On Wed, Jun 11, 2014 at 02:35:38PM +0200, Aurelien Jarno wrote: > On Wed, Jun 11, 2014 at 08:55:42PM +1200, Michael Cree wrote: > > Aurelien: I guess I should file bugs against gcc-4.8 and gcc-4.9 to get > > that fix into Debian's gcc as the commit does not seem to have been > > backported into the upstream 4.8 and 4.9 branches. > > I have just backported the fix into the gcc-4.8 and gcc-4.9 packages. It > will be available with the next upload, but I don't know exactly when it > will happen.
Thanks! > Have you been able to look at the other issue, concerning > __ASSUME_FUTEX_CLOCK_REALTIME? Yes. Just saw this very timely message on the libc-alpha mail list which has given me the clue as to what is wrong: https://sourceware.org/ml/libc-alpha/2014-06/msg00227.html Patch attached fixing lll_futex_timed_wait_bitset macro on alpha. I have confirmed that it fixes the tst-rwlock6, tst-rwlock12, etc., test suite failures on a compile of upstream glibc 2.19 branch. Cheers Michael.
diff --git a/ports/sysdeps/unix/sysv/linux/alpha/nptl/lowlevellock.h b/ports/sysdeps/unix/sysv/linux/alpha/nptl/lowlevellock.h index 361bd342f159..dda2683cb0c6 100644 --- a/ports/sysdeps/unix/sysv/linux/alpha/nptl/lowlevellock.h +++ b/ports/sysdeps/unix/sysv/linux/alpha/nptl/lowlevellock.h @@ -90,7 +90,7 @@ __lll_private_flag (__op, private), \ (val), (timespec), NULL /* Unused. */, \ FUTEX_BITSET_MATCH_ANY); \ - __ret; \ + INTERNAL_SYSCALL_ERROR_P (__ret, __err) ? -__ret : __ret; \ }) #define lll_futex_timed_wait(futexp, val, timespec, private) \
