| CC: | submit [...] bugs.debian.org |
| Subject: | test block fails sometime |
| Date: | Sat, 2 Jun 2007 11:41:18 +0200 |
| To: | bug-Thread-Pool-Simple [...] rt.cpan.org |
| From: | Gonéri Le Bouder <goneri [...] rulezlan.org> |
Hello,
This is a forwarded bug that was open on Debian Bug Tracker.
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=427163
Best regards,
On Sat, Jun 02, 2007 at 10:42:24AM +0200, Lucas Nussbaum wrote:
Show quoted text
> Package: libthread-pool-simple-perl
> Version: 0.23-1
> Severity: serious
>
> Hi,
>
> libthread-pool-simple-perl's tests sometimes go into a deadlock when
> running inside fakeroot.
>
> This can be reproduced by running:
> ~/libthread-pool-simple-perl-0.23# while fakeroot perl t/2.t ; do true; done
>
> The place where it blocks differ each time. For example, I just got it
> to block at the following places:
> ok 33
> ok 12
> ok 1 - use Thread::Pool::Simple;
> ok 901
>
> stracing the sleeping process gives the following results:
> * futex(0x525624, FUTEX_WAIT, 1, NULL
> * futex(0x75f3b4, FUTEX_WAIT, 1, NULL
> * futex(0x525624, FUTEX_WAIT, 9, NULL
> I also got it to block in an msgrcv call, but I couldn't reproduce this
> right now.
>
> I could reproduce it:
> * in an up to date sid i386 chroot running with an amd64 kernel, on a
> 2-CPU system, using fakeroot
> * on an amd64 system, 2-CPU system, using fakeroot (without any chroot
> involved)
>
> I could NOT reproduce it:
> * on my laptop (i386, single CPU), even using fakeroot
> * without using fakeroot, on amd64
>
> So, I think that it is related to
> fakeroot AND (smp OR amd64)
>
> Or maybe it's related to a specific kernel version. My laptop runs
> 2.6.21-1-686, while the AMD64 I tested it on are running 2.6.18-3-amd64.
>
> I don't have access to an i386 SMP or an AMD64 UP system to investigate
> further.
Message body not shown because it is not plain text.