Optimized wcschr and wcscpy for x86-64 and x86-32

1 files changed, 220 insertions, 0 deletions

diff --git a/sysdeps/i386/i686/multiarch/wcschr-sse2.S b/sysdeps/i386/i686/multiarch/wcschr-sse2.S
new file mode 100644
index 0000000000..cc8204cfe3
--- /dev/null
+++ b/sysdeps/i386/i686/multiarch/wcschr-sse2.S
@@ -0,0 +1,220 @@
+/* wcschr with SSE2, without using bsf instructions
+   Copyright (C) 2011 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, write to the Free
+   Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+   02111-1307 USA.  */
+
+#ifndef NOT_IN_libc
+# include <sysdep.h>
+
+# define CFI_PUSH(REG)	\
+	cfi_adjust_cfa_offset (4);	\
+	cfi_rel_offset (REG, 0)
+
+# define CFI_POP(REG)	\
+	cfi_adjust_cfa_offset (-4);	\
+	cfi_restore (REG)
+
+# define PUSH(REG)	pushl REG; CFI_PUSH (REG)
+# define POP(REG)	popl REG; CFI_POP (REG)
+
+# define PARMS	4
+# define STR1	PARMS
+# define STR2	STR1+4
+
+	atom_text_section
+ENTRY (__wcschr_sse2)
+
+	mov	STR1(%esp), %ecx
+	movd	STR2(%esp), %xmm1
+
+	mov	%ecx, %eax
+	punpckldq %xmm1, %xmm1
+	pxor	%xmm2, %xmm2
+	punpckldq %xmm1, %xmm1
+
+	and	$63, %eax
+	cmp	$48, %eax
+	ja	L(cross_cache)
+
+	movdqu	(%ecx), %xmm0
+	pcmpeqd	%xmm0, %xmm2
+	pcmpeqd	%xmm1, %xmm0
+	pmovmskb %xmm2, %edx
+	pmovmskb %xmm0, %eax
+	or	%eax, %edx
+	jnz	L(matches)
+	and	$-16, %ecx
+	jmp	L(loop)
+
+	.p2align 4
+L(cross_cache):
+	PUSH	(%edi)
+	mov	%ecx, %edi
+	mov	%eax, %ecx
+	and	$-16, %edi
+	and	$15, %ecx
+	movdqa	(%edi), %xmm0
+	pcmpeqd	%xmm0, %xmm2
+	pcmpeqd	%xmm1, %xmm0
+	pmovmskb %xmm2, %edx
+	pmovmskb %xmm0, %eax
+
+	sarl	%cl, %edx
+	sarl	%cl, %eax
+	test	%eax, %eax
+	jz	L(unaligned_no_match)
+
+	add	%edi, %ecx
+	POP	(%edi)
+
+	test	%edx, %edx
+	jz	L(match_case1)
+	test	%al, %al
+	jz	L(match_higth_case2)
+	test	$15, %al
+	jnz	L(match_case2_4)
+	test	$15, %dl
+	jnz	L(return_null)
+	lea	4(%ecx), %eax
+	ret
+
+	CFI_PUSH (%edi)
+
+	.p2align 4
+L(unaligned_no_match):
+	mov	%edi, %ecx
+	POP	(%edi)
+
+	test	%edx, %edx
+	jnz	L(return_null)
+
+	pxor	%xmm2, %xmm2
+
+/* Loop start on aligned string.  */
+	.p2align 4
+L(loop):
+	add	$16, %ecx
+	movdqa	(%ecx), %xmm0
+	pcmpeqd	%xmm0, %xmm2
+	pcmpeqd	%xmm1, %xmm0
+	pmovmskb %xmm2, %edx
+	pmovmskb %xmm0, %eax
+	or	%eax, %edx
+	jnz	L(matches)
+	add	$16, %ecx
+
+	movdqa	(%ecx), %xmm0
+	pcmpeqd	%xmm0, %xmm2
+	pcmpeqd	%xmm1, %xmm0
+	pmovmskb %xmm2, %edx
+	pmovmskb %xmm0, %eax
+	or	%eax, %edx
+	jnz	L(matches)
+	add	$16, %ecx
+
+	movdqa	(%ecx), %xmm0
+	pcmpeqd	%xmm0, %xmm2
+	pcmpeqd	%xmm1, %xmm0
+	pmovmskb %xmm2, %edx
+	pmovmskb %xmm0, %eax
+	or	%eax, %edx
+	jnz	L(matches)
+	add	$16, %ecx
+
+	movdqa	(%ecx), %xmm0
+	pcmpeqd	%xmm0, %xmm2
+	pcmpeqd	%xmm1, %xmm0
+	pmovmskb %xmm2, %edx
+	pmovmskb %xmm0, %eax
+	or	%eax, %edx
+	jz	L(loop)
+
+	.p2align 4
+L(matches):
+	pmovmskb %xmm2, %edx
+	test	%eax, %eax
+	jz	L(return_null)
+	test	%edx, %edx
+	jz	L(match_case1)
+
+	.p2align 4
+L(match_case2):
+	test	%al, %al
+	jz	L(match_higth_case2)
+	test	$15, %al
+	jnz	L(match_case2_4)
+	test	$15, %dl
+	jnz	L(return_null)
+	lea	4(%ecx), %eax
+	ret
+
+	.p2align 4
+L(match_case2_4):
+	mov	%ecx, %eax
+	ret
+
+	.p2align 4
+L(match_higth_case2):
+	test	%dl, %dl
+	jnz	L(return_null)
+	test	$15, %ah
+	jnz	L(match_case2_12)
+	test	$15, %dh
+	jnz	L(return_null)
+	lea	12(%ecx), %eax
+	ret
+
+	.p2align 4
+L(match_case2_12):
+	lea	8(%ecx), %eax
+	ret
+
+	.p2align 4
+L(match_case1):
+	test	%al, %al
+	jz	L(match_higth_case1)
+
+	test	$0x01, %al
+	jnz	L(exit0)
+	lea	4(%ecx), %eax
+	ret
+
+	.p2align 4
+L(match_higth_case1):
+	test	$0x01, %ah
+	jnz	L(exit3)
+	lea	12(%ecx), %eax
+	ret
+
+	.p2align 4
+L(exit0):
+	mov	%ecx, %eax
+	ret
+
+	.p2align 4
+L(exit3):
+	lea	8(%ecx), %eax
+	ret
+
+	.p2align 4
+L(return_null):
+	xor	%eax, %eax
+	ret
+
+END (__wcschr_sse2)
+#endif