[SPARC64]: Use saner local label names in Ultra3 copies.

This makes the kernel profiles look much more
meaningful.
Signed-off-by: David S. Miller <davem@redhat.com>

arch/sparc64/lib/U3copy_from_user.S

@@ -140,24 +140,24 @@
 	.globl	U3copy_from_user
 U3copy_from_user:	/* %o0=dst, %o1=src, %o2=len */
 	cmp		%o2, 0
-	be,pn		%XCC, out
+	be,pn		%XCC, 85f
 	 or		%o0, %o1, %o3
 	cmp		%o2, 16
-	bleu,a,pn	%XCC, small_copy
+	bleu,a,pn	%XCC, 80f
 	 or		%o3, %o2, %o3
 
 	cmp		%o2, 256
-	blu,pt		%XCC, medium_copy
+	blu,pt		%XCC, 70f
 	 andcc		%o3, 0x7, %g0
 
-	ba,pt		%xcc, enter
+	ba,pt		%xcc, 1f
 	 andcc		%o0, 0x3f, %g2
 
 	/* Here len >= 256 and condition codes reflect execution
 	 * of "andcc %o0, 0x7, %g2", done by caller.
 	 */
 	.align		64
-enter:
+1:
 	/* Is 'dst' already aligned on an 64-byte boundary? */
 	be,pt		%XCC, 2f
 
@@ -180,11 +180,11 @@ enter:
 
 2:	VISEntryHalf
 	and		%o1, 0x7, %g1
-	ba,pt		%xcc, begin
+	ba,pt		%xcc, 1f
 	 alignaddr	%o1, %g0, %o1
 
 	.align		64
-begin:
+1:
 	membar		#StoreLoad | #StoreStore | #LoadStore
 	prefetcha	[%o1 + 0x000] %asi, #one_read
 	prefetcha	[%o1 + 0x040] %asi, #one_read
@@ -213,11 +213,11 @@ begin:
 
 	sub		%o4, 0x80, %o4
 	add		%o1, 0x40, %o1
-	ba,pt		%xcc, loop
+	ba,pt		%xcc, 1f
 	 srl		%o4, 6, %o3
 
 	.align		64
-loop:
+1:
 	EX3(ldda [%o1 + 0x008] %asi, %f2)
 	faligndata	%f12, %f14, %f28
 	EX3(ldda [%o1 + 0x010] %asi, %f4)
@@ -240,11 +240,10 @@ loop:
 	faligndata	%f10, %f12, %f26
 	subcc		%o3, 0x01, %o3
 	add		%o1, 0x40, %o1
-	bg,pt		%XCC, loop
+	bg,pt		%XCC, 1b
 	 add		%o0, 0x40, %o0
 
 	/* Finally we copy the last full 64-byte block. */
-loopfini:
 	EX3(ldda [%o1 + 0x008] %asi, %f2)
 	faligndata	%f12, %f14, %f28
 	EX3(ldda [%o1 + 0x010] %asi, %f4)
@@ -279,12 +278,11 @@ loopfini:
 	 * Also notice how this code is careful not to perform a
 	 * load past the end of the src buffer.
 	 */
-loopend:
 	and		%o2, 0x3f, %o2
 	andcc		%o2, 0x38, %g2
-	be,pn		%XCC, endcruft
+	be,pn		%XCC, 10f
 	 subcc		%g2, 0x8, %g2
-	be,pn		%XCC, endcruft
+	be,pn		%XCC, 10f
 	 cmp		%g1, 0
 
 	be,a,pt		%XCC, 1f
@@ -296,7 +294,7 @@ loopend:
 	subcc		%g2, 0x8, %g2
 	faligndata	%f0, %f2, %f8
 	std		%f8, [%o0 + 0x00]
-	be,pn		%XCC, endcruft
+	be,pn		%XCC, 10f
 	 add		%o0, 0x8, %o0
 	EX(ldda [%o1 + 0x08] %asi, %f0, add %o2, %g0)
 	add		%o1, 0x8, %o1
@@ -311,15 +309,15 @@ loopend:
 	 * Note that %g1 is (src & 0x3) saved above before the
 	 * alignaddr was performed.
 	 */
-endcruft:
+10:
 	cmp		%o2, 0
 	add		%o1, %g1, %o1
 	VISExitHalf
-	be,pn		%XCC, out
+	be,pn		%XCC, 85f
 	 sub		%o0, %o1, %o3
 
 	andcc		%g1, 0x7, %g0
-	bne,pn		%icc, small_copy_unaligned
+	bne,pn		%icc, 90f
 	 andcc		%o2, 0x8, %g0
 	be,pt		%icc, 1f
 	 nop
@@ -342,17 +340,16 @@ endcruft:
 	add		%o1, 0x2, %o1
 
 1:	andcc		%o2, 0x1, %g0
-	be,pt		%icc, out
+	be,pt		%icc, 85f
 	 nop
 	EXNV(lduba [%o1] %asi, %o5, and %o2, 0x1)
-	ba,pt		%xcc, out
+	ba,pt		%xcc, 85f
 	 stb		%o5, [%o1 + %o3]
 
-medium_copy: /* 16 < len <= 64 */
-	bne,pn		%XCC, small_copy_unaligned
+70: /* 16 < len <= 64 */
+	bne,pn		%XCC, 90f
 	 sub		%o0, %o1, %o3
 
-medium_copy_aligned:
 	andn		%o2, 0x7, %o4
 	and		%o2, 0x7, %o2
 1:	subcc		%o4, 0x8, %o4
@@ -368,32 +365,32 @@ medium_copy_aligned:
 	stw		%o5, [%o1 + %o3]
 	add		%o1, 0x4, %o1
 1:	cmp		%o2, 0
-	be,pt		%XCC, out
+	be,pt		%XCC, 85f
 	 nop
-	ba,pt		%xcc, small_copy_unaligned
+	ba,pt		%xcc, 90f
 	 nop
 
-small_copy: /* 0 < len <= 16 */
+80: /* 0 < len <= 16 */
 	andcc		%o3, 0x3, %g0
-	bne,pn		%XCC, small_copy_unaligned
+	bne,pn		%XCC, 90f
 	 sub		%o0, %o1, %o3
 
-small_copy_aligned:
+1:
 	subcc		%o2, 4, %o2
 	EXNV(lduwa [%o1] %asi, %g1, add %o2, %g0)
 	stw		%g1, [%o1 + %o3]
-	bgu,pt		%XCC, small_copy_aligned
+	bgu,pt		%XCC, 1b
 	 add		%o1, 4, %o1
 
-out:	retl
+85:	retl
 	 clr		%o0
 
 	.align	32
-small_copy_unaligned:
+90:
 	subcc		%o2, 1, %o2
 	EXNV(lduba [%o1] %asi, %g1, add %o2, %g0)
 	stb		%g1, [%o1 + %o3]
-	bgu,pt		%XCC, small_copy_unaligned
+	bgu,pt		%XCC, 90b
 	 add		%o1, 1, %o1
 	retl
 	 clr		%o0

arch/sparc64/lib/U3copy_to_user.S

@@ -159,24 +159,24 @@ U3copy_to_user:	/* %o0=dst, %o1=src, %o2=len */
 	 nop
 
 	cmp		%o2, 0
-	be,pn		%XCC, out
+	be,pn		%XCC, 85f
 	 or		%o0, %o1, %o3
 	cmp		%o2, 16
-	bleu,a,pn	%XCC, small_copy
+	bleu,a,pn	%XCC, 80f
 	 or		%o3, %o2, %o3
 
 	cmp		%o2, 256
-	blu,pt		%XCC, medium_copy
+	blu,pt		%XCC, 70f
 	 andcc		%o3, 0x7, %g0
 
-	ba,pt		%xcc, enter
+	ba,pt		%xcc, 1f
 	 andcc		%o0, 0x3f, %g2
 
 	/* Here len >= 256 and condition codes reflect execution
 	 * of "andcc %o0, 0x7, %g2", done by caller.
 	 */
 	.align		64
-enter:
+1:
 	/* Is 'dst' already aligned on an 64-byte boundary? */
 	be,pt		%XCC, 2f
 
@@ -199,11 +199,11 @@ enter:
 
 2:	VISEntryHalf
 	and		%o1, 0x7, %g1
-	ba,pt		%xcc, begin
+	ba,pt		%xcc, 1f
 	 alignaddr	%o1, %g0, %o1
 
 	.align		64
-begin:
+1:
 	membar		#StoreLoad | #StoreStore | #LoadStore
 	prefetch	[%o1 + 0x000], #one_read
 	prefetch	[%o1 + 0x040], #one_read
@@ -232,11 +232,11 @@ begin:
 
 	sub		%o4, 0x80, %o4
 	add		%o1, 0x40, %o1
-	ba,pt		%xcc, loop
+	ba,pt		%xcc, 1f
 	 srl		%o4, 6, %o3
 
 	.align		64
-loop:
+1:
 	ldd		[%o1 + 0x008], %f2
 	faligndata	%f12, %f14, %f28
 	ldd		[%o1 + 0x010], %f4
@@ -259,11 +259,10 @@ loop:
 	faligndata	%f10, %f12, %f26
 	subcc		%o3, 0x01, %o3
 	add		%o1, 0x40, %o1
-	bg,pt		%XCC, loop
+	bg,pt		%XCC, 1b
 	 add		%o0, 0x40, %o0
 
 	/* Finally we copy the last full 64-byte block. */
-loopfini:
 	ldd		[%o1 + 0x008], %f2
 	faligndata	%f12, %f14, %f28
 	ldd		[%o1 + 0x010], %f4
@@ -298,12 +297,11 @@ loopfini:
 	 * Also notice how this code is careful not to perform a
 	 * load past the end of the src buffer.
 	 */
-loopend:
 	and		%o2, 0x3f, %o2
 	andcc		%o2, 0x38, %g2
-	be,pn		%XCC, endcruft
+	be,pn		%XCC, 2f
 	 subcc		%g2, 0x8, %g2
-	be,pn		%XCC, endcruft
+	be,pn		%XCC, 2f
 	 cmp		%g1, 0
 
 	be,a,pt		%XCC, 1f
@@ -315,7 +313,7 @@ loopend:
 	subcc		%g2, 0x8, %g2
 	faligndata	%f0, %f2, %f8
 	EX(stda %f8, [%o0 + 0x00] %asi, add %o2, 0x8)
-	be,pn		%XCC, endcruft
+	be,pn		%XCC, 2f
 	 add		%o0, 0x8, %o0
 	ldd		[%o1 + 0x08], %f0
 	add		%o1, 0x8, %o1
@@ -330,15 +328,15 @@ loopend:
 	 * Note that %g1 is (src & 0x3) saved above before the
 	 * alignaddr was performed.
 	 */
-endcruft:
+2:
 	cmp		%o2, 0
 	add		%o1, %g1, %o1
 	VISExitHalf
-	be,pn		%XCC, out
+	be,pn		%XCC, 85f
 	 sub		%o0, %o1, %o3
 
 	andcc		%g1, 0x7, %g0
-	bne,pn		%icc, small_copy_unaligned
+	bne,pn		%icc, 90f
 	 andcc		%o2, 0x8, %g0
 	be,pt		%icc, 1f
 	 nop
@@ -361,17 +359,16 @@ endcruft:
 	add		%o1, 0x2, %o1
 
 1:	andcc		%o2, 0x1, %g0
-	be,pt		%icc, out
+	be,pt		%icc, 85f
 	 nop
 	ldub		[%o1], %o5
-	ba,pt		%xcc, out
+	ba,pt		%xcc, 85f
 	 EXNV(stba %o5, [%o1 + %o3] ASI_AIUS, and %o2, 0x1)
 
-medium_copy: /* 16 < len <= 64 */
-	bne,pn		%XCC, small_copy_unaligned
+70: /* 16 < len <= 64 */
+	bne,pn		%XCC, 90f
 	 sub		%o0, %o1, %o3
 
-medium_copy_aligned:
 	andn		%o2, 0x7, %o4
 	and		%o2, 0x7, %o2
 1:	subcc		%o4, 0x8, %o4
@@ -387,32 +384,32 @@ medium_copy_aligned:
 	EXNV3(stwa %o5, [%o1 + %o3] ASI_AIUS, add %o2, %g0)
 	add		%o1, 0x4, %o1
 1:	cmp		%o2, 0
-	be,pt		%XCC, out
+	be,pt		%XCC, 85f
 	 nop
-	ba,pt		%xcc, small_copy_unaligned
+	ba,pt		%xcc, 90f
 	 nop
 
-small_copy: /* 0 < len <= 16 */
+80: /* 0 < len <= 16 */
 	andcc		%o3, 0x3, %g0
-	bne,pn		%XCC, small_copy_unaligned
+	bne,pn		%XCC, 90f
 	 sub		%o0, %o1, %o3
 
-small_copy_aligned:
+1:
 	subcc		%o2, 4, %o2
 	lduw		[%o1], %g1
 	EXNV3(stwa %g1, [%o1 + %o3] ASI_AIUS, add %o2, %g0)
-	bgu,pt		%XCC, small_copy_aligned
+	bgu,pt		%XCC, 1b
 	 add		%o1, 4, %o1
 
-out:	retl
+85:	retl
 	 clr		%o0
 
 	.align	32
-small_copy_unaligned:
+90:
 	subcc		%o2, 1, %o2
 	ldub		[%o1], %g1
 	EXNV2(stba %g1, [%o1 + %o3] ASI_AIUS, add %o2, %g0)
-	bgu,pt		%XCC, small_copy_unaligned
+	bgu,pt		%XCC, 90b
 	 add		%o1, 1, %o1
 	retl
 	 clr		%o0

arch/sparc64/lib/U3memcpy.S

@@ -45,24 +45,24 @@
 U3memcpy:	/* %o0=dst, %o1=src, %o2=len */
 	mov		%o0, %g5
 	cmp		%o2, 0
-	be,pn		%XCC, out
+	be,pn		%XCC, 85f
 	 or		%o0, %o1, %o3
 	cmp		%o2, 16
-	bleu,a,pn	%XCC, small_copy
+	bleu,a,pn	%XCC, 70f
 	 or		%o3, %o2, %o3
 
 	cmp		%o2, 256
-	blu,pt		%XCC, medium_copy
+	blu,pt		%XCC, 80f
 	 andcc		%o3, 0x7, %g0
 
-	ba,pt		%xcc, enter
+	ba,pt		%xcc, 1f
 	 andcc		%o0, 0x3f, %g2
 
 	/* Here len >= 256 and condition codes reflect execution
 	 * of "andcc %o0, 0x7, %g2", done by caller.
 	 */
 	.align		64
-enter:
+1:
 	/* Is 'dst' already aligned on an 64-byte boundary? */
 	be,pt		%XCC, 2f
 
@@ -85,11 +85,11 @@ enter:
 
 2:	VISEntryHalf
 	and		%o1, 0x7, %g1
-	ba,pt		%xcc, begin
+	ba,pt		%xcc, 1f
 	 alignaddr	%o1, %g0, %o1
 
 	.align		64
-begin:
+1:
 	membar		#StoreLoad | #StoreStore | #LoadStore
 	prefetch	[%o1 + 0x000], #one_read
 	prefetch	[%o1 + 0x040], #one_read
@@ -118,11 +118,11 @@ begin:
 
 	sub		%o4, 0x80, %o4
 	add		%o1, 0x40, %o1
-	ba,pt		%xcc, loop
+	ba,pt		%xcc, 1f
 	 srl		%o4, 6, %o3
 
 	.align		64
-loop:
+1:
 	ldd		[%o1 + 0x008], %f2
 	faligndata	%f12, %f14, %f28
 	ldd		[%o1 + 0x010], %f4
@@ -145,11 +145,10 @@ loop:
 	faligndata	%f10, %f12, %f26
 	subcc		%o3, 0x01, %o3
 	add		%o1, 0x40, %o1
-	bg,pt		%XCC, loop
+	bg,pt		%XCC, 1b
 	 add		%o0, 0x40, %o0
 
 	/* Finally we copy the last full 64-byte block. */
-loopfini:
 	ldd		[%o1 + 0x008], %f2
 	faligndata	%f12, %f14, %f28
 	ldd		[%o1 + 0x010], %f4
@@ -183,12 +182,11 @@ loopfini:
 	 * Also notice how this code is careful not to perform a
 	 * load past the end of the src buffer.
 	 */
-loopend:
 	and		%o2, 0x3f, %o2
 	andcc		%o2, 0x38, %g2
-	be,pn		%XCC, endcruft
+	be,pn		%XCC, 2f
 	 subcc		%g2, 0x8, %g2
-	be,pn		%XCC, endcruft
+	be,pn		%XCC, 2f
 	 cmp		%g1, 0
 
 	be,a,pt		%XCC, 1f
@@ -200,7 +198,7 @@ loopend:
 	subcc		%g2, 0x8, %g2
 	faligndata	%f0, %f2, %f8
 	std		%f8, [%o0 + 0x00]
-	be,pn		%XCC, endcruft
+	be,pn		%XCC, 2f
 	 add		%o0, 0x8, %o0
 	ldd		[%o1 + 0x08], %f0
 	add		%o1, 0x8, %o1
@@ -215,15 +213,15 @@ loopend:
 	 * Note that %g1 is (src & 0x3) saved above before the
 	 * alignaddr was performed.
 	 */
-endcruft:
+2:
 	cmp		%o2, 0
 	add		%o1, %g1, %o1
 	VISExitHalf
-	be,pn		%XCC, out
+	be,pn		%XCC, 85f
 	 sub		%o0, %o1, %o3
 
 	andcc		%g1, 0x7, %g0
-	bne,pn		%icc, small_copy_unaligned
+	bne,pn		%icc, 90f
 	 andcc		%o2, 0x8, %g0
 	be,pt		%icc, 1f
 	 nop
@@ -246,17 +244,16 @@ endcruft:
 	add		%o1, 0x2, %o1
 
 1:	andcc		%o2, 0x1, %g0
-	be,pt		%icc, out
+	be,pt		%icc, 85f
 	 nop
 	ldub		[%o1], %o5
-	ba,pt		%xcc, out
+	ba,pt		%xcc, 85f
 	 stb		%o5, [%o1 + %o3]
 
-medium_copy: /* 16 < len <= 64 */
-	bne,pn		%XCC, small_copy_unaligned
+70: /* 16 < len <= 64 */
+	bne,pn		%XCC, 90f
 	 sub		%o0, %o1, %o3
 
-medium_copy_aligned:
 	andn		%o2, 0x7, %o4
 	and		%o2, 0x7, %o2
 1:	subcc		%o4, 0x8, %o4
@@ -272,32 +269,32 @@ medium_copy_aligned:
 	stw		%o5, [%o1 + %o3]
 	add		%o1, 0x4, %o1
 1:	cmp		%o2, 0
-	be,pt		%XCC, out
+	be,pt		%XCC, 85f
 	 nop
-	ba,pt		%xcc, small_copy_unaligned
+	ba,pt		%xcc, 90f
 	 nop
 
-small_copy: /* 0 < len <= 16 */
+80: /* 0 < len <= 16 */
 	andcc		%o3, 0x3, %g0
-	bne,pn		%XCC, small_copy_unaligned
+	bne,pn		%XCC, 90f
 	 sub		%o0, %o1, %o3
 
-small_copy_aligned:
+1:
 	subcc		%o2, 4, %o2
 	lduw		[%o1], %g1
 	stw		%g1, [%o1 + %o3]
-	bgu,pt		%XCC, small_copy_aligned
+	bgu,pt		%XCC, 1b
 	 add		%o1, 4, %o1
 
-out:	retl
+85:	retl
 	 mov		%g5, %o0
 
 	.align	32
-small_copy_unaligned:
+90:
 	subcc		%o2, 1, %o2
 	ldub		[%o1], %g1
 	stb		%g1, [%o1 + %o3]
-	bgu,pt		%XCC, small_copy_unaligned
+	bgu,pt		%XCC, 90b
 	 add		%o1, 1, %o1
 	retl
 	 mov		%g5, %o0