mailing list of musl libc
 help / color / mirror / code / Atom feed
From: Patrick Oppenlander <patrick.oppenlander@gmail.com>
To: musl@lists.openwall.com
Subject: Re: [musl] pthread_cond_wait may has a error
Date: Fri, 11 Aug 2023 07:52:09 +1000	[thread overview]
Message-ID: <CAEg67GnrzEPhR4+t2nuX9_V6Xd9gkSOHSuH3Jqa1HCOUJ5tDtg@mail.gmail.com> (raw)
In-Reply-To: <73135a17-ed7a-4749-aa6e-75faa1c85077.qijinquan@kaihong.com>

[-- Attachment #1: Type: text/plain, Size: 3754 bytes --]

On Thu, 10 Aug 2023, 23:05 祁金全, <qijinquan@kaihong.com> wrote:

>
> Hi,
> In my demo, I create 3 threads, the first and the second thread are wait
> for the global pthread_cond with different pthread_mutex, the last thread
> will call pthread_cond_broadcast.
>
> When I build without musl, the output like this:
> task 3 pthread_cond_broadcast end time 0.000054
> task 2 pthread_cond_wait end time 0.000566
> task 1 pthread_cond_wait end time 0.000612
> task 2 pthread_cond_broadcast end time 1.012233
> task 1 pthread_cond_broadcast end time 1.012247
>
> But when I build with static lib of musl(libc.a) and run , the output like
> this:
> task 3 pthread_cond_broadcast end time 0.000026
> task 1 pthread_cond_wait end time 0.000074
> task 1 pthread_cond_broadcast end time 1.013168
> task 2 pthread_cond_wait end time 1.013245
> task 2 pthread_cond_broadcast end time 2.022992
>
> it seems that the second wakeup thead is waiting until the first wakeup
> thead unlock the mutex(different mute in the two threads).
>
>
>
>
>
> musl version : 1.2.4
>
> pthread_cond_t g_cond = PTHREAD_COND_INITIALIZER;
> pthread_mutex_t g_mutex1 = PTHREAD_MUTEX_INITIALIZER;
> pthread_mutex_t g_mutex2 = PTHREAD_MUTEX_INITIALIZER;
> auto tStart = std::chrono::steady_clock::now();
> void* ThreadTaskOne(void* arg)
> {
>     pthread_mutex_lock(&g_mutex1);
>     pthread_cond_wait(&g_cond, &g_mutex1);
>     auto tEnd = std::chrono::steady_clock::now();
>     auto diff = std::chrono::duration<double>(tEnd - tStart);
>     printf("task 1 pthread_cond_wait end time %f\n", diff.count());
>
>     sleep(1);
>     pthread_mutex_unlock(&g_mutex1);
>     tEnd = std::chrono::steady_clock::now();
>     diff = std::chrono::duration<double>(tEnd - tStart);
>     printf("task 1 pthread_cond_broadcast end time %f\n", diff.count());
>     return nullptr;
> }
>
> void* ThreadTaskTwo(void* arg)
> {
>     pthread_mutex_lock(&g_mutex2);
>     pthread_cond_wait(&g_cond, &g_mutex2);
>     auto tEnd = std::chrono::steady_clock::now();
>     auto diff = std::chrono::duration<double>(tEnd - tStart);
>     printf("task 2 pthread_cond_wait end time %f\n", diff.count());
>
>     sleep(1);
>     pthread_mutex_unlock(&g_mutex2);
>     tEnd = std::chrono::steady_clock::now();
>     diff = std::chrono::duration<double>(tEnd - tStart);
>     printf("task 2 pthread_cond_broadcast end time %f\n", diff.count());
>     return nullptr;
> }
>
> void* BroadcastNotifyMutex(void* arg)
> {
>     tStart = std::chrono::steady_clock::now();
>     pthread_cond_broadcast(&g_cond);
>     auto tEnd = std::chrono::steady_clock::now();
>     auto diff = std::chrono::duration<double>(tEnd - tStart);
>     printf("task 3 pthread_cond_broadcast end time %lf\n", diff.count());
>     return nullptr;
> }
> int main()
> {
>     pthread_t threadOne, threadTwo, threadThree;
>     pthread_create(&threadOne, nullptr, ThreadTaskOne, nullptr);
>     pthread_create(&threadTwo, nullptr, ThreadTaskTwo, nullptr);
>     sleep(3);
>     pthread_create(&threadThree, nullptr, BroadcastNotifyMutex, nullptr);
>     pthread_join(threadOne, nullptr);
>     pthread_join(threadTwo, nullptr);
>     pthread_join(threadThree, nullptr);
>     return 0;
> }
>


I don't think that's a valid test:

https://pubs.opengroup.org/onlinepubs/7908799/xsh/pthread_cond_wait.html

"The effect of using more than one mutex for concurrent
*pthread_cond_wait()* or *pthread_cond_timedwait()* operations on the same
condition variable is undefined; that is, a condition variable becomes
bound to a unique mutex when a thread waits on the condition variable, and
this (dynamic) binding ends when the wait returns."

Patrick

[-- Attachment #2: Type: text/html, Size: 10519 bytes --]

      parent reply	other threads:[~2023-08-10 21:52 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-10  6:22 祁金全
2023-08-10 14:06 ` Alex Xu (Hello71)
2023-08-10 21:52 ` Patrick Oppenlander [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAEg67GnrzEPhR4+t2nuX9_V6Xd9gkSOHSuH3Jqa1HCOUJ5tDtg@mail.gmail.com \
    --to=patrick.oppenlander@gmail.com \
    --cc=musl@lists.openwall.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://git.vuxu.org/mirror/musl/

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).