Problem with malloc changes?

[Thorsten Kukuk]

Hi,

Since we have the following patch in CVS:

2005-09-26  Ulrich Drepper  <drepper@redhat.com>

        [BZ #838]
        * malloc/arena.c (ptmalloc_lock_all): If global lock already taken
        by the same thread, just bump the counter.
        (ptmalloc_unlock_all): If counter for recursive locks hasn't reached
        zero, don't do anything else.

glibc does not work anymore for me, I have a lot of problems with
calls to the wait*() functions on i586.

The best example is nptl/tst-cancel4: I get the message, that the
cancellation of the wait test was successfull, but after that the
test case hangs and will be killed by the timeout. Debugging a little
bit shows that the thread with the wait() calls does not seems to be
completly canceld, it is still running. Reverting exact this patch
solves the problem for me.
Does somebody else is seeing this?

  Thorsten

-- 
Thorsten Kukuk         http://www.suse.de/~kukuk/      kukuk@suse.de
SUSE LINUX Products GmbH       Maxfeldstr. 5       D-90409 Nuernberg
--------------------------------------------------------------------    
Key fingerprint = A368 676B 5E1B 3E46 CFCE  2D97 F8FD 4E23 56C6 FB4B