Bugzilla bug 195219: put the stops (double semicolons) that delimit

instruction groups on the same lines as the last instructions in the
groups.  This is the style used in Intel documentation.
Modified files: os_HPUX_ia64.s os_Linux_ia64.s

nsprpub/pr/src/md/unix/os_HPUX_ia64.s

@@ -46,8 +46,7 @@ _PR_ia64_AtomicIncrement:
 #ifndef _LP64
         addp4 r32 = 0, r32  ;;
 #endif
-        fetchadd4.acq r8 = [r32], 1
-        ;;
+        fetchadd4.acq r8 = [r32], 1  ;;
         adds r8 = 1, r8
         br.ret.sptk.many b0
         .endp _PR_ia64_AtomicIncrement#
@@ -64,8 +63,7 @@ _PR_ia64_AtomicDecrement:
 #ifndef _LP64
         addp4 r32 = 0, r32  ;;
 #endif
-        fetchadd4.rel r8 = [r32], -1
-        ;;
+        fetchadd4.rel r8 = [r32], -1  ;;
         adds r8 = -1, r8
         br.ret.sptk.many b0
         .endp _PR_ia64_AtomicDecrement#
@@ -82,15 +80,12 @@ _PR_ia64_AtomicAdd:
 #ifndef _LP64
         addp4 r32 = 0, r32  ;;
 #endif
-        ld4 r15 = [r32]
-        ;;
+        ld4 r15 = [r32]  ;;
 .L3:
         mov r14 = r15
         mov ar.ccv = r15
-        add r8 = r15, r33
-        ;;
-        cmpxchg4.acq r15 = [r32], r8, ar.ccv
-        ;;
+        add r8 = r15, r33  ;;
+        cmpxchg4.acq r15 = [r32], r8, ar.ccv  ;;
         cmp4.ne p6, p7 =  r15, r14
         (p6) br.cond.dptk .L3
         br.ret.sptk.many b0

nsprpub/pr/src/md/unix/os_Linux_ia64.s

@@ -43,8 +43,7 @@
         .global _PR_ia64_AtomicIncrement#
         .proc _PR_ia64_AtomicIncrement#
 _PR_ia64_AtomicIncrement:
-        fetchadd4.acq r8 = [r32], 1
-        ;;
+        fetchadd4.acq r8 = [r32], 1  ;;
         adds r8 = 1, r8
         br.ret.sptk.many b0
         .endp _PR_ia64_AtomicIncrement#
@@ -58,8 +57,7 @@ _PR_ia64_AtomicIncrement:
         .global _PR_ia64_AtomicDecrement#
         .proc _PR_ia64_AtomicDecrement#
 _PR_ia64_AtomicDecrement:
-        fetchadd4.rel r8 = [r32], -1
-        ;;
+        fetchadd4.rel r8 = [r32], -1  ;;
         adds r8 = -1, r8
         br.ret.sptk.many b0
         .endp _PR_ia64_AtomicDecrement#
@@ -73,15 +71,12 @@ _PR_ia64_AtomicDecrement:
         .global _PR_ia64_AtomicAdd#
         .proc _PR_ia64_AtomicAdd#
 _PR_ia64_AtomicAdd:
-        ld4 r15 = [r32]
-        ;;
+        ld4 r15 = [r32]  ;;
 .L3:
         mov r14 = r15
         mov ar.ccv = r15
-        add r8 = r15, r33
-        ;;
-        cmpxchg4.acq r15 = [r32], r8, ar.ccv
-        ;;
+        add r8 = r15, r33  ;;
+        cmpxchg4.acq r15 = [r32], r8, ar.ccv  ;;
         cmp4.ne p6, p7 =  r15, r14
         (p6) br.cond.dptk .L3
         br.ret.sptk.many b0