On Thu, Mar 10, 2016 at 10:23:17PM +0000, Kári Tristan Helgason wrote: > I think I may have come across a race in the implementation of > `pthread_barrier_wait` in librthread. > > This POC[1] forces the issue with a strategically placed `sleep(3)` call. > This is in essence the same issue as is being reported in glibc here[2]. > > I hope this is the correct place to report things like this. > > Thanks for your time, > Kári > > [1]: https://gist.github.com/kthelgason/687e53091182869b365b > [2]: https://sourceware.org/bugzilla/show_bug.cgi?id=12674 >
Why does your poc contain your own pthread_barrier functions and struct? If you just wanted to add a bunch of sleeps to demonstrate an issue, send a patch against librthread together with a test. Besides this code looks nothing like librthread. I feel like I'm missing something??
