| 1 | // SPDX-License-Identifier: GPL-2.0 |
| 2 | /* |
| 3 | * Bad block management |
| 4 | * |
| 5 | * - Heavily based on MD badblocks code from Neil Brown |
| 6 | * |
| 7 | * Copyright (c) 2015, Intel Corporation. |
| 8 | */ |
| 9 | |
| 10 | #include <linux/badblocks.h> |
| 11 | #include <linux/seqlock.h> |
| 12 | #include <linux/device.h> |
| 13 | #include <linux/kernel.h> |
| 14 | #include <linux/module.h> |
| 15 | #include <linux/stddef.h> |
| 16 | #include <linux/types.h> |
| 17 | #include <linux/slab.h> |
| 18 | |
| 19 | /* |
| 20 | * The purpose of badblocks set/clear is to manage bad blocks ranges which are |
| 21 | * identified by LBA addresses. |
| 22 | * |
| 23 | * When the caller of badblocks_set() wants to set a range of bad blocks, the |
| 24 | * setting range can be acked or unacked. And the setting range may merge, |
| 25 | * overwrite, skip the overlapped already set range, depends on who they are |
| 26 | * overlapped or adjacent, and the acknowledgment type of the ranges. It can be |
| 27 | * more complicated when the setting range covers multiple already set bad block |
| 28 | * ranges, with restrictions of maximum length of each bad range and the bad |
| 29 | * table space limitation. |
| 30 | * |
| 31 | * It is difficult and unnecessary to take care of all the possible situations, |
| 32 | * for setting a large range of bad blocks, we can handle it by dividing the |
| 33 | * large range into smaller ones when encounter overlap, max range length or |
| 34 | * bad table full conditions. Every time only a smaller piece of the bad range |
| 35 | * is handled with a limited number of conditions how it is interacted with |
| 36 | * possible overlapped or adjacent already set bad block ranges. Then the hard |
| 37 | * complicated problem can be much simpler to handle in proper way. |
| 38 | * |
| 39 | * When setting a range of bad blocks to the bad table, the simplified situations |
| 40 | * to be considered are, (The already set bad blocks ranges are naming with |
| 41 | * prefix E, and the setting bad blocks range is naming with prefix S) |
| 42 | * |
| 43 | * 1) A setting range is not overlapped or adjacent to any other already set bad |
| 44 | * block range. |
| 45 | * +--------+ |
| 46 | * | S | |
| 47 | * +--------+ |
| 48 | * +-------------+ +-------------+ |
| 49 | * | E1 | | E2 | |
| 50 | * +-------------+ +-------------+ |
| 51 | * For this situation if the bad blocks table is not full, just allocate a |
| 52 | * free slot from the bad blocks table to mark the setting range S. The |
| 53 | * result is, |
| 54 | * +-------------+ +--------+ +-------------+ |
| 55 | * | E1 | | S | | E2 | |
| 56 | * +-------------+ +--------+ +-------------+ |
| 57 | * 2) A setting range starts exactly at a start LBA of an already set bad blocks |
| 58 | * range. |
| 59 | * 2.1) The setting range size < already set range size |
| 60 | * +--------+ |
| 61 | * | S | |
| 62 | * +--------+ |
| 63 | * +-------------+ |
| 64 | * | E | |
| 65 | * +-------------+ |
| 66 | * 2.1.1) If S and E are both acked or unacked range, the setting range S can |
| 67 | * be merged into existing bad range E. The result is, |
| 68 | * +-------------+ |
| 69 | * | S | |
| 70 | * +-------------+ |
| 71 | * 2.1.2) If S is unacked setting and E is acked, the setting will be denied, and |
| 72 | * the result is, |
| 73 | * +-------------+ |
| 74 | * | E | |
| 75 | * +-------------+ |
| 76 | * 2.1.3) If S is acked setting and E is unacked, range S can overwrite on E. |
| 77 | * An extra slot from the bad blocks table will be allocated for S, and head |
| 78 | * of E will move to end of the inserted range S. The result is, |
| 79 | * +--------+----+ |
| 80 | * | S | E | |
| 81 | * +--------+----+ |
| 82 | * 2.2) The setting range size == already set range size |
| 83 | * 2.2.1) If S and E are both acked or unacked range, the setting range S can |
| 84 | * be merged into existing bad range E. The result is, |
| 85 | * +-------------+ |
| 86 | * | S | |
| 87 | * +-------------+ |
| 88 | * 2.2.2) If S is unacked setting and E is acked, the setting will be denied, and |
| 89 | * the result is, |
| 90 | * +-------------+ |
| 91 | * | E | |
| 92 | * +-------------+ |
| 93 | * 2.2.3) If S is acked setting and E is unacked, range S can overwrite all of |
| 94 | bad blocks range E. The result is, |
| 95 | * +-------------+ |
| 96 | * | S | |
| 97 | * +-------------+ |
| 98 | * 2.3) The setting range size > already set range size |
| 99 | * +-------------------+ |
| 100 | * | S | |
| 101 | * +-------------------+ |
| 102 | * +-------------+ |
| 103 | * | E | |
| 104 | * +-------------+ |
| 105 | * For such situation, the setting range S can be treated as two parts, the |
| 106 | * first part (S1) is as same size as the already set range E, the second |
| 107 | * part (S2) is the rest of setting range. |
| 108 | * +-------------+-----+ +-------------+ +-----+ |
| 109 | * | S1 | S2 | | S1 | | S2 | |
| 110 | * +-------------+-----+ ===> +-------------+ +-----+ |
| 111 | * +-------------+ +-------------+ |
| 112 | * | E | | E | |
| 113 | * +-------------+ +-------------+ |
| 114 | * Now we only focus on how to handle the setting range S1 and already set |
| 115 | * range E, which are already explained in 2.2), for the rest S2 it will be |
| 116 | * handled later in next loop. |
| 117 | * 3) A setting range starts before the start LBA of an already set bad blocks |
| 118 | * range. |
| 119 | * +-------------+ |
| 120 | * | S | |
| 121 | * +-------------+ |
| 122 | * +-------------+ |
| 123 | * | E | |
| 124 | * +-------------+ |
| 125 | * For this situation, the setting range S can be divided into two parts, the |
| 126 | * first (S1) ends at the start LBA of already set range E, the second part |
| 127 | * (S2) starts exactly at a start LBA of the already set range E. |
| 128 | * +----+---------+ +----+ +---------+ |
| 129 | * | S1 | S2 | | S1 | | S2 | |
| 130 | * +----+---------+ ===> +----+ +---------+ |
| 131 | * +-------------+ +-------------+ |
| 132 | * | E | | E | |
| 133 | * +-------------+ +-------------+ |
| 134 | * Now only the first part S1 should be handled in this loop, which is in |
| 135 | * similar condition as 1). The rest part S2 has exact same start LBA address |
| 136 | * of the already set range E, they will be handled in next loop in one of |
| 137 | * situations in 2). |
| 138 | * 4) A setting range starts after the start LBA of an already set bad blocks |
| 139 | * range. |
| 140 | * 4.1) If the setting range S exactly matches the tail part of already set bad |
| 141 | * blocks range E, like the following chart shows, |
| 142 | * +---------+ |
| 143 | * | S | |
| 144 | * +---------+ |
| 145 | * +-------------+ |
| 146 | * | E | |
| 147 | * +-------------+ |
| 148 | * 4.1.1) If range S and E have same acknowledge value (both acked or unacked), |
| 149 | * they will be merged into one, the result is, |
| 150 | * +-------------+ |
| 151 | * | S | |
| 152 | * +-------------+ |
| 153 | * 4.1.2) If range E is acked and the setting range S is unacked, the setting |
| 154 | * request of S will be rejected, the result is, |
| 155 | * +-------------+ |
| 156 | * | E | |
| 157 | * +-------------+ |
| 158 | * 4.1.3) If range E is unacked, and the setting range S is acked, then S may |
| 159 | * overwrite the overlapped range of E, the result is, |
| 160 | * +---+---------+ |
| 161 | * | E | S | |
| 162 | * +---+---------+ |
| 163 | * 4.2) If the setting range S stays in middle of an already set range E, like |
| 164 | * the following chart shows, |
| 165 | * +----+ |
| 166 | * | S | |
| 167 | * +----+ |
| 168 | * +--------------+ |
| 169 | * | E | |
| 170 | * +--------------+ |
| 171 | * 4.2.1) If range S and E have same acknowledge value (both acked or unacked), |
| 172 | * they will be merged into one, the result is, |
| 173 | * +--------------+ |
| 174 | * | S | |
| 175 | * +--------------+ |
| 176 | * 4.2.2) If range E is acked and the setting range S is unacked, the setting |
| 177 | * request of S will be rejected, the result is also, |
| 178 | * +--------------+ |
| 179 | * | E | |
| 180 | * +--------------+ |
| 181 | * 4.2.3) If range E is unacked, and the setting range S is acked, then S will |
| 182 | * inserted into middle of E and split previous range E into two parts (E1 |
| 183 | * and E2), the result is, |
| 184 | * +----+----+----+ |
| 185 | * | E1 | S | E2 | |
| 186 | * +----+----+----+ |
| 187 | * 4.3) If the setting bad blocks range S is overlapped with an already set bad |
| 188 | * blocks range E. The range S starts after the start LBA of range E, and |
| 189 | * ends after the end LBA of range E, as the following chart shows, |
| 190 | * +-------------------+ |
| 191 | * | S | |
| 192 | * +-------------------+ |
| 193 | * +-------------+ |
| 194 | * | E | |
| 195 | * +-------------+ |
| 196 | * For this situation the range S can be divided into two parts, the first |
| 197 | * part (S1) ends at end range E, and the second part (S2) has rest range of |
| 198 | * origin S. |
| 199 | * +---------+---------+ +---------+ +---------+ |
| 200 | * | S1 | S2 | | S1 | | S2 | |
| 201 | * +---------+---------+ ===> +---------+ +---------+ |
| 202 | * +-------------+ +-------------+ |
| 203 | * | E | | E | |
| 204 | * +-------------+ +-------------+ |
| 205 | * Now in this loop the setting range S1 and already set range E can be |
| 206 | * handled as the situations 4.1), the rest range S2 will be handled in next |
| 207 | * loop and ignored in this loop. |
| 208 | * 5) A setting bad blocks range S is adjacent to one or more already set bad |
| 209 | * blocks range(s), and they are all acked or unacked range. |
| 210 | * 5.1) Front merge: If the already set bad blocks range E is before setting |
| 211 | * range S and they are adjacent, |
| 212 | * +------+ |
| 213 | * | S | |
| 214 | * +------+ |
| 215 | * +-------+ |
| 216 | * | E | |
| 217 | * +-------+ |
| 218 | * 5.1.1) When total size of range S and E <= BB_MAX_LEN, and their acknowledge |
| 219 | * values are same, the setting range S can front merges into range E. The |
| 220 | * result is, |
| 221 | * +--------------+ |
| 222 | * | S | |
| 223 | * +--------------+ |
| 224 | * 5.1.2) Otherwise these two ranges cannot merge, just insert the setting |
| 225 | * range S right after already set range E into the bad blocks table. The |
| 226 | * result is, |
| 227 | * +--------+------+ |
| 228 | * | E | S | |
| 229 | * +--------+------+ |
| 230 | * 6) Special cases which above conditions cannot handle |
| 231 | * 6.1) Multiple already set ranges may merge into less ones in a full bad table |
| 232 | * +-------------------------------------------------------+ |
| 233 | * | S | |
| 234 | * +-------------------------------------------------------+ |
| 235 | * |<----- BB_MAX_LEN ----->| |
| 236 | * +-----+ +-----+ +-----+ |
| 237 | * | E1 | | E2 | | E3 | |
| 238 | * +-----+ +-----+ +-----+ |
| 239 | * In the above example, when the bad blocks table is full, inserting the |
| 240 | * first part of setting range S will fail because no more available slot |
| 241 | * can be allocated from bad blocks table. In this situation a proper |
| 242 | * setting method should be go though all the setting bad blocks range and |
| 243 | * look for chance to merge already set ranges into less ones. When there |
| 244 | * is available slot from bad blocks table, re-try again to handle more |
| 245 | * setting bad blocks ranges as many as possible. |
| 246 | * +------------------------+ |
| 247 | * | S3 | |
| 248 | * +------------------------+ |
| 249 | * |<----- BB_MAX_LEN ----->| |
| 250 | * +-----+-----+-----+---+-----+--+ |
| 251 | * | S1 | S2 | |
| 252 | * +-----+-----+-----+---+-----+--+ |
| 253 | * The above chart shows although the first part (S3) cannot be inserted due |
| 254 | * to no-space in bad blocks table, but the following E1, E2 and E3 ranges |
| 255 | * can be merged with rest part of S into less range S1 and S2. Now there is |
| 256 | * 1 free slot in bad blocks table. |
| 257 | * +------------------------+-----+-----+-----+---+-----+--+ |
| 258 | * | S3 | S1 | S2 | |
| 259 | * +------------------------+-----+-----+-----+---+-----+--+ |
| 260 | * Since the bad blocks table is not full anymore, re-try again for the |
| 261 | * origin setting range S. Now the setting range S3 can be inserted into the |
| 262 | * bad blocks table with previous freed slot from multiple ranges merge. |
| 263 | * 6.2) Front merge after overwrite |
| 264 | * In the following example, in bad blocks table, E1 is an acked bad blocks |
| 265 | * range and E2 is an unacked bad blocks range, therefore they are not able |
| 266 | * to merge into a larger range. The setting bad blocks range S is acked, |
| 267 | * therefore part of E2 can be overwritten by S. |
| 268 | * +--------+ |
| 269 | * | S | acknowledged |
| 270 | * +--------+ S: 1 |
| 271 | * +-------+-------------+ E1: 1 |
| 272 | * | E1 | E2 | E2: 0 |
| 273 | * +-------+-------------+ |
| 274 | * With previous simplified routines, after overwriting part of E2 with S, |
| 275 | * the bad blocks table should be (E3 is remaining part of E2 which is not |
| 276 | * overwritten by S), |
| 277 | * acknowledged |
| 278 | * +-------+--------+----+ S: 1 |
| 279 | * | E1 | S | E3 | E1: 1 |
| 280 | * +-------+--------+----+ E3: 0 |
| 281 | * The above result is correct but not perfect. Range E1 and S in the bad |
| 282 | * blocks table are all acked, merging them into a larger one range may |
| 283 | * occupy less bad blocks table space and make badblocks_check() faster. |
| 284 | * Therefore in such situation, after overwriting range S, the previous range |
| 285 | * E1 should be checked for possible front combination. Then the ideal |
| 286 | * result can be, |
| 287 | * +----------------+----+ acknowledged |
| 288 | * | E1 | E3 | E1: 1 |
| 289 | * +----------------+----+ E3: 0 |
| 290 | * 6.3) Behind merge: If the already set bad blocks range E is behind the setting |
| 291 | * range S and they are adjacent. Normally we don't need to care about this |
| 292 | * because front merge handles this while going though range S from head to |
| 293 | * tail, except for the tail part of range S. When the setting range S are |
| 294 | * fully handled, all the above simplified routine doesn't check whether the |
| 295 | * tail LBA of range S is adjacent to the next already set range and not |
| 296 | * merge them even it is possible. |
| 297 | * +------+ |
| 298 | * | S | |
| 299 | * +------+ |
| 300 | * +-------+ |
| 301 | * | E | |
| 302 | * +-------+ |
| 303 | * For the above special situation, when the setting range S are all handled |
| 304 | * and the loop ends, an extra check is necessary for whether next already |
| 305 | * set range E is right after S and mergeable. |
| 306 | * 6.3.1) When total size of range E and S <= BB_MAX_LEN, and their acknowledge |
| 307 | * values are same, the setting range S can behind merges into range E. The |
| 308 | * result is, |
| 309 | * +--------------+ |
| 310 | * | S | |
| 311 | * +--------------+ |
| 312 | * 6.3.2) Otherwise these two ranges cannot merge, just insert the setting range |
| 313 | * S in front of the already set range E in the bad blocks table. The result |
| 314 | * is, |
| 315 | * +------+-------+ |
| 316 | * | S | E | |
| 317 | * +------+-------+ |
| 318 | * |
| 319 | * All the above 5 simplified situations and 3 special cases may cover 99%+ of |
| 320 | * the bad block range setting conditions. Maybe there is some rare corner case |
| 321 | * is not considered and optimized, it won't hurt if badblocks_set() fails due |
| 322 | * to no space, or some ranges are not merged to save bad blocks table space. |
| 323 | * |
| 324 | * Inside badblocks_set() each loop starts by jumping to re_insert label, every |
| 325 | * time for the new loop prev_badblocks() is called to find an already set range |
| 326 | * which starts before or at current setting range. Since the setting bad blocks |
| 327 | * range is handled from head to tail, most of the cases it is unnecessary to do |
| 328 | * the binary search inside prev_badblocks(), it is possible to provide a hint |
| 329 | * to prev_badblocks() for a fast path, then the expensive binary search can be |
| 330 | * avoided. In my test with the hint to prev_badblocks(), except for the first |
| 331 | * loop, all rested calls to prev_badblocks() can go into the fast path and |
| 332 | * return correct bad blocks table index immediately. |
| 333 | * |
| 334 | * |
| 335 | * Clearing a bad blocks range from the bad block table has similar idea as |
| 336 | * setting does, but much more simpler. The only thing needs to be noticed is |
| 337 | * when the clearing range hits middle of a bad block range, the existing bad |
| 338 | * block range will split into two, and one more item should be added into the |
| 339 | * bad block table. The simplified situations to be considered are, (The already |
| 340 | * set bad blocks ranges in bad block table are naming with prefix E, and the |
| 341 | * clearing bad blocks range is naming with prefix C) |
| 342 | * |
| 343 | * 1) A clearing range is not overlapped to any already set ranges in bad block |
| 344 | * table. |
| 345 | * +-----+ | +-----+ | +-----+ |
| 346 | * | C | | | C | | | C | |
| 347 | * +-----+ or +-----+ or +-----+ |
| 348 | * +---+ | +----+ +----+ | +---+ |
| 349 | * | E | | | E1 | | E2 | | | E | |
| 350 | * +---+ | +----+ +----+ | +---+ |
| 351 | * For the above situations, no bad block to be cleared and no failure |
| 352 | * happens, simply returns 0. |
| 353 | * 2) The clearing range hits middle of an already setting bad blocks range in |
| 354 | * the bad block table. |
| 355 | * +---+ |
| 356 | * | C | |
| 357 | * +---+ |
| 358 | * +-----------------+ |
| 359 | * | E | |
| 360 | * +-----------------+ |
| 361 | * In this situation if the bad block table is not full, the range E will be |
| 362 | * split into two ranges E1 and E2. The result is, |
| 363 | * +------+ +------+ |
| 364 | * | E1 | | E2 | |
| 365 | * +------+ +------+ |
| 366 | * 3) The clearing range starts exactly at same LBA as an already set bad block range |
| 367 | * from the bad block table. |
| 368 | * 3.1) Partially covered at head part |
| 369 | * +------------+ |
| 370 | * | C | |
| 371 | * +------------+ |
| 372 | * +-----------------+ |
| 373 | * | E | |
| 374 | * +-----------------+ |
| 375 | * For this situation, the overlapped already set range will update the |
| 376 | * start LBA to end of C and shrink the range to BB_LEN(E) - BB_LEN(C). No |
| 377 | * item deleted from bad block table. The result is, |
| 378 | * +----+ |
| 379 | * | E1 | |
| 380 | * +----+ |
| 381 | * 3.2) Exact fully covered |
| 382 | * +-----------------+ |
| 383 | * | C | |
| 384 | * +-----------------+ |
| 385 | * +-----------------+ |
| 386 | * | E | |
| 387 | * +-----------------+ |
| 388 | * For this situation the whole bad blocks range E will be cleared and its |
| 389 | * corresponded item is deleted from the bad block table. |
| 390 | * 4) The clearing range exactly ends at same LBA as an already set bad block |
| 391 | * range. |
| 392 | * +-------+ |
| 393 | * | C | |
| 394 | * +-------+ |
| 395 | * +-----------------+ |
| 396 | * | E | |
| 397 | * +-----------------+ |
| 398 | * For the above situation, the already set range E is updated to shrink its |
| 399 | * end to the start of C, and reduce its length to BB_LEN(E) - BB_LEN(C). |
| 400 | * The result is, |
| 401 | * +---------+ |
| 402 | * | E | |
| 403 | * +---------+ |
| 404 | * 5) The clearing range is partially overlapped with an already set bad block |
| 405 | * range from the bad block table. |
| 406 | * 5.1) The already set bad block range is front overlapped with the clearing |
| 407 | * range. |
| 408 | * +----------+ |
| 409 | * | C | |
| 410 | * +----------+ |
| 411 | * +------------+ |
| 412 | * | E | |
| 413 | * +------------+ |
| 414 | * For such situation, the clearing range C can be treated as two parts. The |
| 415 | * first part ends at the start LBA of range E, and the second part starts at |
| 416 | * same LBA of range E. |
| 417 | * +----+-----+ +----+ +-----+ |
| 418 | * | C1 | C2 | | C1 | | C2 | |
| 419 | * +----+-----+ ===> +----+ +-----+ |
| 420 | * +------------+ +------------+ |
| 421 | * | E | | E | |
| 422 | * +------------+ +------------+ |
| 423 | * Now the first part C1 can be handled as condition 1), and the second part C2 can be |
| 424 | * handled as condition 3.1) in next loop. |
| 425 | * 5.2) The already set bad block range is behind overlaopped with the clearing |
| 426 | * range. |
| 427 | * +----------+ |
| 428 | * | C | |
| 429 | * +----------+ |
| 430 | * +------------+ |
| 431 | * | E | |
| 432 | * +------------+ |
| 433 | * For such situation, the clearing range C can be treated as two parts. The |
| 434 | * first part C1 ends at same end LBA of range E, and the second part starts |
| 435 | * at end LBA of range E. |
| 436 | * +----+-----+ +----+ +-----+ |
| 437 | * | C1 | C2 | | C1 | | C2 | |
| 438 | * +----+-----+ ===> +----+ +-----+ |
| 439 | * +------------+ +------------+ |
| 440 | * | E | | E | |
| 441 | * +------------+ +------------+ |
| 442 | * Now the first part clearing range C1 can be handled as condition 4), and |
| 443 | * the second part clearing range C2 can be handled as condition 1) in next |
| 444 | * loop. |
| 445 | * |
| 446 | * All bad blocks range clearing can be simplified into the above 5 situations |
| 447 | * by only handling the head part of the clearing range in each run of the |
| 448 | * while-loop. The idea is similar to bad blocks range setting but much |
| 449 | * simpler. |
| 450 | */ |
| 451 | |
| 452 | /* |
| 453 | * Find the range starts at-or-before 's' from bad table. The search |
| 454 | * starts from index 'hint' and stops at index 'hint_end' from the bad |
| 455 | * table. |
| 456 | */ |
| 457 | static int prev_by_hint(struct badblocks *bb, sector_t s, int hint) |
| 458 | { |
| 459 | int hint_end = hint + 2; |
| 460 | u64 *p = bb->page; |
| 461 | int ret = -1; |
| 462 | |
| 463 | while ((hint < hint_end) && ((hint + 1) <= bb->count) && |
| 464 | (BB_OFFSET(p[hint]) <= s)) { |
| 465 | if ((hint + 1) == bb->count || BB_OFFSET(p[hint + 1]) > s) { |
| 466 | ret = hint; |
| 467 | break; |
| 468 | } |
| 469 | hint++; |
| 470 | } |
| 471 | |
| 472 | return ret; |
| 473 | } |
| 474 | |
| 475 | /* |
| 476 | * Find the range starts at-or-before bad->start. If 'hint' is provided |
| 477 | * (hint >= 0) then search in the bad table from hint firstly. It is |
| 478 | * very probably the wanted bad range can be found from the hint index, |
| 479 | * then the unnecessary while-loop iteration can be avoided. |
| 480 | */ |
| 481 | static int prev_badblocks(struct badblocks *bb, struct badblocks_context *bad, |
| 482 | int hint) |
| 483 | { |
| 484 | sector_t s = bad->start; |
| 485 | int ret = -1; |
| 486 | int lo, hi; |
| 487 | u64 *p; |
| 488 | |
| 489 | if (!bb->count) |
| 490 | goto out; |
| 491 | |
| 492 | if (hint >= 0) { |
| 493 | ret = prev_by_hint(bb, s, hint); |
| 494 | if (ret >= 0) |
| 495 | goto out; |
| 496 | } |
| 497 | |
| 498 | lo = 0; |
| 499 | hi = bb->count; |
| 500 | p = bb->page; |
| 501 | |
| 502 | /* The following bisect search might be unnecessary */ |
| 503 | if (BB_OFFSET(p[lo]) > s) |
| 504 | return -1; |
| 505 | if (BB_OFFSET(p[hi - 1]) <= s) |
| 506 | return hi - 1; |
| 507 | |
| 508 | /* Do bisect search in bad table */ |
| 509 | while (hi - lo > 1) { |
| 510 | int mid = (lo + hi)/2; |
| 511 | sector_t a = BB_OFFSET(p[mid]); |
| 512 | |
| 513 | if (a == s) { |
| 514 | ret = mid; |
| 515 | goto out; |
| 516 | } |
| 517 | |
| 518 | if (a < s) |
| 519 | lo = mid; |
| 520 | else |
| 521 | hi = mid; |
| 522 | } |
| 523 | |
| 524 | if (BB_OFFSET(p[lo]) <= s) |
| 525 | ret = lo; |
| 526 | out: |
| 527 | return ret; |
| 528 | } |
| 529 | |
| 530 | /* |
| 531 | * Return 'true' if the range indicated by 'bad' can be backward merged |
| 532 | * with the bad range (from the bad table) index by 'behind'. |
| 533 | */ |
| 534 | static bool can_merge_behind(struct badblocks *bb, |
| 535 | struct badblocks_context *bad, int behind) |
| 536 | { |
| 537 | sector_t sectors = bad->len; |
| 538 | sector_t s = bad->start; |
| 539 | u64 *p = bb->page; |
| 540 | |
| 541 | if ((s < BB_OFFSET(p[behind])) && |
| 542 | ((s + sectors) >= BB_OFFSET(p[behind])) && |
| 543 | ((BB_END(p[behind]) - s) <= BB_MAX_LEN) && |
| 544 | BB_ACK(p[behind]) == bad->ack) |
| 545 | return true; |
| 546 | return false; |
| 547 | } |
| 548 | |
| 549 | /* |
| 550 | * Do backward merge for range indicated by 'bad' and the bad range |
| 551 | * (from the bad table) indexed by 'behind'. The return value is merged |
| 552 | * sectors from bad->len. |
| 553 | */ |
| 554 | static int behind_merge(struct badblocks *bb, struct badblocks_context *bad, |
| 555 | int behind) |
| 556 | { |
| 557 | sector_t sectors = bad->len; |
| 558 | sector_t s = bad->start; |
| 559 | u64 *p = bb->page; |
| 560 | int merged = 0; |
| 561 | |
| 562 | WARN_ON(s >= BB_OFFSET(p[behind])); |
| 563 | WARN_ON((s + sectors) < BB_OFFSET(p[behind])); |
| 564 | |
| 565 | if (s < BB_OFFSET(p[behind])) { |
| 566 | merged = BB_OFFSET(p[behind]) - s; |
| 567 | p[behind] = BB_MAKE(s, BB_LEN(p[behind]) + merged, bad->ack); |
| 568 | |
| 569 | WARN_ON((BB_LEN(p[behind]) + merged) >= BB_MAX_LEN); |
| 570 | } |
| 571 | |
| 572 | return merged; |
| 573 | } |
| 574 | |
| 575 | /* |
| 576 | * Return 'true' if the range indicated by 'bad' can be forward |
| 577 | * merged with the bad range (from the bad table) indexed by 'prev'. |
| 578 | */ |
| 579 | static bool can_merge_front(struct badblocks *bb, int prev, |
| 580 | struct badblocks_context *bad) |
| 581 | { |
| 582 | sector_t s = bad->start; |
| 583 | u64 *p = bb->page; |
| 584 | |
| 585 | if (BB_ACK(p[prev]) == bad->ack && |
| 586 | (s < BB_END(p[prev]) || |
| 587 | (s == BB_END(p[prev]) && (BB_LEN(p[prev]) < BB_MAX_LEN)))) |
| 588 | return true; |
| 589 | return false; |
| 590 | } |
| 591 | |
| 592 | /* |
| 593 | * Do forward merge for range indicated by 'bad' and the bad range |
| 594 | * (from bad table) indexed by 'prev'. The return value is sectors |
| 595 | * merged from bad->len. |
| 596 | */ |
| 597 | static int front_merge(struct badblocks *bb, int prev, struct badblocks_context *bad) |
| 598 | { |
| 599 | sector_t sectors = bad->len; |
| 600 | sector_t s = bad->start; |
| 601 | u64 *p = bb->page; |
| 602 | int merged = 0; |
| 603 | |
| 604 | WARN_ON(s > BB_END(p[prev])); |
| 605 | |
| 606 | if (s < BB_END(p[prev])) { |
| 607 | merged = min_t(sector_t, sectors, BB_END(p[prev]) - s); |
| 608 | } else { |
| 609 | merged = min_t(sector_t, sectors, BB_MAX_LEN - BB_LEN(p[prev])); |
| 610 | if ((prev + 1) < bb->count && |
| 611 | merged > (BB_OFFSET(p[prev + 1]) - BB_END(p[prev]))) { |
| 612 | merged = BB_OFFSET(p[prev + 1]) - BB_END(p[prev]); |
| 613 | } |
| 614 | |
| 615 | p[prev] = BB_MAKE(BB_OFFSET(p[prev]), |
| 616 | BB_LEN(p[prev]) + merged, bad->ack); |
| 617 | } |
| 618 | |
| 619 | return merged; |
| 620 | } |
| 621 | |
| 622 | /* |
| 623 | * 'Combine' is a special case which can_merge_front() is not able to |
| 624 | * handle: If a bad range (indexed by 'prev' from bad table) exactly |
| 625 | * starts as bad->start, and the bad range ahead of 'prev' (indexed by |
| 626 | * 'prev - 1' from bad table) exactly ends at where 'prev' starts, and |
| 627 | * the sum of their lengths does not exceed BB_MAX_LEN limitation, then |
| 628 | * these two bad range (from bad table) can be combined. |
| 629 | * |
| 630 | * Return 'true' if bad ranges indexed by 'prev' and 'prev - 1' from bad |
| 631 | * table can be combined. |
| 632 | */ |
| 633 | static bool can_combine_front(struct badblocks *bb, int prev, |
| 634 | struct badblocks_context *bad) |
| 635 | { |
| 636 | u64 *p = bb->page; |
| 637 | |
| 638 | if ((prev > 0) && |
| 639 | (BB_OFFSET(p[prev]) == bad->start) && |
| 640 | (BB_END(p[prev - 1]) == BB_OFFSET(p[prev])) && |
| 641 | (BB_LEN(p[prev - 1]) + BB_LEN(p[prev]) <= BB_MAX_LEN) && |
| 642 | (BB_ACK(p[prev - 1]) == BB_ACK(p[prev]))) |
| 643 | return true; |
| 644 | return false; |
| 645 | } |
| 646 | |
| 647 | /* |
| 648 | * Combine the bad ranges indexed by 'prev' and 'prev - 1' (from bad |
| 649 | * table) into one larger bad range, and the new range is indexed by |
| 650 | * 'prev - 1'. |
| 651 | * The caller of front_combine() will decrease bb->count, therefore |
| 652 | * it is unnecessary to clear p[perv] after front merge. |
| 653 | */ |
| 654 | static void front_combine(struct badblocks *bb, int prev) |
| 655 | { |
| 656 | u64 *p = bb->page; |
| 657 | |
| 658 | p[prev - 1] = BB_MAKE(BB_OFFSET(p[prev - 1]), |
| 659 | BB_LEN(p[prev - 1]) + BB_LEN(p[prev]), |
| 660 | BB_ACK(p[prev])); |
| 661 | if ((prev + 1) < bb->count) |
| 662 | memmove(p + prev, p + prev + 1, (bb->count - prev - 1) * 8); |
| 663 | } |
| 664 | |
| 665 | /* |
| 666 | * Return 'true' if the range indicated by 'bad' is exactly forward |
| 667 | * overlapped with the bad range (from bad table) indexed by 'front'. |
| 668 | * Exactly forward overlap means the bad range (from bad table) indexed |
| 669 | * by 'prev' does not cover the whole range indicated by 'bad'. |
| 670 | */ |
| 671 | static bool overlap_front(struct badblocks *bb, int front, |
| 672 | struct badblocks_context *bad) |
| 673 | { |
| 674 | u64 *p = bb->page; |
| 675 | |
| 676 | if (bad->start >= BB_OFFSET(p[front]) && |
| 677 | bad->start < BB_END(p[front])) |
| 678 | return true; |
| 679 | return false; |
| 680 | } |
| 681 | |
| 682 | /* |
| 683 | * Return 'true' if the range indicated by 'bad' is exactly backward |
| 684 | * overlapped with the bad range (from bad table) indexed by 'behind'. |
| 685 | */ |
| 686 | static bool overlap_behind(struct badblocks *bb, struct badblocks_context *bad, |
| 687 | int behind) |
| 688 | { |
| 689 | u64 *p = bb->page; |
| 690 | |
| 691 | if (bad->start < BB_OFFSET(p[behind]) && |
| 692 | (bad->start + bad->len) > BB_OFFSET(p[behind])) |
| 693 | return true; |
| 694 | return false; |
| 695 | } |
| 696 | |
| 697 | /* |
| 698 | * Return 'true' if the range indicated by 'bad' can overwrite the bad |
| 699 | * range (from bad table) indexed by 'prev'. |
| 700 | * |
| 701 | * The range indicated by 'bad' can overwrite the bad range indexed by |
| 702 | * 'prev' when, |
| 703 | * 1) The whole range indicated by 'bad' can cover partial or whole bad |
| 704 | * range (from bad table) indexed by 'prev'. |
| 705 | * 2) The ack value of 'bad' is larger or equal to the ack value of bad |
| 706 | * range 'prev'. |
| 707 | * |
| 708 | * If the overwriting doesn't cover the whole bad range (from bad table) |
| 709 | * indexed by 'prev', new range might be split from existing bad range, |
| 710 | * 1) The overwrite covers head or tail part of existing bad range, 1 |
| 711 | * extra bad range will be split and added into the bad table. |
| 712 | * 2) The overwrite covers middle of existing bad range, 2 extra bad |
| 713 | * ranges will be split (ahead and after the overwritten range) and |
| 714 | * added into the bad table. |
| 715 | * The number of extra split ranges of the overwriting is stored in |
| 716 | * 'extra' and returned for the caller. |
| 717 | */ |
| 718 | static bool can_front_overwrite(struct badblocks *bb, int prev, |
| 719 | struct badblocks_context *bad, int *extra) |
| 720 | { |
| 721 | u64 *p = bb->page; |
| 722 | int len; |
| 723 | |
| 724 | WARN_ON(!overlap_front(bb, prev, bad)); |
| 725 | |
| 726 | if (BB_ACK(p[prev]) >= bad->ack) |
| 727 | return false; |
| 728 | |
| 729 | if (BB_END(p[prev]) <= (bad->start + bad->len)) { |
| 730 | len = BB_END(p[prev]) - bad->start; |
| 731 | if (BB_OFFSET(p[prev]) == bad->start) |
| 732 | *extra = 0; |
| 733 | else |
| 734 | *extra = 1; |
| 735 | |
| 736 | bad->len = len; |
| 737 | } else { |
| 738 | if (BB_OFFSET(p[prev]) == bad->start) |
| 739 | *extra = 1; |
| 740 | else |
| 741 | /* |
| 742 | * prev range will be split into two, beside the overwritten |
| 743 | * one, an extra slot needed from bad table. |
| 744 | */ |
| 745 | *extra = 2; |
| 746 | } |
| 747 | |
| 748 | if ((bb->count + (*extra)) >= MAX_BADBLOCKS) |
| 749 | return false; |
| 750 | |
| 751 | return true; |
| 752 | } |
| 753 | |
| 754 | /* |
| 755 | * Do the overwrite from the range indicated by 'bad' to the bad range |
| 756 | * (from bad table) indexed by 'prev'. |
| 757 | * The previously called can_front_overwrite() will provide how many |
| 758 | * extra bad range(s) might be split and added into the bad table. All |
| 759 | * the splitting cases in the bad table will be handled here. |
| 760 | */ |
| 761 | static int front_overwrite(struct badblocks *bb, int prev, |
| 762 | struct badblocks_context *bad, int extra) |
| 763 | { |
| 764 | u64 *p = bb->page; |
| 765 | sector_t orig_end = BB_END(p[prev]); |
| 766 | int orig_ack = BB_ACK(p[prev]); |
| 767 | |
| 768 | switch (extra) { |
| 769 | case 0: |
| 770 | p[prev] = BB_MAKE(BB_OFFSET(p[prev]), BB_LEN(p[prev]), |
| 771 | bad->ack); |
| 772 | break; |
| 773 | case 1: |
| 774 | if (BB_OFFSET(p[prev]) == bad->start) { |
| 775 | p[prev] = BB_MAKE(BB_OFFSET(p[prev]), |
| 776 | bad->len, bad->ack); |
| 777 | memmove(p + prev + 2, p + prev + 1, |
| 778 | (bb->count - prev - 1) * 8); |
| 779 | p[prev + 1] = BB_MAKE(bad->start + bad->len, |
| 780 | orig_end - BB_END(p[prev]), |
| 781 | orig_ack); |
| 782 | } else { |
| 783 | p[prev] = BB_MAKE(BB_OFFSET(p[prev]), |
| 784 | bad->start - BB_OFFSET(p[prev]), |
| 785 | orig_ack); |
| 786 | /* |
| 787 | * prev +2 -> prev + 1 + 1, which is for, |
| 788 | * 1) prev + 1: the slot index of the previous one |
| 789 | * 2) + 1: one more slot for extra being 1. |
| 790 | */ |
| 791 | memmove(p + prev + 2, p + prev + 1, |
| 792 | (bb->count - prev - 1) * 8); |
| 793 | p[prev + 1] = BB_MAKE(bad->start, bad->len, bad->ack); |
| 794 | } |
| 795 | break; |
| 796 | case 2: |
| 797 | p[prev] = BB_MAKE(BB_OFFSET(p[prev]), |
| 798 | bad->start - BB_OFFSET(p[prev]), |
| 799 | orig_ack); |
| 800 | /* |
| 801 | * prev + 3 -> prev + 1 + 2, which is for, |
| 802 | * 1) prev + 1: the slot index of the previous one |
| 803 | * 2) + 2: two more slots for extra being 2. |
| 804 | */ |
| 805 | memmove(p + prev + 3, p + prev + 1, |
| 806 | (bb->count - prev - 1) * 8); |
| 807 | p[prev + 1] = BB_MAKE(bad->start, bad->len, bad->ack); |
| 808 | p[prev + 2] = BB_MAKE(BB_END(p[prev + 1]), |
| 809 | orig_end - BB_END(p[prev + 1]), |
| 810 | orig_ack); |
| 811 | break; |
| 812 | default: |
| 813 | break; |
| 814 | } |
| 815 | |
| 816 | return bad->len; |
| 817 | } |
| 818 | |
| 819 | /* |
| 820 | * Explicitly insert a range indicated by 'bad' to the bad table, where |
| 821 | * the location is indexed by 'at'. |
| 822 | */ |
| 823 | static int insert_at(struct badblocks *bb, int at, struct badblocks_context *bad) |
| 824 | { |
| 825 | u64 *p = bb->page; |
| 826 | int len; |
| 827 | |
| 828 | WARN_ON(badblocks_full(bb)); |
| 829 | |
| 830 | len = min_t(sector_t, bad->len, BB_MAX_LEN); |
| 831 | if (at < bb->count) |
| 832 | memmove(p + at + 1, p + at, (bb->count - at) * 8); |
| 833 | p[at] = BB_MAKE(bad->start, len, bad->ack); |
| 834 | |
| 835 | return len; |
| 836 | } |
| 837 | |
| 838 | static void badblocks_update_acked(struct badblocks *bb) |
| 839 | { |
| 840 | bool unacked = false; |
| 841 | u64 *p = bb->page; |
| 842 | int i; |
| 843 | |
| 844 | if (!bb->unacked_exist) |
| 845 | return; |
| 846 | |
| 847 | for (i = 0; i < bb->count ; i++) { |
| 848 | if (!BB_ACK(p[i])) { |
| 849 | unacked = true; |
| 850 | break; |
| 851 | } |
| 852 | } |
| 853 | |
| 854 | if (!unacked) |
| 855 | bb->unacked_exist = 0; |
| 856 | } |
| 857 | |
| 858 | /* Do exact work to set bad block range into the bad block table */ |
| 859 | static int _badblocks_set(struct badblocks *bb, sector_t s, int sectors, |
| 860 | int acknowledged) |
| 861 | { |
| 862 | int retried = 0, space_desired = 0; |
| 863 | int orig_len, len = 0, added = 0; |
| 864 | struct badblocks_context bad; |
| 865 | int prev = -1, hint = -1; |
| 866 | sector_t orig_start; |
| 867 | unsigned long flags; |
| 868 | int rv = 0; |
| 869 | u64 *p; |
| 870 | |
| 871 | if (bb->shift < 0) |
| 872 | /* badblocks are disabled */ |
| 873 | return 1; |
| 874 | |
| 875 | if (sectors == 0) |
| 876 | /* Invalid sectors number */ |
| 877 | return 1; |
| 878 | |
| 879 | if (bb->shift) { |
| 880 | /* round the start down, and the end up */ |
| 881 | sector_t next = s + sectors; |
| 882 | |
| 883 | rounddown(s, bb->shift); |
| 884 | roundup(next, bb->shift); |
| 885 | sectors = next - s; |
| 886 | } |
| 887 | |
| 888 | write_seqlock_irqsave(&bb->lock, flags); |
| 889 | |
| 890 | orig_start = s; |
| 891 | orig_len = sectors; |
| 892 | bad.ack = acknowledged; |
| 893 | p = bb->page; |
| 894 | |
| 895 | re_insert: |
| 896 | bad.start = s; |
| 897 | bad.len = sectors; |
| 898 | len = 0; |
| 899 | |
| 900 | if (badblocks_empty(bb)) { |
| 901 | len = insert_at(bb, 0, &bad); |
| 902 | bb->count++; |
| 903 | added++; |
| 904 | goto update_sectors; |
| 905 | } |
| 906 | |
| 907 | prev = prev_badblocks(bb, &bad, hint); |
| 908 | |
| 909 | /* start before all badblocks */ |
| 910 | if (prev < 0) { |
| 911 | if (!badblocks_full(bb)) { |
| 912 | /* insert on the first */ |
| 913 | if (bad.len > (BB_OFFSET(p[0]) - bad.start)) |
| 914 | bad.len = BB_OFFSET(p[0]) - bad.start; |
| 915 | len = insert_at(bb, 0, &bad); |
| 916 | bb->count++; |
| 917 | added++; |
| 918 | hint = 0; |
| 919 | goto update_sectors; |
| 920 | } |
| 921 | |
| 922 | /* No sapce, try to merge */ |
| 923 | if (overlap_behind(bb, &bad, 0)) { |
| 924 | if (can_merge_behind(bb, &bad, 0)) { |
| 925 | len = behind_merge(bb, &bad, 0); |
| 926 | added++; |
| 927 | } else { |
| 928 | len = BB_OFFSET(p[0]) - s; |
| 929 | space_desired = 1; |
| 930 | } |
| 931 | hint = 0; |
| 932 | goto update_sectors; |
| 933 | } |
| 934 | |
| 935 | /* no table space and give up */ |
| 936 | goto out; |
| 937 | } |
| 938 | |
| 939 | /* in case p[prev-1] can be merged with p[prev] */ |
| 940 | if (can_combine_front(bb, prev, &bad)) { |
| 941 | front_combine(bb, prev); |
| 942 | bb->count--; |
| 943 | added++; |
| 944 | hint = prev; |
| 945 | goto update_sectors; |
| 946 | } |
| 947 | |
| 948 | if (overlap_front(bb, prev, &bad)) { |
| 949 | if (can_merge_front(bb, prev, &bad)) { |
| 950 | len = front_merge(bb, prev, &bad); |
| 951 | added++; |
| 952 | } else { |
| 953 | int extra = 0; |
| 954 | |
| 955 | if (!can_front_overwrite(bb, prev, &bad, &extra)) { |
| 956 | len = min_t(sector_t, |
| 957 | BB_END(p[prev]) - s, sectors); |
| 958 | hint = prev; |
| 959 | goto update_sectors; |
| 960 | } |
| 961 | |
| 962 | len = front_overwrite(bb, prev, &bad, extra); |
| 963 | added++; |
| 964 | bb->count += extra; |
| 965 | |
| 966 | if (can_combine_front(bb, prev, &bad)) { |
| 967 | front_combine(bb, prev); |
| 968 | bb->count--; |
| 969 | } |
| 970 | } |
| 971 | hint = prev; |
| 972 | goto update_sectors; |
| 973 | } |
| 974 | |
| 975 | if (can_merge_front(bb, prev, &bad)) { |
| 976 | len = front_merge(bb, prev, &bad); |
| 977 | added++; |
| 978 | hint = prev; |
| 979 | goto update_sectors; |
| 980 | } |
| 981 | |
| 982 | /* if no space in table, still try to merge in the covered range */ |
| 983 | if (badblocks_full(bb)) { |
| 984 | /* skip the cannot-merge range */ |
| 985 | if (((prev + 1) < bb->count) && |
| 986 | overlap_behind(bb, &bad, prev + 1) && |
| 987 | ((s + sectors) >= BB_END(p[prev + 1]))) { |
| 988 | len = BB_END(p[prev + 1]) - s; |
| 989 | hint = prev + 1; |
| 990 | goto update_sectors; |
| 991 | } |
| 992 | |
| 993 | /* no retry any more */ |
| 994 | len = sectors; |
| 995 | space_desired = 1; |
| 996 | hint = -1; |
| 997 | goto update_sectors; |
| 998 | } |
| 999 | |
| 1000 | /* cannot merge and there is space in bad table */ |
| 1001 | if ((prev + 1) < bb->count && |
| 1002 | overlap_behind(bb, &bad, prev + 1)) |
| 1003 | bad.len = min_t(sector_t, |
| 1004 | bad.len, BB_OFFSET(p[prev + 1]) - bad.start); |
| 1005 | |
| 1006 | len = insert_at(bb, prev + 1, &bad); |
| 1007 | bb->count++; |
| 1008 | added++; |
| 1009 | hint = prev + 1; |
| 1010 | |
| 1011 | update_sectors: |
| 1012 | s += len; |
| 1013 | sectors -= len; |
| 1014 | |
| 1015 | if (sectors > 0) |
| 1016 | goto re_insert; |
| 1017 | |
| 1018 | WARN_ON(sectors < 0); |
| 1019 | |
| 1020 | /* |
| 1021 | * Check whether the following already set range can be |
| 1022 | * merged. (prev < 0) condition is not handled here, |
| 1023 | * because it's already complicated enough. |
| 1024 | */ |
| 1025 | if (prev >= 0 && |
| 1026 | (prev + 1) < bb->count && |
| 1027 | BB_END(p[prev]) == BB_OFFSET(p[prev + 1]) && |
| 1028 | (BB_LEN(p[prev]) + BB_LEN(p[prev + 1])) <= BB_MAX_LEN && |
| 1029 | BB_ACK(p[prev]) == BB_ACK(p[prev + 1])) { |
| 1030 | p[prev] = BB_MAKE(BB_OFFSET(p[prev]), |
| 1031 | BB_LEN(p[prev]) + BB_LEN(p[prev + 1]), |
| 1032 | BB_ACK(p[prev])); |
| 1033 | |
| 1034 | if ((prev + 2) < bb->count) |
| 1035 | memmove(p + prev + 1, p + prev + 2, |
| 1036 | (bb->count - (prev + 2)) * 8); |
| 1037 | bb->count--; |
| 1038 | } |
| 1039 | |
| 1040 | if (space_desired && !badblocks_full(bb)) { |
| 1041 | s = orig_start; |
| 1042 | sectors = orig_len; |
| 1043 | space_desired = 0; |
| 1044 | if (retried++ < 3) |
| 1045 | goto re_insert; |
| 1046 | } |
| 1047 | |
| 1048 | out: |
| 1049 | if (added) { |
| 1050 | set_changed(bb); |
| 1051 | |
| 1052 | if (!acknowledged) |
| 1053 | bb->unacked_exist = 1; |
| 1054 | else |
| 1055 | badblocks_update_acked(bb); |
| 1056 | } |
| 1057 | |
| 1058 | write_sequnlock_irqrestore(&bb->lock, flags); |
| 1059 | |
| 1060 | if (!added) |
| 1061 | rv = 1; |
| 1062 | |
| 1063 | return rv; |
| 1064 | } |
| 1065 | |
| 1066 | /* |
| 1067 | * Clear the bad block range from bad block table which is front overlapped |
| 1068 | * with the clearing range. The return value is how many sectors from an |
| 1069 | * already set bad block range are cleared. If the whole bad block range is |
| 1070 | * covered by the clearing range and fully cleared, 'delete' is set as 1 for |
| 1071 | * the caller to reduce bb->count. |
| 1072 | */ |
| 1073 | static int front_clear(struct badblocks *bb, int prev, |
| 1074 | struct badblocks_context *bad, int *deleted) |
| 1075 | { |
| 1076 | sector_t sectors = bad->len; |
| 1077 | sector_t s = bad->start; |
| 1078 | u64 *p = bb->page; |
| 1079 | int cleared = 0; |
| 1080 | |
| 1081 | *deleted = 0; |
| 1082 | if (s == BB_OFFSET(p[prev])) { |
| 1083 | if (BB_LEN(p[prev]) > sectors) { |
| 1084 | p[prev] = BB_MAKE(BB_OFFSET(p[prev]) + sectors, |
| 1085 | BB_LEN(p[prev]) - sectors, |
| 1086 | BB_ACK(p[prev])); |
| 1087 | cleared = sectors; |
| 1088 | } else { |
| 1089 | /* BB_LEN(p[prev]) <= sectors */ |
| 1090 | cleared = BB_LEN(p[prev]); |
| 1091 | if ((prev + 1) < bb->count) |
| 1092 | memmove(p + prev, p + prev + 1, |
| 1093 | (bb->count - prev - 1) * 8); |
| 1094 | *deleted = 1; |
| 1095 | } |
| 1096 | } else if (s > BB_OFFSET(p[prev])) { |
| 1097 | if (BB_END(p[prev]) <= (s + sectors)) { |
| 1098 | cleared = BB_END(p[prev]) - s; |
| 1099 | p[prev] = BB_MAKE(BB_OFFSET(p[prev]), |
| 1100 | s - BB_OFFSET(p[prev]), |
| 1101 | BB_ACK(p[prev])); |
| 1102 | } else { |
| 1103 | /* Splitting is handled in front_splitting_clear() */ |
| 1104 | BUG(); |
| 1105 | } |
| 1106 | } |
| 1107 | |
| 1108 | return cleared; |
| 1109 | } |
| 1110 | |
| 1111 | /* |
| 1112 | * Handle the condition that the clearing range hits middle of an already set |
| 1113 | * bad block range from bad block table. In this condition the existing bad |
| 1114 | * block range is split into two after the middle part is cleared. |
| 1115 | */ |
| 1116 | static int front_splitting_clear(struct badblocks *bb, int prev, |
| 1117 | struct badblocks_context *bad) |
| 1118 | { |
| 1119 | u64 *p = bb->page; |
| 1120 | u64 end = BB_END(p[prev]); |
| 1121 | int ack = BB_ACK(p[prev]); |
| 1122 | sector_t sectors = bad->len; |
| 1123 | sector_t s = bad->start; |
| 1124 | |
| 1125 | p[prev] = BB_MAKE(BB_OFFSET(p[prev]), |
| 1126 | s - BB_OFFSET(p[prev]), |
| 1127 | ack); |
| 1128 | memmove(p + prev + 2, p + prev + 1, (bb->count - prev - 1) * 8); |
| 1129 | p[prev + 1] = BB_MAKE(s + sectors, end - s - sectors, ack); |
| 1130 | return sectors; |
| 1131 | } |
| 1132 | |
| 1133 | /* Do the exact work to clear bad block range from the bad block table */ |
| 1134 | static int _badblocks_clear(struct badblocks *bb, sector_t s, int sectors) |
| 1135 | { |
| 1136 | struct badblocks_context bad; |
| 1137 | int prev = -1, hint = -1; |
| 1138 | int len = 0, cleared = 0; |
| 1139 | int rv = 0; |
| 1140 | u64 *p; |
| 1141 | |
| 1142 | if (bb->shift < 0) |
| 1143 | /* badblocks are disabled */ |
| 1144 | return 1; |
| 1145 | |
| 1146 | if (sectors == 0) |
| 1147 | /* Invalid sectors number */ |
| 1148 | return 1; |
| 1149 | |
| 1150 | if (bb->shift) { |
| 1151 | sector_t target; |
| 1152 | |
| 1153 | /* When clearing we round the start up and the end down. |
| 1154 | * This should not matter as the shift should align with |
| 1155 | * the block size and no rounding should ever be needed. |
| 1156 | * However it is better the think a block is bad when it |
| 1157 | * isn't than to think a block is not bad when it is. |
| 1158 | */ |
| 1159 | target = s + sectors; |
| 1160 | roundup(s, bb->shift); |
| 1161 | rounddown(target, bb->shift); |
| 1162 | sectors = target - s; |
| 1163 | } |
| 1164 | |
| 1165 | write_seqlock_irq(&bb->lock); |
| 1166 | |
| 1167 | bad.ack = true; |
| 1168 | p = bb->page; |
| 1169 | |
| 1170 | re_clear: |
| 1171 | bad.start = s; |
| 1172 | bad.len = sectors; |
| 1173 | |
| 1174 | if (badblocks_empty(bb)) { |
| 1175 | len = sectors; |
| 1176 | cleared++; |
| 1177 | goto update_sectors; |
| 1178 | } |
| 1179 | |
| 1180 | |
| 1181 | prev = prev_badblocks(bb, &bad, hint); |
| 1182 | |
| 1183 | /* Start before all badblocks */ |
| 1184 | if (prev < 0) { |
| 1185 | if (overlap_behind(bb, &bad, 0)) { |
| 1186 | len = BB_OFFSET(p[0]) - s; |
| 1187 | hint = 0; |
| 1188 | } else { |
| 1189 | len = sectors; |
| 1190 | } |
| 1191 | /* |
| 1192 | * Both situations are to clear non-bad range, |
| 1193 | * should be treated as successful |
| 1194 | */ |
| 1195 | cleared++; |
| 1196 | goto update_sectors; |
| 1197 | } |
| 1198 | |
| 1199 | /* Start after all badblocks */ |
| 1200 | if ((prev + 1) >= bb->count && !overlap_front(bb, prev, &bad)) { |
| 1201 | len = sectors; |
| 1202 | cleared++; |
| 1203 | goto update_sectors; |
| 1204 | } |
| 1205 | |
| 1206 | /* Clear will split a bad record but the table is full */ |
| 1207 | if (badblocks_full(bb) && (BB_OFFSET(p[prev]) < bad.start) && |
| 1208 | (BB_END(p[prev]) > (bad.start + sectors))) { |
| 1209 | len = sectors; |
| 1210 | goto update_sectors; |
| 1211 | } |
| 1212 | |
| 1213 | if (overlap_front(bb, prev, &bad)) { |
| 1214 | if ((BB_OFFSET(p[prev]) < bad.start) && |
| 1215 | (BB_END(p[prev]) > (bad.start + bad.len))) { |
| 1216 | /* Splitting */ |
| 1217 | if ((bb->count + 1) < MAX_BADBLOCKS) { |
| 1218 | len = front_splitting_clear(bb, prev, &bad); |
| 1219 | bb->count += 1; |
| 1220 | cleared++; |
| 1221 | } else { |
| 1222 | /* No space to split, give up */ |
| 1223 | len = sectors; |
| 1224 | } |
| 1225 | } else { |
| 1226 | int deleted = 0; |
| 1227 | |
| 1228 | len = front_clear(bb, prev, &bad, &deleted); |
| 1229 | bb->count -= deleted; |
| 1230 | cleared++; |
| 1231 | hint = prev; |
| 1232 | } |
| 1233 | |
| 1234 | goto update_sectors; |
| 1235 | } |
| 1236 | |
| 1237 | /* Not front overlap, but behind overlap */ |
| 1238 | if ((prev + 1) < bb->count && overlap_behind(bb, &bad, prev + 1)) { |
| 1239 | len = BB_OFFSET(p[prev + 1]) - bad.start; |
| 1240 | hint = prev + 1; |
| 1241 | /* Clear non-bad range should be treated as successful */ |
| 1242 | cleared++; |
| 1243 | goto update_sectors; |
| 1244 | } |
| 1245 | |
| 1246 | /* Not cover any badblocks range in the table */ |
| 1247 | len = sectors; |
| 1248 | /* Clear non-bad range should be treated as successful */ |
| 1249 | cleared++; |
| 1250 | |
| 1251 | update_sectors: |
| 1252 | s += len; |
| 1253 | sectors -= len; |
| 1254 | |
| 1255 | if (sectors > 0) |
| 1256 | goto re_clear; |
| 1257 | |
| 1258 | WARN_ON(sectors < 0); |
| 1259 | |
| 1260 | if (cleared) { |
| 1261 | badblocks_update_acked(bb); |
| 1262 | set_changed(bb); |
| 1263 | } |
| 1264 | |
| 1265 | write_sequnlock_irq(&bb->lock); |
| 1266 | |
| 1267 | if (!cleared) |
| 1268 | rv = 1; |
| 1269 | |
| 1270 | return rv; |
| 1271 | } |
| 1272 | |
| 1273 | /* Do the exact work to check bad blocks range from the bad block table */ |
| 1274 | static int _badblocks_check(struct badblocks *bb, sector_t s, int sectors, |
| 1275 | sector_t *first_bad, int *bad_sectors) |
| 1276 | { |
| 1277 | int unacked_badblocks, acked_badblocks; |
| 1278 | int prev = -1, hint = -1, set = 0; |
| 1279 | struct badblocks_context bad; |
| 1280 | unsigned int seq; |
| 1281 | int len, rv; |
| 1282 | u64 *p; |
| 1283 | |
| 1284 | WARN_ON(bb->shift < 0 || sectors == 0); |
| 1285 | |
| 1286 | if (bb->shift > 0) { |
| 1287 | sector_t target; |
| 1288 | |
| 1289 | /* round the start down, and the end up */ |
| 1290 | target = s + sectors; |
| 1291 | rounddown(s, bb->shift); |
| 1292 | roundup(target, bb->shift); |
| 1293 | sectors = target - s; |
| 1294 | } |
| 1295 | |
| 1296 | retry: |
| 1297 | seq = read_seqbegin(&bb->lock); |
| 1298 | |
| 1299 | p = bb->page; |
| 1300 | unacked_badblocks = 0; |
| 1301 | acked_badblocks = 0; |
| 1302 | |
| 1303 | re_check: |
| 1304 | bad.start = s; |
| 1305 | bad.len = sectors; |
| 1306 | |
| 1307 | if (badblocks_empty(bb)) { |
| 1308 | len = sectors; |
| 1309 | goto update_sectors; |
| 1310 | } |
| 1311 | |
| 1312 | prev = prev_badblocks(bb, &bad, hint); |
| 1313 | |
| 1314 | /* start after all badblocks */ |
| 1315 | if ((prev >= 0) && |
| 1316 | ((prev + 1) >= bb->count) && !overlap_front(bb, prev, &bad)) { |
| 1317 | len = sectors; |
| 1318 | goto update_sectors; |
| 1319 | } |
| 1320 | |
| 1321 | /* Overlapped with front badblocks record */ |
| 1322 | if ((prev >= 0) && overlap_front(bb, prev, &bad)) { |
| 1323 | if (BB_ACK(p[prev])) |
| 1324 | acked_badblocks++; |
| 1325 | else |
| 1326 | unacked_badblocks++; |
| 1327 | |
| 1328 | if (BB_END(p[prev]) >= (s + sectors)) |
| 1329 | len = sectors; |
| 1330 | else |
| 1331 | len = BB_END(p[prev]) - s; |
| 1332 | |
| 1333 | if (set == 0) { |
| 1334 | *first_bad = BB_OFFSET(p[prev]); |
| 1335 | *bad_sectors = BB_LEN(p[prev]); |
| 1336 | set = 1; |
| 1337 | } |
| 1338 | goto update_sectors; |
| 1339 | } |
| 1340 | |
| 1341 | /* Not front overlap, but behind overlap */ |
| 1342 | if ((prev + 1) < bb->count && overlap_behind(bb, &bad, prev + 1)) { |
| 1343 | len = BB_OFFSET(p[prev + 1]) - bad.start; |
| 1344 | hint = prev + 1; |
| 1345 | goto update_sectors; |
| 1346 | } |
| 1347 | |
| 1348 | /* not cover any badblocks range in the table */ |
| 1349 | len = sectors; |
| 1350 | |
| 1351 | update_sectors: |
| 1352 | s += len; |
| 1353 | sectors -= len; |
| 1354 | |
| 1355 | if (sectors > 0) |
| 1356 | goto re_check; |
| 1357 | |
| 1358 | WARN_ON(sectors < 0); |
| 1359 | |
| 1360 | if (unacked_badblocks > 0) |
| 1361 | rv = -1; |
| 1362 | else if (acked_badblocks > 0) |
| 1363 | rv = 1; |
| 1364 | else |
| 1365 | rv = 0; |
| 1366 | |
| 1367 | if (read_seqretry(&bb->lock, seq)) |
| 1368 | goto retry; |
| 1369 | |
| 1370 | return rv; |
| 1371 | } |
| 1372 | |
| 1373 | /** |
| 1374 | * badblocks_check() - check a given range for bad sectors |
| 1375 | * @bb: the badblocks structure that holds all badblock information |
| 1376 | * @s: sector (start) at which to check for badblocks |
| 1377 | * @sectors: number of sectors to check for badblocks |
| 1378 | * @first_bad: pointer to store location of the first badblock |
| 1379 | * @bad_sectors: pointer to store number of badblocks after @first_bad |
| 1380 | * |
| 1381 | * We can record which blocks on each device are 'bad' and so just |
| 1382 | * fail those blocks, or that stripe, rather than the whole device. |
| 1383 | * Entries in the bad-block table are 64bits wide. This comprises: |
| 1384 | * Length of bad-range, in sectors: 0-511 for lengths 1-512 |
| 1385 | * Start of bad-range, sector offset, 54 bits (allows 8 exbibytes) |
| 1386 | * A 'shift' can be set so that larger blocks are tracked and |
| 1387 | * consequently larger devices can be covered. |
| 1388 | * 'Acknowledged' flag - 1 bit. - the most significant bit. |
| 1389 | * |
| 1390 | * Locking of the bad-block table uses a seqlock so badblocks_check |
| 1391 | * might need to retry if it is very unlucky. |
| 1392 | * We will sometimes want to check for bad blocks in a bi_end_io function, |
| 1393 | * so we use the write_seqlock_irq variant. |
| 1394 | * |
| 1395 | * When looking for a bad block we specify a range and want to |
| 1396 | * know if any block in the range is bad. So we binary-search |
| 1397 | * to the last range that starts at-or-before the given endpoint, |
| 1398 | * (or "before the sector after the target range") |
| 1399 | * then see if it ends after the given start. |
| 1400 | * |
| 1401 | * Return: |
| 1402 | * 0: there are no known bad blocks in the range |
| 1403 | * 1: there are known bad block which are all acknowledged |
| 1404 | * -1: there are bad blocks which have not yet been acknowledged in metadata. |
| 1405 | * plus the start/length of the first bad section we overlap. |
| 1406 | */ |
| 1407 | int badblocks_check(struct badblocks *bb, sector_t s, int sectors, |
| 1408 | sector_t *first_bad, int *bad_sectors) |
| 1409 | { |
| 1410 | return _badblocks_check(bb, s, sectors, first_bad, bad_sectors); |
| 1411 | } |
| 1412 | EXPORT_SYMBOL_GPL(badblocks_check); |
| 1413 | |
| 1414 | /** |
| 1415 | * badblocks_set() - Add a range of bad blocks to the table. |
| 1416 | * @bb: the badblocks structure that holds all badblock information |
| 1417 | * @s: first sector to mark as bad |
| 1418 | * @sectors: number of sectors to mark as bad |
| 1419 | * @acknowledged: weather to mark the bad sectors as acknowledged |
| 1420 | * |
| 1421 | * This might extend the table, or might contract it if two adjacent ranges |
| 1422 | * can be merged. We binary-search to find the 'insertion' point, then |
| 1423 | * decide how best to handle it. |
| 1424 | * |
| 1425 | * Return: |
| 1426 | * 0: success |
| 1427 | * 1: failed to set badblocks (out of space) |
| 1428 | */ |
| 1429 | int badblocks_set(struct badblocks *bb, sector_t s, int sectors, |
| 1430 | int acknowledged) |
| 1431 | { |
| 1432 | return _badblocks_set(bb, s, sectors, acknowledged); |
| 1433 | } |
| 1434 | EXPORT_SYMBOL_GPL(badblocks_set); |
| 1435 | |
| 1436 | /** |
| 1437 | * badblocks_clear() - Remove a range of bad blocks to the table. |
| 1438 | * @bb: the badblocks structure that holds all badblock information |
| 1439 | * @s: first sector to mark as bad |
| 1440 | * @sectors: number of sectors to mark as bad |
| 1441 | * |
| 1442 | * This may involve extending the table if we spilt a region, |
| 1443 | * but it must not fail. So if the table becomes full, we just |
| 1444 | * drop the remove request. |
| 1445 | * |
| 1446 | * Return: |
| 1447 | * 0: success |
| 1448 | * 1: failed to clear badblocks |
| 1449 | */ |
| 1450 | int badblocks_clear(struct badblocks *bb, sector_t s, int sectors) |
| 1451 | { |
| 1452 | return _badblocks_clear(bb, s, sectors); |
| 1453 | } |
| 1454 | EXPORT_SYMBOL_GPL(badblocks_clear); |
| 1455 | |
| 1456 | /** |
| 1457 | * ack_all_badblocks() - Acknowledge all bad blocks in a list. |
| 1458 | * @bb: the badblocks structure that holds all badblock information |
| 1459 | * |
| 1460 | * This only succeeds if ->changed is clear. It is used by |
| 1461 | * in-kernel metadata updates |
| 1462 | */ |
| 1463 | void ack_all_badblocks(struct badblocks *bb) |
| 1464 | { |
| 1465 | if (bb->page == NULL || bb->changed) |
| 1466 | /* no point even trying */ |
| 1467 | return; |
| 1468 | write_seqlock_irq(&bb->lock); |
| 1469 | |
| 1470 | if (bb->changed == 0 && bb->unacked_exist) { |
| 1471 | u64 *p = bb->page; |
| 1472 | int i; |
| 1473 | |
| 1474 | for (i = 0; i < bb->count ; i++) { |
| 1475 | if (!BB_ACK(p[i])) { |
| 1476 | sector_t start = BB_OFFSET(p[i]); |
| 1477 | int len = BB_LEN(p[i]); |
| 1478 | |
| 1479 | p[i] = BB_MAKE(start, len, 1); |
| 1480 | } |
| 1481 | } |
| 1482 | bb->unacked_exist = 0; |
| 1483 | } |
| 1484 | write_sequnlock_irq(&bb->lock); |
| 1485 | } |
| 1486 | EXPORT_SYMBOL_GPL(ack_all_badblocks); |
| 1487 | |
| 1488 | /** |
| 1489 | * badblocks_show() - sysfs access to bad-blocks list |
| 1490 | * @bb: the badblocks structure that holds all badblock information |
| 1491 | * @page: buffer received from sysfs |
| 1492 | * @unack: weather to show unacknowledged badblocks |
| 1493 | * |
| 1494 | * Return: |
| 1495 | * Length of returned data |
| 1496 | */ |
| 1497 | ssize_t badblocks_show(struct badblocks *bb, char *page, int unack) |
| 1498 | { |
| 1499 | size_t len; |
| 1500 | int i; |
| 1501 | u64 *p = bb->page; |
| 1502 | unsigned seq; |
| 1503 | |
| 1504 | if (bb->shift < 0) |
| 1505 | return 0; |
| 1506 | |
| 1507 | retry: |
| 1508 | seq = read_seqbegin(&bb->lock); |
| 1509 | |
| 1510 | len = 0; |
| 1511 | i = 0; |
| 1512 | |
| 1513 | while (len < PAGE_SIZE && i < bb->count) { |
| 1514 | sector_t s = BB_OFFSET(p[i]); |
| 1515 | unsigned int length = BB_LEN(p[i]); |
| 1516 | int ack = BB_ACK(p[i]); |
| 1517 | |
| 1518 | i++; |
| 1519 | |
| 1520 | if (unack && ack) |
| 1521 | continue; |
| 1522 | |
| 1523 | len += snprintf(page+len, PAGE_SIZE-len, "%llu %u\n", |
| 1524 | (unsigned long long)s << bb->shift, |
| 1525 | length << bb->shift); |
| 1526 | } |
| 1527 | if (unack && len == 0) |
| 1528 | bb->unacked_exist = 0; |
| 1529 | |
| 1530 | if (read_seqretry(&bb->lock, seq)) |
| 1531 | goto retry; |
| 1532 | |
| 1533 | return len; |
| 1534 | } |
| 1535 | EXPORT_SYMBOL_GPL(badblocks_show); |
| 1536 | |
| 1537 | /** |
| 1538 | * badblocks_store() - sysfs access to bad-blocks list |
| 1539 | * @bb: the badblocks structure that holds all badblock information |
| 1540 | * @page: buffer received from sysfs |
| 1541 | * @len: length of data received from sysfs |
| 1542 | * @unack: weather to show unacknowledged badblocks |
| 1543 | * |
| 1544 | * Return: |
| 1545 | * Length of the buffer processed or -ve error. |
| 1546 | */ |
| 1547 | ssize_t badblocks_store(struct badblocks *bb, const char *page, size_t len, |
| 1548 | int unack) |
| 1549 | { |
| 1550 | unsigned long long sector; |
| 1551 | int length; |
| 1552 | char newline; |
| 1553 | |
| 1554 | switch (sscanf(page, "%llu %d%c", §or, &length, &newline)) { |
| 1555 | case 3: |
| 1556 | if (newline != '\n') |
| 1557 | return -EINVAL; |
| 1558 | fallthrough; |
| 1559 | case 2: |
| 1560 | if (length <= 0) |
| 1561 | return -EINVAL; |
| 1562 | break; |
| 1563 | default: |
| 1564 | return -EINVAL; |
| 1565 | } |
| 1566 | |
| 1567 | if (badblocks_set(bb, sector, length, !unack)) |
| 1568 | return -ENOSPC; |
| 1569 | else |
| 1570 | return len; |
| 1571 | } |
| 1572 | EXPORT_SYMBOL_GPL(badblocks_store); |
| 1573 | |
| 1574 | static int __badblocks_init(struct device *dev, struct badblocks *bb, |
| 1575 | int enable) |
| 1576 | { |
| 1577 | bb->dev = dev; |
| 1578 | bb->count = 0; |
| 1579 | if (enable) |
| 1580 | bb->shift = 0; |
| 1581 | else |
| 1582 | bb->shift = -1; |
| 1583 | if (dev) |
| 1584 | bb->page = devm_kzalloc(dev, PAGE_SIZE, GFP_KERNEL); |
| 1585 | else |
| 1586 | bb->page = kzalloc(PAGE_SIZE, GFP_KERNEL); |
| 1587 | if (!bb->page) { |
| 1588 | bb->shift = -1; |
| 1589 | return -ENOMEM; |
| 1590 | } |
| 1591 | seqlock_init(&bb->lock); |
| 1592 | |
| 1593 | return 0; |
| 1594 | } |
| 1595 | |
| 1596 | /** |
| 1597 | * badblocks_init() - initialize the badblocks structure |
| 1598 | * @bb: the badblocks structure that holds all badblock information |
| 1599 | * @enable: weather to enable badblocks accounting |
| 1600 | * |
| 1601 | * Return: |
| 1602 | * 0: success |
| 1603 | * -ve errno: on error |
| 1604 | */ |
| 1605 | int badblocks_init(struct badblocks *bb, int enable) |
| 1606 | { |
| 1607 | return __badblocks_init(NULL, bb, enable); |
| 1608 | } |
| 1609 | EXPORT_SYMBOL_GPL(badblocks_init); |
| 1610 | |
| 1611 | int devm_init_badblocks(struct device *dev, struct badblocks *bb) |
| 1612 | { |
| 1613 | if (!bb) |
| 1614 | return -EINVAL; |
| 1615 | return __badblocks_init(dev, bb, 1); |
| 1616 | } |
| 1617 | EXPORT_SYMBOL_GPL(devm_init_badblocks); |
| 1618 | |
| 1619 | /** |
| 1620 | * badblocks_exit() - free the badblocks structure |
| 1621 | * @bb: the badblocks structure that holds all badblock information |
| 1622 | */ |
| 1623 | void badblocks_exit(struct badblocks *bb) |
| 1624 | { |
| 1625 | if (!bb) |
| 1626 | return; |
| 1627 | if (bb->dev) |
| 1628 | devm_kfree(bb->dev, bb->page); |
| 1629 | else |
| 1630 | kfree(bb->page); |
| 1631 | bb->page = NULL; |
| 1632 | } |
| 1633 | EXPORT_SYMBOL_GPL(badblocks_exit); |