openkylin
/
linux
mirror of https://gitee.com/openkylin/linux.git
9
0
Fork 0
Code Issues Projects Releases Wiki Activity
linux/arch/x86/lib/copy_user_nocache_64.S

139 lines
2.4 KiB
ArmAsm
Raw Normal View History

x86: fix copy_user on x86 Switch copy_user_generic_string(), copy_user_generic_unrolled() and __copy_user_nocache() from custom tail handlers to generic copy_user_tail_handle(). Signed-off-by: Vitaly Mayatskikh <v.mayatskih@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-02 21:53:13 +08:00
/*
* Copyright 2008 Vitaly Mayatskikh <vmayatsk@redhat.com>
* Copyright 2002 Andi Kleen, SuSE Labs.
[PATCH] x86-64: Add __copy_from_user_nocache This does user copies in fs write() into the page cache with write combining. This pushes the destination out of the CPU's cache, but allows higher bandwidth in some case. The theory is that the page cache data is usually not touched by the CPU again and it's better to not pollute the cache with it. Also it is a little faster. Signed-off-by: Andi Kleen <ak@suse.de>
2007-02-13 20:26:19 +08:00
* Subject to the GNU Public License v2.
*
* Functions to copy from and to user space.
*/
#include <linux/linkage.h>
#include <asm/dwarf2.h>
#define FIX_ALIGNMENT 1
#include <asm/current.h>
#include <asm/asm-offsets.h>
#include <asm/thread_info.h>
x86: fix copy_user on x86 Switch copy_user_generic_string(), copy_user_generic_unrolled() and __copy_user_nocache() from custom tail handlers to generic copy_user_tail_handle(). Signed-off-by: Vitaly Mayatskikh <v.mayatskih@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-02 21:53:13 +08:00
.macro ALIGN_DESTINATION
[PATCH] x86-64: Add __copy_from_user_nocache This does user copies in fs write() into the page cache with write combining. This pushes the destination out of the CPU's cache, but allows higher bandwidth in some case. The theory is that the page cache data is usually not touched by the CPU again and it's better to not pollute the cache with it. Also it is a little faster. Signed-off-by: Andi Kleen <ak@suse.de>
2007-02-13 20:26:19 +08:00
#ifdef FIX_ALIGNMENT
/* check for bad alignment of destination */
movl %edi,%ecx
andl $7,%ecx
x86: fix copy_user on x86 Switch copy_user_generic_string(), copy_user_generic_unrolled() and __copy_user_nocache() from custom tail handlers to generic copy_user_tail_handle(). Signed-off-by: Vitaly Mayatskikh <v.mayatskih@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-02 21:53:13 +08:00
jz 102f /* already aligned */
subl $8,%ecx
negl %ecx
subl %ecx,%edx
100: movb (%rsi),%al
101: movb %al,(%rdi)
incq %rsi
incq %rdi
decl %ecx
jnz 100b
102:
.section .fixup,"ax"
103: addl %r8d,%edx /* ecx is zerorest also */
jmp copy_user_handle_tail
.previous
[PATCH] x86-64: Add __copy_from_user_nocache This does user copies in fs write() into the page cache with write combining. This pushes the destination out of the CPU's cache, but allows higher bandwidth in some case. The theory is that the page cache data is usually not touched by the CPU again and it's better to not pollute the cache with it. Also it is a little faster. Signed-off-by: Andi Kleen <ak@suse.de>
2007-02-13 20:26:19 +08:00
x86: fix copy_user on x86 Switch copy_user_generic_string(), copy_user_generic_unrolled() and __copy_user_nocache() from custom tail handlers to generic copy_user_tail_handle(). Signed-off-by: Vitaly Mayatskikh <v.mayatskih@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-02 21:53:13 +08:00
.section __ex_table,"a"
.align 8
.quad 100b,103b
.quad 101b,103b
.previous
#endif
.endm
[PATCH] x86-64: Add __copy_from_user_nocache This does user copies in fs write() into the page cache with write combining. This pushes the destination out of the CPU's cache, but allows higher bandwidth in some case. The theory is that the page cache data is usually not touched by the CPU again and it's better to not pollute the cache with it. Also it is a little faster. Signed-off-by: Andi Kleen <ak@suse.de>
2007-02-13 20:26:19 +08:00
x86: fix copy_user on x86 Switch copy_user_generic_string(), copy_user_generic_unrolled() and __copy_user_nocache() from custom tail handlers to generic copy_user_tail_handle(). Signed-off-by: Vitaly Mayatskikh <v.mayatskih@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-02 21:53:13 +08:00
/*
* copy_user_nocache - Uncached memory copy with exception handling
* This will force destination/source out of cache for more performance.
*/
ENTRY(__copy_user_nocache)
CFI_STARTPROC
cmpl $8,%edx
jb 20f /* less then 8 bytes, go to byte copy loop */
ALIGN_DESTINATION
movl %edx,%ecx
andl $63,%edx
shrl $6,%ecx
jz 17f
1: movq (%rsi),%r8
2: movq 1*8(%rsi),%r9
3: movq 2*8(%rsi),%r10
4: movq 3*8(%rsi),%r11
5: movnti %r8,(%rdi)
6: movnti %r9,1*8(%rdi)
7: movnti %r10,2*8(%rdi)
8: movnti %r11,3*8(%rdi)
9: movq 4*8(%rsi),%r8
10: movq 5*8(%rsi),%r9
11: movq 6*8(%rsi),%r10
12: movq 7*8(%rsi),%r11
13: movnti %r8,4*8(%rdi)
14: movnti %r9,5*8(%rdi)
15: movnti %r10,6*8(%rdi)
16: movnti %r11,7*8(%rdi)
[PATCH] x86-64: Add __copy_from_user_nocache This does user copies in fs write() into the page cache with write combining. This pushes the destination out of the CPU's cache, but allows higher bandwidth in some case. The theory is that the page cache data is usually not touched by the CPU again and it's better to not pollute the cache with it. Also it is a little faster. Signed-off-by: Andi Kleen <ak@suse.de>
2007-02-13 20:26:19 +08:00
leaq 64(%rsi),%rsi
leaq 64(%rdi),%rdi
decl %ecx
x86: fix copy_user on x86 Switch copy_user_generic_string(), copy_user_generic_unrolled() and __copy_user_nocache() from custom tail handlers to generic copy_user_tail_handle(). Signed-off-by: Vitaly Mayatskikh <v.mayatskih@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-02 21:53:13 +08:00
jnz 1b
17: movl %edx,%ecx
andl $7,%edx
shrl $3,%ecx
jz 20f
18: movq (%rsi),%r8
19: movnti %r8,(%rdi)
[PATCH] x86-64: Add __copy_from_user_nocache This does user copies in fs write() into the page cache with write combining. This pushes the destination out of the CPU's cache, but allows higher bandwidth in some case. The theory is that the page cache data is usually not touched by the CPU again and it's better to not pollute the cache with it. Also it is a little faster. Signed-off-by: Andi Kleen <ak@suse.de>
2007-02-13 20:26:19 +08:00
leaq 8(%rsi),%rsi
x86: fix copy_user on x86 Switch copy_user_generic_string(), copy_user_generic_unrolled() and __copy_user_nocache() from custom tail handlers to generic copy_user_tail_handle(). Signed-off-by: Vitaly Mayatskikh <v.mayatskih@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-02 21:53:13 +08:00
leaq 8(%rdi),%rdi
decl %ecx
jnz 18b
20: andl %edx,%edx
jz 23f
[PATCH] x86-64: Add __copy_from_user_nocache This does user copies in fs write() into the page cache with write combining. This pushes the destination out of the CPU's cache, but allows higher bandwidth in some case. The theory is that the page cache data is usually not touched by the CPU again and it's better to not pollute the cache with it. Also it is a little faster. Signed-off-by: Andi Kleen <ak@suse.de>
2007-02-13 20:26:19 +08:00
movl %edx,%ecx
x86: fix copy_user on x86 Switch copy_user_generic_string(), copy_user_generic_unrolled() and __copy_user_nocache() from custom tail handlers to generic copy_user_tail_handle(). Signed-off-by: Vitaly Mayatskikh <v.mayatskih@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-02 21:53:13 +08:00
21: movb (%rsi),%al
22: movb %al,(%rdi)
[PATCH] x86-64: Add __copy_from_user_nocache This does user copies in fs write() into the page cache with write combining. This pushes the destination out of the CPU's cache, but allows higher bandwidth in some case. The theory is that the page cache data is usually not touched by the CPU again and it's better to not pollute the cache with it. Also it is a little faster. Signed-off-by: Andi Kleen <ak@suse.de>
2007-02-13 20:26:19 +08:00
incq %rsi
x86: fix copy_user on x86 Switch copy_user_generic_string(), copy_user_generic_unrolled() and __copy_user_nocache() from custom tail handlers to generic copy_user_tail_handle(). Signed-off-by: Vitaly Mayatskikh <v.mayatskih@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-02 21:53:13 +08:00
incq %rdi
[PATCH] x86-64: Add __copy_from_user_nocache This does user copies in fs write() into the page cache with write combining. This pushes the destination out of the CPU's cache, but allows higher bandwidth in some case. The theory is that the page cache data is usually not touched by the CPU again and it's better to not pollute the cache with it. Also it is a little faster. Signed-off-by: Andi Kleen <ak@suse.de>
2007-02-13 20:26:19 +08:00
decl %ecx
x86: fix copy_user on x86 Switch copy_user_generic_string(), copy_user_generic_unrolled() and __copy_user_nocache() from custom tail handlers to generic copy_user_tail_handle(). Signed-off-by: Vitaly Mayatskikh <v.mayatskih@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-02 21:53:13 +08:00
jnz 21b
23: xorl %eax,%eax
x86: fence oostores on 64-bit movnt* instructions are not strongly ordered with respect to other stores, so if we are to assume stores are strongly ordered in the rest of the 64 bit code, we must fence these off (see similar examples in 32 bit code). [ The AMD memory ordering document seems to say that nontemporal stores can also pass earlier regular stores, so maybe we need sfences _before_ movnt* everywhere too? ] Signed-off-by: Nick Piggin <npiggin@suse.de> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-13 09:06:00 +08:00
sfence
[PATCH] x86-64: Add __copy_from_user_nocache This does user copies in fs write() into the page cache with write combining. This pushes the destination out of the CPU's cache, but allows higher bandwidth in some case. The theory is that the page cache data is usually not touched by the CPU again and it's better to not pollute the cache with it. Also it is a little faster. Signed-off-by: Andi Kleen <ak@suse.de>
2007-02-13 20:26:19 +08:00
ret
x86: fix copy_user on x86 Switch copy_user_generic_string(), copy_user_generic_unrolled() and __copy_user_nocache() from custom tail handlers to generic copy_user_tail_handle(). Signed-off-by: Vitaly Mayatskikh <v.mayatskih@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-02 21:53:13 +08:00
.section .fixup,"ax"
30: shll $6,%ecx
addl %ecx,%edx
jmp 60f
x86: fix compile error in current tip.git Gas 2.15 complains about 32-bit registers being used in lea. AS arch/x86/lib/copy_user_64.o /local/scratch-2/jeremy/hg/xen/paravirt/linux/arch/x86/lib/copy_user_64.S: Assembler messages: /local/scratch-2/jeremy/hg/xen/paravirt/linux/arch/x86/lib/copy_user_64.S:188: Error: `(%edx,%ecx,8)' is not a valid 64 bit base/index expression /local/scratch-2/jeremy/hg/xen/paravirt/linux/arch/x86/lib/copy_user_64.S:257: Error: `(%edx,%ecx,8)' is not a valid 64 bit base/index expression AS arch/x86/lib/copy_user_nocache_64.o /local/scratch-2/jeremy/hg/xen/paravirt/linux/arch/x86/lib/copy_user_nocache_64.S: Assembler messages: /local/scratch-2/jeremy/hg/xen/paravirt/linux/arch/x86/lib/copy_user_nocache_64.S:107: Error: `(%edx,%ecx,8)' is not a valid 64 bit base/index expression Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: Vitaly Mayatskikh <v.mayatskih@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-11 03:52:52 +08:00
40: lea (%rdx,%rcx,8),%rdx
x86: fix copy_user on x86 Switch copy_user_generic_string(), copy_user_generic_unrolled() and __copy_user_nocache() from custom tail handlers to generic copy_user_tail_handle(). Signed-off-by: Vitaly Mayatskikh <v.mayatskih@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-02 21:53:13 +08:00
jmp 60f
50: movl %ecx,%edx
60: sfence
movl %r8d,%ecx
jmp copy_user_handle_tail
.previous
[PATCH] x86-64: Add __copy_from_user_nocache This does user copies in fs write() into the page cache with write combining. This pushes the destination out of the CPU's cache, but allows higher bandwidth in some case. The theory is that the page cache data is usually not touched by the CPU again and it's better to not pollute the cache with it. Also it is a little faster. Signed-off-by: Andi Kleen <ak@suse.de>
2007-02-13 20:26:19 +08:00
.section __ex_table,"a"
x86: fix copy_user on x86 Switch copy_user_generic_string(), copy_user_generic_unrolled() and __copy_user_nocache() from custom tail handlers to generic copy_user_tail_handle(). Signed-off-by: Vitaly Mayatskikh <v.mayatskih@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-02 21:53:13 +08:00
.quad 1b,30b
.quad 2b,30b
.quad 3b,30b
.quad 4b,30b
.quad 5b,30b
.quad 6b,30b
.quad 7b,30b
.quad 8b,30b
.quad 9b,30b
.quad 10b,30b
.quad 11b,30b
.quad 12b,30b
.quad 13b,30b
.quad 14b,30b
.quad 15b,30b
.quad 16b,30b
.quad 18b,40b
.quad 19b,40b
.quad 21b,50b
.quad 22b,50b
[PATCH] x86-64: Add __copy_from_user_nocache This does user copies in fs write() into the page cache with write combining. This pushes the destination out of the CPU's cache, but allows higher bandwidth in some case. The theory is that the page cache data is usually not touched by the CPU again and it's better to not pollute the cache with it. Also it is a little faster. Signed-off-by: Andi Kleen <ak@suse.de>
2007-02-13 20:26:19 +08:00
.previous
CFI_ENDPROC
ENDPROC(__copy_user_nocache)