From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <pthreads-win32-return-934-listarch-pthreads-win32=sources.redhat.com@sources.redhat.com>
Received: (qmail 25411 invoked by alias); 3 Mar 2005 09:32:52 -0000
Mailing-List: contact pthreads-win32-help@sources.redhat.com; run by ezmlm
Precedence: bulk
List-Subscribe: <mailto:pthreads-win32-subscribe@sources.redhat.com>
List-Archive: <http://sources.redhat.com/ml/pthreads-win32/>
List-Post: <mailto:pthreads-win32@sources.redhat.com>
List-Help: <mailto:pthreads-win32-help@sources.redhat.com>, <http://sources.redhat.com/ml/#faqs>
Sender: pthreads-win32-owner@sources.redhat.com
Received: (qmail 25351 invoked from network); 3 Mar 2005 09:32:43 -0000
Received: from unknown (HELO wproxy.gmail.com) (64.233.184.207)
  by sourceware.org with SMTP; 3 Mar 2005 09:32:43 -0000
Received: by wproxy.gmail.com with SMTP id 71so393171wri
        for <pthreads-win32@sources.redhat.com>; Thu, 03 Mar 2005 01:32:43 -0800 (PST)
Received: by 10.54.73.11 with SMTP id v11mr74694wra;
        Thu, 03 Mar 2005 01:32:42 -0800 (PST)
Received: by 10.54.7.8 with HTTP; Thu, 3 Mar 2005 01:32:42 -0800 (PST)
Message-ID: <97ffb3105030301323c1bae1@mail.gmail.com>
Date: Thu, 03 Mar 2005 09:32:00 -0000
From: Gottlob Frege <gottlobfrege@gmail.com>
Reply-To: Gottlob Frege <gottlobfrege@gmail.com>
To: pthreads-win32@sources.redhat.com
Subject: starvation in pthread_once?
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
X-SW-Source: 2005/txt/msg00016.txt.bz2

I'm concerned about the Sleep(0) in pthread_once:

  if (!once_control->done)
    {
      if (InterlockedIncrement (&(once_control->started)) == 0)
        {
          /*
           * First thread to increment the started variable
           */
          (*init_routine) ();
          once_control->done = PTW32_TRUE;

        }
      else
        {
          /*
           * Block until other thread finishes executing the onceRoutine
           */
          while (!(once_control->done))
            {
              /*
               * The following gives up CPU cycles without pausing
               * unnecessarily
               */
              Sleep (0);
            }
        }
    }

IIRC, Sleep(0) does not relinquish time slices to lower priority
threads.  (Sleep(n) for n != 0 does, but 0 does not).  So, if a lower
priority thread is first in, followed closely by a higher priority
one, the higher priority thread will spin on Sleep(0) *forever*
because the lower, first thread will never get a chance to set done.

So even Sleep(10) should be good enough.  In theory, there could be
enough higher priority threads in the system that the first thread
still doesn't get in (ever?!), but unlikely.  And that would probably
mean a general design flaw of the calling code, not pthread_once.

?