https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114953
Bug ID: 114953
Summary: libstdc++'s 30_threads/semaphore/try_acquire_posix.cc
can sometimes fail (while running under qemu)
Product: gcc
Version: 15.0
Status: UNCONFIRMED
Keywords: testsuite-fail
Severity: normal
Priority: P3
Component: testsuite
Assignee: unassigned at gcc dot gnu.org
Reporter: pinskia at gcc dot gnu.org
Target Milestone: ---
The testcase has:
```
auto const dur = 250ms;
...
auto const at = std::chrono::system_clock::now() + dur;
auto const t0 = std::chrono::steady_clock::now();
VERIFY( !s._M_try_acquire_until(at) );
auto const diff = std::chrono::steady_clock::now() - t0;
VERIFY( diff >= dur );
```
So the difference here might be less than 250ms because it might be that t0
might have been a few (one or two?) microseconds after at was set.
This testcase failed for me with -std=c++26 but passed for -std=c++20 while
running in the testsuite but on a semi-loaded system and running under qemu. So
it just happened to hit the failure in that case. I suspect it could fail while
not running under qemu but with a heavy load on the system and it just happens
to be time sliced out between the 2 calls to now().
Maybe there should only 1 call to now before until so you don't run into
spurious failure like this.
