Fix the ia32 atomic add and subtract functions so they

do the right thing.  They now properly return
the value after the update.  This also fixes all warnings
reported by the Sun Studio compiler.  George provided the
new assembly routines.  I added some configure code to make
sure the compilers could handle it.

This fixes trac:2560.

This commit was SVN r23721.

The following Trac tickets were found above:
  Ticket 2560 --> https://svn.open-mpi.org/trac/ompi/ticket/2560

opal/config/opal_config_asm.m4

@@ -10,6 +10,7 @@ dnl                         University of Stuttgart.  All rights reserved.
 dnl Copyright (c) 2004-2005 The Regents of the University of California.
 dnl                         All rights reserved.
 dnl Copyright (c) 2008-2009 Cisco Systems, Inc.  All rights reserved.
+dnl Copyright (c) 2010      Oracle and/or its affiliates.  All rights reserved.
 dnl $COPYRIGHT$
 dnl 
 dnl Additional copyrights may follow
@@ -562,6 +563,13 @@ dnl For PowerPC, this would be:
 dnl
 dnl   "li %0,0" : "=&r"(ret)
 dnl
+dnl For testing ia32 assembly, the assembly instruction xaddl is
+dnl tested.  The xaddl instruction is used by some of the atomic
+dnl implementations so it makes sense to test for it.  In addition,
+dnl some compilers (i.e. earlier versions of Sun Studio 12) do not
+dnl necessarily handle xaddl properly, so that needs to be detected
+dnl during configure time.
+dnl
 dnl DEFINE OMPI_GCC_INLINE_ASSEMBLY to 0 or 1 depending on GCC
 dnl                support
 dnl
@@ -588,6 +596,7 @@ AC_DEFUN([OMPI_CHECK_INLINE_C_GCC],[
                         AC_RUN_IFELSE(AC_LANG_PROGRAM([[
 AC_INCLUDES_DEFAULT]],
 [[int ret = 1;
+int negone = -1;
 __asm__ __volatile__ ($assembly);
 return ret;]]),
                     [asm_result="yes"], [asm_result="no"], 
@@ -603,6 +612,7 @@ return ret;]]),
             AC_LINK_IFELSE(AC_LANG_PROGRAM([[
 AC_INCLUDES_DEFAULT]],
 [[int ret = 1;
+int negone = -1;
 __asm__ __volatile__ ($assembly);
 return ret;]]),
             [asm_result="yes"], [asm_result="no"])
@@ -642,6 +652,7 @@ AC_DEFUN([OMPI_CHECK_INLINE_CXX_GCC],[
             AC_RUN_IFELSE(AC_LANG_PROGRAM([[
 AC_INCLUDES_DEFAULT]],
 [[int ret = 1;
+int negone = -1;
 __asm__ __volatile__ ($assembly);
 return ret;]]),
                     [asm_result="yes"], [asm_result="no"], 
@@ -656,6 +667,7 @@ return ret;]]),
         AC_LINK_IFELSE(AC_LANG_PROGRAM([[
 AC_INCLUDES_DEFAULT]],
 [[int ret = 1;
+int negone = -1;
 __asm__ __volatile__ ($assembly);
 return ret;]]),
             [asm_result="yes"], [asm_result="no"])
@@ -873,7 +885,7 @@ AC_DEFUN([OPAL_CONFIG_ASM],[
                 ompi_cv_asm_arch="AMD64"
             fi
             OPAL_ASM_SUPPORT_64BIT=1
-            OMPI_GCC_INLINE_ASSIGN='"movl [$]0, %0" : "=&r"(ret)'
+            OMPI_GCC_INLINE_ASSIGN='"xaddl %1,%0" : "=m"(ret), "+r"(negone)'
             ;;
 
         ia64-*)

opal/include/opal/sys/ia32/atomic.h

@@ -9,7 +9,7 @@
  *                         University of Stuttgart.  All rights reserved.
  * Copyright (c) 2004-2005 The Regents of the University of California.
  *                         All rights reserved.
- * Copyright (c) 2007      Sun Microsystems, Inc.  All rights reserverd.
+ * Copyright (c) 2007-2010 Oracle and/or its affiliates.  All rights reserved.
  * $COPYRIGHT$
  * 
  * Additional copyrights may follow
@@ -164,11 +164,12 @@ static inline int opal_atomic_cmpset_64(volatile int64_t *addr,
  */
 static inline int32_t opal_atomic_add_32(volatile int32_t* v, int i)
 {
+    int ret = i;
    __asm__ __volatile__(
-                        SMPLOCK "addl %1,%0"
-                        :"=m" (*v)
-                        :"ir" (i), "m" (*v));
-   return (*v);  /* should be an atomic operation */
+                        SMPLOCK "xaddl %1,%0"
+                        :"=m" (*v), "+r" (ret)
+                        );
+   return (ret+i);
 }
 
 
@@ -181,11 +182,12 @@ static inline int32_t opal_atomic_add_32(volatile int32_t* v, int i)
  */
 static inline int32_t opal_atomic_sub_32(volatile int32_t* v, int i)
 {
+    int ret = -i;
    __asm__ __volatile__(
-                        SMPLOCK "subl %1,%0"
-                        :"=m" (*v)
-                        :"ir" (i), "m" (*v));
-   return (*v);  /* should be an atomic operation */
+                        SMPLOCK "xaddl %1,%0"
+                        :"=m" (*v), "+r" (ret)
+                        );
+   return (ret-i);
 }
 
 #endif /* OMPI_GCC_INLINE_ASSEMBLY */