Messages in this thread Patch in this message | | | From | Valentin Schneider <> | Subject | [PATCH] sched/core: Print out straggler tasks in sched_cpu_dying() | Date | Wed, 13 Jan 2021 18:31:41 +0000 |
| |
Since commit
1cf12e08bc4d ("sched/hotplug: Consolidate task migration on CPU unplug")
tasks are expected to move themselves out of a out-going CPU. For most tasks this will be done automagically via BALANCE_PUSH, but percpu kthreads will have to cooperate and move themselves away one way or another.
Currently, some percpu kthreads (workqueues being a notable exemple) do not cooperate nicely and can end up on an out-going CPU at the time sched_cpu_dying() is invoked.
Print the dying rq's tasks to shed some light on the stragglers.
Signed-off-by: Valentin Schneider <valentin.schneider@arm.com> --- As Peter pointed out, this should really be caught much earlier than sched_cpu_dying().
If we go down the route of preventing kthreads from being affined to !active CPUs in __set_cpus_allowed_ptr() (genuine percpu kthreads sidestep it via kthread_bind_mask()), then I *think* we could catch this in wakeups, i.e. select_task_rq(). I've been playing around there, but it's not as straightforward as I'd have hoped. --- kernel/sched/core.c | 24 +++++++++++++++++++++++- 1 file changed, 23 insertions(+), 1 deletion(-)
diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 9a08a39d7cdb..d784dd1ae436 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -7564,6 +7564,25 @@ static void calc_load_migrate(struct rq *rq) atomic_long_add(delta, &calc_load_tasks); } +static void dump_rq_tasks(struct rq *rq, const char *loglvl) +{ + struct task_struct *g, *p; + int cpu = cpu_of(rq); + + lockdep_assert_held(&rq->lock); + + printk("%sCPU%d enqueued tasks (%u total):\n", loglvl, cpu, rq->nr_running); + for_each_process_thread(g, p) { + if (task_cpu(p) != cpu) + continue; + + if (!task_on_rq_queued(p)) + continue; + + printk("%s\tpid: %d, name: %s\n", loglvl, p->pid, p->comm); + } +} + int sched_cpu_dying(unsigned int cpu) { struct rq *rq = cpu_rq(cpu); @@ -7573,7 +7592,10 @@ int sched_cpu_dying(unsigned int cpu) sched_tick_stop(cpu); rq_lock_irqsave(rq, &rf); - BUG_ON(rq->nr_running != 1 || rq_has_pinned_tasks(rq)); + if (rq->nr_running != 1 || rq_has_pinned_tasks(rq)) { + WARN(true, "Dying CPU not properly vacated!"); + dump_rq_tasks(rq, KERN_WARNING); + } rq_unlock_irqrestore(rq, &rf); calc_load_migrate(rq); -- 2.27.0
| |