Convert aligned moves to unaligned

Most if not all cpus sold in the past couple of years should see no performance drop from this, and it fixes segfaults in actual user code
This commit is contained in:
Martin Kroeker 2020-04-13 14:52:22 +02:00 committed by GitHub
parent 20d0cb2f65
commit c0c35da831
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 650 additions and 0 deletions

650
kernel/x86_64/copy_sse2.c Normal file
View File

@ -0,0 +1,650 @@
/*********************************************************************/
/* Copyright 2009, 2010 The University of Texas at Austin. */
/* All rights reserved. */
/* */
/* Redistribution and use in source and binary forms, with or */
/* without modification, are permitted provided that the following */
/* conditions are met: */
/* */
/* 1. Redistributions of source code must retain the above */
/* copyright notice, this list of conditions and the following */
/* disclaimer. */
/* */
/* 2. Redistributions in binary form must reproduce the above */
/* copyright notice, this list of conditions and the following */
/* disclaimer in the documentation and/or other materials */
/* provided with the distribution. */
/* */
/* THIS SOFTWARE IS PROVIDED BY THE UNIVERSITY OF TEXAS AT */
/* AUSTIN ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, */
/* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF */
/* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE */
/* DISCLAIMED. IN NO EVENT SHALL THE UNIVERSITY OF TEXAS AT */
/* AUSTIN OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, */
/* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES */
/* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE */
/* GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR */
/* BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF */
/* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT */
/* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT */
/* OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE */
/* POSSIBILITY OF SUCH DAMAGE. */
/* */
/* The views and conclusions contained in the software and */
/* documentation are those of the authors and should not be */
/* interpreted as representing official policies, either expressed */
/* or implied, of The University of Texas at Austin. */
/*********************************************************************/
#define ASSEMBLER
#include "common.h"
#define M ARG1 /* rdi */
#define X ARG2 /* rsi */
#define INCX ARG3 /* rdx */
#define Y ARG4 /* rcx */
#ifndef WINDOWS_ABI
#define INCY ARG5 /* r8 */
#else
#define INCY %r10
#endif
#include "l1param.h"
#ifdef OPTERON
#define LOAD(OFFSET, ADDR, REG) xorps REG, REG; addpd OFFSET(ADDR), REG
#else
#define LOAD(OFFSET, ADDR, REG) movups OFFSET(ADDR), REG
#endif
PROLOGUE
PROFCODE
#ifdef WINDOWS_ABI
movq 40(%rsp), INCY
#endif
SAVEREGISTERS
leaq (, INCX, SIZE), INCX
leaq (, INCY, SIZE), INCY
cmpq $SIZE, INCX
jne .L40
cmpq $SIZE, INCY
jne .L40
#ifdef ALIGNED_ACCESS
testq $SIZE, Y
#else
testq $SIZE, X
#endif
je .L10
movsd (X), %xmm0
movsd %xmm0, (Y)
addq $1 * SIZE, X
addq $1 * SIZE, Y
decq M
jle .L19
ALIGN_4
.L10:
subq $-16 * SIZE, X
subq $-16 * SIZE, Y
#ifdef ALIGNED_ACCESS
testq $SIZE, X
#else
testq $SIZE, Y
#endif
jne .L20
movq M, %rax
sarq $4, %rax
jle .L13
movups -16 * SIZE(X), %xmm0
movups -14 * SIZE(X), %xmm1
movups -12 * SIZE(X), %xmm2
movups -10 * SIZE(X), %xmm3
movups -8 * SIZE(X), %xmm4
movups -6 * SIZE(X), %xmm5
movups -4 * SIZE(X), %xmm6
movups -2 * SIZE(X), %xmm7
decq %rax
jle .L12
ALIGN_3
.L11:
#ifdef PREFETCHW
PREFETCHW (PREFETCHSIZE + 0) - PREOFFSET(Y)
#endif
movups %xmm0, -16 * SIZE(Y)
LOAD( 0 * SIZE, X, %xmm0)
movups %xmm1, -14 * SIZE(Y)
LOAD( 2 * SIZE, X, %xmm1)
#ifdef PREFETCH
PREFETCH (PREFETCHSIZE + 0) - PREOFFSET(X)
#endif
movups %xmm2, -12 * SIZE(Y)
LOAD( 4 * SIZE, X, %xmm2)
movups %xmm3, -10 * SIZE(Y)
LOAD( 6 * SIZE, X, %xmm3)
#if defined(PREFETCHW) && !defined(FETCH128)
PREFETCHW (PREFETCHSIZE + 64) - PREOFFSET(Y)
#endif
movups %xmm4, -8 * SIZE(Y)
LOAD( 8 * SIZE, X, %xmm4)
movups %xmm5, -6 * SIZE(Y)
LOAD(10 * SIZE, X, %xmm5)
#if defined(PREFETCH) && !defined(FETCH128)
PREFETCH (PREFETCHSIZE + 64) - PREOFFSET(X)
#endif
movups %xmm6, -4 * SIZE(Y)
LOAD(12 * SIZE, X, %xmm6)
movups %xmm7, -2 * SIZE(Y)
LOAD(14 * SIZE, X, %xmm7)
subq $-16 * SIZE, Y
subq $-16 * SIZE, X
decq %rax
jg .L11
ALIGN_3
.L12:
movups %xmm0, -16 * SIZE(Y)
movups %xmm1, -14 * SIZE(Y)
movups %xmm2, -12 * SIZE(Y)
movups %xmm3, -10 * SIZE(Y)
movups %xmm4, -8 * SIZE(Y)
movups %xmm5, -6 * SIZE(Y)
movups %xmm6, -4 * SIZE(Y)
movups %xmm7, -2 * SIZE(Y)
subq $-16 * SIZE, Y
subq $-16 * SIZE, X
ALIGN_3
.L13:
testq $8, M
jle .L14
ALIGN_3
movups -16 * SIZE(X), %xmm0
movups -14 * SIZE(X), %xmm1
movups -12 * SIZE(X), %xmm2
movups -10 * SIZE(X), %xmm3
movups %xmm0, -16 * SIZE(Y)
movups %xmm1, -14 * SIZE(Y)
movups %xmm2, -12 * SIZE(Y)
movups %xmm3, -10 * SIZE(Y)
addq $8 * SIZE, X
addq $8 * SIZE, Y
ALIGN_3
.L14:
testq $4, M
jle .L15
ALIGN_3
movups -16 * SIZE(X), %xmm0
movups -14 * SIZE(X), %xmm1
movups %xmm0, -16 * SIZE(Y)
movups %xmm1, -14 * SIZE(Y)
addq $4 * SIZE, X
addq $4 * SIZE, Y
ALIGN_3
.L15:
testq $2, M
jle .L16
ALIGN_3
movups -16 * SIZE(X), %xmm0
movups %xmm0, -16 * SIZE(Y)
addq $2 * SIZE, X
addq $2 * SIZE, Y
ALIGN_3
.L16:
testq $1, M
jle .L19
ALIGN_3
movsd -16 * SIZE(X), %xmm0
movsd %xmm0, -16 * SIZE(Y)
ALIGN_3
.L19:
xorq %rax,%rax
RESTOREREGISTERS
ret
ALIGN_3
.L20:
#ifdef ALIGNED_ACCESS
movhps -16 * SIZE(X), %xmm0
movq M, %rax
sarq $4, %rax
jle .L23
movups -15 * SIZE(X), %xmm1
movups -13 * SIZE(X), %xmm2
movups -11 * SIZE(X), %xmm3
movups -9 * SIZE(X), %xmm4
movups -7 * SIZE(X), %xmm5
movups -5 * SIZE(X), %xmm6
movups -3 * SIZE(X), %xmm7
decq %rax
jle .L22
ALIGN_4
.L21:
#ifdef PREFETCHW
PREFETCHW (PREFETCHSIZE + 0) - PREOFFSET(Y)
#endif
SHUFPD_1 %xmm1, %xmm0
movups %xmm0, -16 * SIZE(Y)
LOAD(-1 * SIZE, X, %xmm0)
SHUFPD_1 %xmm2, %xmm1
movups %xmm1, -14 * SIZE(Y)
LOAD( 1 * SIZE, X, %xmm1)
#ifdef PREFETCH
PREFETCH (PREFETCHSIZE + 0) - PREOFFSET(X)
#endif
SHUFPD_1 %xmm3, %xmm2
movups %xmm2, -12 * SIZE(Y)
LOAD( 3 * SIZE, X, %xmm2)
SHUFPD_1 %xmm4, %xmm3
movups %xmm3, -10 * SIZE(Y)
LOAD( 5 * SIZE, X, %xmm3)
#if defined(PREFETCHW) && !defined(FETCH128)
PREFETCHW (PREFETCHSIZE + 64) - PREOFFSET(Y)
#endif
SHUFPD_1 %xmm5, %xmm4
movups %xmm4, -8 * SIZE(Y)
LOAD( 7 * SIZE, X, %xmm4)
SHUFPD_1 %xmm6, %xmm5
movups %xmm5, -6 * SIZE(Y)
LOAD( 9 * SIZE, X, %xmm5)
#if defined(PREFETCH) && !defined(FETCH128)
PREFETCH (PREFETCHSIZE + 64) - PREOFFSET(X)
#endif
SHUFPD_1 %xmm7, %xmm6
movups %xmm6, -4 * SIZE(Y)
LOAD(11 * SIZE, X, %xmm6)
SHUFPD_1 %xmm0, %xmm7
movups %xmm7, -2 * SIZE(Y)
LOAD(13 * SIZE, X, %xmm7)
subq $-16 * SIZE, X
subq $-16 * SIZE, Y
decq %rax
jg .L21
ALIGN_3
.L22:
SHUFPD_1 %xmm1, %xmm0
movups %xmm0, -16 * SIZE(Y)
LOAD(-1 * SIZE, X, %xmm0)
SHUFPD_1 %xmm2, %xmm1
movups %xmm1, -14 * SIZE(Y)
SHUFPD_1 %xmm3, %xmm2
movups %xmm2, -12 * SIZE(Y)
SHUFPD_1 %xmm4, %xmm3
movups %xmm3, -10 * SIZE(Y)
SHUFPD_1 %xmm5, %xmm4
movups %xmm4, -8 * SIZE(Y)
SHUFPD_1 %xmm6, %xmm5
movups %xmm5, -6 * SIZE(Y)
SHUFPD_1 %xmm7, %xmm6
movups %xmm6, -4 * SIZE(Y)
SHUFPD_1 %xmm0, %xmm7
movups %xmm7, -2 * SIZE(Y)
subq $-16 * SIZE, X
subq $-16 * SIZE, Y
ALIGN_3
.L23:
testq $8, M
jle .L24
ALIGN_3
movups -15 * SIZE(X), %xmm1
movups -13 * SIZE(X), %xmm2
movups -11 * SIZE(X), %xmm3
movups -9 * SIZE(X), %xmm8
SHUFPD_1 %xmm1, %xmm0
movups %xmm0, -16 * SIZE(Y)
SHUFPD_1 %xmm2, %xmm1
movups %xmm1, -14 * SIZE(Y)
SHUFPD_1 %xmm3, %xmm2
movups %xmm2, -12 * SIZE(Y)
SHUFPD_1 %xmm8, %xmm3
movups %xmm3, -10 * SIZE(Y)
movups %xmm8, %xmm0
addq $8 * SIZE, X
addq $8 * SIZE, Y
ALIGN_3
.L24:
testq $4, M
jle .L25
ALIGN_3
movups -15 * SIZE(X), %xmm1
movups -13 * SIZE(X), %xmm2
SHUFPD_1 %xmm1, %xmm0
SHUFPD_1 %xmm2, %xmm1
movups %xmm0, -16 * SIZE(Y)
movups %xmm1, -14 * SIZE(Y)
movups %xmm2, %xmm0
addq $4 * SIZE, X
addq $4 * SIZE, Y
ALIGN_3
.L25:
testq $2, M
jle .L26
ALIGN_3
movups -15 * SIZE(X), %xmm1
SHUFPD_1 %xmm1, %xmm0
movups %xmm0, -16 * SIZE(Y)
addq $2 * SIZE, X
addq $2 * SIZE, Y
ALIGN_3
.L26:
testq $1, M
jle .L29
ALIGN_3
movsd -16 * SIZE(X), %xmm0
movsd %xmm0, -16 * SIZE(Y)
ALIGN_3
.L29:
xorq %rax,%rax
RESTOREREGISTERS
ret
ALIGN_3
#else
movq M, %rax
sarq $4, %rax
jle .L23
movups -16 * SIZE(X), %xmm0
movups -14 * SIZE(X), %xmm1
movups -12 * SIZE(X), %xmm2
movups -10 * SIZE(X), %xmm3
movups -8 * SIZE(X), %xmm4
movups -6 * SIZE(X), %xmm5
movups -4 * SIZE(X), %xmm6
movups -2 * SIZE(X), %xmm7
decq %rax
jle .L22
ALIGN_3
.L21:
#ifdef PREFETCHW
PREFETCHW (PREFETCHSIZE + 0) - PREOFFSET(Y)
#endif
movlps %xmm0, -16 * SIZE(Y)
movhps %xmm0, -15 * SIZE(Y)
LOAD( 0 * SIZE, X, %xmm0)
movlps %xmm1, -14 * SIZE(Y)
movhps %xmm1, -13 * SIZE(Y)
LOAD( 2 * SIZE, X, %xmm1)
#ifdef PREFETCH
PREFETCH (PREFETCHSIZE + 0) - PREOFFSET(X)
#endif
movlps %xmm2, -12 * SIZE(Y)
movhps %xmm2, -11 * SIZE(Y)
LOAD( 4 * SIZE, X, %xmm2)
movlps %xmm3, -10 * SIZE(Y)
movhps %xmm3, -9 * SIZE(Y)
LOAD( 6 * SIZE, X, %xmm3)
#if defined(PREFETCHW) && !defined(FETCH128)
PREFETCHW (PREFETCHSIZE + 64) - PREOFFSET(Y)
#endif
movlps %xmm4, -8 * SIZE(Y)
movhps %xmm4, -7 * SIZE(Y)
LOAD( 8 * SIZE, X, %xmm4)
movlps %xmm5, -6 * SIZE(Y)
movhps %xmm5, -5 * SIZE(Y)
LOAD(10 * SIZE, X, %xmm5)
#if defined(PREFETCH) && !defined(FETCH128)
PREFETCH (PREFETCHSIZE + 64) - PREOFFSET(X)
#endif
movlps %xmm6, -4 * SIZE(Y)
movhps %xmm6, -3 * SIZE(Y)
LOAD(12 * SIZE, X, %xmm6)
movlps %xmm7, -2 * SIZE(Y)
movhps %xmm7, -1 * SIZE(Y)
LOAD(14 * SIZE, X, %xmm7)
subq $-16 * SIZE, Y
subq $-16 * SIZE, X
decq %rax
jg .L21
ALIGN_3
.L22:
movlps %xmm0, -16 * SIZE(Y)
movhps %xmm0, -15 * SIZE(Y)
movlps %xmm1, -14 * SIZE(Y)
movhps %xmm1, -13 * SIZE(Y)
movlps %xmm2, -12 * SIZE(Y)
movhps %xmm2, -11 * SIZE(Y)
movlps %xmm3, -10 * SIZE(Y)
movhps %xmm3, -9 * SIZE(Y)
movlps %xmm4, -8 * SIZE(Y)
movhps %xmm4, -7 * SIZE(Y)
movlps %xmm5, -6 * SIZE(Y)
movhps %xmm5, -5 * SIZE(Y)
movlps %xmm6, -4 * SIZE(Y)
movhps %xmm6, -3 * SIZE(Y)
movlps %xmm7, -2 * SIZE(Y)
movhps %xmm7, -1 * SIZE(Y)
subq $-16 * SIZE, Y
subq $-16 * SIZE, X
ALIGN_3
.L23:
testq $8, M
jle .L24
ALIGN_3
movups -16 * SIZE(X), %xmm0
movlps %xmm0, -16 * SIZE(Y)
movhps %xmm0, -15 * SIZE(Y)
movups -14 * SIZE(X), %xmm1
movlps %xmm1, -14 * SIZE(Y)
movhps %xmm1, -13 * SIZE(Y)
movups -12 * SIZE(X), %xmm2
movlps %xmm2, -12 * SIZE(Y)
movhps %xmm2, -11 * SIZE(Y)
movups -10 * SIZE(X), %xmm3
movlps %xmm3, -10 * SIZE(Y)
movhps %xmm3, -9 * SIZE(Y)
addq $8 * SIZE, X
addq $8 * SIZE, Y
ALIGN_3
.L24:
testq $4, M
jle .L25
ALIGN_3
movups -16 * SIZE(X), %xmm0
movlps %xmm0, -16 * SIZE(Y)
movhps %xmm0, -15 * SIZE(Y)
movups -14 * SIZE(X), %xmm1
movlps %xmm1, -14 * SIZE(Y)
movhps %xmm1, -13 * SIZE(Y)
addq $4 * SIZE, X
addq $4 * SIZE, Y
ALIGN_3
.L25:
testq $2, M
jle .L26
ALIGN_3
movups -16 * SIZE(X), %xmm0
movlps %xmm0, -16 * SIZE(Y)
movhps %xmm0, -15 * SIZE(Y)
addq $2 * SIZE, X
addq $2 * SIZE, Y
ALIGN_3
.L26:
testq $1, M
jle .L29
ALIGN_3
movsd -16 * SIZE(X), %xmm0
movsd %xmm0, -16 * SIZE(Y)
ALIGN_3
.L29:
xorq %rax,%rax
RESTOREREGISTERS
ret
ALIGN_3
#endif
.L40:
movq M, %rax
sarq $3, %rax
jle .L45
ALIGN_3
.L41:
movsd (X), %xmm0
addq INCX, X
movhps (X), %xmm0
addq INCX, X
movsd (X), %xmm1
addq INCX, X
movhps (X), %xmm1
addq INCX, X
movsd (X), %xmm2
addq INCX, X
movhps (X), %xmm2
addq INCX, X
movsd (X), %xmm3
addq INCX, X
movhps (X), %xmm3
addq INCX, X
movlps %xmm0, (Y)
addq INCY, Y
movhps %xmm0, (Y)
addq INCY, Y
movlps %xmm1, (Y)
addq INCY, Y
movhps %xmm1, (Y)
addq INCY, Y
movlps %xmm2, (Y)
addq INCY, Y
movhps %xmm2, (Y)
addq INCY, Y
movlps %xmm3, (Y)
addq INCY, Y
movhps %xmm3, (Y)
addq INCY, Y
decq %rax
jg .L41
ALIGN_3
.L45:
movq M, %rax
andq $7, %rax
jle .L47
ALIGN_3
.L46:
movsd (X), %xmm0
addq INCX, X
movlps %xmm0, (Y)
addq INCY, Y
decq %rax
jg .L46
ALIGN_3
.L47:
xorq %rax, %rax
RESTOREREGISTERS
ret
EPILOGUE