linux-kernel - Re: [PATCH 1/1] sched/core: Fix stuck on completion for affine_move

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <4b0acbb118018bacdcd88979e4083b118c164594.camel@mediatek.com>
Date:   Sun, 1 Oct 2023 15:15:09 +0000
From:   Kuyo Chang (張建文) 
        <Kuyo.Chang@...iatek.com>
To:     "peterz@...radead.org" <peterz@...radead.org>
CC:     "dietmar.eggemann@....com" <dietmar.eggemann@....com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-mediatek@...ts.infradead.org" 
        <linux-mediatek@...ts.infradead.org>,
        "rostedt@...dmis.org" <rostedt@...dmis.org>,
        wsd_upstream <wsd_upstream@...iatek.com>,
        "vschneid@...hat.com" <vschneid@...hat.com>,
        "bristot@...hat.com" <bristot@...hat.com>,
        "juri.lelli@...hat.com" <juri.lelli@...hat.com>,
        "mingo@...hat.com" <mingo@...hat.com>,
        "linux-arm-kernel@...ts.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>,
        "bsegall@...gle.com" <bsegall@...gle.com>,
        "mgorman@...e.de" <mgorman@...e.de>,
        "matthias.bgg@...il.com" <matthias.bgg@...il.com>,
        "vincent.guittot@...aro.org" <vincent.guittot@...aro.org>,
        "angelogioacchino.delregno@...labora.com" 
        <angelogioacchino.delregno@...labora.com>
Subject: Re: [PATCH 1/1] sched/core: Fix stuck on completion for
 affine_move_task() when stopper disable

On Fri, 2023-09-29 at 12:21 +0200, Peter Zijlstra wrote:
>  	 
> External email : Please do not click links or open attachments until
> you have verified the sender or the content.
>  On Wed, Sep 27, 2023 at 03:57:35PM +0000, Kuyo Chang (張建文) wrote:
> 
> > This issue occurs at CPU hotplug/set_affinity stress test.
> > The reproduce ratio is very low(about once a week).
> 
> I'm assuming you're running an arm64 kernel with preempt_full=y (the
> default for arm64).

Yes, the test platform is arm64 with kernel config as below

CONFIG_PREEMPT_BUILD=y
CONFIG_PREEMPT=y
CONFIG_PREEMPT_COUNT=y
CONFIG_PREEMPTION=y
CONFIG_PREEMPT_RCU=y
CONFIG_HAVE_PREEMPT_DYNAMIC=y
CONFIG_HAVE_PREEMPT_DYNAMIC_KEY=y
CONFIG_PREEMPT_NOTIFIERS=y

> Could you please test the below?

Ok, let me run it and report.

> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index d8fd29d66b24..079a63b8a954 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -2645,9 +2645,11 @@ static int migration_cpu_stop(void *data)
>   * it.
>   */
>  WARN_ON_ONCE(!pending->stop_pending);
> +preempt_disable();
>  task_rq_unlock(rq, p, &rf);
>  stop_one_cpu_nowait(task_cpu(p), migration_cpu_stop,
>      &pending->arg, &pending->stop_work);
> +preempt_enable();
>  return 0;
>  }
>  out:
> @@ -2967,12 +2969,13 @@ static int affine_move_task(struct rq *rq,
> struct task_struct *p, struct rq_flag
>  complete = true;
>  }
>  
> +preempt_disable();
>  task_rq_unlock(rq, p, rf);
> -
>  if (push_task) {
>  stop_one_cpu_nowait(rq->cpu, push_cpu_stop,
>      p, &rq->push_work);
>  }
> +preempt_enable();
>  
>  if (complete)
>  complete_all(&pending->done);
> @@ -3038,12 +3041,13 @@ static int affine_move_task(struct rq *rq,
> struct task_struct *p, struct rq_flag
>  if (flags & SCA_MIGRATE_ENABLE)
>  p->migration_flags &= ~MDF_PUSH;
>  
> +preempt_disable();
>  task_rq_unlock(rq, p, rf);
> -
>  if (!stop_pending) {
>  stop_one_cpu_nowait(cpu_of(rq), migration_cpu_stop,
>      &pending->arg, &pending->stop_work);
>  }
> +preempt_enable();
>  
>  if (flags & SCA_MIGRATE_ENABLE)
>  return 0;
> @@ -9459,6 +9461,7 @@ static void balance_push(struct rq *rq)
>   * Temporarily drop rq->lock such that we can wake-up the stop task.
>   * Both preemption and IRQs are still disabled.
>   */
> +preempt_disable();
>  raw_spin_rq_unlock(rq);
>  stop_one_cpu_nowait(rq->cpu, __balance_push_cpu_stop, push_task,
>      this_cpu_ptr(&push_work));
> @@ -9468,6 +9471,7 @@ static void balance_push(struct rq *rq)
>   * which kthread_is_per_cpu() and will push this task away.
>   */
>  raw_spin_rq_lock(rq);
> +preempt_enable();
>  }
>  
>  static void balance_push_set(int cpu, bool on)