This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[COMMITTED 0/4] alpha updates


There are a number of new failures in the math tests, despite the
ulps update.  I suspect some sort of qemu bug, but I've not yet
fully investigated.

The __pointer_chk_guard change fixes some conficting declaration
errors within the testsuite.  Dunno how long I've not noticed that.

Similarly with the backtrace tests.  Somehow I'd thought that alpha
already had unwind tables enabled by default, but clearly not.  I
really should think about fixing that by default in gcc.

Finally, the pthread_once algorithm is identical to the default.
While open-coding the load-lock/store-conditional makes for a bit
smaller code, I can't imagine that benefit is worth maintaining a
separate version.


r~


Richard Henderson (4):
  alpha: Regenerate sysdeps/alpha/libm-test-ulps
  alpha: Fix __pointer_chk_guard definition for the testsuite
  alpha: Enable unwind tables for backtrace.c
  alpha: Remove alpha-linux pthread_once.c

 ChangeLog                                         |   12 +
 sysdeps/alpha/Makefile                            |    5 +
 sysdeps/alpha/fpu/libm-test-ulps                  | 1048 ++++++++++++++++++++-
 sysdeps/unix/alpha/sysdep.h                       |    2 +-
 sysdeps/unix/sysv/linux/alpha/nptl/pthread_once.c |   95 --
 5 files changed, 1051 insertions(+), 111 deletions(-)
 delete mode 100644 sysdeps/unix/sysv/linux/alpha/nptl/pthread_once.c

-- 
1.9.0


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]