Messages in this thread Patch in this message | | | Date | Fri, 2 Jun 2023 12:37:31 +0200 | From | Peter Zijlstra <> | Subject | Re: [PATCH] sched: Consider task_struct::saved_state in wait_task_inactive(). |
| |
On Fri, Jun 02, 2023 at 10:25:03AM +0200, Peter Zijlstra wrote: > On Thu, Jun 01, 2023 at 11:12:34AM +0200, Peter Zijlstra wrote: > > On Fri, May 26, 2023 at 05:13:35PM +0200, Sebastian Andrzej Siewior wrote: > > > On 2023-05-26 10:05:43 [+0200], Peter Zijlstra wrote: > > > > New day, new chances... How's this? Code-gen doesn't look totally > > > > insane, but then, making sense of an optimizing compiler's output is > > > > always a wee challenge. > > > > > > Noticed it too late but looks good. Tested, works. > > > > Excellent; full patch below. Will go stick in tip/sched/core soonish. > > Urgh, so robot kicked me for breaking !SMP. And that made me realize > that UP wait_task_inactive() is broken on PREEMPT_RT. > > Let me figure out what best to do about that..
I'll stick this in front -- see what happens ;-)
--- Subject: sched: Unconditionally use full-fat wait_task_inactive() From: Peter Zijlstra <peterz@infradead.org> Date: Fri Jun 2 10:42:53 CEST 2023
While modifying wait_task_inactive() for PREEMPT_RT; the build robot noted that UP got broken. This led to audit and consideration of the UP implementation of wait_task_inactive().
It looks like the UP implementation is also broken for PREEMPT; consider task_current_syscall() getting preempted between the two calls to wait_task_inactive().
Therefore move the wait_task_inactive() implementation out of CONFIG_SMP and unconditionally use it.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> --- include/linux/sched.h | 7 - kernel/sched/core.c | 216 +++++++++++++++++++++++++------------------------- 2 files changed, 110 insertions(+), 113 deletions(-)
--- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -2006,15 +2006,12 @@ static __always_inline void scheduler_ip */ preempt_fold_need_resched(); } -extern unsigned long wait_task_inactive(struct task_struct *, unsigned int match_state); #else static inline void scheduler_ipi(void) { } -static inline unsigned long wait_task_inactive(struct task_struct *p, unsigned int match_state) -{ - return 1; -} #endif +extern unsigned long wait_task_inactive(struct task_struct *, unsigned int match_state); + /* * Set thread flags in other task's structures. * See asm/thread_info.h for TIF_xxxx flags available: --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2213,6 +2213,114 @@ void check_preempt_curr(struct rq *rq, s rq_clock_skip_update(rq); } +/* + * wait_task_inactive - wait for a thread to unschedule. + * + * Wait for the thread to block in any of the states set in @match_state. + * If it changes, i.e. @p might have woken up, then return zero. When we + * succeed in waiting for @p to be off its CPU, we return a positive number + * (its total switch count). If a second call a short while later returns the + * same number, the caller can be sure that @p has remained unscheduled the + * whole time. + * + * The caller must ensure that the task *will* unschedule sometime soon, + * else this function might spin for a *long* time. This function can't + * be called with interrupts off, or it may introduce deadlock with + * smp_call_function() if an IPI is sent by the same process we are + * waiting to become inactive. + */ +unsigned long wait_task_inactive(struct task_struct *p, unsigned int match_state) +{ + int running, queued; + struct rq_flags rf; + unsigned long ncsw; + struct rq *rq; + + for (;;) { + /* + * We do the initial early heuristics without holding + * any task-queue locks at all. We'll only try to get + * the runqueue lock when things look like they will + * work out! + */ + rq = task_rq(p); + + /* + * If the task is actively running on another CPU + * still, just relax and busy-wait without holding + * any locks. + * + * NOTE! Since we don't hold any locks, it's not + * even sure that "rq" stays as the right runqueue! + * But we don't care, since "task_on_cpu()" will + * return false if the runqueue has changed and p + * is actually now running somewhere else! + */ + while (task_on_cpu(rq, p)) { + if (!(READ_ONCE(p->__state) & match_state)) + return 0; + cpu_relax(); + } + + /* + * Ok, time to look more closely! We need the rq + * lock now, to be *sure*. If we're wrong, we'll + * just go back and repeat. + */ + rq = task_rq_lock(p, &rf); + trace_sched_wait_task(p); + running = task_on_cpu(rq, p); + queued = task_on_rq_queued(p); + ncsw = 0; + if (READ_ONCE(p->__state) & match_state) + ncsw = p->nvcsw | LONG_MIN; /* sets MSB */ + task_rq_unlock(rq, p, &rf); + + /* + * If it changed from the expected state, bail out now. + */ + if (unlikely(!ncsw)) + break; + + /* + * Was it really running after all now that we + * checked with the proper locks actually held? + * + * Oops. Go back and try again.. + */ + if (unlikely(running)) { + cpu_relax(); + continue; + } + + /* + * It's not enough that it's not actively running, + * it must be off the runqueue _entirely_, and not + * preempted! + * + * So if it was still runnable (but just not actively + * running right now), it's preempted, and we should + * yield - it could be a while. + */ + if (unlikely(queued)) { + ktime_t to = NSEC_PER_SEC / HZ; + + set_current_state(TASK_UNINTERRUPTIBLE); + schedule_hrtimeout(&to, HRTIMER_MODE_REL_HARD); + continue; + } + + /* + * Ahh, all good. It wasn't running, and it wasn't + * runnable, which means that it will never become + * running in the future either. We're all done! + */ + break; + } + + return ncsw; +} + #ifdef CONFIG_SMP static void @@ -3341,114 +3449,6 @@ int migrate_swap(struct task_struct *cur } #endif /* CONFIG_NUMA_BALANCING */ -/* - * wait_task_inactive - wait for a thread to unschedule. - * - * Wait for the thread to block in any of the states set in @match_state. - * If it changes, i.e. @p might have woken up, then return zero. When we - * succeed in waiting for @p to be off its CPU, we return a positive number - * (its total switch count). If a second call a short while later returns the - * same number, the caller can be sure that @p has remained unscheduled the - * whole time. - * - * The caller must ensure that the task *will* unschedule sometime soon, - * else this function might spin for a *long* time. This function can't - * be called with interrupts off, or it may introduce deadlock with - * smp_call_function() if an IPI is sent by the same process we are - * waiting to become inactive. - */ -unsigned long wait_task_inactive(struct task_struct *p, unsigned int match_state) -{ - int running, queued; - struct rq_flags rf; - unsigned long ncsw; - struct rq *rq; - - for (;;) { - /* - * We do the initial early heuristics without holding - * any task-queue locks at all. We'll only try to get - * the runqueue lock when things look like they will - * work out! - */ - rq = task_rq(p); - - /* - * If the task is actively running on another CPU - * still, just relax and busy-wait without holding - * any locks. - * - * NOTE! Since we don't hold any locks, it's not - * even sure that "rq" stays as the right runqueue! - * But we don't care, since "task_on_cpu()" will - * return false if the runqueue has changed and p - * is actually now running somewhere else! - */ - while (task_on_cpu(rq, p)) { - if (!(READ_ONCE(p->__state) & match_state)) - return 0; - cpu_relax(); - } - - /* - * Ok, time to look more closely! We need the rq - * lock now, to be *sure*. If we're wrong, we'll - * just go back and repeat. - */ - rq = task_rq_lock(p, &rf); - trace_sched_wait_task(p); - running = task_on_cpu(rq, p); - queued = task_on_rq_queued(p); - ncsw = 0; - if (READ_ONCE(p->__state) & match_state) - ncsw = p->nvcsw | LONG_MIN; /* sets MSB */ - task_rq_unlock(rq, p, &rf); - - /* - * If it changed from the expected state, bail out now. - */ - if (unlikely(!ncsw)) - break; - - /* - * Was it really running after all now that we - * checked with the proper locks actually held? - * - * Oops. Go back and try again.. - */ - if (unlikely(running)) { - cpu_relax(); - continue; - } - - /* - * It's not enough that it's not actively running, - * it must be off the runqueue _entirely_, and not - * preempted! - * - * So if it was still runnable (but just not actively - * running right now), it's preempted, and we should - * yield - it could be a while. - */ - if (unlikely(queued)) { - ktime_t to = NSEC_PER_SEC / HZ; - - set_current_state(TASK_UNINTERRUPTIBLE); - schedule_hrtimeout(&to, HRTIMER_MODE_REL_HARD); - continue; - } - - /* - * Ahh, all good. It wasn't running, and it wasn't - * runnable, which means that it will never become - * running in the future either. We're all done! - */ - break; - } - - return ncsw; -} - /*** * kick_process - kick a running thread to enter/exit the kernel * @p: the to-be-kicked thread
| |