On Fri, Jun 21, 2024 at 08:06:05PM +0200, Mickaël Salaün wrote: > We cannot use CLONE_VFORK because we also need to wait for the timeout > signal. > > Restore tests timeout by using the original fork() call in __run_test() > but also in __TEST_F_IMPL(). Also fix a race condition when waiting for > the test child process. I've given this a quick test both by trying to apply it directly and through yesterday's -next and there's *something* funky going on, the epoll suite which was one of those hanging is somehow not getting run at all although the binaries do appear to be getting built and end up in my kselftest tarball that I'm deploying. However the ptrace test which was also hanging now manages to trigger it's timeout successfully when it that happens so I think whatever is going on with epoll probably an unrelated issue, I didn't get a chance to look at it properly. Tested-by: Mark Brown <broonie@xxxxxxxxxx>
Attachment:
signature.asc
Description: PGP signature
