1// SPDX-License-Identifier: GPL-2.0 2/* 3 * Copyright (c) 2000-2005 Silicon Graphics, Inc. 4 * All Rights Reserved. 5 */ 6#include "xfs.h" 7#include "xfs_fs.h" 8#include "xfs_shared.h" 9#include "xfs_format.h" 10#include "xfs_log_format.h" 11#include "xfs_trans_resv.h" 12#include "xfs_sb.h" 13#include "xfs_mount.h" 14#include "xfs_trans.h" 15#include "xfs_error.h" 16#include "xfs_alloc.h" 17#include "xfs_fsops.h" 18#include "xfs_trans_space.h" 19#include "xfs_log.h" 20#include "xfs_log_priv.h" 21#include "xfs_ag.h" 22#include "xfs_ag_resv.h" 23#include "xfs_trace.h" 24 25/* 26 * Write new AG headers to disk. Non-transactional, but need to be 27 * written and completed prior to the growfs transaction being logged. 28 * To do this, we use a delayed write buffer list and wait for 29 * submission and IO completion of the list as a whole. This allows the 30 * IO subsystem to merge all the AG headers in a single AG into a single 31 * IO and hide most of the latency of the IO from us. 32 * 33 * This also means that if we get an error whilst building the buffer 34 * list to write, we can cancel the entire list without having written 35 * anything. 36 */ 37static int 38xfs_resizefs_init_new_ags( 39 struct xfs_trans *tp, 40 struct aghdr_init_data *id, 41 xfs_agnumber_t oagcount, 42 xfs_agnumber_t nagcount, 43 xfs_rfsblock_t delta, 44 struct xfs_perag *last_pag, 45 bool *lastag_extended) 46{ 47 struct xfs_mount *mp = tp->t_mountp; 48 xfs_rfsblock_t nb = mp->m_sb.sb_dblocks + delta; 49 int error; 50 51 *lastag_extended = false; 52 53 INIT_LIST_HEAD(&id->buffer_list); 54 for (id->agno = nagcount - 1; 55 id->agno >= oagcount; 56 id->agno--, delta -= id->agsize) { 57 58 if (id->agno == nagcount - 1) 59 id->agsize = nb - (id->agno * 60 (xfs_rfsblock_t)mp->m_sb.sb_agblocks); 61 else 62 id->agsize = mp->m_sb.sb_agblocks; 63 64 error = xfs_ag_init_headers(mp, id); 65 if (error) { 66 xfs_buf_delwri_cancel(&id->buffer_list); 67 return error; 68 } 69 } 70 71 error = xfs_buf_delwri_submit(&id->buffer_list); 72 if (error) 73 return error; 74 75 if (delta) { 76 *lastag_extended = true; 77 error = xfs_ag_extend_space(last_pag, tp, delta); 78 } 79 return error; 80} 81 82/* 83 * growfs operations 84 */ 85static int 86xfs_growfs_data_private( 87 struct xfs_mount *mp, /* mount point for filesystem */ 88 struct xfs_growfs_data *in) /* growfs data input struct */ 89{ 90 struct xfs_buf *bp; 91 int error; 92 xfs_agnumber_t nagcount; 93 xfs_agnumber_t nagimax = 0; 94 xfs_rfsblock_t nb, nb_div, nb_mod; 95 int64_t delta; 96 bool lastag_extended = false; 97 xfs_agnumber_t oagcount; 98 struct xfs_trans *tp; 99 struct aghdr_init_data id = {}; 100 struct xfs_perag *last_pag; 101 102 nb = in->newblocks; 103 error = xfs_sb_validate_fsb_count(&mp->m_sb, nb); 104 if (error) 105 return error; 106 107 if (nb > mp->m_sb.sb_dblocks) { 108 error = xfs_buf_read_uncached(mp->m_ddev_targp, 109 XFS_FSB_TO_BB(mp, nb) - XFS_FSS_TO_BB(mp, 1), 110 XFS_FSS_TO_BB(mp, 1), 0, &bp, NULL); 111 if (error) 112 return error; 113 xfs_buf_relse(bp); 114 } 115 116 nb_div = nb; 117 nb_mod = do_div(nb_div, mp->m_sb.sb_agblocks); 118 if (nb_mod && nb_mod >= XFS_MIN_AG_BLOCKS) 119 nb_div++; 120 else if (nb_mod) 121 nb = nb_div * mp->m_sb.sb_agblocks; 122 123 if (nb_div > XFS_MAX_AGNUMBER + 1) { 124 nb_div = XFS_MAX_AGNUMBER + 1; 125 nb = nb_div * mp->m_sb.sb_agblocks; 126 } 127 nagcount = nb_div; 128 delta = nb - mp->m_sb.sb_dblocks; 129 /* 130 * Reject filesystems with a single AG because they are not 131 * supported, and reject a shrink operation that would cause a 132 * filesystem to become unsupported. 133 */ 134 if (delta < 0 && nagcount < 2) 135 return -EINVAL; 136 137 oagcount = mp->m_sb.sb_agcount; 138 /* allocate the new per-ag structures */ 139 if (nagcount > oagcount) { 140 error = xfs_initialize_perag(mp, nagcount, nb, &nagimax); 141 if (error) 142 return error; 143 } else if (nagcount < oagcount) { 144 /* TODO: shrinking the entire AGs hasn't yet completed */ 145 return -EINVAL; 146 } 147 148 if (delta > 0) 149 error = xfs_trans_alloc(mp, &M_RES(mp)->tr_growdata, 150 XFS_GROWFS_SPACE_RES(mp), 0, XFS_TRANS_RESERVE, 151 &tp); 152 else 153 error = xfs_trans_alloc(mp, &M_RES(mp)->tr_growdata, -delta, 0, 154 0, &tp); 155 if (error) 156 return error; 157 158 last_pag = xfs_perag_get(mp, oagcount - 1); 159 if (delta > 0) { 160 error = xfs_resizefs_init_new_ags(tp, &id, oagcount, nagcount, 161 delta, last_pag, &lastag_extended); 162 } else { 163 xfs_warn_mount(mp, XFS_OPSTATE_WARNED_SHRINK, 164 "EXPERIMENTAL online shrink feature in use. Use at your own risk!"); 165 166 error = xfs_ag_shrink_space(last_pag, &tp, -delta); 167 } 168 xfs_perag_put(last_pag); 169 if (error) 170 goto out_trans_cancel; 171 172 /* 173 * Update changed superblock fields transactionally. These are not 174 * seen by the rest of the world until the transaction commit applies 175 * them atomically to the superblock. 176 */ 177 if (nagcount > oagcount) 178 xfs_trans_mod_sb(tp, XFS_TRANS_SB_AGCOUNT, nagcount - oagcount); 179 if (delta) 180 xfs_trans_mod_sb(tp, XFS_TRANS_SB_DBLOCKS, delta); 181 if (id.nfree) 182 xfs_trans_mod_sb(tp, XFS_TRANS_SB_FDBLOCKS, id.nfree); 183 184 /* 185 * Sync sb counters now to reflect the updated values. This is 186 * particularly important for shrink because the write verifier 187 * will fail if sb_fdblocks is ever larger than sb_dblocks. 188 */ 189 if (xfs_has_lazysbcount(mp)) 190 xfs_log_sb(tp); 191 192 xfs_trans_set_sync(tp); 193 error = xfs_trans_commit(tp); 194 if (error) 195 return error; 196 197 /* New allocation groups fully initialized, so update mount struct */ 198 if (nagimax) 199 mp->m_maxagi = nagimax; 200 xfs_set_low_space_thresholds(mp); 201 mp->m_alloc_set_aside = xfs_alloc_set_aside(mp); 202 203 if (delta > 0) { 204 /* 205 * If we expanded the last AG, free the per-AG reservation 206 * so we can reinitialize it with the new size. 207 */ 208 if (lastag_extended) { 209 struct xfs_perag *pag; 210 211 pag = xfs_perag_get(mp, id.agno); 212 error = xfs_ag_resv_free(pag); 213 xfs_perag_put(pag); 214 if (error) 215 return error; 216 } 217 /* 218 * Reserve AG metadata blocks. ENOSPC here does not mean there 219 * was a growfs failure, just that there still isn't space for 220 * new user data after the grow has been run. 221 */ 222 error = xfs_fs_reserve_ag_blocks(mp); 223 if (error == -ENOSPC) 224 error = 0; 225 } 226 return error; 227 228out_trans_cancel: 229 xfs_trans_cancel(tp); 230 return error; 231} 232 233static int 234xfs_growfs_log_private( 235 struct xfs_mount *mp, /* mount point for filesystem */ 236 struct xfs_growfs_log *in) /* growfs log input struct */ 237{ 238 xfs_extlen_t nb; 239 240 nb = in->newblocks; 241 if (nb < XFS_MIN_LOG_BLOCKS || nb < XFS_B_TO_FSB(mp, XFS_MIN_LOG_BYTES)) 242 return -EINVAL; 243 if (nb == mp->m_sb.sb_logblocks && 244 in->isint == (mp->m_sb.sb_logstart != 0)) 245 return -EINVAL; 246 /* 247 * Moving the log is hard, need new interfaces to sync 248 * the log first, hold off all activity while moving it. 249 * Can have shorter or longer log in the same space, 250 * or transform internal to external log or vice versa. 251 */ 252 return -ENOSYS; 253} 254 255static int 256xfs_growfs_imaxpct( 257 struct xfs_mount *mp, 258 __u32 imaxpct) 259{ 260 struct xfs_trans *tp; 261 int dpct; 262 int error; 263 264 if (imaxpct > 100) 265 return -EINVAL; 266 267 error = xfs_trans_alloc(mp, &M_RES(mp)->tr_growdata, 268 XFS_GROWFS_SPACE_RES(mp), 0, XFS_TRANS_RESERVE, &tp); 269 if (error) 270 return error; 271 272 dpct = imaxpct - mp->m_sb.sb_imax_pct; 273 xfs_trans_mod_sb(tp, XFS_TRANS_SB_IMAXPCT, dpct); 274 xfs_trans_set_sync(tp); 275 return xfs_trans_commit(tp); 276} 277 278/* 279 * protected versions of growfs function acquire and release locks on the mount 280 * point - exported through ioctls: XFS_IOC_FSGROWFSDATA, XFS_IOC_FSGROWFSLOG, 281 * XFS_IOC_FSGROWFSRT 282 */ 283int 284xfs_growfs_data( 285 struct xfs_mount *mp, 286 struct xfs_growfs_data *in) 287{ 288 int error = 0; 289 290 if (!capable(CAP_SYS_ADMIN)) 291 return -EPERM; 292 if (!mutex_trylock(&mp->m_growlock)) 293 return -EWOULDBLOCK; 294 295 /* update imaxpct separately to the physical grow of the filesystem */ 296 if (in->imaxpct != mp->m_sb.sb_imax_pct) { 297 error = xfs_growfs_imaxpct(mp, in->imaxpct); 298 if (error) 299 goto out_error; 300 } 301 302 if (in->newblocks != mp->m_sb.sb_dblocks) { 303 error = xfs_growfs_data_private(mp, in); 304 if (error) 305 goto out_error; 306 } 307 308 /* Post growfs calculations needed to reflect new state in operations */ 309 if (mp->m_sb.sb_imax_pct) { 310 uint64_t icount = mp->m_sb.sb_dblocks * mp->m_sb.sb_imax_pct; 311 do_div(icount, 100); 312 M_IGEO(mp)->maxicount = XFS_FSB_TO_INO(mp, icount); 313 } else 314 M_IGEO(mp)->maxicount = 0; 315 316 /* Update secondary superblocks now the physical grow has completed */ 317 error = xfs_update_secondary_sbs(mp); 318 319out_error: 320 /* 321 * Increment the generation unconditionally, the error could be from 322 * updating the secondary superblocks, in which case the new size 323 * is live already. 324 */ 325 mp->m_generation++; 326 mutex_unlock(&mp->m_growlock); 327 return error; 328} 329 330int 331xfs_growfs_log( 332 xfs_mount_t *mp, 333 struct xfs_growfs_log *in) 334{ 335 int error; 336 337 if (!capable(CAP_SYS_ADMIN)) 338 return -EPERM; 339 if (!mutex_trylock(&mp->m_growlock)) 340 return -EWOULDBLOCK; 341 error = xfs_growfs_log_private(mp, in); 342 mutex_unlock(&mp->m_growlock); 343 return error; 344} 345 346/* 347 * exported through ioctl XFS_IOC_FSCOUNTS 348 */ 349 350void 351xfs_fs_counts( 352 xfs_mount_t *mp, 353 xfs_fsop_counts_t *cnt) 354{ 355 cnt->allocino = percpu_counter_read_positive(&mp->m_icount); 356 cnt->freeino = percpu_counter_read_positive(&mp->m_ifree); 357 cnt->freedata = percpu_counter_read_positive(&mp->m_fdblocks) - 358 xfs_fdblocks_unavailable(mp); 359 cnt->freertx = percpu_counter_read_positive(&mp->m_frextents); 360} 361 362/* 363 * exported through ioctl XFS_IOC_SET_RESBLKS & XFS_IOC_GET_RESBLKS 364 * 365 * xfs_reserve_blocks is called to set m_resblks 366 * in the in-core mount table. The number of unused reserved blocks 367 * is kept in m_resblks_avail. 368 * 369 * Reserve the requested number of blocks if available. Otherwise return 370 * as many as possible to satisfy the request. The actual number 371 * reserved are returned in outval 372 * 373 * A null inval pointer indicates that only the current reserved blocks 374 * available should be returned no settings are changed. 375 */ 376 377int 378xfs_reserve_blocks( 379 xfs_mount_t *mp, 380 uint64_t *inval, 381 xfs_fsop_resblks_t *outval) 382{ 383 int64_t lcounter, delta; 384 int64_t fdblks_delta = 0; 385 uint64_t request; 386 int64_t free; 387 int error = 0; 388 389 /* If inval is null, report current values and return */ 390 if (inval == (uint64_t *)NULL) { 391 if (!outval) 392 return -EINVAL; 393 outval->resblks = mp->m_resblks; 394 outval->resblks_avail = mp->m_resblks_avail; 395 return 0; 396 } 397 398 request = *inval; 399 400 /* 401 * With per-cpu counters, this becomes an interesting problem. we need 402 * to work out if we are freeing or allocation blocks first, then we can 403 * do the modification as necessary. 404 * 405 * We do this under the m_sb_lock so that if we are near ENOSPC, we will 406 * hold out any changes while we work out what to do. This means that 407 * the amount of free space can change while we do this, so we need to 408 * retry if we end up trying to reserve more space than is available. 409 */ 410 spin_lock(&mp->m_sb_lock); 411 412 /* 413 * If our previous reservation was larger than the current value, 414 * then move any unused blocks back to the free pool. Modify the resblks 415 * counters directly since we shouldn't have any problems unreserving 416 * space. 417 */ 418 if (mp->m_resblks > request) { 419 lcounter = mp->m_resblks_avail - request; 420 if (lcounter > 0) { /* release unused blocks */ 421 fdblks_delta = lcounter; 422 mp->m_resblks_avail -= lcounter; 423 } 424 mp->m_resblks = request; 425 if (fdblks_delta) { 426 spin_unlock(&mp->m_sb_lock); 427 error = xfs_mod_fdblocks(mp, fdblks_delta, 0); 428 spin_lock(&mp->m_sb_lock); 429 } 430 431 goto out; 432 } 433 434 /* 435 * If the request is larger than the current reservation, reserve the 436 * blocks before we update the reserve counters. Sample m_fdblocks and 437 * perform a partial reservation if the request exceeds free space. 438 * 439 * The code below estimates how many blocks it can request from 440 * fdblocks to stash in the reserve pool. This is a classic TOCTOU 441 * race since fdblocks updates are not always coordinated via 442 * m_sb_lock. Set the reserve size even if there's not enough free 443 * space to fill it because mod_fdblocks will refill an undersized 444 * reserve when it can. 445 */ 446 free = percpu_counter_sum(&mp->m_fdblocks) - 447 xfs_fdblocks_unavailable(mp); 448 delta = request - mp->m_resblks; 449 mp->m_resblks = request; 450 if (delta > 0 && free > 0) { 451 /* 452 * We'll either succeed in getting space from the free block 453 * count or we'll get an ENOSPC. Don't set the reserved flag 454 * here - we don't want to reserve the extra reserve blocks 455 * from the reserve. 456 * 457 * The desired reserve size can change after we drop the lock. 458 * Use mod_fdblocks to put the space into the reserve or into 459 * fdblocks as appropriate. 460 */ 461 fdblks_delta = min(free, delta); 462 spin_unlock(&mp->m_sb_lock); 463 error = xfs_mod_fdblocks(mp, -fdblks_delta, 0); 464 if (!error) 465 xfs_mod_fdblocks(mp, fdblks_delta, 0); 466 spin_lock(&mp->m_sb_lock); 467 } 468out: 469 if (outval) { 470 outval->resblks = mp->m_resblks; 471 outval->resblks_avail = mp->m_resblks_avail; 472 } 473 474 spin_unlock(&mp->m_sb_lock); 475 return error; 476} 477 478int 479xfs_fs_goingdown( 480 xfs_mount_t *mp, 481 uint32_t inflags) 482{ 483 switch (inflags) { 484 case XFS_FSOP_GOING_FLAGS_DEFAULT: { 485 if (!freeze_bdev(mp->m_super->s_bdev)) { 486 xfs_force_shutdown(mp, SHUTDOWN_FORCE_UMOUNT); 487 thaw_bdev(mp->m_super->s_bdev); 488 } 489 break; 490 } 491 case XFS_FSOP_GOING_FLAGS_LOGFLUSH: 492 xfs_force_shutdown(mp, SHUTDOWN_FORCE_UMOUNT); 493 break; 494 case XFS_FSOP_GOING_FLAGS_NOLOGFLUSH: 495 xfs_force_shutdown(mp, 496 SHUTDOWN_FORCE_UMOUNT | SHUTDOWN_LOG_IO_ERROR); 497 break; 498 default: 499 return -EINVAL; 500 } 501 502 return 0; 503} 504 505/* 506 * Force a shutdown of the filesystem instantly while keeping the filesystem 507 * consistent. We don't do an unmount here; just shutdown the shop, make sure 508 * that absolutely nothing persistent happens to this filesystem after this 509 * point. 510 * 511 * The shutdown state change is atomic, resulting in the first and only the 512 * first shutdown call processing the shutdown. This means we only shutdown the 513 * log once as it requires, and we don't spam the logs when multiple concurrent 514 * shutdowns race to set the shutdown flags. 515 */ 516void 517xfs_do_force_shutdown( 518 struct xfs_mount *mp, 519 uint32_t flags, 520 char *fname, 521 int lnnum) 522{ 523 int tag; 524 const char *why; 525 526 527 if (test_and_set_bit(XFS_OPSTATE_SHUTDOWN, &mp->m_opstate)) { 528 xlog_shutdown_wait(mp->m_log); 529 return; 530 } 531 if (mp->m_sb_bp) 532 mp->m_sb_bp->b_flags |= XBF_DONE; 533 534 if (flags & SHUTDOWN_FORCE_UMOUNT) 535 xfs_alert(mp, "User initiated shutdown received."); 536 537 if (xlog_force_shutdown(mp->m_log, flags)) { 538 tag = XFS_PTAG_SHUTDOWN_LOGERROR; 539 why = "Log I/O Error"; 540 } else if (flags & SHUTDOWN_CORRUPT_INCORE) { 541 tag = XFS_PTAG_SHUTDOWN_CORRUPT; 542 why = "Corruption of in-memory data"; 543 } else if (flags & SHUTDOWN_CORRUPT_ONDISK) { 544 tag = XFS_PTAG_SHUTDOWN_CORRUPT; 545 why = "Corruption of on-disk metadata"; 546 } else if (flags & SHUTDOWN_DEVICE_REMOVED) { 547 tag = XFS_PTAG_SHUTDOWN_IOERROR; 548 why = "Block device removal"; 549 } else { 550 tag = XFS_PTAG_SHUTDOWN_IOERROR; 551 why = "Metadata I/O Error"; 552 } 553 554 trace_xfs_force_shutdown(mp, tag, flags, fname, lnnum); 555 556 xfs_alert_tag(mp, tag, 557"%s (0x%x) detected at %pS (%s:%d). Shutting down filesystem.", 558 why, flags, __return_address, fname, lnnum); 559 xfs_alert(mp, 560 "Please unmount the filesystem and rectify the problem(s)"); 561 if (xfs_error_level >= XFS_ERRLEVEL_HIGH) 562 xfs_stack_trace(); 563} 564 565/* 566 * Reserve free space for per-AG metadata. 567 */ 568int 569xfs_fs_reserve_ag_blocks( 570 struct xfs_mount *mp) 571{ 572 xfs_agnumber_t agno; 573 struct xfs_perag *pag; 574 int error = 0; 575 int err2; 576 577 mp->m_finobt_nores = false; 578 for_each_perag(mp, agno, pag) { 579 err2 = xfs_ag_resv_init(pag, NULL); 580 if (err2 && !error) 581 error = err2; 582 } 583 584 if (error && error != -ENOSPC) { 585 xfs_warn(mp, 586 "Error %d reserving per-AG metadata reserve pool.", error); 587 xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); 588 } 589 590 return error; 591} 592 593/* 594 * Free space reserved for per-AG metadata. 595 */ 596int 597xfs_fs_unreserve_ag_blocks( 598 struct xfs_mount *mp) 599{ 600 xfs_agnumber_t agno; 601 struct xfs_perag *pag; 602 int error = 0; 603 int err2; 604 605 for_each_perag(mp, agno, pag) { 606 err2 = xfs_ag_resv_free(pag); 607 if (err2 && !error) 608 error = err2; 609 } 610 611 if (error) 612 xfs_warn(mp, 613 "Error %d freeing per-AG metadata reserve pool.", error); 614 615 return error; 616} 617