https://github.com/torvalds/linux
Revision 4fca50d440cc5d4dc570ad5484cc0b70b381bc2a authored by Jan Kara on 08 September 2022, 09:21:24 UTC, committed by Theodore Ts'o on 22 September 2022, 02:11:34 UTC
One of the side-effects of mb_optimize_scan was that the optimized functions to select next group to try were called even before we tried the goal group. As a result we no longer allocate files close to corresponding inodes as well as we don't try to expand currently allocated extent in the same group. This results in reaim regression with workfile.disk workload of upto 8% with many clients on my test machine: baseline mb_optimize_scan Hmean disk-1 2114.16 ( 0.00%) 2099.37 ( -0.70%) Hmean disk-41 87794.43 ( 0.00%) 83787.47 * -4.56%* Hmean disk-81 148170.73 ( 0.00%) 135527.05 * -8.53%* Hmean disk-121 177506.11 ( 0.00%) 166284.93 * -6.32%* Hmean disk-161 220951.51 ( 0.00%) 207563.39 * -6.06%* Hmean disk-201 208722.74 ( 0.00%) 203235.59 ( -2.63%) Hmean disk-241 222051.60 ( 0.00%) 217705.51 ( -1.96%) Hmean disk-281 252244.17 ( 0.00%) 241132.72 * -4.41%* Hmean disk-321 255844.84 ( 0.00%) 245412.84 * -4.08%* Also this is causing huge regression (time increased by a factor of 5 or so) when untarring archive with lots of small files on some eMMC storage cards. Fix the problem by making sure we try goal group first. Fixes: 196e402adf2e ("ext4: improve cr 0 / cr 1 group scanning") CC: stable@kernel.org Reported-and-tested-by: Stefan Wahren <stefan.wahren@i2se.com> Tested-by: Ojaswin Mujoo <ojaswin@linux.ibm.com> Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com> Link: https://lore.kernel.org/all/20220727105123.ckwrhbilzrxqpt24@quack3/ Link: https://lore.kernel.org/all/0d81a7c2-46b7-6010-62a4-3e6cfc1628d6@i2se.com/ Signed-off-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20220908092136.11770-1-jack@suse.cz Signed-off-by: Theodore Ts'o <tytso@mit.edu>
1 parent 7e18e42
Tip revision: 4fca50d440cc5d4dc570ad5484cc0b70b381bc2a authored by Jan Kara on 08 September 2022, 09:21:24 UTC
ext4: make mballoc try target group first even with mb_optimize_scan
ext4: make mballoc try target group first even with mb_optimize_scan
Tip revision: 4fca50d
sm3.c
/* SPDX-License-Identifier: GPL-2.0-only */
/*
* SM3 secure hash, as specified by OSCCA GM/T 0004-2012 SM3 and described
* at https://datatracker.ietf.org/doc/html/draft-sca-cfrg-sm3-02
*
* Copyright (C) 2017 ARM Limited or its affiliates.
* Copyright (C) 2017 Gilad Ben-Yossef <gilad@benyossef.com>
* Copyright (C) 2021 Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
*/
#include <linux/module.h>
#include <asm/unaligned.h>
#include <crypto/sm3.h>
static const u32 ____cacheline_aligned K[64] = {
0x79cc4519, 0xf3988a32, 0xe7311465, 0xce6228cb,
0x9cc45197, 0x3988a32f, 0x7311465e, 0xe6228cbc,
0xcc451979, 0x988a32f3, 0x311465e7, 0x6228cbce,
0xc451979c, 0x88a32f39, 0x11465e73, 0x228cbce6,
0x9d8a7a87, 0x3b14f50f, 0x7629ea1e, 0xec53d43c,
0xd8a7a879, 0xb14f50f3, 0x629ea1e7, 0xc53d43ce,
0x8a7a879d, 0x14f50f3b, 0x29ea1e76, 0x53d43cec,
0xa7a879d8, 0x4f50f3b1, 0x9ea1e762, 0x3d43cec5,
0x7a879d8a, 0xf50f3b14, 0xea1e7629, 0xd43cec53,
0xa879d8a7, 0x50f3b14f, 0xa1e7629e, 0x43cec53d,
0x879d8a7a, 0x0f3b14f5, 0x1e7629ea, 0x3cec53d4,
0x79d8a7a8, 0xf3b14f50, 0xe7629ea1, 0xcec53d43,
0x9d8a7a87, 0x3b14f50f, 0x7629ea1e, 0xec53d43c,
0xd8a7a879, 0xb14f50f3, 0x629ea1e7, 0xc53d43ce,
0x8a7a879d, 0x14f50f3b, 0x29ea1e76, 0x53d43cec,
0xa7a879d8, 0x4f50f3b1, 0x9ea1e762, 0x3d43cec5
};
/*
* Transform the message X which consists of 16 32-bit-words. See
* GM/T 004-2012 for details.
*/
#define R(i, a, b, c, d, e, f, g, h, t, w1, w2) \
do { \
ss1 = rol32((rol32((a), 12) + (e) + (t)), 7); \
ss2 = ss1 ^ rol32((a), 12); \
d += FF ## i(a, b, c) + ss2 + ((w1) ^ (w2)); \
h += GG ## i(e, f, g) + ss1 + (w1); \
b = rol32((b), 9); \
f = rol32((f), 19); \
h = P0((h)); \
} while (0)
#define R1(a, b, c, d, e, f, g, h, t, w1, w2) \
R(1, a, b, c, d, e, f, g, h, t, w1, w2)
#define R2(a, b, c, d, e, f, g, h, t, w1, w2) \
R(2, a, b, c, d, e, f, g, h, t, w1, w2)
#define FF1(x, y, z) (x ^ y ^ z)
#define FF2(x, y, z) ((x & y) | (x & z) | (y & z))
#define GG1(x, y, z) FF1(x, y, z)
#define GG2(x, y, z) ((x & y) | (~x & z))
/* Message expansion */
#define P0(x) ((x) ^ rol32((x), 9) ^ rol32((x), 17))
#define P1(x) ((x) ^ rol32((x), 15) ^ rol32((x), 23))
#define I(i) (W[i] = get_unaligned_be32(data + i * 4))
#define W1(i) (W[i & 0x0f])
#define W2(i) (W[i & 0x0f] = \
P1(W[i & 0x0f] \
^ W[(i-9) & 0x0f] \
^ rol32(W[(i-3) & 0x0f], 15)) \
^ rol32(W[(i-13) & 0x0f], 7) \
^ W[(i-6) & 0x0f])
static void sm3_transform(struct sm3_state *sctx, u8 const *data, u32 W[16])
{
u32 a, b, c, d, e, f, g, h, ss1, ss2;
a = sctx->state[0];
b = sctx->state[1];
c = sctx->state[2];
d = sctx->state[3];
e = sctx->state[4];
f = sctx->state[5];
g = sctx->state[6];
h = sctx->state[7];
R1(a, b, c, d, e, f, g, h, K[0], I(0), I(4));
R1(d, a, b, c, h, e, f, g, K[1], I(1), I(5));
R1(c, d, a, b, g, h, e, f, K[2], I(2), I(6));
R1(b, c, d, a, f, g, h, e, K[3], I(3), I(7));
R1(a, b, c, d, e, f, g, h, K[4], W1(4), I(8));
R1(d, a, b, c, h, e, f, g, K[5], W1(5), I(9));
R1(c, d, a, b, g, h, e, f, K[6], W1(6), I(10));
R1(b, c, d, a, f, g, h, e, K[7], W1(7), I(11));
R1(a, b, c, d, e, f, g, h, K[8], W1(8), I(12));
R1(d, a, b, c, h, e, f, g, K[9], W1(9), I(13));
R1(c, d, a, b, g, h, e, f, K[10], W1(10), I(14));
R1(b, c, d, a, f, g, h, e, K[11], W1(11), I(15));
R1(a, b, c, d, e, f, g, h, K[12], W1(12), W2(16));
R1(d, a, b, c, h, e, f, g, K[13], W1(13), W2(17));
R1(c, d, a, b, g, h, e, f, K[14], W1(14), W2(18));
R1(b, c, d, a, f, g, h, e, K[15], W1(15), W2(19));
R2(a, b, c, d, e, f, g, h, K[16], W1(16), W2(20));
R2(d, a, b, c, h, e, f, g, K[17], W1(17), W2(21));
R2(c, d, a, b, g, h, e, f, K[18], W1(18), W2(22));
R2(b, c, d, a, f, g, h, e, K[19], W1(19), W2(23));
R2(a, b, c, d, e, f, g, h, K[20], W1(20), W2(24));
R2(d, a, b, c, h, e, f, g, K[21], W1(21), W2(25));
R2(c, d, a, b, g, h, e, f, K[22], W1(22), W2(26));
R2(b, c, d, a, f, g, h, e, K[23], W1(23), W2(27));
R2(a, b, c, d, e, f, g, h, K[24], W1(24), W2(28));
R2(d, a, b, c, h, e, f, g, K[25], W1(25), W2(29));
R2(c, d, a, b, g, h, e, f, K[26], W1(26), W2(30));
R2(b, c, d, a, f, g, h, e, K[27], W1(27), W2(31));
R2(a, b, c, d, e, f, g, h, K[28], W1(28), W2(32));
R2(d, a, b, c, h, e, f, g, K[29], W1(29), W2(33));
R2(c, d, a, b, g, h, e, f, K[30], W1(30), W2(34));
R2(b, c, d, a, f, g, h, e, K[31], W1(31), W2(35));
R2(a, b, c, d, e, f, g, h, K[32], W1(32), W2(36));
R2(d, a, b, c, h, e, f, g, K[33], W1(33), W2(37));
R2(c, d, a, b, g, h, e, f, K[34], W1(34), W2(38));
R2(b, c, d, a, f, g, h, e, K[35], W1(35), W2(39));
R2(a, b, c, d, e, f, g, h, K[36], W1(36), W2(40));
R2(d, a, b, c, h, e, f, g, K[37], W1(37), W2(41));
R2(c, d, a, b, g, h, e, f, K[38], W1(38), W2(42));
R2(b, c, d, a, f, g, h, e, K[39], W1(39), W2(43));
R2(a, b, c, d, e, f, g, h, K[40], W1(40), W2(44));
R2(d, a, b, c, h, e, f, g, K[41], W1(41), W2(45));
R2(c, d, a, b, g, h, e, f, K[42], W1(42), W2(46));
R2(b, c, d, a, f, g, h, e, K[43], W1(43), W2(47));
R2(a, b, c, d, e, f, g, h, K[44], W1(44), W2(48));
R2(d, a, b, c, h, e, f, g, K[45], W1(45), W2(49));
R2(c, d, a, b, g, h, e, f, K[46], W1(46), W2(50));
R2(b, c, d, a, f, g, h, e, K[47], W1(47), W2(51));
R2(a, b, c, d, e, f, g, h, K[48], W1(48), W2(52));
R2(d, a, b, c, h, e, f, g, K[49], W1(49), W2(53));
R2(c, d, a, b, g, h, e, f, K[50], W1(50), W2(54));
R2(b, c, d, a, f, g, h, e, K[51], W1(51), W2(55));
R2(a, b, c, d, e, f, g, h, K[52], W1(52), W2(56));
R2(d, a, b, c, h, e, f, g, K[53], W1(53), W2(57));
R2(c, d, a, b, g, h, e, f, K[54], W1(54), W2(58));
R2(b, c, d, a, f, g, h, e, K[55], W1(55), W2(59));
R2(a, b, c, d, e, f, g, h, K[56], W1(56), W2(60));
R2(d, a, b, c, h, e, f, g, K[57], W1(57), W2(61));
R2(c, d, a, b, g, h, e, f, K[58], W1(58), W2(62));
R2(b, c, d, a, f, g, h, e, K[59], W1(59), W2(63));
R2(a, b, c, d, e, f, g, h, K[60], W1(60), W2(64));
R2(d, a, b, c, h, e, f, g, K[61], W1(61), W2(65));
R2(c, d, a, b, g, h, e, f, K[62], W1(62), W2(66));
R2(b, c, d, a, f, g, h, e, K[63], W1(63), W2(67));
sctx->state[0] ^= a;
sctx->state[1] ^= b;
sctx->state[2] ^= c;
sctx->state[3] ^= d;
sctx->state[4] ^= e;
sctx->state[5] ^= f;
sctx->state[6] ^= g;
sctx->state[7] ^= h;
}
#undef R
#undef R1
#undef R2
#undef I
#undef W1
#undef W2
static inline void sm3_block(struct sm3_state *sctx,
u8 const *data, int blocks, u32 W[16])
{
while (blocks--) {
sm3_transform(sctx, data, W);
data += SM3_BLOCK_SIZE;
}
}
void sm3_update(struct sm3_state *sctx, const u8 *data, unsigned int len)
{
unsigned int partial = sctx->count % SM3_BLOCK_SIZE;
u32 W[16];
sctx->count += len;
if ((partial + len) >= SM3_BLOCK_SIZE) {
int blocks;
if (partial) {
int p = SM3_BLOCK_SIZE - partial;
memcpy(sctx->buffer + partial, data, p);
data += p;
len -= p;
sm3_block(sctx, sctx->buffer, 1, W);
}
blocks = len / SM3_BLOCK_SIZE;
len %= SM3_BLOCK_SIZE;
if (blocks) {
sm3_block(sctx, data, blocks, W);
data += blocks * SM3_BLOCK_SIZE;
}
memzero_explicit(W, sizeof(W));
partial = 0;
}
if (len)
memcpy(sctx->buffer + partial, data, len);
}
EXPORT_SYMBOL_GPL(sm3_update);
void sm3_final(struct sm3_state *sctx, u8 *out)
{
const int bit_offset = SM3_BLOCK_SIZE - sizeof(u64);
__be64 *bits = (__be64 *)(sctx->buffer + bit_offset);
__be32 *digest = (__be32 *)out;
unsigned int partial = sctx->count % SM3_BLOCK_SIZE;
u32 W[16];
int i;
sctx->buffer[partial++] = 0x80;
if (partial > bit_offset) {
memset(sctx->buffer + partial, 0, SM3_BLOCK_SIZE - partial);
partial = 0;
sm3_block(sctx, sctx->buffer, 1, W);
}
memset(sctx->buffer + partial, 0, bit_offset - partial);
*bits = cpu_to_be64(sctx->count << 3);
sm3_block(sctx, sctx->buffer, 1, W);
for (i = 0; i < 8; i++)
put_unaligned_be32(sctx->state[i], digest++);
/* Zeroize sensitive information. */
memzero_explicit(W, sizeof(W));
memzero_explicit(sctx, sizeof(*sctx));
}
EXPORT_SYMBOL_GPL(sm3_final);
MODULE_DESCRIPTION("Generic SM3 library");
MODULE_LICENSE("GPL v2");
Computing file changes ...