]>
Commit | Line | Data |
---|---|---|
34dc7c2f BB |
1 | /* |
2 | * CDDL HEADER START | |
3 | * | |
4 | * The contents of this file are subject to the terms of the | |
5 | * Common Development and Distribution License (the "License"). | |
6 | * You may not use this file except in compliance with the License. | |
7 | * | |
8 | * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE | |
9 | * or http://www.opensolaris.org/os/licensing. | |
10 | * See the License for the specific language governing permissions | |
11 | * and limitations under the License. | |
12 | * | |
13 | * When distributing Covered Code, include this CDDL HEADER in each | |
14 | * file and include the License file at usr/src/OPENSOLARIS.LICENSE. | |
15 | * If applicable, add the following below this CDDL HEADER, with the | |
16 | * fields enclosed by brackets "[]" replaced with your own identifying | |
17 | * information: Portions Copyright [yyyy] [name of copyright owner] | |
18 | * | |
19 | * CDDL HEADER END | |
20 | */ | |
21 | ||
22 | /* | |
428870ff | 23 | * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. |
2e528b49 | 24 | * Copyright (c) 2013 by Delphix. All rights reserved. |
34dc7c2f BB |
25 | */ |
26 | ||
34dc7c2f BB |
27 | #include <sys/zfs_context.h> |
28 | #include <sys/spa.h> | |
29 | #include <sys/vdev_impl.h> | |
30 | #include <sys/zio.h> | |
31 | #include <sys/zio_checksum.h> | |
32 | #include <sys/fs/zfs.h> | |
33 | #include <sys/fm/fs/zfs.h> | |
34 | ||
35 | /* | |
36 | * Virtual device vector for RAID-Z. | |
37 | * | |
45d1cae3 BB |
38 | * This vdev supports single, double, and triple parity. For single parity, |
39 | * we use a simple XOR of all the data columns. For double or triple parity, | |
40 | * we use a special case of Reed-Solomon coding. This extends the | |
41 | * technique described in "The mathematics of RAID-6" by H. Peter Anvin by | |
42 | * drawing on the system described in "A Tutorial on Reed-Solomon Coding for | |
43 | * Fault-Tolerance in RAID-like Systems" by James S. Plank on which the | |
44 | * former is also based. The latter is designed to provide higher performance | |
45 | * for writes. | |
46 | * | |
47 | * Note that the Plank paper claimed to support arbitrary N+M, but was then | |
48 | * amended six years later identifying a critical flaw that invalidates its | |
49 | * claims. Nevertheless, the technique can be adapted to work for up to | |
50 | * triple parity. For additional parity, the amendment "Note: Correction to | |
51 | * the 1997 Tutorial on Reed-Solomon Coding" by James S. Plank and Ying Ding | |
52 | * is viable, but the additional complexity means that write performance will | |
53 | * suffer. | |
54 | * | |
55 | * All of the methods above operate on a Galois field, defined over the | |
56 | * integers mod 2^N. In our case we choose N=8 for GF(8) so that all elements | |
57 | * can be expressed with a single byte. Briefly, the operations on the | |
58 | * field are defined as follows: | |
34dc7c2f BB |
59 | * |
60 | * o addition (+) is represented by a bitwise XOR | |
61 | * o subtraction (-) is therefore identical to addition: A + B = A - B | |
62 | * o multiplication of A by 2 is defined by the following bitwise expression: | |
d3cc8b15 | 63 | * |
34dc7c2f BB |
64 | * (A * 2)_7 = A_6 |
65 | * (A * 2)_6 = A_5 | |
66 | * (A * 2)_5 = A_4 | |
67 | * (A * 2)_4 = A_3 + A_7 | |
68 | * (A * 2)_3 = A_2 + A_7 | |
69 | * (A * 2)_2 = A_1 + A_7 | |
70 | * (A * 2)_1 = A_0 | |
71 | * (A * 2)_0 = A_7 | |
72 | * | |
73 | * In C, multiplying by 2 is therefore ((a << 1) ^ ((a & 0x80) ? 0x1d : 0)). | |
45d1cae3 BB |
74 | * As an aside, this multiplication is derived from the error correcting |
75 | * primitive polynomial x^8 + x^4 + x^3 + x^2 + 1. | |
34dc7c2f BB |
76 | * |
77 | * Observe that any number in the field (except for 0) can be expressed as a | |
78 | * power of 2 -- a generator for the field. We store a table of the powers of | |
79 | * 2 and logs base 2 for quick look ups, and exploit the fact that A * B can | |
80 | * be rewritten as 2^(log_2(A) + log_2(B)) (where '+' is normal addition rather | |
45d1cae3 BB |
81 | * than field addition). The inverse of a field element A (A^-1) is therefore |
82 | * A ^ (255 - 1) = A^254. | |
34dc7c2f | 83 | * |
45d1cae3 BB |
84 | * The up-to-three parity columns, P, Q, R over several data columns, |
85 | * D_0, ... D_n-1, can be expressed by field operations: | |
34dc7c2f BB |
86 | * |
87 | * P = D_0 + D_1 + ... + D_n-2 + D_n-1 | |
88 | * Q = 2^n-1 * D_0 + 2^n-2 * D_1 + ... + 2^1 * D_n-2 + 2^0 * D_n-1 | |
89 | * = ((...((D_0) * 2 + D_1) * 2 + ...) * 2 + D_n-2) * 2 + D_n-1 | |
45d1cae3 BB |
90 | * R = 4^n-1 * D_0 + 4^n-2 * D_1 + ... + 4^1 * D_n-2 + 4^0 * D_n-1 |
91 | * = ((...((D_0) * 4 + D_1) * 4 + ...) * 4 + D_n-2) * 4 + D_n-1 | |
34dc7c2f | 92 | * |
45d1cae3 BB |
93 | * We chose 1, 2, and 4 as our generators because 1 corresponds to the trival |
94 | * XOR operation, and 2 and 4 can be computed quickly and generate linearly- | |
95 | * independent coefficients. (There are no additional coefficients that have | |
96 | * this property which is why the uncorrected Plank method breaks down.) | |
97 | * | |
98 | * See the reconstruction code below for how P, Q and R can used individually | |
99 | * or in concert to recover missing data columns. | |
34dc7c2f BB |
100 | */ |
101 | ||
102 | typedef struct raidz_col { | |
103 | uint64_t rc_devidx; /* child device index for I/O */ | |
104 | uint64_t rc_offset; /* device offset */ | |
105 | uint64_t rc_size; /* I/O size */ | |
106 | void *rc_data; /* I/O data */ | |
428870ff | 107 | void *rc_gdata; /* used to store the "good" version */ |
34dc7c2f BB |
108 | int rc_error; /* I/O error for this device */ |
109 | uint8_t rc_tried; /* Did we attempt this I/O column? */ | |
110 | uint8_t rc_skipped; /* Did we skip this I/O column? */ | |
111 | } raidz_col_t; | |
112 | ||
113 | typedef struct raidz_map { | |
45d1cae3 BB |
114 | uint64_t rm_cols; /* Regular column count */ |
115 | uint64_t rm_scols; /* Count including skipped columns */ | |
34dc7c2f BB |
116 | uint64_t rm_bigcols; /* Number of oversized columns */ |
117 | uint64_t rm_asize; /* Actual total I/O size */ | |
118 | uint64_t rm_missingdata; /* Count of missing data devices */ | |
119 | uint64_t rm_missingparity; /* Count of missing parity devices */ | |
120 | uint64_t rm_firstdatacol; /* First data column/parity count */ | |
428870ff | 121 | uint64_t rm_nskip; /* Skipped sectors for padding */ |
d3cc8b15 | 122 | uint64_t rm_skipstart; /* Column index of padding start */ |
428870ff BB |
123 | void *rm_datacopy; /* rm_asize-buffer of copied data */ |
124 | uintptr_t rm_reports; /* # of referencing checksum reports */ | |
125 | uint8_t rm_freed; /* map no longer has referencing ZIO */ | |
126 | uint8_t rm_ecksuminjected; /* checksum error was injected */ | |
34dc7c2f BB |
127 | raidz_col_t rm_col[1]; /* Flexible array of I/O columns */ |
128 | } raidz_map_t; | |
129 | ||
130 | #define VDEV_RAIDZ_P 0 | |
131 | #define VDEV_RAIDZ_Q 1 | |
45d1cae3 | 132 | #define VDEV_RAIDZ_R 2 |
45d1cae3 BB |
133 | |
134 | #define VDEV_RAIDZ_MUL_2(x) (((x) << 1) ^ (((x) & 0x80) ? 0x1d : 0)) | |
135 | #define VDEV_RAIDZ_MUL_4(x) (VDEV_RAIDZ_MUL_2(VDEV_RAIDZ_MUL_2(x))) | |
136 | ||
137 | /* | |
138 | * We provide a mechanism to perform the field multiplication operation on a | |
139 | * 64-bit value all at once rather than a byte at a time. This works by | |
140 | * creating a mask from the top bit in each byte and using that to | |
141 | * conditionally apply the XOR of 0x1d. | |
142 | */ | |
143 | #define VDEV_RAIDZ_64MUL_2(x, mask) \ | |
144 | { \ | |
145 | (mask) = (x) & 0x8080808080808080ULL; \ | |
146 | (mask) = ((mask) << 1) - ((mask) >> 7); \ | |
147 | (x) = (((x) << 1) & 0xfefefefefefefefeULL) ^ \ | |
c5b3a7bb | 148 | ((mask) & 0x1d1d1d1d1d1d1d1dULL); \ |
45d1cae3 | 149 | } |
34dc7c2f | 150 | |
45d1cae3 BB |
151 | #define VDEV_RAIDZ_64MUL_4(x, mask) \ |
152 | { \ | |
153 | VDEV_RAIDZ_64MUL_2((x), mask); \ | |
154 | VDEV_RAIDZ_64MUL_2((x), mask); \ | |
155 | } | |
34dc7c2f | 156 | |
45d1cae3 BB |
157 | /* |
158 | * Force reconstruction to use the general purpose method. | |
159 | */ | |
160 | int vdev_raidz_default_to_general; | |
34dc7c2f | 161 | |
d3cc8b15 | 162 | /* Powers of 2 in the Galois field defined above. */ |
34dc7c2f BB |
163 | static const uint8_t vdev_raidz_pow2[256] = { |
164 | 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, | |
165 | 0x1d, 0x3a, 0x74, 0xe8, 0xcd, 0x87, 0x13, 0x26, | |
166 | 0x4c, 0x98, 0x2d, 0x5a, 0xb4, 0x75, 0xea, 0xc9, | |
167 | 0x8f, 0x03, 0x06, 0x0c, 0x18, 0x30, 0x60, 0xc0, | |
168 | 0x9d, 0x27, 0x4e, 0x9c, 0x25, 0x4a, 0x94, 0x35, | |
169 | 0x6a, 0xd4, 0xb5, 0x77, 0xee, 0xc1, 0x9f, 0x23, | |
170 | 0x46, 0x8c, 0x05, 0x0a, 0x14, 0x28, 0x50, 0xa0, | |
171 | 0x5d, 0xba, 0x69, 0xd2, 0xb9, 0x6f, 0xde, 0xa1, | |
172 | 0x5f, 0xbe, 0x61, 0xc2, 0x99, 0x2f, 0x5e, 0xbc, | |
173 | 0x65, 0xca, 0x89, 0x0f, 0x1e, 0x3c, 0x78, 0xf0, | |
174 | 0xfd, 0xe7, 0xd3, 0xbb, 0x6b, 0xd6, 0xb1, 0x7f, | |
175 | 0xfe, 0xe1, 0xdf, 0xa3, 0x5b, 0xb6, 0x71, 0xe2, | |
176 | 0xd9, 0xaf, 0x43, 0x86, 0x11, 0x22, 0x44, 0x88, | |
177 | 0x0d, 0x1a, 0x34, 0x68, 0xd0, 0xbd, 0x67, 0xce, | |
178 | 0x81, 0x1f, 0x3e, 0x7c, 0xf8, 0xed, 0xc7, 0x93, | |
179 | 0x3b, 0x76, 0xec, 0xc5, 0x97, 0x33, 0x66, 0xcc, | |
180 | 0x85, 0x17, 0x2e, 0x5c, 0xb8, 0x6d, 0xda, 0xa9, | |
181 | 0x4f, 0x9e, 0x21, 0x42, 0x84, 0x15, 0x2a, 0x54, | |
182 | 0xa8, 0x4d, 0x9a, 0x29, 0x52, 0xa4, 0x55, 0xaa, | |
183 | 0x49, 0x92, 0x39, 0x72, 0xe4, 0xd5, 0xb7, 0x73, | |
184 | 0xe6, 0xd1, 0xbf, 0x63, 0xc6, 0x91, 0x3f, 0x7e, | |
185 | 0xfc, 0xe5, 0xd7, 0xb3, 0x7b, 0xf6, 0xf1, 0xff, | |
186 | 0xe3, 0xdb, 0xab, 0x4b, 0x96, 0x31, 0x62, 0xc4, | |
187 | 0x95, 0x37, 0x6e, 0xdc, 0xa5, 0x57, 0xae, 0x41, | |
188 | 0x82, 0x19, 0x32, 0x64, 0xc8, 0x8d, 0x07, 0x0e, | |
189 | 0x1c, 0x38, 0x70, 0xe0, 0xdd, 0xa7, 0x53, 0xa6, | |
190 | 0x51, 0xa2, 0x59, 0xb2, 0x79, 0xf2, 0xf9, 0xef, | |
191 | 0xc3, 0x9b, 0x2b, 0x56, 0xac, 0x45, 0x8a, 0x09, | |
192 | 0x12, 0x24, 0x48, 0x90, 0x3d, 0x7a, 0xf4, 0xf5, | |
193 | 0xf7, 0xf3, 0xfb, 0xeb, 0xcb, 0x8b, 0x0b, 0x16, | |
194 | 0x2c, 0x58, 0xb0, 0x7d, 0xfa, 0xe9, 0xcf, 0x83, | |
195 | 0x1b, 0x36, 0x6c, 0xd8, 0xad, 0x47, 0x8e, 0x01 | |
196 | }; | |
d3cc8b15 | 197 | /* Logs of 2 in the Galois field defined above. */ |
34dc7c2f BB |
198 | static const uint8_t vdev_raidz_log2[256] = { |
199 | 0x00, 0x00, 0x01, 0x19, 0x02, 0x32, 0x1a, 0xc6, | |
200 | 0x03, 0xdf, 0x33, 0xee, 0x1b, 0x68, 0xc7, 0x4b, | |
201 | 0x04, 0x64, 0xe0, 0x0e, 0x34, 0x8d, 0xef, 0x81, | |
202 | 0x1c, 0xc1, 0x69, 0xf8, 0xc8, 0x08, 0x4c, 0x71, | |
203 | 0x05, 0x8a, 0x65, 0x2f, 0xe1, 0x24, 0x0f, 0x21, | |
204 | 0x35, 0x93, 0x8e, 0xda, 0xf0, 0x12, 0x82, 0x45, | |
205 | 0x1d, 0xb5, 0xc2, 0x7d, 0x6a, 0x27, 0xf9, 0xb9, | |
206 | 0xc9, 0x9a, 0x09, 0x78, 0x4d, 0xe4, 0x72, 0xa6, | |
207 | 0x06, 0xbf, 0x8b, 0x62, 0x66, 0xdd, 0x30, 0xfd, | |
208 | 0xe2, 0x98, 0x25, 0xb3, 0x10, 0x91, 0x22, 0x88, | |
209 | 0x36, 0xd0, 0x94, 0xce, 0x8f, 0x96, 0xdb, 0xbd, | |
210 | 0xf1, 0xd2, 0x13, 0x5c, 0x83, 0x38, 0x46, 0x40, | |
211 | 0x1e, 0x42, 0xb6, 0xa3, 0xc3, 0x48, 0x7e, 0x6e, | |
212 | 0x6b, 0x3a, 0x28, 0x54, 0xfa, 0x85, 0xba, 0x3d, | |
213 | 0xca, 0x5e, 0x9b, 0x9f, 0x0a, 0x15, 0x79, 0x2b, | |
214 | 0x4e, 0xd4, 0xe5, 0xac, 0x73, 0xf3, 0xa7, 0x57, | |
215 | 0x07, 0x70, 0xc0, 0xf7, 0x8c, 0x80, 0x63, 0x0d, | |
216 | 0x67, 0x4a, 0xde, 0xed, 0x31, 0xc5, 0xfe, 0x18, | |
217 | 0xe3, 0xa5, 0x99, 0x77, 0x26, 0xb8, 0xb4, 0x7c, | |
218 | 0x11, 0x44, 0x92, 0xd9, 0x23, 0x20, 0x89, 0x2e, | |
219 | 0x37, 0x3f, 0xd1, 0x5b, 0x95, 0xbc, 0xcf, 0xcd, | |
220 | 0x90, 0x87, 0x97, 0xb2, 0xdc, 0xfc, 0xbe, 0x61, | |
221 | 0xf2, 0x56, 0xd3, 0xab, 0x14, 0x2a, 0x5d, 0x9e, | |
222 | 0x84, 0x3c, 0x39, 0x53, 0x47, 0x6d, 0x41, 0xa2, | |
223 | 0x1f, 0x2d, 0x43, 0xd8, 0xb7, 0x7b, 0xa4, 0x76, | |
224 | 0xc4, 0x17, 0x49, 0xec, 0x7f, 0x0c, 0x6f, 0xf6, | |
225 | 0x6c, 0xa1, 0x3b, 0x52, 0x29, 0x9d, 0x55, 0xaa, | |
226 | 0xfb, 0x60, 0x86, 0xb1, 0xbb, 0xcc, 0x3e, 0x5a, | |
227 | 0xcb, 0x59, 0x5f, 0xb0, 0x9c, 0xa9, 0xa0, 0x51, | |
228 | 0x0b, 0xf5, 0x16, 0xeb, 0x7a, 0x75, 0x2c, 0xd7, | |
229 | 0x4f, 0xae, 0xd5, 0xe9, 0xe6, 0xe7, 0xad, 0xe8, | |
230 | 0x74, 0xd6, 0xf4, 0xea, 0xa8, 0x50, 0x58, 0xaf, | |
231 | }; | |
232 | ||
428870ff BB |
233 | static void vdev_raidz_generate_parity(raidz_map_t *rm); |
234 | ||
34dc7c2f BB |
235 | /* |
236 | * Multiply a given number by 2 raised to the given power. | |
237 | */ | |
238 | static uint8_t | |
239 | vdev_raidz_exp2(uint_t a, int exp) | |
240 | { | |
241 | if (a == 0) | |
242 | return (0); | |
243 | ||
244 | ASSERT(exp >= 0); | |
245 | ASSERT(vdev_raidz_log2[a] > 0 || a == 1); | |
246 | ||
247 | exp += vdev_raidz_log2[a]; | |
248 | if (exp > 255) | |
249 | exp -= 255; | |
250 | ||
251 | return (vdev_raidz_pow2[exp]); | |
252 | } | |
253 | ||
b128c09f | 254 | static void |
428870ff | 255 | vdev_raidz_map_free(raidz_map_t *rm) |
b128c09f | 256 | { |
b128c09f | 257 | int c; |
428870ff | 258 | size_t size; |
b128c09f | 259 | |
428870ff | 260 | for (c = 0; c < rm->rm_firstdatacol; c++) { |
b128c09f BB |
261 | zio_buf_free(rm->rm_col[c].rc_data, rm->rm_col[c].rc_size); |
262 | ||
428870ff BB |
263 | if (rm->rm_col[c].rc_gdata != NULL) |
264 | zio_buf_free(rm->rm_col[c].rc_gdata, | |
265 | rm->rm_col[c].rc_size); | |
266 | } | |
267 | ||
268 | size = 0; | |
269 | for (c = rm->rm_firstdatacol; c < rm->rm_cols; c++) | |
270 | size += rm->rm_col[c].rc_size; | |
271 | ||
272 | if (rm->rm_datacopy != NULL) | |
273 | zio_buf_free(rm->rm_datacopy, size); | |
274 | ||
45d1cae3 | 275 | kmem_free(rm, offsetof(raidz_map_t, rm_col[rm->rm_scols])); |
b128c09f BB |
276 | } |
277 | ||
428870ff BB |
278 | static void |
279 | vdev_raidz_map_free_vsd(zio_t *zio) | |
280 | { | |
281 | raidz_map_t *rm = zio->io_vsd; | |
282 | ||
c99c9001 | 283 | ASSERT0(rm->rm_freed); |
428870ff BB |
284 | rm->rm_freed = 1; |
285 | ||
286 | if (rm->rm_reports == 0) | |
287 | vdev_raidz_map_free(rm); | |
288 | } | |
289 | ||
290 | /*ARGSUSED*/ | |
291 | static void | |
292 | vdev_raidz_cksum_free(void *arg, size_t ignored) | |
293 | { | |
294 | raidz_map_t *rm = arg; | |
295 | ||
296 | ASSERT3U(rm->rm_reports, >, 0); | |
297 | ||
298 | if (--rm->rm_reports == 0 && rm->rm_freed != 0) | |
299 | vdev_raidz_map_free(rm); | |
300 | } | |
301 | ||
302 | static void | |
303 | vdev_raidz_cksum_finish(zio_cksum_report_t *zcr, const void *good_data) | |
304 | { | |
305 | raidz_map_t *rm = zcr->zcr_cbdata; | |
306 | size_t c = zcr->zcr_cbinfo; | |
307 | size_t x; | |
308 | ||
309 | const char *good = NULL; | |
310 | const char *bad = rm->rm_col[c].rc_data; | |
311 | ||
312 | if (good_data == NULL) { | |
313 | zfs_ereport_finish_checksum(zcr, NULL, NULL, B_FALSE); | |
314 | return; | |
315 | } | |
316 | ||
317 | if (c < rm->rm_firstdatacol) { | |
318 | /* | |
319 | * The first time through, calculate the parity blocks for | |
320 | * the good data (this relies on the fact that the good | |
321 | * data never changes for a given logical ZIO) | |
322 | */ | |
323 | if (rm->rm_col[0].rc_gdata == NULL) { | |
324 | char *bad_parity[VDEV_RAIDZ_MAXPARITY]; | |
325 | char *buf; | |
326 | ||
327 | /* | |
328 | * Set up the rm_col[]s to generate the parity for | |
329 | * good_data, first saving the parity bufs and | |
330 | * replacing them with buffers to hold the result. | |
331 | */ | |
332 | for (x = 0; x < rm->rm_firstdatacol; x++) { | |
333 | bad_parity[x] = rm->rm_col[x].rc_data; | |
334 | rm->rm_col[x].rc_data = rm->rm_col[x].rc_gdata = | |
335 | zio_buf_alloc(rm->rm_col[x].rc_size); | |
336 | } | |
337 | ||
338 | /* fill in the data columns from good_data */ | |
339 | buf = (char *)good_data; | |
340 | for (; x < rm->rm_cols; x++) { | |
341 | rm->rm_col[x].rc_data = buf; | |
342 | buf += rm->rm_col[x].rc_size; | |
343 | } | |
344 | ||
345 | /* | |
346 | * Construct the parity from the good data. | |
347 | */ | |
348 | vdev_raidz_generate_parity(rm); | |
349 | ||
350 | /* restore everything back to its original state */ | |
351 | for (x = 0; x < rm->rm_firstdatacol; x++) | |
352 | rm->rm_col[x].rc_data = bad_parity[x]; | |
353 | ||
354 | buf = rm->rm_datacopy; | |
355 | for (x = rm->rm_firstdatacol; x < rm->rm_cols; x++) { | |
356 | rm->rm_col[x].rc_data = buf; | |
357 | buf += rm->rm_col[x].rc_size; | |
358 | } | |
359 | } | |
360 | ||
361 | ASSERT3P(rm->rm_col[c].rc_gdata, !=, NULL); | |
362 | good = rm->rm_col[c].rc_gdata; | |
363 | } else { | |
364 | /* adjust good_data to point at the start of our column */ | |
365 | good = good_data; | |
366 | ||
367 | for (x = rm->rm_firstdatacol; x < c; x++) | |
368 | good += rm->rm_col[x].rc_size; | |
369 | } | |
370 | ||
371 | /* we drop the ereport if it ends up that the data was good */ | |
372 | zfs_ereport_finish_checksum(zcr, good, bad, B_TRUE); | |
373 | } | |
374 | ||
375 | /* | |
376 | * Invoked indirectly by zfs_ereport_start_checksum(), called | |
377 | * below when our read operation fails completely. The main point | |
378 | * is to keep a copy of everything we read from disk, so that at | |
379 | * vdev_raidz_cksum_finish() time we can compare it with the good data. | |
380 | */ | |
381 | static void | |
382 | vdev_raidz_cksum_report(zio_t *zio, zio_cksum_report_t *zcr, void *arg) | |
383 | { | |
384 | size_t c = (size_t)(uintptr_t)arg; | |
385 | caddr_t buf; | |
386 | ||
387 | raidz_map_t *rm = zio->io_vsd; | |
388 | size_t size; | |
389 | ||
390 | /* set up the report and bump the refcount */ | |
391 | zcr->zcr_cbdata = rm; | |
392 | zcr->zcr_cbinfo = c; | |
393 | zcr->zcr_finish = vdev_raidz_cksum_finish; | |
394 | zcr->zcr_free = vdev_raidz_cksum_free; | |
395 | ||
396 | rm->rm_reports++; | |
397 | ASSERT3U(rm->rm_reports, >, 0); | |
398 | ||
399 | if (rm->rm_datacopy != NULL) | |
400 | return; | |
401 | ||
402 | /* | |
403 | * It's the first time we're called for this raidz_map_t, so we need | |
404 | * to copy the data aside; there's no guarantee that our zio's buffer | |
405 | * won't be re-used for something else. | |
406 | * | |
407 | * Our parity data is already in separate buffers, so there's no need | |
408 | * to copy them. | |
409 | */ | |
410 | ||
411 | size = 0; | |
412 | for (c = rm->rm_firstdatacol; c < rm->rm_cols; c++) | |
413 | size += rm->rm_col[c].rc_size; | |
414 | ||
415 | buf = rm->rm_datacopy = zio_buf_alloc(size); | |
416 | ||
417 | for (c = rm->rm_firstdatacol; c < rm->rm_cols; c++) { | |
418 | raidz_col_t *col = &rm->rm_col[c]; | |
419 | ||
420 | bcopy(col->rc_data, buf, col->rc_size); | |
421 | col->rc_data = buf; | |
422 | ||
423 | buf += col->rc_size; | |
424 | } | |
425 | ASSERT3P(buf - (caddr_t)rm->rm_datacopy, ==, size); | |
426 | } | |
427 | ||
428 | static const zio_vsd_ops_t vdev_raidz_vsd_ops = { | |
429 | vdev_raidz_map_free_vsd, | |
430 | vdev_raidz_cksum_report | |
431 | }; | |
432 | ||
e49f1e20 WA |
433 | /* |
434 | * Divides the IO evenly across all child vdevs; usually, dcols is | |
435 | * the number of children in the target vdev. | |
a1687880 BB |
436 | * |
437 | * Avoid inlining the function to keep vdev_raidz_io_start(), which | |
438 | * is this functions only caller, as small as possible on the stack. | |
e49f1e20 | 439 | */ |
a1687880 | 440 | noinline static raidz_map_t * |
34dc7c2f BB |
441 | vdev_raidz_map_alloc(zio_t *zio, uint64_t unit_shift, uint64_t dcols, |
442 | uint64_t nparity) | |
443 | { | |
444 | raidz_map_t *rm; | |
e49f1e20 | 445 | /* The starting RAIDZ (parent) vdev sector of the block. */ |
34dc7c2f | 446 | uint64_t b = zio->io_offset >> unit_shift; |
e49f1e20 | 447 | /* The zio's size in units of the vdev's minimum sector size. */ |
34dc7c2f | 448 | uint64_t s = zio->io_size >> unit_shift; |
e49f1e20 | 449 | /* The first column for this stripe. */ |
34dc7c2f | 450 | uint64_t f = b % dcols; |
e49f1e20 | 451 | /* The starting byte offset on each child vdev. */ |
34dc7c2f | 452 | uint64_t o = (b / dcols) << unit_shift; |
45d1cae3 | 453 | uint64_t q, r, c, bc, col, acols, scols, coff, devidx, asize, tot; |
34dc7c2f | 454 | |
e49f1e20 WA |
455 | /* |
456 | * "Quotient": The number of data sectors for this stripe on all but | |
457 | * the "big column" child vdevs that also contain "remainder" data. | |
458 | */ | |
34dc7c2f | 459 | q = s / (dcols - nparity); |
e49f1e20 WA |
460 | |
461 | /* | |
462 | * "Remainder": The number of partial stripe data sectors in this I/O. | |
463 | * This will add a sector to some, but not all, child vdevs. | |
464 | */ | |
34dc7c2f | 465 | r = s - q * (dcols - nparity); |
e49f1e20 WA |
466 | |
467 | /* The number of "big columns" - those which contain remainder data. */ | |
34dc7c2f | 468 | bc = (r == 0 ? 0 : r + nparity); |
e49f1e20 WA |
469 | |
470 | /* | |
471 | * The total number of data and parity sectors associated with | |
472 | * this I/O. | |
473 | */ | |
45d1cae3 BB |
474 | tot = s + nparity * (q + (r == 0 ? 0 : 1)); |
475 | ||
e49f1e20 WA |
476 | /* acols: The columns that will be accessed. */ |
477 | /* scols: The columns that will be accessed or skipped. */ | |
45d1cae3 | 478 | if (q == 0) { |
e49f1e20 | 479 | /* Our I/O request doesn't span all child vdevs. */ |
45d1cae3 BB |
480 | acols = bc; |
481 | scols = MIN(dcols, roundup(bc, nparity + 1)); | |
482 | } else { | |
483 | acols = dcols; | |
484 | scols = dcols; | |
485 | } | |
34dc7c2f | 486 | |
45d1cae3 | 487 | ASSERT3U(acols, <=, scols); |
34dc7c2f | 488 | |
b8d06fca | 489 | rm = kmem_alloc(offsetof(raidz_map_t, rm_col[scols]), KM_PUSHPAGE); |
34dc7c2f BB |
490 | |
491 | rm->rm_cols = acols; | |
45d1cae3 | 492 | rm->rm_scols = scols; |
34dc7c2f | 493 | rm->rm_bigcols = bc; |
428870ff | 494 | rm->rm_skipstart = bc; |
34dc7c2f BB |
495 | rm->rm_missingdata = 0; |
496 | rm->rm_missingparity = 0; | |
497 | rm->rm_firstdatacol = nparity; | |
428870ff BB |
498 | rm->rm_datacopy = NULL; |
499 | rm->rm_reports = 0; | |
500 | rm->rm_freed = 0; | |
501 | rm->rm_ecksuminjected = 0; | |
34dc7c2f | 502 | |
45d1cae3 BB |
503 | asize = 0; |
504 | ||
505 | for (c = 0; c < scols; c++) { | |
34dc7c2f BB |
506 | col = f + c; |
507 | coff = o; | |
508 | if (col >= dcols) { | |
509 | col -= dcols; | |
510 | coff += 1ULL << unit_shift; | |
511 | } | |
512 | rm->rm_col[c].rc_devidx = col; | |
513 | rm->rm_col[c].rc_offset = coff; | |
34dc7c2f | 514 | rm->rm_col[c].rc_data = NULL; |
428870ff | 515 | rm->rm_col[c].rc_gdata = NULL; |
34dc7c2f BB |
516 | rm->rm_col[c].rc_error = 0; |
517 | rm->rm_col[c].rc_tried = 0; | |
518 | rm->rm_col[c].rc_skipped = 0; | |
45d1cae3 BB |
519 | |
520 | if (c >= acols) | |
521 | rm->rm_col[c].rc_size = 0; | |
522 | else if (c < bc) | |
523 | rm->rm_col[c].rc_size = (q + 1) << unit_shift; | |
524 | else | |
525 | rm->rm_col[c].rc_size = q << unit_shift; | |
526 | ||
527 | asize += rm->rm_col[c].rc_size; | |
34dc7c2f BB |
528 | } |
529 | ||
45d1cae3 BB |
530 | ASSERT3U(asize, ==, tot << unit_shift); |
531 | rm->rm_asize = roundup(asize, (nparity + 1) << unit_shift); | |
428870ff BB |
532 | rm->rm_nskip = roundup(tot, nparity + 1) - tot; |
533 | ASSERT3U(rm->rm_asize - asize, ==, rm->rm_nskip << unit_shift); | |
534 | ASSERT3U(rm->rm_nskip, <=, nparity); | |
34dc7c2f BB |
535 | |
536 | for (c = 0; c < rm->rm_firstdatacol; c++) | |
537 | rm->rm_col[c].rc_data = zio_buf_alloc(rm->rm_col[c].rc_size); | |
538 | ||
539 | rm->rm_col[c].rc_data = zio->io_data; | |
540 | ||
541 | for (c = c + 1; c < acols; c++) | |
542 | rm->rm_col[c].rc_data = (char *)rm->rm_col[c - 1].rc_data + | |
543 | rm->rm_col[c - 1].rc_size; | |
544 | ||
545 | /* | |
546 | * If all data stored spans all columns, there's a danger that parity | |
547 | * will always be on the same device and, since parity isn't read | |
548 | * during normal operation, that that device's I/O bandwidth won't be | |
549 | * used effectively. We therefore switch the parity every 1MB. | |
550 | * | |
551 | * ... at least that was, ostensibly, the theory. As a practical | |
552 | * matter unless we juggle the parity between all devices evenly, we | |
553 | * won't see any benefit. Further, occasional writes that aren't a | |
554 | * multiple of the LCM of the number of children and the minimum | |
555 | * stripe width are sufficient to avoid pessimal behavior. | |
556 | * Unfortunately, this decision created an implicit on-disk format | |
557 | * requirement that we need to support for all eternity, but only | |
558 | * for single-parity RAID-Z. | |
428870ff BB |
559 | * |
560 | * If we intend to skip a sector in the zeroth column for padding | |
561 | * we must make sure to note this swap. We will never intend to | |
562 | * skip the first column since at least one data and one parity | |
563 | * column must appear in each row. | |
34dc7c2f BB |
564 | */ |
565 | ASSERT(rm->rm_cols >= 2); | |
566 | ASSERT(rm->rm_col[0].rc_size == rm->rm_col[1].rc_size); | |
567 | ||
568 | if (rm->rm_firstdatacol == 1 && (zio->io_offset & (1ULL << 20))) { | |
569 | devidx = rm->rm_col[0].rc_devidx; | |
570 | o = rm->rm_col[0].rc_offset; | |
571 | rm->rm_col[0].rc_devidx = rm->rm_col[1].rc_devidx; | |
572 | rm->rm_col[0].rc_offset = rm->rm_col[1].rc_offset; | |
573 | rm->rm_col[1].rc_devidx = devidx; | |
574 | rm->rm_col[1].rc_offset = o; | |
428870ff BB |
575 | |
576 | if (rm->rm_skipstart == 0) | |
577 | rm->rm_skipstart = 1; | |
34dc7c2f BB |
578 | } |
579 | ||
580 | zio->io_vsd = rm; | |
428870ff | 581 | zio->io_vsd_ops = &vdev_raidz_vsd_ops; |
34dc7c2f BB |
582 | return (rm); |
583 | } | |
584 | ||
34dc7c2f BB |
585 | static void |
586 | vdev_raidz_generate_parity_p(raidz_map_t *rm) | |
587 | { | |
588 | uint64_t *p, *src, pcount, ccount, i; | |
589 | int c; | |
590 | ||
591 | pcount = rm->rm_col[VDEV_RAIDZ_P].rc_size / sizeof (src[0]); | |
592 | ||
593 | for (c = rm->rm_firstdatacol; c < rm->rm_cols; c++) { | |
594 | src = rm->rm_col[c].rc_data; | |
595 | p = rm->rm_col[VDEV_RAIDZ_P].rc_data; | |
596 | ccount = rm->rm_col[c].rc_size / sizeof (src[0]); | |
597 | ||
598 | if (c == rm->rm_firstdatacol) { | |
599 | ASSERT(ccount == pcount); | |
45d1cae3 | 600 | for (i = 0; i < ccount; i++, src++, p++) { |
34dc7c2f BB |
601 | *p = *src; |
602 | } | |
603 | } else { | |
604 | ASSERT(ccount <= pcount); | |
45d1cae3 | 605 | for (i = 0; i < ccount; i++, src++, p++) { |
34dc7c2f BB |
606 | *p ^= *src; |
607 | } | |
608 | } | |
609 | } | |
610 | } | |
611 | ||
612 | static void | |
613 | vdev_raidz_generate_parity_pq(raidz_map_t *rm) | |
614 | { | |
45d1cae3 | 615 | uint64_t *p, *q, *src, pcnt, ccnt, mask, i; |
34dc7c2f BB |
616 | int c; |
617 | ||
45d1cae3 | 618 | pcnt = rm->rm_col[VDEV_RAIDZ_P].rc_size / sizeof (src[0]); |
34dc7c2f BB |
619 | ASSERT(rm->rm_col[VDEV_RAIDZ_P].rc_size == |
620 | rm->rm_col[VDEV_RAIDZ_Q].rc_size); | |
621 | ||
622 | for (c = rm->rm_firstdatacol; c < rm->rm_cols; c++) { | |
623 | src = rm->rm_col[c].rc_data; | |
624 | p = rm->rm_col[VDEV_RAIDZ_P].rc_data; | |
625 | q = rm->rm_col[VDEV_RAIDZ_Q].rc_data; | |
45d1cae3 BB |
626 | |
627 | ccnt = rm->rm_col[c].rc_size / sizeof (src[0]); | |
34dc7c2f BB |
628 | |
629 | if (c == rm->rm_firstdatacol) { | |
45d1cae3 BB |
630 | ASSERT(ccnt == pcnt || ccnt == 0); |
631 | for (i = 0; i < ccnt; i++, src++, p++, q++) { | |
34dc7c2f | 632 | *p = *src; |
45d1cae3 | 633 | *q = *src; |
34dc7c2f | 634 | } |
45d1cae3 | 635 | for (; i < pcnt; i++, src++, p++, q++) { |
34dc7c2f | 636 | *p = 0; |
45d1cae3 | 637 | *q = 0; |
34dc7c2f BB |
638 | } |
639 | } else { | |
45d1cae3 | 640 | ASSERT(ccnt <= pcnt); |
34dc7c2f BB |
641 | |
642 | /* | |
45d1cae3 BB |
643 | * Apply the algorithm described above by multiplying |
644 | * the previous result and adding in the new value. | |
34dc7c2f | 645 | */ |
45d1cae3 BB |
646 | for (i = 0; i < ccnt; i++, src++, p++, q++) { |
647 | *p ^= *src; | |
648 | ||
649 | VDEV_RAIDZ_64MUL_2(*q, mask); | |
34dc7c2f | 650 | *q ^= *src; |
45d1cae3 BB |
651 | } |
652 | ||
653 | /* | |
654 | * Treat short columns as though they are full of 0s. | |
655 | * Note that there's therefore nothing needed for P. | |
656 | */ | |
657 | for (; i < pcnt; i++, q++) { | |
658 | VDEV_RAIDZ_64MUL_2(*q, mask); | |
659 | } | |
660 | } | |
661 | } | |
662 | } | |
663 | ||
664 | static void | |
665 | vdev_raidz_generate_parity_pqr(raidz_map_t *rm) | |
666 | { | |
667 | uint64_t *p, *q, *r, *src, pcnt, ccnt, mask, i; | |
668 | int c; | |
669 | ||
670 | pcnt = rm->rm_col[VDEV_RAIDZ_P].rc_size / sizeof (src[0]); | |
671 | ASSERT(rm->rm_col[VDEV_RAIDZ_P].rc_size == | |
672 | rm->rm_col[VDEV_RAIDZ_Q].rc_size); | |
673 | ASSERT(rm->rm_col[VDEV_RAIDZ_P].rc_size == | |
674 | rm->rm_col[VDEV_RAIDZ_R].rc_size); | |
675 | ||
676 | for (c = rm->rm_firstdatacol; c < rm->rm_cols; c++) { | |
677 | src = rm->rm_col[c].rc_data; | |
678 | p = rm->rm_col[VDEV_RAIDZ_P].rc_data; | |
679 | q = rm->rm_col[VDEV_RAIDZ_Q].rc_data; | |
680 | r = rm->rm_col[VDEV_RAIDZ_R].rc_data; | |
681 | ||
682 | ccnt = rm->rm_col[c].rc_size / sizeof (src[0]); | |
683 | ||
684 | if (c == rm->rm_firstdatacol) { | |
685 | ASSERT(ccnt == pcnt || ccnt == 0); | |
686 | for (i = 0; i < ccnt; i++, src++, p++, q++, r++) { | |
687 | *p = *src; | |
688 | *q = *src; | |
689 | *r = *src; | |
690 | } | |
691 | for (; i < pcnt; i++, src++, p++, q++, r++) { | |
692 | *p = 0; | |
693 | *q = 0; | |
694 | *r = 0; | |
695 | } | |
696 | } else { | |
697 | ASSERT(ccnt <= pcnt); | |
698 | ||
699 | /* | |
700 | * Apply the algorithm described above by multiplying | |
701 | * the previous result and adding in the new value. | |
702 | */ | |
703 | for (i = 0; i < ccnt; i++, src++, p++, q++, r++) { | |
34dc7c2f | 704 | *p ^= *src; |
45d1cae3 BB |
705 | |
706 | VDEV_RAIDZ_64MUL_2(*q, mask); | |
707 | *q ^= *src; | |
708 | ||
709 | VDEV_RAIDZ_64MUL_4(*r, mask); | |
710 | *r ^= *src; | |
34dc7c2f BB |
711 | } |
712 | ||
713 | /* | |
714 | * Treat short columns as though they are full of 0s. | |
45d1cae3 | 715 | * Note that there's therefore nothing needed for P. |
34dc7c2f | 716 | */ |
45d1cae3 BB |
717 | for (; i < pcnt; i++, q++, r++) { |
718 | VDEV_RAIDZ_64MUL_2(*q, mask); | |
719 | VDEV_RAIDZ_64MUL_4(*r, mask); | |
34dc7c2f BB |
720 | } |
721 | } | |
722 | } | |
723 | } | |
724 | ||
45d1cae3 BB |
725 | /* |
726 | * Generate RAID parity in the first virtual columns according to the number of | |
727 | * parity columns available. | |
728 | */ | |
34dc7c2f | 729 | static void |
45d1cae3 BB |
730 | vdev_raidz_generate_parity(raidz_map_t *rm) |
731 | { | |
732 | switch (rm->rm_firstdatacol) { | |
733 | case 1: | |
734 | vdev_raidz_generate_parity_p(rm); | |
735 | break; | |
736 | case 2: | |
737 | vdev_raidz_generate_parity_pq(rm); | |
738 | break; | |
739 | case 3: | |
740 | vdev_raidz_generate_parity_pqr(rm); | |
741 | break; | |
742 | default: | |
743 | cmn_err(CE_PANIC, "invalid RAID-Z configuration"); | |
744 | } | |
745 | } | |
746 | ||
747 | static int | |
748 | vdev_raidz_reconstruct_p(raidz_map_t *rm, int *tgts, int ntgts) | |
34dc7c2f BB |
749 | { |
750 | uint64_t *dst, *src, xcount, ccount, count, i; | |
45d1cae3 | 751 | int x = tgts[0]; |
34dc7c2f BB |
752 | int c; |
753 | ||
45d1cae3 BB |
754 | ASSERT(ntgts == 1); |
755 | ASSERT(x >= rm->rm_firstdatacol); | |
756 | ASSERT(x < rm->rm_cols); | |
757 | ||
34dc7c2f BB |
758 | xcount = rm->rm_col[x].rc_size / sizeof (src[0]); |
759 | ASSERT(xcount <= rm->rm_col[VDEV_RAIDZ_P].rc_size / sizeof (src[0])); | |
760 | ASSERT(xcount > 0); | |
761 | ||
762 | src = rm->rm_col[VDEV_RAIDZ_P].rc_data; | |
763 | dst = rm->rm_col[x].rc_data; | |
764 | for (i = 0; i < xcount; i++, dst++, src++) { | |
765 | *dst = *src; | |
766 | } | |
767 | ||
768 | for (c = rm->rm_firstdatacol; c < rm->rm_cols; c++) { | |
769 | src = rm->rm_col[c].rc_data; | |
770 | dst = rm->rm_col[x].rc_data; | |
771 | ||
772 | if (c == x) | |
773 | continue; | |
774 | ||
775 | ccount = rm->rm_col[c].rc_size / sizeof (src[0]); | |
776 | count = MIN(ccount, xcount); | |
777 | ||
778 | for (i = 0; i < count; i++, dst++, src++) { | |
779 | *dst ^= *src; | |
780 | } | |
781 | } | |
45d1cae3 BB |
782 | |
783 | return (1 << VDEV_RAIDZ_P); | |
34dc7c2f BB |
784 | } |
785 | ||
45d1cae3 BB |
786 | static int |
787 | vdev_raidz_reconstruct_q(raidz_map_t *rm, int *tgts, int ntgts) | |
34dc7c2f BB |
788 | { |
789 | uint64_t *dst, *src, xcount, ccount, count, mask, i; | |
790 | uint8_t *b; | |
45d1cae3 | 791 | int x = tgts[0]; |
34dc7c2f BB |
792 | int c, j, exp; |
793 | ||
45d1cae3 BB |
794 | ASSERT(ntgts == 1); |
795 | ||
34dc7c2f BB |
796 | xcount = rm->rm_col[x].rc_size / sizeof (src[0]); |
797 | ASSERT(xcount <= rm->rm_col[VDEV_RAIDZ_Q].rc_size / sizeof (src[0])); | |
798 | ||
799 | for (c = rm->rm_firstdatacol; c < rm->rm_cols; c++) { | |
800 | src = rm->rm_col[c].rc_data; | |
801 | dst = rm->rm_col[x].rc_data; | |
802 | ||
803 | if (c == x) | |
804 | ccount = 0; | |
805 | else | |
806 | ccount = rm->rm_col[c].rc_size / sizeof (src[0]); | |
807 | ||
808 | count = MIN(ccount, xcount); | |
809 | ||
810 | if (c == rm->rm_firstdatacol) { | |
811 | for (i = 0; i < count; i++, dst++, src++) { | |
812 | *dst = *src; | |
813 | } | |
814 | for (; i < xcount; i++, dst++) { | |
815 | *dst = 0; | |
816 | } | |
817 | ||
818 | } else { | |
34dc7c2f | 819 | for (i = 0; i < count; i++, dst++, src++) { |
45d1cae3 | 820 | VDEV_RAIDZ_64MUL_2(*dst, mask); |
34dc7c2f BB |
821 | *dst ^= *src; |
822 | } | |
823 | ||
824 | for (; i < xcount; i++, dst++) { | |
45d1cae3 | 825 | VDEV_RAIDZ_64MUL_2(*dst, mask); |
34dc7c2f BB |
826 | } |
827 | } | |
828 | } | |
829 | ||
830 | src = rm->rm_col[VDEV_RAIDZ_Q].rc_data; | |
831 | dst = rm->rm_col[x].rc_data; | |
832 | exp = 255 - (rm->rm_cols - 1 - x); | |
833 | ||
834 | for (i = 0; i < xcount; i++, dst++, src++) { | |
835 | *dst ^= *src; | |
836 | for (j = 0, b = (uint8_t *)dst; j < 8; j++, b++) { | |
837 | *b = vdev_raidz_exp2(*b, exp); | |
838 | } | |
839 | } | |
45d1cae3 BB |
840 | |
841 | return (1 << VDEV_RAIDZ_Q); | |
34dc7c2f BB |
842 | } |
843 | ||
45d1cae3 BB |
844 | static int |
845 | vdev_raidz_reconstruct_pq(raidz_map_t *rm, int *tgts, int ntgts) | |
34dc7c2f BB |
846 | { |
847 | uint8_t *p, *q, *pxy, *qxy, *xd, *yd, tmp, a, b, aexp, bexp; | |
848 | void *pdata, *qdata; | |
849 | uint64_t xsize, ysize, i; | |
45d1cae3 BB |
850 | int x = tgts[0]; |
851 | int y = tgts[1]; | |
34dc7c2f | 852 | |
45d1cae3 | 853 | ASSERT(ntgts == 2); |
34dc7c2f BB |
854 | ASSERT(x < y); |
855 | ASSERT(x >= rm->rm_firstdatacol); | |
856 | ASSERT(y < rm->rm_cols); | |
857 | ||
858 | ASSERT(rm->rm_col[x].rc_size >= rm->rm_col[y].rc_size); | |
859 | ||
860 | /* | |
861 | * Move the parity data aside -- we're going to compute parity as | |
862 | * though columns x and y were full of zeros -- Pxy and Qxy. We want to | |
863 | * reuse the parity generation mechanism without trashing the actual | |
864 | * parity so we make those columns appear to be full of zeros by | |
865 | * setting their lengths to zero. | |
866 | */ | |
867 | pdata = rm->rm_col[VDEV_RAIDZ_P].rc_data; | |
868 | qdata = rm->rm_col[VDEV_RAIDZ_Q].rc_data; | |
869 | xsize = rm->rm_col[x].rc_size; | |
870 | ysize = rm->rm_col[y].rc_size; | |
871 | ||
872 | rm->rm_col[VDEV_RAIDZ_P].rc_data = | |
873 | zio_buf_alloc(rm->rm_col[VDEV_RAIDZ_P].rc_size); | |
874 | rm->rm_col[VDEV_RAIDZ_Q].rc_data = | |
875 | zio_buf_alloc(rm->rm_col[VDEV_RAIDZ_Q].rc_size); | |
876 | rm->rm_col[x].rc_size = 0; | |
877 | rm->rm_col[y].rc_size = 0; | |
878 | ||
879 | vdev_raidz_generate_parity_pq(rm); | |
880 | ||
881 | rm->rm_col[x].rc_size = xsize; | |
882 | rm->rm_col[y].rc_size = ysize; | |
883 | ||
884 | p = pdata; | |
885 | q = qdata; | |
886 | pxy = rm->rm_col[VDEV_RAIDZ_P].rc_data; | |
887 | qxy = rm->rm_col[VDEV_RAIDZ_Q].rc_data; | |
888 | xd = rm->rm_col[x].rc_data; | |
889 | yd = rm->rm_col[y].rc_data; | |
890 | ||
891 | /* | |
892 | * We now have: | |
893 | * Pxy = P + D_x + D_y | |
894 | * Qxy = Q + 2^(ndevs - 1 - x) * D_x + 2^(ndevs - 1 - y) * D_y | |
895 | * | |
896 | * We can then solve for D_x: | |
897 | * D_x = A * (P + Pxy) + B * (Q + Qxy) | |
898 | * where | |
899 | * A = 2^(x - y) * (2^(x - y) + 1)^-1 | |
900 | * B = 2^(ndevs - 1 - x) * (2^(x - y) + 1)^-1 | |
901 | * | |
902 | * With D_x in hand, we can easily solve for D_y: | |
903 | * D_y = P + Pxy + D_x | |
904 | */ | |
905 | ||
906 | a = vdev_raidz_pow2[255 + x - y]; | |
907 | b = vdev_raidz_pow2[255 - (rm->rm_cols - 1 - x)]; | |
908 | tmp = 255 - vdev_raidz_log2[a ^ 1]; | |
909 | ||
910 | aexp = vdev_raidz_log2[vdev_raidz_exp2(a, tmp)]; | |
911 | bexp = vdev_raidz_log2[vdev_raidz_exp2(b, tmp)]; | |
912 | ||
913 | for (i = 0; i < xsize; i++, p++, q++, pxy++, qxy++, xd++, yd++) { | |
914 | *xd = vdev_raidz_exp2(*p ^ *pxy, aexp) ^ | |
915 | vdev_raidz_exp2(*q ^ *qxy, bexp); | |
916 | ||
917 | if (i < ysize) | |
918 | *yd = *p ^ *pxy ^ *xd; | |
919 | } | |
920 | ||
921 | zio_buf_free(rm->rm_col[VDEV_RAIDZ_P].rc_data, | |
922 | rm->rm_col[VDEV_RAIDZ_P].rc_size); | |
923 | zio_buf_free(rm->rm_col[VDEV_RAIDZ_Q].rc_data, | |
924 | rm->rm_col[VDEV_RAIDZ_Q].rc_size); | |
925 | ||
926 | /* | |
927 | * Restore the saved parity data. | |
928 | */ | |
929 | rm->rm_col[VDEV_RAIDZ_P].rc_data = pdata; | |
930 | rm->rm_col[VDEV_RAIDZ_Q].rc_data = qdata; | |
45d1cae3 BB |
931 | |
932 | return ((1 << VDEV_RAIDZ_P) | (1 << VDEV_RAIDZ_Q)); | |
933 | } | |
934 | ||
935 | /* BEGIN CSTYLED */ | |
936 | /* | |
937 | * In the general case of reconstruction, we must solve the system of linear | |
938 | * equations defined by the coeffecients used to generate parity as well as | |
939 | * the contents of the data and parity disks. This can be expressed with | |
940 | * vectors for the original data (D) and the actual data (d) and parity (p) | |
941 | * and a matrix composed of the identity matrix (I) and a dispersal matrix (V): | |
942 | * | |
943 | * __ __ __ __ | |
944 | * | | __ __ | p_0 | | |
945 | * | V | | D_0 | | p_m-1 | | |
946 | * | | x | : | = | d_0 | | |
947 | * | I | | D_n-1 | | : | | |
948 | * | | ~~ ~~ | d_n-1 | | |
949 | * ~~ ~~ ~~ ~~ | |
950 | * | |
951 | * I is simply a square identity matrix of size n, and V is a vandermonde | |
952 | * matrix defined by the coeffecients we chose for the various parity columns | |
953 | * (1, 2, 4). Note that these values were chosen both for simplicity, speedy | |
954 | * computation as well as linear separability. | |
955 | * | |
956 | * __ __ __ __ | |
957 | * | 1 .. 1 1 1 | | p_0 | | |
958 | * | 2^n-1 .. 4 2 1 | __ __ | : | | |
959 | * | 4^n-1 .. 16 4 1 | | D_0 | | p_m-1 | | |
960 | * | 1 .. 0 0 0 | | D_1 | | d_0 | | |
961 | * | 0 .. 0 0 0 | x | D_2 | = | d_1 | | |
962 | * | : : : : | | : | | d_2 | | |
963 | * | 0 .. 1 0 0 | | D_n-1 | | : | | |
964 | * | 0 .. 0 1 0 | ~~ ~~ | : | | |
965 | * | 0 .. 0 0 1 | | d_n-1 | | |
966 | * ~~ ~~ ~~ ~~ | |
967 | * | |
968 | * Note that I, V, d, and p are known. To compute D, we must invert the | |
969 | * matrix and use the known data and parity values to reconstruct the unknown | |
970 | * data values. We begin by removing the rows in V|I and d|p that correspond | |
971 | * to failed or missing columns; we then make V|I square (n x n) and d|p | |
972 | * sized n by removing rows corresponding to unused parity from the bottom up | |
973 | * to generate (V|I)' and (d|p)'. We can then generate the inverse of (V|I)' | |
974 | * using Gauss-Jordan elimination. In the example below we use m=3 parity | |
975 | * columns, n=8 data columns, with errors in d_1, d_2, and p_1: | |
976 | * __ __ | |
977 | * | 1 1 1 1 1 1 1 1 | | |
978 | * | 128 64 32 16 8 4 2 1 | <-----+-+-- missing disks | |
979 | * | 19 205 116 29 64 16 4 1 | / / | |
980 | * | 1 0 0 0 0 0 0 0 | / / | |
981 | * | 0 1 0 0 0 0 0 0 | <--' / | |
982 | * (V|I) = | 0 0 1 0 0 0 0 0 | <---' | |
983 | * | 0 0 0 1 0 0 0 0 | | |
984 | * | 0 0 0 0 1 0 0 0 | | |
985 | * | 0 0 0 0 0 1 0 0 | | |
986 | * | 0 0 0 0 0 0 1 0 | | |
987 | * | 0 0 0 0 0 0 0 1 | | |
988 | * ~~ ~~ | |
989 | * __ __ | |
990 | * | 1 1 1 1 1 1 1 1 | | |
991 | * | 128 64 32 16 8 4 2 1 | | |
992 | * | 19 205 116 29 64 16 4 1 | | |
993 | * | 1 0 0 0 0 0 0 0 | | |
994 | * | 0 1 0 0 0 0 0 0 | | |
995 | * (V|I)' = | 0 0 1 0 0 0 0 0 | | |
996 | * | 0 0 0 1 0 0 0 0 | | |
997 | * | 0 0 0 0 1 0 0 0 | | |
998 | * | 0 0 0 0 0 1 0 0 | | |
999 | * | 0 0 0 0 0 0 1 0 | | |
1000 | * | 0 0 0 0 0 0 0 1 | | |
1001 | * ~~ ~~ | |
1002 | * | |
1003 | * Here we employ Gauss-Jordan elimination to find the inverse of (V|I)'. We | |
1004 | * have carefully chosen the seed values 1, 2, and 4 to ensure that this | |
1005 | * matrix is not singular. | |
1006 | * __ __ | |
1007 | * | 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 | | |
1008 | * | 19 205 116 29 64 16 4 1 0 1 0 0 0 0 0 0 | | |
1009 | * | 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 | | |
1010 | * | 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 | | |
1011 | * | 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 | | |
1012 | * | 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 | | |
1013 | * | 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 | | |
1014 | * | 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 | | |
1015 | * ~~ ~~ | |
1016 | * __ __ | |
1017 | * | 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 | | |
1018 | * | 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 | | |
1019 | * | 19 205 116 29 64 16 4 1 0 1 0 0 0 0 0 0 | | |
1020 | * | 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 | | |
1021 | * | 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 | | |
1022 | * | 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 | | |
1023 | * | 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 | | |
1024 | * | 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 | | |
1025 | * ~~ ~~ | |
1026 | * __ __ | |
1027 | * | 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 | | |
1028 | * | 0 1 1 0 0 0 0 0 1 0 1 1 1 1 1 1 | | |
1029 | * | 0 205 116 0 0 0 0 0 0 1 19 29 64 16 4 1 | | |
1030 | * | 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 | | |
1031 | * | 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 | | |
1032 | * | 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 | | |
1033 | * | 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 | | |
1034 | * | 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 | | |
1035 | * ~~ ~~ | |
1036 | * __ __ | |
1037 | * | 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 | | |
1038 | * | 0 1 1 0 0 0 0 0 1 0 1 1 1 1 1 1 | | |
1039 | * | 0 0 185 0 0 0 0 0 205 1 222 208 141 221 201 204 | | |
1040 | * | 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 | | |
1041 | * | 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 | | |
1042 | * | 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 | | |
1043 | * | 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 | | |
1044 | * | 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 | | |
1045 | * ~~ ~~ | |
1046 | * __ __ | |
1047 | * | 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 | | |
1048 | * | 0 1 1 0 0 0 0 0 1 0 1 1 1 1 1 1 | | |
1049 | * | 0 0 1 0 0 0 0 0 166 100 4 40 158 168 216 209 | | |
1050 | * | 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 | | |
1051 | * | 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 | | |
1052 | * | 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 | | |
1053 | * | 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 | | |
1054 | * | 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 | | |
1055 | * ~~ ~~ | |
1056 | * __ __ | |
1057 | * | 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 | | |
1058 | * | 0 1 0 0 0 0 0 0 167 100 5 41 159 169 217 208 | | |
1059 | * | 0 0 1 0 0 0 0 0 166 100 4 40 158 168 216 209 | | |
1060 | * | 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 | | |
1061 | * | 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 | | |
1062 | * | 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 | | |
1063 | * | 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 | | |
1064 | * | 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 | | |
1065 | * ~~ ~~ | |
1066 | * __ __ | |
1067 | * | 0 0 1 0 0 0 0 0 | | |
1068 | * | 167 100 5 41 159 169 217 208 | | |
1069 | * | 166 100 4 40 158 168 216 209 | | |
1070 | * (V|I)'^-1 = | 0 0 0 1 0 0 0 0 | | |
1071 | * | 0 0 0 0 1 0 0 0 | | |
1072 | * | 0 0 0 0 0 1 0 0 | | |
1073 | * | 0 0 0 0 0 0 1 0 | | |
1074 | * | 0 0 0 0 0 0 0 1 | | |
1075 | * ~~ ~~ | |
1076 | * | |
1077 | * We can then simply compute D = (V|I)'^-1 x (d|p)' to discover the values | |
1078 | * of the missing data. | |
1079 | * | |
1080 | * As is apparent from the example above, the only non-trivial rows in the | |
1081 | * inverse matrix correspond to the data disks that we're trying to | |
1082 | * reconstruct. Indeed, those are the only rows we need as the others would | |
1083 | * only be useful for reconstructing data known or assumed to be valid. For | |
1084 | * that reason, we only build the coefficients in the rows that correspond to | |
1085 | * targeted columns. | |
1086 | */ | |
1087 | /* END CSTYLED */ | |
1088 | ||
1089 | static void | |
1090 | vdev_raidz_matrix_init(raidz_map_t *rm, int n, int nmap, int *map, | |
1091 | uint8_t **rows) | |
1092 | { | |
1093 | int i, j; | |
1094 | int pow; | |
1095 | ||
1096 | ASSERT(n == rm->rm_cols - rm->rm_firstdatacol); | |
1097 | ||
1098 | /* | |
1099 | * Fill in the missing rows of interest. | |
1100 | */ | |
1101 | for (i = 0; i < nmap; i++) { | |
1102 | ASSERT3S(0, <=, map[i]); | |
1103 | ASSERT3S(map[i], <=, 2); | |
1104 | ||
1105 | pow = map[i] * n; | |
1106 | if (pow > 255) | |
1107 | pow -= 255; | |
1108 | ASSERT(pow <= 255); | |
1109 | ||
1110 | for (j = 0; j < n; j++) { | |
1111 | pow -= map[i]; | |
1112 | if (pow < 0) | |
1113 | pow += 255; | |
1114 | rows[i][j] = vdev_raidz_pow2[pow]; | |
1115 | } | |
1116 | } | |
1117 | } | |
1118 | ||
1119 | static void | |
1120 | vdev_raidz_matrix_invert(raidz_map_t *rm, int n, int nmissing, int *missing, | |
1121 | uint8_t **rows, uint8_t **invrows, const uint8_t *used) | |
1122 | { | |
1123 | int i, j, ii, jj; | |
1124 | uint8_t log; | |
1125 | ||
1126 | /* | |
1127 | * Assert that the first nmissing entries from the array of used | |
1128 | * columns correspond to parity columns and that subsequent entries | |
1129 | * correspond to data columns. | |
1130 | */ | |
1131 | for (i = 0; i < nmissing; i++) { | |
1132 | ASSERT3S(used[i], <, rm->rm_firstdatacol); | |
1133 | } | |
1134 | for (; i < n; i++) { | |
1135 | ASSERT3S(used[i], >=, rm->rm_firstdatacol); | |
1136 | } | |
1137 | ||
1138 | /* | |
1139 | * First initialize the storage where we'll compute the inverse rows. | |
1140 | */ | |
1141 | for (i = 0; i < nmissing; i++) { | |
1142 | for (j = 0; j < n; j++) { | |
1143 | invrows[i][j] = (i == j) ? 1 : 0; | |
1144 | } | |
1145 | } | |
1146 | ||
1147 | /* | |
1148 | * Subtract all trivial rows from the rows of consequence. | |
1149 | */ | |
1150 | for (i = 0; i < nmissing; i++) { | |
1151 | for (j = nmissing; j < n; j++) { | |
1152 | ASSERT3U(used[j], >=, rm->rm_firstdatacol); | |
1153 | jj = used[j] - rm->rm_firstdatacol; | |
1154 | ASSERT3S(jj, <, n); | |
1155 | invrows[i][j] = rows[i][jj]; | |
1156 | rows[i][jj] = 0; | |
1157 | } | |
1158 | } | |
1159 | ||
1160 | /* | |
1161 | * For each of the rows of interest, we must normalize it and subtract | |
1162 | * a multiple of it from the other rows. | |
1163 | */ | |
1164 | for (i = 0; i < nmissing; i++) { | |
1165 | for (j = 0; j < missing[i]; j++) { | |
c99c9001 | 1166 | ASSERT0(rows[i][j]); |
45d1cae3 BB |
1167 | } |
1168 | ASSERT3U(rows[i][missing[i]], !=, 0); | |
1169 | ||
1170 | /* | |
1171 | * Compute the inverse of the first element and multiply each | |
1172 | * element in the row by that value. | |
1173 | */ | |
1174 | log = 255 - vdev_raidz_log2[rows[i][missing[i]]]; | |
1175 | ||
1176 | for (j = 0; j < n; j++) { | |
1177 | rows[i][j] = vdev_raidz_exp2(rows[i][j], log); | |
1178 | invrows[i][j] = vdev_raidz_exp2(invrows[i][j], log); | |
1179 | } | |
1180 | ||
1181 | for (ii = 0; ii < nmissing; ii++) { | |
1182 | if (i == ii) | |
1183 | continue; | |
1184 | ||
1185 | ASSERT3U(rows[ii][missing[i]], !=, 0); | |
1186 | ||
1187 | log = vdev_raidz_log2[rows[ii][missing[i]]]; | |
1188 | ||
1189 | for (j = 0; j < n; j++) { | |
1190 | rows[ii][j] ^= | |
1191 | vdev_raidz_exp2(rows[i][j], log); | |
1192 | invrows[ii][j] ^= | |
1193 | vdev_raidz_exp2(invrows[i][j], log); | |
1194 | } | |
1195 | } | |
1196 | } | |
1197 | ||
1198 | /* | |
1199 | * Verify that the data that is left in the rows are properly part of | |
1200 | * an identity matrix. | |
1201 | */ | |
1202 | for (i = 0; i < nmissing; i++) { | |
1203 | for (j = 0; j < n; j++) { | |
1204 | if (j == missing[i]) { | |
1205 | ASSERT3U(rows[i][j], ==, 1); | |
1206 | } else { | |
c99c9001 | 1207 | ASSERT0(rows[i][j]); |
45d1cae3 BB |
1208 | } |
1209 | } | |
1210 | } | |
1211 | } | |
1212 | ||
1213 | static void | |
1214 | vdev_raidz_matrix_reconstruct(raidz_map_t *rm, int n, int nmissing, | |
1215 | int *missing, uint8_t **invrows, const uint8_t *used) | |
1216 | { | |
1217 | int i, j, x, cc, c; | |
1218 | uint8_t *src; | |
1219 | uint64_t ccount; | |
1220 | uint8_t *dst[VDEV_RAIDZ_MAXPARITY]; | |
1221 | uint64_t dcount[VDEV_RAIDZ_MAXPARITY]; | |
a117a6d6 GW |
1222 | uint8_t log = 0; |
1223 | uint8_t val; | |
45d1cae3 BB |
1224 | int ll; |
1225 | uint8_t *invlog[VDEV_RAIDZ_MAXPARITY]; | |
1226 | uint8_t *p, *pp; | |
1227 | size_t psize; | |
1228 | ||
1229 | psize = sizeof (invlog[0][0]) * n * nmissing; | |
b8d06fca | 1230 | p = kmem_alloc(psize, KM_PUSHPAGE); |
45d1cae3 BB |
1231 | |
1232 | for (pp = p, i = 0; i < nmissing; i++) { | |
1233 | invlog[i] = pp; | |
1234 | pp += n; | |
1235 | } | |
1236 | ||
1237 | for (i = 0; i < nmissing; i++) { | |
1238 | for (j = 0; j < n; j++) { | |
1239 | ASSERT3U(invrows[i][j], !=, 0); | |
1240 | invlog[i][j] = vdev_raidz_log2[invrows[i][j]]; | |
1241 | } | |
1242 | } | |
1243 | ||
1244 | for (i = 0; i < n; i++) { | |
1245 | c = used[i]; | |
1246 | ASSERT3U(c, <, rm->rm_cols); | |
1247 | ||
1248 | src = rm->rm_col[c].rc_data; | |
1249 | ccount = rm->rm_col[c].rc_size; | |
1250 | for (j = 0; j < nmissing; j++) { | |
1251 | cc = missing[j] + rm->rm_firstdatacol; | |
1252 | ASSERT3U(cc, >=, rm->rm_firstdatacol); | |
1253 | ASSERT3U(cc, <, rm->rm_cols); | |
1254 | ASSERT3U(cc, !=, c); | |
1255 | ||
1256 | dst[j] = rm->rm_col[cc].rc_data; | |
1257 | dcount[j] = rm->rm_col[cc].rc_size; | |
1258 | } | |
1259 | ||
1260 | ASSERT(ccount >= rm->rm_col[missing[0]].rc_size || i > 0); | |
1261 | ||
1262 | for (x = 0; x < ccount; x++, src++) { | |
1263 | if (*src != 0) | |
1264 | log = vdev_raidz_log2[*src]; | |
1265 | ||
1266 | for (cc = 0; cc < nmissing; cc++) { | |
1267 | if (x >= dcount[cc]) | |
1268 | continue; | |
1269 | ||
1270 | if (*src == 0) { | |
1271 | val = 0; | |
1272 | } else { | |
1273 | if ((ll = log + invlog[cc][i]) >= 255) | |
1274 | ll -= 255; | |
1275 | val = vdev_raidz_pow2[ll]; | |
1276 | } | |
1277 | ||
1278 | if (i == 0) | |
1279 | dst[cc][x] = val; | |
1280 | else | |
1281 | dst[cc][x] ^= val; | |
1282 | } | |
1283 | } | |
1284 | } | |
1285 | ||
1286 | kmem_free(p, psize); | |
1287 | } | |
1288 | ||
1289 | static int | |
1290 | vdev_raidz_reconstruct_general(raidz_map_t *rm, int *tgts, int ntgts) | |
1291 | { | |
1292 | int n, i, c, t, tt; | |
1293 | int nmissing_rows; | |
1294 | int missing_rows[VDEV_RAIDZ_MAXPARITY]; | |
1295 | int parity_map[VDEV_RAIDZ_MAXPARITY]; | |
1296 | ||
1297 | uint8_t *p, *pp; | |
1298 | size_t psize; | |
1299 | ||
1300 | uint8_t *rows[VDEV_RAIDZ_MAXPARITY]; | |
1301 | uint8_t *invrows[VDEV_RAIDZ_MAXPARITY]; | |
1302 | uint8_t *used; | |
1303 | ||
1304 | int code = 0; | |
1305 | ||
1306 | ||
1307 | n = rm->rm_cols - rm->rm_firstdatacol; | |
1308 | ||
1309 | /* | |
1310 | * Figure out which data columns are missing. | |
1311 | */ | |
1312 | nmissing_rows = 0; | |
1313 | for (t = 0; t < ntgts; t++) { | |
1314 | if (tgts[t] >= rm->rm_firstdatacol) { | |
1315 | missing_rows[nmissing_rows++] = | |
1316 | tgts[t] - rm->rm_firstdatacol; | |
1317 | } | |
1318 | } | |
1319 | ||
1320 | /* | |
1321 | * Figure out which parity columns to use to help generate the missing | |
1322 | * data columns. | |
1323 | */ | |
1324 | for (tt = 0, c = 0, i = 0; i < nmissing_rows; c++) { | |
1325 | ASSERT(tt < ntgts); | |
1326 | ASSERT(c < rm->rm_firstdatacol); | |
1327 | ||
1328 | /* | |
1329 | * Skip any targeted parity columns. | |
1330 | */ | |
1331 | if (c == tgts[tt]) { | |
1332 | tt++; | |
1333 | continue; | |
1334 | } | |
1335 | ||
1336 | code |= 1 << c; | |
1337 | ||
1338 | parity_map[i] = c; | |
1339 | i++; | |
1340 | } | |
1341 | ||
1342 | ASSERT(code != 0); | |
1343 | ASSERT3U(code, <, 1 << VDEV_RAIDZ_MAXPARITY); | |
1344 | ||
1345 | psize = (sizeof (rows[0][0]) + sizeof (invrows[0][0])) * | |
1346 | nmissing_rows * n + sizeof (used[0]) * n; | |
b8d06fca | 1347 | p = kmem_alloc(psize, KM_PUSHPAGE); |
45d1cae3 BB |
1348 | |
1349 | for (pp = p, i = 0; i < nmissing_rows; i++) { | |
1350 | rows[i] = pp; | |
1351 | pp += n; | |
1352 | invrows[i] = pp; | |
1353 | pp += n; | |
1354 | } | |
1355 | used = pp; | |
1356 | ||
1357 | for (i = 0; i < nmissing_rows; i++) { | |
1358 | used[i] = parity_map[i]; | |
1359 | } | |
1360 | ||
1361 | for (tt = 0, c = rm->rm_firstdatacol; c < rm->rm_cols; c++) { | |
1362 | if (tt < nmissing_rows && | |
1363 | c == missing_rows[tt] + rm->rm_firstdatacol) { | |
1364 | tt++; | |
1365 | continue; | |
1366 | } | |
1367 | ||
1368 | ASSERT3S(i, <, n); | |
1369 | used[i] = c; | |
1370 | i++; | |
1371 | } | |
1372 | ||
1373 | /* | |
1374 | * Initialize the interesting rows of the matrix. | |
1375 | */ | |
1376 | vdev_raidz_matrix_init(rm, n, nmissing_rows, parity_map, rows); | |
1377 | ||
1378 | /* | |
1379 | * Invert the matrix. | |
1380 | */ | |
1381 | vdev_raidz_matrix_invert(rm, n, nmissing_rows, missing_rows, rows, | |
1382 | invrows, used); | |
1383 | ||
1384 | /* | |
1385 | * Reconstruct the missing data using the generated matrix. | |
1386 | */ | |
1387 | vdev_raidz_matrix_reconstruct(rm, n, nmissing_rows, missing_rows, | |
1388 | invrows, used); | |
1389 | ||
1390 | kmem_free(p, psize); | |
1391 | ||
1392 | return (code); | |
34dc7c2f BB |
1393 | } |
1394 | ||
45d1cae3 BB |
1395 | static int |
1396 | vdev_raidz_reconstruct(raidz_map_t *rm, int *t, int nt) | |
1397 | { | |
1398 | int tgts[VDEV_RAIDZ_MAXPARITY], *dt; | |
1399 | int ntgts; | |
1400 | int i, c; | |
1401 | int code; | |
1402 | int nbadparity, nbaddata; | |
1403 | int parity_valid[VDEV_RAIDZ_MAXPARITY]; | |
1404 | ||
1405 | /* | |
1406 | * The tgts list must already be sorted. | |
1407 | */ | |
1408 | for (i = 1; i < nt; i++) { | |
1409 | ASSERT(t[i] > t[i - 1]); | |
1410 | } | |
1411 | ||
1412 | nbadparity = rm->rm_firstdatacol; | |
1413 | nbaddata = rm->rm_cols - nbadparity; | |
1414 | ntgts = 0; | |
1415 | for (i = 0, c = 0; c < rm->rm_cols; c++) { | |
1416 | if (c < rm->rm_firstdatacol) | |
1417 | parity_valid[c] = B_FALSE; | |
1418 | ||
1419 | if (i < nt && c == t[i]) { | |
1420 | tgts[ntgts++] = c; | |
1421 | i++; | |
1422 | } else if (rm->rm_col[c].rc_error != 0) { | |
1423 | tgts[ntgts++] = c; | |
1424 | } else if (c >= rm->rm_firstdatacol) { | |
1425 | nbaddata--; | |
1426 | } else { | |
1427 | parity_valid[c] = B_TRUE; | |
1428 | nbadparity--; | |
1429 | } | |
1430 | } | |
1431 | ||
1432 | ASSERT(ntgts >= nt); | |
1433 | ASSERT(nbaddata >= 0); | |
1434 | ASSERT(nbaddata + nbadparity == ntgts); | |
1435 | ||
1436 | dt = &tgts[nbadparity]; | |
1437 | ||
1438 | /* | |
1439 | * See if we can use any of our optimized reconstruction routines. | |
1440 | */ | |
1441 | if (!vdev_raidz_default_to_general) { | |
1442 | switch (nbaddata) { | |
1443 | case 1: | |
1444 | if (parity_valid[VDEV_RAIDZ_P]) | |
1445 | return (vdev_raidz_reconstruct_p(rm, dt, 1)); | |
1446 | ||
1447 | ASSERT(rm->rm_firstdatacol > 1); | |
1448 | ||
1449 | if (parity_valid[VDEV_RAIDZ_Q]) | |
1450 | return (vdev_raidz_reconstruct_q(rm, dt, 1)); | |
1451 | ||
1452 | ASSERT(rm->rm_firstdatacol > 2); | |
1453 | break; | |
1454 | ||
1455 | case 2: | |
1456 | ASSERT(rm->rm_firstdatacol > 1); | |
1457 | ||
1458 | if (parity_valid[VDEV_RAIDZ_P] && | |
1459 | parity_valid[VDEV_RAIDZ_Q]) | |
1460 | return (vdev_raidz_reconstruct_pq(rm, dt, 2)); | |
1461 | ||
1462 | ASSERT(rm->rm_firstdatacol > 2); | |
1463 | ||
1464 | break; | |
1465 | } | |
1466 | } | |
1467 | ||
1468 | code = vdev_raidz_reconstruct_general(rm, tgts, ntgts); | |
1469 | ASSERT(code < (1 << VDEV_RAIDZ_MAXPARITY)); | |
1470 | ASSERT(code > 0); | |
1471 | return (code); | |
1472 | } | |
34dc7c2f BB |
1473 | |
1474 | static int | |
1bd201e7 CS |
1475 | vdev_raidz_open(vdev_t *vd, uint64_t *asize, uint64_t *max_asize, |
1476 | uint64_t *ashift) | |
34dc7c2f BB |
1477 | { |
1478 | vdev_t *cvd; | |
1479 | uint64_t nparity = vd->vdev_nparity; | |
45d1cae3 | 1480 | int c; |
34dc7c2f BB |
1481 | int lasterror = 0; |
1482 | int numerrors = 0; | |
1483 | ||
1484 | ASSERT(nparity > 0); | |
1485 | ||
1486 | if (nparity > VDEV_RAIDZ_MAXPARITY || | |
1487 | vd->vdev_children < nparity + 1) { | |
1488 | vd->vdev_stat.vs_aux = VDEV_AUX_BAD_LABEL; | |
2e528b49 | 1489 | return (SET_ERROR(EINVAL)); |
34dc7c2f BB |
1490 | } |
1491 | ||
45d1cae3 BB |
1492 | vdev_open_children(vd); |
1493 | ||
34dc7c2f BB |
1494 | for (c = 0; c < vd->vdev_children; c++) { |
1495 | cvd = vd->vdev_child[c]; | |
1496 | ||
45d1cae3 BB |
1497 | if (cvd->vdev_open_error != 0) { |
1498 | lasterror = cvd->vdev_open_error; | |
34dc7c2f BB |
1499 | numerrors++; |
1500 | continue; | |
1501 | } | |
1502 | ||
1503 | *asize = MIN(*asize - 1, cvd->vdev_asize - 1) + 1; | |
1bd201e7 | 1504 | *max_asize = MIN(*max_asize - 1, cvd->vdev_max_asize - 1) + 1; |
34dc7c2f BB |
1505 | *ashift = MAX(*ashift, cvd->vdev_ashift); |
1506 | } | |
1507 | ||
1508 | *asize *= vd->vdev_children; | |
1bd201e7 | 1509 | *max_asize *= vd->vdev_children; |
34dc7c2f BB |
1510 | |
1511 | if (numerrors > nparity) { | |
1512 | vd->vdev_stat.vs_aux = VDEV_AUX_NO_REPLICAS; | |
1513 | return (lasterror); | |
1514 | } | |
1515 | ||
1516 | return (0); | |
1517 | } | |
1518 | ||
1519 | static void | |
1520 | vdev_raidz_close(vdev_t *vd) | |
1521 | { | |
1522 | int c; | |
1523 | ||
1524 | for (c = 0; c < vd->vdev_children; c++) | |
1525 | vdev_close(vd->vdev_child[c]); | |
1526 | } | |
1527 | ||
1528 | static uint64_t | |
1529 | vdev_raidz_asize(vdev_t *vd, uint64_t psize) | |
1530 | { | |
1531 | uint64_t asize; | |
1532 | uint64_t ashift = vd->vdev_top->vdev_ashift; | |
1533 | uint64_t cols = vd->vdev_children; | |
1534 | uint64_t nparity = vd->vdev_nparity; | |
1535 | ||
1536 | asize = ((psize - 1) >> ashift) + 1; | |
1537 | asize += nparity * ((asize + cols - nparity - 1) / (cols - nparity)); | |
1538 | asize = roundup(asize, nparity + 1) << ashift; | |
1539 | ||
1540 | return (asize); | |
1541 | } | |
1542 | ||
1543 | static void | |
1544 | vdev_raidz_child_done(zio_t *zio) | |
1545 | { | |
1546 | raidz_col_t *rc = zio->io_private; | |
1547 | ||
1548 | rc->rc_error = zio->io_error; | |
1549 | rc->rc_tried = 1; | |
1550 | rc->rc_skipped = 0; | |
1551 | } | |
1552 | ||
e49f1e20 WA |
1553 | /* |
1554 | * Start an IO operation on a RAIDZ VDev | |
1555 | * | |
1556 | * Outline: | |
1557 | * - For write operations: | |
1558 | * 1. Generate the parity data | |
1559 | * 2. Create child zio write operations to each column's vdev, for both | |
1560 | * data and parity. | |
1561 | * 3. If the column skips any sectors for padding, create optional dummy | |
1562 | * write zio children for those areas to improve aggregation continuity. | |
1563 | * - For read operations: | |
1564 | * 1. Create child zio read operations to each data column's vdev to read | |
1565 | * the range of data required for zio. | |
1566 | * 2. If this is a scrub or resilver operation, or if any of the data | |
1567 | * vdevs have had errors, then create zio read operations to the parity | |
1568 | * columns' VDevs as well. | |
1569 | */ | |
34dc7c2f BB |
1570 | static int |
1571 | vdev_raidz_io_start(zio_t *zio) | |
1572 | { | |
1573 | vdev_t *vd = zio->io_vd; | |
1574 | vdev_t *tvd = vd->vdev_top; | |
1575 | vdev_t *cvd; | |
34dc7c2f BB |
1576 | raidz_map_t *rm; |
1577 | raidz_col_t *rc; | |
45d1cae3 | 1578 | int c, i; |
34dc7c2f BB |
1579 | |
1580 | rm = vdev_raidz_map_alloc(zio, tvd->vdev_ashift, vd->vdev_children, | |
1581 | vd->vdev_nparity); | |
1582 | ||
1583 | ASSERT3U(rm->rm_asize, ==, vdev_psize_to_asize(vd, zio->io_size)); | |
1584 | ||
1585 | if (zio->io_type == ZIO_TYPE_WRITE) { | |
45d1cae3 | 1586 | vdev_raidz_generate_parity(rm); |
34dc7c2f BB |
1587 | |
1588 | for (c = 0; c < rm->rm_cols; c++) { | |
1589 | rc = &rm->rm_col[c]; | |
1590 | cvd = vd->vdev_child[rc->rc_devidx]; | |
1591 | zio_nowait(zio_vdev_child_io(zio, NULL, cvd, | |
1592 | rc->rc_offset, rc->rc_data, rc->rc_size, | |
b128c09f | 1593 | zio->io_type, zio->io_priority, 0, |
34dc7c2f BB |
1594 | vdev_raidz_child_done, rc)); |
1595 | } | |
1596 | ||
45d1cae3 BB |
1597 | /* |
1598 | * Generate optional I/Os for any skipped sectors to improve | |
1599 | * aggregation contiguity. | |
1600 | */ | |
428870ff | 1601 | for (c = rm->rm_skipstart, i = 0; i < rm->rm_nskip; c++, i++) { |
45d1cae3 BB |
1602 | ASSERT(c <= rm->rm_scols); |
1603 | if (c == rm->rm_scols) | |
1604 | c = 0; | |
1605 | rc = &rm->rm_col[c]; | |
1606 | cvd = vd->vdev_child[rc->rc_devidx]; | |
1607 | zio_nowait(zio_vdev_child_io(zio, NULL, cvd, | |
1608 | rc->rc_offset + rc->rc_size, NULL, | |
1609 | 1 << tvd->vdev_ashift, | |
1610 | zio->io_type, zio->io_priority, | |
1611 | ZIO_FLAG_NODATA | ZIO_FLAG_OPTIONAL, NULL, NULL)); | |
1612 | } | |
1613 | ||
b128c09f | 1614 | return (ZIO_PIPELINE_CONTINUE); |
34dc7c2f BB |
1615 | } |
1616 | ||
1617 | ASSERT(zio->io_type == ZIO_TYPE_READ); | |
1618 | ||
1619 | /* | |
1620 | * Iterate over the columns in reverse order so that we hit the parity | |
45d1cae3 | 1621 | * last -- any errors along the way will force us to read the parity. |
34dc7c2f BB |
1622 | */ |
1623 | for (c = rm->rm_cols - 1; c >= 0; c--) { | |
1624 | rc = &rm->rm_col[c]; | |
1625 | cvd = vd->vdev_child[rc->rc_devidx]; | |
1626 | if (!vdev_readable(cvd)) { | |
1627 | if (c >= rm->rm_firstdatacol) | |
1628 | rm->rm_missingdata++; | |
1629 | else | |
1630 | rm->rm_missingparity++; | |
2e528b49 | 1631 | rc->rc_error = SET_ERROR(ENXIO); |
34dc7c2f BB |
1632 | rc->rc_tried = 1; /* don't even try */ |
1633 | rc->rc_skipped = 1; | |
1634 | continue; | |
1635 | } | |
428870ff | 1636 | if (vdev_dtl_contains(cvd, DTL_MISSING, zio->io_txg, 1)) { |
34dc7c2f BB |
1637 | if (c >= rm->rm_firstdatacol) |
1638 | rm->rm_missingdata++; | |
1639 | else | |
1640 | rm->rm_missingparity++; | |
2e528b49 | 1641 | rc->rc_error = SET_ERROR(ESTALE); |
34dc7c2f BB |
1642 | rc->rc_skipped = 1; |
1643 | continue; | |
1644 | } | |
1645 | if (c >= rm->rm_firstdatacol || rm->rm_missingdata > 0 || | |
9babb374 | 1646 | (zio->io_flags & (ZIO_FLAG_SCRUB | ZIO_FLAG_RESILVER))) { |
34dc7c2f BB |
1647 | zio_nowait(zio_vdev_child_io(zio, NULL, cvd, |
1648 | rc->rc_offset, rc->rc_data, rc->rc_size, | |
b128c09f | 1649 | zio->io_type, zio->io_priority, 0, |
34dc7c2f BB |
1650 | vdev_raidz_child_done, rc)); |
1651 | } | |
1652 | } | |
1653 | ||
b128c09f | 1654 | return (ZIO_PIPELINE_CONTINUE); |
34dc7c2f BB |
1655 | } |
1656 | ||
428870ff | 1657 | |
34dc7c2f BB |
1658 | /* |
1659 | * Report a checksum error for a child of a RAID-Z device. | |
1660 | */ | |
1661 | static void | |
428870ff | 1662 | raidz_checksum_error(zio_t *zio, raidz_col_t *rc, void *bad_data) |
34dc7c2f BB |
1663 | { |
1664 | vdev_t *vd = zio->io_vd->vdev_child[rc->rc_devidx]; | |
34dc7c2f BB |
1665 | |
1666 | if (!(zio->io_flags & ZIO_FLAG_SPECULATIVE)) { | |
428870ff BB |
1667 | zio_bad_cksum_t zbc; |
1668 | raidz_map_t *rm = zio->io_vsd; | |
1669 | ||
34dc7c2f BB |
1670 | mutex_enter(&vd->vdev_stat_lock); |
1671 | vd->vdev_stat.vs_checksum_errors++; | |
1672 | mutex_exit(&vd->vdev_stat_lock); | |
428870ff BB |
1673 | |
1674 | zbc.zbc_has_cksum = 0; | |
1675 | zbc.zbc_injected = rm->rm_ecksuminjected; | |
1676 | ||
1677 | zfs_ereport_post_checksum(zio->io_spa, vd, zio, | |
1678 | rc->rc_offset, rc->rc_size, rc->rc_data, bad_data, | |
1679 | &zbc); | |
34dc7c2f | 1680 | } |
428870ff BB |
1681 | } |
1682 | ||
1683 | /* | |
1684 | * We keep track of whether or not there were any injected errors, so that | |
1685 | * any ereports we generate can note it. | |
1686 | */ | |
1687 | static int | |
1688 | raidz_checksum_verify(zio_t *zio) | |
1689 | { | |
1690 | zio_bad_cksum_t zbc; | |
1691 | raidz_map_t *rm = zio->io_vsd; | |
d4ed6673 | 1692 | int ret; |
428870ff | 1693 | |
d4ed6673 BB |
1694 | bzero(&zbc, sizeof (zio_bad_cksum_t)); |
1695 | ||
1696 | ret = zio_checksum_error(zio, &zbc); | |
428870ff BB |
1697 | if (ret != 0 && zbc.zbc_injected != 0) |
1698 | rm->rm_ecksuminjected = 1; | |
34dc7c2f | 1699 | |
428870ff | 1700 | return (ret); |
34dc7c2f BB |
1701 | } |
1702 | ||
1703 | /* | |
1704 | * Generate the parity from the data columns. If we tried and were able to | |
1705 | * read the parity without error, verify that the generated parity matches the | |
1706 | * data we read. If it doesn't, we fire off a checksum error. Return the | |
1707 | * number such failures. | |
1708 | */ | |
1709 | static int | |
1710 | raidz_parity_verify(zio_t *zio, raidz_map_t *rm) | |
1711 | { | |
1712 | void *orig[VDEV_RAIDZ_MAXPARITY]; | |
1713 | int c, ret = 0; | |
1714 | raidz_col_t *rc; | |
1715 | ||
1716 | for (c = 0; c < rm->rm_firstdatacol; c++) { | |
1717 | rc = &rm->rm_col[c]; | |
1718 | if (!rc->rc_tried || rc->rc_error != 0) | |
1719 | continue; | |
1720 | orig[c] = zio_buf_alloc(rc->rc_size); | |
1721 | bcopy(rc->rc_data, orig[c], rc->rc_size); | |
1722 | } | |
1723 | ||
45d1cae3 | 1724 | vdev_raidz_generate_parity(rm); |
34dc7c2f BB |
1725 | |
1726 | for (c = 0; c < rm->rm_firstdatacol; c++) { | |
1727 | rc = &rm->rm_col[c]; | |
1728 | if (!rc->rc_tried || rc->rc_error != 0) | |
1729 | continue; | |
1730 | if (bcmp(orig[c], rc->rc_data, rc->rc_size) != 0) { | |
428870ff | 1731 | raidz_checksum_error(zio, rc, orig[c]); |
2e528b49 | 1732 | rc->rc_error = SET_ERROR(ECKSUM); |
34dc7c2f BB |
1733 | ret++; |
1734 | } | |
1735 | zio_buf_free(orig[c], rc->rc_size); | |
1736 | } | |
1737 | ||
1738 | return (ret); | |
1739 | } | |
1740 | ||
45d1cae3 BB |
1741 | /* |
1742 | * Keep statistics on all the ways that we used parity to correct data. | |
1743 | */ | |
1744 | static uint64_t raidz_corrected[1 << VDEV_RAIDZ_MAXPARITY]; | |
34dc7c2f BB |
1745 | |
1746 | static int | |
b128c09f BB |
1747 | vdev_raidz_worst_error(raidz_map_t *rm) |
1748 | { | |
d6320ddb | 1749 | int c, error = 0; |
b128c09f | 1750 | |
d6320ddb | 1751 | for (c = 0; c < rm->rm_cols; c++) |
b128c09f BB |
1752 | error = zio_worst_error(error, rm->rm_col[c].rc_error); |
1753 | ||
1754 | return (error); | |
1755 | } | |
1756 | ||
45d1cae3 BB |
1757 | /* |
1758 | * Iterate over all combinations of bad data and attempt a reconstruction. | |
1759 | * Note that the algorithm below is non-optimal because it doesn't take into | |
1760 | * account how reconstruction is actually performed. For example, with | |
1761 | * triple-parity RAID-Z the reconstruction procedure is the same if column 4 | |
1762 | * is targeted as invalid as if columns 1 and 4 are targeted since in both | |
1763 | * cases we'd only use parity information in column 0. | |
1764 | */ | |
1765 | static int | |
1766 | vdev_raidz_combrec(zio_t *zio, int total_errors, int data_errors) | |
1767 | { | |
1768 | raidz_map_t *rm = zio->io_vsd; | |
1769 | raidz_col_t *rc; | |
1770 | void *orig[VDEV_RAIDZ_MAXPARITY]; | |
1771 | int tstore[VDEV_RAIDZ_MAXPARITY + 2]; | |
1772 | int *tgts = &tstore[1]; | |
5631c038 | 1773 | int curr, next, i, c, n; |
45d1cae3 BB |
1774 | int code, ret = 0; |
1775 | ||
1776 | ASSERT(total_errors < rm->rm_firstdatacol); | |
1777 | ||
1778 | /* | |
1779 | * This simplifies one edge condition. | |
1780 | */ | |
1781 | tgts[-1] = -1; | |
1782 | ||
1783 | for (n = 1; n <= rm->rm_firstdatacol - total_errors; n++) { | |
1784 | /* | |
1785 | * Initialize the targets array by finding the first n columns | |
1786 | * that contain no error. | |
1787 | * | |
1788 | * If there were no data errors, we need to ensure that we're | |
1789 | * always explicitly attempting to reconstruct at least one | |
1790 | * data column. To do this, we simply push the highest target | |
1791 | * up into the data columns. | |
1792 | */ | |
1793 | for (c = 0, i = 0; i < n; i++) { | |
1794 | if (i == n - 1 && data_errors == 0 && | |
1795 | c < rm->rm_firstdatacol) { | |
1796 | c = rm->rm_firstdatacol; | |
1797 | } | |
1798 | ||
1799 | while (rm->rm_col[c].rc_error != 0) { | |
1800 | c++; | |
1801 | ASSERT3S(c, <, rm->rm_cols); | |
1802 | } | |
1803 | ||
1804 | tgts[i] = c++; | |
1805 | } | |
1806 | ||
1807 | /* | |
1808 | * Setting tgts[n] simplifies the other edge condition. | |
1809 | */ | |
1810 | tgts[n] = rm->rm_cols; | |
1811 | ||
1812 | /* | |
1813 | * These buffers were allocated in previous iterations. | |
1814 | */ | |
1815 | for (i = 0; i < n - 1; i++) { | |
1816 | ASSERT(orig[i] != NULL); | |
1817 | } | |
1818 | ||
1819 | orig[n - 1] = zio_buf_alloc(rm->rm_col[0].rc_size); | |
1820 | ||
5631c038 BB |
1821 | curr = 0; |
1822 | next = tgts[curr]; | |
45d1cae3 | 1823 | |
5631c038 BB |
1824 | while (curr != n) { |
1825 | tgts[curr] = next; | |
1826 | curr = 0; | |
45d1cae3 BB |
1827 | |
1828 | /* | |
1829 | * Save off the original data that we're going to | |
1830 | * attempt to reconstruct. | |
1831 | */ | |
1832 | for (i = 0; i < n; i++) { | |
1833 | ASSERT(orig[i] != NULL); | |
1834 | c = tgts[i]; | |
1835 | ASSERT3S(c, >=, 0); | |
1836 | ASSERT3S(c, <, rm->rm_cols); | |
1837 | rc = &rm->rm_col[c]; | |
1838 | bcopy(rc->rc_data, orig[i], rc->rc_size); | |
1839 | } | |
1840 | ||
1841 | /* | |
1842 | * Attempt a reconstruction and exit the outer loop on | |
1843 | * success. | |
1844 | */ | |
1845 | code = vdev_raidz_reconstruct(rm, tgts, n); | |
428870ff | 1846 | if (raidz_checksum_verify(zio) == 0) { |
45d1cae3 BB |
1847 | atomic_inc_64(&raidz_corrected[code]); |
1848 | ||
1849 | for (i = 0; i < n; i++) { | |
1850 | c = tgts[i]; | |
1851 | rc = &rm->rm_col[c]; | |
1852 | ASSERT(rc->rc_error == 0); | |
1853 | if (rc->rc_tried) | |
428870ff BB |
1854 | raidz_checksum_error(zio, rc, |
1855 | orig[i]); | |
2e528b49 | 1856 | rc->rc_error = SET_ERROR(ECKSUM); |
45d1cae3 BB |
1857 | } |
1858 | ||
1859 | ret = code; | |
1860 | goto done; | |
1861 | } | |
1862 | ||
1863 | /* | |
1864 | * Restore the original data. | |
1865 | */ | |
1866 | for (i = 0; i < n; i++) { | |
1867 | c = tgts[i]; | |
1868 | rc = &rm->rm_col[c]; | |
1869 | bcopy(orig[i], rc->rc_data, rc->rc_size); | |
1870 | } | |
1871 | ||
1872 | do { | |
1873 | /* | |
5631c038 | 1874 | * Find the next valid column after the curr |
45d1cae3 BB |
1875 | * position.. |
1876 | */ | |
5631c038 | 1877 | for (next = tgts[curr] + 1; |
45d1cae3 BB |
1878 | next < rm->rm_cols && |
1879 | rm->rm_col[next].rc_error != 0; next++) | |
1880 | continue; | |
1881 | ||
5631c038 | 1882 | ASSERT(next <= tgts[curr + 1]); |
45d1cae3 BB |
1883 | |
1884 | /* | |
1885 | * If that spot is available, we're done here. | |
1886 | */ | |
5631c038 | 1887 | if (next != tgts[curr + 1]) |
45d1cae3 BB |
1888 | break; |
1889 | ||
1890 | /* | |
1891 | * Otherwise, find the next valid column after | |
1892 | * the previous position. | |
1893 | */ | |
5631c038 | 1894 | for (c = tgts[curr - 1] + 1; |
45d1cae3 BB |
1895 | rm->rm_col[c].rc_error != 0; c++) |
1896 | continue; | |
1897 | ||
5631c038 BB |
1898 | tgts[curr] = c; |
1899 | curr++; | |
45d1cae3 | 1900 | |
5631c038 | 1901 | } while (curr != n); |
45d1cae3 BB |
1902 | } |
1903 | } | |
1904 | n--; | |
1905 | done: | |
1906 | for (i = 0; i < n; i++) { | |
1907 | zio_buf_free(orig[i], rm->rm_col[0].rc_size); | |
1908 | } | |
1909 | ||
1910 | return (ret); | |
1911 | } | |
1912 | ||
e49f1e20 WA |
1913 | /* |
1914 | * Complete an IO operation on a RAIDZ VDev | |
1915 | * | |
1916 | * Outline: | |
1917 | * - For write operations: | |
1918 | * 1. Check for errors on the child IOs. | |
1919 | * 2. Return, setting an error code if too few child VDevs were written | |
1920 | * to reconstruct the data later. Note that partial writes are | |
1921 | * considered successful if they can be reconstructed at all. | |
1922 | * - For read operations: | |
1923 | * 1. Check for errors on the child IOs. | |
1924 | * 2. If data errors occurred: | |
1925 | * a. Try to reassemble the data from the parity available. | |
1926 | * b. If we haven't yet read the parity drives, read them now. | |
1927 | * c. If all parity drives have been read but the data still doesn't | |
1928 | * reassemble with a correct checksum, then try combinatorial | |
1929 | * reconstruction. | |
1930 | * d. If that doesn't work, return an error. | |
1931 | * 3. If there were unexpected errors or this is a resilver operation, | |
1932 | * rewrite the vdevs that had errors. | |
1933 | */ | |
b128c09f | 1934 | static void |
34dc7c2f BB |
1935 | vdev_raidz_io_done(zio_t *zio) |
1936 | { | |
1937 | vdev_t *vd = zio->io_vd; | |
1938 | vdev_t *cvd; | |
1939 | raidz_map_t *rm = zio->io_vsd; | |
d4ed6673 | 1940 | raidz_col_t *rc = NULL; |
34dc7c2f BB |
1941 | int unexpected_errors = 0; |
1942 | int parity_errors = 0; | |
1943 | int parity_untried = 0; | |
1944 | int data_errors = 0; | |
b128c09f | 1945 | int total_errors = 0; |
45d1cae3 BB |
1946 | int n, c; |
1947 | int tgts[VDEV_RAIDZ_MAXPARITY]; | |
1948 | int code; | |
34dc7c2f BB |
1949 | |
1950 | ASSERT(zio->io_bp != NULL); /* XXX need to add code to enforce this */ | |
1951 | ||
34dc7c2f BB |
1952 | ASSERT(rm->rm_missingparity <= rm->rm_firstdatacol); |
1953 | ASSERT(rm->rm_missingdata <= rm->rm_cols - rm->rm_firstdatacol); | |
1954 | ||
1955 | for (c = 0; c < rm->rm_cols; c++) { | |
1956 | rc = &rm->rm_col[c]; | |
1957 | ||
34dc7c2f | 1958 | if (rc->rc_error) { |
b128c09f | 1959 | ASSERT(rc->rc_error != ECKSUM); /* child has no bp */ |
34dc7c2f BB |
1960 | |
1961 | if (c < rm->rm_firstdatacol) | |
1962 | parity_errors++; | |
1963 | else | |
1964 | data_errors++; | |
1965 | ||
1966 | if (!rc->rc_skipped) | |
1967 | unexpected_errors++; | |
1968 | ||
b128c09f | 1969 | total_errors++; |
34dc7c2f BB |
1970 | } else if (c < rm->rm_firstdatacol && !rc->rc_tried) { |
1971 | parity_untried++; | |
1972 | } | |
1973 | } | |
1974 | ||
1975 | if (zio->io_type == ZIO_TYPE_WRITE) { | |
1976 | /* | |
b128c09f BB |
1977 | * XXX -- for now, treat partial writes as a success. |
1978 | * (If we couldn't write enough columns to reconstruct | |
1979 | * the data, the I/O failed. Otherwise, good enough.) | |
1980 | * | |
1981 | * Now that we support write reallocation, it would be better | |
1982 | * to treat partial failure as real failure unless there are | |
1983 | * no non-degraded top-level vdevs left, and not update DTLs | |
1984 | * if we intend to reallocate. | |
34dc7c2f BB |
1985 | */ |
1986 | /* XXPOLICY */ | |
b128c09f BB |
1987 | if (total_errors > rm->rm_firstdatacol) |
1988 | zio->io_error = vdev_raidz_worst_error(rm); | |
34dc7c2f | 1989 | |
b128c09f | 1990 | return; |
34dc7c2f BB |
1991 | } |
1992 | ||
1993 | ASSERT(zio->io_type == ZIO_TYPE_READ); | |
1994 | /* | |
1995 | * There are three potential phases for a read: | |
1996 | * 1. produce valid data from the columns read | |
1997 | * 2. read all disks and try again | |
1998 | * 3. perform combinatorial reconstruction | |
1999 | * | |
2000 | * Each phase is progressively both more expensive and less likely to | |
2001 | * occur. If we encounter more errors than we can repair or all phases | |
2002 | * fail, we have no choice but to return an error. | |
2003 | */ | |
2004 | ||
2005 | /* | |
2006 | * If the number of errors we saw was correctable -- less than or equal | |
2007 | * to the number of parity disks read -- attempt to produce data that | |
2008 | * has a valid checksum. Naturally, this case applies in the absence of | |
2009 | * any errors. | |
2010 | */ | |
b128c09f | 2011 | if (total_errors <= rm->rm_firstdatacol - parity_untried) { |
45d1cae3 | 2012 | if (data_errors == 0) { |
428870ff | 2013 | if (raidz_checksum_verify(zio) == 0) { |
34dc7c2f BB |
2014 | /* |
2015 | * If we read parity information (unnecessarily | |
2016 | * as it happens since no reconstruction was | |
2017 | * needed) regenerate and verify the parity. | |
2018 | * We also regenerate parity when resilvering | |
2019 | * so we can write it out to the failed device | |
2020 | * later. | |
2021 | */ | |
2022 | if (parity_errors + parity_untried < | |
2023 | rm->rm_firstdatacol || | |
2024 | (zio->io_flags & ZIO_FLAG_RESILVER)) { | |
2025 | n = raidz_parity_verify(zio, rm); | |
2026 | unexpected_errors += n; | |
2027 | ASSERT(parity_errors + n <= | |
2028 | rm->rm_firstdatacol); | |
2029 | } | |
2030 | goto done; | |
2031 | } | |
45d1cae3 | 2032 | } else { |
34dc7c2f BB |
2033 | /* |
2034 | * We either attempt to read all the parity columns or | |
2035 | * none of them. If we didn't try to read parity, we | |
2036 | * wouldn't be here in the correctable case. There must | |
2037 | * also have been fewer parity errors than parity | |
2038 | * columns or, again, we wouldn't be in this code path. | |
2039 | */ | |
2040 | ASSERT(parity_untried == 0); | |
2041 | ASSERT(parity_errors < rm->rm_firstdatacol); | |
2042 | ||
2043 | /* | |
45d1cae3 | 2044 | * Identify the data columns that reported an error. |
34dc7c2f | 2045 | */ |
45d1cae3 | 2046 | n = 0; |
34dc7c2f BB |
2047 | for (c = rm->rm_firstdatacol; c < rm->rm_cols; c++) { |
2048 | rc = &rm->rm_col[c]; | |
45d1cae3 BB |
2049 | if (rc->rc_error != 0) { |
2050 | ASSERT(n < VDEV_RAIDZ_MAXPARITY); | |
2051 | tgts[n++] = c; | |
2052 | } | |
34dc7c2f | 2053 | } |
34dc7c2f | 2054 | |
45d1cae3 BB |
2055 | ASSERT(rm->rm_firstdatacol >= n); |
2056 | ||
2057 | code = vdev_raidz_reconstruct(rm, tgts, n); | |
34dc7c2f | 2058 | |
428870ff | 2059 | if (raidz_checksum_verify(zio) == 0) { |
45d1cae3 | 2060 | atomic_inc_64(&raidz_corrected[code]); |
34dc7c2f BB |
2061 | |
2062 | /* | |
45d1cae3 BB |
2063 | * If we read more parity disks than were used |
2064 | * for reconstruction, confirm that the other | |
2065 | * parity disks produced correct data. This | |
2066 | * routine is suboptimal in that it regenerates | |
2067 | * the parity that we already used in addition | |
2068 | * to the parity that we're attempting to | |
2069 | * verify, but this should be a relatively | |
2070 | * uncommon case, and can be optimized if it | |
2071 | * becomes a problem. Note that we regenerate | |
2072 | * parity when resilvering so we can write it | |
2073 | * out to failed devices later. | |
34dc7c2f | 2074 | */ |
45d1cae3 | 2075 | if (parity_errors < rm->rm_firstdatacol - n || |
34dc7c2f BB |
2076 | (zio->io_flags & ZIO_FLAG_RESILVER)) { |
2077 | n = raidz_parity_verify(zio, rm); | |
2078 | unexpected_errors += n; | |
2079 | ASSERT(parity_errors + n <= | |
2080 | rm->rm_firstdatacol); | |
2081 | } | |
2082 | ||
2083 | goto done; | |
2084 | } | |
34dc7c2f BB |
2085 | } |
2086 | } | |
2087 | ||
2088 | /* | |
2089 | * This isn't a typical situation -- either we got a read error or | |
2090 | * a child silently returned bad data. Read every block so we can | |
2091 | * try again with as much data and parity as we can track down. If | |
2092 | * we've already been through once before, all children will be marked | |
2093 | * as tried so we'll proceed to combinatorial reconstruction. | |
2094 | */ | |
2095 | unexpected_errors = 1; | |
2096 | rm->rm_missingdata = 0; | |
2097 | rm->rm_missingparity = 0; | |
2098 | ||
2099 | for (c = 0; c < rm->rm_cols; c++) { | |
2100 | if (rm->rm_col[c].rc_tried) | |
2101 | continue; | |
2102 | ||
34dc7c2f BB |
2103 | zio_vdev_io_redone(zio); |
2104 | do { | |
2105 | rc = &rm->rm_col[c]; | |
2106 | if (rc->rc_tried) | |
2107 | continue; | |
2108 | zio_nowait(zio_vdev_child_io(zio, NULL, | |
2109 | vd->vdev_child[rc->rc_devidx], | |
2110 | rc->rc_offset, rc->rc_data, rc->rc_size, | |
b128c09f | 2111 | zio->io_type, zio->io_priority, 0, |
34dc7c2f BB |
2112 | vdev_raidz_child_done, rc)); |
2113 | } while (++c < rm->rm_cols); | |
34dc7c2f | 2114 | |
b128c09f | 2115 | return; |
34dc7c2f BB |
2116 | } |
2117 | ||
2118 | /* | |
2119 | * At this point we've attempted to reconstruct the data given the | |
2120 | * errors we detected, and we've attempted to read all columns. There | |
2121 | * must, therefore, be one or more additional problems -- silent errors | |
2122 | * resulting in invalid data rather than explicit I/O errors resulting | |
45d1cae3 BB |
2123 | * in absent data. We check if there is enough additional data to |
2124 | * possibly reconstruct the data and then perform combinatorial | |
2125 | * reconstruction over all possible combinations. If that fails, | |
2126 | * we're cooked. | |
34dc7c2f | 2127 | */ |
428870ff | 2128 | if (total_errors > rm->rm_firstdatacol) { |
b128c09f | 2129 | zio->io_error = vdev_raidz_worst_error(rm); |
34dc7c2f | 2130 | |
428870ff BB |
2131 | } else if (total_errors < rm->rm_firstdatacol && |
2132 | (code = vdev_raidz_combrec(zio, total_errors, data_errors)) != 0) { | |
34dc7c2f | 2133 | /* |
45d1cae3 BB |
2134 | * If we didn't use all the available parity for the |
2135 | * combinatorial reconstruction, verify that the remaining | |
2136 | * parity is correct. | |
34dc7c2f | 2137 | */ |
45d1cae3 BB |
2138 | if (code != (1 << rm->rm_firstdatacol) - 1) |
2139 | (void) raidz_parity_verify(zio, rm); | |
2140 | } else { | |
34dc7c2f | 2141 | /* |
428870ff BB |
2142 | * We're here because either: |
2143 | * | |
2144 | * total_errors == rm_first_datacol, or | |
2145 | * vdev_raidz_combrec() failed | |
2146 | * | |
2147 | * In either case, there is enough bad data to prevent | |
2148 | * reconstruction. | |
2149 | * | |
2150 | * Start checksum ereports for all children which haven't | |
2151 | * failed, and the IO wasn't speculative. | |
34dc7c2f | 2152 | */ |
2e528b49 | 2153 | zio->io_error = SET_ERROR(ECKSUM); |
34dc7c2f | 2154 | |
45d1cae3 BB |
2155 | if (!(zio->io_flags & ZIO_FLAG_SPECULATIVE)) { |
2156 | for (c = 0; c < rm->rm_cols; c++) { | |
2157 | rc = &rm->rm_col[c]; | |
428870ff BB |
2158 | if (rc->rc_error == 0) { |
2159 | zio_bad_cksum_t zbc; | |
2160 | zbc.zbc_has_cksum = 0; | |
2161 | zbc.zbc_injected = | |
2162 | rm->rm_ecksuminjected; | |
2163 | ||
2164 | zfs_ereport_start_checksum( | |
2165 | zio->io_spa, | |
2166 | vd->vdev_child[rc->rc_devidx], | |
2167 | zio, rc->rc_offset, rc->rc_size, | |
2168 | (void *)(uintptr_t)c, &zbc); | |
2169 | } | |
34dc7c2f | 2170 | } |
34dc7c2f BB |
2171 | } |
2172 | } | |
2173 | ||
2174 | done: | |
2175 | zio_checksum_verified(zio); | |
2176 | ||
fb5f0bc8 | 2177 | if (zio->io_error == 0 && spa_writeable(zio->io_spa) && |
34dc7c2f | 2178 | (unexpected_errors || (zio->io_flags & ZIO_FLAG_RESILVER))) { |
34dc7c2f BB |
2179 | /* |
2180 | * Use the good data we have in hand to repair damaged children. | |
34dc7c2f | 2181 | */ |
34dc7c2f BB |
2182 | for (c = 0; c < rm->rm_cols; c++) { |
2183 | rc = &rm->rm_col[c]; | |
2184 | cvd = vd->vdev_child[rc->rc_devidx]; | |
2185 | ||
2186 | if (rc->rc_error == 0) | |
2187 | continue; | |
2188 | ||
b128c09f | 2189 | zio_nowait(zio_vdev_child_io(zio, NULL, cvd, |
34dc7c2f | 2190 | rc->rc_offset, rc->rc_data, rc->rc_size, |
e8b96c60 | 2191 | ZIO_TYPE_WRITE, ZIO_PRIORITY_ASYNC_WRITE, |
fb5f0bc8 BB |
2192 | ZIO_FLAG_IO_REPAIR | (unexpected_errors ? |
2193 | ZIO_FLAG_SELF_HEAL : 0), NULL, NULL)); | |
34dc7c2f | 2194 | } |
34dc7c2f | 2195 | } |
34dc7c2f BB |
2196 | } |
2197 | ||
2198 | static void | |
2199 | vdev_raidz_state_change(vdev_t *vd, int faulted, int degraded) | |
2200 | { | |
2201 | if (faulted > vd->vdev_nparity) | |
2202 | vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, | |
2203 | VDEV_AUX_NO_REPLICAS); | |
2204 | else if (degraded + faulted != 0) | |
2205 | vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED, VDEV_AUX_NONE); | |
2206 | else | |
2207 | vdev_set_state(vd, B_FALSE, VDEV_STATE_HEALTHY, VDEV_AUX_NONE); | |
2208 | } | |
2209 | ||
2210 | vdev_ops_t vdev_raidz_ops = { | |
2211 | vdev_raidz_open, | |
2212 | vdev_raidz_close, | |
34dc7c2f BB |
2213 | vdev_raidz_asize, |
2214 | vdev_raidz_io_start, | |
2215 | vdev_raidz_io_done, | |
2216 | vdev_raidz_state_change, | |
428870ff BB |
2217 | NULL, |
2218 | NULL, | |
34dc7c2f BB |
2219 | VDEV_TYPE_RAIDZ, /* name of this vdev type */ |
2220 | B_FALSE /* not a leaf vdev */ | |
2221 | }; |