https://bugs.kde.org/show_bug.cgi?id=445504
--- Comment #7 from Paul Floyd <pjfl...@wanadoo.fr> ---
At lunch time even. Reproduced with Fedora 34 and g++ 11.2.1
If I run the executable through ltrace I see
paulf> /usr/bin/ltrace ./valgrind_test 2>&1 | c++filt | grep pthread
pthread_mutex_lock(0x7ffd797174b0, 0x7ffd797174b0, 0x4022d0, 0x4043a0) = 0
pthread_mutex_unlock(0x7ffd797174b0, 0x7ffd797173f8, 0x4022d0, 5582) = 0
Seems strange to me - I'm expecting 2 locks and 2 unlocks.
and
paulf> nm -C valgrind_test | grep pthread
U pthread_cond_clockwait@GLIBC_2.30
U pthread_create@GLIBC_2.2.5
w __pthread_key_create@GLIBC_2.2.5
w pthread_mutex_lock@GLIBC_2.2.5
w pthread_mutex_unlock@GLIBC_2.2.5
w pthread_once@GLIBC_2.2.5
00000000004029a1 t __gthread_mutex_lock(pthread_mutex_t*)
00000000004029d0 t __gthread_mutex_unlock(pthread_mutex_t*)
I'm a bit suspicious of those weak symbols and _gthread wrappers. Next thing is
to check when these errors are not present.
--
You are receiving this mail because:
You are watching all bug changes.
