Messages in this thread Patch in this message | | | Date | Mon, 20 Nov 2017 23:57:51 -0500 | From | Steven Rostedt <> | Subject | Re: [PATCH] 4.4.86-rt99: fix sync breakage between nr_cpus_allowed and cpus_allowed |
| |
On Mon, 20 Nov 2017 23:02:07 -0500 Steven Rostedt <rostedt@goodmis.org> wrote:
> Ideally, I would like to stay close to what upstream -rt does. Would > you be able to backport the 4.11-rt patch? > > I'm currently working on releasing 4.9-rt and 4.4-rt with the latest > backports. I could easily add this one too.
Speaking of which. I just backported this patch to 4.4-rt. Is this what you are talking about?
-- Steve
From 1dc89be37874bfc7bb4a0ea7c45492d7db39f62b Mon Sep 17 00:00:00 2001 From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Date: Mon, 19 Jun 2017 09:55:47 +0200 Subject: [PATCH] sched/migrate disable: handle updated task-mask mg-dis section
If task's cpumask changes while in the task is in a migrate_disable() section then we don't react on it after a migrate_enable(). It matters however if current CPU is no longer part of the cpumask. We also miss the ->set_cpus_allowed() callback. This patch fixes it by setting task->migrate_disable_update once we this "delayed" hook. This bug was introduced while fixing unrelated issue in migrate_disable() in v4.4-rt3 (update_migrate_disable() got removed during that).
Cc: stable-rt@vger.kernel.org Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org> --- include/linux/sched.h | 1 kernel/sched/core.c | 59 ++++++++++++++++++++++++++++++++++++++++++++------ 2 files changed, 54 insertions(+), 6 deletions(-)
Index: stable-rt.git/include/linux/sched.h =================================================================== --- stable-rt.git.orig/include/linux/sched.h 2017-11-20 23:43:24.214077537 -0500 +++ stable-rt.git/include/linux/sched.h 2017-11-20 23:43:24.154079278 -0500 @@ -1438,6 +1438,7 @@ struct task_struct { unsigned int policy; #ifdef CONFIG_PREEMPT_RT_FULL int migrate_disable; + int migrate_disable_update; # ifdef CONFIG_SCHED_DEBUG int migrate_disable_atomic; # endif Index: stable-rt.git/kernel/sched/core.c =================================================================== --- stable-rt.git.orig/kernel/sched/core.c 2017-11-20 23:43:24.214077537 -0500 +++ stable-rt.git/kernel/sched/core.c 2017-11-20 23:56:05.071687323 -0500 @@ -1212,18 +1212,14 @@ void set_cpus_allowed_common(struct task p->nr_cpus_allowed = cpumask_weight(new_mask); } -void do_set_cpus_allowed(struct task_struct *p, const struct cpumask *new_mask) +static void __do_set_cpus_allowed_tail(struct task_struct *p, + const struct cpumask *new_mask) { struct rq *rq = task_rq(p); bool queued, running; lockdep_assert_held(&p->pi_lock); - if (__migrate_disabled(p)) { - cpumask_copy(&p->cpus_allowed, new_mask); - return; - } - queued = task_on_rq_queued(p); running = task_current(rq, p); @@ -1246,6 +1242,20 @@ void do_set_cpus_allowed(struct task_str enqueue_task(rq, p, ENQUEUE_RESTORE); } +void do_set_cpus_allowed(struct task_struct *p, const struct cpumask *new_mask) +{ + if (__migrate_disabled(p)) { + lockdep_assert_held(&p->pi_lock); + + cpumask_copy(&p->cpus_allowed, new_mask); +#if defined(CONFIG_PREEMPT_RT_FULL) && defined(CONFIG_SMP) + p->migrate_disable_update = 1; +#endif + return; + } + __do_set_cpus_allowed_tail(p, new_mask); +} + static DEFINE_PER_CPU(struct cpumask, sched_cpumasks); static DEFINE_MUTEX(sched_down_mutex); static cpumask_t sched_down_cpumask; @@ -3231,6 +3241,43 @@ void migrate_enable(void) */ p->migrate_disable = 0; + if (p->migrate_disable_update) { + unsigned long flags; + struct rq *rq; + + rq = task_rq_lock(p, &flags); + update_rq_clock(rq); + + __do_set_cpus_allowed_tail(p, &p->cpus_allowed); + task_rq_unlock(rq, p, &flags); + + p->migrate_disable_update = 0; + + WARN_ON(smp_processor_id() != task_cpu(p)); + if (!cpumask_test_cpu(task_cpu(p), &p->cpus_allowed)) { + const struct cpumask *cpu_valid_mask = cpu_active_mask; + struct migration_arg arg; + unsigned int dest_cpu; + + if (p->flags & PF_KTHREAD) { + /* + * Kernel threads are allowed on online && !active CPUs + */ + cpu_valid_mask = cpu_online_mask; + } + dest_cpu = cpumask_any_and(cpu_valid_mask, &p->cpus_allowed); + arg.task = p; + arg.dest_cpu = dest_cpu; + + unpin_current_cpu(); + preempt_lazy_enable(); + preempt_enable(); + stop_one_cpu(task_cpu(p), migration_cpu_stop, &arg); + tlb_migrate_finish(p->mm); + return; + } + } + unpin_current_cpu(); preempt_enable(); preempt_lazy_enable();
| |