Re: [PATCH glibc-2.44] Fix __pthread_timedblock for hurd
Samuel Thibault <[email protected]>
| Newsgroups | gmane.os.hurd.bugs |
|---|---|
| Organization | I am not organized |
| Message-ID | <aodxjxt-r9VvCs7p@end> |
Hello, Damien Zammit, le mer. 19 août 2026 00:14:46 +0000, a ecrit: > #3 0x0000000001145061 in __pthread_cond_timedwait_internal (cond=0x40b8348, mutex=<optimized out>, clockid=<optimized out>, abstime=<optimized out>) at ../sysdeps/htl/pt-cond-timedwait.c:164 > #4 0x00000000023b3828 in os::Linux::safe_cond_timedwait (_cond=0x40b8348, _mutex=<optimized out>, _abstime=0x3fc2db0) at /part4/git/icedtea/openjdk-boot/hotspot/src/os/linux/vm/os_linux.cpp:5321 > #5 os::PlatformEvent::park (this=0x40b8300, millis=<optimized out>) at /part4/git/icedtea/openjdk-boot/hotspot/src/os/linux/vm/os_linux.cpp:6040 This uses compute_abstime to produce abst. compute_abstime does really compute an absolute time, not a relative time, by using either clock_gettime(CLOCK_MONOTONIC) (if supports_monotonic_clock() returns true), or gettimeofday() (i.e. CLOCK_REALTIME) otherwise. So this looks all coherent in using an absolute time. If you have a hang, you should check that the various values are coherently computed. Samuel