^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0+
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) * Copyright (C) 2018 Oracle. All Rights Reserved.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) * Author: Darrick J. Wong <darrick.wong@oracle.com>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) #include "xfs.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) #include "xfs_fs.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) #include "xfs_shared.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) #include "xfs_format.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) #include "xfs_trans_resv.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) #include "xfs_mount.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) #include "xfs_btree.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) #include "xfs_log_format.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #include "xfs_trans.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #include "xfs_sb.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) #include "xfs_inode.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) #include "xfs_alloc.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) #include "xfs_alloc_btree.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) #include "xfs_ialloc.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) #include "xfs_ialloc_btree.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) #include "xfs_rmap.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) #include "xfs_rmap_btree.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) #include "xfs_refcount_btree.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) #include "xfs_extent_busy.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) #include "xfs_ag_resv.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) #include "xfs_quota.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) #include "scrub/scrub.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) #include "scrub/common.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) #include "scrub/trace.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) #include "scrub/repair.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) #include "scrub/bitmap.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) * Attempt to repair some metadata, if the metadata is corrupt and userspace
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) * told us to fix it. This function returns -EAGAIN to mean "re-run scrub",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) * and will set *fixed to true if it thinks it repaired anything.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) xrep_attempt(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) struct xfs_scrub *sc)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) trace_xrep_attempt(ip, sc->sm, error);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) xchk_ag_btcur_free(&sc->sa);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) /* Repair whatever's broken. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) ASSERT(sc->ops->repair);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) error = sc->ops->repair(sc);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) trace_xrep_done(ip, sc->sm, error);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) switch (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) case 0:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) * Repair succeeded. Commit the fixes and perform a second
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) * scrub so that we can tell userspace if we fixed the problem.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) sc->sm->sm_flags &= ~XFS_SCRUB_FLAGS_OUT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) sc->flags |= XREP_ALREADY_FIXED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) return -EAGAIN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) case -EDEADLOCK:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) case -EAGAIN:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) /* Tell the caller to try again having grabbed all the locks. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) if (!(sc->flags & XCHK_TRY_HARDER)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) sc->flags |= XCHK_TRY_HARDER;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) return -EAGAIN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) * We tried harder but still couldn't grab all the resources
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) * we needed to fix it. The corruption has not been fixed,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) * so report back to userspace.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) return -EFSCORRUPTED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) * Complain about unfixable problems in the filesystem. We don't log
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) * corruptions when IFLAG_REPAIR wasn't set on the assumption that the driver
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) * program is xfs_scrub, which will call back with IFLAG_REPAIR set if the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) * administrator isn't running xfs_scrub in no-repairs mode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) * Use this helper function because _ratelimited silently declares a static
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) * structure to track rate limiting information.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) xrep_failure(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) struct xfs_mount *mp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) xfs_alert_ratelimited(mp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) "Corruption not fixed during online repair. Unmount and run xfs_repair.");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) * Repair probe -- userspace uses this to probe if we're willing to repair a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) * given mountpoint.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) xrep_probe(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) struct xfs_scrub *sc)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) if (xchk_should_terminate(sc, &error))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) * Roll a transaction, keeping the AG headers locked and reinitializing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) * the btree cursors.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) xrep_roll_ag_trans(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) struct xfs_scrub *sc)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) /* Keep the AG header buffers locked so we can keep going. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) if (sc->sa.agi_bp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) xfs_trans_bhold(sc->tp, sc->sa.agi_bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) if (sc->sa.agf_bp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) xfs_trans_bhold(sc->tp, sc->sa.agf_bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) if (sc->sa.agfl_bp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) xfs_trans_bhold(sc->tp, sc->sa.agfl_bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) * Roll the transaction. We still own the buffer and the buffer lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) * regardless of whether or not the roll succeeds. If the roll fails,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) * the buffers will be released during teardown on our way out of the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) * kernel. If it succeeds, we join them to the new transaction and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) * move on.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) error = xfs_trans_roll(&sc->tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) /* Join AG headers to the new transaction. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) if (sc->sa.agi_bp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) xfs_trans_bjoin(sc->tp, sc->sa.agi_bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) if (sc->sa.agf_bp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) xfs_trans_bjoin(sc->tp, sc->sa.agf_bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) if (sc->sa.agfl_bp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) xfs_trans_bjoin(sc->tp, sc->sa.agfl_bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) * Does the given AG have enough space to rebuild a btree? Neither AG
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) * reservation can be critical, and we must have enough space (factoring
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) * in AG reservations) to construct a whole btree.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) bool
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) xrep_ag_has_space(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) struct xfs_perag *pag,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) xfs_extlen_t nr_blocks,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) enum xfs_ag_resv_type type)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) return !xfs_ag_resv_critical(pag, XFS_AG_RESV_RMAPBT) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) !xfs_ag_resv_critical(pag, XFS_AG_RESV_METADATA) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) pag->pagf_freeblks > xfs_ag_resv_needed(pag, type) + nr_blocks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) * Figure out how many blocks to reserve for an AG repair. We calculate the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) * worst case estimate for the number of blocks we'd need to rebuild one of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) * any type of per-AG btree.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) xfs_extlen_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) xrep_calc_ag_resblks(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) struct xfs_scrub *sc)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) struct xfs_mount *mp = sc->mp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) struct xfs_scrub_metadata *sm = sc->sm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) struct xfs_perag *pag;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) struct xfs_buf *bp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) xfs_agino_t icount = NULLAGINO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) xfs_extlen_t aglen = NULLAGBLOCK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) xfs_extlen_t usedlen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) xfs_extlen_t freelen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) xfs_extlen_t bnobt_sz;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) xfs_extlen_t inobt_sz;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) xfs_extlen_t rmapbt_sz;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) xfs_extlen_t refcbt_sz;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) if (!(sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) pag = xfs_perag_get(mp, sm->sm_agno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) if (pag->pagi_init) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) /* Use in-core icount if possible. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) icount = pag->pagi_count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) /* Try to get the actual counters from disk. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) error = xfs_ialloc_read_agi(mp, NULL, sm->sm_agno, &bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) if (!error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) icount = pag->pagi_count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) xfs_buf_relse(bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) /* Now grab the block counters from the AGF. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) error = xfs_alloc_read_agf(mp, NULL, sm->sm_agno, 0, &bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) if (!error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) struct xfs_agf *agf = bp->b_addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) aglen = be32_to_cpu(agf->agf_length);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) freelen = be32_to_cpu(agf->agf_freeblks);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) usedlen = aglen - freelen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) xfs_buf_relse(bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) xfs_perag_put(pag);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) /* If the icount is impossible, make some worst-case assumptions. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) if (icount == NULLAGINO ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) !xfs_verify_agino(mp, sm->sm_agno, icount)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) xfs_agino_t first, last;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) xfs_agino_range(mp, sm->sm_agno, &first, &last);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) icount = last - first + 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) /* If the block counts are impossible, make worst-case assumptions. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) if (aglen == NULLAGBLOCK ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) aglen != xfs_ag_block_count(mp, sm->sm_agno) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) freelen >= aglen) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) aglen = xfs_ag_block_count(mp, sm->sm_agno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) freelen = aglen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) usedlen = aglen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) trace_xrep_calc_ag_resblks(mp, sm->sm_agno, icount, aglen,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) freelen, usedlen);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) * Figure out how many blocks we'd need worst case to rebuild
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) * each type of btree. Note that we can only rebuild the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) * bnobt/cntbt or inobt/finobt as pairs.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) bnobt_sz = 2 * xfs_allocbt_calc_size(mp, freelen);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) if (xfs_sb_version_hassparseinodes(&mp->m_sb))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) inobt_sz = xfs_iallocbt_calc_size(mp, icount /
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) XFS_INODES_PER_HOLEMASK_BIT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) inobt_sz = xfs_iallocbt_calc_size(mp, icount /
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) XFS_INODES_PER_CHUNK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) if (xfs_sb_version_hasfinobt(&mp->m_sb))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) inobt_sz *= 2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) if (xfs_sb_version_hasreflink(&mp->m_sb))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) refcbt_sz = xfs_refcountbt_calc_size(mp, usedlen);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) refcbt_sz = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) if (xfs_sb_version_hasrmapbt(&mp->m_sb)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) * Guess how many blocks we need to rebuild the rmapbt.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) * For non-reflink filesystems we can't have more records than
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) * used blocks. However, with reflink it's possible to have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) * more than one rmap record per AG block. We don't know how
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) * many rmaps there could be in the AG, so we start off with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) * what we hope is an generous over-estimation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) if (xfs_sb_version_hasreflink(&mp->m_sb))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) rmapbt_sz = xfs_rmapbt_calc_size(mp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) (unsigned long long)aglen * 2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) rmapbt_sz = xfs_rmapbt_calc_size(mp, usedlen);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) rmapbt_sz = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) trace_xrep_calc_ag_resblks_btsize(mp, sm->sm_agno, bnobt_sz,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) inobt_sz, rmapbt_sz, refcbt_sz);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) return max(max(bnobt_sz, inobt_sz), max(rmapbt_sz, refcbt_sz));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) /* Allocate a block in an AG. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) xrep_alloc_ag_block(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) struct xfs_scrub *sc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) const struct xfs_owner_info *oinfo,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) xfs_fsblock_t *fsbno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) enum xfs_ag_resv_type resv)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) struct xfs_alloc_arg args = {0};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) xfs_agblock_t bno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) switch (resv) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) case XFS_AG_RESV_AGFL:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) case XFS_AG_RESV_RMAPBT:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) error = xfs_alloc_get_freelist(sc->tp, sc->sa.agf_bp, &bno, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) if (bno == NULLAGBLOCK)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) return -ENOSPC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) xfs_extent_busy_reuse(sc->mp, sc->sa.agno, bno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) 1, false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) *fsbno = XFS_AGB_TO_FSB(sc->mp, sc->sa.agno, bno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) if (resv == XFS_AG_RESV_RMAPBT)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) xfs_ag_resv_rmapbt_alloc(sc->mp, sc->sa.agno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) args.tp = sc->tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) args.mp = sc->mp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) args.oinfo = *oinfo;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) args.fsbno = XFS_AGB_TO_FSB(args.mp, sc->sa.agno, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) args.minlen = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) args.maxlen = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) args.prod = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) args.type = XFS_ALLOCTYPE_THIS_AG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) args.resv = resv;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) error = xfs_alloc_vextent(&args);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) if (args.fsbno == NULLFSBLOCK)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) return -ENOSPC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) ASSERT(args.len == 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) *fsbno = args.fsbno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) /* Initialize a new AG btree root block with zero entries. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) xrep_init_btblock(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) struct xfs_scrub *sc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) xfs_fsblock_t fsb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) struct xfs_buf **bpp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) xfs_btnum_t btnum,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) const struct xfs_buf_ops *ops)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) struct xfs_trans *tp = sc->tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) struct xfs_mount *mp = sc->mp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) struct xfs_buf *bp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) trace_xrep_init_btblock(mp, XFS_FSB_TO_AGNO(mp, fsb),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) XFS_FSB_TO_AGBNO(mp, fsb), btnum);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) ASSERT(XFS_FSB_TO_AGNO(mp, fsb) == sc->sa.agno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) error = xfs_trans_get_buf(tp, mp->m_ddev_targp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) XFS_FSB_TO_DADDR(mp, fsb), XFS_FSB_TO_BB(mp, 1), 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) &bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) xfs_buf_zero(bp, 0, BBTOB(bp->b_length));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) xfs_btree_init_block(mp, bp, btnum, 0, 0, sc->sa.agno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) xfs_trans_buf_set_type(tp, bp, XFS_BLFT_BTREE_BUF);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) xfs_trans_log_buf(tp, bp, 0, BBTOB(bp->b_length) - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) bp->b_ops = ops;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) *bpp = bp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368) * Reconstructing per-AG Btrees
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) * When a space btree is corrupt, we don't bother trying to fix it. Instead,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371) * we scan secondary space metadata to derive the records that should be in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) * the damaged btree, initialize a fresh btree root, and insert the records.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) * Note that for rebuilding the rmapbt we scan all the primary data to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) * generate the new records.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376) * However, that leaves the matter of removing all the metadata describing the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377) * old broken structure. For primary metadata we use the rmap data to collect
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) * every extent with a matching rmap owner (bitmap); we then iterate all other
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379) * metadata structures with the same rmap owner to collect the extents that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380) * cannot be removed (sublist). We then subtract sublist from bitmap to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) * derive the blocks that were used by the old btree. These blocks can be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) * reaped.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) * For rmapbt reconstructions we must use different tactics for extent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385) * collection. First we iterate all primary metadata (this excludes the old
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) * rmapbt, obviously) to generate new rmap records. The gaps in the rmap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) * records are collected as bitmap. The bnobt records are collected as
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388) * sublist. As with the other btrees we subtract sublist from bitmap, and the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) * result (since the rmapbt lives in the free space) are the blocks from the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390) * old rmapbt.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) * Disposal of Blocks from Old per-AG Btrees
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) * Now that we've constructed a new btree to replace the damaged one, we want
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) * to dispose of the blocks that (we think) the old btree was using.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) * Previously, we used the rmapbt to collect the extents (bitmap) with the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397) * rmap owner corresponding to the tree we rebuilt, collected extents for any
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) * blocks with the same rmap owner that are owned by another data structure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) * (sublist), and subtracted sublist from bitmap. In theory the extents
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400) * remaining in bitmap are the old btree's blocks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402) * Unfortunately, it's possible that the btree was crosslinked with other
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) * blocks on disk. The rmap data can tell us if there are multiple owners, so
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404) * if the rmapbt says there is an owner of this block other than @oinfo, then
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) * the block is crosslinked. Remove the reverse mapping and continue.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407) * If there is one rmap record, we can free the block, which removes the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408) * reverse mapping but doesn't add the block to the free space. Our repair
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) * strategy is to hope the other metadata objects crosslinked on this block
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) * will be rebuilt (atop different blocks), thereby removing all the cross
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) * links.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413) * If there are no rmap records at all, we also free the block. If the btree
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) * being rebuilt lives in the free space (bnobt/cntbt/rmapbt) then there isn't
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) * supposed to be a rmap record and everything is ok. For other btrees there
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416) * had to have been an rmap entry for the block to have ended up on @bitmap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417) * so if it's gone now there's something wrong and the fs will shut down.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) * Note: If there are multiple rmap records with only the same rmap owner as
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420) * the btree we're trying to rebuild and the block is indeed owned by another
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421) * data structure with the same rmap owner, then the block will be in sublist
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422) * and therefore doesn't need disposal. If there are multiple rmap records
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) * with only the same rmap owner but the block is not owned by something with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424) * the same rmap owner, the block will be freed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426) * The caller is responsible for locking the AG headers for the entire rebuild
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) * operation so that nothing else can sneak in and change the AG state while
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) * we're not looking. We also assume that the caller already invalidated any
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429) * buffers associated with @bitmap.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) * Invalidate buffers for per-AG btree blocks we're dumping. This function
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) * is not intended for use with file data repairs; we have bunmapi for that.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) xrep_invalidate_blocks(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438) struct xfs_scrub *sc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439) struct xbitmap *bitmap)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441) struct xbitmap_range *bmr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442) struct xbitmap_range *n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) struct xfs_buf *bp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) xfs_fsblock_t fsbno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) * For each block in each extent, see if there's an incore buffer for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) * exactly that block; if so, invalidate it. The buffer cache only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449) * lets us look for one buffer at a time, so we have to look one block
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450) * at a time. Avoid invalidating AG headers and post-EOFS blocks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) * because we never own those; and if we can't TRYLOCK the buffer we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452) * assume it's owned by someone else.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454) for_each_xbitmap_block(fsbno, bmr, n, bitmap) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) /* Skip AG headers and post-EOFS blocks */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) if (!xfs_verify_fsbno(sc->mp, fsbno))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458) bp = xfs_buf_incore(sc->mp->m_ddev_targp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) XFS_FSB_TO_DADDR(sc->mp, fsbno),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) XFS_FSB_TO_BB(sc->mp, 1), XBF_TRYLOCK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461) if (bp) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) xfs_trans_bjoin(sc->tp, bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) xfs_trans_binval(sc->tp, bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) /* Ensure the freelist is the correct size. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472) xrep_fix_freelist(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473) struct xfs_scrub *sc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474) bool can_shrink)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476) struct xfs_alloc_arg args = {0};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478) args.mp = sc->mp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) args.tp = sc->tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480) args.agno = sc->sa.agno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481) args.alignment = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482) args.pag = sc->sa.pag;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484) return xfs_alloc_fix_freelist(&args,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) can_shrink ? 0 : XFS_ALLOC_FLAG_NOSHRINK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489) * Put a block back on the AGFL.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) xrep_put_freelist(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493) struct xfs_scrub *sc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494) xfs_agblock_t agbno)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498) /* Make sure there's space on the freelist. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) error = xrep_fix_freelist(sc, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) * Since we're "freeing" a lost block onto the AGFL, we have to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505) * create an rmap for the block prior to merging it or else other
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) * parts will break.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508) error = xfs_rmap_alloc(sc->tp, sc->sa.agf_bp, sc->sa.agno, agbno, 1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509) &XFS_RMAP_OINFO_AG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513) /* Put the block on the AGFL. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514) error = xfs_alloc_put_freelist(sc->tp, sc->sa.agf_bp, sc->sa.agfl_bp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) agbno, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518) xfs_extent_busy_insert(sc->tp, sc->sa.agno, agbno, 1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) XFS_EXTENT_BUSY_SKIP_DISCARD);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524) /* Dispose of a single block. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526) xrep_reap_block(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527) struct xfs_scrub *sc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) xfs_fsblock_t fsbno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529) const struct xfs_owner_info *oinfo,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530) enum xfs_ag_resv_type resv)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532) struct xfs_btree_cur *cur;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533) struct xfs_buf *agf_bp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534) xfs_agnumber_t agno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) xfs_agblock_t agbno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536) bool has_other_rmap;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) agno = XFS_FSB_TO_AGNO(sc->mp, fsbno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540) agbno = XFS_FSB_TO_AGBNO(sc->mp, fsbno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) * If we are repairing per-inode metadata, we need to read in the AGF
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544) * buffer. Otherwise, we're repairing a per-AG structure, so reuse
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545) * the AGF buffer that the setup functions already grabbed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547) if (sc->ip) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548) error = xfs_alloc_read_agf(sc->mp, sc->tp, agno, 0, &agf_bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552) agf_bp = sc->sa.agf_bp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) cur = xfs_rmapbt_init_cursor(sc->mp, sc->tp, agf_bp, agno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) /* Can we find any other rmappings? */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557) error = xfs_rmap_has_other_keys(cur, agbno, 1, oinfo, &has_other_rmap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558) xfs_btree_del_cursor(cur, error);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560) goto out_free;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563) * If there are other rmappings, this block is cross linked and must
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564) * not be freed. Remove the reverse mapping and move on. Otherwise,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565) * we were the only owner of the block, so free the extent, which will
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566) * also remove the rmap.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568) * XXX: XFS doesn't support detecting the case where a single block
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569) * metadata structure is crosslinked with a multi-block structure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) * because the buffer cache doesn't detect aliasing problems, so we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571) * can't fix 100% of crosslinking problems (yet). The verifiers will
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572) * blow on writeout, the filesystem will shut down, and the admin gets
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573) * to run xfs_repair.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575) if (has_other_rmap)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576) error = xfs_rmap_free(sc->tp, agf_bp, agno, agbno, 1, oinfo);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577) else if (resv == XFS_AG_RESV_AGFL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) error = xrep_put_freelist(sc, agbno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580) error = xfs_free_extent(sc->tp, fsbno, 1, oinfo, resv);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581) if (agf_bp != sc->sa.agf_bp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582) xfs_trans_brelse(sc->tp, agf_bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586) if (sc->ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587) return xfs_trans_roll_inode(&sc->tp, sc->ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) return xrep_roll_ag_trans(sc);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590) out_free:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591) if (agf_bp != sc->sa.agf_bp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592) xfs_trans_brelse(sc->tp, agf_bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 596) /* Dispose of every block of every extent in the bitmap. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 597) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 598) xrep_reap_extents(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599) struct xfs_scrub *sc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600) struct xbitmap *bitmap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) const struct xfs_owner_info *oinfo,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) enum xfs_ag_resv_type type)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604) struct xbitmap_range *bmr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605) struct xbitmap_range *n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606) xfs_fsblock_t fsbno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 608)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 609) ASSERT(xfs_sb_version_hasrmapbt(&sc->mp->m_sb));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 610)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 611) for_each_xbitmap_block(fsbno, bmr, n, bitmap) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 612) ASSERT(sc->ip != NULL ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 613) XFS_FSB_TO_AGNO(sc->mp, fsbno) == sc->sa.agno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 614) trace_xrep_dispose_btree_extent(sc->mp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 615) XFS_FSB_TO_AGNO(sc->mp, fsbno),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 616) XFS_FSB_TO_AGBNO(sc->mp, fsbno), 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 617)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 618) error = xrep_reap_block(sc, fsbno, oinfo, type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 619) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 620) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 621) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 622)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 623) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 624) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 625)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 626) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 627) * Finding per-AG Btree Roots for AGF/AGI Reconstruction
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 628) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 629) * If the AGF or AGI become slightly corrupted, it may be necessary to rebuild
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 630) * the AG headers by using the rmap data to rummage through the AG looking for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 631) * btree roots. This is not guaranteed to work if the AG is heavily damaged
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 632) * or the rmap data are corrupt.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 633) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 634) * Callers of xrep_find_ag_btree_roots must lock the AGF and AGFL
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 635) * buffers if the AGF is being rebuilt; or the AGF and AGI buffers if the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 636) * AGI is being rebuilt. It must maintain these locks until it's safe for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 637) * other threads to change the btrees' shapes. The caller provides
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 638) * information about the btrees to look for by passing in an array of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 639) * xrep_find_ag_btree with the (rmap owner, buf_ops, magic) fields set.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 640) * The (root, height) fields will be set on return if anything is found. The
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 641) * last element of the array should have a NULL buf_ops to mark the end of the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 642) * array.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 643) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 644) * For every rmapbt record matching any of the rmap owners in btree_info,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 645) * read each block referenced by the rmap record. If the block is a btree
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 646) * block from this filesystem matching any of the magic numbers and has a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 647) * level higher than what we've already seen, remember the block and the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 648) * height of the tree required to have such a block. When the call completes,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 649) * we return the highest block we've found for each btree description; those
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 650) * should be the roots.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 651) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 652)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 653) struct xrep_findroot {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 654) struct xfs_scrub *sc;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 655) struct xfs_buf *agfl_bp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 656) struct xfs_agf *agf;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 657) struct xrep_find_ag_btree *btree_info;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 658) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 659)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 660) /* See if our block is in the AGFL. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 661) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 662) xrep_findroot_agfl_walk(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 663) struct xfs_mount *mp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 664) xfs_agblock_t bno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 665) void *priv)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 666) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 667) xfs_agblock_t *agbno = priv;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 668)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 669) return (*agbno == bno) ? -ECANCELED : 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 670) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 671)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 672) /* Does this block match the btree information passed in? */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 673) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 674) xrep_findroot_block(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 675) struct xrep_findroot *ri,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 676) struct xrep_find_ag_btree *fab,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 677) uint64_t owner,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 678) xfs_agblock_t agbno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 679) bool *done_with_block)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 680) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 681) struct xfs_mount *mp = ri->sc->mp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 682) struct xfs_buf *bp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 683) struct xfs_btree_block *btblock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 684) xfs_daddr_t daddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 685) int block_level;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 686) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 687)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 688) daddr = XFS_AGB_TO_DADDR(mp, ri->sc->sa.agno, agbno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 689)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 690) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 691) * Blocks in the AGFL have stale contents that might just happen to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 692) * have a matching magic and uuid. We don't want to pull these blocks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 693) * in as part of a tree root, so we have to filter out the AGFL stuff
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 694) * here. If the AGFL looks insane we'll just refuse to repair.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 695) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 696) if (owner == XFS_RMAP_OWN_AG) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 697) error = xfs_agfl_walk(mp, ri->agf, ri->agfl_bp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 698) xrep_findroot_agfl_walk, &agbno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 699) if (error == -ECANCELED)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 700) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 701) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 702) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 703) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 704)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 705) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 706) * Read the buffer into memory so that we can see if it's a match for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 707) * our btree type. We have no clue if it is beforehand, and we want to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 708) * avoid xfs_trans_read_buf's behavior of dumping the DONE state (which
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 709) * will cause needless disk reads in subsequent calls to this function)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 710) * and logging metadata verifier failures.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 711) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 712) * Therefore, pass in NULL buffer ops. If the buffer was already in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 713) * memory from some other caller it will already have b_ops assigned.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 714) * If it was in memory from a previous unsuccessful findroot_block
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 715) * call, the buffer won't have b_ops but it should be clean and ready
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 716) * for us to try to verify if the read call succeeds. The same applies
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 717) * if the buffer wasn't in memory at all.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 718) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 719) * Note: If we never match a btree type with this buffer, it will be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 720) * left in memory with NULL b_ops. This shouldn't be a problem unless
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 721) * the buffer gets written.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 722) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 723) error = xfs_trans_read_buf(mp, ri->sc->tp, mp->m_ddev_targp, daddr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 724) mp->m_bsize, 0, &bp, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 725) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 726) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 727)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 728) /* Ensure the block magic matches the btree type we're looking for. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 729) btblock = XFS_BUF_TO_BLOCK(bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 730) ASSERT(fab->buf_ops->magic[1] != 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 731) if (btblock->bb_magic != fab->buf_ops->magic[1])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 732) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 733)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 734) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 735) * If the buffer already has ops applied and they're not the ones for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 736) * this btree type, we know this block doesn't match the btree and we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 737) * can bail out.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 738) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 739) * If the buffer ops match ours, someone else has already validated
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 740) * the block for us, so we can move on to checking if this is a root
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 741) * block candidate.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 742) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 743) * If the buffer does not have ops, nobody has successfully validated
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 744) * the contents and the buffer cannot be dirty. If the magic, uuid,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 745) * and structure match this btree type then we'll move on to checking
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 746) * if it's a root block candidate. If there is no match, bail out.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 747) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 748) if (bp->b_ops) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 749) if (bp->b_ops != fab->buf_ops)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 750) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 751) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 752) ASSERT(!xfs_trans_buf_is_dirty(bp));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 753) if (!uuid_equal(&btblock->bb_u.s.bb_uuid,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 754) &mp->m_sb.sb_meta_uuid))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 755) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 756) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 757) * Read verifiers can reference b_ops, so we set the pointer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 758) * here. If the verifier fails we'll reset the buffer state
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 759) * to what it was before we touched the buffer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 760) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 761) bp->b_ops = fab->buf_ops;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 762) fab->buf_ops->verify_read(bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 763) if (bp->b_error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 764) bp->b_ops = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 765) bp->b_error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 766) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 767) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 768)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 769) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 770) * Some read verifiers will (re)set b_ops, so we must be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 771) * careful not to change b_ops after running the verifier.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 772) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 773) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 774)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 775) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 776) * This block passes the magic/uuid and verifier tests for this btree
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 777) * type. We don't need the caller to try the other tree types.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 778) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 779) *done_with_block = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 780)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 781) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 782) * Compare this btree block's level to the height of the current
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 783) * candidate root block.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 784) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 785) * If the level matches the root we found previously, throw away both
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 786) * blocks because there can't be two candidate roots.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 787) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 788) * If level is lower in the tree than the root we found previously,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 789) * ignore this block.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 790) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 791) block_level = xfs_btree_get_level(btblock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 792) if (block_level + 1 == fab->height) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 793) fab->root = NULLAGBLOCK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 794) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 795) } else if (block_level < fab->height) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 796) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 797) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 798)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 799) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 800) * This is the highest block in the tree that we've found so far.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 801) * Update the btree height to reflect what we've learned from this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 802) * block.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 803) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 804) fab->height = block_level + 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 805)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 806) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 807) * If this block doesn't have sibling pointers, then it's the new root
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 808) * block candidate. Otherwise, the root will be found farther up the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 809) * tree.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 810) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 811) if (btblock->bb_u.s.bb_leftsib == cpu_to_be32(NULLAGBLOCK) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 812) btblock->bb_u.s.bb_rightsib == cpu_to_be32(NULLAGBLOCK))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 813) fab->root = agbno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 814) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 815) fab->root = NULLAGBLOCK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 816)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 817) trace_xrep_findroot_block(mp, ri->sc->sa.agno, agbno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 818) be32_to_cpu(btblock->bb_magic), fab->height - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 819) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 820) xfs_trans_brelse(ri->sc->tp, bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 821) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 822) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 823)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 824) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 825) * Do any of the blocks in this rmap record match one of the btrees we're
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 826) * looking for?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 827) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 828) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 829) xrep_findroot_rmap(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 830) struct xfs_btree_cur *cur,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 831) struct xfs_rmap_irec *rec,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 832) void *priv)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 833) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 834) struct xrep_findroot *ri = priv;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 835) struct xrep_find_ag_btree *fab;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 836) xfs_agblock_t b;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 837) bool done;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 838) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 839)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 840) /* Ignore anything that isn't AG metadata. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 841) if (!XFS_RMAP_NON_INODE_OWNER(rec->rm_owner))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 842) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 843)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 844) /* Otherwise scan each block + btree type. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 845) for (b = 0; b < rec->rm_blockcount; b++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 846) done = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 847) for (fab = ri->btree_info; fab->buf_ops; fab++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 848) if (rec->rm_owner != fab->rmap_owner)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 849) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 850) error = xrep_findroot_block(ri, fab,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 851) rec->rm_owner, rec->rm_startblock + b,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 852) &done);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 853) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 854) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 855) if (done)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 856) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 857) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 858) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 859)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 860) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 861) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 862)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 863) /* Find the roots of the per-AG btrees described in btree_info. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 864) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 865) xrep_find_ag_btree_roots(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 866) struct xfs_scrub *sc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 867) struct xfs_buf *agf_bp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 868) struct xrep_find_ag_btree *btree_info,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 869) struct xfs_buf *agfl_bp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 870) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 871) struct xfs_mount *mp = sc->mp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 872) struct xrep_findroot ri;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 873) struct xrep_find_ag_btree *fab;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 874) struct xfs_btree_cur *cur;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 875) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 876)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 877) ASSERT(xfs_buf_islocked(agf_bp));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 878) ASSERT(agfl_bp == NULL || xfs_buf_islocked(agfl_bp));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 879)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 880) ri.sc = sc;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 881) ri.btree_info = btree_info;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 882) ri.agf = agf_bp->b_addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 883) ri.agfl_bp = agfl_bp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 884) for (fab = btree_info; fab->buf_ops; fab++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 885) ASSERT(agfl_bp || fab->rmap_owner != XFS_RMAP_OWN_AG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 886) ASSERT(XFS_RMAP_NON_INODE_OWNER(fab->rmap_owner));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 887) fab->root = NULLAGBLOCK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 888) fab->height = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 889) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 890)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 891) cur = xfs_rmapbt_init_cursor(mp, sc->tp, agf_bp, sc->sa.agno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 892) error = xfs_rmap_query_all(cur, xrep_findroot_rmap, &ri);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 893) xfs_btree_del_cursor(cur, error);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 894)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 895) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 896) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 897)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 898) /* Force a quotacheck the next time we mount. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 899) void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 900) xrep_force_quotacheck(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 901) struct xfs_scrub *sc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 902) xfs_dqtype_t type)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 903) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 904) uint flag;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 905)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 906) flag = xfs_quota_chkd_flag(type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 907) if (!(flag & sc->mp->m_qflags))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 908) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 909)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 910) sc->mp->m_qflags &= ~flag;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 911) spin_lock(&sc->mp->m_sb_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 912) sc->mp->m_sb.sb_qflags &= ~flag;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 913) spin_unlock(&sc->mp->m_sb_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 914) xfs_log_sb(sc->tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 915) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 916)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 917) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 918) * Attach dquots to this inode, or schedule quotacheck to fix them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 919) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 920) * This function ensures that the appropriate dquots are attached to an inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 921) * We cannot allow the dquot code to allocate an on-disk dquot block here
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 922) * because we're already in transaction context with the inode locked. The
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 923) * on-disk dquot should already exist anyway. If the quota code signals
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 924) * corruption or missing quota information, schedule quotacheck, which will
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 925) * repair corruptions in the quota metadata.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 926) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 927) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 928) xrep_ino_dqattach(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 929) struct xfs_scrub *sc)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 930) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 931) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 932)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 933) error = xfs_qm_dqattach_locked(sc->ip, false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 934) switch (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 935) case -EFSBADCRC:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 936) case -EFSCORRUPTED:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 937) case -ENOENT:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 938) xfs_err_ratelimited(sc->mp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 939) "inode %llu repair encountered quota error %d, quotacheck forced.",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 940) (unsigned long long)sc->ip->i_ino, error);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 941) if (XFS_IS_UQUOTA_ON(sc->mp) && !sc->ip->i_udquot)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 942) xrep_force_quotacheck(sc, XFS_DQTYPE_USER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 943) if (XFS_IS_GQUOTA_ON(sc->mp) && !sc->ip->i_gdquot)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 944) xrep_force_quotacheck(sc, XFS_DQTYPE_GROUP);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 945) if (XFS_IS_PQUOTA_ON(sc->mp) && !sc->ip->i_pdquot)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 946) xrep_force_quotacheck(sc, XFS_DQTYPE_PROJ);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 947) /* fall through */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 948) case -ESRCH:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 949) error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 950) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 951) default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 952) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 953) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 954)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 955) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 956) }