LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* Re: Possible deadlock errors in tools/perf/builtin-sched.c
       [not found] <CAAo+4rXNLdgvAiT2-B8cWtLNPnWoGo9RWMW=8SPchzRgxJ4BhA@mail.gmail.com>
@ 2021-08-28  8:06 ` 叶澄锋
  2021-08-31 18:46 ` Jiri Olsa
  1 sibling, 0 replies; 3+ messages in thread
From: 叶澄锋 @ 2021-08-28  8:06 UTC (permalink / raw)
  To: peterz, mingo, acme
  Cc: mark.rutland, alexander.shishkin, jolsa, namhyung,
	linux-perf-users, linux-kernel, CAI Yuandao

Dear developers:

Thank you for your checking.

It seems there are two deadlock errors on the
locksched->work_done_wait_mutexandsched->start_work_mutex.

They are triggered due to one thread(A) runs function run_one_test
locating in a loop and unreleasing the two locks in
thewait_for_tasksfunction, and another thread(B) runs function
thread_func acquiring the two locks.

Because the two locks are not properly released in thread A, there
will be a  deadlock problem if thread B acquires the two locks.

The related codes are below:

Thread A:

static void create_tasks(struct perf_sched *sched)
{
     ...;
    err = pthread_mutex_lock(&sched->start_work_mutex);
    ...;
    err = pthread_mutex_lock(&sched->work_done_wait_mutex);
    ...;
}

static int perf_sched__replay(struct perf_sched *sched)
{
   ...;
     create_tasks(sched);
     printf("------------------------------------------------------------\n");
     for (i = 0; i < sched->replay_repeat; i++)
          run_one_test(sched);   // multiple reacquisition on the lock
sched->work_done_wait_mutex and sched->start_work_mutex

   return 0;
}

static void run_one_test(struct perf_sched *sched)
{
        ...;
      wait_for_tasks(sched);
        ...;
}

static void wait_for_tasks(struct perf_sched *sched)
{
      ...;
     pthread_mutex_unlock(&sched->work_done_wait_mutex);
      ...;
     ret = pthread_mutex_lock(&sched->work_done_wait_mutex);
     ...;
     pthread_mutex_unlock(&sched->start_work_mutex);
     ...;

    ret = pthread_mutex_lock(&sched->start_work_mutex);
    ....;
}

Thread B:
static void *thread_func(void *ctx)
{
...;
ret = pthread_mutex_lock(&sched->start_work_mutex);
...;
ret = pthread_mutex_unlock(&sched->start_work_mutex);

...;

ret = pthread_mutex_lock(&sched->work_done_wait_mutex);
...;
ret = pthread_mutex_unlock(&sched->work_done_wait_mutex);
..;

}

PS: The previous email fails to reach the maillist, so I send it again.

Thanks,

叶澄锋 <dg573847474@gmail.com> 于2021年8月28日周六 下午3:57写道:
>
> Dear developers:
>
> Thank you for your checking.
>
> It seems there are two deadlock errors on the locksched->work_done_wait_mutexandsched->start_work_mutex.
>
> They are triggered due to one thread(A) runs function run_one_test locating in a loop and unreleasing the two locks in thewait_for_tasksfunction, and another thread(B) runs function thread_func acquiring the two locks.
>
> Because the two locks are not properly released in thread A, there will be a  deadlock problem if thread B acquires the two locks.
>
> The related codes are below:
>
> Thread A:
>
> static void create_tasks(struct perf_sched *sched)
> {
>      ...;
>   err = pthread_mutex_lock(&sched->start_work_mutex);
>     ...;
>  err = pthread_mutex_lock(&sched->work_done_wait_mutex);
>        ...;
> }
> static int perf_sched__replay(struct perf_sched *sched)
> {
>    ...;
>
>         create_tasks(sched);
>      printf("------------------------------------------------------------\n");
>      for (i = 0; i < sched->replay_repeat; i++)
>           run_one_test(sched);   // multiple reacquisition on the lock sched->work_done_wait_mutex and sched->start_work_mutex
>
>    return 0;
> }
>
> static void run_one_test(struct perf_sched *sched)
> {
>  ...;
>       wait_for_tasks(sched);
>         ...;
> }
> static void wait_for_tasks(struct perf_sched *sched)
> {
>        ...;
>     pthread_mutex_unlock(&sched->work_done_wait_mutex);
>
>   ...;
>        ret = pthread_mutex_lock(&sched->work_done_wait_mutex);
>      ...;
>   pthread_mutex_unlock(&sched->start_work_mutex);
>
>     ...;
>
>  ret = pthread_mutex_lock(&sched->start_work_mutex);
>    ....;
> }
>
> Thread B:
>
> static void *thread_func(void *ctx)
> {
>
> ...;
> ret = pthread_mutex_lock(&sched->start_work_mutex);
> ...;
> ret = pthread_mutex_unlock(&sched->start_work_mutex);
>
> ...;
>
> ret = pthread_mutex_lock(&sched->work_done_wait_mutex);
> ...;
> ret = pthread_mutex_unlock(&sched->work_done_wait_mutex);
> ..;
>
> }
>
>
> Thanks,

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Possible deadlock errors in tools/perf/builtin-sched.c
       [not found] <CAAo+4rXNLdgvAiT2-B8cWtLNPnWoGo9RWMW=8SPchzRgxJ4BhA@mail.gmail.com>
  2021-08-28  8:06 ` Possible deadlock errors in tools/perf/builtin-sched.c 叶澄锋
@ 2021-08-31 18:46 ` Jiri Olsa
  2021-09-01  5:19   ` CAI Yuandao
  1 sibling, 1 reply; 3+ messages in thread
From: Jiri Olsa @ 2021-08-31 18:46 UTC (permalink / raw)
  To: 叶澄锋
  Cc: peterz, mingo, acme, mark.rutland, alexander.shishkin, namhyung,
	linux-perf-users, linux-kernel, CAI Yuandao

On Sat, Aug 28, 2021 at 03:57:17PM +0800, 叶澄锋 wrote:
> Dear developers:
> 
> Thank you for your checking.
> 
> It seems there are two deadlock errors on the lock
> *sched->work_done_wait_mutex*and*sched->start_work_mutex.*
> 
> They are triggered due to one thread(A) runs function *run_one_test* locating
> in a loop and unreleasing the two locks in the*wait_for_tasks*function, and
> another thread(B) runs function *thread_func *acquiring the two locks.
> 
> Because the two locks are not properly released in thread A, there will be
> a  deadlock problem if thread B acquires the two locks.

hi,
do you have a way to reproduce this?

thanks,
jirka

> 
> The related codes are below:
> 
> Thread A:
> 
> static void create_tasks(struct perf_sched *sched)
> {
>      ...;
>   err = pthread_mutex_lock(&sched->start_work_mutex);
>     ...;
>  err = pthread_mutex_lock(&sched->work_done_wait_mutex);
>        ...;
> }
> static int perf_sched__replay(struct perf_sched *sched)
> {
>    ...;
> 
>         create_tasks(sched);
>      printf("------------------------------------------------------------\n");
>      for (i = 0; i < sched->replay_repeat; i++)
>           run_one_test(sched);   // multiple reacquisition on the lock
> sched->work_done_wait_mutex and sched->start_work_mutex
> 
>    return 0;
> }
> 
> static void run_one_test(struct perf_sched *sched)
> {
>  ...;
>       wait_for_tasks(sched);
>         ...;
> }
> static void wait_for_tasks(struct perf_sched *sched)
> {
>        ...;
>     pthread_mutex_unlock(&sched->work_done_wait_mutex);
> 
>   ...;
>        ret = pthread_mutex_lock(&sched->work_done_wait_mutex);
>      ...;
>   pthread_mutex_unlock(&sched->start_work_mutex);
> 
>     ...;
> 
>  ret = pthread_mutex_lock(&sched->start_work_mutex);
>    ....;
> }
> 
> Thread B:
> 
> static void *thread_func(void *ctx)
> {
> 
> ...;
> ret = pthread_mutex_lock(&sched->start_work_mutex);
> ...;
> ret = pthread_mutex_unlock(&sched->start_work_mutex);
> 
> ...;
> 
> ret = pthread_mutex_lock(&sched->work_done_wait_mutex);
> ...;
> ret = pthread_mutex_unlock(&sched->work_done_wait_mutex);
> ..;
> 
> }
> 
> 
> Thanks,


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Possible deadlock errors in tools/perf/builtin-sched.c
  2021-08-31 18:46 ` Jiri Olsa
@ 2021-09-01  5:19   ` CAI Yuandao
  0 siblings, 0 replies; 3+ messages in thread
From: CAI Yuandao @ 2021-09-01  5:19 UTC (permalink / raw)
  To: Jiri Olsa
  Cc: peterz, mingo, acme, mark.rutland, alexander.shishkin, namhyung,
	linux-perf-users, linux-kernel, 叶澄锋

Hi Jirka,

        They were found by our static code analyzer and manually checked as true positives by us. Are they real bugs?

Best,
Ryan

On 1/9/2021, 2:46 AM, "Jiri Olsa" <jolsa@redhat.com> wrote:

    On Sat, Aug 28, 2021 at 03:57:17PM +0800, 叶澄锋 wrote:
    > Dear developers:
    > 
    > Thank you for your checking.
    > 
    > It seems there are two deadlock errors on the lock
    > *sched->work_done_wait_mutex*and*sched->start_work_mutex.*
    > 
    > They are triggered due to one thread(A) runs function *run_one_test* locating
    > in a loop and unreleasing the two locks in the*wait_for_tasks*function, and
    > another thread(B) runs function *thread_func *acquiring the two locks.
    > 
    > Because the two locks are not properly released in thread A, there will be
    > a  deadlock problem if thread B acquires the two locks.

    hi,
    do you have a way to reproduce this?

    thanks,
    jirka

    > 
    > The related codes are below:
    > 
    > Thread A:
    > 
    > static void create_tasks(struct perf_sched *sched)
    > {
    >      ...;
    >   err = pthread_mutex_lock(&sched->start_work_mutex);
    >     ...;
    >  err = pthread_mutex_lock(&sched->work_done_wait_mutex);
    >        ...;
    > }
    > static int perf_sched__replay(struct perf_sched *sched)
    > {
    >    ...;
    > 
    >         create_tasks(sched);
    >      printf("------------------------------------------------------------\n");
    >      for (i = 0; i < sched->replay_repeat; i++)
    >           run_one_test(sched);   // multiple reacquisition on the lock
    > sched->work_done_wait_mutex and sched->start_work_mutex
    > 
    >    return 0;
    > }
    > 
    > static void run_one_test(struct perf_sched *sched)
    > {
    >  ...;
    >       wait_for_tasks(sched);
    >         ...;
    > }
    > static void wait_for_tasks(struct perf_sched *sched)
    > {
    >        ...;
    >     pthread_mutex_unlock(&sched->work_done_wait_mutex);
    > 
    >   ...;
    >        ret = pthread_mutex_lock(&sched->work_done_wait_mutex);
    >      ...;
    >   pthread_mutex_unlock(&sched->start_work_mutex);
    > 
    >     ...;
    > 
    >  ret = pthread_mutex_lock(&sched->start_work_mutex);
    >    ....;
    > }
    > 
    > Thread B:
    > 
    > static void *thread_func(void *ctx)
    > {
    > 
    > ...;
    > ret = pthread_mutex_lock(&sched->start_work_mutex);
    > ...;
    > ret = pthread_mutex_unlock(&sched->start_work_mutex);
    > 
    > ...;
    > 
    > ret = pthread_mutex_lock(&sched->work_done_wait_mutex);
    > ...;
    > ret = pthread_mutex_unlock(&sched->work_done_wait_mutex);
    > ..;
    > 
    > }
    > 
    > 
    > Thanks,



^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2021-09-01  5:28 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <CAAo+4rXNLdgvAiT2-B8cWtLNPnWoGo9RWMW=8SPchzRgxJ4BhA@mail.gmail.com>
2021-08-28  8:06 ` Possible deadlock errors in tools/perf/builtin-sched.c 叶澄锋
2021-08-31 18:46 ` Jiri Olsa
2021-09-01  5:19   ` CAI Yuandao

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).