korg/linux
0
0
Fork 0
mirror of https://mirrors.bfsu.edu.cn/git/linux.git synced 2024-11-20 02:34:23 +08:00
Code Issues Packages Projects Releases Wiki Activity
2d4d1eea54
linux/lib/raid6/avx2.c
Markus Stockhausen fe5cbc6e06 md/raid6 algorithms: delta syndrome functions 
v3: s-o-b comment, explanation of performance and descision for
the start/stop implementation

Implementing rmw functionality for RAID6 requires optimized syndrome
calculation. Up to now we can only generate a complete syndrome. The
target P/Q pages are always overwritten. With this patch we provide
a framework for inplace P/Q modification. In the first place simply
fill those functions with NULL values.

xor_syndrome() has two additional parameters: start & stop. These
will indicate the first and last page that are changing during a
rmw run. That makes it possible to avoid several unneccessary loops
and speed up calculation. The caller needs to implement the following
logic to make the functions work.

1) xor_syndrome(disks, start, stop, ...): "Remove" all data of source
blocks inside P/Q between (and including) start and end.

2) modify any block with start <= block <= stop

3) xor_syndrome(disks, start, stop, ...): "Reinsert" all data of
source blocks into P/Q between (and including) start and end.

Pages between start and stop that won't be changed should be filled
with a pointer to the kernel zero page. The reasons for not taking NULL
pages are:

1) Algorithms cross the whole source data line by line. Thus avoid
additional branches.

2) Having a NULL page avoids calculating the XOR P parity but still
need calulation steps for the Q parity. Depending on the algorithm
unrolling that might be only a difference of 2 instructions per loop.

The benchmark numbers of the gen_syndrome() functions are displayed in
the kernel log. Do the same for the xor_syndrome() functions. This
will help to analyze performance problems and give an rough estimate
how well the algorithm works. The choice of the fastest algorithm will
still depend on the gen_syndrome() performance.

With the start/stop page implementation the speed can vary a lot in real
life. E.g. a change of page 0 & page 15 on a stripe will be harder to
compute than the case where page 0 & page 1 are XOR candidates. To be not
to enthusiatic about the expected speeds we will run a worse case test
that simulates a change on the upper half of the stripe. So we do:

1) calculation of P/Q for the upper pages

2) continuation of Q for the lower (empty) pages

Signed-off-by: Markus Stockhausen <stockhausen@collogia.de>
Signed-off-by: NeilBrown <neilb@suse.de>
2015-04-22 08:00:41 +10:00

255 lines
8.6 KiB
C
Raw Blame History

/* -*- linux-c -*- ------------------------------------------------------- *
*
* Copyright (C) 2012 Intel Corporation
* Author: Yuanhan Liu <yuanhan.liu@linux.intel.com>
*
* Based on sse2.c: Copyright 2002 H. Peter Anvin - All Rights Reserved
*
*
* 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, Inc., 53 Temple Place Ste 330,
* Boston MA 02111-1307, USA; either version 2 of the License, or
* (at your option) any later version; incorporated herein by reference.
*
* ----------------------------------------------------------------------- */
/*
* AVX2 implementation of RAID-6 syndrome functions
*
*/
#ifdef CONFIG_AS_AVX2
#include <linux/raid/pq.h>
#include "x86.h"
static const struct raid6_avx2_constants {
u64 x1d[4];
} raid6_avx2_constants __aligned(32) = {
{ 0x1d1d1d1d1d1d1d1dULL, 0x1d1d1d1d1d1d1d1dULL,
0x1d1d1d1d1d1d1d1dULL, 0x1d1d1d1d1d1d1d1dULL,},
};
static int raid6_have_avx2(void)
{
return boot_cpu_has(X86_FEATURE_AVX2) && boot_cpu_has(X86_FEATURE_AVX);
}
/*
* Plain AVX2 implementation
*/
static void raid6_avx21_gen_syndrome(int disks, size_t bytes, void **ptrs)
{
u8 **dptr = (u8 **)ptrs;
u8 *p, *q;
int d, z, z0;
z0 = disks - 3; /* Highest data disk */
p = dptr[z0+1]; /* XOR parity */
q = dptr[z0+2]; /* RS syndrome */
kernel_fpu_begin();
asm volatile("vmovdqa %0,%%ymm0" : : "m" (raid6_avx2_constants.x1d[0]));
asm volatile("vpxor %ymm3,%ymm3,%ymm3"); /* Zero temp */
for (d = 0; d < bytes; d += 32) {
asm volatile("prefetchnta %0" : : "m" (dptr[z0][d]));
asm volatile("vmovdqa %0,%%ymm2" : : "m" (dptr[z0][d]));/* P[0] */
asm volatile("prefetchnta %0" : : "m" (dptr[z0-1][d]));
asm volatile("vmovdqa %ymm2,%ymm4");/* Q[0] */
asm volatile("vmovdqa %0,%%ymm6" : : "m" (dptr[z0-1][d]));
for (z = z0-2; z >= 0; z--) {
asm volatile("prefetchnta %0" : : "m" (dptr[z][d]));
asm volatile("vpcmpgtb %ymm4,%ymm3,%ymm5");
asm volatile("vpaddb %ymm4,%ymm4,%ymm4");
asm volatile("vpand %ymm0,%ymm5,%ymm5");
asm volatile("vpxor %ymm5,%ymm4,%ymm4");
asm volatile("vpxor %ymm6,%ymm2,%ymm2");
asm volatile("vpxor %ymm6,%ymm4,%ymm4");
asm volatile("vmovdqa %0,%%ymm6" : : "m" (dptr[z][d]));
}
asm volatile("vpcmpgtb %ymm4,%ymm3,%ymm5");
asm volatile("vpaddb %ymm4,%ymm4,%ymm4");
asm volatile("vpand %ymm0,%ymm5,%ymm5");
asm volatile("vpxor %ymm5,%ymm4,%ymm4");
asm volatile("vpxor %ymm6,%ymm2,%ymm2");
asm volatile("vpxor %ymm6,%ymm4,%ymm4");
asm volatile("vmovntdq %%ymm2,%0" : "=m" (p[d]));
asm volatile("vpxor %ymm2,%ymm2,%ymm2");
asm volatile("vmovntdq %%ymm4,%0" : "=m" (q[d]));
asm volatile("vpxor %ymm4,%ymm4,%ymm4");
}
asm volatile("sfence" : : : "memory");
kernel_fpu_end();
}
const struct raid6_calls raid6_avx2x1 = {
raid6_avx21_gen_syndrome,
NULL, /* XOR not yet implemented */
raid6_have_avx2,
"avx2x1",
1 /* Has cache hints */
};
/*
* Unrolled-by-2 AVX2 implementation
*/
static void raid6_avx22_gen_syndrome(int disks, size_t bytes, void **ptrs)
{
u8 **dptr = (u8 **)ptrs;
u8 *p, *q;
int d, z, z0;
z0 = disks - 3; /* Highest data disk */
p = dptr[z0+1]; /* XOR parity */
q = dptr[z0+2]; /* RS syndrome */
kernel_fpu_begin();
asm volatile("vmovdqa %0,%%ymm0" : : "m" (raid6_avx2_constants.x1d[0]));
asm volatile("vpxor %ymm1,%ymm1,%ymm1"); /* Zero temp */
/* We uniformly assume a single prefetch covers at least 32 bytes */
for (d = 0; d < bytes; d += 64) {
asm volatile("prefetchnta %0" : : "m" (dptr[z0][d]));
asm volatile("prefetchnta %0" : : "m" (dptr[z0][d+32]));
asm volatile("vmovdqa %0,%%ymm2" : : "m" (dptr[z0][d]));/* P[0] */
asm volatile("vmovdqa %0,%%ymm3" : : "m" (dptr[z0][d+32]));/* P[1] */
asm volatile("vmovdqa %ymm2,%ymm4"); /* Q[0] */
asm volatile("vmovdqa %ymm3,%ymm6"); /* Q[1] */
for (z = z0-1; z >= 0; z--) {
asm volatile("prefetchnta %0" : : "m" (dptr[z][d]));
asm volatile("prefetchnta %0" : : "m" (dptr[z][d+32]));
asm volatile("vpcmpgtb %ymm4,%ymm1,%ymm5");
asm volatile("vpcmpgtb %ymm6,%ymm1,%ymm7");
asm volatile("vpaddb %ymm4,%ymm4,%ymm4");
asm volatile("vpaddb %ymm6,%ymm6,%ymm6");
asm volatile("vpand %ymm0,%ymm5,%ymm5");
asm volatile("vpand %ymm0,%ymm7,%ymm7");
asm volatile("vpxor %ymm5,%ymm4,%ymm4");
asm volatile("vpxor %ymm7,%ymm6,%ymm6");
asm volatile("vmovdqa %0,%%ymm5" : : "m" (dptr[z][d]));
asm volatile("vmovdqa %0,%%ymm7" : : "m" (dptr[z][d+32]));
asm volatile("vpxor %ymm5,%ymm2,%ymm2");
asm volatile("vpxor %ymm7,%ymm3,%ymm3");
asm volatile("vpxor %ymm5,%ymm4,%ymm4");
asm volatile("vpxor %ymm7,%ymm6,%ymm6");
}
asm volatile("vmovntdq %%ymm2,%0" : "=m" (p[d]));
asm volatile("vmovntdq %%ymm3,%0" : "=m" (p[d+32]));
asm volatile("vmovntdq %%ymm4,%0" : "=m" (q[d]));
asm volatile("vmovntdq %%ymm6,%0" : "=m" (q[d+32]));
}
asm volatile("sfence" : : : "memory");
kernel_fpu_end();
}
const struct raid6_calls raid6_avx2x2 = {
raid6_avx22_gen_syndrome,
NULL, /* XOR not yet implemented */
raid6_have_avx2,
"avx2x2",
1 /* Has cache hints */
};
#ifdef CONFIG_X86_64
/*
* Unrolled-by-4 AVX2 implementation
*/
static void raid6_avx24_gen_syndrome(int disks, size_t bytes, void **ptrs)
{
u8 **dptr = (u8 **)ptrs;
u8 *p, *q;
int d, z, z0;
z0 = disks - 3; /* Highest data disk */
p = dptr[z0+1]; /* XOR parity */
q = dptr[z0+2]; /* RS syndrome */
kernel_fpu_begin();
asm volatile("vmovdqa %0,%%ymm0" : : "m" (raid6_avx2_constants.x1d[0]));
asm volatile("vpxor %ymm1,%ymm1,%ymm1"); /* Zero temp */
asm volatile("vpxor %ymm2,%ymm2,%ymm2"); /* P[0] */
asm volatile("vpxor %ymm3,%ymm3,%ymm3"); /* P[1] */
asm volatile("vpxor %ymm4,%ymm4,%ymm4"); /* Q[0] */
asm volatile("vpxor %ymm6,%ymm6,%ymm6"); /* Q[1] */
asm volatile("vpxor %ymm10,%ymm10,%ymm10"); /* P[2] */
asm volatile("vpxor %ymm11,%ymm11,%ymm11"); /* P[3] */
asm volatile("vpxor %ymm12,%ymm12,%ymm12"); /* Q[2] */
asm volatile("vpxor %ymm14,%ymm14,%ymm14"); /* Q[3] */
for (d = 0; d < bytes; d += 128) {
for (z = z0; z >= 0; z--) {
asm volatile("prefetchnta %0" : : "m" (dptr[z][d]));
asm volatile("prefetchnta %0" : : "m" (dptr[z][d+32]));
asm volatile("prefetchnta %0" : : "m" (dptr[z][d+64]));
asm volatile("prefetchnta %0" : : "m" (dptr[z][d+96]));
asm volatile("vpcmpgtb %ymm4,%ymm1,%ymm5");
asm volatile("vpcmpgtb %ymm6,%ymm1,%ymm7");
asm volatile("vpcmpgtb %ymm12,%ymm1,%ymm13");
asm volatile("vpcmpgtb %ymm14,%ymm1,%ymm15");
asm volatile("vpaddb %ymm4,%ymm4,%ymm4");
asm volatile("vpaddb %ymm6,%ymm6,%ymm6");
asm volatile("vpaddb %ymm12,%ymm12,%ymm12");
asm volatile("vpaddb %ymm14,%ymm14,%ymm14");
asm volatile("vpand %ymm0,%ymm5,%ymm5");
asm volatile("vpand %ymm0,%ymm7,%ymm7");
asm volatile("vpand %ymm0,%ymm13,%ymm13");
asm volatile("vpand %ymm0,%ymm15,%ymm15");
asm volatile("vpxor %ymm5,%ymm4,%ymm4");
asm volatile("vpxor %ymm7,%ymm6,%ymm6");
asm volatile("vpxor %ymm13,%ymm12,%ymm12");
asm volatile("vpxor %ymm15,%ymm14,%ymm14");
asm volatile("vmovdqa %0,%%ymm5" : : "m" (dptr[z][d]));
asm volatile("vmovdqa %0,%%ymm7" : : "m" (dptr[z][d+32]));
asm volatile("vmovdqa %0,%%ymm13" : : "m" (dptr[z][d+64]));
asm volatile("vmovdqa %0,%%ymm15" : : "m" (dptr[z][d+96]));
asm volatile("vpxor %ymm5,%ymm2,%ymm2");
asm volatile("vpxor %ymm7,%ymm3,%ymm3");
asm volatile("vpxor %ymm13,%ymm10,%ymm10");
asm volatile("vpxor %ymm15,%ymm11,%ymm11");
asm volatile("vpxor %ymm5,%ymm4,%ymm4");
asm volatile("vpxor %ymm7,%ymm6,%ymm6");
asm volatile("vpxor %ymm13,%ymm12,%ymm12");
asm volatile("vpxor %ymm15,%ymm14,%ymm14");
}
asm volatile("vmovntdq %%ymm2,%0" : "=m" (p[d]));
asm volatile("vpxor %ymm2,%ymm2,%ymm2");
asm volatile("vmovntdq %%ymm3,%0" : "=m" (p[d+32]));
asm volatile("vpxor %ymm3,%ymm3,%ymm3");
asm volatile("vmovntdq %%ymm10,%0" : "=m" (p[d+64]));
asm volatile("vpxor %ymm10,%ymm10,%ymm10");
asm volatile("vmovntdq %%ymm11,%0" : "=m" (p[d+96]));
asm volatile("vpxor %ymm11,%ymm11,%ymm11");
asm volatile("vmovntdq %%ymm4,%0" : "=m" (q[d]));
asm volatile("vpxor %ymm4,%ymm4,%ymm4");
asm volatile("vmovntdq %%ymm6,%0" : "=m" (q[d+32]));
asm volatile("vpxor %ymm6,%ymm6,%ymm6");
asm volatile("vmovntdq %%ymm12,%0" : "=m" (q[d+64]));
asm volatile("vpxor %ymm12,%ymm12,%ymm12");
asm volatile("vmovntdq %%ymm14,%0" : "=m" (q[d+96]));
asm volatile("vpxor %ymm14,%ymm14,%ymm14");
}
asm volatile("sfence" : : : "memory");
kernel_fpu_end();
}
const struct raid6_calls raid6_avx2x4 = {
raid6_avx24_gen_syndrome,
NULL, /* XOR not yet implemented */
raid6_have_avx2,
"avx2x4",
1 /* Has cache hints */
};
#endif
#endif /* CONFIG_AS_AVX2 */
Reference in New Issue View Git Blame Copy Permalink