Subtle bug in LinuxThreads pthread_once().

Kaz Kylheku kaz@ashi.footprints.net
Wed Mar 15 10:20:00 GMT 2000


On 15 Mar 2000, Andreas Jaeger wrote:

> Date: 15 Mar 2000 11:52:47 +0100
> From: Andreas Jaeger <aj@suse.de>
> To: Ulrich Drepper <drepper@cygnus.com>
> Cc: Kaz Kylheku <kaz@ashi.footprints.net>, libc-alpha@sourceware.cygnus.com
> Subject: Re: Subtle bug in LinuxThreads pthread_once().
> 
> >>>>> Ulrich Drepper writes:
> 
> Uli> Thanks, I've applied the patch.
> 
> I'm not convinced that the patch is correct (Please note that I didn't
> review it so far).  I've updated from CVS and run make and make check
> and got a problem.
> 
> The test linuxthreads/ex2 is broken - it's running for 10 minutes now
> and didn't terminate.
> 
> Kaz and  Uli, did you run the testsuite with the patch applied?  Have
> you been successful?

I'm ashamed to say that I did not run the testsuite. Note that the patch I
submitted does not affect ex2.c, because ex2.c doesn't do anything with
pthread_once.  The only example which exercises pthread_once is ex4.c.  So this
breakage may be caused by the other change that Ulrich put in under the same
commit, related to checking the ownership of a mutex in pthread_cond_wait.

I will investigate.

Cheers.



More information about the Libc-alpha mailing list