powerpc: Improve memcmp performance for POWER8

Submitted by Rajalakshmi Srinivasaraghavan on April 6, 2017, 7:29 a.m.

Details

Message ID 1491463760-1004-1-git-send-email-raji@linux.vnet.ibm.com
State New
Headers show

Commit Message

Rajalakshmi Srinivasaraghavan April 6, 2017, 7:29 a.m.
Vectorization improves performance over the current implementation.
Tested on powerpc64 and powerpc64le.

2017-04-06  Rajalakshmi Srinivasaraghavan  <raji@linux.vnet.ibm.com>

	* sysdeps/powerpc/powerpc64/multiarch/Makefile
	(sysdep_routines): Add memcmp-power8.
	* sysdeps/powerpc/powerpc64/multiarch/ifunc-impl-list.c
	(memcmp): Add __memcmp_power8 to list of memcmp functions.
	* sysdeps/powerpc/powerpc64/multiarch/memcmp.c
	(memcmp): Add __memcmp_power8 to ifunc list.
	* sysdeps/powerpc/powerpc64/multiarch/memcmp-power8.S: New file.
	* sysdeps/powerpc/powerpc64/power8/memcmp.S: New file.
---
 sysdeps/powerpc/powerpc64/multiarch/Makefile       |    5 +-
 .../powerpc/powerpc64/multiarch/ifunc-impl-list.c  |    2 +
 .../powerpc/powerpc64/multiarch/memcmp-power8.S    |   42 +
 sysdeps/powerpc/powerpc64/multiarch/memcmp.c       |    3 +
 sysdeps/powerpc/powerpc64/power8/memcmp.S          | 1445 ++++++++++++++++++++
 5 files changed, 1495 insertions(+), 2 deletions(-)
 create mode 100644 sysdeps/powerpc/powerpc64/multiarch/memcmp-power8.S
 create mode 100644 sysdeps/powerpc/powerpc64/power8/memcmp.S

Comments

Tulio Magno Quites Machado Filho May 17, 2017, 9:50 p.m.
Rajalakshmi Srinivasaraghavan <raji@linux.vnet.ibm.com> writes:

> diff --git a/sysdeps/powerpc/powerpc64/multiarch/memcmp-power8.S b/sysdeps/powerpc/powerpc64/multiarch/memcmp-power8.S
> new file mode 100644
> index 0000000..a568be4
> --- /dev/null
> +++ b/sysdeps/powerpc/powerpc64/multiarch/memcmp-power8.S
> @@ -0,0 +1,42 @@
> +/* Optimized memcmp implementation for PowerPC64/POWER8.
> +   Copyright (C) 2017 Free Software Foundation, Inc.
> +   This file is part of the GNU C Library.
> +
> +   The GNU C Library is free software; you can redistribute it and/or
> +   modify it under the terms of the GNU Lesser General Public
> +   License as published by the Free Software Foundation; either
> +   version 2.1 of the License, or (at your option) any later version.
> +
> +   The GNU C Library is distributed in the hope that it will be useful,
> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> +   Lesser General Public License for more details.
> +
> +   You should have received a copy of the GNU Lesser General Public
> +   License along with the GNU C Library; if not, see
> +   <http://www.gnu.org/licenses/>.  */
> +
> +#include <sysdep.h>
> +
> +#undef EALIGN
> +#define EALIGN(name, alignt, words)				\
> +  .section ".text";						\
> +  ENTRY_2(__memcmp_power8)					\
> +  .align ALIGNARG(alignt);					\
> +  EALIGN_W_##words;						\
> +  BODY_LABEL(__memcmp_power8):					\
> +  cfi_startproc;						\
> +  LOCALENTRY(__memcmp_power8)
> +
> +#undef END
> +#define END(name)						\
> +  cfi_endproc;							\
> +  TRACEBACK(__memcmp_power8)					\
> +  END_2(__memcmp_power8)
> +
> +#undef libc_hidden_builtin_def
> +#define libc_hidden_builtin_def(name)
> +#undef weak_alias
> +#define weak_alias(name,alias)
> +
> +#include <sysdeps/powerpc/powerpc64/power8/memcmp.S>

The patch LGTM in overall, except for this file.
I think it should be implemented as in commit b6a66222096c.

LGTM with that change.

Patch hide | download patch | download mbox

diff --git a/sysdeps/powerpc/powerpc64/multiarch/Makefile b/sysdeps/powerpc/powerpc64/multiarch/Makefile
index 38233a7..5f21062 100644
--- a/sysdeps/powerpc/powerpc64/multiarch/Makefile
+++ b/sysdeps/powerpc/powerpc64/multiarch/Makefile
@@ -1,7 +1,8 @@ 
 ifeq ($(subdir),string)
 sysdep_routines += memcpy-power7 memcpy-a2 memcpy-power6 memcpy-cell \
-		   memcpy-power4 memcpy-ppc64 memcmp-power7 memcmp-power4 \
-		   memcmp-ppc64 memset-power7 memset-power6 memset-power4 \
+		   memcpy-power4 memcpy-ppc64 \
+		   memcmp-power8 memcmp-power7 memcmp-power4 memcmp-ppc64 \
+		   memset-power7 memset-power6 memset-power4 \
 		   memset-ppc64 memset-power8 \
 		   mempcpy-power7 mempcpy-ppc64 memchr-power7 memchr-ppc64 \
 		   memrchr-power7 memrchr-ppc64 rawmemchr-power7 \
diff --git a/sysdeps/powerpc/powerpc64/multiarch/ifunc-impl-list.c b/sysdeps/powerpc/powerpc64/multiarch/ifunc-impl-list.c
index 30a0133..90ea1e5 100644
--- a/sysdeps/powerpc/powerpc64/multiarch/ifunc-impl-list.c
+++ b/sysdeps/powerpc/powerpc64/multiarch/ifunc-impl-list.c
@@ -146,6 +146,8 @@  __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array,
 
   /* Support sysdeps/powerpc/powerpc64/multiarch/memcmp.c.  */
   IFUNC_IMPL (i, name, memcmp,
+	      IFUNC_IMPL_ADD (array, i, memcmp, hwcap2 & PPC_FEATURE2_ARCH_2_07,
+			      __memcmp_power8)
 	      IFUNC_IMPL_ADD (array, i, memcmp, hwcap & PPC_FEATURE_HAS_VSX,
 			      __memcmp_power7)
 	      IFUNC_IMPL_ADD (array, i, memcmp, hwcap & PPC_FEATURE_POWER4,
diff --git a/sysdeps/powerpc/powerpc64/multiarch/memcmp-power8.S b/sysdeps/powerpc/powerpc64/multiarch/memcmp-power8.S
new file mode 100644
index 0000000..a568be4
--- /dev/null
+++ b/sysdeps/powerpc/powerpc64/multiarch/memcmp-power8.S
@@ -0,0 +1,42 @@ 
+/* Optimized memcmp implementation for PowerPC64/POWER8.
+   Copyright (C) 2017 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <http://www.gnu.org/licenses/>.  */
+
+#include <sysdep.h>
+
+#undef EALIGN
+#define EALIGN(name, alignt, words)				\
+  .section ".text";						\
+  ENTRY_2(__memcmp_power8)					\
+  .align ALIGNARG(alignt);					\
+  EALIGN_W_##words;						\
+  BODY_LABEL(__memcmp_power8):					\
+  cfi_startproc;						\
+  LOCALENTRY(__memcmp_power8)
+
+#undef END
+#define END(name)						\
+  cfi_endproc;							\
+  TRACEBACK(__memcmp_power8)					\
+  END_2(__memcmp_power8)
+
+#undef libc_hidden_builtin_def
+#define libc_hidden_builtin_def(name)
+#undef weak_alias
+#define weak_alias(name,alias)
+
+#include <sysdeps/powerpc/powerpc64/power8/memcmp.S>
diff --git a/sysdeps/powerpc/powerpc64/multiarch/memcmp.c b/sysdeps/powerpc/powerpc64/multiarch/memcmp.c
index ee69e22..0d315d5 100644
--- a/sysdeps/powerpc/powerpc64/multiarch/memcmp.c
+++ b/sysdeps/powerpc/powerpc64/multiarch/memcmp.c
@@ -26,11 +26,14 @@ 
 extern __typeof (memcmp) __memcmp_ppc attribute_hidden;
 extern __typeof (memcmp) __memcmp_power4 attribute_hidden;
 extern __typeof (memcmp) __memcmp_power7 attribute_hidden;
+extern __typeof (memcmp) __memcmp_power8 attribute_hidden;
 # undef memcmp
 
 /* Avoid DWARF definition DIE on ifunc symbol so that GDB can handle
    ifunc symbol properly.  */
 libc_ifunc_redirected (__redirect_memcmp, memcmp,
+		       (hwcap2 & PPC_FEATURE2_ARCH_2_07)
+		       ? __memcmp_power8 :
 		       (hwcap & PPC_FEATURE_HAS_VSX)
 		       ? __memcmp_power7
 		       : (hwcap & PPC_FEATURE_POWER4)
diff --git a/sysdeps/powerpc/powerpc64/power8/memcmp.S b/sysdeps/powerpc/powerpc64/power8/memcmp.S
new file mode 100644
index 0000000..f26d59d
--- /dev/null
+++ b/sysdeps/powerpc/powerpc64/power8/memcmp.S
@@ -0,0 +1,1445 @@ 
+/* Optimized memcmp implementation for POWER7/PowerPC64.
+   Copyright (C) 2010-2017 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <http://www.gnu.org/licenses/>.  */
+
+#include <sysdep.h>
+
+/* int [r3] memcmp (const char *s1 [r3],
+		    const char *s2 [r4],
+		    size_t size [r5])  */
+
+/* TODO: change these to the actual instructions when the minimum required
+   binutils allows it.  */
+#define MFVRD(r,v)	.long (0x7c000067 | ((v)<<(32-11)) | ((r)<<(32-16)))
+
+	.machine power7
+EALIGN (memcmp, 4, 0)
+	CALL_MCOUNT 3
+
+#define rRTN		r3
+#define rSTR1		r3	/* First string arg.  */
+#define rSTR2		r4	/* Second string arg.  */
+#define rN		r5	/* Max string length.  */
+#define rWORD1		r6	/* Current word in s1.  */
+#define rWORD2		r7	/* Current word in s2.  */
+#define rWORD3		r8	/* Next word in s1.  */
+#define rWORD4		r9	/* Next word in s2.  */
+#define rWORD5		r10	/* Next word in s1.  */
+#define rWORD6		r11	/* Next word in s2.  */
+
+#define rOFF8		r20	/* 8 bytes offset.  */
+#define rOFF16  	r21	/* 16 bytes offset.  */
+#define rOFF24		r22	/* 24 bytes offset.  */
+#define rOFF32		r23	/* 24 bytes offset.  */
+#define rWORD6_SHIFT	r24	/* Left rotation temp for rWORD8.  */
+#define rWORD4_SHIFT	r25	/* Left rotation temp for rWORD6.  */
+#define rWORD2_SHIFT	r26	/* Left rotation temp for rWORD4.  */
+#define rWORD8_SHIFT	r27	/* Left rotation temp for rWORD2.  */
+#define rSHR		r28	/* Unaligned shift right count.  */
+#define rSHL		r29	/* Unaligned shift left count.  */
+#define rWORD7		r30	/* Next word in s1.  */
+#define rWORD8		r31	/* Next word in s2.  */
+
+#define rWORD8SAVE	(-8)
+#define rWORD7SAVE	(-16)
+#define rOFF8SAVE	(-24)
+#define rOFF16SAVE	(-32)
+#define rOFF24SAVE	(-40)
+#define rOFF32SAVE	(-48)
+#define rSHRSAVE	(-56)
+#define rSHLSAVE	(-64)
+#define rWORD8SHIFTSAVE	(-72)
+#define rWORD2SHIFTSAVE	(-80)
+#define rWORD4SHIFTSAVE	(-88)
+#define rWORD6SHIFTSAVE	(-96)
+
+#ifdef __LITTLE_ENDIAN__
+# define LD	ldbrx
+#else
+# define LD	ldx
+#endif
+
+	xor	r10, rSTR2, rSTR1
+	cmpldi	cr6, rN, 0
+	cmpldi	cr1, rN, 8
+	clrldi.	r0, r10, 61
+	clrldi	r12, rSTR1, 61
+	cmpldi	cr5, r12, 0
+	beq-	cr6, L(zeroLength)
+	dcbt	0, rSTR1
+	dcbt	0, rSTR2
+	/* If less than 8 bytes or not aligned, use the unaligned
+	   byte loop.  */
+	blt	cr1, L(bytealigned)
+	bne	L(unalignedqw)
+/* At this point we know both strings have the same alignment and the
+   compare length is at least 8 bytes.  r12 contains the low order
+   3 bits of rSTR1 and cr5 contains the result of the logical compare
+   of r12 to 0.  If r12 == 0 then we are already double word
+   aligned and can perform the DW aligned loop.  */
+
+	.align	4
+L(samealignment):
+	or	r11, rSTR2, rSTR1
+	clrldi.	r11, r11, 60
+	beq	L(qw_align)
+	/* Try to align to QW else proceed to DW loop.  */
+	clrldi.	r10, r10, 60
+	bne	L(DW)
+	/* For the difference to reach QW alignment, load as DW.  */
+	clrrdi	rSTR1, rSTR1, 3
+	clrrdi	rSTR2, rSTR2, 3
+	subfic	r10, r12, 8
+	LD	rWORD1, 0, rSTR1
+	LD	rWORD2, 0, rSTR2
+	sldi	r9, r10, 3
+	subfic	r9, r9, 64
+	sld	rWORD1, rWORD1, r9
+	sld	rWORD2, rWORD2, r9
+	cmpld	cr6, rWORD1, rWORD2
+	addi	rSTR1, rSTR1, 8
+	addi	rSTR2, rSTR2, 8
+	bne	cr6, L(ret_diff)
+	subf	rN, r10, rN
+
+	cmpld	cr6, r11, r12
+	bgt	cr6, L(qw_align)
+	LD	rWORD1, 0, rSTR1
+	LD	rWORD2, 0, rSTR2
+	cmpld	cr6, rWORD1, rWORD2
+	addi	rSTR1, rSTR1, 8
+	addi	rSTR2, rSTR2, 8
+	bne	cr6, L(different)
+	cmpldi	cr6, rN, 8
+	ble	cr6, L(zeroLength)
+	addi	rN, rN, -8
+	/* Now both rSTR1 and rSTR2 are aligned to QW.  */
+	.align	4
+L(qw_align):
+	vspltisb	v0, 0
+	srdi.	r6, rN, 6
+	li	r8, 16
+	li	r10, 32
+	li	r11, 48
+	ble	cr0, L(lessthan64)
+	mtctr	r6
+	vspltisb	v8, 0
+	vspltisb	v6, 0
+	/* Aligned vector loop.  */
+	.align	4
+L(aligned_loop):
+	lvx	v4, 0, rSTR1
+	lvx	v5, 0, rSTR2
+	vcmpequb.	v7, v6, v8
+	bnl	cr6, L(different3)
+	lvx	v6, rSTR1, r8
+	lvx	v8, rSTR2, r8
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different2)
+	lvx	v4, rSTR1, r10
+	lvx	v5, rSTR2, r10
+	vcmpequb.	v7, v6, v8
+	bnl	cr6, L(different3)
+	lvx	v6, rSTR1, r11
+	lvx	v8, rSTR2, r11
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different2)
+	addi	rSTR1, rSTR1, 64
+	addi	rSTR2, rSTR2, 64
+	bdnz	L(aligned_loop)
+	vcmpequb.	v7, v6, v8
+	bnl	cr6, L(different3)
+	clrldi	rN, rN, 58
+	/* Handle remainder for aligned loop.  */
+	.align	4
+L(lessthan64):
+	mr	r9, rSTR1
+	cmpdi	cr6, rN, 0
+	li	rSTR1, 0
+	blelr	cr6
+	lvx	v4, 0, r9
+	lvx	v5, 0, rSTR2
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different1)
+	addi	rN, rN, -16
+
+	cmpdi	cr6, rN, 0
+	blelr	cr6
+	lvx	v4, r9, r8
+	lvx	v5, rSTR2, r8
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different1)
+	addi	rN, rN, -16
+
+	cmpdi	cr6, rN, 0
+	blelr	cr6
+	lvx	v4, r9, r10
+	lvx	v5, rSTR2, r10
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different1)
+	addi	rN, rN, -16
+
+	cmpdi	cr6, rN, 0
+	blelr	cr6
+	lvx	v4, r9, r11
+	lvx	v5, rSTR2, r11
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different1)
+	blr
+
+	/* Calculate and return the difference.  */
+	.align 4
+L(different1):
+	cmpdi	cr6, rN, 16
+	bge	cr6, L(different2)
+	/* Discard unwanted bytes.  */
+#ifdef __LITTLE_ENDIAN__
+	lvsr	v1, 0, rN
+	vperm	v4, v4, v0, v1
+	vperm	v5, v5, v0, v1
+#else
+	lvsl	v1, 0, rN
+	vperm	v4, v0, v4, v1
+	vperm	v5, v0, v5, v1
+#endif
+	vcmpequb.	v7, v4, v5
+	li	rRTN, 0
+	bltlr	cr6
+	.align 4
+L(different2):
+#ifdef __LITTLE_ENDIAN__
+	/* Reverse bytes for direct comparison.  */
+	lvsl	v10, r0, r0
+	vspltisb	v8, 15
+	vsububm	v9, v8, v10
+	vperm	v4, v4, v0, v9
+	vperm	v5, v5, v0, v9
+#endif
+	MFVRD(r7, v4)
+	MFVRD(r9, v5)
+	cmpld	cr6, r7, r9
+	bne	cr6, L(ret_diff)
+	/* Difference in second DW.  */
+	vsldoi	v4, v4, v4, 8
+	vsldoi	v5, v5, v5, 8
+	MFVRD(r7, v4)
+	MFVRD(r9, v5)
+	cmpld	cr6, r7, r9
+L(ret_diff):
+	li	rRTN, 1
+	bgtlr	cr6
+	li	rRTN, -1
+	blr
+	.align	4
+L(different3):
+#ifdef __LITTLE_ENDIAN__
+	/* Reverse bytes for direct comparison.  */
+	vspltisb	v9, 15
+	lvsl	v10, r0, r0
+	vsububm	v9, v9, v10
+	vperm	v6, v6, v0, v9
+	vperm	v8, v8, v0, v9
+#endif
+	MFVRD(r7, v6)
+	MFVRD(r9, v8)
+	cmpld	cr6, r7, r9
+	bne	cr6, L(ret_diff)
+	/* Difference in second DW.  */
+	vsldoi	v6, v6, v6, 8
+	vsldoi	v8, v8, v8, 8
+	MFVRD(r7, v6)
+	MFVRD(r9, v8)
+	cmpld	cr6, r7, r9
+	li	rRTN, 1
+	bgtlr	cr6
+	li	rRTN, -1
+	blr
+
+	.align 4
+L(different):
+	cmpldi	cr7, rN, 8
+	bgt	cr7, L(end)
+	/* Skip unwanted bytes.  */
+	sldi	r8, rN, 3
+	subfic	r8, r8, 64
+	srd	rWORD1, rWORD1, r8
+	srd	rWORD2, rWORD2, r8
+	cmpld	cr6, rWORD1, rWORD2
+	li	rRTN, 0
+	beqlr	cr6
+L(end):
+	li	rRTN, 1
+	bgtlr	cr6
+	li	rRTN, -1
+	blr
+
+	.align	4
+L(unalignedqw):
+	/* Proceed to DW unaligned loop,if there is a chance of pagecross.  */
+	rldicl	r9, rSTR1, 0, 52
+	add	r9, r9, rN
+	cmpldi	cr0, r9, 4096-16
+	bgt	cr0, L(unaligned)
+	rldicl	r9, rSTR2, 0, 52
+	add	r9, r9, rN
+	cmpldi	cr0, r9, 4096-16
+	bgt	cr0, L(unaligned)
+	li	r0, 0
+	li	r8, 16
+	vspltisb	v0, 0
+	/* Check if rSTR1 is aligned to QW.  */
+	andi.	r11, rSTR1, 0xF
+	beq	L(s1_align)
+
+	/* Compare 16B and align S1 to QW.  */
+#ifdef __LITTLE_ENDIAN__
+	lvsr	v10, 0, rSTR1	/* Compute mask.  */
+	lvsr	v6, 0, rSTR2	/* Compute mask.  */
+#else
+	lvsl	v10, 0, rSTR1	/* Compute mask.  */
+	lvsl	v6, 0, rSTR2	/* Compute mask.  */
+#endif
+	lvx	v5, 0, rSTR2
+	lvx	v9, rSTR2, r8
+#ifdef __LITTLE_ENDIAN__
+	vperm	v5, v9, v5, v6
+#else
+	vperm	v5, v5, v9, v6
+#endif
+	lvx	v4, 0, rSTR1
+	lvx	v9, rSTR1, r8
+#ifdef __LITTLE_ENDIAN__
+	vperm	v4, v9, v4, v10
+#else
+	vperm	v4, v4, v9, v10
+#endif
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different1)
+	cmpldi	cr6, rN, 16
+	ble	cr6, L(zeroLength)
+	subfic	r11, r11, 16
+	subf	rN, r11, rN
+	add	rSTR1, rSTR1, r11
+	add	rSTR2, rSTR2, r11
+
+	/* As s1 is QW aligned prepare for unaligned loop.  */
+	.align	4
+L(s1_align):
+#ifdef __LITTLE_ENDIAN__
+	lvsr	v6, 0, rSTR2
+#else
+	lvsl	v6, 0, rSTR2
+#endif
+	lvx	v5, 0, rSTR2
+	srdi.	r6, rN, 6
+	li	r10, 32
+	li	r11, 48
+	ble	cr0, L(lessthan64_unalign)
+	mtctr	r6
+	li 	r9, 64
+	/* Unaligned vector loop.  */
+	.align	4
+L(unalign_qwloop):
+	lvx	v4, 0, rSTR1
+	lvx	v10, rSTR2, r8
+#ifdef __LITTLE_ENDIAN__
+	vperm	v5, v10, v5, v6
+#else
+	vperm	v5, v5, v10, v6
+#endif
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different2)
+	vor	v5, v10, v10
+	lvx	v4, rSTR1, r8
+	lvx	v10, rSTR2, r10
+#ifdef __LITTLE_ENDIAN__
+	vperm	v5, v10, v5, v6
+#else
+	vperm	v5, v5, v10, v6
+#endif
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different2)
+	vor	v5, v10, v10
+	lvx	v4, rSTR1, r10
+	lvx	v10, rSTR2, r11
+#ifdef __LITTLE_ENDIAN__
+	vperm	v5, v10, v5, v6
+#else
+	vperm	v5, v5, v10, v6
+#endif
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different2)
+	vor	v5, v10, v10
+	lvx	v4, rSTR1, r11
+	lvx	v10, rSTR2, r9
+#ifdef __LITTLE_ENDIAN__
+	vperm	v5, v10, v5, v6
+#else
+	vperm	v5, v5, v10, v6
+#endif
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different2)
+	vor	v5, v10, v10
+	addi	rSTR1, rSTR1, 64
+	addi	rSTR2, rSTR2, 64
+	bdnz	L(unalign_qwloop)
+	clrldi	rN, rN, 58
+	/* Handle remainder for unaligned loop.  */
+	.align	4
+L(lessthan64_unalign):
+	mr	r9, rSTR1
+	cmpdi	cr6, rN, 0
+	li	rSTR1, 0
+	blelr	cr6
+	lvx	v4, 0, r9
+	lvx     v10, rSTR2, r8
+#ifdef __LITTLE_ENDIAN__
+	vperm	v5, v10, v5, v6
+#else
+	vperm	v5, v5, v10, v6
+#endif
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different1)
+	vor	v5, v10, v10
+	addi	rN, rN, -16
+
+	cmpdi	cr6, rN, 0
+	blelr	cr6
+	lvx	v4, r9, r8
+	lvx	v10, rSTR2, r10
+#ifdef __LITTLE_ENDIAN__
+	vperm	v5, v10, v5, v6
+#else
+	vperm	v5, v5, v10, v6
+#endif
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different1)
+	vor	v5, v10, v10
+	addi	rN, rN, -16
+
+	cmpdi	cr6, rN, 0
+	blelr	cr6
+	lvx	v4, r9, r10
+	lvx	v10, rSTR2, r11
+#ifdef __LITTLE_ENDIAN__
+	vperm	v5, v10, v5, v6
+#else
+	vperm	v5, v5, v10, v6
+#endif
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different1)
+	vor	v5, v10, v10
+	addi	rN, rN, -16
+
+	cmpdi	cr6, rN, 0
+	blelr	cr6
+	lvx	v4, r9, r11
+	addi	r11, r11, 16
+	lvx	v10, rSTR2, r11
+#ifdef __LITTLE_ENDIAN__
+	vperm	v5, v10, v5, v6
+#else
+	vperm	v5, v5, v10, v6
+#endif
+	vcmpequb.	v7, v5, v4
+	bnl	cr6, L(different1)
+	blr
+
+/* Otherwise we know the two strings have the same alignment (but not
+   yet DW).  So we force the string addresses to the next lower DW
+   boundary and special case this first DW using shift left to
+   eliminate bits preceding the first byte.  Since we want to join the
+   normal (DW aligned) compare loop, starting at the second double word,
+   we need to adjust the length (rN) and special case the loop
+   versioning for the first DW.  This ensures that the loop count is
+   correct and the first DW (shifted) is in the expected register pair.  */
+	.align	4
+L(DW):
+	std	rWORD8, rWORD8SAVE(r1)
+	std	rWORD7, rWORD7SAVE(r1)
+	std	rOFF8, rOFF8SAVE(r1)
+	std	rOFF16, rOFF16SAVE(r1)
+	std	rOFF24, rOFF24SAVE(r1)
+	std	rOFF32, rOFF32SAVE(r1)
+	cfi_offset(rWORD8, rWORD8SAVE)
+	cfi_offset(rWORD7, rWORD7SAVE)
+	cfi_offset(rOFF8, rOFF8SAVE)
+	cfi_offset(rOFF16, rOFF16SAVE)
+	cfi_offset(rOFF24, rOFF24SAVE)
+	cfi_offset(rOFF32, rOFF32SAVE)
+
+	li	rOFF8,8
+	li	rOFF16,16
+	li	rOFF24,24
+	li	rOFF32,32
+	clrrdi	rSTR1, rSTR1, 3
+	clrrdi	rSTR2, rSTR2, 3
+	beq	cr5, L(DWaligned)
+	add	rN, rN, r12
+	sldi	rWORD6, r12, 3
+	srdi	r0, rN, 5	/* Divide by 32.  */
+	andi.	r12, rN, 24	/* Get the DW remainder.  */
+	LD	rWORD1, 0, rSTR1
+	LD	rWORD2, 0, rSTR2
+	cmpldi	cr1, r12, 16
+	cmpldi	cr7, rN, 32
+	clrldi	rN, rN, 61
+	beq	L(dPs4)
+	mtctr	r0
+	bgt	cr1, L(dPs3)
+	beq	cr1, L(dPs2)
+
+/* Remainder is 8.  */
+	.align	3
+L(dsP1):
+	sld	rWORD5, rWORD1, rWORD6
+	sld	rWORD6, rWORD2, rWORD6
+	cmpld	cr5, rWORD5, rWORD6
+	blt	cr7, L(dP1x)
+/* Do something useful in this cycle since we have to branch anyway.  */
+	LD	rWORD1, rOFF8, rSTR1
+	LD	rWORD2, rOFF8, rSTR2
+	cmpld	cr7, rWORD1, rWORD2
+	b	L(dP1e)
+/* Remainder is 16.  */
+	.align	4
+L(dPs2):
+	sld	rWORD5, rWORD1, rWORD6
+	sld	rWORD6, rWORD2, rWORD6
+	cmpld	cr6, rWORD5, rWORD6
+	blt	cr7, L(dP2x)
+/* Do something useful in this cycle since we have to branch anyway.  */
+	LD	rWORD7, rOFF8, rSTR1
+	LD	rWORD8, rOFF8, rSTR2
+	cmpld	cr5, rWORD7, rWORD8
+	b	L(dP2e)
+/* Remainder is 24.  */
+	.align	4
+L(dPs3):
+	sld	rWORD3, rWORD1, rWORD6
+	sld	rWORD4, rWORD2, rWORD6
+	cmpld	cr1, rWORD3, rWORD4
+	b	L(dP3e)
+/* Count is a multiple of 32, remainder is 0.  */
+	.align	4
+L(dPs4):
+	mtctr	r0
+	sld	rWORD1, rWORD1, rWORD6
+	sld	rWORD2, rWORD2, rWORD6
+	cmpld	cr7, rWORD1, rWORD2
+	b	L(dP4e)
+
+/* At this point we know both strings are double word aligned and the
+   compare length is at least 8 bytes.  */
+	.align	4
+L(DWaligned):
+	andi.	r12, rN, 24	/* Get the DW remainder.  */
+	srdi	r0, rN, 5	/* Divide by 32.  */
+	cmpldi	cr1, r12, 16
+	cmpldi	cr7, rN, 32
+	clrldi	rN, rN, 61
+	beq	L(dP4)
+	bgt	cr1, L(dP3)
+	beq	cr1, L(dP2)
+
+/* Remainder is 8.  */
+	.align	4
+L(dP1):
+	mtctr	r0
+/* Normally we'd use rWORD7/rWORD8 here, but since we might exit early
+   (8-15 byte compare), we want to use only volatile registers.  This
+   means we can avoid restoring non-volatile registers since we did not
+   change any on the early exit path.  The key here is the non-early
+   exit path only cares about the condition code (cr5), not about which
+   register pair was used.  */
+	LD	rWORD5, 0, rSTR1
+	LD	rWORD6, 0, rSTR2
+	cmpld	cr5, rWORD5, rWORD6
+	blt	cr7, L(dP1x)
+	LD	rWORD1, rOFF8, rSTR1
+	LD	rWORD2, rOFF8, rSTR2
+	cmpld	cr7, rWORD1, rWORD2
+L(dP1e):
+	LD	rWORD3, rOFF16, rSTR1
+	LD	rWORD4, rOFF16, rSTR2
+	cmpld	cr1, rWORD3, rWORD4
+	LD	rWORD5, rOFF24, rSTR1
+	LD	rWORD6, rOFF24, rSTR2
+	cmpld	cr6, rWORD5, rWORD6
+	bne	cr5, L(dLcr5x)
+	bne	cr7, L(dLcr7x)
+
+	LD	rWORD7, rOFF32, rSTR1
+	LD	rWORD8, rOFF32, rSTR2
+	addi	rSTR1, rSTR1, 32
+	addi	rSTR2, rSTR2, 32
+	bne	cr1, L(dLcr1)
+	cmpld	cr5, rWORD7, rWORD8
+	bdnz	L(dLoop)
+	bne	cr6, L(dLcr6)
+	ld	rWORD8, rWORD8SAVE(r1)
+	ld	rWORD7, rWORD7SAVE(r1)
+	.align	3
+L(dP1x):
+	sldi.	r12, rN, 3
+	bne	cr5, L(dLcr5x)
+	subfic	rN, r12, 64	/* Shift count is 64 - (rN * 8).  */
+	bne	L(d00)
+	ld	rOFF8,  rOFF8SAVE(r1)
+	ld	rOFF16, rOFF16SAVE(r1)
+	ld	rOFF24, rOFF24SAVE(r1)
+	ld	rOFF32, rOFF32SAVE(r1)
+	li	rRTN, 0
+	blr
+
+/* Remainder is 16.  */
+	.align	4
+L(dP2):
+	mtctr	r0
+	LD	rWORD5, 0, rSTR1
+	LD	rWORD6, 0, rSTR2
+	cmpld	cr6, rWORD5, rWORD6
+	blt	cr7, L(dP2x)
+	LD	rWORD7, rOFF8, rSTR1
+	LD	rWORD8, rOFF8, rSTR2
+	cmpld	cr5, rWORD7, rWORD8
+L(dP2e):
+	LD	rWORD1, rOFF16, rSTR1
+	LD	rWORD2, rOFF16, rSTR2
+	cmpld	cr7, rWORD1, rWORD2
+	LD	rWORD3, rOFF24, rSTR1
+	LD	rWORD4, rOFF24, rSTR2
+	cmpld	cr1, rWORD3, rWORD4
+	addi	rSTR1, rSTR1, 8
+	addi	rSTR2, rSTR2, 8
+	bne	cr6, L(dLcr6)
+	bne	cr5, L(dLcr5)
+	b	L(dLoop2)
+	.align	4
+L(dP2x):
+	LD	rWORD3, rOFF8, rSTR1
+	LD	rWORD4, rOFF8, rSTR2
+	cmpld	cr1, rWORD3, rWORD4
+	sldi.	r12, rN, 3
+	bne	cr6, L(dLcr6x)
+	addi	rSTR1, rSTR1, 8
+	addi	rSTR2, rSTR2, 8
+	bne	cr1, L(dLcr1x)
+	subfic	rN, r12, 64	/* Shift count is 64 - (rN * 8).  */
+	bne	L(d00)
+	ld	rOFF8,  rOFF8SAVE(r1)
+	ld	rOFF16, rOFF16SAVE(r1)
+	ld	rOFF24, rOFF24SAVE(r1)
+	ld	rOFF32, rOFF32SAVE(r1)
+	li	rRTN, 0
+	blr
+
+/* Remainder is 24.  */
+	.align	4
+L(dP3):
+	mtctr	r0
+	LD	rWORD3, 0, rSTR1
+	LD	rWORD4, 0, rSTR2
+	cmpld	cr1, rWORD3, rWORD4
+L(dP3e):
+	LD	rWORD5, rOFF8, rSTR1
+	LD	rWORD6, rOFF8, rSTR2
+	cmpld	cr6, rWORD5, rWORD6
+	blt	cr7, L(dP3x)
+	LD	rWORD7, rOFF16, rSTR1
+	LD	rWORD8, rOFF16, rSTR2
+	cmpld	cr5, rWORD7, rWORD8
+	LD	rWORD1, rOFF24, rSTR1
+	LD	rWORD2, rOFF24, rSTR2
+	cmpld	cr7, rWORD1, rWORD2
+	addi	rSTR1, rSTR1, 16
+	addi	rSTR2, rSTR2, 16
+	bne	cr1, L(dLcr1)
+	bne	cr6, L(dLcr6)
+	b	L(dLoop1)
+/* Again we are on a early exit path (24-31 byte compare), we want to
+   only use volatile registers and avoid restoring non-volatile
+   registers.  */
+	.align	4
+L(dP3x):
+	LD	rWORD1, rOFF16, rSTR1
+	LD	rWORD2, rOFF16, rSTR2
+	cmpld	cr7, rWORD1, rWORD2
+	sldi.	r12, rN, 3
+	bne	cr1, L(dLcr1x)
+	addi	rSTR1, rSTR1, 16
+	addi	rSTR2, rSTR2, 16
+	bne	cr6, L(dLcr6x)
+	subfic	rN, r12, 64	/* Shift count is 64 - (rN * 8).  */
+	bne	cr7, L(dLcr7x)
+	bne	L(d00)
+	ld	rOFF8,  rOFF8SAVE(r1)
+	ld	rOFF16, rOFF16SAVE(r1)
+	ld	rOFF24, rOFF24SAVE(r1)
+	ld	rOFF32, rOFF32SAVE(r1)
+	li	rRTN, 0
+	blr
+
+/* Count is a multiple of 32, remainder is 0.  */
+	.align	4
+L(dP4):
+	mtctr	r0
+	LD	rWORD1, 0, rSTR1
+	LD	rWORD2, 0, rSTR2
+	cmpld	cr7, rWORD1, rWORD2
+L(dP4e):
+	LD	rWORD3, rOFF8, rSTR1
+	LD	rWORD4, rOFF8, rSTR2
+	cmpld	cr1, rWORD3, rWORD4
+	LD	rWORD5, rOFF16, rSTR1
+	LD	rWORD6, rOFF16, rSTR2
+	cmpld	cr6, rWORD5, rWORD6
+	LD	rWORD7, rOFF24, rSTR1
+	LD	rWORD8, rOFF24, rSTR2
+	addi	rSTR1, rSTR1, 24
+	addi	rSTR2, rSTR2, 24
+	cmpld	cr5, rWORD7, rWORD8
+	bne	cr7, L(dLcr7)
+	bne	cr1, L(dLcr1)
+	bdz-	L(d24)		/* Adjust CTR as we start with +4.  */
+/* This is the primary loop.  */
+	.align	4
+L(dLoop):
+	LD	rWORD1, rOFF8, rSTR1
+	LD	rWORD2, rOFF8, rSTR2
+	cmpld	cr1, rWORD3, rWORD4
+	bne	cr6, L(dLcr6)
+L(dLoop1):
+	LD	rWORD3, rOFF16, rSTR1
+	LD	rWORD4, rOFF16, rSTR2
+	cmpld	cr6, rWORD5, rWORD6
+	bne	cr5, L(dLcr5)
+L(dLoop2):
+	LD	rWORD5, rOFF24, rSTR1
+	LD	rWORD6, rOFF24, rSTR2
+	cmpld	cr5, rWORD7, rWORD8
+	bne	cr7, L(dLcr7)
+L(dLoop3):
+	LD	rWORD7, rOFF32, rSTR1
+	LD	rWORD8, rOFF32, rSTR2
+	addi	rSTR1, rSTR1, 32
+	addi	rSTR2, rSTR2, 32
+	bne	cr1, L(dLcr1)
+	cmpld	cr7, rWORD1, rWORD2
+	bdnz	L(dLoop)
+
+L(dL4):
+	cmpld	cr1, rWORD3, rWORD4
+	bne	cr6, L(dLcr6)
+	cmpld	cr6, rWORD5, rWORD6
+	bne	cr5, L(dLcr5)
+	cmpld	cr5, rWORD7, rWORD8
+L(d44):
+	bne	cr7, L(dLcr7)
+L(d34):
+	bne	cr1, L(dLcr1)
+L(d24):
+	bne	cr6, L(dLcr6)
+L(d14):
+	sldi.	r12, rN, 3
+	bne	cr5, L(dLcr5)
+L(d04):
+	ld	rWORD8, rWORD8SAVE(r1)
+	ld	rWORD7, rWORD7SAVE(r1)
+	subfic	rN, r12, 64	/* Shift count is 64 - (rN * 8).  */
+	beq	L(duzeroLength)
+/* At this point we have a remainder of 1 to 7 bytes to compare.  Since
+   we are aligned it is safe to load the whole double word, and use
+   shift right double to eliminate bits beyond the compare length.  */
+L(d00):
+	LD	rWORD1, rOFF8, rSTR1
+	LD	rWORD2, rOFF8, rSTR2
+	srd	rWORD1, rWORD1, rN
+	srd	rWORD2, rWORD2, rN
+	cmpld	cr7, rWORD1, rWORD2
+	bne	cr7, L(dLcr7x)
+	ld	rOFF8,  rOFF8SAVE(r1)
+	ld	rOFF16, rOFF16SAVE(r1)
+	ld	rOFF24, rOFF24SAVE(r1)
+	ld	rOFF32, rOFF32SAVE(r1)
+	li	rRTN, 0
+	blr
+
+	.align	4
+L(dLcr7):
+	ld	rWORD8, rWORD8SAVE(r1)
+	ld	rWORD7, rWORD7SAVE(r1)
+L(dLcr7x):
+	ld	rOFF8,  rOFF8SAVE(r1)
+	ld	rOFF16, rOFF16SAVE(r1)
+	ld	rOFF24, rOFF24SAVE(r1)
+	ld	rOFF32, rOFF32SAVE(r1)
+	li	rRTN, 1
+	bgtlr	cr7
+	li	rRTN, -1
+	blr
+	.align	4
+L(dLcr1):
+	ld	rWORD8, rWORD8SAVE(r1)
+	ld	rWORD7, rWORD7SAVE(r1)
+L(dLcr1x):
+	ld	rOFF8,  rOFF8SAVE(r1)
+	ld	rOFF16, rOFF16SAVE(r1)
+	ld	rOFF24, rOFF24SAVE(r1)
+	ld	rOFF32, rOFF32SAVE(r1)
+	li	rRTN, 1
+	bgtlr	cr1
+	li	rRTN, -1
+	blr
+	.align	4
+L(dLcr6):
+	ld	rWORD8, rWORD8SAVE(r1)
+	ld	rWORD7, rWORD7SAVE(r1)
+L(dLcr6x):
+	ld	rOFF8,  rOFF8SAVE(r1)
+	ld	rOFF16, rOFF16SAVE(r1)
+	ld	rOFF24, rOFF24SAVE(r1)
+	ld	rOFF32, rOFF32SAVE(r1)
+	li	rRTN, 1
+	bgtlr	cr6
+	li	rRTN, -1
+	blr
+	.align	4
+L(dLcr5):
+	ld	rWORD8, rWORD8SAVE(r1)
+	ld	rWORD7, rWORD7SAVE(r1)
+L(dLcr5x):
+	ld	rOFF8,  rOFF8SAVE(r1)
+	ld	rOFF16, rOFF16SAVE(r1)
+	ld	rOFF24, rOFF24SAVE(r1)
+	ld	rOFF32, rOFF32SAVE(r1)
+	li	rRTN, 1
+	bgtlr	cr5
+	li	rRTN, -1
+	blr
+
+	.align	4
+L(bytealigned):
+	mtctr	rN
+
+/* We need to prime this loop.  This loop is swing modulo scheduled
+   to avoid pipe delays.  The dependent instruction latencies (load to
+   compare to conditional branch) is 2 to 3 cycles.  In this loop each
+   dispatch group ends in a branch and takes 1 cycle.  Effectively
+   the first iteration of the loop only serves to load operands and
+   branches based on compares are delayed until the next loop.
+
+   So we must precondition some registers and condition codes so that
+   we don't exit the loop early on the first iteration.  */
+
+	lbz	rWORD1, 0(rSTR1)
+	lbz	rWORD2, 0(rSTR2)
+	bdz	L(b11)
+	cmpld	cr7, rWORD1, rWORD2
+	lbz	rWORD3, 1(rSTR1)
+	lbz	rWORD4, 1(rSTR2)
+	bdz	L(b12)
+	cmpld	cr1, rWORD3, rWORD4
+	lbzu	rWORD5, 2(rSTR1)
+	lbzu	rWORD6, 2(rSTR2)
+	bdz	L(b13)
+	.align	4
+L(bLoop):
+	lbzu	rWORD1, 1(rSTR1)
+	lbzu	rWORD2, 1(rSTR2)
+	bne	cr7, L(bLcr7)
+
+	cmpld	cr6, rWORD5, rWORD6
+	bdz	L(b3i)
+
+	lbzu	rWORD3, 1(rSTR1)
+	lbzu	rWORD4, 1(rSTR2)
+	bne	cr1, L(bLcr1)
+
+	cmpld	cr7, rWORD1, rWORD2
+	bdz	L(b2i)
+
+	lbzu	rWORD5, 1(rSTR1)
+	lbzu	rWORD6, 1(rSTR2)
+	bne	cr6, L(bLcr6)
+
+	cmpld	cr1, rWORD3, rWORD4
+	bdnz	L(bLoop)
+
+/* We speculatively loading bytes before we have tested the previous
+   bytes.  But we must avoid overrunning the length (in the ctr) to
+   prevent these speculative loads from causing a segfault.  In this
+   case the loop will exit early (before the all pending bytes are
+   tested.  In this case we must complete the pending operations
+   before returning.  */
+L(b1i):
+	bne	cr7, L(bLcr7)
+	bne	cr1, L(bLcr1)
+	b	L(bx56)
+	.align	4
+L(b2i):
+	bne	cr6, L(bLcr6)
+	bne	cr7, L(bLcr7)
+	b	L(bx34)
+	.align	4
+L(b3i):
+	bne	cr1, L(bLcr1)
+	bne	cr6, L(bLcr6)
+	b	L(bx12)
+	.align	4
+L(bLcr7):
+	li	rRTN, 1
+	bgtlr	cr7
+	li	rRTN, -1
+	blr
+L(bLcr1):
+	li	rRTN, 1
+	bgtlr	cr1
+	li	rRTN, -1
+	blr
+L(bLcr6):
+	li	rRTN, 1
+	bgtlr	cr6
+	li	rRTN, -1
+	blr
+
+L(b13):
+	bne	cr7, L(bx12)
+	bne	cr1, L(bx34)
+L(bx56):
+	sub	rRTN, rWORD5, rWORD6
+	blr
+	nop
+L(b12):
+	bne	cr7, L(bx12)
+L(bx34):
+	sub	rRTN, rWORD3, rWORD4
+	blr
+L(b11):
+L(bx12):
+	sub	rRTN, rWORD1, rWORD2
+	blr
+
+	.align	4
+L(zeroLength):
+	li	rRTN, 0
+	blr
+
+	.align	4
+/* At this point we know the strings have different alignment and the
+   compare length is at least 8 bytes.  r12 contains the low order
+   3 bits of rSTR1 and cr5 contains the result of the logical compare
+   of r12 to 0.  If r12 == 0 then rStr1 is double word
+   aligned and can perform the DWunaligned loop.
+
+   Otherwise we know that rSTR1 is not already DW aligned yet.
+   So we can force the string addresses to the next lower DW
+   boundary and special case this first DW using shift left to
+   eliminate bits preceding the first byte.  Since we want to join the
+   normal (DWaligned) compare loop, starting at the second double word,
+   we need to adjust the length (rN) and special case the loop
+   versioning for the first DW.  This ensures that the loop count is
+   correct and the first DW (shifted) is in the expected resister pair.  */
+L(unaligned):
+	std	rWORD8, rWORD8SAVE(r1)
+	std	rWORD7, rWORD7SAVE(r1)
+	std	rOFF8, rOFF8SAVE(r1)
+	std	rOFF16, rOFF16SAVE(r1)
+	std	rOFF24, rOFF24SAVE(r1)
+	std	rOFF32, rOFF32SAVE(r1)
+	cfi_offset(rWORD8, rWORD8SAVE)
+	cfi_offset(rWORD7, rWORD7SAVE)
+	cfi_offset(rOFF8, rOFF8SAVE)
+	cfi_offset(rOFF16, rOFF16SAVE)
+	cfi_offset(rOFF24, rOFF24SAVE)
+	cfi_offset(rOFF32, rOFF32SAVE)
+	li	rOFF8,8
+	li	rOFF16,16
+	li	rOFF24,24
+	li	rOFF32,32
+	std	rSHL, rSHLSAVE(r1)
+	cfi_offset(rSHL, rSHLSAVE)
+	clrldi	rSHL, rSTR2, 61
+	beq	cr6, L(duzeroLength)
+	std	rSHR, rSHRSAVE(r1)
+	cfi_offset(rSHR, rSHRSAVE)
+	beq	cr5, L(DWunaligned)
+	std	rWORD8_SHIFT, rWORD8SHIFTSAVE(r1)
+	cfi_offset(rWORD8_SHIFT, rWORD8SHIFTSAVE)
+/* Adjust the logical start of rSTR2 to compensate for the extra bits
+   in the 1st rSTR1 DW.  */
+	sub	rWORD8_SHIFT, rSTR2, r12
+/* But do not attempt to address the DW before that DW that contains
+   the actual start of rSTR2.  */
+	clrrdi	rSTR2, rSTR2, 3
+	std	rWORD2_SHIFT, rWORD2SHIFTSAVE(r1)
+/* Compute the left/right shift counts for the unaligned rSTR2,
+   compensating for the logical (DW aligned) start of rSTR1.  */
+	clrldi	rSHL, rWORD8_SHIFT, 61
+	clrrdi	rSTR1, rSTR1, 3
+	std	rWORD4_SHIFT, rWORD4SHIFTSAVE(r1)
+	sldi	rSHL, rSHL, 3
+	cmpld	cr5, rWORD8_SHIFT, rSTR2
+	add	rN, rN, r12
+	sldi	rWORD6, r12, 3
+	std	rWORD6_SHIFT, rWORD6SHIFTSAVE(r1)
+	cfi_offset(rWORD2_SHIFT, rWORD2SHIFTSAVE)
+	cfi_offset(rWORD4_SHIFT, rWORD4SHIFTSAVE)
+	cfi_offset(rWORD6_SHIFT, rWORD6SHIFTSAVE)
+	subfic	rSHR, rSHL, 64
+	srdi	r0, rN, 5	/* Divide by 32.  */
+	andi.	r12, rN, 24	/* Get the DW remainder.  */
+/* We normally need to load 2 DWs to start the unaligned rSTR2, but in
+   this special case those bits may be discarded anyway.  Also we
+   must avoid loading a DW where none of the bits are part of rSTR2 as
+   this may cross a page boundary and cause a page fault.  */
+	li	rWORD8, 0
+	blt	cr5, L(dus0)
+	LD	rWORD8, 0, rSTR2
+	addi	rSTR2, rSTR2, 8
+	sld	rWORD8, rWORD8, rSHL
+
+L(dus0):
+	LD	rWORD1, 0, rSTR1
+	LD	rWORD2, 0, rSTR2
+	cmpldi	cr1, r12, 16
+	cmpldi	cr7, rN, 32
+	srd	r12, rWORD2, rSHR
+	clrldi	rN, rN, 61
+	beq	L(duPs4)
+	mtctr	r0
+	or	rWORD8, r12, rWORD8
+	bgt	cr1, L(duPs3)
+	beq	cr1, L(duPs2)
+
+/* Remainder is 8.  */
+	.align	4
+L(dusP1):
+	sld	rWORD8_SHIFT, rWORD2, rSHL
+	sld	rWORD7, rWORD1, rWORD6
+	sld	rWORD8, rWORD8, rWORD6
+	bge	cr7, L(duP1e)
+/* At this point we exit early with the first double word compare
+   complete and remainder of 0 to 7 bytes.  See L(du14) for details on
+   how we handle the remaining bytes.  */
+	cmpld	cr5, rWORD7, rWORD8
+	sldi.	rN, rN, 3
+	bne	cr5, L(duLcr5)
+	cmpld	cr7, rN, rSHR
+	beq	L(duZeroReturn)
+	li	r0, 0
+	ble	cr7, L(dutrim)
+	LD	rWORD2, rOFF8, rSTR2
+	srd	r0, rWORD2, rSHR
+	b	L(dutrim)
+/* Remainder is 16.  */
+	.align	4
+L(duPs2):
+	sld	rWORD6_SHIFT, rWORD2, rSHL
+	sld	rWORD5, rWORD1, rWORD6
+	sld	rWORD6, rWORD8, rWORD6
+	b	L(duP2e)
+/* Remainder is 24.  */
+	.align	4
+L(duPs3):
+	sld	rWORD4_SHIFT, rWORD2, rSHL
+	sld	rWORD3, rWORD1, rWORD6
+	sld	rWORD4, rWORD8, rWORD6
+	b	L(duP3e)
+/* Count is a multiple of 32, remainder is 0.  */
+	.align	4
+L(duPs4):
+	mtctr	r0
+	or	rWORD8, r12, rWORD8
+	sld	rWORD2_SHIFT, rWORD2, rSHL
+	sld	rWORD1, rWORD1, rWORD6
+	sld	rWORD2, rWORD8, rWORD6
+	b	L(duP4e)
+
+/* At this point we know rSTR1 is double word aligned and the
+   compare length is at least 8 bytes.  */
+	.align	4
+L(DWunaligned):
+	std	rWORD8_SHIFT, rWORD8SHIFTSAVE(r1)
+	clrrdi	rSTR2, rSTR2, 3
+	std	rWORD2_SHIFT, rWORD2SHIFTSAVE(r1)
+	srdi	r0, rN, 5	/* Divide by 32.  */
+	std	rWORD4_SHIFT, rWORD4SHIFTSAVE(r1)
+	andi.	r12, rN, 24	/* Get the DW remainder.  */
+	std	rWORD6_SHIFT, rWORD6SHIFTSAVE(r1)
+	cfi_offset(rWORD8_SHIFT, rWORD8SHIFTSAVE)
+	cfi_offset(rWORD2_SHIFT, rWORD2SHIFTSAVE)
+	cfi_offset(rWORD4_SHIFT, rWORD4SHIFTSAVE)
+	cfi_offset(rWORD6_SHIFT, rWORD6SHIFTSAVE)
+	sldi	rSHL, rSHL, 3
+	LD	rWORD6, 0, rSTR2
+	LD	rWORD8, rOFF8, rSTR2
+	addi	rSTR2, rSTR2, 8
+	cmpldi	cr1, r12, 16
+	cmpldi	cr7, rN, 32
+	clrldi	rN, rN, 61
+	subfic	rSHR, rSHL, 64
+	sld	rWORD6_SHIFT, rWORD6, rSHL
+	beq	L(duP4)
+	mtctr	r0
+	bgt	cr1, L(duP3)
+	beq	cr1, L(duP2)
+
+/* Remainder is 8.  */
+	.align	4
+L(duP1):
+	srd	r12, rWORD8, rSHR
+	LD	rWORD7, 0, rSTR1
+	sld	rWORD8_SHIFT, rWORD8, rSHL
+	or	rWORD8, r12, rWORD6_SHIFT
+	blt	cr7, L(duP1x)
+L(duP1e):
+	LD	rWORD1, rOFF8, rSTR1
+	LD	rWORD2, rOFF8, rSTR2
+	cmpld	cr5, rWORD7, rWORD8
+	srd	r0, rWORD2, rSHR
+	sld	rWORD2_SHIFT, rWORD2, rSHL
+	or	rWORD2, r0, rWORD8_SHIFT
+	LD	rWORD3, rOFF16, rSTR1
+	LD	rWORD4, rOFF16, rSTR2
+	cmpld	cr7, rWORD1, rWORD2
+	srd	r12, rWORD4, rSHR
+	sld	rWORD4_SHIFT, rWORD4, rSHL
+	bne	cr5, L(duLcr5)
+	or	rWORD4, r12, rWORD2_SHIFT
+	LD	rWORD5, rOFF24, rSTR1
+	LD	rWORD6, rOFF24, rSTR2
+	cmpld	cr1, rWORD3, rWORD4
+	srd	r0, rWORD6, rSHR
+	sld	rWORD6_SHIFT, rWORD6, rSHL
+	bne	cr7, L(duLcr7)
+	or	rWORD6, r0, rWORD4_SHIFT
+	cmpld	cr6, rWORD5, rWORD6
+	b	L(duLoop3)
+	.align	4
+/* At this point we exit early with the first double word compare
+   complete and remainder of 0 to 7 bytes.  See L(du14) for details on
+   how we handle the remaining bytes.  */
+L(duP1x):
+	cmpld	cr5, rWORD7, rWORD8
+	sldi.	rN, rN, 3
+	bne	cr5, L(duLcr5)
+	cmpld	cr7, rN, rSHR
+	beq	L(duZeroReturn)
+	li	r0, 0
+	ble	cr7, L(dutrim)
+	LD	rWORD2, rOFF8, rSTR2
+	srd	r0, rWORD2, rSHR
+	b	L(dutrim)
+/* Remainder is 16.  */
+	.align	4
+L(duP2):
+	srd	r0, rWORD8, rSHR
+	LD	rWORD5, 0, rSTR1
+	or	rWORD6, r0, rWORD6_SHIFT
+	sld	rWORD6_SHIFT, rWORD8, rSHL
+L(duP2e):
+	LD	rWORD7, rOFF8, rSTR1
+	LD	rWORD8, rOFF8, rSTR2
+	cmpld	cr6, rWORD5, rWORD6
+	srd	r12, rWORD8, rSHR
+	sld	rWORD8_SHIFT, rWORD8, rSHL
+	or	rWORD8, r12, rWORD6_SHIFT
+	blt	cr7, L(duP2x)
+	LD	rWORD1, rOFF16, rSTR1
+	LD	rWORD2, rOFF16, rSTR2
+	cmpld	cr5, rWORD7, rWORD8
+	bne	cr6, L(duLcr6)
+	srd	r0, rWORD2, rSHR
+	sld	rWORD2_SHIFT, rWORD2, rSHL
+	or	rWORD2, r0, rWORD8_SHIFT
+	LD	rWORD3, rOFF24, rSTR1
+	LD	rWORD4, rOFF24, rSTR2
+	cmpld	cr7, rWORD1, rWORD2
+	bne	cr5, L(duLcr5)
+	srd	r12, rWORD4, rSHR
+	sld	rWORD4_SHIFT, rWORD4, rSHL
+	or	rWORD4, r12, rWORD2_SHIFT
+	addi	rSTR1, rSTR1, 8
+	addi	rSTR2, rSTR2, 8
+	cmpld	cr1, rWORD3, rWORD4
+	b	L(duLoop2)
+	.align	4
+L(duP2x):
+	cmpld	cr5, rWORD7, rWORD8
+	addi	rSTR1, rSTR1, 8
+	addi	rSTR2, rSTR2, 8
+	bne	cr6, L(duLcr6)
+	sldi.	rN, rN, 3
+	bne	cr5, L(duLcr5)
+	cmpld	cr7, rN, rSHR
+	beq	L(duZeroReturn)
+	li	r0, 0
+	ble	cr7, L(dutrim)
+	LD	rWORD2, rOFF8, rSTR2
+	srd	r0, rWORD2, rSHR
+	b	L(dutrim)
+
+/* Remainder is 24.  */
+	.align	4
+L(duP3):
+	srd	r12, rWORD8, rSHR
+	LD	rWORD3, 0, rSTR1
+	sld	rWORD4_SHIFT, rWORD8, rSHL
+	or	rWORD4, r12, rWORD6_SHIFT
+L(duP3e):
+	LD	rWORD5, rOFF8, rSTR1
+	LD	rWORD6, rOFF8, rSTR2
+	cmpld	cr1, rWORD3, rWORD4
+	srd	r0, rWORD6, rSHR
+	sld	rWORD6_SHIFT, rWORD6, rSHL
+	or	rWORD6, r0, rWORD4_SHIFT
+	LD	rWORD7, rOFF16, rSTR1
+	LD	rWORD8, rOFF16, rSTR2
+	cmpld	cr6, rWORD5, rWORD6
+	bne	cr1, L(duLcr1)
+	srd	r12, rWORD8, rSHR
+	sld	rWORD8_SHIFT, rWORD8, rSHL
+	or	rWORD8, r12, rWORD6_SHIFT
+	blt	cr7, L(duP3x)
+	LD	rWORD1, rOFF24, rSTR1
+	LD	rWORD2, rOFF24, rSTR2
+	cmpld	cr5, rWORD7, rWORD8
+	bne	cr6, L(duLcr6)
+	srd	r0, rWORD2, rSHR
+	sld	rWORD2_SHIFT, rWORD2, rSHL
+	or	rWORD2, r0, rWORD8_SHIFT
+	addi	rSTR1, rSTR1, 16
+	addi	rSTR2, rSTR2, 16
+	cmpld	cr7, rWORD1, rWORD2
+	b	L(duLoop1)
+	.align	4
+L(duP3x):
+	addi	rSTR1, rSTR1, 16
+	addi	rSTR2, rSTR2, 16
+	cmpld	cr5, rWORD7, rWORD8
+	bne	cr6, L(duLcr6)
+	sldi.	rN, rN, 3
+	bne	cr5, L(duLcr5)
+	cmpld	cr7, rN, rSHR
+	beq	L(duZeroReturn)
+	li	r0, 0
+	ble	cr7, L(dutrim)
+	LD	rWORD2, rOFF8, rSTR2
+	srd	r0, rWORD2, rSHR
+	b	L(dutrim)
+
+/* Count is a multiple of 32, remainder is 0.  */
+	.align	4
+L(duP4):
+	mtctr	r0
+	srd	r0, rWORD8, rSHR
+	LD	rWORD1, 0, rSTR1
+	sld	rWORD2_SHIFT, rWORD8, rSHL
+	or	rWORD2, r0, rWORD6_SHIFT
+L(duP4e):
+	LD	rWORD3, rOFF8, rSTR1
+	LD	rWORD4, rOFF8, rSTR2
+	cmpld	cr7, rWORD1, rWORD2
+	srd	r12, rWORD4, rSHR
+	sld	rWORD4_SHIFT, rWORD4, rSHL
+	or	rWORD4, r12, rWORD2_SHIFT
+	LD	rWORD5, rOFF16, rSTR1
+	LD	rWORD6, rOFF16, rSTR2
+	cmpld	cr1, rWORD3, rWORD4
+	bne	cr7, L(duLcr7)
+	srd	r0, rWORD6, rSHR
+	sld	rWORD6_SHIFT, rWORD6, rSHL
+	or	rWORD6, r0, rWORD4_SHIFT
+	LD	rWORD7, rOFF24, rSTR1
+	LD	rWORD8, rOFF24, rSTR2
+	addi	rSTR1, rSTR1, 24
+	addi	rSTR2, rSTR2, 24
+	cmpld	cr6, rWORD5, rWORD6
+	bne	cr1, L(duLcr1)
+	srd	r12, rWORD8, rSHR
+	sld	rWORD8_SHIFT, rWORD8, rSHL
+	or	rWORD8, r12, rWORD6_SHIFT
+	cmpld	cr5, rWORD7, rWORD8
+	bdz	L(du24)		/* Adjust CTR as we start with +4.  */
+/* This is the primary loop.  */
+	.align	4
+L(duLoop):
+	LD	rWORD1, rOFF8, rSTR1
+	LD	rWORD2, rOFF8, rSTR2
+	cmpld	cr1, rWORD3, rWORD4
+	bne	cr6, L(duLcr6)
+	srd	r0, rWORD2, rSHR
+	sld	rWORD2_SHIFT, rWORD2, rSHL
+	or	rWORD2, r0, rWORD8_SHIFT
+L(duLoop1):
+	LD	rWORD3, rOFF16, rSTR1
+	LD	rWORD4, rOFF16, rSTR2
+	cmpld	cr6, rWORD5, rWORD6
+	bne	cr5, L(duLcr5)
+	srd	r12, rWORD4, rSHR
+	sld	rWORD4_SHIFT, rWORD4, rSHL
+	or	rWORD4, r12, rWORD2_SHIFT
+L(duLoop2):
+	LD	rWORD5, rOFF24, rSTR1
+	LD	rWORD6, rOFF24, rSTR2
+	cmpld	cr5, rWORD7, rWORD8
+	bne	cr7, L(duLcr7)
+	srd	r0, rWORD6, rSHR
+	sld	rWORD6_SHIFT, rWORD6, rSHL
+	or	rWORD6, r0, rWORD4_SHIFT
+L(duLoop3):
+	LD	rWORD7, rOFF32, rSTR1
+	LD	rWORD8, rOFF32, rSTR2
+	addi	rSTR1, rSTR1, 32
+	addi	rSTR2, rSTR2, 32
+	cmpld	cr7, rWORD1, rWORD2
+	bne	cr1, L(duLcr1)
+	srd	r12, rWORD8, rSHR
+	sld	rWORD8_SHIFT, rWORD8, rSHL
+	or	rWORD8, r12, rWORD6_SHIFT
+	bdnz	L(duLoop)
+
+L(duL4):
+	cmpld	cr1, rWORD3, rWORD4
+	bne	cr6, L(duLcr6)
+	cmpld	cr6, rWORD5, rWORD6
+	bne	cr5, L(duLcr5)
+	cmpld	cr5, rWORD7, rWORD8
+L(du44):
+	bne	cr7, L(duLcr7)
+L(du34):
+	bne	cr1, L(duLcr1)
+L(du24):
+	bne	cr6, L(duLcr6)
+L(du14):
+	sldi.	rN, rN, 3
+	bne	cr5, L(duLcr5)
+/* At this point we have a remainder of 1 to 7 bytes to compare.  We use
+   shift right double to eliminate bits beyond the compare length.
+
+   However it may not be safe to load rWORD2 which may be beyond the
+   string length.  So we compare the bit length of the remainder to
+   the right shift count (rSHR).  If the bit count is less than or equal
+   we do not need to load rWORD2 (all significant bits are already in
+   rWORD8_SHIFT).  */
+	cmpld	cr7, rN, rSHR
+	beq	L(duZeroReturn)
+	li	r0, 0
+	ble	cr7, L(dutrim)
+	LD	rWORD2, rOFF8, rSTR2
+	srd	r0, rWORD2, rSHR
+	.align	4
+L(dutrim):
+	LD	rWORD1, rOFF8, rSTR1
+	ld	rWORD8, -8(r1)
+	subfic	rN, rN, 64	/* Shift count is 64 - (rN * 8).  */
+	or	rWORD2, r0, rWORD8_SHIFT
+	ld	rWORD7, rWORD7SAVE(r1)
+	ld	rSHL, rSHLSAVE(r1)
+	srd	rWORD1, rWORD1, rN
+	srd	rWORD2, rWORD2, rN
+	ld	rSHR, rSHRSAVE(r1)
+	ld	rWORD8_SHIFT, rWORD8SHIFTSAVE(r1)
+	li	rRTN, 0
+	cmpld	cr7, rWORD1, rWORD2
+	ld	rWORD2_SHIFT, rWORD2SHIFTSAVE(r1)
+	ld	rWORD4_SHIFT, rWORD4SHIFTSAVE(r1)
+	beq	cr7, L(dureturn24)
+	li	rRTN, 1
+	ld	rWORD6_SHIFT, rWORD6SHIFTSAVE(r1)
+	ld	rOFF8,  rOFF8SAVE(r1)
+	ld	rOFF16, rOFF16SAVE(r1)
+	ld	rOFF24, rOFF24SAVE(r1)
+	ld	rOFF32, rOFF32SAVE(r1)
+	bgtlr	cr7
+	li	rRTN, -1
+	blr
+	.align	4
+L(duLcr7):
+	ld	rWORD8, rWORD8SAVE(r1)
+	ld	rWORD7, rWORD7SAVE(r1)
+	li	rRTN, 1
+	bgt	cr7, L(dureturn29)
+	ld	rSHL, rSHLSAVE(r1)
+	ld	rSHR, rSHRSAVE(r1)
+	li	rRTN, -1
+	b	L(dureturn27)
+	.align	4
+L(duLcr1):
+	ld	rWORD8, rWORD8SAVE(r1)
+	ld	rWORD7, rWORD7SAVE(r1)
+	li	rRTN, 1
+	bgt	cr1, L(dureturn29)
+	ld	rSHL, rSHLSAVE(r1)
+	ld	rSHR, rSHRSAVE(r1)
+	li	rRTN, -1
+	b	L(dureturn27)
+	.align	4
+L(duLcr6):
+	ld	rWORD8, rWORD8SAVE(r1)
+	ld	rWORD7, rWORD7SAVE(r1)
+	li	rRTN, 1
+	bgt	cr6, L(dureturn29)
+	ld	rSHL, rSHLSAVE(r1)
+	ld	rSHR, rSHRSAVE(r1)
+	li	rRTN, -1
+	b	L(dureturn27)
+	.align	4
+L(duLcr5):
+	ld	rWORD8, rWORD8SAVE(r1)
+	ld	rWORD7, rWORD7SAVE(r1)
+	li	rRTN, 1
+	bgt	cr5, L(dureturn29)
+	ld	rSHL, rSHLSAVE(r1)
+	ld	rSHR, rSHRSAVE(r1)
+	li	rRTN, -1
+	b	L(dureturn27)
+
+	.align	3
+L(duZeroReturn):
+	li	rRTN, 0
+	.align	4
+L(dureturn):
+	ld	rWORD8, rWORD8SAVE(r1)
+	ld	rWORD7, rWORD7SAVE(r1)
+L(dureturn29):
+	ld	rSHL, rSHLSAVE(r1)
+	ld	rSHR, rSHRSAVE(r1)
+L(dureturn27):
+	ld	rWORD8_SHIFT, rWORD8SHIFTSAVE(r1)
+	ld	rWORD2_SHIFT, rWORD2SHIFTSAVE(r1)
+	ld	rWORD4_SHIFT, rWORD4SHIFTSAVE(r1)
+L(dureturn24):
+	ld	rWORD6_SHIFT, rWORD6SHIFTSAVE(r1)
+	ld	rOFF8,  rOFF8SAVE(r1)
+	ld	rOFF16, rOFF16SAVE(r1)
+	ld	rOFF24, rOFF24SAVE(r1)
+	ld	rOFF32, rOFF32SAVE(r1)
+	blr
+
+L(duzeroLength):
+	ld	rOFF8,  rOFF8SAVE(r1)
+	ld	rOFF16, rOFF16SAVE(r1)
+	ld	rOFF24, rOFF24SAVE(r1)
+	ld	rOFF32, rOFF32SAVE(r1)
+	li	rRTN, 0
+	blr
+
+END (memcmp)
+libc_hidden_builtin_def (memcmp)
+weak_alias (memcmp, bcmp)