Gentoo Archives: gentoo-commits

From: "Mike Frysinger (vapier)" <vapier@g.o>
To: gentoo-commits@l.g.o
Subject: [gentoo-commits] gentoo-x86 commit in dev-libs/libatomic_ops/files: libatomic_ops-7.2_alpha4-x32.patch
Date: Wed, 01 Feb 2012 20:53:58
Message-Id: 20120201205348.BCE612004C@flycatcher.gentoo.org
vapier      12/02/01 20:53:48

  Added:                libatomic_ops-7.2_alpha4-x32.patch
  Log:
  Version bump #353240 by Kevin Bowling.
  
  (Portage version: 2.2.0_alpha84/cvs/Linux x86_64)

Revision  Changes    Path
1.1                  dev-libs/libatomic_ops/files/libatomic_ops-7.2_alpha4-x32.patch

file : http://sources.gentoo.org/viewvc.cgi/gentoo-x86/dev-libs/libatomic_ops/files/libatomic_ops-7.2_alpha4-x32.patch?rev=1.1&view=markup
plain: http://sources.gentoo.org/viewvc.cgi/gentoo-x86/dev-libs/libatomic_ops/files/libatomic_ops-7.2_alpha4-x32.patch?rev=1.1&content-type=text/plain

Index: libatomic_ops-7.2_alpha4-x32.patch
===================================================================
http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=e3a384578f677c05d812d99c2c92aa13670bd06a

Upstream-Status: Pending

Remove the `q' suffix on x86-64 atomic instructions.

We don't need the `q' suffix on x86_64 atomic instructions for AO_t,
which is defined as "unsigned long".  "unsigned long" is 32bit for x32
and 64bit for x86-64. The register operand in x86-64 atomic instructions
is sufficient to properly determine the register size.

Received this patch from H.J. Lu <hjl.tools@×××××.com>
Signed-Off-By: Nitin A Kamble <nitin.a.kamble@×××××.com> 2011/12/02

--- a/src/atomic_ops/sysdeps/gcc/x86_64.h
+++ b/src/atomic_ops/sysdeps/gcc/x86_64.h
@@ -56,7 +56,7 @@ AO_fetch_and_add_full (volatile AO_t *p,
 {
   AO_t result;
 
-  __asm__ __volatile__ ("lock; xaddq %0, %1" :
+  __asm__ __volatile__ ("lock; xadd %0, %1" :
                         "=r" (result), "=m" (*p) : "0" (incr), "m" (*p)
                         : "memory");
   return result;
@@ -106,7 +106,7 @@ AO_int_fetch_and_add_full (volatile unsi
 AO_INLINE void
 AO_or_full (volatile AO_t *p, AO_t incr)
 {
-  __asm__ __volatile__ ("lock; orq %1, %0" :
+  __asm__ __volatile__ ("lock; or %1, %0" :
                         "=m" (*p) : "r" (incr), "m" (*p) : "memory");
 }
 
@@ -131,7 +131,7 @@ AO_compare_and_swap_full(volatile AO_t *
                          AO_t old, AO_t new_val)
 {
   char result;
-  __asm__ __volatile__("lock; cmpxchgq %3, %0; setz %1"
+  __asm__ __volatile__("lock; cmpxchg %3, %0; setz %1"
                        : "=m"(*addr), "=q"(result)
                        : "m"(*addr), "r" (new_val), "a"(old) : "memory");
   return (int) result;