control: forwarded -1 https://sourceware.org/bugzilla/show_bug.cgi?id=27458
Hi, On 2026-10-28 13:35, Tobias Schülein wrote: > Package: libc6 > Version: 2.41-12+deb13u3 > Severity: normal > > A standalone pthread reproducer shows unexpected behavior for a mutex > configured with PTHREAD_MUTEX_RECURSIVE and PTHREAD_MUTEX_ROBUST. > > An owner thread exits while holding the mutex. The next thread receives > EOWNERDEAD and unlocks without calling pthread_mutex_consistent(). > The mutex should then remain permanently unrecoverable. > > Expected: > Both subsequent pthread_mutex_trylock() calls return ENOTRECOVERABLE. > > Observed: > The first returns ENOTRECOVERABLE (131), but the second returns success (0). > > Build: > gcc -std=c11 -D_POSIX_C_SOURCE=200809L -Wall -Wextra -pthread repro.c -o repro Thanks for the bug report and the reproducer. This corresponds to upstream bug BZ#27458, which already got a fixed in master: https://sourceware.org/bugzilla/show_bug.cgi?id=27458 The bug was opened for for more than 5 years until someone reported a similar issue using WSL2. I noticed that you also use WSL2. I am not sure if it is a coincidence or if there is some application on WSL2 that is affected by this bug. Do you mind providing a bit more details on how you found this bug and what are the consequences, it might also help to decide if this bug needs to be fixed in trixie or not, as it is technically not a regression from bookworm or even bullseye. Regards Aurelien -- Aurelien Jarno GPG: 4096R/1DDD8C9B [email protected] http://aurel32.net
