pthread_mutex_destroy returns EBUSY, but the mutex isn't locked

[Carlos O'Donell]

On Wed, Jun 4, 2008 at 10:15 AM, Adrian Ludwin <adrian.ludwin@gmail.com> wrote:
> Thanks, but that seems unlikely for a reason I forgot to mention: all
> threads except the main thread have been successfully joined *before*
> I attempt to destroy any of the mutexes. In addition, I never create
> more than three additional threads for this algorithm (I'm running on
> a quad-core machine and the main thread remains active) and I'm
> occasionally seeing nusers go as high as 10.

Can you reproduce this using glibc cvs head?

It could be a glibc or kernel bug.

Cheers,
Carlos.