curr)) && - (curr->nr_cpus_allowed < 2 || curr->prio <= p->prio); + unlikely(rt_task(donor)) && + (curr->nr_cpus_allowed < 2 || donor->prio <= p->prio); if (test || !rt_task_fits_capacity(p, cpu)) { int target = find_lowest_rq(p); @@ -1606,12 +1607,8 @@ out: static void check_preempt_equal_prio(struct rq *rq, struct task_struct *p) { - /* - * Current can't be migrated, useless to reschedule, - * let's hope p can move out. - */ if (rq->curr->nr_cpus_allowed == 1 || - !cpupri_find(&rq->rd->cpupri, rq->curr, NULL)) + !cpupri_find(&rq->rd->cpupri, rq->donor, NULL)) return; /* @@ -1654,7 +1651,9 @@ static int balance_rt(struct rq *rq, struct task_struct *p, struct rq_flags *rf) */ static void wakeup_preempt_rt(struct rq *rq, struct task_struct *p, int flags) { - if (p->prio < rq->curr->prio) { + struct task_struct *donor = rq->donor; + + if (p->prio < donor->prio) { resched_curr(rq); return; } @@ -1672,7 +1671,7 @@ static void wakeup_preempt_rt(struct rq *rq, struct task_struct *p, int flags) * to move current somewhere else, making room for our non-migratable * task. */ - if (p->prio == rq->curr->prio && !test_tsk_need_resched(rq->curr)) + if (p->prio == donor->prio && !test_tsk_need_resched(rq->curr)) check_preempt_equal_prio(rq, p); #endif } @@ -1697,7 +1696,7 @@ static inline void set_next_task_rt(struct rq *rq, struct task_struct *p, bool f * utilization. We only care of the case where we start to schedule a * rt task */ - if (rq->curr->sched_class != &rt_sched_class) + if (rq->donor->sched_class != &rt_sched_class) update_rt_rq_load_avg(rq_clock_pelt(rq), rq, 0); rt_queue_push_tasks(rq); @@ -1959,6 +1958,7 @@ static struct task_struct *pick_next_pushable_task(struct rq *rq) BUG_ON(rq->cpu != task_cpu(p)); BUG_ON(task_current(rq, p)); + BUG_ON(task_current_donor(rq, p)); BUG_ON(p->nr_cpus_allowed <= 1); BUG_ON(!task_on_rq_queued(p)); @@ -1991,7 +1991,7 @@ retry: * higher priority than current. If that's the case * just reschedule current. */ - if (unlikely(next_task->prio < rq->curr->prio)) { + if (unlikely(next_task->prio < rq->donor->prio)) { resched_curr(rq); return 0; } @@ -2012,7 +2012,7 @@ retry: * Note that the stoppers are masqueraded as SCHED_FIFO * (cf. sched_set_stop_task()), so we can't rely on rt_task(). */ - if (rq->curr->sched_class != &rt_sched_class) + if (rq->donor->sched_class != &rt_sched_class) return 0; cpu = find_lowest_rq(rq->curr); @@ -2344,7 +2344,7 @@ static void pull_rt_task(struct rq *this_rq) * p if it is lower in priority than the * current task on the run queue */ - if (p->prio < src_rq->curr->prio) + if (p->prio < src_rq->donor->prio) goto skip; if (is_migration_disabled(p)) { @@ -2386,9 +2386,9 @@ static void task_woken_rt(struct rq *rq, struct task_struct *p) bool need_to_push = !task_on_cpu(rq, p) && !test_tsk_need_resched(rq->curr) && p->nr_cpus_allowed > 1 && - (dl_task(rq->curr) || rt_task(rq->curr)) && + (dl_task(rq->donor) || rt_task(rq->donor)) && (rq->curr->nr_cpus_allowed < 2 || - rq->curr->prio <= p->prio); + rq->donor->prio <= p->prio); if (need_to_push) push_rt_tasks(rq); @@ -2472,7 +2472,7 @@ static void switched_to_rt(struct rq *rq, struct task_struct *p) if (p->nr_cpus_allowed > 1 && rq->rt.overloaded) rt_queue_push_tasks(rq); #endif /* CONFIG_SMP */ - if (p->prio < rq->curr->prio && cpu_online(cpu_of(rq))) + if (p->prio < rq->donor->prio && cpu_online(cpu_of(rq))) resched_curr(rq); } } @@ -2487,7 +2487,7 @@ prio_changed_rt(struct rq *rq, struct task_struct *p, int oldprio) if (!task_on_rq_queued(p)) return; - if (task_current(rq, p)) { + if (task_current_donor(rq, p)) { #ifdef CONFIG_SMP /* * If our priority decreases while running, we @@ -2513,7 +2513,7 @@ prio_changed_rt(struct rq *rq, struct task_struct *p, int oldprio) * greater than the current running task * then reschedule. */ - if (p->prio < rq->curr->prio) + if (p->prio < rq->donor->prio) resched_curr(rq); } } diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 449335299d2f..5f3de82ec9cf 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -1147,7 +1147,10 @@ struct rq { */ unsigned int nr_uninterruptible; - struct task_struct __rcu *curr; + union { + struct task_struct __rcu *donor; /* Scheduler context */ + struct task_struct __rcu *curr; /* Execution context */ + }; struct sched_dl_entity *dl_server; struct task_struct *idle; struct task_struct *stop; @@ -1344,6 +1347,11 @@ DECLARE_PER_CPU_SHARED_ALIGNED(struct rq, runqueues); #define cpu_curr(cpu) (cpu_rq(cpu)->curr) #define raw_rq() raw_cpu_ptr(&runqueues) +static inline void rq_set_donor(struct rq *rq, struct task_struct *t) +{ + /* Do nothing */ +} + #ifdef CONFIG_SCHED_CORE static inline struct cpumask *sched_group_span(struct sched_group *sg); @@ -2260,11 +2268,25 @@ static inline u64 global_rt_runtime(void) return (u64)sysctl_sched_rt_runtime * NSEC_PER_USEC; } +/* + * Is p the current execution context? + */ static inline int task_current(struct rq *rq, struct task_struct *p) { return rq->curr == p; } +/* + * Is p the current scheduling context? + * + * Note that it might be the current execution context at the same time if + * rq->curr == rq->donor == p. + */ +static inline int task_current_donor(struct rq *rq, struct task_struct *p) +{ + return rq->donor == p; +} + static inline int task_on_cpu(struct rq *rq, struct task_struct *p) { #ifdef CONFIG_SMP @@ -2448,7 +2470,7 @@ struct sched_class { static inline void put_prev_task(struct rq *rq, struct task_struct *prev) { - WARN_ON_ONCE(rq->curr != prev); + WARN_ON_ONCE(rq->donor != prev); prev->sched_class->put_prev_task(rq, prev, NULL); } @@ -2612,7 +2634,7 @@ static inline cpumask_t *alloc_user_cpus_ptr(int node) static inline struct task_struct *get_push_task(struct rq *rq) { - struct task_struct *p = rq->curr; + struct task_struct *p = rq->donor; lockdep_assert_rq_held(rq); diff --git a/kernel/sched/syscalls.c b/kernel/sched/syscalls.c index 0470bcc3d204..60e66c31e85b 100644 --- a/kernel/sched/syscalls.c +++ b/kernel/sched/syscalls.c @@ -91,7 +91,7 @@ void set_user_nice(struct task_struct *p, long nice) } queued = task_on_rq_queued(p); - running = task_current(rq, p); + running = task_current_donor(rq, p); if (queued) dequeue_task(rq, p, DEQUEUE_SAVE | DEQUEUE_NOCLOCK); if (running) @@ -713,7 +713,7 @@ change: dequeue_task(rq, p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK); queued = task_on_rq_queued(p); - running = task_current(rq, p); + running = task_current_donor(rq, p); if (queued) dequeue_task(rq, p, queue_flags); if (running) -- 2.51.0