diff options
Diffstat (limited to 'fs/xfs/xfs_mount.c')
| -rw-r--r-- | fs/xfs/xfs_mount.c | 1496 |
1 files changed, 869 insertions, 627 deletions
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index b4d8c318be3c..0953f6ae94ab 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -12,9 +12,6 @@ #include "xfs_bit.h" #include "xfs_sb.h" #include "xfs_mount.h" -#include "xfs_defer.h" -#include "xfs_da_format.h" -#include "xfs_da_btree.h" #include "xfs_inode.h" #include "xfs_dir2.h" #include "xfs_ialloc.h" @@ -24,17 +21,26 @@ #include "xfs_trans.h" #include "xfs_trans_priv.h" #include "xfs_log.h" +#include "xfs_log_priv.h" #include "xfs_error.h" #include "xfs_quota.h" #include "xfs_fsops.h" -#include "xfs_trace.h" #include "xfs_icache.h" #include "xfs_sysfs.h" #include "xfs_rmap_btree.h" #include "xfs_refcount_btree.h" #include "xfs_reflink.h" #include "xfs_extent_busy.h" - +#include "xfs_health.h" +#include "xfs_trace.h" +#include "xfs_ag.h" +#include "xfs_rtbitmap.h" +#include "xfs_metafile.h" +#include "xfs_rtgroup.h" +#include "xfs_rtrmap_btree.h" +#include "xfs_rtrefcount_btree.h" +#include "scrub/stats.h" +#include "xfs_zone_alloc.h" static DEFINE_MUTEX(xfs_uuid_table_mutex); static int xfs_uuid_table_size; @@ -45,7 +51,7 @@ xfs_uuid_table_free(void) { if (xfs_uuid_table_size == 0) return; - kmem_free(xfs_uuid_table); + kfree(xfs_uuid_table); xfs_uuid_table = NULL; xfs_uuid_table_size = 0; } @@ -62,9 +68,9 @@ xfs_uuid_mount( int hole, i; /* Publish UUID in struct super_block */ - uuid_copy(&mp->m_super->s_uuid, uuid); + super_set_uuid(mp->m_super, uuid->b, sizeof(*uuid)); - if (mp->m_flags & XFS_MOUNT_NOUUID) + if (xfs_has_nouuid(mp)) return 0; if (uuid_is_null(uuid)) { @@ -83,9 +89,9 @@ xfs_uuid_mount( } if (hole < 0) { - xfs_uuid_table = kmem_realloc(xfs_uuid_table, + xfs_uuid_table = krealloc(xfs_uuid_table, (xfs_uuid_table_size + 1) * sizeof(*xfs_uuid_table), - KM_SLEEP); + GFP_KERNEL | __GFP_NOFAIL); hole = xfs_uuid_table_size++; } xfs_uuid_table[hole] = *uuid; @@ -106,7 +112,7 @@ xfs_uuid_unmount( uuid_t *uuid = &mp->m_sb.sb_uuid; int i; - if (mp->m_flags & XFS_MOUNT_NOUUID) + if (xfs_has_nouuid(mp)) return; mutex_lock(&xfs_uuid_table_mutex); @@ -122,39 +128,6 @@ xfs_uuid_unmount( mutex_unlock(&xfs_uuid_table_mutex); } - -STATIC void -__xfs_free_perag( - struct rcu_head *head) -{ - struct xfs_perag *pag = container_of(head, struct xfs_perag, rcu_head); - - ASSERT(atomic_read(&pag->pag_ref) == 0); - kmem_free(pag); -} - -/* - * Free up the per-ag resources associated with the mount structure. - */ -STATIC void -xfs_free_perag( - xfs_mount_t *mp) -{ - xfs_agnumber_t agno; - struct xfs_perag *pag; - - for (agno = 0; agno < mp->m_sb.sb_agcount; agno++) { - spin_lock(&mp->m_perag_lock); - pag = radix_tree_delete(&mp->m_perag_tree, agno); - spin_unlock(&mp->m_perag_lock); - ASSERT(pag); - ASSERT(atomic_read(&pag->pag_ref) == 0); - xfs_buf_hash_destroy(pag); - mutex_destroy(&pag->pag_ici_reclaim_lock); - call_rcu(&pag->rcu_head, __xfs_free_perag); - } -} - /* * Check size of device based on the (data/realtime) block count. * Note: this check is used by the growfs code as well as mount. @@ -164,95 +137,17 @@ xfs_sb_validate_fsb_count( xfs_sb_t *sbp, uint64_t nblocks) { - ASSERT(PAGE_SHIFT >= sbp->sb_blocklog); + uint64_t max_bytes; + ASSERT(sbp->sb_blocklog >= BBSHIFT); - /* Limited by ULONG_MAX of page cache index */ - if (nblocks >> (PAGE_SHIFT - sbp->sb_blocklog) > ULONG_MAX) + if (check_shl_overflow(nblocks, sbp->sb_blocklog, &max_bytes)) return -EFBIG; - return 0; -} - -int -xfs_initialize_perag( - xfs_mount_t *mp, - xfs_agnumber_t agcount, - xfs_agnumber_t *maxagi) -{ - xfs_agnumber_t index; - xfs_agnumber_t first_initialised = NULLAGNUMBER; - xfs_perag_t *pag; - int error = -ENOMEM; - - /* - * Walk the current per-ag tree so we don't try to initialise AGs - * that already exist (growfs case). Allocate and insert all the - * AGs we don't find ready for initialisation. - */ - for (index = 0; index < agcount; index++) { - pag = xfs_perag_get(mp, index); - if (pag) { - xfs_perag_put(pag); - continue; - } - - pag = kmem_zalloc(sizeof(*pag), KM_MAYFAIL); - if (!pag) - goto out_unwind_new_pags; - pag->pag_agno = index; - pag->pag_mount = mp; - spin_lock_init(&pag->pag_ici_lock); - mutex_init(&pag->pag_ici_reclaim_lock); - INIT_RADIX_TREE(&pag->pag_ici_root, GFP_ATOMIC); - if (xfs_buf_hash_init(pag)) - goto out_free_pag; - init_waitqueue_head(&pag->pagb_wait); - spin_lock_init(&pag->pagb_lock); - pag->pagb_count = 0; - pag->pagb_tree = RB_ROOT; - - if (radix_tree_preload(GFP_NOFS)) - goto out_hash_destroy; - - spin_lock(&mp->m_perag_lock); - if (radix_tree_insert(&mp->m_perag_tree, index, pag)) { - BUG(); - spin_unlock(&mp->m_perag_lock); - radix_tree_preload_end(); - error = -EEXIST; - goto out_hash_destroy; - } - spin_unlock(&mp->m_perag_lock); - radix_tree_preload_end(); - /* first new pag is fully initialized */ - if (first_initialised == NULLAGNUMBER) - first_initialised = index; - } - - index = xfs_set_inode_alloc(mp, agcount); - if (maxagi) - *maxagi = index; - - mp->m_ag_prealloc_blocks = xfs_prealloc_blocks(mp); + /* Limited by ULONG_MAX of page cache index */ + if (max_bytes >> PAGE_SHIFT > ULONG_MAX) + return -EFBIG; return 0; - -out_hash_destroy: - xfs_buf_hash_destroy(pag); -out_free_pag: - mutex_destroy(&pag->pag_ici_reclaim_lock); - kmem_free(pag); -out_unwind_new_pags: - /* unwind any prior newly initialized pags */ - for (index = first_initialised; index < agcount; index++) { - pag = radix_tree_delete(&mp->m_perag_tree, index); - if (!pag) - break; - xfs_buf_hash_destroy(pag); - mutex_destroy(&pag->pag_ici_reclaim_lock); - kmem_free(pag); - } - return error; } /* @@ -276,25 +171,19 @@ xfs_readsb( ASSERT(mp->m_ddev_targp != NULL); /* - * For the initial read, we must guess at the sector - * size based on the block device. It's enough to - * get the sb_sectsize out of the superblock and - * then reread with the proper length. - * We don't verify it yet, because it may not be complete. + * In the first pass, use the device sector size to just read enough + * of the superblock to extract the XFS sector size. + * + * The device sector size must be smaller than or equal to the XFS + * sector size and thus we can always read the superblock. Once we know + * the XFS sector size, re-read it and run the buffer verifier. */ - sector_size = xfs_getsize_buftarg(mp->m_ddev_targp); + sector_size = mp->m_ddev_targp->bt_logical_sectorsize; buf_ops = NULL; - /* - * Allocate a (locked) buffer to hold the superblock. This will be kept - * around at all times to optimize access to the superblock. Therefore, - * set XBF_NO_IOACCT to make sure it doesn't hold the buftarg count - * elevated. - */ reread: error = xfs_buf_read_uncached(mp->m_ddev_targp, XFS_SB_DADDR, - BTOBB(sector_size), XBF_NO_IOACCT, &bp, - buf_ops); + BTOBB(sector_size), &bp, buf_ops); if (error) { if (loud) xfs_warn(mp, "SB validate failed with error %d.", error); @@ -307,7 +196,7 @@ reread: /* * Initialize the mount structure from the superblock. */ - xfs_sb_from_disk(sbp, XFS_BUF_TO_SBP(bp)); + xfs_sb_from_disk(sbp, bp->b_addr); /* * If we haven't validated the superblock, do so now before we try @@ -342,11 +231,23 @@ reread: goto reread; } + mp->m_features |= xfs_sb_version_to_features(sbp); xfs_reinit_percpu_counters(mp); + /* + * If logged xattrs are enabled after log recovery finishes, then set + * the opstate so that log recovery will work properly. + */ + if (xfs_sb_version_haslogxattrs(&mp->m_sb)) + xfs_set_using_logged_xattrs(mp); + /* no need to be quiet anymore, so reset the buf ops */ bp->b_ops = &xfs_sb_buf_ops; + /* + * Keep a pointer of the sb buffer around instead of caching it in the + * buffer cache because we access it frequently. + */ mp->m_sb_bp = bp; xfs_buf_unlock(bp); return 0; @@ -357,132 +258,123 @@ release_buf: } /* - * Update alignment values based on mount options and sb values + * If the sunit/swidth change would move the precomputed root inode value, we + * must reject the ondisk change because repair will stumble over that. + * However, we allow the mount to proceed because we never rejected this + * combination before. Returns true to update the sb, false otherwise. */ -STATIC int -xfs_update_alignment(xfs_mount_t *mp) +static inline int +xfs_check_new_dalign( + struct xfs_mount *mp, + int new_dalign, + bool *update_sb) { - xfs_sb_t *sbp = &(mp->m_sb); + struct xfs_sb *sbp = &mp->m_sb; + xfs_ino_t calc_ino; - if (mp->m_dalign) { - /* - * If stripe unit and stripe width are not multiples - * of the fs blocksize turn off alignment. - */ - if ((BBTOB(mp->m_dalign) & mp->m_blockmask) || - (BBTOB(mp->m_swidth) & mp->m_blockmask)) { - xfs_warn(mp, - "alignment check failed: sunit/swidth vs. blocksize(%d)", - sbp->sb_blocksize); - return -EINVAL; - } else { - /* - * Convert the stripe unit and width to FSBs. - */ - mp->m_dalign = XFS_BB_TO_FSBT(mp, mp->m_dalign); - if (mp->m_dalign && (sbp->sb_agblocks % mp->m_dalign)) { - xfs_warn(mp, - "alignment check failed: sunit/swidth vs. agsize(%d)", - sbp->sb_agblocks); - return -EINVAL; - } else if (mp->m_dalign) { - mp->m_swidth = XFS_BB_TO_FSBT(mp, mp->m_swidth); - } else { - xfs_warn(mp, - "alignment check failed: sunit(%d) less than bsize(%d)", - mp->m_dalign, sbp->sb_blocksize); - return -EINVAL; - } - } + calc_ino = xfs_ialloc_calc_rootino(mp, new_dalign); + trace_xfs_check_new_dalign(mp, new_dalign, calc_ino); - /* - * Update superblock with new values - * and log changes - */ - if (xfs_sb_version_hasdalign(sbp)) { - if (sbp->sb_unit != mp->m_dalign) { - sbp->sb_unit = mp->m_dalign; - mp->m_update_sb = true; - } - if (sbp->sb_width != mp->m_swidth) { - sbp->sb_width = mp->m_swidth; - mp->m_update_sb = true; - } - } else { - xfs_warn(mp, - "cannot change alignment: superblock does not support data alignment"); - return -EINVAL; - } - } else if ((mp->m_flags & XFS_MOUNT_NOALIGN) != XFS_MOUNT_NOALIGN && - xfs_sb_version_hasdalign(&mp->m_sb)) { - mp->m_dalign = sbp->sb_unit; - mp->m_swidth = sbp->sb_width; + if (sbp->sb_rootino == calc_ino) { + *update_sb = true; + return 0; } + xfs_warn(mp, +"Cannot change stripe alignment; would require moving root inode."); + + /* + * XXX: Next time we add a new incompat feature, this should start + * returning -EINVAL to fail the mount. Until then, spit out a warning + * that we're ignoring the administrator's instructions. + */ + xfs_warn(mp, "Skipping superblock stripe alignment update."); + *update_sb = false; return 0; } /* - * Set the maximum inode count for this filesystem + * If we were provided with new sunit/swidth values as mount options, make sure + * that they pass basic alignment and superblock feature checks, and convert + * them into the same units (FSB) that everything else expects. This step + * /must/ be done before computing the inode geometry. */ -STATIC void -xfs_set_maxicount(xfs_mount_t *mp) +STATIC int +xfs_validate_new_dalign( + struct xfs_mount *mp) { - xfs_sb_t *sbp = &(mp->m_sb); - uint64_t icount; + if (mp->m_dalign == 0) + return 0; - if (sbp->sb_imax_pct) { - /* - * Make sure the maximum inode count is a multiple - * of the units we allocate inodes in. - */ - icount = sbp->sb_dblocks * sbp->sb_imax_pct; - do_div(icount, 100); - do_div(icount, mp->m_ialloc_blks); - mp->m_maxicount = (icount * mp->m_ialloc_blks) << - sbp->sb_inopblog; - } else { - mp->m_maxicount = 0; + /* + * If stripe unit and stripe width are not multiples + * of the fs blocksize turn off alignment. + */ + if ((BBTOB(mp->m_dalign) & mp->m_blockmask) || + (BBTOB(mp->m_swidth) & mp->m_blockmask)) { + xfs_warn(mp, + "alignment check failed: sunit/swidth vs. blocksize(%d)", + mp->m_sb.sb_blocksize); + return -EINVAL; } -} -/* - * Set the default minimum read and write sizes unless - * already specified in a mount option. - * We use smaller I/O sizes when the file system - * is being used for NFS service (wsync mount option). - */ -STATIC void -xfs_set_rw_sizes(xfs_mount_t *mp) -{ - xfs_sb_t *sbp = &(mp->m_sb); - int readio_log, writeio_log; + /* + * Convert the stripe unit and width to FSBs. + */ + mp->m_dalign = XFS_BB_TO_FSBT(mp, mp->m_dalign); + if (mp->m_dalign && (mp->m_sb.sb_agblocks % mp->m_dalign)) { + xfs_warn(mp, + "alignment check failed: sunit/swidth vs. agsize(%d)", + mp->m_sb.sb_agblocks); + return -EINVAL; + } - if (!(mp->m_flags & XFS_MOUNT_DFLT_IOSIZE)) { - if (mp->m_flags & XFS_MOUNT_WSYNC) { - readio_log = XFS_WSYNC_READIO_LOG; - writeio_log = XFS_WSYNC_WRITEIO_LOG; - } else { - readio_log = XFS_READIO_LOG_LARGE; - writeio_log = XFS_WRITEIO_LOG_LARGE; - } - } else { - readio_log = mp->m_readio_log; - writeio_log = mp->m_writeio_log; + if (!mp->m_dalign) { + xfs_warn(mp, + "alignment check failed: sunit(%d) less than bsize(%d)", + mp->m_dalign, mp->m_sb.sb_blocksize); + return -EINVAL; } - if (sbp->sb_blocklog > readio_log) { - mp->m_readio_log = sbp->sb_blocklog; - } else { - mp->m_readio_log = readio_log; + mp->m_swidth = XFS_BB_TO_FSBT(mp, mp->m_swidth); + + if (!xfs_has_dalign(mp)) { + xfs_warn(mp, +"cannot change alignment: superblock does not support data alignment"); + return -EINVAL; } - mp->m_readio_blocks = 1 << (mp->m_readio_log - sbp->sb_blocklog); - if (sbp->sb_blocklog > writeio_log) { - mp->m_writeio_log = sbp->sb_blocklog; - } else { - mp->m_writeio_log = writeio_log; + + return 0; +} + +/* Update alignment values based on mount options and sb values. */ +STATIC int +xfs_update_alignment( + struct xfs_mount *mp) +{ + struct xfs_sb *sbp = &mp->m_sb; + + if (mp->m_dalign) { + bool update_sb; + int error; + + if (sbp->sb_unit == mp->m_dalign && + sbp->sb_width == mp->m_swidth) + return 0; + + error = xfs_check_new_dalign(mp, mp->m_dalign, &update_sb); + if (error || !update_sb) + return error; + + sbp->sb_unit = mp->m_dalign; + sbp->sb_width = mp->m_swidth; + mp->m_update_sb = true; + } else if (!xfs_has_noalign(mp) && xfs_has_dalign(mp)) { + mp->m_dalign = sbp->sb_unit; + mp->m_swidth = sbp->sb_width; } - mp->m_writeio_blocks = 1 << (mp->m_writeio_log - sbp->sb_blocklog); + + return 0; } /* @@ -492,39 +384,19 @@ void xfs_set_low_space_thresholds( struct xfs_mount *mp) { - int i; + uint64_t dblocks = mp->m_sb.sb_dblocks; + uint64_t rtexts = mp->m_sb.sb_rextents; + int i; - for (i = 0; i < XFS_LOWSP_MAX; i++) { - uint64_t space = mp->m_sb.sb_dblocks; + do_div(dblocks, 100); + do_div(rtexts, 100); - do_div(space, 100); - mp->m_low_space[i] = space * (i + 1); + for (i = 0; i < XFS_LOWSP_MAX; i++) { + mp->m_low_space[i] = dblocks * (i + 1); + mp->m_low_rtexts[i] = rtexts * (i + 1); } } - -/* - * Set whether we're using inode alignment. - */ -STATIC void -xfs_set_inoalignment(xfs_mount_t *mp) -{ - if (xfs_sb_version_hasalign(&mp->m_sb) && - mp->m_sb.sb_inoalignmt >= xfs_icluster_size_fsb(mp)) - mp->m_inoalign_mask = mp->m_sb.sb_inoalignmt - 1; - else - mp->m_inoalign_mask = 0; - /* - * If we are using stripe alignment, check whether - * the stripe unit is a multiple of the inode alignment - */ - if (mp->m_dalign && mp->m_inoalign_mask && - !(mp->m_dalign & mp->m_inoalign_mask)) - mp->m_sinoalign = mp->m_dalign; - else - mp->m_sinoalign = 0; -} - /* * Check that the data (and log if separate) is an ok size. */ @@ -543,7 +415,7 @@ xfs_check_sizes( } error = xfs_buf_read_uncached(mp->m_ddev_targp, d - XFS_FSS_TO_BB(mp, 1), - XFS_FSS_TO_BB(mp, 1), 0, &bp, NULL); + XFS_FSS_TO_BB(mp, 1), &bp, NULL); if (error) { xfs_warn(mp, "last sector read failed"); return error; @@ -560,7 +432,7 @@ xfs_check_sizes( } error = xfs_buf_read_uncached(mp->m_logdev_targp, d - XFS_FSB_TO_BB(mp, 1), - XFS_FSB_TO_BB(mp, 1), 0, &bp, NULL); + XFS_FSB_TO_BB(mp, 1), &bp, NULL); if (error) { xfs_warn(mp, "log device read failed"); return error; @@ -591,22 +463,38 @@ xfs_mount_reset_sbqflags( return xfs_sync_sb(mp, false); } +static const char *const xfs_free_pool_name[] = { + [XC_FREE_BLOCKS] = "free blocks", + [XC_FREE_RTEXTENTS] = "free rt extents", + [XC_FREE_RTAVAILABLE] = "available rt extents", +}; + uint64_t -xfs_default_resblks(xfs_mount_t *mp) +xfs_default_resblks( + struct xfs_mount *mp, + enum xfs_free_counter ctr) { - uint64_t resblks; - - /* - * We default to 5% or 8192 fsbs of space reserved, whichever is - * smaller. This is intended to cover concurrent allocation - * transactions when we initially hit enospc. These each require a 4 - * block reservation. Hence by default we cover roughly 2000 concurrent - * allocation reservations. - */ - resblks = mp->m_sb.sb_dblocks; - do_div(resblks, 20); - resblks = min_t(uint64_t, resblks, 8192); - return resblks; + switch (ctr) { + case XC_FREE_BLOCKS: + /* + * Default to 5% or 8192 FSBs of space reserved, whichever is + * smaller. + * + * This is intended to cover concurrent allocation transactions + * when we initially hit ENOSPC. These each require a 4 block + * reservation. Hence by default we cover roughly 2000 + * concurrent allocation reservations. + */ + return min(div_u64(mp->m_sb.sb_dblocks, 20), 8192ULL); + case XC_FREE_RTEXTENTS: + case XC_FREE_RTAVAILABLE: + if (IS_ENABLED(CONFIG_XFS_RT) && xfs_has_zoned(mp)) + return xfs_zoned_default_resblks(mp, ctr); + return 0; + default: + ASSERT(0); + return 0; + } } /* Ensure the summary counts are correct. */ @@ -614,6 +502,8 @@ STATIC int xfs_check_summary_counts( struct xfs_mount *mp) { + int error = 0; + /* * The AG0 superblock verifier rejects in-progress filesystems, * so we should never see the flag set this far into mounting. @@ -635,11 +525,11 @@ xfs_check_summary_counts( * counters. If any of them are obviously incorrect, we can recompute * them from the AGF headers in the next step. */ - if (XFS_LAST_UNMOUNT_WAS_CLEAN(mp) && + if (xfs_is_clean(mp) && (mp->m_sb.sb_fdblocks > mp->m_sb.sb_dblocks || !xfs_verify_icount(mp, mp->m_sb.sb_icount) || mp->m_sb.sb_ifree > mp->m_sb.sb_icount)) - mp->m_flags |= XFS_MOUNT_BAD_SUMMARY; + xfs_fs_mark_sick(mp, XFS_SICK_FS_COUNTERS); /* * We can safely re-initialise incore superblock counters from the @@ -652,12 +542,284 @@ xfs_check_summary_counts( * superblock to be correct and we don't need to do anything here. * Otherwise, recalculate the summary counters. */ - if ((!xfs_sb_version_haslazysbcount(&mp->m_sb) || - XFS_LAST_UNMOUNT_WAS_CLEAN(mp)) && - !(mp->m_flags & XFS_MOUNT_BAD_SUMMARY)) + if ((xfs_has_lazysbcount(mp) && !xfs_is_clean(mp)) || + xfs_fs_has_sickness(mp, XFS_SICK_FS_COUNTERS)) { + error = xfs_initialize_perag_data(mp, mp->m_sb.sb_agcount); + if (error) + return error; + } + + /* + * Older kernels misused sb_frextents to reflect both incore + * reservations made by running transactions and the actual count of + * free rt extents in the ondisk metadata. Transactions committed + * during runtime can therefore contain a superblock update that + * undercounts the number of free rt extents tracked in the rt bitmap. + * A clean unmount record will have the correct frextents value since + * there can be no other transactions running at that point. + * + * If we're mounting the rt volume after recovering the log, recompute + * frextents from the rtbitmap file to fix the inconsistency. + */ + if (xfs_has_realtime(mp) && !xfs_has_zoned(mp) && !xfs_is_clean(mp)) { + error = xfs_rtalloc_reinit_frextents(mp); + if (error) + return error; + } + + return 0; +} + +static void +xfs_unmount_check( + struct xfs_mount *mp) +{ + if (xfs_is_shutdown(mp)) + return; + + if (percpu_counter_sum(&mp->m_ifree) > + percpu_counter_sum(&mp->m_icount)) { + xfs_alert(mp, "ifree/icount mismatch at unmount"); + xfs_fs_mark_sick(mp, XFS_SICK_FS_COUNTERS); + } +} + +/* + * Flush and reclaim dirty inodes in preparation for unmount. Inodes and + * internal inode structures can be sitting in the CIL and AIL at this point, + * so we need to unpin them, write them back and/or reclaim them before unmount + * can proceed. In other words, callers are required to have inactivated all + * inodes. + * + * An inode cluster that has been freed can have its buffer still pinned in + * memory because the transaction is still sitting in a iclog. The stale inodes + * on that buffer will be pinned to the buffer until the transaction hits the + * disk and the callbacks run. Pushing the AIL will skip the stale inodes and + * may never see the pinned buffer, so nothing will push out the iclog and + * unpin the buffer. + * + * Hence we need to force the log to unpin everything first. However, log + * forces don't wait for the discards they issue to complete, so we have to + * explicitly wait for them to complete here as well. + * + * Then we can tell the world we are unmounting so that error handling knows + * that the filesystem is going away and we should error out anything that we + * have been retrying in the background. This will prevent never-ending + * retries in AIL pushing from hanging the unmount. + * + * Finally, we can push the AIL to clean all the remaining dirty objects, then + * reclaim the remaining inodes that are still in memory at this point in time. + */ +static void +xfs_unmount_flush_inodes( + struct xfs_mount *mp) +{ + xfs_log_force(mp, XFS_LOG_SYNC); + xfs_extent_busy_wait_all(mp); + flush_workqueue(xfs_discard_wq); + + xfs_set_unmounting(mp); + + xfs_ail_push_all_sync(mp->m_ail); + xfs_inodegc_stop(mp); + cancel_delayed_work_sync(&mp->m_reclaim_work); + xfs_reclaim_inodes(mp); + xfs_health_unmount(mp); +} + +static void +xfs_mount_setup_inode_geom( + struct xfs_mount *mp) +{ + struct xfs_ino_geometry *igeo = M_IGEO(mp); + + igeo->attr_fork_offset = xfs_bmap_compute_attr_offset(mp); + ASSERT(igeo->attr_fork_offset < XFS_LITINO(mp)); + + xfs_ialloc_setup_geometry(mp); +} + +/* Mount the metadata directory tree root. */ +STATIC int +xfs_mount_setup_metadir( + struct xfs_mount *mp) +{ + int error; + + /* Load the metadata directory root inode into memory. */ + error = xfs_metafile_iget(mp, mp->m_sb.sb_metadirino, XFS_METAFILE_DIR, + &mp->m_metadirip); + if (error) + xfs_warn(mp, "Failed to load metadir root directory, error %d", + error); + return error; +} + +/* Compute maximum possible height for per-AG btree types for this fs. */ +static inline void +xfs_agbtree_compute_maxlevels( + struct xfs_mount *mp) +{ + unsigned int levels; + + levels = max(mp->m_alloc_maxlevels, M_IGEO(mp)->inobt_maxlevels); + levels = max(levels, mp->m_rmap_maxlevels); + mp->m_agbtree_maxlevels = max(levels, mp->m_refc_maxlevels); +} + +/* Maximum atomic write IO size that the kernel allows. */ +static inline xfs_extlen_t xfs_calc_atomic_write_max(struct xfs_mount *mp) +{ + return rounddown_pow_of_two(XFS_B_TO_FSB(mp, MAX_RW_COUNT)); +} + +/* + * If the underlying device advertises atomic write support, limit the size of + * atomic writes to the greatest power-of-two factor of the group size so + * that every atomic write unit aligns with the start of every group. This is + * required so that the allocations for an atomic write will always be + * aligned compatibly with the alignment requirements of the storage. + * + * If the device doesn't advertise atomic writes, then there are no alignment + * restrictions and the largest out-of-place write we can do ourselves is the + * number of blocks that user files can allocate from any group. + */ +static xfs_extlen_t +xfs_calc_group_awu_max( + struct xfs_mount *mp, + enum xfs_group_type type) +{ + struct xfs_groups *g = &mp->m_groups[type]; + struct xfs_buftarg *btp = xfs_group_type_buftarg(mp, type); + + if (g->blocks == 0) return 0; + if (btp && btp->bt_awu_min > 0) + return max_pow_of_two_factor(g->blocks); + return rounddown_pow_of_two(g->blocks); +} + +/* Compute the maximum atomic write unit size for each section. */ +static inline void +xfs_calc_atomic_write_unit_max( + struct xfs_mount *mp, + enum xfs_group_type type) +{ + struct xfs_groups *g = &mp->m_groups[type]; - return xfs_initialize_perag_data(mp, mp->m_sb.sb_agcount); + const xfs_extlen_t max_write = xfs_calc_atomic_write_max(mp); + const xfs_extlen_t max_ioend = xfs_reflink_max_atomic_cow(mp); + const xfs_extlen_t max_gsize = xfs_calc_group_awu_max(mp, type); + + g->awu_max = min3(max_write, max_ioend, max_gsize); + trace_xfs_calc_atomic_write_unit_max(mp, type, max_write, max_ioend, + max_gsize, g->awu_max); +} + +/* + * Try to set the atomic write maximum to a new value that we got from + * userspace via mount option. + */ +int +xfs_set_max_atomic_write_opt( + struct xfs_mount *mp, + unsigned long long new_max_bytes) +{ + const xfs_filblks_t new_max_fsbs = XFS_B_TO_FSBT(mp, new_max_bytes); + const xfs_extlen_t max_write = xfs_calc_atomic_write_max(mp); + const xfs_extlen_t max_group = + max(mp->m_groups[XG_TYPE_AG].blocks, + mp->m_groups[XG_TYPE_RTG].blocks); + const xfs_extlen_t max_group_write = + max(xfs_calc_group_awu_max(mp, XG_TYPE_AG), + xfs_calc_group_awu_max(mp, XG_TYPE_RTG)); + int error; + + if (new_max_bytes == 0) + goto set_limit; + + ASSERT(max_write <= U32_MAX); + + /* generic_atomic_write_valid enforces power of two length */ + if (!is_power_of_2(new_max_bytes)) { + xfs_warn(mp, + "max atomic write size of %llu bytes is not a power of 2", + new_max_bytes); + return -EINVAL; + } + + if (new_max_bytes & mp->m_blockmask) { + xfs_warn(mp, + "max atomic write size of %llu bytes not aligned with fsblock", + new_max_bytes); + return -EINVAL; + } + + if (new_max_fsbs > max_write) { + xfs_warn(mp, + "max atomic write size of %lluk cannot be larger than max write size %lluk", + new_max_bytes >> 10, + XFS_FSB_TO_B(mp, max_write) >> 10); + return -EINVAL; + } + + if (new_max_fsbs > max_group) { + xfs_warn(mp, + "max atomic write size of %lluk cannot be larger than allocation group size %lluk", + new_max_bytes >> 10, + XFS_FSB_TO_B(mp, max_group) >> 10); + return -EINVAL; + } + + if (new_max_fsbs > max_group_write) { + xfs_warn(mp, + "max atomic write size of %lluk cannot be larger than max allocation group write size %lluk", + new_max_bytes >> 10, + XFS_FSB_TO_B(mp, max_group_write) >> 10); + return -EINVAL; + } + + if (xfs_has_reflink(mp)) + goto set_limit; + + if (new_max_fsbs == 1) { + if (mp->m_ddev_targp->bt_awu_max || + (mp->m_rtdev_targp && mp->m_rtdev_targp->bt_awu_max)) { + } else { + xfs_warn(mp, + "cannot support atomic writes of size %lluk with no reflink or HW support", + new_max_bytes >> 10); + return -EINVAL; + } + } else { + xfs_warn(mp, + "cannot support atomic writes of size %lluk with no reflink support", + new_max_bytes >> 10); + return -EINVAL; + } + +set_limit: + error = xfs_calc_atomic_write_reservation(mp, new_max_fsbs); + if (error) { + xfs_warn(mp, + "cannot support completing atomic writes of %lluk", + new_max_bytes >> 10); + return error; + } + + xfs_calc_atomic_write_unit_max(mp, XG_TYPE_AG); + xfs_calc_atomic_write_unit_max(mp, XG_TYPE_RTG); + mp->m_awu_max_bytes = new_max_bytes; + return 0; +} + +/* Compute maximum possible height for realtime btree types for this fs. */ +static inline void +xfs_rtbtree_compute_maxlevels( + struct xfs_mount *mp) +{ + mp->m_rtbtree_maxlevels = max(mp->m_rtrmap_maxlevels, + mp->m_rtrefc_maxlevels); } /* @@ -676,10 +838,11 @@ xfs_mountfs( { struct xfs_sb *sbp = &(mp->m_sb); struct xfs_inode *rip; - uint64_t resblks; + struct xfs_ino_geometry *igeo = M_IGEO(mp); uint quotamount = 0; uint quotaflags = 0; int error = 0; + int i; xfs_sb_mount_common(mp, sbp); @@ -703,128 +866,94 @@ xfs_mountfs( xfs_warn(mp, "correcting sb_features alignment problem"); sbp->sb_features2 |= sbp->sb_bad_features2; mp->m_update_sb = true; - - /* - * Re-check for ATTR2 in case it was found in bad_features2 - * slot. - */ - if (xfs_sb_version_hasattr2(&mp->m_sb) && - !(mp->m_flags & XFS_MOUNT_NOATTR2)) - mp->m_flags |= XFS_MOUNT_ATTR2; } - if (xfs_sb_version_hasattr2(&mp->m_sb) && - (mp->m_flags & XFS_MOUNT_NOATTR2)) { - xfs_sb_version_removeattr2(&mp->m_sb); - mp->m_update_sb = true; - - /* update sb_versionnum for the clearing of the morebits */ - if (!sbp->sb_features2) - mp->m_update_sb = true; - } /* always use v2 inodes by default now */ if (!(mp->m_sb.sb_versionnum & XFS_SB_VERSION_NLINKBIT)) { mp->m_sb.sb_versionnum |= XFS_SB_VERSION_NLINKBIT; + mp->m_features |= XFS_FEAT_NLINK; mp->m_update_sb = true; } /* - * Check if sb_agblocks is aligned at stripe boundary - * If sb_agblocks is NOT aligned turn off m_dalign since - * allocator alignment is within an ag, therefore ag has - * to be aligned at stripe boundary. + * If we were given new sunit/swidth options, do some basic validation + * checks and convert the incore dalign and swidth values to the + * same units (FSB) that everything else uses. This /must/ happen + * before computing the inode geometry. */ - error = xfs_update_alignment(mp); + error = xfs_validate_new_dalign(mp); if (error) goto out; xfs_alloc_compute_maxlevels(mp); xfs_bmap_compute_maxlevels(mp, XFS_DATA_FORK); xfs_bmap_compute_maxlevels(mp, XFS_ATTR_FORK); - xfs_ialloc_compute_maxlevels(mp); + xfs_mount_setup_inode_geom(mp); xfs_rmapbt_compute_maxlevels(mp); + xfs_rtrmapbt_compute_maxlevels(mp); xfs_refcountbt_compute_maxlevels(mp); + xfs_rtrefcountbt_compute_maxlevels(mp); - xfs_set_maxicount(mp); - - /* enable fail_at_unmount as default */ - mp->m_fail_unmount = true; + xfs_agbtree_compute_maxlevels(mp); + xfs_rtbtree_compute_maxlevels(mp); - error = xfs_sysfs_init(&mp->m_kobj, &xfs_mp_ktype, NULL, mp->m_fsname); + /* + * Check if sb_agblocks is aligned at stripe boundary. If sb_agblocks + * is NOT aligned turn off m_dalign since allocator alignment is within + * an ag, therefore ag has to be aligned at stripe boundary. Note that + * we must compute the free space and rmap btree geometry before doing + * this. + */ + error = xfs_update_alignment(mp); if (error) goto out; - error = xfs_sysfs_init(&mp->m_stats.xs_kobj, &xfs_stats_ktype, - &mp->m_kobj, "stats"); - if (error) - goto out_remove_sysfs; + /* enable fail_at_unmount as default */ + mp->m_fail_unmount = true; - error = xfs_error_sysfs_init(mp); + error = xfs_mount_sysfs_init(mp); if (error) - goto out_del_stats; + goto out_remove_scrub_stats; + + xchk_stats_register(mp->m_scrub_stats, mp->m_debugfs); error = xfs_errortag_init(mp); if (error) - goto out_remove_error_sysfs; + goto out_remove_sysfs; error = xfs_uuid_mount(mp); if (error) goto out_remove_errortag; /* - * Set the minimum read and write sizes + * Update the preferred write size based on the information from the + * on-disk superblock. */ - xfs_set_rw_sizes(mp); + mp->m_allocsize_log = + max_t(uint32_t, sbp->sb_blocklog, mp->m_allocsize_log); + mp->m_allocsize_blocks = 1U << (mp->m_allocsize_log - sbp->sb_blocklog); /* set the low space thresholds for dynamic preallocation */ xfs_set_low_space_thresholds(mp); /* - * Set the inode cluster size. - * This may still be overridden by the file system - * block size if it is larger than the chosen cluster size. - * - * For v5 filesystems, scale the cluster size with the inode size to - * keep a constant ratio of inode per cluster buffer, but only if mkfs - * has set the inode alignment value appropriately for larger cluster - * sizes. - */ - mp->m_inode_cluster_size = XFS_INODE_BIG_CLUSTER_SIZE; - if (xfs_sb_version_hascrc(&mp->m_sb)) { - int new_size = mp->m_inode_cluster_size; - - new_size *= mp->m_sb.sb_inodesize / XFS_DINODE_MIN_SIZE; - if (mp->m_sb.sb_inoalignmt >= XFS_B_TO_FSBT(mp, new_size)) - mp->m_inode_cluster_size = new_size; - } - mp->m_blocks_per_cluster = xfs_icluster_size_fsb(mp); - mp->m_inodes_per_cluster = XFS_FSB_TO_INO(mp, mp->m_blocks_per_cluster); - mp->m_cluster_align = xfs_ialloc_cluster_alignment(mp); - mp->m_cluster_align_inodes = XFS_FSB_TO_INO(mp, mp->m_cluster_align); - - /* * If enabled, sparse inode chunk alignment is expected to match the * cluster size. Full inode chunk alignment must match the chunk size, * but that is checked on sb read verification... */ - if (xfs_sb_version_hassparseinodes(&mp->m_sb) && + if (xfs_has_sparseinodes(mp) && mp->m_sb.sb_spino_align != - XFS_B_TO_FSBT(mp, mp->m_inode_cluster_size)) { + XFS_B_TO_FSBT(mp, igeo->inode_cluster_size_raw)) { xfs_warn(mp, "Sparse inode block alignment (%u) must match cluster size (%llu).", mp->m_sb.sb_spino_align, - XFS_B_TO_FSBT(mp, mp->m_inode_cluster_size)); + XFS_B_TO_FSBT(mp, igeo->inode_cluster_size_raw)); error = -EINVAL; goto out_remove_uuid; } /* - * Set inode alignment fields - */ - xfs_set_inoalignment(mp); - - /* * Check that the data (and log if separate) is an ok size. */ error = xfs_check_sizes(mp); @@ -863,19 +992,37 @@ xfs_mountfs( /* * Allocate and initialize the per-ag data. */ - error = xfs_initialize_perag(mp, sbp->sb_agcount, &mp->m_maxagi); + error = xfs_initialize_perag(mp, 0, sbp->sb_agcount, + mp->m_sb.sb_dblocks, &mp->m_maxagi); if (error) { xfs_warn(mp, "Failed per-ag init: %d", error); goto out_free_dir; } - if (!sbp->sb_logblocks) { + error = xfs_initialize_rtgroups(mp, 0, sbp->sb_rgcount, + mp->m_sb.sb_rextents); + if (error) { + xfs_warn(mp, "Failed rtgroup init: %d", error); + goto out_free_perag; + } + + if (XFS_IS_CORRUPT(mp, !sbp->sb_logblocks)) { xfs_warn(mp, "no log defined"); - XFS_ERROR_REPORT("xfs_mountfs", XFS_ERRLEVEL_LOW, mp); error = -EFSCORRUPTED; - goto out_free_perag; + goto out_free_rtgroup; } + error = xfs_inodegc_register_shrinker(mp); + if (error) + goto out_fail_wait; + + /* + * If we're resuming quota status, pick up the preliminary qflags from + * the ondisk superblock so that we know if we should recover dquots. + */ + if (xfs_is_resuming_quotaon(mp)) + xfs_qm_resume_quotaon(mp); + /* * Log's mount-time initialization. The first part of recovery can place * some items on the AIL, to be handled when recovery is finished or @@ -886,13 +1033,35 @@ xfs_mountfs( XFS_FSB_TO_BB(mp, sbp->sb_logblocks)); if (error) { xfs_warn(mp, "log mount failed"); - goto out_fail_wait; + goto out_inodegc_shrinker; } - /* Make sure the summary counts are ok. */ - error = xfs_check_summary_counts(mp); - if (error) - goto out_log_dealloc; + /* + * If we're resuming quota status and recovered the log, re-sample the + * qflags from the ondisk superblock now that we've recovered it, just + * in case someone shut down enforcement just before a crash. + */ + if (xfs_clear_resuming_quotaon(mp) && xlog_recovery_needed(mp->m_log)) + xfs_qm_resume_quotaon(mp); + + /* + * If logged xattrs are still enabled after log recovery finishes, then + * they'll be available until unmount. Otherwise, turn them off. + */ + if (xfs_sb_version_haslogxattrs(&mp->m_sb)) + xfs_set_using_logged_xattrs(mp); + else + xfs_clear_using_logged_xattrs(mp); + + /* Enable background inode inactivation workers. */ + xfs_inodegc_start(mp); + xfs_blockgc_start(mp); + + if (xfs_has_metadir(mp)) { + error = xfs_mount_setup_metadir(mp); + if (error) + goto out_free_metadir; + } /* * Get and sanity-check the root inode. @@ -904,17 +1073,15 @@ xfs_mountfs( xfs_warn(mp, "Failed to read root inode 0x%llx, error %d", sbp->sb_rootino, -error); - goto out_log_dealloc; + goto out_free_metadir; } ASSERT(rip != NULL); - if (unlikely(!S_ISDIR(VFS_I(rip)->i_mode))) { + if (XFS_IS_CORRUPT(mp, !S_ISDIR(VFS_I(rip)->i_mode))) { xfs_warn(mp, "corrupted root inode %llu: not a directory", (unsigned long long)rip->i_ino); xfs_iunlock(rip, XFS_ILOCK_EXCL); - XFS_ERROR_REPORT("xfs_mountfs_int(2)", XFS_ERRLEVEL_LOW, - mp); error = -EFSCORRUPTED; goto out_rele_rip; } @@ -934,12 +1101,17 @@ xfs_mountfs( goto out_rele_rip; } + /* Make sure the summary counts are ok. */ + error = xfs_check_summary_counts(mp); + if (error) + goto out_rtunmount; + /* * If this is a read-only mount defer the superblock updates until * the next remount into writeable mode. Otherwise we would never * perform the update e.g. for the root filesystem. */ - if (mp->m_update_sb && !(mp->m_flags & XFS_MOUNT_RDONLY)) { + if (mp->m_update_sb && !xfs_is_readonly(mp)) { error = xfs_sync_sb(mp, false); if (error) { xfs_warn(mp, "failed to write sb changes"); @@ -950,13 +1122,11 @@ xfs_mountfs( /* * Initialise the XFS quota management subsystem for this mount */ - if (XFS_IS_QUOTA_RUNNING(mp)) { + if (XFS_IS_QUOTA_ON(mp)) { error = xfs_qm_newmount(mp, "amount, "aflags); if (error) goto out_rtunmount; } else { - ASSERT(!XFS_IS_QUOTA_ON(mp)); - /* * If a file system had quotas running earlier, but decided to * mount without -o uquota/pquota/gquota options, revoke the @@ -973,9 +1143,17 @@ xfs_mountfs( /* * Finish recovering the file system. This part needed to be delayed * until after the root and real-time bitmap inodes were consistently - * read in. + * read in. Temporarily create per-AG space reservations for metadata + * btree shape changes because space freeing transactions (for inode + * inactivation) require the per-AG reservation in lieu of reserving + * blocks. */ + error = xfs_fs_reserve_ag_blocks(mp); + if (error && error == -ENOSPC) + xfs_warn(mp, + "ENOSPC reserving per-AG metadata pool, log recovery may fail."); error = xfs_log_mount_finish(mp); + xfs_fs_unreserve_ag_blocks(mp); if (error) { xfs_warn(mp, "log mount finish failed"); goto out_rtunmount; @@ -990,9 +1168,13 @@ xfs_mountfs( * We use the same quiesce mechanism as the rw->ro remount, as they are * semantically identical operations. */ - if ((mp->m_flags & (XFS_MOUNT_RDONLY|XFS_MOUNT_NORECOVERY)) == - XFS_MOUNT_RDONLY) { - xfs_quiesce_attr(mp); + if (xfs_is_readonly(mp) && !xfs_has_norecovery(mp)) + xfs_log_clean(mp); + + if (xfs_has_zoned(mp)) { + error = xfs_mount_zones(mp); + if (error) + goto out_rtunmount; } /* @@ -1010,48 +1192,66 @@ xfs_mountfs( * privileged transactions. This is needed so that transaction * space required for critical operations can dip into this pool * when at ENOSPC. This is needed for operations like create with - * attr, unwritten extent conversion at ENOSPC, etc. Data allocations - * are not allowed to use this reserved space. + * attr, unwritten extent conversion at ENOSPC, garbage collection + * etc. Data allocations are not allowed to use this reserved space. * * This may drive us straight to ENOSPC on mount, but that implies * we were already there on the last unmount. Warn if this occurs. */ - if (!(mp->m_flags & XFS_MOUNT_RDONLY)) { - resblks = xfs_default_resblks(mp); - error = xfs_reserve_blocks(mp, &resblks, NULL); - if (error) - xfs_warn(mp, - "Unable to allocate reserve blocks. Continuing without reserve pool."); - - /* Recover any CoW blocks that never got remapped. */ - error = xfs_reflink_recover_cow(mp); - if (error) { - xfs_err(mp, - "Error %d recovering leftover CoW allocations.", error); - xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); - goto out_quota; + if (!xfs_is_readonly(mp)) { + for (i = 0; i < XC_FREE_NR; i++) { + error = xfs_reserve_blocks(mp, i, + xfs_default_resblks(mp, i)); + if (error) + xfs_warn(mp, +"Unable to allocate reserve blocks. Continuing without reserve pool for %s.", + xfs_free_pool_name[i]); } /* Reserve AG blocks for future btree expansion. */ error = xfs_fs_reserve_ag_blocks(mp); if (error && error != -ENOSPC) goto out_agresv; + + xfs_zone_gc_start(mp); } + /* + * Pre-calculate atomic write unit max. This involves computations + * derived from transaction reservations, so we must do this after the + * log is fully initialized. + */ + error = xfs_set_max_atomic_write_opt(mp, mp->m_awu_max_bytes); + if (error) + goto out_agresv; + return 0; out_agresv: xfs_fs_unreserve_ag_blocks(mp); - out_quota: xfs_qm_unmount_quotas(mp); + if (xfs_has_zoned(mp)) + xfs_unmount_zones(mp); out_rtunmount: xfs_rtunmount_inodes(mp); out_rele_rip: xfs_irele(rip); /* Clean out dquots that might be in memory after quotacheck. */ xfs_qm_unmount(mp); + out_free_metadir: + if (mp->m_metadirip) + xfs_irele(mp->m_metadirip); + /* - * Cancel all delayed reclaim work and reclaim the inodes directly. + * Inactivate all inodes that might still be in memory after a log + * intent recovery failure so that reclaim can free them. Metadata + * inodes and the root directory shouldn't need inactivation, but the + * mount failed for some reason, so pull down all the state and flee. + */ + xfs_inodegc_flush(mp); + + /* + * Flush all inode reclamation work and flush the log. * We have to do this /after/ rtunmount and qm_unmount because those * two will have scheduled delayed reclaim for the rt/quota inodes. * @@ -1061,29 +1261,28 @@ xfs_mountfs( * qm_unmount_quotas and therefore rely on qm_unmount to release the * quota inodes. */ - cancel_delayed_work_sync(&mp->m_reclaim_work); - xfs_reclaim_inodes(mp, SYNC_WAIT); - out_log_dealloc: - mp->m_flags |= XFS_MOUNT_UNMOUNTING; + xfs_unmount_flush_inodes(mp); xfs_log_mount_cancel(mp); + out_inodegc_shrinker: + shrinker_free(mp->m_inodegc_shrinker); out_fail_wait: if (mp->m_logdev_targp && mp->m_logdev_targp != mp->m_ddev_targp) - xfs_wait_buftarg(mp->m_logdev_targp); - xfs_wait_buftarg(mp->m_ddev_targp); + xfs_buftarg_drain(mp->m_logdev_targp); + xfs_buftarg_drain(mp->m_ddev_targp); + out_free_rtgroup: + xfs_free_rtgroups(mp, 0, mp->m_sb.sb_rgcount); out_free_perag: - xfs_free_perag(mp); + xfs_free_perag_range(mp, 0, mp->m_sb.sb_agcount); out_free_dir: xfs_da_unmount(mp); out_remove_uuid: xfs_uuid_unmount(mp); out_remove_errortag: xfs_errortag_del(mp); - out_remove_error_sysfs: - xfs_error_sysfs_del(mp); - out_del_stats: - xfs_sysfs_del(&mp->m_stats.xs_kobj); out_remove_sysfs: - xfs_sysfs_del(&mp->m_kobj); + xfs_mount_sysfs_del(mp); + out_remove_scrub_stats: + xchk_stats_unregister(mp->m_scrub_stats); out: return error; } @@ -1096,55 +1295,31 @@ void xfs_unmountfs( struct xfs_mount *mp) { - uint64_t resblks; int error; - xfs_icache_disable_reclaim(mp); + /* + * Perform all on-disk metadata updates required to inactivate inodes + * that the VFS evicted earlier in the unmount process. Freeing inodes + * and discarding CoW fork preallocations can cause shape changes to + * the free inode and refcount btrees, respectively, so we must finish + * this before we discard the metadata space reservations. Metadata + * inodes and the root directory do not require inactivation. + */ + xfs_inodegc_flush(mp); + + xfs_blockgc_stop(mp); + if (!test_bit(XFS_OPSTATE_READONLY, &mp->m_opstate)) + xfs_zone_gc_stop(mp); xfs_fs_unreserve_ag_blocks(mp); xfs_qm_unmount_quotas(mp); + if (xfs_has_zoned(mp)) + xfs_unmount_zones(mp); xfs_rtunmount_inodes(mp); xfs_irele(mp->m_rootip); + if (mp->m_metadirip) + xfs_irele(mp->m_metadirip); - /* - * We can potentially deadlock here if we have an inode cluster - * that has been freed has its buffer still pinned in memory because - * the transaction is still sitting in a iclog. The stale inodes - * on that buffer will have their flush locks held until the - * transaction hits the disk and the callbacks run. the inode - * flush takes the flush lock unconditionally and with nothing to - * push out the iclog we will never get that unlocked. hence we - * need to force the log first. - */ - xfs_log_force(mp, XFS_LOG_SYNC); - - /* - * Wait for all busy extents to be freed, including completion of - * any discard operation. - */ - xfs_extent_busy_wait_all(mp); - flush_workqueue(xfs_discard_wq); - - /* - * We now need to tell the world we are unmounting. This will allow - * us to detect that the filesystem is going away and we should error - * out anything that we have been retrying in the background. This will - * prevent neverending retries in AIL pushing from hanging the unmount. - */ - mp->m_flags |= XFS_MOUNT_UNMOUNTING; - - /* - * Flush all pending changes from the AIL. - */ - xfs_ail_push_all_sync(mp->m_ail); - - /* - * And reclaim all inodes. At this point there should be no dirty - * inodes and none should be pinned or locked, but use synchronous - * reclaim just to be sure. We can stop background inode reclaim - * here as well if it is still running. - */ - cancel_delayed_work_sync(&mp->m_reclaim_work); - xfs_reclaim_inodes(mp, SYNC_WAIT); + xfs_unmount_flush_inodes(mp); xfs_qm_unmount(mp); @@ -1162,18 +1337,17 @@ xfs_unmountfs( * we only every apply deltas to the superblock and hence the incore * value does not matter.... */ - resblks = 0; - error = xfs_reserve_blocks(mp, &resblks, NULL); + error = xfs_reserve_blocks(mp, XC_FREE_BLOCKS, 0); if (error) xfs_warn(mp, "Unable to free reserved block pool. " "Freespace may not be correct on next mount."); + xfs_unmount_check(mp); - error = xfs_log_sbcount(mp); - if (error) - xfs_warn(mp, "Unable to update superblock counters. " - "Freespace may not be correct on next mount."); - - + /* + * Indicate that it's ok to clear log incompat bits before cleaning + * the log and writing the unmount record. + */ + xfs_set_done_with_log_incompat(mp); xfs_log_unmount(mp); xfs_da_unmount(mp); xfs_uuid_unmount(mp); @@ -1181,12 +1355,12 @@ xfs_unmountfs( #if defined(DEBUG) xfs_errortag_clearall(mp); #endif - xfs_free_perag(mp); - + shrinker_free(mp->m_inodegc_shrinker); + xfs_free_rtgroups(mp, 0, mp->m_sb.sb_rgcount); + xfs_free_perag_range(mp, 0, mp->m_sb.sb_agcount); xfs_errortag_del(mp); - xfs_error_sysfs_del(mp); - xfs_sysfs_del(&mp->m_stats.xs_kobj); - xfs_sysfs_del(&mp->m_kobj); + xchk_stats_unregister(mp->m_scrub_stats); + xfs_mount_sysfs_del(mp); } /* @@ -1202,112 +1376,73 @@ xfs_fs_writable( { ASSERT(level > SB_UNFROZEN); if ((mp->m_super->s_writers.frozen >= level) || - XFS_FORCED_SHUTDOWN(mp) || (mp->m_flags & XFS_MOUNT_RDONLY)) + xfs_is_shutdown(mp) || xfs_is_readonly(mp)) return false; return true; } /* - * xfs_log_sbcount - * - * Sync the superblock counters to disk. + * Estimate the amount of free space that is not available to userspace and is + * not explicitly reserved from the incore fdblocks. This includes: * - * Note this code can be called during the process of freezing, so we use the - * transaction allocator that does not block when the transaction subsystem is - * in its frozen state. + * - The minimum number of blocks needed to support splitting a bmap btree + * - The blocks currently in use by the freespace btrees because they record + * the actual blocks that will fill per-AG metadata space reservations */ -int -xfs_log_sbcount(xfs_mount_t *mp) +uint64_t +xfs_freecounter_unavailable( + struct xfs_mount *mp, + enum xfs_free_counter ctr) { - /* allow this to proceed during the freeze sequence... */ - if (!xfs_fs_writable(mp, SB_FREEZE_COMPLETE)) - return 0; - - /* - * we don't need to do this if we are updating the superblock - * counters on every modification. - */ - if (!xfs_sb_version_haslazysbcount(&mp->m_sb)) + if (ctr != XC_FREE_BLOCKS) return 0; - - return xfs_sync_sb(mp, true); + return mp->m_alloc_set_aside + atomic64_read(&mp->m_allocbt_blks); } -/* - * Deltas for the inode count are +/-64, hence we use a large batch size - * of 128 so we don't need to take the counter lock on every update. - */ -#define XFS_ICOUNT_BATCH 128 -int -xfs_mod_icount( +void +xfs_add_freecounter( struct xfs_mount *mp, - int64_t delta) + enum xfs_free_counter ctr, + uint64_t delta) { - percpu_counter_add_batch(&mp->m_icount, delta, XFS_ICOUNT_BATCH); - if (__percpu_counter_compare(&mp->m_icount, 0, XFS_ICOUNT_BATCH) < 0) { - ASSERT(0); - percpu_counter_add(&mp->m_icount, -delta); - return -EINVAL; + struct xfs_freecounter *counter = &mp->m_free[ctr]; + uint64_t res_used; + + /* + * If the reserve pool is depleted, put blocks back into it first. + * Most of the time the pool is full. + */ + if (likely(counter->res_avail == counter->res_total)) { + percpu_counter_add(&counter->count, delta); + return; } - return 0; -} -int -xfs_mod_ifree( - struct xfs_mount *mp, - int64_t delta) -{ - percpu_counter_add(&mp->m_ifree, delta); - if (percpu_counter_compare(&mp->m_ifree, 0) < 0) { - ASSERT(0); - percpu_counter_add(&mp->m_ifree, -delta); - return -EINVAL; + spin_lock(&mp->m_sb_lock); + res_used = counter->res_total - counter->res_avail; + if (res_used > delta) { + counter->res_avail += delta; + } else { + delta -= res_used; + counter->res_avail = counter->res_total; + percpu_counter_add(&counter->count, delta); } - return 0; + spin_unlock(&mp->m_sb_lock); } -/* - * Deltas for the block count can vary from 1 to very large, but lock contention - * only occurs on frequent small block count updates such as in the delayed - * allocation path for buffered writes (page a time updates). Hence we set - * a large batch count (1024) to minimise global counter updates except when - * we get near to ENOSPC and we have to be very accurate with our updates. - */ -#define XFS_FDBLOCKS_BATCH 1024 + +/* Adjust in-core free blocks or RT extents. */ int -xfs_mod_fdblocks( +xfs_dec_freecounter( struct xfs_mount *mp, - int64_t delta, + enum xfs_free_counter ctr, + uint64_t delta, bool rsvd) { - int64_t lcounter; - long long res_used; + struct xfs_freecounter *counter = &mp->m_free[ctr]; s32 batch; - if (delta > 0) { - /* - * If the reserve pool is depleted, put blocks back into it - * first. Most of the time the pool is full. - */ - if (likely(mp->m_resblks == mp->m_resblks_avail)) { - percpu_counter_add(&mp->m_fdblocks, delta); - return 0; - } - - spin_lock(&mp->m_sb_lock); - res_used = (long long)(mp->m_resblks - mp->m_resblks_avail); - - if (res_used > delta) { - mp->m_resblks_avail += delta; - } else { - delta -= res_used; - mp->m_resblks_avail = mp->m_resblks; - percpu_counter_add(&mp->m_fdblocks, delta); - } - spin_unlock(&mp->m_sb_lock); - return 0; - } + ASSERT(ctr < XC_FREE_NR); /* * Taking blocks away, need to be more accurate the closer we @@ -1317,86 +1452,53 @@ xfs_mod_fdblocks( * then make everything serialise as we are real close to * ENOSPC. */ - if (__percpu_counter_compare(&mp->m_fdblocks, 2 * XFS_FDBLOCKS_BATCH, + if (__percpu_counter_compare(&counter->count, 2 * XFS_FDBLOCKS_BATCH, XFS_FDBLOCKS_BATCH) < 0) batch = 1; else batch = XFS_FDBLOCKS_BATCH; - percpu_counter_add_batch(&mp->m_fdblocks, delta, batch); - if (__percpu_counter_compare(&mp->m_fdblocks, mp->m_alloc_set_aside, - XFS_FDBLOCKS_BATCH) >= 0) { - /* we had space! */ - return 0; - } - /* - * lock up the sb for dipping into reserves before releasing the space - * that took us to ENOSPC. + * Set aside allocbt blocks because these blocks are tracked as free + * space but not available for allocation. Technically this means that a + * single reservation cannot consume all remaining free space, but the + * ratio of allocbt blocks to usable free blocks should be rather small. + * The tradeoff without this is that filesystems that maintain high + * perag block reservations can over reserve physical block availability + * and fail physical allocation, which leads to much more serious + * problems (i.e. transaction abort, pagecache discards, etc.) than + * slightly premature -ENOSPC. */ - spin_lock(&mp->m_sb_lock); - percpu_counter_add(&mp->m_fdblocks, -delta); - if (!rsvd) - goto fdblocks_enospc; - - lcounter = (long long)mp->m_resblks_avail + delta; - if (lcounter >= 0) { - mp->m_resblks_avail = lcounter; + percpu_counter_add_batch(&counter->count, -((int64_t)delta), batch); + if (__percpu_counter_compare(&counter->count, + xfs_freecounter_unavailable(mp, ctr), + XFS_FDBLOCKS_BATCH) < 0) { + /* + * Lock up the sb for dipping into reserves before releasing the + * space that took us to ENOSPC. + */ + spin_lock(&mp->m_sb_lock); + percpu_counter_add(&counter->count, delta); + if (!rsvd) + goto fdblocks_enospc; + if (delta > counter->res_avail) { + if (ctr == XC_FREE_BLOCKS) + xfs_warn_once(mp, +"Reserve blocks depleted! Consider increasing reserve pool size."); + goto fdblocks_enospc; + } + counter->res_avail -= delta; + trace_xfs_freecounter_reserved(mp, ctr, delta, _RET_IP_); spin_unlock(&mp->m_sb_lock); - return 0; } - printk_once(KERN_WARNING - "Filesystem \"%s\": reserve blocks depleted! " - "Consider increasing reserve pool size.", - mp->m_fsname); -fdblocks_enospc: - spin_unlock(&mp->m_sb_lock); - return -ENOSPC; -} -int -xfs_mod_frextents( - struct xfs_mount *mp, - int64_t delta) -{ - int64_t lcounter; - int ret = 0; + /* we had space! */ + return 0; - spin_lock(&mp->m_sb_lock); - lcounter = mp->m_sb.sb_frextents + delta; - if (lcounter < 0) - ret = -ENOSPC; - else - mp->m_sb.sb_frextents = lcounter; +fdblocks_enospc: + trace_xfs_freecounter_enospc(mp, ctr, delta, _RET_IP_); spin_unlock(&mp->m_sb_lock); - return ret; -} - -/* - * xfs_getsb() is called to obtain the buffer for the superblock. - * The buffer is returned locked and read in from disk. - * The buffer should be released with a call to xfs_brelse(). - * - * If the flags parameter is BUF_TRYLOCK, then we'll only return - * the superblock buffer if it can be locked without sleeping. - * If it can't then we'll return NULL. - */ -struct xfs_buf * -xfs_getsb( - struct xfs_mount *mp, - int flags) -{ - struct xfs_buf *bp = mp->m_sb_bp; - - if (!xfs_buf_trylock(bp)) { - if (flags & XBF_TRYLOCK) - return NULL; - xfs_buf_lock(bp); - } - - xfs_buf_hold(bp); - ASSERT(bp->b_flags & XBF_DONE); - return bp; + return -ENOSPC; } /* @@ -1437,10 +1539,150 @@ void xfs_force_summary_recalc( struct xfs_mount *mp) { - if (!xfs_sb_version_haslazysbcount(&mp->m_sb)) + if (!xfs_has_lazysbcount(mp)) return; - spin_lock(&mp->m_sb_lock); - mp->m_flags |= XFS_MOUNT_BAD_SUMMARY; - spin_unlock(&mp->m_sb_lock); + xfs_fs_mark_sick(mp, XFS_SICK_FS_COUNTERS); +} + +/* + * Enable a log incompat feature flag in the primary superblock. The caller + * cannot have any other transactions in progress. + */ +int +xfs_add_incompat_log_feature( + struct xfs_mount *mp, + uint32_t feature) +{ + struct xfs_dsb *dsb; + int error; + + ASSERT(hweight32(feature) == 1); + ASSERT(!(feature & XFS_SB_FEAT_INCOMPAT_LOG_UNKNOWN)); + + /* + * Force the log to disk and kick the background AIL thread to reduce + * the chances that the bwrite will stall waiting for the AIL to unpin + * the primary superblock buffer. This isn't a data integrity + * operation, so we don't need a synchronous push. + */ + error = xfs_log_force(mp, XFS_LOG_SYNC); + if (error) + return error; + xfs_ail_push_all(mp->m_ail); + + /* + * Lock the primary superblock buffer to serialize all callers that + * are trying to set feature bits. + */ + xfs_buf_lock(mp->m_sb_bp); + xfs_buf_hold(mp->m_sb_bp); + + if (xfs_is_shutdown(mp)) { + error = -EIO; + goto rele; + } + + if (xfs_sb_has_incompat_log_feature(&mp->m_sb, feature)) + goto rele; + + /* + * Write the primary superblock to disk immediately, because we need + * the log_incompat bit to be set in the primary super now to protect + * the log items that we're going to commit later. + */ + dsb = mp->m_sb_bp->b_addr; + xfs_sb_to_disk(dsb, &mp->m_sb); + dsb->sb_features_log_incompat |= cpu_to_be32(feature); + error = xfs_bwrite(mp->m_sb_bp); + if (error) + goto shutdown; + + /* + * Add the feature bits to the incore superblock before we unlock the + * buffer. + */ + xfs_sb_add_incompat_log_features(&mp->m_sb, feature); + xfs_buf_relse(mp->m_sb_bp); + + /* Log the superblock to disk. */ + return xfs_sync_sb(mp, false); +shutdown: + xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR); +rele: + xfs_buf_relse(mp->m_sb_bp); + return error; +} + +/* + * Clear all the log incompat flags from the superblock. + * + * The caller cannot be in a transaction, must ensure that the log does not + * contain any log items protected by any log incompat bit, and must ensure + * that there are no other threads that depend on the state of the log incompat + * feature flags in the primary super. + * + * Returns true if the superblock is dirty. + */ +bool +xfs_clear_incompat_log_features( + struct xfs_mount *mp) +{ + bool ret = false; + + if (!xfs_has_crc(mp) || + !xfs_sb_has_incompat_log_feature(&mp->m_sb, + XFS_SB_FEAT_INCOMPAT_LOG_ALL) || + xfs_is_shutdown(mp) || + !xfs_is_done_with_log_incompat(mp)) + return false; + + /* + * Update the incore superblock. We synchronize on the primary super + * buffer lock to be consistent with the add function, though at least + * in theory this shouldn't be necessary. + */ + xfs_buf_lock(mp->m_sb_bp); + xfs_buf_hold(mp->m_sb_bp); + + if (xfs_sb_has_incompat_log_feature(&mp->m_sb, + XFS_SB_FEAT_INCOMPAT_LOG_ALL)) { + xfs_sb_remove_incompat_log_features(&mp->m_sb); + ret = true; + } + + xfs_buf_relse(mp->m_sb_bp); + return ret; +} + +/* + * Update the in-core delayed block counter. + * + * We prefer to update the counter without having to take a spinlock for every + * counter update (i.e. batching). Each change to delayed allocation + * reservations can change can easily exceed the default percpu counter + * batching, so we use a larger batch factor here. + * + * Note that we don't currently have any callers requiring fast summation + * (e.g. percpu_counter_read) so we can use a big batch value here. + */ +#define XFS_DELALLOC_BATCH (4096) +void +xfs_mod_delalloc( + struct xfs_inode *ip, + int64_t data_delta, + int64_t ind_delta) +{ + struct xfs_mount *mp = ip->i_mount; + + if (XFS_IS_REALTIME_INODE(ip)) { + percpu_counter_add_batch(&mp->m_delalloc_rtextents, + xfs_blen_to_rtbxlen(mp, data_delta), + XFS_DELALLOC_BATCH); + if (!ind_delta) + return; + data_delta = 0; + } + percpu_counter_add_batch(&mp->m_delalloc_blks, data_delta + ind_delta, + XFS_DELALLOC_BATCH); } |