This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.
Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]
[PING][PATCH neleai/string-x64] Improve strcpy sse2 and avx2 implementation

From: OndÅej BÃlka <neleai at seznam dot cz>
To: libc-alpha at sourceware dot org
Date: Wed, 24 Jun 2015 10:13:31 +0200
Subject: [PING][PATCH neleai/string-x64] Improve strcpy sse2 and avx2 implementation
Authentication-results: sourceware.org; auth=none
References: <20150617180105 dot GA26497 at domone>
On Wed, Jun 17, 2015 at 08:01:05PM +0200, OndÅej BÃlka wrote:
> Hi,
> 
> I wrote new strcpy on x64 and for some reason I thought that I had
> commited it and forgot to ping it.
> 
> As there are other routines that I could improve I will use branch
> neleai/string-x64 to collect these.
> 
> Here is revised version of what I sumbitted in 2013. Main change is that
> I now target i7 instead core2 That simplifies things as unaligned loads
> are cheap instead bit slower than aligned ones on core2. That mainly
> concerns header as for core2 you could get better performance by
> aligning loads or stores to 16 bytes after first bytes were read. I do
> not know whats better I would need to test it.
> 
> That also makes less important support of ssse3 variant. I could send it
> but it was one of my list on TODO list that now probably lost
> importance. Problem is that on x64 for aligning by ssse3 or sse2 with
> shifts you need to make 16 loops for each alignment as you don't have
> variable shift. Also it needs to use jump table thats very expensive
> For strcpy thats dubious as it increases instruction cache pressure 
> and most copies are small. You would need to do switching from unaligned 
> loads to aligning. I needed to do profiling to select correct treshold.
> 
> If somebody is interested in optimizing old pentiums4, athlon64 I will
> provide a ssse3 variant that is also 50% faster than current one.
> That is also reason why I omitted drawing current ssse3 implementation
> performance.
> 
> 
> In this version header first checks 128 bytes unaligned unless they
> cross page boundary. That allows more effective loop as then at end of
> loop we could simply write last 64 bytes instead specialcasing to avoid
> writing before start.
> 
> I tried several variants of header, as we first read 16 bytes to xmm0
> register question is if they could be reused. I used evolver to select
> best variant, there was almost no difference in performance between
> these.
> 
> Now I do checks for bytes 0-15, then 16-31, then 32-63, then 64-128.
> There is possibility to get some cycles with different grouping, I will
> post later improvement if I could find something. 
> 
> 
> First problem was reading ahead. A rereading 8 bytes looked bit faster
> than move from xmm.
> 
> Then I tried when to reuse/reread. In 4-7 byte case it was faster reread
> than using bit shifts to get second half. For 1-3 bytes I use following
> copy with s[0] and s[1] from rdx register with byte shifts.
> 
>   Test branch vs this branchless that works for i 0,1,2
>    d[i] = 0;
>    d[i/2] = s[1];
>    d[0] = s[0];
> 
> I also added a avx2 loop. Problem why I shouldn't use them in headers
> was high latency. I could test if using them for bytes 64-128 would give
> speedup.
> 
> As technical issues go I needed to move old strcpy_sse_unaligned
> implementation into strncpy_sse2_unaligned as strncpy is function that
> should be optimized for size, not performance. For now I this will keep
> these unchanged.
> 
> As performance these are 15%-30% faster than current one for gcc workload on
> haswell and ivy bridge. 
> 
> As avx2 version its currently 6% on this workload mainly as its bash and
> has lot of large loads so avx2 loop helps.
> 
> I used my profiler to show improvement, see here
> 
> http://kam.mff.cuni.cz/~ondra/benchmark_string/strcpy_profile.html
> 
> and source is here
> 
> http://kam.mff.cuni.cz/~ondra/benchmark_string/strcpy_profile170615.tar.bz2
> 
> Comments?
> 
>         * sysdeps/x86_64/multiarch/ifunc-impl-list.c (__libc_ifunc_impl_list):
> 	Add __strcpy_avx2 and __stpcpy_avx2
>         * sysdeps/x86_64/multiarch/Makefile (routines): Add stpcpy_avx2.S and 
> 	strcpy_avx2.S
>         * sysdeps/x86_64/multiarch/stpcpy-avx2.S: New file
>         * sysdeps/x86_64/multiarch/strcpy-avx2.S: Likewise.
>         * sysdeps/x86_64/multiarch/stpcpy-sse2-unaligned.S: Refactored
> 	implementation.
>         * sysdeps/x86_64/multiarch/strcpy.S: Updated ifunc.
>         * sysdeps/x86_64/multiarch/strncpy.S: Moved from strcpy.S.
>         * sysdeps/x86_64/multiarch/strncpy-sse2-unaligned.S: Moved
> 	strcpy-sse2-unaligned.S here.
>         * sysdeps/x86_64/multiarch/strcpy-sse2-unaligned.S: Likewise.
>         * sysdeps/x86_64/multiarch/stpncpy-sse2-unaligned.S: Redirect
> 	from strcpy-sse2-unaligned.S to strncpy-sse2-unaligned.S 
>         * sysdeps/x86_64/multiarch/stpncpy.S: Likewise.
>         * sysdeps/x86_64/multiarch/strcat-sse2-unaligned.S: Likewise.
> 
> ---
>  sysdeps/x86_64/multiarch/Makefile                 |    2 +-
>  sysdeps/x86_64/multiarch/ifunc-impl-list.c        |    2 +
>  sysdeps/x86_64/multiarch/stpcpy-avx2.S            |    3 +
>  sysdeps/x86_64/multiarch/stpcpy-sse2-unaligned.S  |  439 ++++-
>  sysdeps/x86_64/multiarch/stpncpy-sse2-unaligned.S |    3 +-
>  sysdeps/x86_64/multiarch/stpncpy.S                |    5 +-
>  sysdeps/x86_64/multiarch/strcat-sse2-unaligned.S  |    2 +-
>  sysdeps/x86_64/multiarch/strcpy-avx2.S            |    4 +
>  sysdeps/x86_64/multiarch/strcpy-sse2-unaligned.S  | 1890 +-------------------
>  sysdeps/x86_64/multiarch/strcpy.S                 |   22 +-
>  sysdeps/x86_64/multiarch/strncpy-sse2-unaligned.S | 1891 ++++++++++++++++++++-
>  sysdeps/x86_64/multiarch/strncpy.S                |   88 +-
>  14 files changed, 2435 insertions(+), 1921 deletions(-)
>  create mode 100644 sysdeps/x86_64/multiarch/stpcpy-avx2.S
>  create mode 100644 sysdeps/x86_64/multiarch/strcpy-avx2.S
> 
> 
> diff --git a/sysdeps/x86_64/multiarch/Makefile b/sysdeps/x86_64/multiarch/Makefile
> index d7002a9..c573744 100644
> --- a/sysdeps/x86_64/multiarch/Makefile
> +++ b/sysdeps/x86_64/multiarch/Makefile
> @@ -29,7 +29,7 @@ CFLAGS-strspn-c.c += -msse4
>  endif
>  
>  ifeq (yes,$(config-cflags-avx2))
> -sysdep_routines += memset-avx2
> +sysdep_routines += memset-avx2 strcpy-avx2 stpcpy-avx2
>  endif
>  endif
>  
> diff --git a/sysdeps/x86_64/multiarch/ifunc-impl-list.c b/sysdeps/x86_64/multiarch/ifunc-impl-list.c
> index b64e4f1..d398e43 100644
> --- a/sysdeps/x86_64/multiarch/ifunc-impl-list.c
> +++ b/sysdeps/x86_64/multiarch/ifunc-impl-list.c
> @@ -88,6 +88,7 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array,
>  
>    /* Support sysdeps/x86_64/multiarch/stpcpy.S.  */
>    IFUNC_IMPL (i, name, stpcpy,
> +	      IFUNC_IMPL_ADD (array, i, strcpy, HAS_AVX2, __stpcpy_avx2)
>  	      IFUNC_IMPL_ADD (array, i, stpcpy, HAS_SSSE3, __stpcpy_ssse3)
>  	      IFUNC_IMPL_ADD (array, i, stpcpy, 1, __stpcpy_sse2_unaligned)
>  	      IFUNC_IMPL_ADD (array, i, stpcpy, 1, __stpcpy_sse2))
> @@ -137,6 +138,7 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array,
>  
>    /* Support sysdeps/x86_64/multiarch/strcpy.S.  */
>    IFUNC_IMPL (i, name, strcpy,
> +	      IFUNC_IMPL_ADD (array, i, strcpy, HAS_AVX2, __strcpy_avx2)
>  	      IFUNC_IMPL_ADD (array, i, strcpy, HAS_SSSE3, __strcpy_ssse3)
>  	      IFUNC_IMPL_ADD (array, i, strcpy, 1, __strcpy_sse2_unaligned)
>  	      IFUNC_IMPL_ADD (array, i, strcpy, 1, __strcpy_sse2))
> diff --git a/sysdeps/x86_64/multiarch/stpcpy-avx2.S b/sysdeps/x86_64/multiarch/stpcpy-avx2.S
> new file mode 100644
> index 0000000..bd30ef6
> --- /dev/null
> +++ b/sysdeps/x86_64/multiarch/stpcpy-avx2.S
> @@ -0,0 +1,3 @@
> +#define USE_AVX2
> +#define STPCPY __stpcpy_avx2
> +#include "stpcpy-sse2-unaligned.S"
> diff --git a/sysdeps/x86_64/multiarch/stpcpy-sse2-unaligned.S b/sysdeps/x86_64/multiarch/stpcpy-sse2-unaligned.S
> index 34231f8..695a236 100644
> --- a/sysdeps/x86_64/multiarch/stpcpy-sse2-unaligned.S
> +++ b/sysdeps/x86_64/multiarch/stpcpy-sse2-unaligned.S
> @@ -1,3 +1,436 @@
> -#define USE_AS_STPCPY
> -#define STRCPY __stpcpy_sse2_unaligned
> -#include "strcpy-sse2-unaligned.S"
> +/* stpcpy with SSE2 and unaligned load
> +   Copyright (C) 2015 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>
> +#ifndef STPCPY
> +# define STPCPY __stpcpy_sse2_unaligned
> +#endif
> +
> +ENTRY(STPCPY)
> +	mov	%esi, %edx
> +#ifdef AS_STRCPY
> +	movq    %rdi, %rax
> +#endif
> +	pxor	%xmm4, %xmm4
> +	pxor	%xmm5, %xmm5
> +	andl	$4095, %edx
> +	cmp	$3968, %edx
> +	ja	L(cross_page)
> +
> +	movdqu	(%rsi), %xmm0
> +	pcmpeqb	%xmm0, %xmm4
> +	pmovmskb %xmm4, %edx
> +	testl	%edx, %edx
> +	je	L(more16bytes)
> +	bsf	%edx, %ecx
> +#ifndef AS_STRCPY
> +	lea	(%rdi, %rcx), %rax
> +#endif
> +	cmp	$7, %ecx
> +	movq	(%rsi), %rdx
> +	jb	L(less_8_bytesb)
> +L(8bytes_from_cross):
> +	movq	-7(%rsi, %rcx), %rsi
> +	movq	%rdx, (%rdi)
> +#ifdef AS_STRCPY
> +	movq    %rsi, -7(%rdi, %rcx)
> +#else
> +	movq	%rsi, -7(%rax)
> +#endif
> +	ret
> +
> +	.p2align 4
> +L(less_8_bytesb):
> +	cmp	$2, %ecx
> +	jbe	L(less_4_bytes)
> +L(4bytes_from_cross):
> +	mov	-3(%rsi, %rcx), %esi
> +	mov	%edx, (%rdi)
> +#ifdef AS_STRCPY
> +        mov     %esi, -3(%rdi, %rcx)
> +#else
> +	mov	%esi, -3(%rax)
> +#endif
> +	ret
> +
> +.p2align 4
> + L(less_4_bytes):
> + /*
> +  Test branch vs this branchless that works for i 0,1,2
> +   d[i] = 0;
> +   d[i/2] = s[1];
> +   d[0] = s[0];
> +  */
> +#ifdef AS_STRCPY
> +	movb	$0, (%rdi, %rcx)
> +#endif
> +
> +	shr	$1, %ecx
> +	mov	%edx, %esi
> +	shr	$8, %edx
> +	movb	%dl, (%rdi, %rcx)
> +#ifndef AS_STRCPY
> +	movb	$0, (%rax)
> +#endif
> +	movb	%sil, (%rdi)
> +	ret
> +
> +
> +
> +
> +
> +	.p2align 4
> +L(more16bytes):
> +	pxor	%xmm6, %xmm6
> +	movdqu	16(%rsi), %xmm1
> +	pxor	%xmm7, %xmm7
> +	pcmpeqb	%xmm1, %xmm5
> +	pmovmskb %xmm5, %edx
> +	testl	%edx, %edx
> +	je	L(more32bytes)
> +	bsf	%edx, %edx
> +#ifdef AS_STRCPY
> +        movdqu  1(%rsi, %rdx), %xmm1
> +        movdqu  %xmm0, (%rdi)
> +	movdqu  %xmm1, 1(%rdi, %rdx)
> +#else
> +	lea	16(%rdi, %rdx), %rax
> +	movdqu	1(%rsi, %rdx), %xmm1
> +	movdqu	%xmm0, (%rdi)
> +	movdqu	%xmm1, -15(%rax)
> +#endif
> +	ret
> +
> +	.p2align 4
> +L(more32bytes):
> +	movdqu	32(%rsi), %xmm2
> +	movdqu	48(%rsi), %xmm3
> +
> +	pcmpeqb	%xmm2, %xmm6
> +	pcmpeqb	%xmm3, %xmm7
> +	pmovmskb %xmm7, %edx
> +	shl	$16, %edx
> +	pmovmskb %xmm6, %ecx
> +	or	%ecx, %edx
> +	je	L(more64bytes)
> +	bsf	%edx, %edx
> +#ifndef AS_STRCPY
> +	lea	32(%rdi, %rdx), %rax
> +#endif
> +	movdqu	1(%rsi, %rdx), %xmm2
> +	movdqu	17(%rsi, %rdx), %xmm3
> +	movdqu	%xmm0, (%rdi)
> +	movdqu	%xmm1, 16(%rdi)
> +#ifdef AS_STRCPY
> +        movdqu  %xmm2, 1(%rdi, %rdx)
> +        movdqu  %xmm3, 17(%rdi, %rdx)
> +#else
> +	movdqu	%xmm2, -31(%rax)
> +	movdqu	%xmm3, -15(%rax)
> +#endif
> +	ret
> +
> +	.p2align 4
> +L(more64bytes):
> +	movdqu	%xmm0, (%rdi)
> +	movdqu	%xmm1, 16(%rdi)
> +	movdqu	%xmm2, 32(%rdi)
> +	movdqu	%xmm3, 48(%rdi)
> +	movdqu	64(%rsi), %xmm0
> +	movdqu	80(%rsi), %xmm1
> +	movdqu	96(%rsi), %xmm2
> +	movdqu	112(%rsi), %xmm3
> +
> +	pcmpeqb	%xmm0, %xmm4
> +	pcmpeqb	%xmm1, %xmm5
> +	pcmpeqb	%xmm2, %xmm6
> +	pcmpeqb	%xmm3, %xmm7
> +	pmovmskb %xmm4, %ecx
> +	pmovmskb %xmm5, %edx
> +	pmovmskb %xmm6, %r8d
> +	pmovmskb %xmm7, %r9d
> +	shl	$16, %edx
> +	or	%ecx, %edx
> +	shl	$32, %r8
> +	shl	$48, %r9
> +	or	%r8, %rdx
> +	or	%r9, %rdx
> +	test	%rdx, %rdx
> +	je	L(prepare_loop)
> +	bsf	%rdx, %rdx
> +#ifndef AS_STRCPY
> +	lea	64(%rdi, %rdx), %rax
> +#endif
> +	movdqu	1(%rsi, %rdx), %xmm0
> +	movdqu	17(%rsi, %rdx), %xmm1
> +	movdqu	33(%rsi, %rdx), %xmm2
> +	movdqu	49(%rsi, %rdx), %xmm3
> +#ifdef AS_STRCPY
> +        movdqu  %xmm0, 1(%rdi, %rdx)
> +        movdqu  %xmm1, 17(%rdi, %rdx)
> +        movdqu  %xmm2, 33(%rdi, %rdx)
> +        movdqu  %xmm3, 49(%rdi, %rdx)
> +#else
> +	movdqu	%xmm0, -63(%rax)
> +	movdqu	%xmm1, -47(%rax)
> +	movdqu	%xmm2, -31(%rax)
> +	movdqu	%xmm3, -15(%rax)
> +#endif
> +	ret
> +
> +
> +	.p2align 4
> +L(prepare_loop):
> +	movdqu	%xmm0, 64(%rdi)
> +	movdqu	%xmm1, 80(%rdi)
> +	movdqu	%xmm2, 96(%rdi)
> +	movdqu	%xmm3, 112(%rdi)
> +
> +	subq	%rsi, %rdi
> +	add	$64, %rsi
> +	andq	$-64, %rsi
> +	addq	%rsi, %rdi
> +	jmp	L(loop_entry)
> +
> +#ifdef USE_AVX2
> +	.p2align 4
> +L(loop):
> +	vmovdqu	%ymm1, (%rdi)
> +	vmovdqu	%ymm3, 32(%rdi)
> +L(loop_entry):
> +	vmovdqa	96(%rsi), %ymm3
> +	vmovdqa	64(%rsi), %ymm1
> +	vpminub	%ymm3, %ymm1, %ymm2
> +	addq	$64, %rsi
> +	addq	$64, %rdi
> +	vpcmpeqb %ymm5, %ymm2, %ymm0
> +	vpmovmskb %ymm0, %edx
> +	test	%edx, %edx
> +	je	L(loop)
> +	salq	$32, %rdx
> +	vpcmpeqb %ymm5, %ymm1, %ymm4
> +	vpmovmskb %ymm4, %ecx
> +	or	%rcx, %rdx
> +	bsfq	%rdx, %rdx
> +#ifndef AS_STRCPY
> +	lea	(%rdi, %rdx), %rax
> +#endif
> +	vmovdqu	-63(%rsi, %rdx), %ymm0
> +	vmovdqu	-31(%rsi, %rdx), %ymm2
> +#ifdef AS_STRCPY
> +        vmovdqu  %ymm0, -63(%rdi, %rdx)
> +        vmovdqu  %ymm2, -31(%rdi, %rdx)
> +#else
> +	vmovdqu	%ymm0, -63(%rax)
> +	vmovdqu	%ymm2, -31(%rax)
> +#endif
> +	vzeroupper
> +	ret
> +#else
> +	.p2align 4
> +L(loop):
> +	movdqu	%xmm1, (%rdi)
> +	movdqu	%xmm2, 16(%rdi)
> +	movdqu	%xmm3, 32(%rdi)
> +	movdqu	%xmm4, 48(%rdi)
> +L(loop_entry):
> +	movdqa	96(%rsi), %xmm3
> +	movdqa	112(%rsi), %xmm4
> +	movdqa	%xmm3, %xmm0
> +	movdqa	80(%rsi), %xmm2
> +	pminub	%xmm4, %xmm0
> +	movdqa	64(%rsi), %xmm1
> +	pminub	%xmm2, %xmm0
> +	pminub	%xmm1, %xmm0
> +	addq	$64, %rsi
> +	addq	$64, %rdi
> +	pcmpeqb	%xmm5, %xmm0
> +	pmovmskb %xmm0, %edx
> +	test	%edx, %edx
> +	je	L(loop)
> +	salq	$48, %rdx
> +	pcmpeqb	%xmm1, %xmm5
> +	pcmpeqb	%xmm2, %xmm6
> +	pmovmskb %xmm5, %ecx
> +#ifdef AS_STRCPY
> +	pmovmskb %xmm6, %r8d
> +	pcmpeqb	%xmm3, %xmm7
> +	pmovmskb %xmm7, %r9d
> +	sal	$16, %r8d
> +	or	%r8d, %ecx
> +#else
> +	pmovmskb %xmm6, %eax
> +	pcmpeqb	%xmm3, %xmm7
> +	pmovmskb %xmm7, %r9d
> +	sal	$16, %eax
> +	or	%eax, %ecx
> +#endif
> +	salq	$32, %r9
> +	orq	%rcx, %rdx
> +	orq	%r9, %rdx
> +	bsfq	%rdx, %rdx
> +#ifndef AS_STRCPY
> +	lea	(%rdi, %rdx), %rax
> +#endif
> +	movdqu	-63(%rsi, %rdx), %xmm0
> +	movdqu	-47(%rsi, %rdx), %xmm1
> +	movdqu	-31(%rsi, %rdx), %xmm2
> +	movdqu	-15(%rsi, %rdx), %xmm3
> +#ifdef AS_STRCPY
> +        movdqu  %xmm0, -63(%rdi, %rdx)
> +        movdqu  %xmm1, -47(%rdi, %rdx)
> +        movdqu  %xmm2, -31(%rdi, %rdx)
> +        movdqu  %xmm3, -15(%rdi, %rdx)
> +#else
> +	movdqu	%xmm0, -63(%rax)
> +	movdqu	%xmm1, -47(%rax)
> +	movdqu	%xmm2, -31(%rax)
> +	movdqu	%xmm3, -15(%rax)
> +#endif
> +	ret
> +#endif
> +
> +	.p2align 4
> +L(cross_page):
> +	movq	%rsi, %rcx
> +	pxor	%xmm0, %xmm0
> +	and	$15, %ecx
> +	movq	%rsi, %r9
> +	movq	%rdi, %r10
> +	subq	%rcx, %rsi
> +	subq	%rcx, %rdi
> +	movdqa	(%rsi), %xmm1
> +	pcmpeqb	%xmm0, %xmm1
> +	pmovmskb %xmm1, %edx
> +	shr	%cl, %edx
> +	shl	%cl, %edx
> +	test	%edx, %edx
> +	jne	L(less_32_cross)
> +
> +	addq	$16, %rsi
> +	addq	$16, %rdi
> +	movdqa	(%rsi), %xmm1
> +	pcmpeqb	%xmm1, %xmm0
> +	pmovmskb %xmm0, %edx
> +	test	%edx, %edx
> +	jne	L(less_32_cross)
> +	movdqu	%xmm1, (%rdi)
> +
> +	movdqu	(%r9), %xmm0
> +	movdqu	%xmm0, (%r10)
> +
> +	mov	$8, %rcx
> +L(cross_loop):
> +	addq	$16, %rsi
> +	addq	$16, %rdi
> +	pxor	%xmm0, %xmm0
> +	movdqa	(%rsi), %xmm1
> +	pcmpeqb	%xmm1, %xmm0
> +	pmovmskb %xmm0, %edx
> +	test	%edx, %edx
> +	jne	L(return_cross)
> +	movdqu	%xmm1, (%rdi)
> +	sub	$1, %rcx
> +	ja	L(cross_loop)
> +
> +	pxor	%xmm5, %xmm5
> +	pxor	%xmm6, %xmm6
> +	pxor	%xmm7, %xmm7
> +
> +	lea	-64(%rsi), %rdx
> +	andq	$-64, %rdx
> +	addq	%rdx, %rdi
> +	subq	%rsi, %rdi
> +	movq	%rdx, %rsi
> +	jmp	L(loop_entry)
> +
> +	.p2align 4
> +L(return_cross):
> +	bsf	%edx, %edx
> +#ifdef AS_STRCPY
> +        movdqu  -15(%rsi, %rdx), %xmm0
> +        movdqu  %xmm0, -15(%rdi, %rdx)
> +#else
> +	lea	(%rdi, %rdx), %rax
> +	movdqu	-15(%rsi, %rdx), %xmm0
> +	movdqu	%xmm0, -15(%rax)
> +#endif
> +	ret
> +
> +	.p2align 4
> +L(less_32_cross):
> +	bsf	%rdx, %rdx
> +	lea	(%rdi, %rdx), %rcx
> +#ifndef AS_STRCPY
> +	mov	%rcx, %rax
> +#endif
> +	mov	%r9, %rsi
> +	mov	%r10, %rdi
> +	sub	%rdi, %rcx
> +	cmp	$15, %ecx
> +	jb	L(less_16_cross)
> +	movdqu	(%rsi), %xmm0
> +	movdqu	-15(%rsi, %rcx), %xmm1
> +	movdqu	%xmm0, (%rdi)
> +#ifdef AS_STRCPY
> +	movdqu  %xmm1, -15(%rdi, %rcx)
> +#else
> +	movdqu	%xmm1, -15(%rax)
> +#endif
> +	ret
> +
> +L(less_16_cross):
> +	cmp	$7, %ecx
> +	jb	L(less_8_bytes_cross)
> +	movq	(%rsi), %rdx
> +	jmp	L(8bytes_from_cross)
> +
> +L(less_8_bytes_cross):
> +	cmp	$2, %ecx
> +	jbe	L(3_bytes_cross)
> +	mov	(%rsi), %edx
> +	jmp	L(4bytes_from_cross)
> +
> +L(3_bytes_cross):
> +	jb	L(1_2bytes_cross)
> +	movzwl	(%rsi), %edx
> +	jmp	L(_3_bytesb)
> +
> +L(1_2bytes_cross):
> +	movb	(%rsi), %dl
> +	jmp	L(0_2bytes_from_cross)
> +
> +	.p2align 4
> +L(less_4_bytesb):
> +	je	L(_3_bytesb)
> +L(0_2bytes_from_cross):
> +	movb	%dl, (%rdi)
> +#ifdef AS_STRCPY
> +	movb    $0, (%rdi, %rcx)
> +#else
> +	movb	$0, (%rax)
> +#endif
> +	ret
> +
> +	.p2align 4
> +L(_3_bytesb):
> +	movw	%dx, (%rdi)
> +	movb	$0, 2(%rdi)
> +	ret
> +
> +END(STPCPY)
> diff --git a/sysdeps/x86_64/multiarch/stpncpy-sse2-unaligned.S b/sysdeps/x86_64/multiarch/stpncpy-sse2-unaligned.S
> index 658520f..3f35068 100644
> --- a/sysdeps/x86_64/multiarch/stpncpy-sse2-unaligned.S
> +++ b/sysdeps/x86_64/multiarch/stpncpy-sse2-unaligned.S
> @@ -1,4 +1,3 @@
>  #define USE_AS_STPCPY
> -#define USE_AS_STRNCPY
>  #define STRCPY __stpncpy_sse2_unaligned
> -#include "strcpy-sse2-unaligned.S"
> +#include "strncpy-sse2-unaligned.S"
> diff --git a/sysdeps/x86_64/multiarch/stpncpy.S b/sysdeps/x86_64/multiarch/stpncpy.S
> index 2698ca6..159604a 100644
> --- a/sysdeps/x86_64/multiarch/stpncpy.S
> +++ b/sysdeps/x86_64/multiarch/stpncpy.S
> @@ -1,8 +1,7 @@
>  /* Multiple versions of stpncpy
>     All versions must be listed in ifunc-impl-list.c.  */
> -#define STRCPY __stpncpy
> +#define STRNCPY __stpncpy
>  #define USE_AS_STPCPY
> -#define USE_AS_STRNCPY
> -#include "strcpy.S"
> +#include "strncpy.S"
>  
>  weak_alias (__stpncpy, stpncpy)
> diff --git a/sysdeps/x86_64/multiarch/strcat-sse2-unaligned.S b/sysdeps/x86_64/multiarch/strcat-sse2-unaligned.S
> index 81f1b40..1faa49d 100644
> --- a/sysdeps/x86_64/multiarch/strcat-sse2-unaligned.S
> +++ b/sysdeps/x86_64/multiarch/strcat-sse2-unaligned.S
> @@ -275,5 +275,5 @@ L(StartStrcpyPart):
>  #  define USE_AS_STRNCPY
>  # endif
>  
> -# include "strcpy-sse2-unaligned.S"
> +# include "strncpy-sse2-unaligned.S"
>  #endif
> diff --git a/sysdeps/x86_64/multiarch/strcpy-avx2.S b/sysdeps/x86_64/multiarch/strcpy-avx2.S
> new file mode 100644
> index 0000000..a3133a4
> --- /dev/null
> +++ b/sysdeps/x86_64/multiarch/strcpy-avx2.S
> @@ -0,0 +1,4 @@
> +#define USE_AVX2
> +#define AS_STRCPY
> +#define STPCPY __strcpy_avx2
> +#include "stpcpy-sse2-unaligned.S"
> diff --git a/sysdeps/x86_64/multiarch/strcpy-sse2-unaligned.S b/sysdeps/x86_64/multiarch/strcpy-sse2-unaligned.S
> index 8f03d1d..310e4fa 100644
> --- a/sysdeps/x86_64/multiarch/strcpy-sse2-unaligned.S
> +++ b/sysdeps/x86_64/multiarch/strcpy-sse2-unaligned.S
> @@ -1,1887 +1,3 @@
> -/* strcpy with SSE2 and unaligned load
> -   Copyright (C) 2011-2015 Free Software Foundation, Inc.
> -   Contributed by Intel Corporation.
> -   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/>.  */
> -
> -#if IS_IN (libc)
> -
> -# ifndef USE_AS_STRCAT
> -#  include <sysdep.h>
> -
> -#  ifndef STRCPY
> -#   define STRCPY  __strcpy_sse2_unaligned
> -#  endif
> -
> -# endif
> -
> -# define JMPTBL(I, B)	I - B
> -# define BRANCH_TO_JMPTBL_ENTRY(TABLE, INDEX, SCALE)             \
> -	lea	TABLE(%rip), %r11;                              \
> -	movslq	(%r11, INDEX, SCALE), %rcx;                     \
> -	lea	(%r11, %rcx), %rcx;                             \
> -	jmp	*%rcx
> -
> -# ifndef USE_AS_STRCAT
> -
> -.text
> -ENTRY (STRCPY)
> -#  ifdef USE_AS_STRNCPY
> -	mov	%rdx, %r8
> -	test	%r8, %r8
> -	jz	L(ExitZero)
> -#  endif
> -	mov	%rsi, %rcx
> -#  ifndef USE_AS_STPCPY
> -	mov	%rdi, %rax      /* save result */
> -#  endif
> -
> -# endif
> -
> -	and	$63, %rcx
> -	cmp	$32, %rcx
> -	jbe	L(SourceStringAlignmentLess32)
> -
> -	and	$-16, %rsi
> -	and	$15, %rcx
> -	pxor	%xmm0, %xmm0
> -	pxor	%xmm1, %xmm1
> -
> -	pcmpeqb	(%rsi), %xmm1
> -	pmovmskb %xmm1, %rdx
> -	shr	%cl, %rdx
> -
> -# ifdef USE_AS_STRNCPY
> -#  if defined USE_AS_STPCPY || defined USE_AS_STRCAT
> -	mov	$16, %r10
> -	sub	%rcx, %r10
> -	cmp	%r10, %r8
> -#  else
> -	mov	$17, %r10
> -	sub	%rcx, %r10
> -	cmp	%r10, %r8
> -#  endif
> -	jbe	L(CopyFrom1To16BytesTailCase2OrCase3)
> -# endif
> -	test	%rdx, %rdx
> -	jnz	L(CopyFrom1To16BytesTail)
> -
> -	pcmpeqb	16(%rsi), %xmm0
> -	pmovmskb %xmm0, %rdx
> -
> -# ifdef USE_AS_STRNCPY
> -	add	$16, %r10
> -	cmp	%r10, %r8
> -	jbe	L(CopyFrom1To32BytesCase2OrCase3)
> -# endif
> -	test	%rdx, %rdx
> -	jnz	L(CopyFrom1To32Bytes)
> -
> -	movdqu	(%rsi, %rcx), %xmm1   /* copy 16 bytes */
> -	movdqu	%xmm1, (%rdi)
> -
> -/* If source address alignment != destination address alignment */
> -	.p2align 4
> -L(Unalign16Both):
> -	sub	%rcx, %rdi
> -# ifdef USE_AS_STRNCPY
> -	add	%rcx, %r8
> -# endif
> -	mov	$16, %rcx
> -	movdqa	(%rsi, %rcx), %xmm1
> -	movaps	16(%rsi, %rcx), %xmm2
> -	movdqu	%xmm1, (%rdi, %rcx)
> -	pcmpeqb	%xmm2, %xmm0
> -	pmovmskb %xmm0, %rdx
> -	add	$16, %rcx
> -# ifdef USE_AS_STRNCPY
> -	sub	$48, %r8
> -	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> -# endif
> -	test	%rdx, %rdx
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	jnz	L(CopyFrom1To16BytesUnalignedXmm2)
> -# else
> -	jnz	L(CopyFrom1To16Bytes)
> -# endif
> -
> -	movaps	16(%rsi, %rcx), %xmm3
> -	movdqu	%xmm2, (%rdi, %rcx)
> -	pcmpeqb	%xmm3, %xmm0
> -	pmovmskb %xmm0, %rdx
> -	add	$16, %rcx
> -# ifdef USE_AS_STRNCPY
> -	sub	$16, %r8
> -	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> -# endif
> -	test	%rdx, %rdx
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	jnz	L(CopyFrom1To16BytesUnalignedXmm3)
> -# else
> -	jnz	L(CopyFrom1To16Bytes)
> -# endif
> -
> -	movaps	16(%rsi, %rcx), %xmm4
> -	movdqu	%xmm3, (%rdi, %rcx)
> -	pcmpeqb	%xmm4, %xmm0
> -	pmovmskb %xmm0, %rdx
> -	add	$16, %rcx
> -# ifdef USE_AS_STRNCPY
> -	sub	$16, %r8
> -	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> -# endif
> -	test	%rdx, %rdx
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	jnz	L(CopyFrom1To16BytesUnalignedXmm4)
> -# else
> -	jnz	L(CopyFrom1To16Bytes)
> -# endif
> -
> -	movaps	16(%rsi, %rcx), %xmm1
> -	movdqu	%xmm4, (%rdi, %rcx)
> -	pcmpeqb	%xmm1, %xmm0
> -	pmovmskb %xmm0, %rdx
> -	add	$16, %rcx
> -# ifdef USE_AS_STRNCPY
> -	sub	$16, %r8
> -	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> -# endif
> -	test	%rdx, %rdx
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	jnz	L(CopyFrom1To16BytesUnalignedXmm1)
> -# else
> -	jnz	L(CopyFrom1To16Bytes)
> -# endif
> -
> -	movaps	16(%rsi, %rcx), %xmm2
> -	movdqu	%xmm1, (%rdi, %rcx)
> -	pcmpeqb	%xmm2, %xmm0
> -	pmovmskb %xmm0, %rdx
> -	add	$16, %rcx
> -# ifdef USE_AS_STRNCPY
> -	sub	$16, %r8
> -	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> -# endif
> -	test	%rdx, %rdx
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	jnz	L(CopyFrom1To16BytesUnalignedXmm2)
> -# else
> -	jnz	L(CopyFrom1To16Bytes)
> -# endif
> -
> -	movaps	16(%rsi, %rcx), %xmm3
> -	movdqu	%xmm2, (%rdi, %rcx)
> -	pcmpeqb	%xmm3, %xmm0
> -	pmovmskb %xmm0, %rdx
> -	add	$16, %rcx
> -# ifdef USE_AS_STRNCPY
> -	sub	$16, %r8
> -	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> -# endif
> -	test	%rdx, %rdx
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	jnz	L(CopyFrom1To16BytesUnalignedXmm3)
> -# else
> -	jnz	L(CopyFrom1To16Bytes)
> -# endif
> -
> -	movdqu	%xmm3, (%rdi, %rcx)
> -	mov	%rsi, %rdx
> -	lea	16(%rsi, %rcx), %rsi
> -	and	$-0x40, %rsi
> -	sub	%rsi, %rdx
> -	sub	%rdx, %rdi
> -# ifdef USE_AS_STRNCPY
> -	lea	128(%r8, %rdx), %r8
> -# endif
> -L(Unaligned64Loop):
> -	movaps	(%rsi), %xmm2
> -	movaps	%xmm2, %xmm4
> -	movaps	16(%rsi), %xmm5
> -	movaps	32(%rsi), %xmm3
> -	movaps	%xmm3, %xmm6
> -	movaps	48(%rsi), %xmm7
> -	pminub	%xmm5, %xmm2
> -	pminub	%xmm7, %xmm3
> -	pminub	%xmm2, %xmm3
> -	pcmpeqb	%xmm0, %xmm3
> -	pmovmskb %xmm3, %rdx
> -# ifdef USE_AS_STRNCPY
> -	sub	$64, %r8
> -	jbe	L(UnalignedLeaveCase2OrCase3)
> -# endif
> -	test	%rdx, %rdx
> -	jnz	L(Unaligned64Leave)
> -
> -L(Unaligned64Loop_start):
> -	add	$64, %rdi
> -	add	$64, %rsi
> -	movdqu	%xmm4, -64(%rdi)
> -	movaps	(%rsi), %xmm2
> -	movdqa	%xmm2, %xmm4
> -	movdqu	%xmm5, -48(%rdi)
> -	movaps	16(%rsi), %xmm5
> -	pminub	%xmm5, %xmm2
> -	movaps	32(%rsi), %xmm3
> -	movdqu	%xmm6, -32(%rdi)
> -	movaps	%xmm3, %xmm6
> -	movdqu	%xmm7, -16(%rdi)
> -	movaps	48(%rsi), %xmm7
> -	pminub	%xmm7, %xmm3
> -	pminub	%xmm2, %xmm3
> -	pcmpeqb	%xmm0, %xmm3
> -	pmovmskb %xmm3, %rdx
> -# ifdef USE_AS_STRNCPY
> -	sub	$64, %r8
> -	jbe	L(UnalignedLeaveCase2OrCase3)
> -# endif
> -	test	%rdx, %rdx
> -	jz	L(Unaligned64Loop_start)
> -
> -L(Unaligned64Leave):
> -	pxor	%xmm1, %xmm1
> -
> -	pcmpeqb	%xmm4, %xmm0
> -	pcmpeqb	%xmm5, %xmm1
> -	pmovmskb %xmm0, %rdx
> -	pmovmskb %xmm1, %rcx
> -	test	%rdx, %rdx
> -	jnz	L(CopyFrom1To16BytesUnaligned_0)
> -	test	%rcx, %rcx
> -	jnz	L(CopyFrom1To16BytesUnaligned_16)
> -
> -	pcmpeqb	%xmm6, %xmm0
> -	pcmpeqb	%xmm7, %xmm1
> -	pmovmskb %xmm0, %rdx
> -	pmovmskb %xmm1, %rcx
> -	test	%rdx, %rdx
> -	jnz	L(CopyFrom1To16BytesUnaligned_32)
> -
> -	bsf	%rcx, %rdx
> -	movdqu	%xmm4, (%rdi)
> -	movdqu	%xmm5, 16(%rdi)
> -	movdqu	%xmm6, 32(%rdi)
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -# ifdef USE_AS_STPCPY
> -	lea	48(%rdi, %rdx), %rax
> -# endif
> -	movdqu	%xmm7, 48(%rdi)
> -	add	$15, %r8
> -	sub	%rdx, %r8
> -	lea	49(%rdi, %rdx), %rdi
> -	jmp	L(StrncpyFillTailWithZero)
> -# else
> -	add	$48, %rsi
> -	add	$48, %rdi
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> -# endif
> -
> -/* If source address alignment == destination address alignment */
> -
> -L(SourceStringAlignmentLess32):
> -	pxor	%xmm0, %xmm0
> -	movdqu	(%rsi), %xmm1
> -	movdqu	16(%rsi), %xmm2
> -	pcmpeqb	%xmm1, %xmm0
> -	pmovmskb %xmm0, %rdx
> -
> -# ifdef USE_AS_STRNCPY
> -#  if defined USE_AS_STPCPY || defined USE_AS_STRCAT
> -	cmp	$16, %r8
> -#  else
> -	cmp	$17, %r8
> -#  endif
> -	jbe	L(CopyFrom1To16BytesTail1Case2OrCase3)
> -# endif
> -	test	%rdx, %rdx
> -	jnz	L(CopyFrom1To16BytesTail1)
> -
> -	pcmpeqb	%xmm2, %xmm0
> -	movdqu	%xmm1, (%rdi)
> -	pmovmskb %xmm0, %rdx
> -
> -# ifdef USE_AS_STRNCPY
> -#  if defined USE_AS_STPCPY || defined USE_AS_STRCAT
> -	cmp	$32, %r8
> -#  else
> -	cmp	$33, %r8
> -#  endif
> -	jbe	L(CopyFrom1To32Bytes1Case2OrCase3)
> -# endif
> -	test	%rdx, %rdx
> -	jnz	L(CopyFrom1To32Bytes1)
> -
> -	and	$-16, %rsi
> -	and	$15, %rcx
> -	jmp	L(Unalign16Both)
> -
> -/*------End of main part with loops---------------------*/
> -
> -/* Case1 */
> -
> -# if (!defined USE_AS_STRNCPY) || (defined USE_AS_STRCAT)
> -	.p2align 4
> -L(CopyFrom1To16Bytes):
> -	add	%rcx, %rdi
> -	add	%rcx, %rsi
> -	bsf	%rdx, %rdx
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> -# endif
> -	.p2align 4
> -L(CopyFrom1To16BytesTail):
> -	add	%rcx, %rsi
> -	bsf	%rdx, %rdx
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> -
> -	.p2align 4
> -L(CopyFrom1To32Bytes1):
> -	add	$16, %rsi
> -	add	$16, %rdi
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$16, %r8
> -# endif
> -L(CopyFrom1To16BytesTail1):
> -	bsf	%rdx, %rdx
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> -
> -	.p2align 4
> -L(CopyFrom1To32Bytes):
> -	bsf	%rdx, %rdx
> -	add	%rcx, %rsi
> -	add	$16, %rdx
> -	sub	%rcx, %rdx
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> -
> -	.p2align 4
> -L(CopyFrom1To16BytesUnaligned_0):
> -	bsf	%rdx, %rdx
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -# ifdef USE_AS_STPCPY
> -	lea	(%rdi, %rdx), %rax
> -# endif
> -	movdqu	%xmm4, (%rdi)
> -	add	$63, %r8
> -	sub	%rdx, %r8
> -	lea	1(%rdi, %rdx), %rdi
> -	jmp	L(StrncpyFillTailWithZero)
> -# else
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> -# endif
> -
> -	.p2align 4
> -L(CopyFrom1To16BytesUnaligned_16):
> -	bsf	%rcx, %rdx
> -	movdqu	%xmm4, (%rdi)
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -# ifdef USE_AS_STPCPY
> -	lea	16(%rdi, %rdx), %rax
> -# endif
> -	movdqu	%xmm5, 16(%rdi)
> -	add	$47, %r8
> -	sub	%rdx, %r8
> -	lea	17(%rdi, %rdx), %rdi
> -	jmp	L(StrncpyFillTailWithZero)
> -# else
> -	add	$16, %rsi
> -	add	$16, %rdi
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> -# endif
> -
> -	.p2align 4
> -L(CopyFrom1To16BytesUnaligned_32):
> -	bsf	%rdx, %rdx
> -	movdqu	%xmm4, (%rdi)
> -	movdqu	%xmm5, 16(%rdi)
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -# ifdef USE_AS_STPCPY
> -	lea	32(%rdi, %rdx), %rax
> -# endif
> -	movdqu	%xmm6, 32(%rdi)
> -	add	$31, %r8
> -	sub	%rdx, %r8
> -	lea	33(%rdi, %rdx), %rdi
> -	jmp	L(StrncpyFillTailWithZero)
> -# else
> -	add	$32, %rsi
> -	add	$32, %rdi
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> -# endif
> -
> -# ifdef USE_AS_STRNCPY
> -#  ifndef USE_AS_STRCAT
> -	.p2align 4
> -L(CopyFrom1To16BytesUnalignedXmm6):
> -	movdqu	%xmm6, (%rdi, %rcx)
> -	jmp	L(CopyFrom1To16BytesXmmExit)
> -
> -	.p2align 4
> -L(CopyFrom1To16BytesUnalignedXmm5):
> -	movdqu	%xmm5, (%rdi, %rcx)
> -	jmp	L(CopyFrom1To16BytesXmmExit)
> -
> -	.p2align 4
> -L(CopyFrom1To16BytesUnalignedXmm4):
> -	movdqu	%xmm4, (%rdi, %rcx)
> -	jmp	L(CopyFrom1To16BytesXmmExit)
> -
> -	.p2align 4
> -L(CopyFrom1To16BytesUnalignedXmm3):
> -	movdqu	%xmm3, (%rdi, %rcx)
> -	jmp	L(CopyFrom1To16BytesXmmExit)
> -
> -	.p2align 4
> -L(CopyFrom1To16BytesUnalignedXmm1):
> -	movdqu	%xmm1, (%rdi, %rcx)
> -	jmp	L(CopyFrom1To16BytesXmmExit)
> -#  endif
> -
> -	.p2align 4
> -L(CopyFrom1To16BytesExit):
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> -
> -/* Case2 */
> -
> -	.p2align 4
> -L(CopyFrom1To16BytesCase2):
> -	add	$16, %r8
> -	add	%rcx, %rdi
> -	add	%rcx, %rsi
> -	bsf	%rdx, %rdx
> -	cmp	%r8, %rdx
> -	jb	L(CopyFrom1To16BytesExit)
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> -
> -	.p2align 4
> -L(CopyFrom1To32BytesCase2):
> -	add	%rcx, %rsi
> -	bsf	%rdx, %rdx
> -	add	$16, %rdx
> -	sub	%rcx, %rdx
> -	cmp	%r8, %rdx
> -	jb	L(CopyFrom1To16BytesExit)
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> -
> -L(CopyFrom1To16BytesTailCase2):
> -	add	%rcx, %rsi
> -	bsf	%rdx, %rdx
> -	cmp	%r8, %rdx
> -	jb	L(CopyFrom1To16BytesExit)
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> -
> -L(CopyFrom1To16BytesTail1Case2):
> -	bsf	%rdx, %rdx
> -	cmp	%r8, %rdx
> -	jb	L(CopyFrom1To16BytesExit)
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> -
> -/* Case2 or Case3,  Case3 */
> -
> -	.p2align 4
> -L(CopyFrom1To16BytesCase2OrCase3):
> -	test	%rdx, %rdx
> -	jnz	L(CopyFrom1To16BytesCase2)
> -L(CopyFrom1To16BytesCase3):
> -	add	$16, %r8
> -	add	%rcx, %rdi
> -	add	%rcx, %rsi
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> -
> -	.p2align 4
> -L(CopyFrom1To32BytesCase2OrCase3):
> -	test	%rdx, %rdx
> -	jnz	L(CopyFrom1To32BytesCase2)
> -	add	%rcx, %rsi
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> -
> -	.p2align 4
> -L(CopyFrom1To16BytesTailCase2OrCase3):
> -	test	%rdx, %rdx
> -	jnz	L(CopyFrom1To16BytesTailCase2)
> -	add	%rcx, %rsi
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> -
> -	.p2align 4
> -L(CopyFrom1To32Bytes1Case2OrCase3):
> -	add	$16, %rdi
> -	add	$16, %rsi
> -	sub	$16, %r8
> -L(CopyFrom1To16BytesTail1Case2OrCase3):
> -	test	%rdx, %rdx
> -	jnz	L(CopyFrom1To16BytesTail1Case2)
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> -
> -# endif
> -
> -/*------------End labels regarding with copying 1-16 bytes--and 1-32 bytes----*/
> -
> -	.p2align 4
> -L(Exit1):
> -	mov	%dh, (%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$1, %r8
> -	lea	1(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit2):
> -	mov	(%rsi), %dx
> -	mov	%dx, (%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	1(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$2, %r8
> -	lea	2(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit3):
> -	mov	(%rsi), %cx
> -	mov	%cx, (%rdi)
> -	mov	%dh, 2(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	2(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$3, %r8
> -	lea	3(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit4):
> -	mov	(%rsi), %edx
> -	mov	%edx, (%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	3(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$4, %r8
> -	lea	4(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit5):
> -	mov	(%rsi), %ecx
> -	mov	%dh, 4(%rdi)
> -	mov	%ecx, (%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	4(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$5, %r8
> -	lea	5(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit6):
> -	mov	(%rsi), %ecx
> -	mov	4(%rsi), %dx
> -	mov	%ecx, (%rdi)
> -	mov	%dx, 4(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	5(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$6, %r8
> -	lea	6(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit7):
> -	mov	(%rsi), %ecx
> -	mov	3(%rsi), %edx
> -	mov	%ecx, (%rdi)
> -	mov	%edx, 3(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	6(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$7, %r8
> -	lea	7(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit8):
> -	mov	(%rsi), %rdx
> -	mov	%rdx, (%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	7(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$8, %r8
> -	lea	8(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit9):
> -	mov	(%rsi), %rcx
> -	mov	%dh, 8(%rdi)
> -	mov	%rcx, (%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	8(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$9, %r8
> -	lea	9(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit10):
> -	mov	(%rsi), %rcx
> -	mov	8(%rsi), %dx
> -	mov	%rcx, (%rdi)
> -	mov	%dx, 8(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	9(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$10, %r8
> -	lea	10(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit11):
> -	mov	(%rsi), %rcx
> -	mov	7(%rsi), %edx
> -	mov	%rcx, (%rdi)
> -	mov	%edx, 7(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	10(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$11, %r8
> -	lea	11(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit12):
> -	mov	(%rsi), %rcx
> -	mov	8(%rsi), %edx
> -	mov	%rcx, (%rdi)
> -	mov	%edx, 8(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	11(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$12, %r8
> -	lea	12(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit13):
> -	mov	(%rsi), %rcx
> -	mov	5(%rsi), %rdx
> -	mov	%rcx, (%rdi)
> -	mov	%rdx, 5(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	12(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$13, %r8
> -	lea	13(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit14):
> -	mov	(%rsi), %rcx
> -	mov	6(%rsi), %rdx
> -	mov	%rcx, (%rdi)
> -	mov	%rdx, 6(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	13(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$14, %r8
> -	lea	14(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit15):
> -	mov	(%rsi), %rcx
> -	mov	7(%rsi), %rdx
> -	mov	%rcx, (%rdi)
> -	mov	%rdx, 7(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	14(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$15, %r8
> -	lea	15(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit16):
> -	movdqu	(%rsi), %xmm0
> -	movdqu	%xmm0, (%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	15(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$16, %r8
> -	lea	16(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit17):
> -	movdqu	(%rsi), %xmm0
> -	movdqu	%xmm0, (%rdi)
> -	mov	%dh, 16(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	16(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$17, %r8
> -	lea	17(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit18):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %cx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%cx, 16(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	17(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$18, %r8
> -	lea	18(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit19):
> -	movdqu	(%rsi), %xmm0
> -	mov	15(%rsi), %ecx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%ecx, 15(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	18(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$19, %r8
> -	lea	19(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit20):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %ecx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%ecx, 16(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	19(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$20, %r8
> -	lea	20(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit21):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %ecx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%ecx, 16(%rdi)
> -	mov	%dh, 20(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	20(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$21, %r8
> -	lea	21(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit22):
> -	movdqu	(%rsi), %xmm0
> -	mov	14(%rsi), %rcx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rcx, 14(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	21(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$22, %r8
> -	lea	22(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit23):
> -	movdqu	(%rsi), %xmm0
> -	mov	15(%rsi), %rcx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rcx, 15(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	22(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$23, %r8
> -	lea	23(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit24):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %rcx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rcx, 16(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	23(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$24, %r8
> -	lea	24(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit25):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %rcx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rcx, 16(%rdi)
> -	mov	%dh, 24(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	24(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$25, %r8
> -	lea	25(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit26):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %rdx
> -	mov	24(%rsi), %cx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rdx, 16(%rdi)
> -	mov	%cx, 24(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	25(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$26, %r8
> -	lea	26(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit27):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %rdx
> -	mov	23(%rsi), %ecx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rdx, 16(%rdi)
> -	mov	%ecx, 23(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	26(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$27, %r8
> -	lea	27(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit28):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %rdx
> -	mov	24(%rsi), %ecx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rdx, 16(%rdi)
> -	mov	%ecx, 24(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	27(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$28, %r8
> -	lea	28(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit29):
> -	movdqu	(%rsi), %xmm0
> -	movdqu	13(%rsi), %xmm2
> -	movdqu	%xmm0, (%rdi)
> -	movdqu	%xmm2, 13(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	28(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$29, %r8
> -	lea	29(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit30):
> -	movdqu	(%rsi), %xmm0
> -	movdqu	14(%rsi), %xmm2
> -	movdqu	%xmm0, (%rdi)
> -	movdqu	%xmm2, 14(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	29(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$30, %r8
> -	lea	30(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit31):
> -	movdqu	(%rsi), %xmm0
> -	movdqu	15(%rsi), %xmm2
> -	movdqu	%xmm0, (%rdi)
> -	movdqu	%xmm2, 15(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	30(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$31, %r8
> -	lea	31(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -	.p2align 4
> -L(Exit32):
> -	movdqu	(%rsi), %xmm0
> -	movdqu	16(%rsi), %xmm2
> -	movdqu	%xmm0, (%rdi)
> -	movdqu	%xmm2, 16(%rdi)
> -# ifdef USE_AS_STPCPY
> -	lea	31(%rdi), %rax
> -# endif
> -# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> -	sub	$32, %r8
> -	lea	32(%rdi), %rdi
> -	jnz	L(StrncpyFillTailWithZero)
> -# endif
> -	ret
> -
> -# ifdef USE_AS_STRNCPY
> -
> -	.p2align 4
> -L(StrncpyExit0):
> -#  ifdef USE_AS_STPCPY
> -	mov	%rdi, %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, (%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit1):
> -	mov	(%rsi), %dl
> -	mov	%dl, (%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	1(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 1(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit2):
> -	mov	(%rsi), %dx
> -	mov	%dx, (%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	2(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 2(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit3):
> -	mov	(%rsi), %cx
> -	mov	2(%rsi), %dl
> -	mov	%cx, (%rdi)
> -	mov	%dl, 2(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	3(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 3(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit4):
> -	mov	(%rsi), %edx
> -	mov	%edx, (%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	4(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 4(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit5):
> -	mov	(%rsi), %ecx
> -	mov	4(%rsi), %dl
> -	mov	%ecx, (%rdi)
> -	mov	%dl, 4(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	5(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 5(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit6):
> -	mov	(%rsi), %ecx
> -	mov	4(%rsi), %dx
> -	mov	%ecx, (%rdi)
> -	mov	%dx, 4(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	6(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 6(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit7):
> -	mov	(%rsi), %ecx
> -	mov	3(%rsi), %edx
> -	mov	%ecx, (%rdi)
> -	mov	%edx, 3(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	7(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 7(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit8):
> -	mov	(%rsi), %rdx
> -	mov	%rdx, (%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	8(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 8(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit9):
> -	mov	(%rsi), %rcx
> -	mov	8(%rsi), %dl
> -	mov	%rcx, (%rdi)
> -	mov	%dl, 8(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	9(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 9(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit10):
> -	mov	(%rsi), %rcx
> -	mov	8(%rsi), %dx
> -	mov	%rcx, (%rdi)
> -	mov	%dx, 8(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	10(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 10(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit11):
> -	mov	(%rsi), %rcx
> -	mov	7(%rsi), %edx
> -	mov	%rcx, (%rdi)
> -	mov	%edx, 7(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	11(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 11(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit12):
> -	mov	(%rsi), %rcx
> -	mov	8(%rsi), %edx
> -	mov	%rcx, (%rdi)
> -	mov	%edx, 8(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	12(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 12(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit13):
> -	mov	(%rsi), %rcx
> -	mov	5(%rsi), %rdx
> -	mov	%rcx, (%rdi)
> -	mov	%rdx, 5(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	13(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 13(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit14):
> -	mov	(%rsi), %rcx
> -	mov	6(%rsi), %rdx
> -	mov	%rcx, (%rdi)
> -	mov	%rdx, 6(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	14(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 14(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit15):
> -	mov	(%rsi), %rcx
> -	mov	7(%rsi), %rdx
> -	mov	%rcx, (%rdi)
> -	mov	%rdx, 7(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	15(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 15(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit16):
> -	movdqu	(%rsi), %xmm0
> -	movdqu	%xmm0, (%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	16(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 16(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit17):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %cl
> -	movdqu	%xmm0, (%rdi)
> -	mov	%cl, 16(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	17(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 17(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit18):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %cx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%cx, 16(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	18(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 18(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit19):
> -	movdqu	(%rsi), %xmm0
> -	mov	15(%rsi), %ecx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%ecx, 15(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	19(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 19(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit20):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %ecx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%ecx, 16(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	20(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 20(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit21):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %ecx
> -	mov	20(%rsi), %dl
> -	movdqu	%xmm0, (%rdi)
> -	mov	%ecx, 16(%rdi)
> -	mov	%dl, 20(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	21(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 21(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit22):
> -	movdqu	(%rsi), %xmm0
> -	mov	14(%rsi), %rcx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rcx, 14(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	22(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 22(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit23):
> -	movdqu	(%rsi), %xmm0
> -	mov	15(%rsi), %rcx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rcx, 15(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	23(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 23(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit24):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %rcx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rcx, 16(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	24(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 24(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit25):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %rdx
> -	mov	24(%rsi), %cl
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rdx, 16(%rdi)
> -	mov	%cl, 24(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	25(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 25(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit26):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %rdx
> -	mov	24(%rsi), %cx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rdx, 16(%rdi)
> -	mov	%cx, 24(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	26(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 26(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit27):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %rdx
> -	mov	23(%rsi), %ecx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rdx, 16(%rdi)
> -	mov	%ecx, 23(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	27(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 27(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit28):
> -	movdqu	(%rsi), %xmm0
> -	mov	16(%rsi), %rdx
> -	mov	24(%rsi), %ecx
> -	movdqu	%xmm0, (%rdi)
> -	mov	%rdx, 16(%rdi)
> -	mov	%ecx, 24(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	28(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 28(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit29):
> -	movdqu	(%rsi), %xmm0
> -	movdqu	13(%rsi), %xmm2
> -	movdqu	%xmm0, (%rdi)
> -	movdqu	%xmm2, 13(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	29(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 29(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit30):
> -	movdqu	(%rsi), %xmm0
> -	movdqu	14(%rsi), %xmm2
> -	movdqu	%xmm0, (%rdi)
> -	movdqu	%xmm2, 14(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	30(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 30(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit31):
> -	movdqu	(%rsi), %xmm0
> -	movdqu	15(%rsi), %xmm2
> -	movdqu	%xmm0, (%rdi)
> -	movdqu	%xmm2, 15(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	31(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 31(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit32):
> -	movdqu	(%rsi), %xmm0
> -	movdqu	16(%rsi), %xmm2
> -	movdqu	%xmm0, (%rdi)
> -	movdqu	%xmm2, 16(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	32(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 32(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(StrncpyExit33):
> -	movdqu	(%rsi), %xmm0
> -	movdqu	16(%rsi), %xmm2
> -	mov	32(%rsi), %cl
> -	movdqu	%xmm0, (%rdi)
> -	movdqu	%xmm2, 16(%rdi)
> -	mov	%cl, 32(%rdi)
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 33(%rdi)
> -#  endif
> -	ret
> -
> -#  ifndef USE_AS_STRCAT
> -
> -	.p2align 4
> -L(Fill0):
> -	ret
> -
> -	.p2align 4
> -L(Fill1):
> -	mov	%dl, (%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill2):
> -	mov	%dx, (%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill3):
> -	mov	%edx, -1(%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill4):
> -	mov	%edx, (%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill5):
> -	mov	%edx, (%rdi)
> -	mov	%dl, 4(%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill6):
> -	mov	%edx, (%rdi)
> -	mov	%dx, 4(%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill7):
> -	mov	%rdx, -1(%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill8):
> -	mov	%rdx, (%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill9):
> -	mov	%rdx, (%rdi)
> -	mov	%dl, 8(%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill10):
> -	mov	%rdx, (%rdi)
> -	mov	%dx, 8(%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill11):
> -	mov	%rdx, (%rdi)
> -	mov	%edx, 7(%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill12):
> -	mov	%rdx, (%rdi)
> -	mov	%edx, 8(%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill13):
> -	mov	%rdx, (%rdi)
> -	mov	%rdx, 5(%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill14):
> -	mov	%rdx, (%rdi)
> -	mov	%rdx, 6(%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill15):
> -	movdqu	%xmm0, -1(%rdi)
> -	ret
> -
> -	.p2align 4
> -L(Fill16):
> -	movdqu	%xmm0, (%rdi)
> -	ret
> -
> -	.p2align 4
> -L(CopyFrom1To16BytesUnalignedXmm2):
> -	movdqu	%xmm2, (%rdi, %rcx)
> -
> -	.p2align 4
> -L(CopyFrom1To16BytesXmmExit):
> -	bsf	%rdx, %rdx
> -	add	$15, %r8
> -	add	%rcx, %rdi
> -#   ifdef USE_AS_STPCPY
> -	lea	(%rdi, %rdx), %rax
> -#   endif
> -	sub	%rdx, %r8
> -	lea	1(%rdi, %rdx), %rdi
> -
> -	.p2align 4
> -L(StrncpyFillTailWithZero):
> -	pxor	%xmm0, %xmm0
> -	xor	%rdx, %rdx
> -	sub	$16, %r8
> -	jbe	L(StrncpyFillExit)
> -
> -	movdqu	%xmm0, (%rdi)
> -	add	$16, %rdi
> -
> -	mov	%rdi, %rsi
> -	and	$0xf, %rsi
> -	sub	%rsi, %rdi
> -	add	%rsi, %r8
> -	sub	$64, %r8
> -	jb	L(StrncpyFillLess64)
> -
> -L(StrncpyFillLoopMovdqa):
> -	movdqa	%xmm0, (%rdi)
> -	movdqa	%xmm0, 16(%rdi)
> -	movdqa	%xmm0, 32(%rdi)
> -	movdqa	%xmm0, 48(%rdi)
> -	add	$64, %rdi
> -	sub	$64, %r8
> -	jae	L(StrncpyFillLoopMovdqa)
> -
> -L(StrncpyFillLess64):
> -	add	$32, %r8
> -	jl	L(StrncpyFillLess32)
> -	movdqa	%xmm0, (%rdi)
> -	movdqa	%xmm0, 16(%rdi)
> -	add	$32, %rdi
> -	sub	$16, %r8
> -	jl	L(StrncpyFillExit)
> -	movdqa	%xmm0, (%rdi)
> -	add	$16, %rdi
> -	BRANCH_TO_JMPTBL_ENTRY (L(FillTable), %r8, 4)
> -
> -L(StrncpyFillLess32):
> -	add	$16, %r8
> -	jl	L(StrncpyFillExit)
> -	movdqa	%xmm0, (%rdi)
> -	add	$16, %rdi
> -	BRANCH_TO_JMPTBL_ENTRY (L(FillTable), %r8, 4)
> -
> -L(StrncpyFillExit):
> -	add	$16, %r8
> -	BRANCH_TO_JMPTBL_ENTRY (L(FillTable), %r8, 4)
> -
> -/* end of ifndef USE_AS_STRCAT */
> -#  endif
> -
> -	.p2align 4
> -L(UnalignedLeaveCase2OrCase3):
> -	test	%rdx, %rdx
> -	jnz	L(Unaligned64LeaveCase2)
> -L(Unaligned64LeaveCase3):
> -	lea	64(%r8), %rcx
> -	and	$-16, %rcx
> -	add	$48, %r8
> -	jl	L(CopyFrom1To16BytesCase3)
> -	movdqu	%xmm4, (%rdi)
> -	sub	$16, %r8
> -	jb	L(CopyFrom1To16BytesCase3)
> -	movdqu	%xmm5, 16(%rdi)
> -	sub	$16, %r8
> -	jb	L(CopyFrom1To16BytesCase3)
> -	movdqu	%xmm6, 32(%rdi)
> -	sub	$16, %r8
> -	jb	L(CopyFrom1To16BytesCase3)
> -	movdqu	%xmm7, 48(%rdi)
> -#  ifdef USE_AS_STPCPY
> -	lea	64(%rdi), %rax
> -#  endif
> -#  ifdef USE_AS_STRCAT
> -	xor	%ch, %ch
> -	movb	%ch, 64(%rdi)
> -#  endif
> -	ret
> -
> -	.p2align 4
> -L(Unaligned64LeaveCase2):
> -	xor	%rcx, %rcx
> -	pcmpeqb	%xmm4, %xmm0
> -	pmovmskb %xmm0, %rdx
> -	add	$48, %r8
> -	jle	L(CopyFrom1To16BytesCase2OrCase3)
> -	test	%rdx, %rdx
> -#  ifndef USE_AS_STRCAT
> -	jnz	L(CopyFrom1To16BytesUnalignedXmm4)
> -#  else
> -	jnz	L(CopyFrom1To16Bytes)
> -#  endif
> -	pcmpeqb	%xmm5, %xmm0
> -	pmovmskb %xmm0, %rdx
> -	movdqu	%xmm4, (%rdi)
> -	add	$16, %rcx
> -	sub	$16, %r8
> -	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> -	test	%rdx, %rdx
> -#  ifndef USE_AS_STRCAT
> -	jnz	L(CopyFrom1To16BytesUnalignedXmm5)
> -#  else
> -	jnz	L(CopyFrom1To16Bytes)
> -#  endif
> -
> -	pcmpeqb	%xmm6, %xmm0
> -	pmovmskb %xmm0, %rdx
> -	movdqu	%xmm5, 16(%rdi)
> -	add	$16, %rcx
> -	sub	$16, %r8
> -	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> -	test	%rdx, %rdx
> -#  ifndef USE_AS_STRCAT
> -	jnz	L(CopyFrom1To16BytesUnalignedXmm6)
> -#  else
> -	jnz	L(CopyFrom1To16Bytes)
> -#  endif
> -
> -	pcmpeqb	%xmm7, %xmm0
> -	pmovmskb %xmm0, %rdx
> -	movdqu	%xmm6, 32(%rdi)
> -	lea	16(%rdi, %rcx), %rdi
> -	lea	16(%rsi, %rcx), %rsi
> -	bsf	%rdx, %rdx
> -	cmp	%r8, %rdx
> -	jb	L(CopyFrom1To16BytesExit)
> -	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> -
> -	.p2align 4
> -L(ExitZero):
> -#  ifndef USE_AS_STRCAT
> -	mov	%rdi, %rax
> -#  endif
> -	ret
> -
> -# endif
> -
> -# ifndef USE_AS_STRCAT
> -END (STRCPY)
> -# else
> -END (STRCAT)
> -# endif
> -	.p2align 4
> -	.section .rodata
> -L(ExitTable):
> -	.int	JMPTBL(L(Exit1), L(ExitTable))
> -	.int	JMPTBL(L(Exit2), L(ExitTable))
> -	.int	JMPTBL(L(Exit3), L(ExitTable))
> -	.int	JMPTBL(L(Exit4), L(ExitTable))
> -	.int	JMPTBL(L(Exit5), L(ExitTable))
> -	.int	JMPTBL(L(Exit6), L(ExitTable))
> -	.int	JMPTBL(L(Exit7), L(ExitTable))
> -	.int	JMPTBL(L(Exit8), L(ExitTable))
> -	.int	JMPTBL(L(Exit9), L(ExitTable))
> -	.int	JMPTBL(L(Exit10), L(ExitTable))
> -	.int	JMPTBL(L(Exit11), L(ExitTable))
> -	.int	JMPTBL(L(Exit12), L(ExitTable))
> -	.int	JMPTBL(L(Exit13), L(ExitTable))
> -	.int	JMPTBL(L(Exit14), L(ExitTable))
> -	.int	JMPTBL(L(Exit15), L(ExitTable))
> -	.int	JMPTBL(L(Exit16), L(ExitTable))
> -	.int	JMPTBL(L(Exit17), L(ExitTable))
> -	.int	JMPTBL(L(Exit18), L(ExitTable))
> -	.int	JMPTBL(L(Exit19), L(ExitTable))
> -	.int	JMPTBL(L(Exit20), L(ExitTable))
> -	.int	JMPTBL(L(Exit21), L(ExitTable))
> -	.int	JMPTBL(L(Exit22), L(ExitTable))
> -	.int    JMPTBL(L(Exit23), L(ExitTable))
> -	.int	JMPTBL(L(Exit24), L(ExitTable))
> -	.int	JMPTBL(L(Exit25), L(ExitTable))
> -	.int	JMPTBL(L(Exit26), L(ExitTable))
> -	.int	JMPTBL(L(Exit27), L(ExitTable))
> -	.int	JMPTBL(L(Exit28), L(ExitTable))
> -	.int	JMPTBL(L(Exit29), L(ExitTable))
> -	.int	JMPTBL(L(Exit30), L(ExitTable))
> -	.int	JMPTBL(L(Exit31), L(ExitTable))
> -	.int	JMPTBL(L(Exit32), L(ExitTable))
> -# ifdef USE_AS_STRNCPY
> -L(ExitStrncpyTable):
> -	.int	JMPTBL(L(StrncpyExit0), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit1), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit2), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit3), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit4), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit5), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit6), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit7), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit8), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit9), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit10), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit11), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit12), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit13), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit14), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit15), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit16), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit17), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit18), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit19), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit20), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit21), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit22), L(ExitStrncpyTable))
> -	.int    JMPTBL(L(StrncpyExit23), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit24), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit25), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit26), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit27), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit28), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit29), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit30), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit31), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit32), L(ExitStrncpyTable))
> -	.int	JMPTBL(L(StrncpyExit33), L(ExitStrncpyTable))
> -#  ifndef USE_AS_STRCAT
> -	.p2align 4
> -L(FillTable):
> -	.int	JMPTBL(L(Fill0), L(FillTable))
> -	.int	JMPTBL(L(Fill1), L(FillTable))
> -	.int	JMPTBL(L(Fill2), L(FillTable))
> -	.int	JMPTBL(L(Fill3), L(FillTable))
> -	.int	JMPTBL(L(Fill4), L(FillTable))
> -	.int	JMPTBL(L(Fill5), L(FillTable))
> -	.int	JMPTBL(L(Fill6), L(FillTable))
> -	.int	JMPTBL(L(Fill7), L(FillTable))
> -	.int	JMPTBL(L(Fill8), L(FillTable))
> -	.int	JMPTBL(L(Fill9), L(FillTable))
> -	.int	JMPTBL(L(Fill10), L(FillTable))
> -	.int	JMPTBL(L(Fill11), L(FillTable))
> -	.int	JMPTBL(L(Fill12), L(FillTable))
> -	.int	JMPTBL(L(Fill13), L(FillTable))
> -	.int	JMPTBL(L(Fill14), L(FillTable))
> -	.int	JMPTBL(L(Fill15), L(FillTable))
> -	.int	JMPTBL(L(Fill16), L(FillTable))
> -#  endif
> -# endif
> -#endif
> +#define AS_STRCPY
> +#define STPCPY __strcpy_sse2_unaligned
> +#include "stpcpy-sse2-unaligned.S"
> diff --git a/sysdeps/x86_64/multiarch/strcpy.S b/sysdeps/x86_64/multiarch/strcpy.S
> index 9464ee8..92be04c 100644
> --- a/sysdeps/x86_64/multiarch/strcpy.S
> +++ b/sysdeps/x86_64/multiarch/strcpy.S
> @@ -28,31 +28,18 @@
>  #endif
>  
>  #ifdef USE_AS_STPCPY
> -# ifdef USE_AS_STRNCPY
> -#  define STRCPY_SSSE3		__stpncpy_ssse3
> -#  define STRCPY_SSE2		__stpncpy_sse2
> -#  define STRCPY_SSE2_UNALIGNED __stpncpy_sse2_unaligned
> -#  define __GI_STRCPY		__GI_stpncpy
> -#  define __GI___STRCPY		__GI___stpncpy
> -# else
>  #  define STRCPY_SSSE3		__stpcpy_ssse3
>  #  define STRCPY_SSE2		__stpcpy_sse2
> +#  define STRCPY_AVX2		__stpcpy_avx2
>  #  define STRCPY_SSE2_UNALIGNED	__stpcpy_sse2_unaligned
>  #  define __GI_STRCPY		__GI_stpcpy
>  #  define __GI___STRCPY		__GI___stpcpy
> -# endif
>  #else
> -# ifdef USE_AS_STRNCPY
> -#  define STRCPY_SSSE3		__strncpy_ssse3
> -#  define STRCPY_SSE2		__strncpy_sse2
> -#  define STRCPY_SSE2_UNALIGNED	__strncpy_sse2_unaligned
> -#  define __GI_STRCPY		__GI_strncpy
> -# else
>  #  define STRCPY_SSSE3		__strcpy_ssse3
> +#  define STRCPY_AVX2		__strcpy_avx2
>  #  define STRCPY_SSE2		__strcpy_sse2
>  #  define STRCPY_SSE2_UNALIGNED	__strcpy_sse2_unaligned
>  #  define __GI_STRCPY		__GI_strcpy
> -# endif
>  #endif
>  
>  
> @@ -64,7 +51,10 @@ ENTRY(STRCPY)
>  	cmpl	$0, __cpu_features+KIND_OFFSET(%rip)
>  	jne	1f
>  	call	__init_cpu_features
> -1:	leaq	STRCPY_SSE2_UNALIGNED(%rip), %rax
> +1:	leaq	STRCPY_AVX2(%rip), %rax
> +	testl   $bit_AVX_Fast_Unaligned_Load, __cpu_features+FEATURE_OFFSET+index_AVX_Fast_Unaligned_Load(%rip)
> +	jnz	2f
> +	leaq	STRCPY_SSE2_UNALIGNED(%rip), %rax
>  	testl	$bit_Fast_Unaligned_Load, __cpu_features+FEATURE_OFFSET+index_Fast_Unaligned_Load(%rip)
>  	jnz	2f
>  	leaq	STRCPY_SSE2(%rip), %rax
> diff --git a/sysdeps/x86_64/multiarch/strncpy-sse2-unaligned.S b/sysdeps/x86_64/multiarch/strncpy-sse2-unaligned.S
> index fcc23a7..e4c98e7 100644
> --- a/sysdeps/x86_64/multiarch/strncpy-sse2-unaligned.S
> +++ b/sysdeps/x86_64/multiarch/strncpy-sse2-unaligned.S
> @@ -1,3 +1,1888 @@
> -#define USE_AS_STRNCPY
> -#define STRCPY __strncpy_sse2_unaligned
> -#include "strcpy-sse2-unaligned.S"
> +/* strcpy with SSE2 and unaligned load
> +   Copyright (C) 2011-2015 Free Software Foundation, Inc.
> +   Contributed by Intel Corporation.
> +   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/>.  */
> +
> +#if IS_IN (libc)
> +
> +# ifndef USE_AS_STRCAT
> +#  include <sysdep.h>
> +
> +#  ifndef STRCPY
> +#   define STRCPY  __strncpy_sse2_unaligned
> +#  endif
> +
> +# define USE_AS_STRNCPY
> +# endif
> +
> +# define JMPTBL(I, B)	I - B
> +# define BRANCH_TO_JMPTBL_ENTRY(TABLE, INDEX, SCALE)             \
> +	lea	TABLE(%rip), %r11;                              \
> +	movslq	(%r11, INDEX, SCALE), %rcx;                     \
> +	lea	(%r11, %rcx), %rcx;                             \
> +	jmp	*%rcx
> +
> +# ifndef USE_AS_STRCAT
> +
> +.text
> +ENTRY (STRCPY)
> +#  ifdef USE_AS_STRNCPY
> +	mov	%rdx, %r8
> +	test	%r8, %r8
> +	jz	L(ExitZero)
> +#  endif
> +	mov	%rsi, %rcx
> +#  ifndef USE_AS_STPCPY
> +	mov	%rdi, %rax      /* save result */
> +#  endif
> +
> +# endif
> +
> +	and	$63, %rcx
> +	cmp	$32, %rcx
> +	jbe	L(SourceStringAlignmentLess32)
> +
> +	and	$-16, %rsi
> +	and	$15, %rcx
> +	pxor	%xmm0, %xmm0
> +	pxor	%xmm1, %xmm1
> +
> +	pcmpeqb	(%rsi), %xmm1
> +	pmovmskb %xmm1, %rdx
> +	shr	%cl, %rdx
> +
> +# ifdef USE_AS_STRNCPY
> +#  if defined USE_AS_STPCPY || defined USE_AS_STRCAT
> +	mov	$16, %r10
> +	sub	%rcx, %r10
> +	cmp	%r10, %r8
> +#  else
> +	mov	$17, %r10
> +	sub	%rcx, %r10
> +	cmp	%r10, %r8
> +#  endif
> +	jbe	L(CopyFrom1To16BytesTailCase2OrCase3)
> +# endif
> +	test	%rdx, %rdx
> +	jnz	L(CopyFrom1To16BytesTail)
> +
> +	pcmpeqb	16(%rsi), %xmm0
> +	pmovmskb %xmm0, %rdx
> +
> +# ifdef USE_AS_STRNCPY
> +	add	$16, %r10
> +	cmp	%r10, %r8
> +	jbe	L(CopyFrom1To32BytesCase2OrCase3)
> +# endif
> +	test	%rdx, %rdx
> +	jnz	L(CopyFrom1To32Bytes)
> +
> +	movdqu	(%rsi, %rcx), %xmm1   /* copy 16 bytes */
> +	movdqu	%xmm1, (%rdi)
> +
> +/* If source address alignment != destination address alignment */
> +	.p2align 4
> +L(Unalign16Both):
> +	sub	%rcx, %rdi
> +# ifdef USE_AS_STRNCPY
> +	add	%rcx, %r8
> +# endif
> +	mov	$16, %rcx
> +	movdqa	(%rsi, %rcx), %xmm1
> +	movaps	16(%rsi, %rcx), %xmm2
> +	movdqu	%xmm1, (%rdi, %rcx)
> +	pcmpeqb	%xmm2, %xmm0
> +	pmovmskb %xmm0, %rdx
> +	add	$16, %rcx
> +# ifdef USE_AS_STRNCPY
> +	sub	$48, %r8
> +	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> +# endif
> +	test	%rdx, %rdx
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	jnz	L(CopyFrom1To16BytesUnalignedXmm2)
> +# else
> +	jnz	L(CopyFrom1To16Bytes)
> +# endif
> +
> +	movaps	16(%rsi, %rcx), %xmm3
> +	movdqu	%xmm2, (%rdi, %rcx)
> +	pcmpeqb	%xmm3, %xmm0
> +	pmovmskb %xmm0, %rdx
> +	add	$16, %rcx
> +# ifdef USE_AS_STRNCPY
> +	sub	$16, %r8
> +	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> +# endif
> +	test	%rdx, %rdx
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	jnz	L(CopyFrom1To16BytesUnalignedXmm3)
> +# else
> +	jnz	L(CopyFrom1To16Bytes)
> +# endif
> +
> +	movaps	16(%rsi, %rcx), %xmm4
> +	movdqu	%xmm3, (%rdi, %rcx)
> +	pcmpeqb	%xmm4, %xmm0
> +	pmovmskb %xmm0, %rdx
> +	add	$16, %rcx
> +# ifdef USE_AS_STRNCPY
> +	sub	$16, %r8
> +	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> +# endif
> +	test	%rdx, %rdx
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	jnz	L(CopyFrom1To16BytesUnalignedXmm4)
> +# else
> +	jnz	L(CopyFrom1To16Bytes)
> +# endif
> +
> +	movaps	16(%rsi, %rcx), %xmm1
> +	movdqu	%xmm4, (%rdi, %rcx)
> +	pcmpeqb	%xmm1, %xmm0
> +	pmovmskb %xmm0, %rdx
> +	add	$16, %rcx
> +# ifdef USE_AS_STRNCPY
> +	sub	$16, %r8
> +	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> +# endif
> +	test	%rdx, %rdx
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	jnz	L(CopyFrom1To16BytesUnalignedXmm1)
> +# else
> +	jnz	L(CopyFrom1To16Bytes)
> +# endif
> +
> +	movaps	16(%rsi, %rcx), %xmm2
> +	movdqu	%xmm1, (%rdi, %rcx)
> +	pcmpeqb	%xmm2, %xmm0
> +	pmovmskb %xmm0, %rdx
> +	add	$16, %rcx
> +# ifdef USE_AS_STRNCPY
> +	sub	$16, %r8
> +	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> +# endif
> +	test	%rdx, %rdx
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	jnz	L(CopyFrom1To16BytesUnalignedXmm2)
> +# else
> +	jnz	L(CopyFrom1To16Bytes)
> +# endif
> +
> +	movaps	16(%rsi, %rcx), %xmm3
> +	movdqu	%xmm2, (%rdi, %rcx)
> +	pcmpeqb	%xmm3, %xmm0
> +	pmovmskb %xmm0, %rdx
> +	add	$16, %rcx
> +# ifdef USE_AS_STRNCPY
> +	sub	$16, %r8
> +	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> +# endif
> +	test	%rdx, %rdx
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	jnz	L(CopyFrom1To16BytesUnalignedXmm3)
> +# else
> +	jnz	L(CopyFrom1To16Bytes)
> +# endif
> +
> +	movdqu	%xmm3, (%rdi, %rcx)
> +	mov	%rsi, %rdx
> +	lea	16(%rsi, %rcx), %rsi
> +	and	$-0x40, %rsi
> +	sub	%rsi, %rdx
> +	sub	%rdx, %rdi
> +# ifdef USE_AS_STRNCPY
> +	lea	128(%r8, %rdx), %r8
> +# endif
> +L(Unaligned64Loop):
> +	movaps	(%rsi), %xmm2
> +	movaps	%xmm2, %xmm4
> +	movaps	16(%rsi), %xmm5
> +	movaps	32(%rsi), %xmm3
> +	movaps	%xmm3, %xmm6
> +	movaps	48(%rsi), %xmm7
> +	pminub	%xmm5, %xmm2
> +	pminub	%xmm7, %xmm3
> +	pminub	%xmm2, %xmm3
> +	pcmpeqb	%xmm0, %xmm3
> +	pmovmskb %xmm3, %rdx
> +# ifdef USE_AS_STRNCPY
> +	sub	$64, %r8
> +	jbe	L(UnalignedLeaveCase2OrCase3)
> +# endif
> +	test	%rdx, %rdx
> +	jnz	L(Unaligned64Leave)
> +
> +L(Unaligned64Loop_start):
> +	add	$64, %rdi
> +	add	$64, %rsi
> +	movdqu	%xmm4, -64(%rdi)
> +	movaps	(%rsi), %xmm2
> +	movdqa	%xmm2, %xmm4
> +	movdqu	%xmm5, -48(%rdi)
> +	movaps	16(%rsi), %xmm5
> +	pminub	%xmm5, %xmm2
> +	movaps	32(%rsi), %xmm3
> +	movdqu	%xmm6, -32(%rdi)
> +	movaps	%xmm3, %xmm6
> +	movdqu	%xmm7, -16(%rdi)
> +	movaps	48(%rsi), %xmm7
> +	pminub	%xmm7, %xmm3
> +	pminub	%xmm2, %xmm3
> +	pcmpeqb	%xmm0, %xmm3
> +	pmovmskb %xmm3, %rdx
> +# ifdef USE_AS_STRNCPY
> +	sub	$64, %r8
> +	jbe	L(UnalignedLeaveCase2OrCase3)
> +# endif
> +	test	%rdx, %rdx
> +	jz	L(Unaligned64Loop_start)
> +
> +L(Unaligned64Leave):
> +	pxor	%xmm1, %xmm1
> +
> +	pcmpeqb	%xmm4, %xmm0
> +	pcmpeqb	%xmm5, %xmm1
> +	pmovmskb %xmm0, %rdx
> +	pmovmskb %xmm1, %rcx
> +	test	%rdx, %rdx
> +	jnz	L(CopyFrom1To16BytesUnaligned_0)
> +	test	%rcx, %rcx
> +	jnz	L(CopyFrom1To16BytesUnaligned_16)
> +
> +	pcmpeqb	%xmm6, %xmm0
> +	pcmpeqb	%xmm7, %xmm1
> +	pmovmskb %xmm0, %rdx
> +	pmovmskb %xmm1, %rcx
> +	test	%rdx, %rdx
> +	jnz	L(CopyFrom1To16BytesUnaligned_32)
> +
> +	bsf	%rcx, %rdx
> +	movdqu	%xmm4, (%rdi)
> +	movdqu	%xmm5, 16(%rdi)
> +	movdqu	%xmm6, 32(%rdi)
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +# ifdef USE_AS_STPCPY
> +	lea	48(%rdi, %rdx), %rax
> +# endif
> +	movdqu	%xmm7, 48(%rdi)
> +	add	$15, %r8
> +	sub	%rdx, %r8
> +	lea	49(%rdi, %rdx), %rdi
> +	jmp	L(StrncpyFillTailWithZero)
> +# else
> +	add	$48, %rsi
> +	add	$48, %rdi
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> +# endif
> +
> +/* If source address alignment == destination address alignment */
> +
> +L(SourceStringAlignmentLess32):
> +	pxor	%xmm0, %xmm0
> +	movdqu	(%rsi), %xmm1
> +	movdqu	16(%rsi), %xmm2
> +	pcmpeqb	%xmm1, %xmm0
> +	pmovmskb %xmm0, %rdx
> +
> +# ifdef USE_AS_STRNCPY
> +#  if defined USE_AS_STPCPY || defined USE_AS_STRCAT
> +	cmp	$16, %r8
> +#  else
> +	cmp	$17, %r8
> +#  endif
> +	jbe	L(CopyFrom1To16BytesTail1Case2OrCase3)
> +# endif
> +	test	%rdx, %rdx
> +	jnz	L(CopyFrom1To16BytesTail1)
> +
> +	pcmpeqb	%xmm2, %xmm0
> +	movdqu	%xmm1, (%rdi)
> +	pmovmskb %xmm0, %rdx
> +
> +# ifdef USE_AS_STRNCPY
> +#  if defined USE_AS_STPCPY || defined USE_AS_STRCAT
> +	cmp	$32, %r8
> +#  else
> +	cmp	$33, %r8
> +#  endif
> +	jbe	L(CopyFrom1To32Bytes1Case2OrCase3)
> +# endif
> +	test	%rdx, %rdx
> +	jnz	L(CopyFrom1To32Bytes1)
> +
> +	and	$-16, %rsi
> +	and	$15, %rcx
> +	jmp	L(Unalign16Both)
> +
> +/*------End of main part with loops---------------------*/
> +
> +/* Case1 */
> +
> +# if (!defined USE_AS_STRNCPY) || (defined USE_AS_STRCAT)
> +	.p2align 4
> +L(CopyFrom1To16Bytes):
> +	add	%rcx, %rdi
> +	add	%rcx, %rsi
> +	bsf	%rdx, %rdx
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> +# endif
> +	.p2align 4
> +L(CopyFrom1To16BytesTail):
> +	add	%rcx, %rsi
> +	bsf	%rdx, %rdx
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> +
> +	.p2align 4
> +L(CopyFrom1To32Bytes1):
> +	add	$16, %rsi
> +	add	$16, %rdi
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$16, %r8
> +# endif
> +L(CopyFrom1To16BytesTail1):
> +	bsf	%rdx, %rdx
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> +
> +	.p2align 4
> +L(CopyFrom1To32Bytes):
> +	bsf	%rdx, %rdx
> +	add	%rcx, %rsi
> +	add	$16, %rdx
> +	sub	%rcx, %rdx
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> +
> +	.p2align 4
> +L(CopyFrom1To16BytesUnaligned_0):
> +	bsf	%rdx, %rdx
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +# ifdef USE_AS_STPCPY
> +	lea	(%rdi, %rdx), %rax
> +# endif
> +	movdqu	%xmm4, (%rdi)
> +	add	$63, %r8
> +	sub	%rdx, %r8
> +	lea	1(%rdi, %rdx), %rdi
> +	jmp	L(StrncpyFillTailWithZero)
> +# else
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> +# endif
> +
> +	.p2align 4
> +L(CopyFrom1To16BytesUnaligned_16):
> +	bsf	%rcx, %rdx
> +	movdqu	%xmm4, (%rdi)
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +# ifdef USE_AS_STPCPY
> +	lea	16(%rdi, %rdx), %rax
> +# endif
> +	movdqu	%xmm5, 16(%rdi)
> +	add	$47, %r8
> +	sub	%rdx, %r8
> +	lea	17(%rdi, %rdx), %rdi
> +	jmp	L(StrncpyFillTailWithZero)
> +# else
> +	add	$16, %rsi
> +	add	$16, %rdi
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> +# endif
> +
> +	.p2align 4
> +L(CopyFrom1To16BytesUnaligned_32):
> +	bsf	%rdx, %rdx
> +	movdqu	%xmm4, (%rdi)
> +	movdqu	%xmm5, 16(%rdi)
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +# ifdef USE_AS_STPCPY
> +	lea	32(%rdi, %rdx), %rax
> +# endif
> +	movdqu	%xmm6, 32(%rdi)
> +	add	$31, %r8
> +	sub	%rdx, %r8
> +	lea	33(%rdi, %rdx), %rdi
> +	jmp	L(StrncpyFillTailWithZero)
> +# else
> +	add	$32, %rsi
> +	add	$32, %rdi
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> +# endif
> +
> +# ifdef USE_AS_STRNCPY
> +#  ifndef USE_AS_STRCAT
> +	.p2align 4
> +L(CopyFrom1To16BytesUnalignedXmm6):
> +	movdqu	%xmm6, (%rdi, %rcx)
> +	jmp	L(CopyFrom1To16BytesXmmExit)
> +
> +	.p2align 4
> +L(CopyFrom1To16BytesUnalignedXmm5):
> +	movdqu	%xmm5, (%rdi, %rcx)
> +	jmp	L(CopyFrom1To16BytesXmmExit)
> +
> +	.p2align 4
> +L(CopyFrom1To16BytesUnalignedXmm4):
> +	movdqu	%xmm4, (%rdi, %rcx)
> +	jmp	L(CopyFrom1To16BytesXmmExit)
> +
> +	.p2align 4
> +L(CopyFrom1To16BytesUnalignedXmm3):
> +	movdqu	%xmm3, (%rdi, %rcx)
> +	jmp	L(CopyFrom1To16BytesXmmExit)
> +
> +	.p2align 4
> +L(CopyFrom1To16BytesUnalignedXmm1):
> +	movdqu	%xmm1, (%rdi, %rcx)
> +	jmp	L(CopyFrom1To16BytesXmmExit)
> +#  endif
> +
> +	.p2align 4
> +L(CopyFrom1To16BytesExit):
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitTable), %rdx, 4)
> +
> +/* Case2 */
> +
> +	.p2align 4
> +L(CopyFrom1To16BytesCase2):
> +	add	$16, %r8
> +	add	%rcx, %rdi
> +	add	%rcx, %rsi
> +	bsf	%rdx, %rdx
> +	cmp	%r8, %rdx
> +	jb	L(CopyFrom1To16BytesExit)
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> +
> +	.p2align 4
> +L(CopyFrom1To32BytesCase2):
> +	add	%rcx, %rsi
> +	bsf	%rdx, %rdx
> +	add	$16, %rdx
> +	sub	%rcx, %rdx
> +	cmp	%r8, %rdx
> +	jb	L(CopyFrom1To16BytesExit)
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> +
> +L(CopyFrom1To16BytesTailCase2):
> +	add	%rcx, %rsi
> +	bsf	%rdx, %rdx
> +	cmp	%r8, %rdx
> +	jb	L(CopyFrom1To16BytesExit)
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> +
> +L(CopyFrom1To16BytesTail1Case2):
> +	bsf	%rdx, %rdx
> +	cmp	%r8, %rdx
> +	jb	L(CopyFrom1To16BytesExit)
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> +
> +/* Case2 or Case3,  Case3 */
> +
> +	.p2align 4
> +L(CopyFrom1To16BytesCase2OrCase3):
> +	test	%rdx, %rdx
> +	jnz	L(CopyFrom1To16BytesCase2)
> +L(CopyFrom1To16BytesCase3):
> +	add	$16, %r8
> +	add	%rcx, %rdi
> +	add	%rcx, %rsi
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> +
> +	.p2align 4
> +L(CopyFrom1To32BytesCase2OrCase3):
> +	test	%rdx, %rdx
> +	jnz	L(CopyFrom1To32BytesCase2)
> +	add	%rcx, %rsi
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> +
> +	.p2align 4
> +L(CopyFrom1To16BytesTailCase2OrCase3):
> +	test	%rdx, %rdx
> +	jnz	L(CopyFrom1To16BytesTailCase2)
> +	add	%rcx, %rsi
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> +
> +	.p2align 4
> +L(CopyFrom1To32Bytes1Case2OrCase3):
> +	add	$16, %rdi
> +	add	$16, %rsi
> +	sub	$16, %r8
> +L(CopyFrom1To16BytesTail1Case2OrCase3):
> +	test	%rdx, %rdx
> +	jnz	L(CopyFrom1To16BytesTail1Case2)
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> +
> +# endif
> +
> +/*------------End labels regarding with copying 1-16 bytes--and 1-32 bytes----*/
> +
> +	.p2align 4
> +L(Exit1):
> +	mov	%dh, (%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$1, %r8
> +	lea	1(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit2):
> +	mov	(%rsi), %dx
> +	mov	%dx, (%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	1(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$2, %r8
> +	lea	2(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit3):
> +	mov	(%rsi), %cx
> +	mov	%cx, (%rdi)
> +	mov	%dh, 2(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	2(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$3, %r8
> +	lea	3(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit4):
> +	mov	(%rsi), %edx
> +	mov	%edx, (%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	3(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$4, %r8
> +	lea	4(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit5):
> +	mov	(%rsi), %ecx
> +	mov	%dh, 4(%rdi)
> +	mov	%ecx, (%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	4(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$5, %r8
> +	lea	5(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit6):
> +	mov	(%rsi), %ecx
> +	mov	4(%rsi), %dx
> +	mov	%ecx, (%rdi)
> +	mov	%dx, 4(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	5(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$6, %r8
> +	lea	6(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit7):
> +	mov	(%rsi), %ecx
> +	mov	3(%rsi), %edx
> +	mov	%ecx, (%rdi)
> +	mov	%edx, 3(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	6(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$7, %r8
> +	lea	7(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit8):
> +	mov	(%rsi), %rdx
> +	mov	%rdx, (%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	7(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$8, %r8
> +	lea	8(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit9):
> +	mov	(%rsi), %rcx
> +	mov	%dh, 8(%rdi)
> +	mov	%rcx, (%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	8(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$9, %r8
> +	lea	9(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit10):
> +	mov	(%rsi), %rcx
> +	mov	8(%rsi), %dx
> +	mov	%rcx, (%rdi)
> +	mov	%dx, 8(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	9(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$10, %r8
> +	lea	10(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit11):
> +	mov	(%rsi), %rcx
> +	mov	7(%rsi), %edx
> +	mov	%rcx, (%rdi)
> +	mov	%edx, 7(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	10(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$11, %r8
> +	lea	11(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit12):
> +	mov	(%rsi), %rcx
> +	mov	8(%rsi), %edx
> +	mov	%rcx, (%rdi)
> +	mov	%edx, 8(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	11(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$12, %r8
> +	lea	12(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit13):
> +	mov	(%rsi), %rcx
> +	mov	5(%rsi), %rdx
> +	mov	%rcx, (%rdi)
> +	mov	%rdx, 5(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	12(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$13, %r8
> +	lea	13(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit14):
> +	mov	(%rsi), %rcx
> +	mov	6(%rsi), %rdx
> +	mov	%rcx, (%rdi)
> +	mov	%rdx, 6(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	13(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$14, %r8
> +	lea	14(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit15):
> +	mov	(%rsi), %rcx
> +	mov	7(%rsi), %rdx
> +	mov	%rcx, (%rdi)
> +	mov	%rdx, 7(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	14(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$15, %r8
> +	lea	15(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit16):
> +	movdqu	(%rsi), %xmm0
> +	movdqu	%xmm0, (%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	15(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$16, %r8
> +	lea	16(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit17):
> +	movdqu	(%rsi), %xmm0
> +	movdqu	%xmm0, (%rdi)
> +	mov	%dh, 16(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	16(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$17, %r8
> +	lea	17(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit18):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %cx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%cx, 16(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	17(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$18, %r8
> +	lea	18(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit19):
> +	movdqu	(%rsi), %xmm0
> +	mov	15(%rsi), %ecx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%ecx, 15(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	18(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$19, %r8
> +	lea	19(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit20):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %ecx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%ecx, 16(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	19(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$20, %r8
> +	lea	20(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit21):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %ecx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%ecx, 16(%rdi)
> +	mov	%dh, 20(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	20(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$21, %r8
> +	lea	21(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit22):
> +	movdqu	(%rsi), %xmm0
> +	mov	14(%rsi), %rcx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rcx, 14(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	21(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$22, %r8
> +	lea	22(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit23):
> +	movdqu	(%rsi), %xmm0
> +	mov	15(%rsi), %rcx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rcx, 15(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	22(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$23, %r8
> +	lea	23(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit24):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %rcx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rcx, 16(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	23(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$24, %r8
> +	lea	24(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit25):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %rcx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rcx, 16(%rdi)
> +	mov	%dh, 24(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	24(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$25, %r8
> +	lea	25(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit26):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %rdx
> +	mov	24(%rsi), %cx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rdx, 16(%rdi)
> +	mov	%cx, 24(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	25(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$26, %r8
> +	lea	26(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit27):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %rdx
> +	mov	23(%rsi), %ecx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rdx, 16(%rdi)
> +	mov	%ecx, 23(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	26(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$27, %r8
> +	lea	27(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit28):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %rdx
> +	mov	24(%rsi), %ecx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rdx, 16(%rdi)
> +	mov	%ecx, 24(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	27(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$28, %r8
> +	lea	28(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit29):
> +	movdqu	(%rsi), %xmm0
> +	movdqu	13(%rsi), %xmm2
> +	movdqu	%xmm0, (%rdi)
> +	movdqu	%xmm2, 13(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	28(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$29, %r8
> +	lea	29(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit30):
> +	movdqu	(%rsi), %xmm0
> +	movdqu	14(%rsi), %xmm2
> +	movdqu	%xmm0, (%rdi)
> +	movdqu	%xmm2, 14(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	29(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$30, %r8
> +	lea	30(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit31):
> +	movdqu	(%rsi), %xmm0
> +	movdqu	15(%rsi), %xmm2
> +	movdqu	%xmm0, (%rdi)
> +	movdqu	%xmm2, 15(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	30(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$31, %r8
> +	lea	31(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +	.p2align 4
> +L(Exit32):
> +	movdqu	(%rsi), %xmm0
> +	movdqu	16(%rsi), %xmm2
> +	movdqu	%xmm0, (%rdi)
> +	movdqu	%xmm2, 16(%rdi)
> +# ifdef USE_AS_STPCPY
> +	lea	31(%rdi), %rax
> +# endif
> +# if defined USE_AS_STRNCPY && !defined USE_AS_STRCAT
> +	sub	$32, %r8
> +	lea	32(%rdi), %rdi
> +	jnz	L(StrncpyFillTailWithZero)
> +# endif
> +	ret
> +
> +# ifdef USE_AS_STRNCPY
> +
> +	.p2align 4
> +L(StrncpyExit0):
> +#  ifdef USE_AS_STPCPY
> +	mov	%rdi, %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, (%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit1):
> +	mov	(%rsi), %dl
> +	mov	%dl, (%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	1(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 1(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit2):
> +	mov	(%rsi), %dx
> +	mov	%dx, (%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	2(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 2(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit3):
> +	mov	(%rsi), %cx
> +	mov	2(%rsi), %dl
> +	mov	%cx, (%rdi)
> +	mov	%dl, 2(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	3(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 3(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit4):
> +	mov	(%rsi), %edx
> +	mov	%edx, (%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	4(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 4(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit5):
> +	mov	(%rsi), %ecx
> +	mov	4(%rsi), %dl
> +	mov	%ecx, (%rdi)
> +	mov	%dl, 4(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	5(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 5(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit6):
> +	mov	(%rsi), %ecx
> +	mov	4(%rsi), %dx
> +	mov	%ecx, (%rdi)
> +	mov	%dx, 4(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	6(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 6(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit7):
> +	mov	(%rsi), %ecx
> +	mov	3(%rsi), %edx
> +	mov	%ecx, (%rdi)
> +	mov	%edx, 3(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	7(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 7(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit8):
> +	mov	(%rsi), %rdx
> +	mov	%rdx, (%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	8(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 8(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit9):
> +	mov	(%rsi), %rcx
> +	mov	8(%rsi), %dl
> +	mov	%rcx, (%rdi)
> +	mov	%dl, 8(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	9(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 9(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit10):
> +	mov	(%rsi), %rcx
> +	mov	8(%rsi), %dx
> +	mov	%rcx, (%rdi)
> +	mov	%dx, 8(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	10(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 10(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit11):
> +	mov	(%rsi), %rcx
> +	mov	7(%rsi), %edx
> +	mov	%rcx, (%rdi)
> +	mov	%edx, 7(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	11(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 11(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit12):
> +	mov	(%rsi), %rcx
> +	mov	8(%rsi), %edx
> +	mov	%rcx, (%rdi)
> +	mov	%edx, 8(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	12(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 12(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit13):
> +	mov	(%rsi), %rcx
> +	mov	5(%rsi), %rdx
> +	mov	%rcx, (%rdi)
> +	mov	%rdx, 5(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	13(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 13(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit14):
> +	mov	(%rsi), %rcx
> +	mov	6(%rsi), %rdx
> +	mov	%rcx, (%rdi)
> +	mov	%rdx, 6(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	14(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 14(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit15):
> +	mov	(%rsi), %rcx
> +	mov	7(%rsi), %rdx
> +	mov	%rcx, (%rdi)
> +	mov	%rdx, 7(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	15(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 15(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit16):
> +	movdqu	(%rsi), %xmm0
> +	movdqu	%xmm0, (%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	16(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 16(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit17):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %cl
> +	movdqu	%xmm0, (%rdi)
> +	mov	%cl, 16(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	17(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 17(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit18):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %cx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%cx, 16(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	18(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 18(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit19):
> +	movdqu	(%rsi), %xmm0
> +	mov	15(%rsi), %ecx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%ecx, 15(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	19(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 19(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit20):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %ecx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%ecx, 16(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	20(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 20(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit21):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %ecx
> +	mov	20(%rsi), %dl
> +	movdqu	%xmm0, (%rdi)
> +	mov	%ecx, 16(%rdi)
> +	mov	%dl, 20(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	21(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 21(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit22):
> +	movdqu	(%rsi), %xmm0
> +	mov	14(%rsi), %rcx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rcx, 14(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	22(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 22(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit23):
> +	movdqu	(%rsi), %xmm0
> +	mov	15(%rsi), %rcx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rcx, 15(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	23(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 23(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit24):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %rcx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rcx, 16(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	24(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 24(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit25):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %rdx
> +	mov	24(%rsi), %cl
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rdx, 16(%rdi)
> +	mov	%cl, 24(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	25(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 25(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit26):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %rdx
> +	mov	24(%rsi), %cx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rdx, 16(%rdi)
> +	mov	%cx, 24(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	26(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 26(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit27):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %rdx
> +	mov	23(%rsi), %ecx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rdx, 16(%rdi)
> +	mov	%ecx, 23(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	27(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 27(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit28):
> +	movdqu	(%rsi), %xmm0
> +	mov	16(%rsi), %rdx
> +	mov	24(%rsi), %ecx
> +	movdqu	%xmm0, (%rdi)
> +	mov	%rdx, 16(%rdi)
> +	mov	%ecx, 24(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	28(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 28(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit29):
> +	movdqu	(%rsi), %xmm0
> +	movdqu	13(%rsi), %xmm2
> +	movdqu	%xmm0, (%rdi)
> +	movdqu	%xmm2, 13(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	29(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 29(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit30):
> +	movdqu	(%rsi), %xmm0
> +	movdqu	14(%rsi), %xmm2
> +	movdqu	%xmm0, (%rdi)
> +	movdqu	%xmm2, 14(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	30(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 30(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit31):
> +	movdqu	(%rsi), %xmm0
> +	movdqu	15(%rsi), %xmm2
> +	movdqu	%xmm0, (%rdi)
> +	movdqu	%xmm2, 15(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	31(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 31(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit32):
> +	movdqu	(%rsi), %xmm0
> +	movdqu	16(%rsi), %xmm2
> +	movdqu	%xmm0, (%rdi)
> +	movdqu	%xmm2, 16(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	32(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 32(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(StrncpyExit33):
> +	movdqu	(%rsi), %xmm0
> +	movdqu	16(%rsi), %xmm2
> +	mov	32(%rsi), %cl
> +	movdqu	%xmm0, (%rdi)
> +	movdqu	%xmm2, 16(%rdi)
> +	mov	%cl, 32(%rdi)
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 33(%rdi)
> +#  endif
> +	ret
> +
> +#  ifndef USE_AS_STRCAT
> +
> +	.p2align 4
> +L(Fill0):
> +	ret
> +
> +	.p2align 4
> +L(Fill1):
> +	mov	%dl, (%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill2):
> +	mov	%dx, (%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill3):
> +	mov	%edx, -1(%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill4):
> +	mov	%edx, (%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill5):
> +	mov	%edx, (%rdi)
> +	mov	%dl, 4(%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill6):
> +	mov	%edx, (%rdi)
> +	mov	%dx, 4(%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill7):
> +	mov	%rdx, -1(%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill8):
> +	mov	%rdx, (%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill9):
> +	mov	%rdx, (%rdi)
> +	mov	%dl, 8(%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill10):
> +	mov	%rdx, (%rdi)
> +	mov	%dx, 8(%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill11):
> +	mov	%rdx, (%rdi)
> +	mov	%edx, 7(%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill12):
> +	mov	%rdx, (%rdi)
> +	mov	%edx, 8(%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill13):
> +	mov	%rdx, (%rdi)
> +	mov	%rdx, 5(%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill14):
> +	mov	%rdx, (%rdi)
> +	mov	%rdx, 6(%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill15):
> +	movdqu	%xmm0, -1(%rdi)
> +	ret
> +
> +	.p2align 4
> +L(Fill16):
> +	movdqu	%xmm0, (%rdi)
> +	ret
> +
> +	.p2align 4
> +L(CopyFrom1To16BytesUnalignedXmm2):
> +	movdqu	%xmm2, (%rdi, %rcx)
> +
> +	.p2align 4
> +L(CopyFrom1To16BytesXmmExit):
> +	bsf	%rdx, %rdx
> +	add	$15, %r8
> +	add	%rcx, %rdi
> +#   ifdef USE_AS_STPCPY
> +	lea	(%rdi, %rdx), %rax
> +#   endif
> +	sub	%rdx, %r8
> +	lea	1(%rdi, %rdx), %rdi
> +
> +	.p2align 4
> +L(StrncpyFillTailWithZero):
> +	pxor	%xmm0, %xmm0
> +	xor	%rdx, %rdx
> +	sub	$16, %r8
> +	jbe	L(StrncpyFillExit)
> +
> +	movdqu	%xmm0, (%rdi)
> +	add	$16, %rdi
> +
> +	mov	%rdi, %rsi
> +	and	$0xf, %rsi
> +	sub	%rsi, %rdi
> +	add	%rsi, %r8
> +	sub	$64, %r8
> +	jb	L(StrncpyFillLess64)
> +
> +L(StrncpyFillLoopMovdqa):
> +	movdqa	%xmm0, (%rdi)
> +	movdqa	%xmm0, 16(%rdi)
> +	movdqa	%xmm0, 32(%rdi)
> +	movdqa	%xmm0, 48(%rdi)
> +	add	$64, %rdi
> +	sub	$64, %r8
> +	jae	L(StrncpyFillLoopMovdqa)
> +
> +L(StrncpyFillLess64):
> +	add	$32, %r8
> +	jl	L(StrncpyFillLess32)
> +	movdqa	%xmm0, (%rdi)
> +	movdqa	%xmm0, 16(%rdi)
> +	add	$32, %rdi
> +	sub	$16, %r8
> +	jl	L(StrncpyFillExit)
> +	movdqa	%xmm0, (%rdi)
> +	add	$16, %rdi
> +	BRANCH_TO_JMPTBL_ENTRY (L(FillTable), %r8, 4)
> +
> +L(StrncpyFillLess32):
> +	add	$16, %r8
> +	jl	L(StrncpyFillExit)
> +	movdqa	%xmm0, (%rdi)
> +	add	$16, %rdi
> +	BRANCH_TO_JMPTBL_ENTRY (L(FillTable), %r8, 4)
> +
> +L(StrncpyFillExit):
> +	add	$16, %r8
> +	BRANCH_TO_JMPTBL_ENTRY (L(FillTable), %r8, 4)
> +
> +/* end of ifndef USE_AS_STRCAT */
> +#  endif
> +
> +	.p2align 4
> +L(UnalignedLeaveCase2OrCase3):
> +	test	%rdx, %rdx
> +	jnz	L(Unaligned64LeaveCase2)
> +L(Unaligned64LeaveCase3):
> +	lea	64(%r8), %rcx
> +	and	$-16, %rcx
> +	add	$48, %r8
> +	jl	L(CopyFrom1To16BytesCase3)
> +	movdqu	%xmm4, (%rdi)
> +	sub	$16, %r8
> +	jb	L(CopyFrom1To16BytesCase3)
> +	movdqu	%xmm5, 16(%rdi)
> +	sub	$16, %r8
> +	jb	L(CopyFrom1To16BytesCase3)
> +	movdqu	%xmm6, 32(%rdi)
> +	sub	$16, %r8
> +	jb	L(CopyFrom1To16BytesCase3)
> +	movdqu	%xmm7, 48(%rdi)
> +#  ifdef USE_AS_STPCPY
> +	lea	64(%rdi), %rax
> +#  endif
> +#  ifdef USE_AS_STRCAT
> +	xor	%ch, %ch
> +	movb	%ch, 64(%rdi)
> +#  endif
> +	ret
> +
> +	.p2align 4
> +L(Unaligned64LeaveCase2):
> +	xor	%rcx, %rcx
> +	pcmpeqb	%xmm4, %xmm0
> +	pmovmskb %xmm0, %rdx
> +	add	$48, %r8
> +	jle	L(CopyFrom1To16BytesCase2OrCase3)
> +	test	%rdx, %rdx
> +#  ifndef USE_AS_STRCAT
> +	jnz	L(CopyFrom1To16BytesUnalignedXmm4)
> +#  else
> +	jnz	L(CopyFrom1To16Bytes)
> +#  endif
> +	pcmpeqb	%xmm5, %xmm0
> +	pmovmskb %xmm0, %rdx
> +	movdqu	%xmm4, (%rdi)
> +	add	$16, %rcx
> +	sub	$16, %r8
> +	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> +	test	%rdx, %rdx
> +#  ifndef USE_AS_STRCAT
> +	jnz	L(CopyFrom1To16BytesUnalignedXmm5)
> +#  else
> +	jnz	L(CopyFrom1To16Bytes)
> +#  endif
> +
> +	pcmpeqb	%xmm6, %xmm0
> +	pmovmskb %xmm0, %rdx
> +	movdqu	%xmm5, 16(%rdi)
> +	add	$16, %rcx
> +	sub	$16, %r8
> +	jbe	L(CopyFrom1To16BytesCase2OrCase3)
> +	test	%rdx, %rdx
> +#  ifndef USE_AS_STRCAT
> +	jnz	L(CopyFrom1To16BytesUnalignedXmm6)
> +#  else
> +	jnz	L(CopyFrom1To16Bytes)
> +#  endif
> +
> +	pcmpeqb	%xmm7, %xmm0
> +	pmovmskb %xmm0, %rdx
> +	movdqu	%xmm6, 32(%rdi)
> +	lea	16(%rdi, %rcx), %rdi
> +	lea	16(%rsi, %rcx), %rsi
> +	bsf	%rdx, %rdx
> +	cmp	%r8, %rdx
> +	jb	L(CopyFrom1To16BytesExit)
> +	BRANCH_TO_JMPTBL_ENTRY (L(ExitStrncpyTable), %r8, 4)
> +
> +	.p2align 4
> +L(ExitZero):
> +#  ifndef USE_AS_STRCAT
> +	mov	%rdi, %rax
> +#  endif
> +	ret
> +
> +# endif
> +
> +# ifndef USE_AS_STRCAT
> +END (STRCPY)
> +# else
> +END (STRCAT)
> +# endif
> +	.p2align 4
> +	.section .rodata
> +L(ExitTable):
> +	.int	JMPTBL(L(Exit1), L(ExitTable))
> +	.int	JMPTBL(L(Exit2), L(ExitTable))
> +	.int	JMPTBL(L(Exit3), L(ExitTable))
> +	.int	JMPTBL(L(Exit4), L(ExitTable))
> +	.int	JMPTBL(L(Exit5), L(ExitTable))
> +	.int	JMPTBL(L(Exit6), L(ExitTable))
> +	.int	JMPTBL(L(Exit7), L(ExitTable))
> +	.int	JMPTBL(L(Exit8), L(ExitTable))
> +	.int	JMPTBL(L(Exit9), L(ExitTable))
> +	.int	JMPTBL(L(Exit10), L(ExitTable))
> +	.int	JMPTBL(L(Exit11), L(ExitTable))
> +	.int	JMPTBL(L(Exit12), L(ExitTable))
> +	.int	JMPTBL(L(Exit13), L(ExitTable))
> +	.int	JMPTBL(L(Exit14), L(ExitTable))
> +	.int	JMPTBL(L(Exit15), L(ExitTable))
> +	.int	JMPTBL(L(Exit16), L(ExitTable))
> +	.int	JMPTBL(L(Exit17), L(ExitTable))
> +	.int	JMPTBL(L(Exit18), L(ExitTable))
> +	.int	JMPTBL(L(Exit19), L(ExitTable))
> +	.int	JMPTBL(L(Exit20), L(ExitTable))
> +	.int	JMPTBL(L(Exit21), L(ExitTable))
> +	.int	JMPTBL(L(Exit22), L(ExitTable))
> +	.int    JMPTBL(L(Exit23), L(ExitTable))
> +	.int	JMPTBL(L(Exit24), L(ExitTable))
> +	.int	JMPTBL(L(Exit25), L(ExitTable))
> +	.int	JMPTBL(L(Exit26), L(ExitTable))
> +	.int	JMPTBL(L(Exit27), L(ExitTable))
> +	.int	JMPTBL(L(Exit28), L(ExitTable))
> +	.int	JMPTBL(L(Exit29), L(ExitTable))
> +	.int	JMPTBL(L(Exit30), L(ExitTable))
> +	.int	JMPTBL(L(Exit31), L(ExitTable))
> +	.int	JMPTBL(L(Exit32), L(ExitTable))
> +# ifdef USE_AS_STRNCPY
> +L(ExitStrncpyTable):
> +	.int	JMPTBL(L(StrncpyExit0), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit1), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit2), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit3), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit4), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit5), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit6), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit7), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit8), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit9), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit10), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit11), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit12), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit13), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit14), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit15), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit16), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit17), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit18), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit19), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit20), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit21), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit22), L(ExitStrncpyTable))
> +	.int    JMPTBL(L(StrncpyExit23), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit24), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit25), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit26), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit27), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit28), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit29), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit30), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit31), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit32), L(ExitStrncpyTable))
> +	.int	JMPTBL(L(StrncpyExit33), L(ExitStrncpyTable))
> +#  ifndef USE_AS_STRCAT
> +	.p2align 4
> +L(FillTable):
> +	.int	JMPTBL(L(Fill0), L(FillTable))
> +	.int	JMPTBL(L(Fill1), L(FillTable))
> +	.int	JMPTBL(L(Fill2), L(FillTable))
> +	.int	JMPTBL(L(Fill3), L(FillTable))
> +	.int	JMPTBL(L(Fill4), L(FillTable))
> +	.int	JMPTBL(L(Fill5), L(FillTable))
> +	.int	JMPTBL(L(Fill6), L(FillTable))
> +	.int	JMPTBL(L(Fill7), L(FillTable))
> +	.int	JMPTBL(L(Fill8), L(FillTable))
> +	.int	JMPTBL(L(Fill9), L(FillTable))
> +	.int	JMPTBL(L(Fill10), L(FillTable))
> +	.int	JMPTBL(L(Fill11), L(FillTable))
> +	.int	JMPTBL(L(Fill12), L(FillTable))
> +	.int	JMPTBL(L(Fill13), L(FillTable))
> +	.int	JMPTBL(L(Fill14), L(FillTable))
> +	.int	JMPTBL(L(Fill15), L(FillTable))
> +	.int	JMPTBL(L(Fill16), L(FillTable))
> +#  endif
> +# endif
> +#endif
> diff --git a/sysdeps/x86_64/multiarch/strncpy.S b/sysdeps/x86_64/multiarch/strncpy.S
> index 6d87a0b..afbd870 100644
> --- a/sysdeps/x86_64/multiarch/strncpy.S
> +++ b/sysdeps/x86_64/multiarch/strncpy.S
> @@ -1,5 +1,85 @@
> -/* Multiple versions of strncpy
> -   All versions must be listed in ifunc-impl-list.c.  */
> -#define STRCPY strncpy
> +/* Multiple versions of strcpy
> +   All versions must be listed in ifunc-impl-list.c.
> +   Copyright (C) 2009-2015 Free Software Foundation, Inc.
> +   Contributed by Intel Corporation.
> +   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>
> +#include <init-arch.h>
> +
>  #define USE_AS_STRNCPY
> -#include "strcpy.S"
> +#ifndef STRNCPY
> +#define STRNCPY strncpy
> +#endif
> +
> +#ifdef USE_AS_STPCPY
> +#  define STRNCPY_SSSE3		__stpncpy_ssse3
> +#  define STRNCPY_SSE2		__stpncpy_sse2
> +#  define STRNCPY_SSE2_UNALIGNED __stpncpy_sse2_unaligned
> +#  define __GI_STRNCPY		__GI_stpncpy
> +#  define __GI___STRNCPY		__GI___stpncpy
> +#else
> +#  define STRNCPY_SSSE3		__strncpy_ssse3
> +#  define STRNCPY_SSE2		__strncpy_sse2
> +#  define STRNCPY_SSE2_UNALIGNED	__strncpy_sse2_unaligned
> +#  define __GI_STRNCPY		__GI_strncpy
> +#endif
> +
> +
> +/* Define multiple versions only for the definition in libc.  */
> +#if IS_IN (libc)
> +	.text
> +ENTRY(STRNCPY)
> +	.type	STRNCPY, @gnu_indirect_function
> +	cmpl	$0, __cpu_features+KIND_OFFSET(%rip)
> +	jne	1f
> +	call	__init_cpu_features
> +1:	leaq	STRNCPY_SSE2_UNALIGNED(%rip), %rax
> +	testl	$bit_Fast_Unaligned_Load, __cpu_features+FEATURE_OFFSET+index_Fast_Unaligned_Load(%rip)
> +	jnz	2f
> +	leaq	STRNCPY_SSE2(%rip), %rax
> +	testl	$bit_SSSE3, __cpu_features+CPUID_OFFSET+index_SSSE3(%rip)
> +	jz	2f
> +	leaq	STRNCPY_SSSE3(%rip), %rax
> +2:	ret
> +END(STRNCPY)
> +
> +# undef ENTRY
> +# define ENTRY(name) \
> +	.type STRNCPY_SSE2, @function; \
> +	.align 16; \
> +	.globl STRNCPY_SSE2; \
> +	.hidden STRNCPY_SSE2; \
> +	STRNCPY_SSE2: cfi_startproc; \
> +	CALL_MCOUNT
> +# undef END
> +# define END(name) \
> +	cfi_endproc; .size STRNCPY_SSE2, .-STRNCPY_SSE2
> +# undef libc_hidden_builtin_def
> +/* It doesn't make sense to send libc-internal strcpy calls through a PLT.
> +   The speedup we get from using SSSE3 instruction is likely eaten away
> +   by the indirect call in the PLT.  */
> +# define libc_hidden_builtin_def(name) \
> +	.globl __GI_STRNCPY; __GI_STRNCPY = STRNCPY_SSE2
> +# undef libc_hidden_def
> +# define libc_hidden_def(name) \
> +	.globl __GI___STRNCPY; __GI___STRNCPY = STRNCPY_SSE2
> +#endif
> +
> +#ifndef USE_AS_STRNCPY
> +#include "../strcpy.S"
> +#endif
> -- 
> 1.8.4.rc3

-- 

Communications satellite used by the military for star wars.
References:
- [PATCH neleai/string-x64] Improve strcpy sse2 and avx2 implementation
  - From: OndÅej BÃlka
Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]