/* Copyright (C) 2011-2018 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>

#ifdef USE_AS_WMEMCHR

# define MEMCHR		wmemchr

# define PCMPEQ		pcmpeqd

#else

# define MEMCHR		memchr

# define PCMPEQ		pcmpeqb

#endif

/* fast SSE2 version with using pmaxub and 64 byte loop */

	.text

ENTRY(MEMCHR)

	movd	%esi, %xmm1

	mov	%edi, %ecx

#ifdef USE_AS_WMEMCHR

	test	%rdx, %rdx

	jz	L(return_null)

	shl	$2, %rdx

#else

	punpcklbw %xmm1, %xmm1

	test	%rdx, %rdx

	jz	L(return_null)

	punpcklbw %xmm1, %xmm1

#endif

	and	$63, %ecx

	pshufd	$0, %xmm1, %xmm1

	cmp	$48, %ecx

	ja	L(crosscache)

	movdqu	(%rdi), %xmm0

	PCMPEQ	%xmm1, %xmm0

	pmovmskb %xmm0, %eax

	test	%eax, %eax

	jnz	L(matches_1)

	sub	$16, %rdx

	jbe	L(return_null)

	add	$16, %rdi

	and	$15, %ecx

	and	$-16, %rdi

	add	%rcx, %rdx

	sub	$64, %rdx

	jbe	L(exit_loop)

	jmp	L(loop_prolog)

	.p2align 4

L(crosscache):

	and	$15, %ecx

	and	$-16, %rdi

	movdqa	(%rdi), %xmm0

	PCMPEQ	%xmm1, %xmm0

/* Check if there is a match.  */

	pmovmskb %xmm0, %eax

/* Remove the leading bytes.  */

	sar	%cl, %eax

	test	%eax, %eax

	je	L(unaligned_no_match)

/* Check which byte is a match.  */

	bsf	%eax, %eax

	sub	%rax, %rdx

	jbe	L(return_null)

	add	%rdi, %rax

	add	%rcx, %rax

	ret

	.p2align 4

L(unaligned_no_match):

        /* "rcx" is less than 16.  Calculate "rdx + rcx - 16" by using

	   "rdx - (16 - rcx)" instead of "(rdx + rcx) - 16" to void

	   possible addition overflow.  */

	neg	%rcx

	add	$16, %rcx

	sub	%rcx, %rdx

	jbe	L(return_null)

	add	$16, %rdi

	sub	$64, %rdx

	jbe	L(exit_loop)

	.p2align 4

L(loop_prolog):

	movdqa	(%rdi), %xmm0

	PCMPEQ	%xmm1, %xmm0

	pmovmskb %xmm0, %eax

	test	%eax, %eax

	jnz	L(matches)

	movdqa	16(%rdi), %xmm2

	PCMPEQ	%xmm1, %xmm2

	pmovmskb %xmm2, %eax

	test	%eax, %eax

	jnz	L(matches16)

	movdqa	32(%rdi), %xmm3

	PCMPEQ	%xmm1, %xmm3

	pmovmskb %xmm3, %eax

	test	%eax, %eax

	jnz	L(matches32)

	movdqa	48(%rdi), %xmm4

	PCMPEQ	%xmm1, %xmm4

	add	$64, %rdi

	pmovmskb %xmm4, %eax

	test	%eax, %eax

	jnz	L(matches0)

	test	$0x3f, %rdi

	jz	L(align64_loop)

	sub	$64, %rdx

	jbe	L(exit_loop)

	movdqa	(%rdi), %xmm0

	PCMPEQ	%xmm1, %xmm0

	pmovmskb %xmm0, %eax

	test	%eax, %eax

	jnz	L(matches)

	movdqa	16(%rdi), %xmm2

	PCMPEQ	%xmm1, %xmm2

	pmovmskb %xmm2, %eax

	test	%eax, %eax

	jnz	L(matches16)

	movdqa	32(%rdi), %xmm3

	PCMPEQ	%xmm1, %xmm3

	pmovmskb %xmm3, %eax

	test	%eax, %eax

	jnz	L(matches32)

	movdqa	48(%rdi), %xmm3

	PCMPEQ	%xmm1, %xmm3

	pmovmskb %xmm3, %eax

	add	$64, %rdi

	test	%eax, %eax

	jnz	L(matches0)

	mov	%rdi, %rcx

	and	$-64, %rdi

	and	$63, %ecx

	add	%rcx, %rdx

	.p2align 4

L(align64_loop):

	sub	$64, %rdx

	jbe	L(exit_loop)

	movdqa	(%rdi), %xmm0

	movdqa	16(%rdi), %xmm2

	movdqa	32(%rdi), %xmm3

	movdqa	48(%rdi), %xmm4

	PCMPEQ	%xmm1, %xmm0

	PCMPEQ	%xmm1, %xmm2

	PCMPEQ	%xmm1, %xmm3

	PCMPEQ	%xmm1, %xmm4

	pmaxub	%xmm0, %xmm3

	pmaxub	%xmm2, %xmm4

	pmaxub	%xmm3, %xmm4

	pmovmskb %xmm4, %eax

	add	$64, %rdi

	test	%eax, %eax

	jz	L(align64_loop)

	sub	$64, %rdi

	pmovmskb %xmm0, %eax

	test	%eax, %eax

	jnz	L(matches)

	pmovmskb %xmm2, %eax

	test	%eax, %eax

	jnz	L(matches16)

	movdqa	32(%rdi), %xmm3

	PCMPEQ	%xmm1, %xmm3

	PCMPEQ	48(%rdi), %xmm1

	pmovmskb %xmm3, %eax

	test	%eax, %eax

	jnz	L(matches32)

	pmovmskb %xmm1, %eax

	bsf	%eax, %eax

	lea	48(%rdi, %rax), %rax

	ret

	.p2align 4

L(exit_loop):

	add	$32, %edx

	jle	L(exit_loop_32)

	movdqa	(%rdi), %xmm0

	PCMPEQ	%xmm1, %xmm0

	pmovmskb %xmm0, %eax

	test	%eax, %eax

	jnz	L(matches)

	movdqa	16(%rdi), %xmm2

	PCMPEQ	%xmm1, %xmm2

	pmovmskb %xmm2, %eax

	test	%eax, %eax

	jnz	L(matches16)

	movdqa	32(%rdi), %xmm3

	PCMPEQ	%xmm1, %xmm3

	pmovmskb %xmm3, %eax

	test	%eax, %eax

	jnz	L(matches32_1)

	sub	$16, %edx

	jle	L(return_null)

	PCMPEQ	48(%rdi), %xmm1

	pmovmskb %xmm1, %eax

	test	%eax, %eax

	jnz	L(matches48_1)

	xor	%eax, %eax

	ret

	.p2align 4

L(exit_loop_32):

	add	$32, %edx

	movdqa	(%rdi), %xmm0

	PCMPEQ	%xmm1, %xmm0

	pmovmskb %xmm0, %eax

	test	%eax, %eax

	jnz	L(matches_1)

	sub	$16, %edx

	jbe	L(return_null)

	PCMPEQ	16(%rdi), %xmm1

	pmovmskb %xmm1, %eax

	test	%eax, %eax

	jnz	L(matches16_1)

	xor	%eax, %eax

	ret

	.p2align 4

L(matches0):

	bsf	%eax, %eax

	lea	-16(%rax, %rdi), %rax

	ret

	.p2align 4

L(matches):

	bsf	%eax, %eax

	add	%rdi, %rax

	ret

	.p2align 4

L(matches16):

	bsf	%eax, %eax

	lea	16(%rax, %rdi), %rax

	ret

	.p2align 4

L(matches32):

	bsf	%eax, %eax

	lea	32(%rax, %rdi), %rax

	ret

	.p2align 4

L(matches_1):

	bsf	%eax, %eax

	sub	%rax, %rdx

	jbe	L(return_null)

	add	%rdi, %rax

	ret

	.p2align 4

L(matches16_1):

	bsf	%eax, %eax

	sub	%rax, %rdx

	jbe	L(return_null)

	lea	16(%rdi, %rax), %rax

	ret

	.p2align 4

L(matches32_1):

	bsf	%eax, %eax

	sub	%rax, %rdx

	jbe	L(return_null)

	lea	32(%rdi, %rax), %rax

	ret

	.p2align 4

L(matches48_1):

	bsf	%eax, %eax

	sub	%rax, %rdx

	jbe	L(return_null)

	lea	48(%rdi, %rax), %rax

	ret

	.p2align 4

L(return_null):

	xor	%eax, %eax

	ret

END(MEMCHR)

#ifndef USE_AS_WMEMCHR

strong_alias (memchr, __memchr)

libc_hidden_builtin_def(memchr)

#endif