AuxXxilium/linux_dsm_epyc7002
1
0
Fork 0
mirror of https://github.com/AuxXxilium/linux_dsm_epyc7002.git synced 2024-12-23 01:24:47 +07:00
Code Issues Actions Packages Projects Releases Wiki Activity
6d97d7a28f
linux_dsm_epyc7002/arch/powerpc/lib/copypage_64.S

113 lines
2.1 KiB
ArmAsm
Raw Normal View History

powerpc: Merge enough to start building in arch/powerpc. This creates the directory structure under arch/powerpc and a bunch of Kconfig files. It does a first-cut merge of arch/powerpc/mm, arch/powerpc/lib and arch/powerpc/platforms/powermac. This is enough to build a 32-bit powermac kernel with ARCH=powerpc. For now we are getting some unmerged files from arch/ppc/kernel and arch/ppc/syslib, or arch/ppc64/kernel. This makes some minor changes to files in those directories and files outside arch/powerpc. The boot directory is still not merged. That's going to be interesting. Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-09-26 13:04:21 +07:00
/*
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
* Copyright (C) 2008 Mark Nelson, IBM Corp.
powerpc: Merge enough to start building in arch/powerpc. This creates the directory structure under arch/powerpc and a bunch of Kconfig files. It does a first-cut merge of arch/powerpc/mm, arch/powerpc/lib and arch/powerpc/platforms/powermac. This is enough to build a 32-bit powermac kernel with ARCH=powerpc. For now we are getting some unmerged files from arch/ppc/kernel and arch/ppc/syslib, or arch/ppc64/kernel. This makes some minor changes to files in those directories and files outside arch/powerpc. The boot directory is still not merged. That's going to be interesting. Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-09-26 13:04:21 +07:00
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*/
powerpc: Simplify 4k/64k copy_page logic To make it easier to add optimised versions of copy_page, remove the 4kB loop for 64kB pages and just do all the work in copy_page. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2011-05-09 04:18:38 +07:00
#include <asm/page.h>
powerpc: Merge enough to start building in arch/powerpc. This creates the directory structure under arch/powerpc and a bunch of Kconfig files. It does a first-cut merge of arch/powerpc/mm, arch/powerpc/lib and arch/powerpc/platforms/powermac. This is enough to build a 32-bit powermac kernel with ARCH=powerpc. For now we are getting some unmerged files from arch/ppc/kernel and arch/ppc/syslib, or arch/ppc64/kernel. This makes some minor changes to files in those directories and files outside arch/powerpc. The boot directory is still not merged. That's going to be interesting. Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-09-26 13:04:21 +07:00
#include <asm/processor.h>
#include <asm/ppc_asm.h>
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
#include <asm/asm-offsets.h>
.section ".toc","aw"
PPC64_CACHES:
.tc ppc64_caches[TC],ppc64_caches
.section ".text"
powerpc: Simplify 4k/64k copy_page logic To make it easier to add optimised versions of copy_page, remove the 4kB loop for 64kB pages and just do all the work in copy_page. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2011-05-09 04:18:38 +07:00
_GLOBAL(copy_page)
powerpc: POWER7 optimised copy_page using VMX and enhanced prefetch Implement a POWER7 optimised copy_page using VMX and enhanced prefetch instructions. We use enhanced prefetch hints to prefetch both the load and store side. We copy a cacheline at a time and fall back to regular loads and stores if we are unable to use VMX (eg we are in an interrupt). The following microbenchmark was used to assess the impact of the patch: http://ozlabs.org/~anton/junkcode/page_fault_file.c We test MAP_PRIVATE page faults across a 1GB file, 100 times: # time ./page_fault_file -p -l 1G -i 100 Before: 22.25s After: 18.89s 17% faster Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2012-05-30 02:33:12 +07:00
BEGIN_FTR_SECTION
powerpc: Simplify 4k/64k copy_page logic To make it easier to add optimised versions of copy_page, remove the 4kB loop for 64kB pages and just do all the work in copy_page. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2011-05-09 04:18:38 +07:00
lis r5,PAGE_SIZE@h
powerpc: POWER7 optimised copy_page using VMX and enhanced prefetch Implement a POWER7 optimised copy_page using VMX and enhanced prefetch instructions. We use enhanced prefetch hints to prefetch both the load and store side. We copy a cacheline at a time and fall back to regular loads and stores if we are unable to use VMX (eg we are in an interrupt). The following microbenchmark was used to assess the impact of the patch: http://ozlabs.org/~anton/junkcode/page_fault_file.c We test MAP_PRIVATE page faults across a 1GB file, 100 times: # time ./page_fault_file -p -l 1G -i 100 Before: 22.25s After: 18.89s 17% faster Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2012-05-30 02:33:12 +07:00
FTR_SECTION_ELSE
powerpc: No need to use dot symbols when branching to a function binutils is smart enough to know that a branch to a function descriptor is actually a branch to the functions text address. Alan tells me that binutils has been doing this for 9 years. Signed-off-by: Anton Blanchard <anton@samba.org>
2014-02-04 12:04:35 +07:00
b copypage_power7
powerpc: POWER7 optimised copy_page using VMX and enhanced prefetch Implement a POWER7 optimised copy_page using VMX and enhanced prefetch instructions. We use enhanced prefetch hints to prefetch both the load and store side. We copy a cacheline at a time and fall back to regular loads and stores if we are unable to use VMX (eg we are in an interrupt). The following microbenchmark was used to assess the impact of the patch: http://ozlabs.org/~anton/junkcode/page_fault_file.c We test MAP_PRIVATE page faults across a 1GB file, 100 times: # time ./page_fault_file -p -l 1G -i 100 Before: 22.25s After: 18.89s 17% faster Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2012-05-30 02:33:12 +07:00
ALT_FTR_SECTION_END_IFCLR(CPU_FTR_VMX_COPY)
powerpc: Simplify 4k/64k copy_page logic To make it easier to add optimised versions of copy_page, remove the 4kB loop for 64kB pages and just do all the work in copy_page. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2011-05-09 04:18:38 +07:00
ori r5,r5,PAGE_SIZE@l
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
BEGIN_FTR_SECTION
ld r10,PPC64_CACHES@toc(r2)
lwz r11,DCACHEL1LOGLINESIZE(r10) /* log2 of cache line size */
lwz r12,DCACHEL1LINESIZE(r10) /* get cache line size */
li r9,0
srd r8,r5,r11
mtctr r8
powerpc: perf_event: Cleanup copy_page output by hiding setup symbol A lot of hits in "setup" doesn't make much sense, so hide this symbol and allow all the hits to end up in copy_4k_page. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2009-10-18 08:23:28 +07:00
.Lsetup:
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
dcbt r9,r4
dcbz r9,r3
add r9,r9,r12
powerpc: perf_event: Cleanup copy_page output by hiding setup symbol A lot of hits in "setup" doesn't make much sense, so hide this symbol and allow all the hits to end up in copy_4k_page. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2009-10-18 08:23:28 +07:00
bdnz .Lsetup
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
END_FTR_SECTION_IFSET(CPU_FTR_CP_USE_DCBTZ)
powerpc: Merge enough to start building in arch/powerpc. This creates the directory structure under arch/powerpc and a bunch of Kconfig files. It does a first-cut merge of arch/powerpc/mm, arch/powerpc/lib and arch/powerpc/platforms/powermac. This is enough to build a 32-bit powermac kernel with ARCH=powerpc. For now we are getting some unmerged files from arch/ppc/kernel and arch/ppc/syslib, or arch/ppc64/kernel. This makes some minor changes to files in those directories and files outside arch/powerpc. The boot directory is still not merged. That's going to be interesting. Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-09-26 13:04:21 +07:00
addi r3,r3,-8
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
srdi r8,r5,7 /* page is copied in 128 byte strides */
addi r8,r8,-1 /* one stride copied outside loop */
mtctr r8
ld r5,0(r4)
ld r6,8(r4)
ld r7,16(r4)
ldu r8,24(r4)
1: std r5,8(r3)
std r6,16(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r9,8(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ld r10,16(r4)
std r7,24(r3)
std r8,32(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r11,24(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ld r12,32(r4)
std r9,40(r3)
std r10,48(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r5,40(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ld r6,48(r4)
std r11,56(r3)
std r12,64(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r7,56(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ld r8,64(r4)
std r5,72(r3)
std r6,80(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r9,72(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ld r10,80(r4)
std r7,88(r3)
std r8,96(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r11,88(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ld r12,96(r4)
std r9,104(r3)
std r10,112(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r5,104(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ld r6,112(r4)
std r11,120(r3)
stdu r12,128(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r7,120(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ldu r8,128(r4)
powerpc: Merge enough to start building in arch/powerpc. This creates the directory structure under arch/powerpc and a bunch of Kconfig files. It does a first-cut merge of arch/powerpc/mm, arch/powerpc/lib and arch/powerpc/platforms/powermac. This is enough to build a 32-bit powermac kernel with ARCH=powerpc. For now we are getting some unmerged files from arch/ppc/kernel and arch/ppc/syslib, or arch/ppc64/kernel. This makes some minor changes to files in those directories and files outside arch/powerpc. The boot directory is still not merged. That's going to be interesting. Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-09-26 13:04:21 +07:00
bdnz 1b
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
std r5,8(r3)
std r6,16(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r9,8(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ld r10,16(r4)
std r7,24(r3)
std r8,32(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r11,24(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ld r12,32(r4)
std r9,40(r3)
std r10,48(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r5,40(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ld r6,48(r4)
std r11,56(r3)
std r12,64(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r7,56(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ld r8,64(r4)
std r5,72(r3)
std r6,80(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r9,72(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ld r10,80(r4)
std r7,88(r3)
std r8,96(r3)
powerpc: Pair loads and stores in copy_4k_page A number of our chips like loads and stores to be paired. A small kernel module testcase shows the improvement of pairing loads and stores in copy_4k_page: POWER6: +9% POWER7: +1.5% #include <linux/module.h> #include <linux/mm.h> #define ITERATIONS 10000000 static int __init copypage_init(void) { struct timespec before, after; unsigned long i; struct page *destpage, *srcpage; char *dest, *src; destpage = alloc_page(GFP_KERNEL); srcpage = alloc_page(GFP_KERNEL); dest = page_address(destpage); src = page_address(srcpage); getnstimeofday(&before); for (i = 0; i < ITERATIONS; i++) copy_4K_page(dest, src); getnstimeofday(&after); free_page((unsigned long)dest); free_page((unsigned long)src); printk(KERN_DEBUG "copy_4K_page loop took %lu ns\n", (after.tv_sec - before.tv_sec) * NSEC_PER_SEC + (after.tv_nsec - before.tv_nsec)); return 0; } static void __exit copypage_exit(void) { } module_init(copypage_init) module_exit(copypage_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Anton Blanchard"); Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2010-02-11 01:07:54 +07:00
ld r11,88(r4)
powerpc: New copy_4K_page() This new copy_4K_page() function was originally tuned for the best performance on the Cell processor, but after testing on more 64bit powerpc chips it was found that with a small modification it either matched the performance offered by the current mainline version or bettered it by a small amount. It was found that on a Cell-based QS22 blade the amount of system time measured when compiling a 2.6.26 pseries_defconfig decreased by 4%. Using the same test, a 4-way 970MP machine saw a decrease of 2% in system time. No noticeable change was seen on Power4, Power5 or Power6. The 4096 byte page is copied in thirty-two 128 byte strides. An initial setup loop executes dcbt instructions for the whole source page and dcbz instructions for the whole destination page. To do this, the cache line size is retrieved from ppc64_caches. A new CPU feature bit, CPU_FTR_CP_USE_DCBTZ, (introduced in the previous patch) is used to make the modification to this new copy routine - on Power4, 970 and Cell the feature bit is set so the setup loop is executed, but on all other 64bit chips the setup loop is nop'ed out. Signed-off-by: Mark Nelson <markn@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-08-22 11:39:00 +07:00
ld r12,96(r4)
std r9,104(r3)
std r10,112(r3)
std r11,120(r3)
std r12,128(r3)
powerpc: Merge enough to start building in arch/powerpc. This creates the directory structure under arch/powerpc and a bunch of Kconfig files. It does a first-cut merge of arch/powerpc/mm, arch/powerpc/lib and arch/powerpc/platforms/powermac. This is enough to build a 32-bit powermac kernel with ARCH=powerpc. For now we are getting some unmerged files from arch/ppc/kernel and arch/ppc/syslib, or arch/ppc64/kernel. This makes some minor changes to files in those directories and files outside arch/powerpc. The boot directory is still not merged. That's going to be interesting. Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-09-26 13:04:21 +07:00
blr
Reference in New Issue Copy Permalink