From 211d022c43cac3aecbe967fcaf9b10156bfa63ad Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 11 Apr 2013 22:09:56 +0200 Subject: [PATCH 01/61] xfs: Avoid pathological backwards allocation Writing a large file using direct IO in 16 MB chunks sometimes results in a pathological allocation pattern where 16 MB chunks of large free extent are allocated to a file in a reversed order. So extents of a file look for example as: ext logical physical expected length flags 0 0 13 4550656 1 4550656 188136807 4550668 12562432 2 17113088 200699240 200699238 622592 3 17735680 182046055 201321831 4096 4 17739776 182041959 182050150 4096 5 17743872 182037863 182046054 4096 6 17747968 182033767 182041958 4096 7 17752064 182029671 182037862 4096 ... 6757 45400064 154381644 154389835 4096 6758 45404160 154377548 154385739 4096 6759 45408256 252951571 154381643 73728 eof This happens because XFS_ALLOCTYPE_THIS_BNO allocation fails (the last extent in the file cannot be further extended) so we fall back to XFS_ALLOCTYPE_NEAR_BNO allocation which picks end of a large free extent as the best place to continue the file. Since the chunk at the end of the free extent again cannot be further extended, this behavior repeats until the whole free extent is consumed in a reversed order. For data allocations this backward allocation isn't beneficial so make xfs_alloc_compute_diff() pick start of a free extent instead of its end for them. That avoids the backward allocation pattern. See thread at http://oss.sgi.com/archives/xfs/2013-03/msg00144.html for more details about the reproduction case and why this solution was chosen. Based on idea by Dave Chinner . CC: Dave Chinner Signed-off-by: Jan Kara Reviewed-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_alloc.c | 24 ++++++++++++++++++------ 1 file changed, 18 insertions(+), 6 deletions(-) diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index 5673bcfda2f0..71596e57283a 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -175,6 +175,7 @@ xfs_alloc_compute_diff( xfs_agblock_t wantbno, /* target starting block */ xfs_extlen_t wantlen, /* target length */ xfs_extlen_t alignment, /* target alignment */ + char userdata, /* are we allocating data? */ xfs_agblock_t freebno, /* freespace's starting block */ xfs_extlen_t freelen, /* freespace's length */ xfs_agblock_t *newbnop) /* result: best start block from free */ @@ -189,7 +190,14 @@ xfs_alloc_compute_diff( ASSERT(freelen >= wantlen); freeend = freebno + freelen; wantend = wantbno + wantlen; - if (freebno >= wantbno) { + /* + * We want to allocate from the start of a free extent if it is past + * the desired block or if we are allocating user data and the free + * extent is before desired block. The second case is there to allow + * for contiguous allocation from the remaining free space if the file + * grows in the short term. + */ + if (freebno >= wantbno || (userdata && freeend < wantend)) { if ((newbno1 = roundup(freebno, alignment)) >= freeend) newbno1 = NULLAGBLOCK; } else if (freeend >= wantend && alignment > 1) { @@ -805,7 +813,8 @@ xfs_alloc_find_best_extent( xfs_alloc_fix_len(args); sdiff = xfs_alloc_compute_diff(args->agbno, args->len, - args->alignment, *sbnoa, + args->alignment, + args->userdata, *sbnoa, *slena, &new); /* @@ -976,7 +985,8 @@ restart: if (args->len < blen) continue; ltdiff = xfs_alloc_compute_diff(args->agbno, args->len, - args->alignment, ltbnoa, ltlena, <new); + args->alignment, args->userdata, ltbnoa, + ltlena, <new); if (ltnew != NULLAGBLOCK && (args->len > blen || ltdiff < bdiff)) { bdiff = ltdiff; @@ -1128,7 +1138,8 @@ restart: args->len = XFS_EXTLEN_MIN(ltlena, args->maxlen); xfs_alloc_fix_len(args); ltdiff = xfs_alloc_compute_diff(args->agbno, args->len, - args->alignment, ltbnoa, ltlena, <new); + args->alignment, args->userdata, ltbnoa, + ltlena, <new); error = xfs_alloc_find_best_extent(args, &bno_cur_lt, &bno_cur_gt, @@ -1144,7 +1155,8 @@ restart: args->len = XFS_EXTLEN_MIN(gtlena, args->maxlen); xfs_alloc_fix_len(args); gtdiff = xfs_alloc_compute_diff(args->agbno, args->len, - args->alignment, gtbnoa, gtlena, >new); + args->alignment, args->userdata, gtbnoa, + gtlena, >new); error = xfs_alloc_find_best_extent(args, &bno_cur_gt, &bno_cur_lt, @@ -1203,7 +1215,7 @@ restart: } rlen = args->len; (void)xfs_alloc_compute_diff(args->agbno, rlen, args->alignment, - ltbnoa, ltlena, <new); + args->userdata, ltbnoa, ltlena, <new); ASSERT(ltnew >= ltbno); ASSERT(ltnew + rlen <= ltbnoa + ltlena); ASSERT(ltnew + rlen <= be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length)); From 49b137cbbcc836ef231866c137d24f42c42bb483 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 20 May 2013 09:51:08 +1000 Subject: [PATCH 02/61] xfs: fix sub-page blocksize data integrity writes FSX on 512 byte block size filesystems has been failing for some time with corrupted data. The fault dates back to the change in the writeback data integrity algorithm that uses a mark-and-sweep approach to avoid data writeback livelocks. Unfortunately, a side effect of this mark-and-sweep approach is that each page will only be written once for a data integrity sync, and there is a condition in writeback in XFS where a page may require two writeback attempts to be fully written. As a result of the high level change, we now only get a partial page writeback during the integrity sync because the first pass through writeback clears the mark left on the page index to tell writeback that the page needs writeback.... The cause is writing a partial page in the clustering code. This can happen when a mapping boundary falls in the middle of a page - we end up writing back the first part of the page that the mapping covers, but then never revisit the page to have the remainder mapped and written. The fix is simple - if the mapping boundary falls inside a page, then simple abort clustering without touching the page. This means that the next ->writepage entry that write_cache_pages() will make is the page we aborted on, and xfs_vm_writepage() will map all sections of the page correctly. This behaviour is also optimal for non-data integrity writes, as it results in contiguous sequential writeback of the file rather than missing small holes and having to write them a "random" writes in a future pass. With this fix, all the fsx tests in xfstests now pass on a 512 byte block size filesystem on a 4k page machine. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Ben Myers --- fs/xfs/xfs_aops.c | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index 2b2691b73428..41a695048be7 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -725,6 +725,25 @@ xfs_convert_page( (xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT, i_size_read(inode)); + /* + * If the current map does not span the entire page we are about to try + * to write, then give up. The only way we can write a page that spans + * multiple mappings in a single writeback iteration is via the + * xfs_vm_writepage() function. Data integrity writeback requires the + * entire page to be written in a single attempt, otherwise the part of + * the page we don't write here doesn't get written as part of the data + * integrity sync. + * + * For normal writeback, we also don't attempt to write partial pages + * here as it simply means that write_cache_pages() will see it under + * writeback and ignore the page until some point in the future, at + * which time this will be the only page in the file that needs + * writeback. Hence for more optimal IO patterns, we should always + * avoid partial page writeback due to multiple mappings on a page here. + */ + if (!xfs_imap_valid(inode, imap, end_offset)) + goto fail_unlock_page; + len = 1 << inode->i_blkbits; p_offset = min_t(unsigned long, end_offset & (PAGE_CACHE_SIZE - 1), PAGE_CACHE_SIZE); From 28ca489c63e9aceed8801d2f82d731b3c9aa50f5 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 20 May 2013 09:51:09 +1000 Subject: [PATCH 03/61] xfs: fix rounding in xfs_free_file_space The offset passed into xfs_free_file_space() needs to be rounded down to a certain size, but the rounding mask is built by a 32 bit variable. Hence the mask will always mask off the upper 32 bits of the offset and lead to incorrect writeback and invalidation ranges. This is not actually exposed as a bug because we writeback and invalidate from the rounded offset to the end of the file, and hence the offset we are actually punching a hole out of will always be covered by the code. This needs fixing, however, if we ever want to use exact ranges for writeback/invalidation here... Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Ben Myers --- fs/xfs/xfs_vnodeops.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 1501f4fa51a6..0176bb21f09a 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -1453,7 +1453,7 @@ xfs_free_file_space( xfs_mount_t *mp; int nimap; uint resblks; - uint rounding; + xfs_off_t rounding; int rt; xfs_fileoff_t startoffset_fsb; xfs_trans_t *tp; @@ -1482,7 +1482,7 @@ xfs_free_file_space( inode_dio_wait(VFS_I(ip)); } - rounding = max_t(uint, 1 << mp->m_sb.sb_blocklog, PAGE_CACHE_SIZE); + rounding = max_t(xfs_off_t, 1 << mp->m_sb.sb_blocklog, PAGE_CACHE_SIZE); ioffset = offset & ~(rounding - 1); error = -filemap_write_and_wait_range(VFS_I(ip)->i_mapping, ioffset, -1); From 52c24ad39ff02d7bd73c92eb0c926fb44984a41d Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 20 May 2013 09:51:10 +1000 Subject: [PATCH 04/61] xfs: Don't reference the EFI after it is freed Checking the EFI for whether it is being released from recovery after we've already released the known active reference is a mistake worthy of a brown paper bag. Fix the (now) obvious use after free that it can cause. Reported-by: Dave Jones Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Ben Myers --- fs/xfs/xfs_extfree_item.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/fs/xfs/xfs_extfree_item.c b/fs/xfs/xfs_extfree_item.c index c0f375087efc..452920a3f03f 100644 --- a/fs/xfs/xfs_extfree_item.c +++ b/fs/xfs/xfs_extfree_item.c @@ -305,11 +305,12 @@ xfs_efi_release(xfs_efi_log_item_t *efip, { ASSERT(atomic_read(&efip->efi_next_extent) >= nextents); if (atomic_sub_and_test(nextents, &efip->efi_next_extent)) { - __xfs_efi_release(efip); - /* recovery needs us to drop the EFI reference, too */ if (test_bit(XFS_EFI_RECOVERED, &efip->efi_flags)) __xfs_efi_release(efip); + + __xfs_efi_release(efip); + /* efip may now have been freed, do not reference it again. */ } } From ac14876cf9255175bf3bdad645bf8aa2b8fb2d7c Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 20 May 2013 09:51:12 +1000 Subject: [PATCH 05/61] xfs: fix missing KM_NOFS tags to keep lockdep happy There are several places where we use KM_SLEEP allocation contexts and use the fact that they are called from transaction context to add KM_NOFS where appropriate. Unfortunately, there are several places where the code makes this assumption but can be called from outside transaction context but with filesystem locks held. These places need explicit KM_NOFS annotations to avoid lockdep complaining about reclaim contexts. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_buf.c | 2 +- fs/xfs/xfs_da_btree.c | 6 ++++-- fs/xfs/xfs_dir2_leaf.c | 2 +- fs/xfs/xfs_log_cil.c | 2 +- 4 files changed, 7 insertions(+), 5 deletions(-) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 82b70bda9f47..0d2554299688 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1649,7 +1649,7 @@ xfs_alloc_buftarg( { xfs_buftarg_t *btp; - btp = kmem_zalloc(sizeof(*btp), KM_SLEEP); + btp = kmem_zalloc(sizeof(*btp), KM_SLEEP | KM_NOFS); btp->bt_mount = mp; btp->bt_dev = bdev->bd_dev; diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c index 9b26a99ebfe9..41ea7e14a7b6 100644 --- a/fs/xfs/xfs_da_btree.c +++ b/fs/xfs/xfs_da_btree.c @@ -2464,7 +2464,8 @@ xfs_buf_map_from_irec( ASSERT(nirecs >= 1); if (nirecs > 1) { - map = kmem_zalloc(nirecs * sizeof(struct xfs_buf_map), KM_SLEEP); + map = kmem_zalloc(nirecs * sizeof(struct xfs_buf_map), + KM_SLEEP | KM_NOFS); if (!map) return ENOMEM; *mapp = map; @@ -2520,7 +2521,8 @@ xfs_dabuf_map( * Optimize the one-block case. */ if (nfsb != 1) - irecs = kmem_zalloc(sizeof(irec) * nfsb, KM_SLEEP); + irecs = kmem_zalloc(sizeof(irec) * nfsb, + KM_SLEEP | KM_NOFS); nirecs = nfsb; error = xfs_bmapi_read(dp, (xfs_fileoff_t)bno, nfsb, irecs, diff --git a/fs/xfs/xfs_dir2_leaf.c b/fs/xfs/xfs_dir2_leaf.c index 721ba2fe8e54..da71a1819d78 100644 --- a/fs/xfs/xfs_dir2_leaf.c +++ b/fs/xfs/xfs_dir2_leaf.c @@ -1336,7 +1336,7 @@ xfs_dir2_leaf_getdents( mp->m_sb.sb_blocksize); map_info = kmem_zalloc(offsetof(struct xfs_dir2_leaf_map_info, map) + (length * sizeof(struct xfs_bmbt_irec)), - KM_SLEEP); + KM_SLEEP | KM_NOFS); map_info->map_size = length; /* diff --git a/fs/xfs/xfs_log_cil.c b/fs/xfs/xfs_log_cil.c index e3d0b85d852b..d0833b54e55d 100644 --- a/fs/xfs/xfs_log_cil.c +++ b/fs/xfs/xfs_log_cil.c @@ -139,7 +139,7 @@ xlog_cil_prepare_log_vecs( new_lv = kmem_zalloc(sizeof(*new_lv) + niovecs * sizeof(struct xfs_log_iovec), - KM_SLEEP); + KM_SLEEP|KM_NOFS); /* The allocated iovec region lies beyond the log vector. */ new_lv->lv_iovecp = (struct xfs_log_iovec *)&new_lv[1]; From 72916fb8cbcf0c2928f56cdc2fbe8c7bf5517758 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 20 May 2013 09:51:13 +1000 Subject: [PATCH 06/61] xfs: xfs_da3_node_read_verify() doesn't handle XFS_ATTR3_LEAF_MAGIC Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_da_btree.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c index 41ea7e14a7b6..0b8b2a13cd24 100644 --- a/fs/xfs/xfs_da_btree.c +++ b/fs/xfs/xfs_da_btree.c @@ -270,6 +270,7 @@ xfs_da3_node_read_verify( break; return; case XFS_ATTR_LEAF_MAGIC: + case XFS_ATTR3_LEAF_MAGIC: bp->b_ops = &xfs_attr3_leaf_buf_ops; bp->b_ops->verify_read(bp); return; From b38958d715316031fe9ea0cc6c22043072a55f49 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 20 May 2013 09:51:14 +1000 Subject: [PATCH 07/61] xfs: xfs_attr_shortform_allfit() does not handle attr3 format. xfstests generic/117 fails with: XFS: Assertion failed: leaf->hdr.info.magic == cpu_to_be16(XFS_ATTR_LEAF_MAGIC) indicating a function that does not handle the attr3 format correctly. Fix it. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_attr_leaf.c | 24 +++++++++++++----------- 1 file changed, 13 insertions(+), 11 deletions(-) diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index 08d5457c948e..8eeb88fb3201 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -931,20 +931,22 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context) */ int xfs_attr_shortform_allfit( - struct xfs_buf *bp, - struct xfs_inode *dp) + struct xfs_buf *bp, + struct xfs_inode *dp) { - xfs_attr_leafblock_t *leaf; - xfs_attr_leaf_entry_t *entry; + struct xfs_attr_leafblock *leaf; + struct xfs_attr_leaf_entry *entry; xfs_attr_leaf_name_local_t *name_loc; - int bytes, i; + struct xfs_attr3_icleaf_hdr leafhdr; + int bytes; + int i; leaf = bp->b_addr; - ASSERT(leaf->hdr.info.magic == cpu_to_be16(XFS_ATTR_LEAF_MAGIC)); + xfs_attr3_leaf_hdr_from_disk(&leafhdr, leaf); + entry = xfs_attr3_leaf_entryp(leaf); - entry = &leaf->entries[0]; bytes = sizeof(struct xfs_attr_sf_hdr); - for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) { + for (i = 0; i < leafhdr.count; entry++, i++) { if (entry->flags & XFS_ATTR_INCOMPLETE) continue; /* don't copy partial entries */ if (!(entry->flags & XFS_ATTR_LOCAL)) @@ -954,15 +956,15 @@ xfs_attr_shortform_allfit( return(0); if (be16_to_cpu(name_loc->valuelen) >= XFS_ATTR_SF_ENTSIZE_MAX) return(0); - bytes += sizeof(struct xfs_attr_sf_entry)-1 + bytes += sizeof(struct xfs_attr_sf_entry) - 1 + name_loc->namelen + be16_to_cpu(name_loc->valuelen); } if ((dp->i_mount->m_flags & XFS_MOUNT_ATTR2) && (dp->i_d.di_format != XFS_DINODE_FMT_BTREE) && (bytes == sizeof(struct xfs_attr_sf_hdr))) - return(-1); - return(xfs_attr_shortform_bytesfit(dp, bytes)); + return -1; + return xfs_attr_shortform_bytesfit(dp, bytes); } /* From e461fcb194172b3f709e0b478d2ac1bdac7ab9a3 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 20 May 2013 09:51:16 +1000 Subject: [PATCH 08/61] xfs: remote attribute lookups require the value length When reading a remote attribute, to correctly calculate the length of the data buffer for CRC enable filesystems, we need to know the length of the attribute data. We get this information when we look up the attribute, but we don't store it in the args structure along with the other remote attr information we get from the lookup. Add this information to the args structure so we can use it appropriately. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_attr_leaf.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index 8eeb88fb3201..0bce1b348580 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -2332,9 +2332,10 @@ xfs_attr3_leaf_lookup_int( if (!xfs_attr_namesp_match(args->flags, entry->flags)) continue; args->index = probe; + args->valuelen = be32_to_cpu(name_rmt->valuelen); args->rmtblkno = be32_to_cpu(name_rmt->valueblk); args->rmtblkcnt = XFS_B_TO_FSB(args->dp->i_mount, - be32_to_cpu(name_rmt->valuelen)); + args->valuelen); return XFS_ERROR(EEXIST); } } From f648167f3ac79018c210112508c732ea9bf67c7b Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 21 May 2013 18:02:00 +1000 Subject: [PATCH 09/61] xfs: avoid nesting transactions in xfs_qm_scall_setqlim() Lockdep reports: ============================================= [ INFO: possible recursive locking detected ] 3.9.0+ #3 Not tainted --------------------------------------------- setquota/28368 is trying to acquire lock: (sb_internal){++++.?}, at: [] xfs_trans_alloc+0x26/0x50 but task is already holding lock: (sb_internal){++++.?}, at: [] xfs_trans_alloc+0x26/0x50 from xfs_qm_scall_setqlim()->xfs_dqread() when a dquot needs to be allocated. xfs_qm_scall_setqlim() is starting a transaction and then not passing it into xfs_qm_dqet() and so it starts it's own transaction when allocating the dquot. Splat! Fix this by not allocating the dquot in xfs_qm_scall_setqlim() inside the setqlim transaction. This requires getting the dquot first (and allocating it if necessary) then dropping and relocking the dquot before joining it to the setqlim transaction. Reported-by: Michael L. Semon Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_qm_syscalls.c | 46 +++++++++++++++++++++++----------------- 1 file changed, 26 insertions(+), 20 deletions(-) diff --git a/fs/xfs/xfs_qm_syscalls.c b/fs/xfs/xfs_qm_syscalls.c index c41190cad6e9..6cdf6ffc36a1 100644 --- a/fs/xfs/xfs_qm_syscalls.c +++ b/fs/xfs/xfs_qm_syscalls.c @@ -489,31 +489,36 @@ xfs_qm_scall_setqlim( if ((newlim->d_fieldmask & XFS_DQ_MASK) == 0) return 0; + /* + * We don't want to race with a quotaoff so take the quotaoff lock. + * We don't hold an inode lock, so there's nothing else to stop + * a quotaoff from happening. + */ + mutex_lock(&q->qi_quotaofflock); + + /* + * Get the dquot (locked) before we start, as we need to do a + * transaction to allocate it if it doesn't exist. Once we have the + * dquot, unlock it so we can start the next transaction safely. We hold + * a reference to the dquot, so it's safe to do this unlock/lock without + * it being reclaimed in the mean time. + */ + error = xfs_qm_dqget(mp, NULL, id, type, XFS_QMOPT_DQALLOC, &dqp); + if (error) { + ASSERT(error != ENOENT); + goto out_unlock; + } + xfs_dqunlock(dqp); + tp = xfs_trans_alloc(mp, XFS_TRANS_QM_SETQLIM); error = xfs_trans_reserve(tp, 0, XFS_QM_SETQLIM_LOG_RES(mp), 0, 0, XFS_DEFAULT_LOG_COUNT); if (error) { xfs_trans_cancel(tp, 0); - return (error); + goto out_rele; } - /* - * We don't want to race with a quotaoff so take the quotaoff lock. - * (We don't hold an inode lock, so there's nothing else to stop - * a quotaoff from happening). (XXXThis doesn't currently happen - * because we take the vfslock before calling xfs_qm_sysent). - */ - mutex_lock(&q->qi_quotaofflock); - - /* - * Get the dquot (locked), and join it to the transaction. - * Allocate the dquot if this doesn't exist. - */ - if ((error = xfs_qm_dqget(mp, NULL, id, type, XFS_QMOPT_DQALLOC, &dqp))) { - xfs_trans_cancel(tp, XFS_TRANS_ABORT); - ASSERT(error != ENOENT); - goto out_unlock; - } + xfs_dqlock(dqp); xfs_trans_dqjoin(tp, dqp); ddq = &dqp->q_core; @@ -621,9 +626,10 @@ xfs_qm_scall_setqlim( xfs_trans_log_dquot(tp, dqp); error = xfs_trans_commit(tp, 0); - xfs_qm_dqrele(dqp); - out_unlock: +out_rele: + xfs_qm_dqrele(dqp); +out_unlock: mutex_unlock(&q->qi_quotaofflock); return error; } From 90253cf142469a40f89f989904abf0a1e500e1a6 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 21 May 2013 18:02:01 +1000 Subject: [PATCH 10/61] xfs: remote attribute allocation may be contiguous When CRCs are enabled, there may be multiple allocations made if the headers cause a length overflow. This, however, does not mean that the number of headers required increases, as the second and subsequent extents may be contiguous with the previous extent. Hence when we map the extents to write the attribute data, we may end up with less extents than allocations made. Hence the assertion that we consume the number of headers we calculated in the allocation loop is incorrect and needs to be removed. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_attr_remote.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/fs/xfs/xfs_attr_remote.c b/fs/xfs/xfs_attr_remote.c index dee84466dcc9..aad95b08e76b 100644 --- a/fs/xfs/xfs_attr_remote.c +++ b/fs/xfs/xfs_attr_remote.c @@ -359,6 +359,11 @@ xfs_attr_rmtval_set( * into requiring more blocks. e.g. for 512 byte blocks, we'll * spill for another block every 9 headers we require in this * loop. + * + * Note that this can result in contiguous allocation of blocks, + * so we don't use all the space we allocate for headers as we + * have one less header for each contiguous allocation that + * occurs in the map/write loop below. */ if (crcs && blkcnt == 0) { int total_len; @@ -439,7 +444,6 @@ xfs_attr_rmtval_set( lblkno += map.br_blockcount; } ASSERT(valuelen == 0); - ASSERT(hdrcnt == 0); return 0; } From 913e96bc292e1bb248854686c79d6545ef3ee720 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 21 May 2013 18:02:02 +1000 Subject: [PATCH 11/61] xfs: remote attribute read too short Reading a maximally size remote attribute fails when CRCs are enabled with this verification error: XFS (vdb): remote attribute header does not match required off/len/owner) There are two reasons for this, the first being that the length of the buffer being read is determined from the args->rmtblkcnt which doesn't take into account CRC headers. Hence the mapped length ends up being too short and so we need to calculate it directly from the value length. The second is that the byte count of valid data within a buffer is capped by the length of the data and so doesn't take into account that the buffer might be longer due to headers. Hence we need to calculate the data space in the buffer first before calculating the actual byte count of data. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_attr_remote.c | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) diff --git a/fs/xfs/xfs_attr_remote.c b/fs/xfs/xfs_attr_remote.c index aad95b08e76b..bcdc07c4e8f4 100644 --- a/fs/xfs/xfs_attr_remote.c +++ b/fs/xfs/xfs_attr_remote.c @@ -52,9 +52,11 @@ xfs_attr3_rmt_blocks( struct xfs_mount *mp, int attrlen) { - int buflen = XFS_ATTR3_RMT_BUF_SPACE(mp, - mp->m_sb.sb_blocksize); - return (attrlen + buflen - 1) / buflen; + if (xfs_sb_version_hascrc(&mp->m_sb)) { + int buflen = XFS_ATTR3_RMT_BUF_SPACE(mp, mp->m_sb.sb_blocksize); + return (attrlen + buflen - 1) / buflen; + } + return XFS_B_TO_FSB(mp, attrlen); } static bool @@ -206,8 +208,9 @@ xfs_attr_rmtval_get( while (valuelen > 0) { nmap = ATTR_RMTVALUE_MAPSIZE; + blkcnt = xfs_attr3_rmt_blocks(mp, valuelen); error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno, - args->rmtblkcnt, map, &nmap, + blkcnt, map, &nmap, XFS_BMAPI_ATTRFORK); if (error) return error; @@ -227,8 +230,8 @@ xfs_attr_rmtval_get( if (error) return error; - byte_cnt = min_t(int, valuelen, BBTOB(bp->b_length)); - byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, byte_cnt); + byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, BBTOB(bp->b_length)); + byte_cnt = min_t(int, valuelen, byte_cnt); src = bp->b_addr; if (xfs_sb_version_hascrc(&mp->m_sb)) { From 4af3644c9a53eb2f1ecf69cc53576561b64be4c6 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 21 May 2013 18:02:03 +1000 Subject: [PATCH 12/61] xfs: remote attribute tail zeroing does too much When an attribute data does not fill then entire remote block, we zero the remaining part of the buffer. This, however, needs to take into account that the buffer has a header, and so the offset where zeroing starts and the length of zeroing need to take this into account. Otherwise we end up with zeros over the end of the attribute value when CRCs are enabled. While there, make sure we only ask to map an extent that covers the remaining range of the attribute, rather than asking every time for the full length of remote data. If the remote attribute blocks are contiguous with other parts of the attribute tree, it will map those blocks as well and we can potentially zero them incorrectly. We can also get buffer size mistmatches when trying to read or remove the remote attribute, and this can lead to not finding the correct buffer when looking it up in cache. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_attr_remote.c | 39 +++++++++++++++++++-------------------- 1 file changed, 19 insertions(+), 20 deletions(-) diff --git a/fs/xfs/xfs_attr_remote.c b/fs/xfs/xfs_attr_remote.c index bcdc07c4e8f4..e207bf0004b6 100644 --- a/fs/xfs/xfs_attr_remote.c +++ b/fs/xfs/xfs_attr_remote.c @@ -296,10 +296,7 @@ xfs_attr_rmtval_set( * and we may not need that many, so we have to handle this when * allocating the blocks below. */ - if (!crcs) - blkcnt = XFS_B_TO_FSB(mp, args->valuelen); - else - blkcnt = xfs_attr3_rmt_blocks(mp, args->valuelen); + blkcnt = xfs_attr3_rmt_blocks(mp, args->valuelen); error = xfs_bmap_first_unused(args->trans, args->dp, blkcnt, &lfileoff, XFS_ATTR_FORK); @@ -394,8 +391,11 @@ xfs_attr_rmtval_set( */ lblkno = args->rmtblkno; valuelen = args->valuelen; + blkcnt = args->rmtblkcnt; while (valuelen > 0) { int byte_cnt; + int hdr_size; + int dblkcnt; char *buf; /* @@ -404,7 +404,7 @@ xfs_attr_rmtval_set( xfs_bmap_init(args->flist, args->firstblock); nmap = 1; error = xfs_bmapi_read(dp, (xfs_fileoff_t)lblkno, - args->rmtblkcnt, &map, &nmap, + blkcnt, &map, &nmap, XFS_BMAPI_ATTRFORK); if (error) return(error); @@ -413,26 +413,25 @@ xfs_attr_rmtval_set( (map.br_startblock != HOLESTARTBLOCK)); dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock), - blkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount); + dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount); - bp = xfs_buf_get(mp->m_ddev_targp, dblkno, blkcnt, 0); + bp = xfs_buf_get(mp->m_ddev_targp, dblkno, dblkcnt, 0); if (!bp) return ENOMEM; bp->b_ops = &xfs_attr3_rmt_buf_ops; - - byte_cnt = BBTOB(bp->b_length); - byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, byte_cnt); - if (valuelen < byte_cnt) - byte_cnt = valuelen; - buf = bp->b_addr; - buf += xfs_attr3_rmt_hdr_set(mp, dp->i_ino, offset, - byte_cnt, bp); - memcpy(buf, src, byte_cnt); - if (byte_cnt < BBTOB(bp->b_length)) - xfs_buf_zero(bp, byte_cnt, - BBTOB(bp->b_length) - byte_cnt); + byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, BBTOB(bp->b_length)); + byte_cnt = min_t(int, valuelen, byte_cnt); + hdr_size = xfs_attr3_rmt_hdr_set(mp, dp->i_ino, offset, + byte_cnt, bp); + ASSERT(hdr_size + byte_cnt <= BBTOB(bp->b_length)); + + memcpy(buf + hdr_size, src, byte_cnt); + + if (byte_cnt + hdr_size < BBTOB(bp->b_length)) + xfs_buf_zero(bp, byte_cnt + hdr_size, + BBTOB(bp->b_length) - byte_cnt - hdr_size); error = xfs_bwrite(bp); /* GROT: NOTE: synchronous write */ xfs_buf_relse(bp); @@ -442,9 +441,9 @@ xfs_attr_rmtval_set( src += byte_cnt; valuelen -= byte_cnt; offset += byte_cnt; - hdrcnt--; lblkno += map.br_blockcount; + blkcnt -= map.br_blockcount; } ASSERT(valuelen == 0); return 0; From 6863ef8449f1908c19f43db572e4474f24a1e9da Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 21 May 2013 18:02:04 +1000 Subject: [PATCH 13/61] xfs: correctly map remote attr buffers during removal If we don't map the buffers correctly (same as for get/set operations) then the incore buffer lookup will fail. If a block number matches but a length is wrong, then debug kernels will ASSERT fail in _xfs_buf_find() due to the length mismatch. Ensure that we map the buffers correctly by basing the length of the buffer on the attribute data length rather than the remote block count. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_attr_remote.c | 27 ++++++++++++++++++--------- 1 file changed, 18 insertions(+), 9 deletions(-) diff --git a/fs/xfs/xfs_attr_remote.c b/fs/xfs/xfs_attr_remote.c index e207bf0004b6..d8bcb2d742d1 100644 --- a/fs/xfs/xfs_attr_remote.c +++ b/fs/xfs/xfs_attr_remote.c @@ -468,19 +468,25 @@ xfs_attr_rmtval_remove(xfs_da_args_t *args) mp = args->dp->i_mount; /* - * Roll through the "value", invalidating the attribute value's - * blocks. + * Roll through the "value", invalidating the attribute value's blocks. + * Note that args->rmtblkcnt is the minimum number of data blocks we'll + * see for a CRC enabled remote attribute. Each extent will have a + * header, and so we may have more blocks than we realise here. If we + * fail to map the blocks correctly, we'll have problems with the buffer + * lookups. */ lblkno = args->rmtblkno; - valuelen = args->rmtblkcnt; + valuelen = args->valuelen; + blkcnt = xfs_attr3_rmt_blocks(mp, valuelen); while (valuelen > 0) { + int dblkcnt; + /* * Try to remember where we decided to put the value. */ nmap = 1; error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno, - args->rmtblkcnt, &map, &nmap, - XFS_BMAPI_ATTRFORK); + blkcnt, &map, &nmap, XFS_BMAPI_ATTRFORK); if (error) return(error); ASSERT(nmap == 1); @@ -488,28 +494,31 @@ xfs_attr_rmtval_remove(xfs_da_args_t *args) (map.br_startblock != HOLESTARTBLOCK)); dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock), - blkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount); + dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount); /* * If the "remote" value is in the cache, remove it. */ - bp = xfs_incore(mp->m_ddev_targp, dblkno, blkcnt, XBF_TRYLOCK); + bp = xfs_incore(mp->m_ddev_targp, dblkno, dblkcnt, XBF_TRYLOCK); if (bp) { xfs_buf_stale(bp); xfs_buf_relse(bp); bp = NULL; } - valuelen -= map.br_blockcount; + valuelen -= XFS_ATTR3_RMT_BUF_SPACE(mp, + XFS_FSB_TO_B(mp, map.br_blockcount)); lblkno += map.br_blockcount; + blkcnt -= map.br_blockcount; + blkcnt = max(blkcnt, xfs_attr3_rmt_blocks(mp, valuelen)); } /* * Keep de-allocating extents until the remote-value region is gone. */ + blkcnt = lblkno - args->rmtblkno; lblkno = args->rmtblkno; - blkcnt = args->rmtblkcnt; done = 0; while (!done) { xfs_bmap_init(args->flist, args->firstblock); From 8517de2a81da830f5d90da66b4799f4040c76dc9 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 21 May 2013 18:02:05 +1000 Subject: [PATCH 14/61] xfs: fully initialise temp leaf in xfs_attr3_leaf_unbalance xfs_attr3_leaf_unbalance() uses a temporary buffer for recombining the entries in two leaves when the destination leaf requires compaction. The temporary buffer ends up being copied back over the original destination buffer, so the header in the temporary buffer needs to contain all the information that is in the destination buffer. To make sure the temporary buffer is fully initialised, once we've set up the temporary incore header appropriately, write is back to the temporary buffer before starting to move entries around. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_attr_leaf.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index 0bce1b348580..79ece72976ae 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -2181,14 +2181,24 @@ xfs_attr3_leaf_unbalance( struct xfs_attr_leafblock *tmp_leaf; struct xfs_attr3_icleaf_hdr tmphdr; - tmp_leaf = kmem_alloc(state->blocksize, KM_SLEEP); - memset(tmp_leaf, 0, state->blocksize); - memset(&tmphdr, 0, sizeof(tmphdr)); + tmp_leaf = kmem_zalloc(state->blocksize, KM_SLEEP); + /* + * Copy the header into the temp leaf so that all the stuff + * not in the incore header is present and gets copied back in + * once we've moved all the entries. + */ + memcpy(tmp_leaf, save_leaf, xfs_attr3_leaf_hdr_size(save_leaf)); + + memset(&tmphdr, 0, sizeof(tmphdr)); tmphdr.magic = savehdr.magic; tmphdr.forw = savehdr.forw; tmphdr.back = savehdr.back; tmphdr.firstused = state->blocksize; + + /* write the header to the temp buffer to initialise it */ + xfs_attr3_leaf_hdr_to_disk(tmp_leaf, &tmphdr); + if (xfs_attr3_leaf_order(save_blk->bp, &savehdr, drop_blk->bp, &drophdr)) { xfs_attr3_leaf_moveents(drop_leaf, &drophdr, 0, From d4c712bcf26a25c2b67c90e44e0b74c7993b5334 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 21 May 2013 18:02:06 +1000 Subject: [PATCH 15/61] xfs: fully initialise temp leaf in xfs_attr3_leaf_compact xfs_attr3_leaf_compact() uses a temporary buffer for compacting the the entries in a leaf. It copies the the original buffer into the temporary buffer, then zeros the original buffer completely. It then copies the entries back into the original buffer. However, the original buffer has not been correctly initialised, and so the movement of the entries goes horribly wrong. Make sure the zeroed destination buffer is fully initialised, and once we've set up the destination incore header appropriately, write is back to the buffer before starting to move entries around. While debugging this, the _d/_s prefixes weren't sufficient to remind me what buffer was what, so rename then all _src/_dst. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_attr_leaf.c | 42 ++++++++++++++++++++++++++---------------- 1 file changed, 26 insertions(+), 16 deletions(-) diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index 79ece72976ae..5b03d15b707b 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -1445,11 +1445,12 @@ xfs_attr3_leaf_add_work( STATIC void xfs_attr3_leaf_compact( struct xfs_da_args *args, - struct xfs_attr3_icleaf_hdr *ichdr_d, + struct xfs_attr3_icleaf_hdr *ichdr_dst, struct xfs_buf *bp) { - xfs_attr_leafblock_t *leaf_s, *leaf_d; - struct xfs_attr3_icleaf_hdr ichdr_s; + struct xfs_attr_leafblock *leaf_src; + struct xfs_attr_leafblock *leaf_dst; + struct xfs_attr3_icleaf_hdr ichdr_src; struct xfs_trans *trans = args->trans; struct xfs_mount *mp = trans->t_mountp; char *tmpbuffer; @@ -1457,29 +1458,38 @@ xfs_attr3_leaf_compact( trace_xfs_attr_leaf_compact(args); tmpbuffer = kmem_alloc(XFS_LBSIZE(mp), KM_SLEEP); - ASSERT(tmpbuffer != NULL); memcpy(tmpbuffer, bp->b_addr, XFS_LBSIZE(mp)); memset(bp->b_addr, 0, XFS_LBSIZE(mp)); + leaf_src = (xfs_attr_leafblock_t *)tmpbuffer; + leaf_dst = bp->b_addr; /* - * Copy basic information + * Copy the on-disk header back into the destination buffer to ensure + * all the information in the header that is not part of the incore + * header structure is preserved. */ - leaf_s = (xfs_attr_leafblock_t *)tmpbuffer; - leaf_d = bp->b_addr; - ichdr_s = *ichdr_d; /* struct copy */ - ichdr_d->firstused = XFS_LBSIZE(mp); - ichdr_d->usedbytes = 0; - ichdr_d->count = 0; - ichdr_d->holes = 0; - ichdr_d->freemap[0].base = xfs_attr3_leaf_hdr_size(leaf_s); - ichdr_d->freemap[0].size = ichdr_d->firstused - ichdr_d->freemap[0].base; + memcpy(bp->b_addr, tmpbuffer, xfs_attr3_leaf_hdr_size(leaf_src)); + + /* Initialise the incore headers */ + ichdr_src = *ichdr_dst; /* struct copy */ + ichdr_dst->firstused = XFS_LBSIZE(mp); + ichdr_dst->usedbytes = 0; + ichdr_dst->count = 0; + ichdr_dst->holes = 0; + ichdr_dst->freemap[0].base = xfs_attr3_leaf_hdr_size(leaf_src); + ichdr_dst->freemap[0].size = ichdr_dst->firstused - + ichdr_dst->freemap[0].base; + + + /* write the header back to initialise the underlying buffer */ + xfs_attr3_leaf_hdr_to_disk(leaf_dst, ichdr_dst); /* * Copy all entry's in the same (sorted) order, * but allocate name/value pairs packed and in sequence. */ - xfs_attr3_leaf_moveents(leaf_s, &ichdr_s, 0, leaf_d, ichdr_d, 0, - ichdr_s.count, mp); + xfs_attr3_leaf_moveents(leaf_src, &ichdr_src, 0, leaf_dst, ichdr_dst, 0, + ichdr_src.count, mp); /* * this logs the entire buffer, but the caller must write the header * back to the buffer when it is finished modifying it. From ad1858d77771172e08016890f0eb2faedec3ecee Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 21 May 2013 18:02:08 +1000 Subject: [PATCH 16/61] xfs: rework remote attr CRCs Note: this changes the on-disk remote attribute format. I assert that this is OK to do as CRCs are marked experimental and the first kernel it is included in has not yet reached release yet. Further, the userspace utilities are still evolving and so anyone using this stuff right now is a developer or tester using volatile filesystems for testing this feature. Hence changing the format right now to save longer term pain is the right thing to do. The fundamental change is to move from a header per extent in the attribute to a header per filesytem block in the attribute. This means there are more header blocks and the parsing of the attribute data is slightly more complex, but it has the advantage that we always know the size of the attribute on disk based on the length of the data it contains. This is where the header-per-extent method has problems. We don't know the size of the attribute on disk without first knowing how many extents are used to hold it. And we can't tell from a mapping lookup, either, because remote attributes can be allocated contiguously with other attribute blocks and so there is no obvious way of determining the actual size of the atribute on disk short of walking and mapping buffers. The problem with this approach is that if we map a buffer incorrectly (e.g. we make the last buffer for the attribute data too long), we then get buffer cache lookup failure when we map it correctly. i.e. we get a size mismatch on lookup. This is not necessarily fatal, but it's a cache coherency problem that can lead to returning the wrong data to userspace or writing the wrong data to disk. And debug kernels will assert fail if this occurs. I found lots of niggly little problems trying to fix this issue on a 4k block size filesystem, finally getting it to pass with lots of fixes. The thing is, 1024 byte filesystems still failed, and it was getting really complex handling all the corner cases that were showing up. And there were clearly more that I hadn't found yet. It is complex, fragile code, and if we don't fix it now, it will be complex, fragile code forever more. Hence the simple fix is to add a header to each filesystem block. This gives us the same relationship between the attribute data length and the number of blocks on disk as we have without CRCs - it's a linear mapping and doesn't require us to guess anything. It is simple to implement, too - the remote block count calculated at lookup time can be used by the remote attribute set/get/remove code without modification for both CRC and non-CRC filesystems. The world becomes sane again. Because the copy-in and copy-out now need to iterate over each filesystem block, I moved them into helper functions so we separate the block mapping and buffer manupulations from the attribute data and CRC header manipulations. The code becomes much clearer as a result, and it is a lot easier to understand and debug. It also appears to be much more robust - once it worked on 4k block size filesystems, it has worked without failure on 1k block size filesystems, too. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_attr_leaf.c | 13 +- fs/xfs/xfs_attr_remote.c | 381 +++++++++++++++++++++++---------------- fs/xfs/xfs_attr_remote.h | 10 + fs/xfs/xfs_buf.c | 1 + 4 files changed, 247 insertions(+), 158 deletions(-) diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index 5b03d15b707b..d788302e506a 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -1412,7 +1412,7 @@ xfs_attr3_leaf_add_work( name_rmt->valuelen = 0; name_rmt->valueblk = 0; args->rmtblkno = 1; - args->rmtblkcnt = XFS_B_TO_FSB(mp, args->valuelen); + args->rmtblkcnt = xfs_attr3_rmt_blocks(mp, args->valuelen); } xfs_trans_log_buf(args->trans, bp, XFS_DA_LOGRANGE(leaf, xfs_attr3_leaf_name(leaf, args->index), @@ -2354,8 +2354,9 @@ xfs_attr3_leaf_lookup_int( args->index = probe; args->valuelen = be32_to_cpu(name_rmt->valuelen); args->rmtblkno = be32_to_cpu(name_rmt->valueblk); - args->rmtblkcnt = XFS_B_TO_FSB(args->dp->i_mount, - args->valuelen); + args->rmtblkcnt = xfs_attr3_rmt_blocks( + args->dp->i_mount, + args->valuelen); return XFS_ERROR(EEXIST); } } @@ -2406,7 +2407,8 @@ xfs_attr3_leaf_getvalue( ASSERT(memcmp(args->name, name_rmt->name, args->namelen) == 0); valuelen = be32_to_cpu(name_rmt->valuelen); args->rmtblkno = be32_to_cpu(name_rmt->valueblk); - args->rmtblkcnt = XFS_B_TO_FSB(args->dp->i_mount, valuelen); + args->rmtblkcnt = xfs_attr3_rmt_blocks(args->dp->i_mount, + valuelen); if (args->flags & ATTR_KERNOVAL) { args->valuelen = valuelen; return 0; @@ -2732,7 +2734,8 @@ xfs_attr3_leaf_list_int( args.valuelen = valuelen; args.value = kmem_alloc(valuelen, KM_SLEEP | KM_NOFS); args.rmtblkno = be32_to_cpu(name_rmt->valueblk); - args.rmtblkcnt = XFS_B_TO_FSB(args.dp->i_mount, valuelen); + args.rmtblkcnt = xfs_attr3_rmt_blocks( + args.dp->i_mount, valuelen); retval = xfs_attr_rmtval_get(&args); if (retval) return retval; diff --git a/fs/xfs/xfs_attr_remote.c b/fs/xfs/xfs_attr_remote.c index d8bcb2d742d1..ef6b0c124528 100644 --- a/fs/xfs/xfs_attr_remote.c +++ b/fs/xfs/xfs_attr_remote.c @@ -47,7 +47,7 @@ * Each contiguous block has a header, so it is not just a simple attribute * length to FSB conversion. */ -static int +int xfs_attr3_rmt_blocks( struct xfs_mount *mp, int attrlen) @@ -59,12 +59,43 @@ xfs_attr3_rmt_blocks( return XFS_B_TO_FSB(mp, attrlen); } +/* + * Checking of the remote attribute header is split into two parts. The verifier + * does CRC, location and bounds checking, the unpacking function checks the + * attribute parameters and owner. + */ +static bool +xfs_attr3_rmt_hdr_ok( + struct xfs_mount *mp, + void *ptr, + xfs_ino_t ino, + uint32_t offset, + uint32_t size, + xfs_daddr_t bno) +{ + struct xfs_attr3_rmt_hdr *rmt = ptr; + + if (bno != be64_to_cpu(rmt->rm_blkno)) + return false; + if (offset != be32_to_cpu(rmt->rm_offset)) + return false; + if (size != be32_to_cpu(rmt->rm_bytes)) + return false; + if (ino != be64_to_cpu(rmt->rm_owner)) + return false; + + /* ok */ + return true; +} + static bool xfs_attr3_rmt_verify( - struct xfs_buf *bp) + struct xfs_mount *mp, + void *ptr, + int fsbsize, + xfs_daddr_t bno) { - struct xfs_mount *mp = bp->b_target->bt_mount; - struct xfs_attr3_rmt_hdr *rmt = bp->b_addr; + struct xfs_attr3_rmt_hdr *rmt = ptr; if (!xfs_sb_version_hascrc(&mp->m_sb)) return false; @@ -72,7 +103,9 @@ xfs_attr3_rmt_verify( return false; if (!uuid_equal(&rmt->rm_uuid, &mp->m_sb.sb_uuid)) return false; - if (bp->b_bn != be64_to_cpu(rmt->rm_blkno)) + if (be64_to_cpu(rmt->rm_blkno) != bno) + return false; + if (be32_to_cpu(rmt->rm_bytes) > fsbsize - sizeof(*rmt)) return false; if (be32_to_cpu(rmt->rm_offset) + be32_to_cpu(rmt->rm_bytes) >= XATTR_SIZE_MAX) @@ -88,17 +121,40 @@ xfs_attr3_rmt_read_verify( struct xfs_buf *bp) { struct xfs_mount *mp = bp->b_target->bt_mount; + char *ptr; + int len; + bool corrupt = false; + xfs_daddr_t bno; /* no verification of non-crc buffers */ if (!xfs_sb_version_hascrc(&mp->m_sb)) return; - if (!xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_ATTR3_RMT_CRC_OFF) || - !xfs_attr3_rmt_verify(bp)) { + ptr = bp->b_addr; + bno = bp->b_bn; + len = BBTOB(bp->b_length); + ASSERT(len >= XFS_LBSIZE(mp)); + + while (len > 0) { + if (!xfs_verify_cksum(ptr, XFS_LBSIZE(mp), + XFS_ATTR3_RMT_CRC_OFF)) { + corrupt = true; + break; + } + if (!xfs_attr3_rmt_verify(mp, ptr, XFS_LBSIZE(mp), bno)) { + corrupt = true; + break; + } + len -= XFS_LBSIZE(mp); + ptr += XFS_LBSIZE(mp); + bno += mp->m_bsize; + } + + if (corrupt) { XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); - } + } else + ASSERT(len == 0); } static void @@ -107,23 +163,39 @@ xfs_attr3_rmt_write_verify( { struct xfs_mount *mp = bp->b_target->bt_mount; struct xfs_buf_log_item *bip = bp->b_fspriv; + char *ptr; + int len; + xfs_daddr_t bno; /* no verification of non-crc buffers */ if (!xfs_sb_version_hascrc(&mp->m_sb)) return; - if (!xfs_attr3_rmt_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); - xfs_buf_ioerror(bp, EFSCORRUPTED); - return; - } + ptr = bp->b_addr; + bno = bp->b_bn; + len = BBTOB(bp->b_length); + ASSERT(len >= XFS_LBSIZE(mp)); - if (bip) { - struct xfs_attr3_rmt_hdr *rmt = bp->b_addr; - rmt->rm_lsn = cpu_to_be64(bip->bli_item.li_lsn); + while (len > 0) { + if (!xfs_attr3_rmt_verify(mp, ptr, XFS_LBSIZE(mp), bno)) { + XFS_CORRUPTION_ERROR(__func__, + XFS_ERRLEVEL_LOW, mp, bp->b_addr); + xfs_buf_ioerror(bp, EFSCORRUPTED); + return; + } + if (bip) { + struct xfs_attr3_rmt_hdr *rmt; + + rmt = (struct xfs_attr3_rmt_hdr *)ptr; + rmt->rm_lsn = cpu_to_be64(bip->bli_item.li_lsn); + } + xfs_update_cksum(ptr, XFS_LBSIZE(mp), XFS_ATTR3_RMT_CRC_OFF); + + len -= XFS_LBSIZE(mp); + ptr += XFS_LBSIZE(mp); + bno += mp->m_bsize; } - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_ATTR3_RMT_CRC_OFF); + ASSERT(len == 0); } const struct xfs_buf_ops xfs_attr3_rmt_buf_ops = { @@ -131,15 +203,16 @@ const struct xfs_buf_ops xfs_attr3_rmt_buf_ops = { .verify_write = xfs_attr3_rmt_write_verify, }; -static int +STATIC int xfs_attr3_rmt_hdr_set( struct xfs_mount *mp, + void *ptr, xfs_ino_t ino, uint32_t offset, uint32_t size, - struct xfs_buf *bp) + xfs_daddr_t bno) { - struct xfs_attr3_rmt_hdr *rmt = bp->b_addr; + struct xfs_attr3_rmt_hdr *rmt = ptr; if (!xfs_sb_version_hascrc(&mp->m_sb)) return 0; @@ -149,36 +222,107 @@ xfs_attr3_rmt_hdr_set( rmt->rm_bytes = cpu_to_be32(size); uuid_copy(&rmt->rm_uuid, &mp->m_sb.sb_uuid); rmt->rm_owner = cpu_to_be64(ino); - rmt->rm_blkno = cpu_to_be64(bp->b_bn); - bp->b_ops = &xfs_attr3_rmt_buf_ops; + rmt->rm_blkno = cpu_to_be64(bno); return sizeof(struct xfs_attr3_rmt_hdr); } /* - * Checking of the remote attribute header is split into two parts. the verifier - * does CRC, location and bounds checking, the unpacking function checks the - * attribute parameters and owner. + * Helper functions to copy attribute data in and out of the one disk extents */ -static bool -xfs_attr3_rmt_hdr_ok( - struct xfs_mount *mp, - xfs_ino_t ino, - uint32_t offset, - uint32_t size, - struct xfs_buf *bp) +STATIC int +xfs_attr_rmtval_copyout( + struct xfs_mount *mp, + struct xfs_buf *bp, + xfs_ino_t ino, + int *offset, + int *valuelen, + char **dst) { - struct xfs_attr3_rmt_hdr *rmt = bp->b_addr; + char *src = bp->b_addr; + xfs_daddr_t bno = bp->b_bn; + int len = BBTOB(bp->b_length); - if (offset != be32_to_cpu(rmt->rm_offset)) - return false; - if (size != be32_to_cpu(rmt->rm_bytes)) - return false; - if (ino != be64_to_cpu(rmt->rm_owner)) - return false; + ASSERT(len >= XFS_LBSIZE(mp)); - /* ok */ - return true; + while (len > 0 && *valuelen > 0) { + int hdr_size = 0; + int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, XFS_LBSIZE(mp)); + + byte_cnt = min_t(int, *valuelen, byte_cnt); + + if (xfs_sb_version_hascrc(&mp->m_sb)) { + if (!xfs_attr3_rmt_hdr_ok(mp, src, ino, *offset, + byte_cnt, bno)) { + xfs_alert(mp, +"remote attribute header mismatch bno/off/len/owner (0x%llx/0x%x/Ox%x/0x%llx)", + bno, *offset, byte_cnt, ino); + return EFSCORRUPTED; + } + hdr_size = sizeof(struct xfs_attr3_rmt_hdr); + } + + memcpy(*dst, src + hdr_size, byte_cnt); + + /* roll buffer forwards */ + len -= XFS_LBSIZE(mp); + src += XFS_LBSIZE(mp); + bno += mp->m_bsize; + + /* roll attribute data forwards */ + *valuelen -= byte_cnt; + *dst += byte_cnt; + *offset += byte_cnt; + } + return 0; +} + +STATIC void +xfs_attr_rmtval_copyin( + struct xfs_mount *mp, + struct xfs_buf *bp, + xfs_ino_t ino, + int *offset, + int *valuelen, + char **src) +{ + char *dst = bp->b_addr; + xfs_daddr_t bno = bp->b_bn; + int len = BBTOB(bp->b_length); + + ASSERT(len >= XFS_LBSIZE(mp)); + + while (len > 0 && *valuelen > 0) { + int hdr_size; + int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, XFS_LBSIZE(mp)); + + byte_cnt = min(*valuelen, byte_cnt); + hdr_size = xfs_attr3_rmt_hdr_set(mp, dst, ino, *offset, + byte_cnt, bno); + + memcpy(dst + hdr_size, *src, byte_cnt); + + /* + * If this is the last block, zero the remainder of it. + * Check that we are actually the last block, too. + */ + if (byte_cnt + hdr_size < XFS_LBSIZE(mp)) { + ASSERT(*valuelen - byte_cnt == 0); + ASSERT(len == XFS_LBSIZE(mp)); + memset(dst + hdr_size + byte_cnt, 0, + XFS_LBSIZE(mp) - hdr_size - byte_cnt); + } + + /* roll buffer forwards */ + len -= XFS_LBSIZE(mp); + dst += XFS_LBSIZE(mp); + bno += mp->m_bsize; + + /* roll attribute data forwards */ + *valuelen -= byte_cnt; + *src += byte_cnt; + *offset += byte_cnt; + } } /* @@ -192,13 +336,12 @@ xfs_attr_rmtval_get( struct xfs_bmbt_irec map[ATTR_RMTVALUE_MAPSIZE]; struct xfs_mount *mp = args->dp->i_mount; struct xfs_buf *bp; - xfs_daddr_t dblkno; xfs_dablk_t lblkno = args->rmtblkno; - void *dst = args->value; + char *dst = args->value; int valuelen = args->valuelen; int nmap; int error; - int blkcnt; + int blkcnt = args->rmtblkcnt; int i; int offset = 0; @@ -208,7 +351,6 @@ xfs_attr_rmtval_get( while (valuelen > 0) { nmap = ATTR_RMTVALUE_MAPSIZE; - blkcnt = xfs_attr3_rmt_blocks(mp, valuelen); error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno, blkcnt, map, &nmap, XFS_BMAPI_ATTRFORK); @@ -217,45 +359,29 @@ xfs_attr_rmtval_get( ASSERT(nmap >= 1); for (i = 0; (i < nmap) && (valuelen > 0); i++) { - int byte_cnt; - char *src; + xfs_daddr_t dblkno; + int dblkcnt; ASSERT((map[i].br_startblock != DELAYSTARTBLOCK) && (map[i].br_startblock != HOLESTARTBLOCK)); dblkno = XFS_FSB_TO_DADDR(mp, map[i].br_startblock); - blkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount); + dblkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount); error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp, - dblkno, blkcnt, 0, &bp, + dblkno, dblkcnt, 0, &bp, &xfs_attr3_rmt_buf_ops); if (error) return error; - byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, BBTOB(bp->b_length)); - byte_cnt = min_t(int, valuelen, byte_cnt); - - src = bp->b_addr; - if (xfs_sb_version_hascrc(&mp->m_sb)) { - if (!xfs_attr3_rmt_hdr_ok(mp, args->dp->i_ino, - offset, byte_cnt, bp)) { - xfs_alert(mp, -"remote attribute header does not match required off/len/owner (0x%x/Ox%x,0x%llx)", - offset, byte_cnt, args->dp->i_ino); - xfs_buf_relse(bp); - return EFSCORRUPTED; - - } - - src += sizeof(struct xfs_attr3_rmt_hdr); - } - - memcpy(dst, src, byte_cnt); + error = xfs_attr_rmtval_copyout(mp, bp, args->dp->i_ino, + &offset, &valuelen, + &dst); xfs_buf_relse(bp); + if (error) + return error; - offset += byte_cnt; - dst += byte_cnt; - valuelen -= byte_cnt; - + /* roll attribute extent map forwards */ lblkno += map[i].br_blockcount; + blkcnt -= map[i].br_blockcount; } } ASSERT(valuelen == 0); @@ -273,17 +399,13 @@ xfs_attr_rmtval_set( struct xfs_inode *dp = args->dp; struct xfs_mount *mp = dp->i_mount; struct xfs_bmbt_irec map; - struct xfs_buf *bp; - xfs_daddr_t dblkno; xfs_dablk_t lblkno; xfs_fileoff_t lfileoff = 0; - void *src = args->value; + char *src = args->value; int blkcnt; int valuelen; int nmap; int error; - int hdrcnt = 0; - bool crcs = xfs_sb_version_hascrc(&mp->m_sb); int offset = 0; trace_xfs_attr_rmtval_set(args); @@ -292,21 +414,14 @@ xfs_attr_rmtval_set( * Find a "hole" in the attribute address space large enough for * us to drop the new attribute's value into. Because CRC enable * attributes have headers, we can't just do a straight byte to FSB - * conversion. We calculate the worst case block count in this case - * and we may not need that many, so we have to handle this when - * allocating the blocks below. + * conversion and have to take the header space into account. */ blkcnt = xfs_attr3_rmt_blocks(mp, args->valuelen); - error = xfs_bmap_first_unused(args->trans, args->dp, blkcnt, &lfileoff, XFS_ATTR_FORK); if (error) return error; - /* Start with the attribute data. We'll allocate the rest afterwards. */ - if (crcs) - blkcnt = XFS_B_TO_FSB(mp, args->valuelen); - args->rmtblkno = lblkno = (xfs_dablk_t)lfileoff; args->rmtblkcnt = blkcnt; @@ -349,31 +464,6 @@ xfs_attr_rmtval_set( (map.br_startblock != HOLESTARTBLOCK)); lblkno += map.br_blockcount; blkcnt -= map.br_blockcount; - hdrcnt++; - - /* - * If we have enough blocks for the attribute data, calculate - * how many extra blocks we need for headers. We might run - * through this multiple times in the case that the additional - * headers in the blocks needed for the data fragments spills - * into requiring more blocks. e.g. for 512 byte blocks, we'll - * spill for another block every 9 headers we require in this - * loop. - * - * Note that this can result in contiguous allocation of blocks, - * so we don't use all the space we allocate for headers as we - * have one less header for each contiguous allocation that - * occurs in the map/write loop below. - */ - if (crcs && blkcnt == 0) { - int total_len; - - total_len = args->valuelen + - hdrcnt * sizeof(struct xfs_attr3_rmt_hdr); - blkcnt = XFS_B_TO_FSB(mp, total_len); - blkcnt -= args->rmtblkcnt; - args->rmtblkcnt += blkcnt; - } /* * Start the next trans in the chain. @@ -390,17 +480,15 @@ xfs_attr_rmtval_set( * the INCOMPLETE flag. */ lblkno = args->rmtblkno; - valuelen = args->valuelen; blkcnt = args->rmtblkcnt; + valuelen = args->valuelen; while (valuelen > 0) { - int byte_cnt; - int hdr_size; - int dblkcnt; - char *buf; + struct xfs_buf *bp; + xfs_daddr_t dblkno; + int dblkcnt; + + ASSERT(blkcnt > 0); - /* - * Try to remember where we decided to put the value. - */ xfs_bmap_init(args->flist, args->firstblock); nmap = 1; error = xfs_bmapi_read(dp, (xfs_fileoff_t)lblkno, @@ -419,29 +507,17 @@ xfs_attr_rmtval_set( if (!bp) return ENOMEM; bp->b_ops = &xfs_attr3_rmt_buf_ops; - buf = bp->b_addr; - byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, BBTOB(bp->b_length)); - byte_cnt = min_t(int, valuelen, byte_cnt); - hdr_size = xfs_attr3_rmt_hdr_set(mp, dp->i_ino, offset, - byte_cnt, bp); - ASSERT(hdr_size + byte_cnt <= BBTOB(bp->b_length)); - - memcpy(buf + hdr_size, src, byte_cnt); - - if (byte_cnt + hdr_size < BBTOB(bp->b_length)) - xfs_buf_zero(bp, byte_cnt + hdr_size, - BBTOB(bp->b_length) - byte_cnt - hdr_size); + xfs_attr_rmtval_copyin(mp, bp, args->dp->i_ino, &offset, + &valuelen, &src); error = xfs_bwrite(bp); /* GROT: NOTE: synchronous write */ xfs_buf_relse(bp); if (error) return error; - src += byte_cnt; - valuelen -= byte_cnt; - offset += byte_cnt; + /* roll attribute extent map forwards */ lblkno += map.br_blockcount; blkcnt -= map.br_blockcount; } @@ -454,19 +530,17 @@ xfs_attr_rmtval_set( * out-of-line buffer that it is stored on. */ int -xfs_attr_rmtval_remove(xfs_da_args_t *args) +xfs_attr_rmtval_remove( + struct xfs_da_args *args) { - xfs_mount_t *mp; - xfs_bmbt_irec_t map; - xfs_buf_t *bp; - xfs_daddr_t dblkno; - xfs_dablk_t lblkno; - int valuelen, blkcnt, nmap, error, done, committed; + struct xfs_mount *mp = args->dp->i_mount; + xfs_dablk_t lblkno; + int blkcnt; + int error; + int done; trace_xfs_attr_rmtval_remove(args); - mp = args->dp->i_mount; - /* * Roll through the "value", invalidating the attribute value's blocks. * Note that args->rmtblkcnt is the minimum number of data blocks we'll @@ -476,10 +550,13 @@ xfs_attr_rmtval_remove(xfs_da_args_t *args) * lookups. */ lblkno = args->rmtblkno; - valuelen = args->valuelen; - blkcnt = xfs_attr3_rmt_blocks(mp, valuelen); - while (valuelen > 0) { - int dblkcnt; + blkcnt = args->rmtblkcnt; + while (blkcnt > 0) { + struct xfs_bmbt_irec map; + struct xfs_buf *bp; + xfs_daddr_t dblkno; + int dblkcnt; + int nmap; /* * Try to remember where we decided to put the value. @@ -506,21 +583,19 @@ xfs_attr_rmtval_remove(xfs_da_args_t *args) bp = NULL; } - valuelen -= XFS_ATTR3_RMT_BUF_SPACE(mp, - XFS_FSB_TO_B(mp, map.br_blockcount)); - lblkno += map.br_blockcount; blkcnt -= map.br_blockcount; - blkcnt = max(blkcnt, xfs_attr3_rmt_blocks(mp, valuelen)); } /* * Keep de-allocating extents until the remote-value region is gone. */ - blkcnt = lblkno - args->rmtblkno; lblkno = args->rmtblkno; + blkcnt = args->rmtblkcnt; done = 0; while (!done) { + int committed; + xfs_bmap_init(args->flist, args->firstblock); error = xfs_bunmapi(args->trans, args->dp, lblkno, blkcnt, XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA, diff --git a/fs/xfs/xfs_attr_remote.h b/fs/xfs/xfs_attr_remote.h index c7cca60a062a..92a8fd7977cc 100644 --- a/fs/xfs/xfs_attr_remote.h +++ b/fs/xfs/xfs_attr_remote.h @@ -20,6 +20,14 @@ #define XFS_ATTR3_RMT_MAGIC 0x5841524d /* XARM */ +/* + * There is one of these headers per filesystem block in a remote attribute. + * This is done to ensure there is a 1:1 mapping between the attribute value + * length and the number of blocks needed to store the attribute. This makes the + * verification of a buffer a little more complex, but greatly simplifies the + * allocation, reading and writing of these attributes as we don't have to guess + * the number of blocks needed to store the attribute data. + */ struct xfs_attr3_rmt_hdr { __be32 rm_magic; __be32 rm_offset; @@ -39,6 +47,8 @@ struct xfs_attr3_rmt_hdr { extern const struct xfs_buf_ops xfs_attr3_rmt_buf_ops; +int xfs_attr3_rmt_blocks(struct xfs_mount *mp, int attrlen); + int xfs_attr_rmtval_get(struct xfs_da_args *args); int xfs_attr_rmtval_set(struct xfs_da_args *args); int xfs_attr_rmtval_remove(struct xfs_da_args *args); diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 0d2554299688..1b2472a46e46 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -513,6 +513,7 @@ _xfs_buf_find( xfs_alert(btp->bt_mount, "%s: Block out of range: block 0x%llx, EOFS 0x%llx ", __func__, blkno, eofs); + WARN_ON(1); return NULL; } From 34510185abeaa5be9b178a41c0a03d30aec3db7e Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 27 May 2013 16:38:19 +1000 Subject: [PATCH 17/61] xfs: don't emit v5 superblock warnings on write We write the superblock every 30s or so which results in the verifier being called. Right now that results in this output every 30s: XFS (vda): Version 5 superblock detected. This kernel has EXPERIMENTAL support enabled! Use of these features in this kernel is at your own risk! And spamming the logs. We don't need to check for whether we support v5 superblocks or whether there are feature bits we don't support set as these are only relevant when we first mount the filesytem. i.e. on superblock read. Hence for the write verification we can just skip all the checks (and hence verbose output) altogether. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Ben Myers --- fs/xfs/xfs_mount.c | 18 +++++++++++------- 1 file changed, 11 insertions(+), 7 deletions(-) diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index f6bfbd734669..e8e310c05097 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -314,7 +314,8 @@ STATIC int xfs_mount_validate_sb( xfs_mount_t *mp, xfs_sb_t *sbp, - bool check_inprogress) + bool check_inprogress, + bool check_version) { /* @@ -337,9 +338,10 @@ xfs_mount_validate_sb( /* * Version 5 superblock feature mask validation. Reject combinations the - * kernel cannot support up front before checking anything else. + * kernel cannot support up front before checking anything else. For + * write validation, we don't need to check feature masks. */ - if (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5) { + if (check_version && XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5) { xfs_alert(mp, "Version 5 superblock detected. This kernel has EXPERIMENTAL support enabled!\n" "Use of these features in this kernel is at your own risk!"); @@ -675,7 +677,8 @@ xfs_sb_to_disk( static int xfs_sb_verify( - struct xfs_buf *bp) + struct xfs_buf *bp, + bool check_version) { struct xfs_mount *mp = bp->b_target->bt_mount; struct xfs_sb sb; @@ -686,7 +689,8 @@ xfs_sb_verify( * Only check the in progress field for the primary superblock as * mkfs.xfs doesn't clear it from secondary superblocks. */ - return xfs_mount_validate_sb(mp, &sb, bp->b_bn == XFS_SB_DADDR); + return xfs_mount_validate_sb(mp, &sb, bp->b_bn == XFS_SB_DADDR, + check_version); } /* @@ -719,7 +723,7 @@ xfs_sb_read_verify( goto out_error; } } - error = xfs_sb_verify(bp); + error = xfs_sb_verify(bp, true); out_error: if (error) { @@ -758,7 +762,7 @@ xfs_sb_write_verify( struct xfs_buf_log_item *bip = bp->b_fspriv; int error; - error = xfs_sb_verify(bp); + error = xfs_sb_verify(bp, false); if (error) { XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, error); From 321a95839e65db3759a07a3655184b0283af90fe Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 27 May 2013 16:38:20 +1000 Subject: [PATCH 18/61] xfs: fix incorrect remote symlink block count When CRCs are enabled, the number of blocks needed to hold a remote symlink on a 1k block size filesystem may be 2 instead of 1. The transaction reservation for the allocated blocks was not taking this into account and only allocating one block. Hence when trying to read or invalidate such symlinks, we are mapping a hole where there should be a block and things go bad at that point. Fix the reservation to use the correct block count, clean up the block count calculation similar to the remote attribute calculation, and add a debug guard to detect when we don't write the entire symlink to disk. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Reviewed-by: Brian Foster Signed-off-by: Ben Myers --- fs/xfs/xfs_symlink.c | 20 ++++++-------------- 1 file changed, 6 insertions(+), 14 deletions(-) diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c index 5f234389327c..195a403e1522 100644 --- a/fs/xfs/xfs_symlink.c +++ b/fs/xfs/xfs_symlink.c @@ -56,16 +56,9 @@ xfs_symlink_blocks( struct xfs_mount *mp, int pathlen) { - int fsblocks = 0; - int len = pathlen; + int buflen = XFS_SYMLINK_BUF_SPACE(mp, mp->m_sb.sb_blocksize); - do { - fsblocks++; - len -= XFS_SYMLINK_BUF_SPACE(mp, mp->m_sb.sb_blocksize); - } while (len > 0); - - ASSERT(fsblocks <= XFS_SYMLINK_MAPS); - return fsblocks; + return (pathlen + buflen - 1) / buflen; } static int @@ -405,7 +398,7 @@ xfs_symlink( if (pathlen <= XFS_LITINO(mp, dp->i_d.di_version)) fs_blocks = 0; else - fs_blocks = XFS_B_TO_FSB(mp, pathlen); + fs_blocks = xfs_symlink_blocks(mp, pathlen); resblks = XFS_SYMLINK_SPACE_RES(mp, link_name->len, fs_blocks); error = xfs_trans_reserve(tp, resblks, XFS_SYMLINK_LOG_RES(mp), 0, XFS_TRANS_PERM_LOG_RES, XFS_SYMLINK_LOG_COUNT); @@ -512,7 +505,7 @@ xfs_symlink( cur_chunk = target_path; offset = 0; for (n = 0; n < nmaps; n++) { - char *buf; + char *buf; d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock); byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount); @@ -525,9 +518,7 @@ xfs_symlink( bp->b_ops = &xfs_symlink_buf_ops; byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt); - if (pathlen < byte_cnt) { - byte_cnt = pathlen; - } + byte_cnt = min(byte_cnt, pathlen); buf = bp->b_addr; buf += xfs_symlink_hdr_set(mp, ip->i_ino, offset, @@ -542,6 +533,7 @@ xfs_symlink( xfs_trans_log_buf(tp, bp, 0, (buf + byte_cnt - 1) - (char *)bp->b_addr); } + ASSERT(pathlen == 0); } /* From 709da6a61aaf12181a8eea8443919ae5fc1b731d Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 27 May 2013 16:38:23 +1000 Subject: [PATCH 19/61] xfs: fix split buffer vector log recovery support A long time ago in a galaxy far away.... .. the was a commit made to fix some ilinux specific "fragmented buffer" log recovery problem: http://oss.sgi.com/cgi-bin/gitweb.cgi?p=archive/xfs-import.git;a=commitdiff;h=b29c0bece51da72fb3ff3b61391a391ea54e1603 That problem occurred when a contiguous dirty region of a buffer was split across across two pages of an unmapped buffer. It's been a long time since that has been done in XFS, and the changes to log the entire inode buffers for CRC enabled filesystems has re-introduced that corner case. And, of course, it turns out that the above commit didn't actually fix anything - it just ensured that log recovery is guaranteed to fail when this situation occurs. And now for the gory details. xfstest xfs/085 is failing with this assert: XFS (vdb): bad number of regions (0) in inode log format XFS: Assertion failed: 0, file: fs/xfs/xfs_log_recover.c, line: 1583 Largely undocumented factoid #1: Log recovery depends on all log buffer format items starting with this format: struct foo_log_format { __uint16_t type; __uint16_t size; .... As recoery uses the size field and assumptions about 32 bit alignment in decoding format items. So don't pay much attention to the fact log recovery thinks that it decoding an inode log format item - it just uses them to determine what the size of the item is. But why would it see a log format item with a zero size? Well, luckily enough xfs_logprint uses the same code and gives the same error, so with a bit of gdb magic, it turns out that it isn't a log format that is being decoded. What logprint tells us is this: Oper (130): tid: a0375e1a len: 28 clientid: TRANS flags: none BUF: #regs: 2 start blkno: 144 (0x90) len: 16 bmap size: 2 flags: 0x4000 Oper (131): tid: a0375e1a len: 4096 clientid: TRANS flags: none BUF DATA ---------------------------------------------------------------------------- Oper (132): tid: a0375e1a len: 4096 clientid: TRANS flags: none xfs_logprint: unknown log operation type (4e49) ********************************************************************** * ERROR: data block=2 * ********************************************************************** That we've got a buffer format item (oper 130) that has two regions; the format item itself and one dirty region. The subsequent region after the buffer format item and it's data is them what we are tripping over, and the first bytes of it at an inode magic number. Not a log opheader like there is supposed to be. That means there's a problem with the buffer format item. It's dirty data region is 4096 bytes, and it contains - you guessed it - initialised inodes. But inode buffers are 8k, not 4k, and we log them in their entirety. So something is wrong here. The buffer format item contains: (gdb) p /x *(struct xfs_buf_log_format *)in_f $22 = {blf_type = 0x123c, blf_size = 0x2, blf_flags = 0x4000, blf_len = 0x10, blf_blkno = 0x90, blf_map_size = 0x2, blf_data_map = {0xffffffff, 0xffffffff, .... }} Two regions, and a signle dirty contiguous region of 64 bits. 64 * 128 = 8k, so this should be followed by a single 8k region of data. And the blf_flags tell us that the type of buffer is a XFS_BLFT_DINO_BUF. It contains inodes. And because it doesn't have the XFS_BLF_INODE_BUF flag set, that means it's an inode allocation buffer. So, it should be followed by 8k of inode data. But we know that the next region has a header of: (gdb) p /x *ohead $25 = {oh_tid = 0x1a5e37a0, oh_len = 0x100000, oh_clientid = 0x69, oh_flags = 0x0, oh_res2 = 0x0} and so be32_to_cpu(oh_len) = 0x1000 = 4096 bytes. It's simply not long enough to hold all the logged data. There must be another region. There is - there's a following opheader for another 4k of data that contains the other half of the inode cluster data - the one we assert fail on because it's not a log format header. So why is the second part of the data not being accounted to the correct buffer log format structure? It took a little more work with gdb to work out that the buffer log format structure was both expecting it to be there but hadn't accounted for it. It was at that point I went to the kernel code, as clearly this wasn't a bug in xfs_logprint and the kernel was writing bad stuff to the log. First port of call was the buffer item formatting code, and the discontiguous memory/contiguous dirty region handling code immediately stood out. I've wondered for a long time why the code had this comment in it: vecp->i_addr = xfs_buf_offset(bp, buffer_offset); vecp->i_len = nbits * XFS_BLF_CHUNK; vecp->i_type = XLOG_REG_TYPE_BCHUNK; /* * You would think we need to bump the nvecs here too, but we do not * this number is used by recovery, and it gets confused by the boundary * split here * nvecs++; */ vecp++; And it didn't account for the extra vector pointer. The case being handled here is that a contiguous dirty region lies across a boundary that cannot be memcpy()d across, and so has to be split into two separate operations for xlog_write() to perform. What this code assumes is that what is written to the log is two consecutive blocks of data that are accounted in the buf log format item as the same contiguous dirty region and so will get decoded as such by the log recovery code. The thing is, xlog_write() knows nothing about this, and so just does it's normal thing of adding an opheader for each vector. That means the 8k region gets written to the log as two separate regions of 4k each, but because nvecs has not been incremented, the buf log format item accounts for only one of them. Hence when we come to log recovery, we process the first 4k region and then expect to come across a new item that starts with a log format structure of some kind that tells us whenteh next data is going to be. Instead, we hit raw buffer data and things go bad real quick. So, the commit from 2002 that commented out nvecs++ is just plain wrong. It breaks log recovery completely, and it would seem the only reason this hasn't been since then is that we don't log large contigous regions of multi-page unmapped buffers very often. Never would be a closer estimate, at least until the CRC code came along.... So, lets fix that by restoring the nvecs accounting for the extra region when we hit this case..... .... and there's the problemin log recovery it is apparently working around: XFS: Assertion failed: i == item->ri_total, file: fs/xfs/xfs_log_recover.c, line: 2135 Yup, xlog_recover_do_reg_buffer() doesn't handle contigous dirty regions being broken up into multiple regions by the log formatting code. That's an easy fix, though - if the number of contiguous dirty bits exceeds the length of the region being copied out of the log, only account for the number of dirty bits that region covers, and then loop again and copy more from the next region. It's a 2 line fix. Now xfstests xfs/085 passes, we have one less piece of mystery code, and one more important piece of knowledge about how to structure new log format items.. Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_buf_item.c | 7 +------ fs/xfs/xfs_log_recover.c | 11 +++++++++++ 2 files changed, 12 insertions(+), 6 deletions(-) diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index cf263476d6b4..4ec431777048 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -262,12 +262,7 @@ xfs_buf_item_format_segment( vecp->i_addr = xfs_buf_offset(bp, buffer_offset); vecp->i_len = nbits * XFS_BLF_CHUNK; vecp->i_type = XLOG_REG_TYPE_BCHUNK; -/* - * You would think we need to bump the nvecs here too, but we do not - * this number is used by recovery, and it gets confused by the boundary - * split here - * nvecs++; - */ + nvecs++; vecp++; first_bit = next_bit; last_bit = next_bit; diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 93f03ec17eec..d9e4d3c3991a 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -2096,6 +2096,17 @@ xlog_recover_do_reg_buffer( ASSERT(BBTOB(bp->b_io_length) >= ((uint)bit << XFS_BLF_SHIFT) + (nbits << XFS_BLF_SHIFT)); + /* + * The dirty regions logged in the buffer, even though + * contiguous, may span multiple chunks. This is because the + * dirty region may span a physical page boundary in a buffer + * and hence be split into two separate vectors for writing into + * the log. Hence we need to trim nbits back to the length of + * the current region being copied out of the log. + */ + if (item->ri_buf[i].i_len < (nbits << XFS_BLF_SHIFT)) + nbits = item->ri_buf[i].i_len >> XFS_BLF_SHIFT; + /* * Do a sanity check if this is a dquot buffer. Just checking * the first dquot in the buffer should do. XXXThis is From 02f75405a75eadfb072609f6bf839e027de6a29a Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 27 May 2013 16:38:24 +1000 Subject: [PATCH 20/61] xfs: disable swap extents ioctl on CRC enabled filesystems Currently, swapping extents from one inode to another is a simple act of switching data and attribute forks from one inode to another. This, unfortunately in no longer so simple with CRC enabled filesystems as there is owner information embedded into the BMBT blocks that are swapped between inodes. Hence swapping the forks between inodes results in the inodes having mapping blocks that point to the wrong owner and hence are considered corrupt. To fix this we need an extent tree block or record based swap algorithm so that the BMBT block owner information can be updated atomically in the swap transaction. This is a significant piece of new work, so for the moment simply don't allow swap extent operations to succeed on CRC enabled filesystems. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Reviewed-by: Brian Foster Signed-off-by: Ben Myers --- fs/xfs/xfs_dfrag.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/fs/xfs/xfs_dfrag.c b/fs/xfs/xfs_dfrag.c index f852b082a084..c407e1ccff43 100644 --- a/fs/xfs/xfs_dfrag.c +++ b/fs/xfs/xfs_dfrag.c @@ -219,6 +219,14 @@ xfs_swap_extents( int taforkblks = 0; __uint64_t tmp; + /* + * We have no way of updating owner information in the BMBT blocks for + * each inode on CRC enabled filesystems, so to avoid corrupting the + * this metadata we simply don't allow extent swaps to occur. + */ + if (xfs_sb_version_hascrc(&mp->m_sb)) + return XFS_ERROR(EINVAL); + tempifp = kmem_alloc(sizeof(xfs_ifork_t), KM_MAYFAIL); if (!tempifp) { error = XFS_ERROR(ENOMEM); From 74137fff067961c9aca1e14d073805c3de8549bd Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 27 May 2013 16:38:26 +1000 Subject: [PATCH 21/61] xfs: add fsgeom flag for v5 superblock support. Currently userspace has no way of determining that a filesystem is CRC enabled. Add a flag to the XFS_IOC_FSGEOMETRY ioctl output to indicate that the filesystem has v5 superblock support enabled. This will allow xfs_info to correctly report the state of the filesystem. Signed-off-by: Dave Chinner Reviewed-by: Eric Sandeen Reviewed-by: Brian Foster Signed-off-by: Ben Myers --- fs/xfs/xfs_fs.h | 1 + fs/xfs/xfs_fsops.c | 4 +++- 2 files changed, 4 insertions(+), 1 deletion(-) diff --git a/fs/xfs/xfs_fs.h b/fs/xfs/xfs_fs.h index 6dda3f949b04..d04695545397 100644 --- a/fs/xfs/xfs_fs.h +++ b/fs/xfs/xfs_fs.h @@ -236,6 +236,7 @@ typedef struct xfs_fsop_resblks { #define XFS_FSOP_GEOM_FLAGS_PROJID32 0x0800 /* 32-bit project IDs */ #define XFS_FSOP_GEOM_FLAGS_DIRV2CI 0x1000 /* ASCII only CI names */ #define XFS_FSOP_GEOM_FLAGS_LAZYSB 0x4000 /* lazy superblock counters */ +#define XFS_FSOP_GEOM_FLAGS_V5SB 0x8000 /* version 5 superblock */ /* diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index 87595b211da1..3c3644ea825b 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -99,7 +99,9 @@ xfs_fs_geometry( (xfs_sb_version_hasattr2(&mp->m_sb) ? XFS_FSOP_GEOM_FLAGS_ATTR2 : 0) | (xfs_sb_version_hasprojid32bit(&mp->m_sb) ? - XFS_FSOP_GEOM_FLAGS_PROJID32 : 0); + XFS_FSOP_GEOM_FLAGS_PROJID32 : 0) | + (xfs_sb_version_hascrc(&mp->m_sb) ? + XFS_FSOP_GEOM_FLAGS_V5SB : 0); geo->logsectsize = xfs_sb_version_hassector(&mp->m_sb) ? mp->m_sb.sb_logsectsize : BBSIZE; geo->rtsectsize = mp->m_sb.sb_blocksize; From 56c19e89b38618390addfc743d822f99519055c6 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 27 May 2013 16:38:25 +1000 Subject: [PATCH 22/61] xfs: kill suid/sgid through the truncate path. XFS has failed to kill suid/sgid bits correctly when truncating files of non-zero size since commit c4ed4243 ("xfs: split xfs_setattr") introduced in the 3.1 kernel. Fix it. Fix it. cc: stable kernel Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Ben Myers --- fs/xfs/xfs_iops.c | 47 ++++++++++++++++++++++++++++++++--------------- 1 file changed, 32 insertions(+), 15 deletions(-) diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index d82efaa2ac73..ca9ecaa81112 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -455,6 +455,28 @@ xfs_vn_getattr( return 0; } +static void +xfs_setattr_mode( + struct xfs_trans *tp, + struct xfs_inode *ip, + struct iattr *iattr) +{ + struct inode *inode = VFS_I(ip); + umode_t mode = iattr->ia_mode; + + ASSERT(tp); + ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + + if (!in_group_p(inode->i_gid) && !capable(CAP_FSETID)) + mode &= ~S_ISGID; + + ip->i_d.di_mode &= S_IFMT; + ip->i_d.di_mode |= mode & ~S_IFMT; + + inode->i_mode &= S_IFMT; + inode->i_mode |= mode & ~S_IFMT; +} + int xfs_setattr_nonsize( struct xfs_inode *ip, @@ -606,18 +628,8 @@ xfs_setattr_nonsize( /* * Change file access modes. */ - if (mask & ATTR_MODE) { - umode_t mode = iattr->ia_mode; - - if (!in_group_p(inode->i_gid) && !capable(CAP_FSETID)) - mode &= ~S_ISGID; - - ip->i_d.di_mode &= S_IFMT; - ip->i_d.di_mode |= mode & ~S_IFMT; - - inode->i_mode &= S_IFMT; - inode->i_mode |= mode & ~S_IFMT; - } + if (mask & ATTR_MODE) + xfs_setattr_mode(tp, ip, iattr); /* * Change file access or modified times. @@ -714,9 +726,8 @@ xfs_setattr_size( return XFS_ERROR(error); ASSERT(S_ISREG(ip->i_d.di_mode)); - ASSERT((mask & (ATTR_MODE|ATTR_UID|ATTR_GID|ATTR_ATIME|ATTR_ATIME_SET| - ATTR_MTIME_SET|ATTR_KILL_SUID|ATTR_KILL_SGID| - ATTR_KILL_PRIV|ATTR_TIMES_SET)) == 0); + ASSERT((mask & (ATTR_UID|ATTR_GID|ATTR_ATIME|ATTR_ATIME_SET| + ATTR_MTIME_SET|ATTR_KILL_PRIV|ATTR_TIMES_SET)) == 0); if (!(flags & XFS_ATTR_NOLOCK)) { lock_flags |= XFS_IOLOCK_EXCL; @@ -860,6 +871,12 @@ xfs_setattr_size( xfs_inode_clear_eofblocks_tag(ip); } + /* + * Change file access modes. + */ + if (mask & ATTR_MODE) + xfs_setattr_mode(tp, ip, iattr); + if (mask & ATTR_CTIME) { inode->i_ctime = iattr->ia_ctime; ip->i_d.di_ctime.t_sec = iattr->ia_ctime.tv_sec; From 5ae6e6a401957698f2bd8c9f4a86d86d02199fea Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 28 May 2013 18:37:17 +1000 Subject: [PATCH 23/61] xfs: fix dir3 freespace block corruption When the directory freespace index grows to a second block (2017 4k data blocks in the directory), the initialisation of the second new block header goes wrong. The write verifier fires a corruption error indicating that the block number in the header is zero. This was being tripped by xfs/110. The problem is that the initialisation of the new block is done just fine in xfs_dir3_free_get_buf(), but the caller then users a dirv2 structure to zero on-disk header fields that xfs_dir3_free_get_buf() has already zeroed. These lined up with the block number in the dir v3 header format. While looking at this, I noticed that the struct xfs_dir3_free_hdr() had 4 bytes of padding in it that wasn't defined as padding or being zeroed by the initialisation. Add a pad field declaration and fully zero the on disk and in-core headers in xfs_dir3_free_get_buf() so that this is never an issue in the future. Note that this doesn't change the on-disk layout, just makes the 32 bits of padding in the layout explicit. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_dir2_format.h | 1 + fs/xfs/xfs_dir2_node.c | 13 ++++++------- 2 files changed, 7 insertions(+), 7 deletions(-) diff --git a/fs/xfs/xfs_dir2_format.h b/fs/xfs/xfs_dir2_format.h index a3b1bd841a80..995f1f505a52 100644 --- a/fs/xfs/xfs_dir2_format.h +++ b/fs/xfs/xfs_dir2_format.h @@ -715,6 +715,7 @@ struct xfs_dir3_free_hdr { __be32 firstdb; /* db of first entry */ __be32 nvalid; /* count of valid entries */ __be32 nused; /* count of used entries */ + __be32 pad; /* 64 bit alignment. */ }; struct xfs_dir3_free { diff --git a/fs/xfs/xfs_dir2_node.c b/fs/xfs/xfs_dir2_node.c index 5246de4912d4..2226a00acd15 100644 --- a/fs/xfs/xfs_dir2_node.c +++ b/fs/xfs/xfs_dir2_node.c @@ -263,18 +263,19 @@ xfs_dir3_free_get_buf( * Initialize the new block to be empty, and remember * its first slot as our empty slot. */ - hdr.magic = XFS_DIR2_FREE_MAGIC; - hdr.firstdb = 0; - hdr.nused = 0; - hdr.nvalid = 0; + memset(bp->b_addr, 0, sizeof(struct xfs_dir3_free_hdr)); + memset(&hdr, 0, sizeof(hdr)); + if (xfs_sb_version_hascrc(&mp->m_sb)) { struct xfs_dir3_free_hdr *hdr3 = bp->b_addr; hdr.magic = XFS_DIR3_FREE_MAGIC; + hdr3->hdr.blkno = cpu_to_be64(bp->b_bn); hdr3->hdr.owner = cpu_to_be64(dp->i_ino); uuid_copy(&hdr3->hdr.uuid, &mp->m_sb.sb_uuid); - } + } else + hdr.magic = XFS_DIR2_FREE_MAGIC; xfs_dir3_free_hdr_to_disk(bp->b_addr, &hdr); *bpp = bp; return 0; @@ -1921,8 +1922,6 @@ xfs_dir2_node_addname_int( */ freehdr.firstdb = (fbno - XFS_DIR2_FREE_FIRSTDB(mp)) * xfs_dir3_free_max_bests(mp); - free->hdr.nvalid = 0; - free->hdr.nused = 0; } else { free = fbp->b_addr; bests = xfs_dir3_free_bests_p(mp, free); From 6fcdc59de28817d1fbf1bd58cc01f4f3fac858fb Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 3 Jun 2013 15:28:46 +1000 Subject: [PATCH 24/61] xfs: rework dquot CRCs Calculating dquot CRCs when the backing buffer is written back just doesn't work reliably. There are several places which manipulate dquots directly in the buffers, and they don't calculate CRCs appropriately, nor do they always set the buffer up to calculate CRCs appropriately. Firstly, if we log a dquot buffer (e.g. during allocation) it gets logged without valid CRC, and so on recovery we end up with a dquot that is not valid. Secondly, if we recover/repair a dquot, we don't have a verifier attached to the buffer and hence CRCs are not calculated on the way down to disk. Thirdly, calculating the CRC after we've changed the contents means that if we re-read the dquot from the buffer, we cannot verify the contents of the dquot are valid, as the CRC is invalid. So, to avoid all the dquot CRC errors that are being detected by the read verifier, change to using the same model as for inodes. That is, dquot CRCs are calculated and written to the backing buffer at the time the dquot is flushed to the backing buffer. If we modify the dquot directly in the backing buffer, calculate the CRC immediately after the modification is complete. Hence the dquot in the on-disk buffer should always have a valid CRC. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_dquot.c | 37 ++++++++++++++++--------------------- fs/xfs/xfs_log_recover.c | 10 ++++++++++ fs/xfs/xfs_qm.c | 40 ++++++++++++++++++++++++++++++---------- fs/xfs/xfs_quota.h | 2 ++ 4 files changed, 58 insertions(+), 31 deletions(-) diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index a41f8bf1da37..044e97a33c8d 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -249,8 +249,11 @@ xfs_qm_init_dquot_blk( d->dd_diskdq.d_version = XFS_DQUOT_VERSION; d->dd_diskdq.d_id = cpu_to_be32(curid); d->dd_diskdq.d_flags = type; - if (xfs_sb_version_hascrc(&mp->m_sb)) + if (xfs_sb_version_hascrc(&mp->m_sb)) { uuid_copy(&d->dd_uuid, &mp->m_sb.sb_uuid); + xfs_update_cksum((char *)d, sizeof(struct xfs_dqblk), + XFS_DQUOT_CRC_OFF); + } } xfs_trans_dquot_buf(tp, bp, @@ -286,23 +289,6 @@ xfs_dquot_set_prealloc_limits(struct xfs_dquot *dqp) dqp->q_low_space[XFS_QLOWSP_5_PCNT] = space * 5; } -STATIC void -xfs_dquot_buf_calc_crc( - struct xfs_mount *mp, - struct xfs_buf *bp) -{ - struct xfs_dqblk *d = (struct xfs_dqblk *)bp->b_addr; - int i; - - if (!xfs_sb_version_hascrc(&mp->m_sb)) - return; - - for (i = 0; i < mp->m_quotainfo->qi_dqperchunk; i++, d++) { - xfs_update_cksum((char *)d, sizeof(struct xfs_dqblk), - offsetof(struct xfs_dqblk, dd_crc)); - } -} - STATIC bool xfs_dquot_buf_verify_crc( struct xfs_mount *mp, @@ -328,12 +314,11 @@ xfs_dquot_buf_verify_crc( for (i = 0; i < ndquots; i++, d++) { if (!xfs_verify_cksum((char *)d, sizeof(struct xfs_dqblk), - offsetof(struct xfs_dqblk, dd_crc))) + XFS_DQUOT_CRC_OFF)) return false; if (!uuid_equal(&d->dd_uuid, &mp->m_sb.sb_uuid)) return false; } - return true; } @@ -393,6 +378,11 @@ xfs_dquot_buf_read_verify( } } +/* + * we don't calculate the CRC here as that is done when the dquot is flushed to + * the buffer after the update is done. This ensures that the dquot in the + * buffer always has an up-to-date CRC value. + */ void xfs_dquot_buf_write_verify( struct xfs_buf *bp) @@ -404,7 +394,6 @@ xfs_dquot_buf_write_verify( xfs_buf_ioerror(bp, EFSCORRUPTED); return; } - xfs_dquot_buf_calc_crc(mp, bp); } const struct xfs_buf_ops xfs_dquot_buf_ops = { @@ -1151,11 +1140,17 @@ xfs_qm_dqflush( * copy the lsn into the on-disk dquot now while we have the in memory * dquot here. This can't be done later in the write verifier as we * can't get access to the log item at that point in time. + * + * We also calculate the CRC here so that the on-disk dquot in the + * buffer always has a valid CRC. This ensures there is no possibility + * of a dquot without an up-to-date CRC getting to disk. */ if (xfs_sb_version_hascrc(&mp->m_sb)) { struct xfs_dqblk *dqb = (struct xfs_dqblk *)ddqp; dqb->dd_lsn = cpu_to_be64(dqp->q_logitem.qli_item.li_lsn); + xfs_update_cksum((char *)dqb, sizeof(struct xfs_dqblk), + XFS_DQUOT_CRC_OFF); } /* diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index d9e4d3c3991a..d6204d1ac47f 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -2266,6 +2266,12 @@ xfs_qm_dqcheck( d->dd_diskdq.d_flags = type; d->dd_diskdq.d_id = cpu_to_be32(id); + if (xfs_sb_version_hascrc(&mp->m_sb)) { + uuid_copy(&d->dd_uuid, &mp->m_sb.sb_uuid); + xfs_update_cksum((char *)d, sizeof(struct xfs_dqblk), + XFS_DQUOT_CRC_OFF); + } + return errs; } @@ -2793,6 +2799,10 @@ xlog_recover_dquot_pass2( } memcpy(ddq, recddq, item->ri_buf[1].i_len); + if (xfs_sb_version_hascrc(&mp->m_sb)) { + xfs_update_cksum((char *)ddq, sizeof(struct xfs_dqblk), + XFS_DQUOT_CRC_OFF); + } ASSERT(dq_f->qlf_size == 2); ASSERT(bp->b_target->bt_mount == mp); diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index f41702b43003..b75c9bb6e71e 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -41,6 +41,7 @@ #include "xfs_qm.h" #include "xfs_trace.h" #include "xfs_icache.h" +#include "xfs_cksum.h" /* * The global quota manager. There is only one of these for the entire @@ -839,7 +840,7 @@ xfs_qm_reset_dqcounts( xfs_dqid_t id, uint type) { - xfs_disk_dquot_t *ddq; + struct xfs_dqblk *dqb; int j; trace_xfs_reset_dqcounts(bp, _RET_IP_); @@ -853,8 +854,12 @@ xfs_qm_reset_dqcounts( do_div(j, sizeof(xfs_dqblk_t)); ASSERT(mp->m_quotainfo->qi_dqperchunk == j); #endif - ddq = bp->b_addr; + dqb = bp->b_addr; for (j = 0; j < mp->m_quotainfo->qi_dqperchunk; j++) { + struct xfs_disk_dquot *ddq; + + ddq = (struct xfs_disk_dquot *)&dqb[j]; + /* * Do a sanity check, and if needed, repair the dqblk. Don't * output any warnings because it's perfectly possible to @@ -871,7 +876,12 @@ xfs_qm_reset_dqcounts( ddq->d_bwarns = 0; ddq->d_iwarns = 0; ddq->d_rtbwarns = 0; - ddq = (xfs_disk_dquot_t *) ((xfs_dqblk_t *)ddq + 1); + + if (xfs_sb_version_hascrc(&mp->m_sb)) { + xfs_update_cksum((char *)&dqb[j], + sizeof(struct xfs_dqblk), + XFS_DQUOT_CRC_OFF); + } } } @@ -907,19 +917,29 @@ xfs_qm_dqiter_bufs( XFS_FSB_TO_DADDR(mp, bno), mp->m_quotainfo->qi_dqchunklen, 0, &bp, &xfs_dquot_buf_ops); + + /* + * CRC and validation errors will return a EFSCORRUPTED here. If + * this occurs, re-read without CRC validation so that we can + * repair the damage via xfs_qm_reset_dqcounts(). This process + * will leave a trace in the log indicating corruption has + * been detected. + */ + if (error == EFSCORRUPTED) { + error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp, + XFS_FSB_TO_DADDR(mp, bno), + mp->m_quotainfo->qi_dqchunklen, 0, &bp, + NULL); + } + if (error) break; - /* - * XXX(hch): need to figure out if it makes sense to validate - * the CRC here. - */ xfs_qm_reset_dqcounts(mp, bp, firstid, type); xfs_buf_delwri_queue(bp, buffer_list); xfs_buf_relse(bp); - /* - * goto the next block. - */ + + /* goto the next block. */ bno++; firstid += mp->m_quotainfo->qi_dqperchunk; } diff --git a/fs/xfs/xfs_quota.h b/fs/xfs/xfs_quota.h index c61e31c7d997..c38068f26c55 100644 --- a/fs/xfs/xfs_quota.h +++ b/fs/xfs/xfs_quota.h @@ -87,6 +87,8 @@ typedef struct xfs_dqblk { uuid_t dd_uuid; /* location information */ } xfs_dqblk_t; +#define XFS_DQUOT_CRC_OFF offsetof(struct xfs_dqblk, dd_crc) + /* * flags for q_flags field in the dquot. */ From 59913f14dfe8eb772ff93eb442947451b4416329 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 3 Jun 2013 15:28:49 +1000 Subject: [PATCH 25/61] xfs: fix remote attribute invalidation for a leaf When invalidating an attribute leaf block block, there might be remote attributes that it points to. With the recent rework of the remote attribute format, we have to make sure we calculate the length of the attribute correctly. We aren't doing that in xfs_attr3_leaf_inactive(), so fix it. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_attr_leaf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index d788302e506a..31d3cd129269 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -3258,7 +3258,7 @@ xfs_attr3_leaf_inactive( name_rmt = xfs_attr3_leaf_name_remote(leaf, i); if (name_rmt->valueblk) { lp->valueblk = be32_to_cpu(name_rmt->valueblk); - lp->valuelen = XFS_B_TO_FSB(dp->i_mount, + lp->valuelen = xfs_attr3_rmt_blocks(dp->i_mount, be32_to_cpu(name_rmt->valuelen)); lp++; } From a775ad778073d55744ed6709ccede36310638911 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 5 Jun 2013 12:09:07 +1000 Subject: [PATCH 26/61] xfs: fix log recovery transaction item reordering There are several constraints that inode allocation and unlink logging impose on log recovery. These all stem from the fact that inode alloc/unlink are logged in buffers, but all other inode changes are logged in inode items. Hence there are ordering constraints that recovery must follow to ensure the correct result occurs. As it turns out, this ordering has been working mostly by chance than good management. The existing code moves all buffers except cancelled buffers to the head of the list, and everything else to the tail of the list. The problem with this is that is interleaves inode items with the buffer cancellation items, and hence whether the inode item in an cancelled buffer gets replayed is essentially left to chance. Further, this ordering causes problems for log recovery when inode CRCs are enabled. It typically replays the inode unlink buffer long before it replays the inode core changes, and so the CRC recorded in an unlink buffer is going to be invalid and hence any attempt to validate the inode in the buffer is going to fail. Hence we really need to enforce the ordering that the inode alloc/unlink code has expected log recovery to have since inode chunk de-allocation was introduced back in 2003... Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_log_recover.c | 65 +++++++++++++++++++++++++++++++++++----- 1 file changed, 58 insertions(+), 7 deletions(-) diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index d6204d1ac47f..83088d96e6c4 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -1599,10 +1599,43 @@ xlog_recover_add_to_trans( } /* - * Sort the log items in the transaction. Cancelled buffers need - * to be put first so they are processed before any items that might - * modify the buffers. If they are cancelled, then the modifications - * don't need to be replayed. + * Sort the log items in the transaction. + * + * The ordering constraints are defined by the inode allocation and unlink + * behaviour. The rules are: + * + * 1. Every item is only logged once in a given transaction. Hence it + * represents the last logged state of the item. Hence ordering is + * dependent on the order in which operations need to be performed so + * required initial conditions are always met. + * + * 2. Cancelled buffers are recorded in pass 1 in a separate table and + * there's nothing to replay from them so we can simply cull them + * from the transaction. However, we can't do that until after we've + * replayed all the other items because they may be dependent on the + * cancelled buffer and replaying the cancelled buffer can remove it + * form the cancelled buffer table. Hence they have tobe done last. + * + * 3. Inode allocation buffers must be replayed before inode items that + * read the buffer and replay changes into it. + * + * 4. Inode unlink buffers must be replayed after inode items are replayed. + * This ensures that inodes are completely flushed to the inode buffer + * in a "free" state before we remove the unlinked inode list pointer. + * + * Hence the ordering needs to be inode allocation buffers first, inode items + * second, inode unlink buffers third and cancelled buffers last. + * + * But there's a problem with that - we can't tell an inode allocation buffer + * apart from a regular buffer, so we can't separate them. We can, however, + * tell an inode unlink buffer from the others, and so we can separate them out + * from all the other buffers and move them to last. + * + * Hence, 4 lists, in order from head to tail: + * - buffer_list for all buffers except cancelled/inode unlink buffers + * - item_list for all non-buffer items + * - inode_buffer_list for inode unlink buffers + * - cancel_list for the cancelled buffers */ STATIC int xlog_recover_reorder_trans( @@ -1612,6 +1645,10 @@ xlog_recover_reorder_trans( { xlog_recover_item_t *item, *n; LIST_HEAD(sort_list); + LIST_HEAD(cancel_list); + LIST_HEAD(buffer_list); + LIST_HEAD(inode_buffer_list); + LIST_HEAD(inode_list); list_splice_init(&trans->r_itemq, &sort_list); list_for_each_entry_safe(item, n, &sort_list, ri_list) { @@ -1619,12 +1656,18 @@ xlog_recover_reorder_trans( switch (ITEM_TYPE(item)) { case XFS_LI_BUF: - if (!(buf_f->blf_flags & XFS_BLF_CANCEL)) { + if (buf_f->blf_flags & XFS_BLF_CANCEL) { trace_xfs_log_recover_item_reorder_head(log, trans, item, pass); - list_move(&item->ri_list, &trans->r_itemq); + list_move(&item->ri_list, &cancel_list); break; } + if (buf_f->blf_flags & XFS_BLF_INODE_BUF) { + list_move(&item->ri_list, &inode_buffer_list); + break; + } + list_move_tail(&item->ri_list, &buffer_list); + break; case XFS_LI_INODE: case XFS_LI_DQUOT: case XFS_LI_QUOTAOFF: @@ -1632,7 +1675,7 @@ xlog_recover_reorder_trans( case XFS_LI_EFI: trace_xfs_log_recover_item_reorder_tail(log, trans, item, pass); - list_move_tail(&item->ri_list, &trans->r_itemq); + list_move_tail(&item->ri_list, &inode_list); break; default: xfs_warn(log->l_mp, @@ -1643,6 +1686,14 @@ xlog_recover_reorder_trans( } } ASSERT(list_empty(&sort_list)); + if (!list_empty(&buffer_list)) + list_splice(&buffer_list, &trans->r_itemq); + if (!list_empty(&inode_list)) + list_splice_tail(&inode_list, &trans->r_itemq); + if (!list_empty(&inode_buffer_list)) + list_splice_tail(&inode_buffer_list, &trans->r_itemq); + if (!list_empty(&cancel_list)) + list_splice_tail(&cancel_list, &trans->r_itemq); return 0; } From 0a32c26e720a8b38971d0685976f4a7d63f9e2ef Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 5 Jun 2013 12:09:08 +1000 Subject: [PATCH 27/61] xfs: inode unlinked list needs to recalculate the inode CRC The inode unlinked list manipulations operate directly on the inode buffer, and so bypass the inode CRC calculation mechanisms. Hence an inode on the unlinked list has an invalid CRC. Fix this by recalculating the CRC whenever we modify an unlinked list pointer in an inode, ncluding during log recovery. This is trivial to do and results in unlinked list operations always leaving a consistent inode in the buffer. Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_inode.c | 16 ++++++++++++++++ fs/xfs/xfs_log_recover.c | 9 +++++++++ 2 files changed, 25 insertions(+) diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index efbe1accb6ca..7f7be5f98f52 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1638,6 +1638,10 @@ xfs_iunlink( dip->di_next_unlinked = agi->agi_unlinked[bucket_index]; offset = ip->i_imap.im_boffset + offsetof(xfs_dinode_t, di_next_unlinked); + + /* need to recalc the inode CRC if appropriate */ + xfs_dinode_calc_crc(mp, dip); + xfs_trans_inode_buf(tp, ibp); xfs_trans_log_buf(tp, ibp, offset, (offset + sizeof(xfs_agino_t) - 1)); @@ -1723,6 +1727,10 @@ xfs_iunlink_remove( dip->di_next_unlinked = cpu_to_be32(NULLAGINO); offset = ip->i_imap.im_boffset + offsetof(xfs_dinode_t, di_next_unlinked); + + /* need to recalc the inode CRC if appropriate */ + xfs_dinode_calc_crc(mp, dip); + xfs_trans_inode_buf(tp, ibp); xfs_trans_log_buf(tp, ibp, offset, (offset + sizeof(xfs_agino_t) - 1)); @@ -1796,6 +1804,10 @@ xfs_iunlink_remove( dip->di_next_unlinked = cpu_to_be32(NULLAGINO); offset = ip->i_imap.im_boffset + offsetof(xfs_dinode_t, di_next_unlinked); + + /* need to recalc the inode CRC if appropriate */ + xfs_dinode_calc_crc(mp, dip); + xfs_trans_inode_buf(tp, ibp); xfs_trans_log_buf(tp, ibp, offset, (offset + sizeof(xfs_agino_t) - 1)); @@ -1809,6 +1821,10 @@ xfs_iunlink_remove( last_dip->di_next_unlinked = cpu_to_be32(next_agino); ASSERT(next_agino != 0); offset = last_offset + offsetof(xfs_dinode_t, di_next_unlinked); + + /* need to recalc the inode CRC if appropriate */ + xfs_dinode_calc_crc(mp, last_dip); + xfs_trans_inode_buf(tp, last_ibp); xfs_trans_log_buf(tp, last_ibp, offset, (offset + sizeof(xfs_agino_t) - 1)); diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 83088d96e6c4..45a85ff84da1 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -1912,6 +1912,15 @@ xlog_recover_do_inode_buffer( buffer_nextp = (xfs_agino_t *)xfs_buf_offset(bp, next_unlinked_offset); *buffer_nextp = *logged_nextp; + + /* + * If necessary, recalculate the CRC in the on-disk inode. We + * have to leave the inode in a consistent state for whoever + * reads it next.... + */ + xfs_dinode_calc_crc(mp, (struct xfs_dinode *) + xfs_buf_offset(bp, i * mp->m_sb.sb_inodesize)); + } return 0; From d3eaace84e40bf946129e516dcbd617173c1cf14 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 5 Jun 2013 12:09:09 +1000 Subject: [PATCH 28/61] xfs: disable noattr2/attr2 mount options for CRC enabled filesystems attr2 format is always enabled for v5 superblock filesystems, so the mount options to enable or disable it need to be cause mount errors. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Ben Myers --- Documentation/filesystems/xfs.txt | 3 +++ fs/xfs/xfs_super.c | 11 +++++++++++ 2 files changed, 14 insertions(+) diff --git a/Documentation/filesystems/xfs.txt b/Documentation/filesystems/xfs.txt index 3e4b3dd1e046..83577f0232a0 100644 --- a/Documentation/filesystems/xfs.txt +++ b/Documentation/filesystems/xfs.txt @@ -33,6 +33,9 @@ When mounting an XFS filesystem, the following options are accepted. removing extended attributes) the on-disk superblock feature bit field will be updated to reflect this format being in use. + CRC enabled filesystems always use the attr2 format, and so + will reject the noattr2 mount option if it is set. + barrier Enables the use of block layer write barriers for writes into the journal and unwritten extent conversion. This allows for diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index ea341cea68cb..3033ba5e9762 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -1372,6 +1372,17 @@ xfs_finish_flags( } } + /* + * V5 filesystems always use attr2 format for attributes. + */ + if (xfs_sb_version_hascrc(&mp->m_sb) && + (mp->m_flags & XFS_MOUNT_NOATTR2)) { + xfs_warn(mp, +"Cannot mount a V5 filesystem as %s. %s is always enabled for V5 filesystems.", + MNTOPT_NOATTR2, MNTOPT_ATTR2); + return XFS_ERROR(EINVAL); + } + /* * mkfs'ed attr2 will turn on attr2 mount unless explicitly * told by noattr2 to turn it off From 5c87d4bc1a86bd6e6754ac3d6e111d776ddcfe57 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 5 Jun 2013 12:09:10 +1000 Subject: [PATCH 29/61] xfs: increase number of ACL entries for V5 superblocks The limit of 25 ACL entries is arbitrary, but baked into the on-disk format. For version 5 superblocks, increase it to the maximum nuber of ACLs that can fit into a single xattr. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_acl.c | 31 ++++++++++++++++++------------- fs/xfs/xfs_acl.h | 33 +++++++++++++++++++++++++-------- 2 files changed, 43 insertions(+), 21 deletions(-) diff --git a/fs/xfs/xfs_acl.c b/fs/xfs/xfs_acl.c index 1d32f1d52763..306d883d89bc 100644 --- a/fs/xfs/xfs_acl.c +++ b/fs/xfs/xfs_acl.c @@ -21,6 +21,8 @@ #include "xfs_bmap_btree.h" #include "xfs_inode.h" #include "xfs_vnodeops.h" +#include "xfs_sb.h" +#include "xfs_mount.h" #include "xfs_trace.h" #include #include @@ -34,7 +36,9 @@ */ STATIC struct posix_acl * -xfs_acl_from_disk(struct xfs_acl *aclp) +xfs_acl_from_disk( + struct xfs_acl *aclp, + int max_entries) { struct posix_acl_entry *acl_e; struct posix_acl *acl; @@ -42,7 +46,7 @@ xfs_acl_from_disk(struct xfs_acl *aclp) unsigned int count, i; count = be32_to_cpu(aclp->acl_cnt); - if (count > XFS_ACL_MAX_ENTRIES) + if (count > max_entries) return ERR_PTR(-EFSCORRUPTED); acl = posix_acl_alloc(count, GFP_KERNEL); @@ -108,9 +112,9 @@ xfs_get_acl(struct inode *inode, int type) struct xfs_inode *ip = XFS_I(inode); struct posix_acl *acl; struct xfs_acl *xfs_acl; - int len = sizeof(struct xfs_acl); unsigned char *ea_name; int error; + int len; acl = get_cached_acl(inode, type); if (acl != ACL_NOT_CACHED) @@ -133,8 +137,8 @@ xfs_get_acl(struct inode *inode, int type) * If we have a cached ACLs value just return it, not need to * go out to the disk. */ - - xfs_acl = kzalloc(sizeof(struct xfs_acl), GFP_KERNEL); + len = XFS_ACL_MAX_SIZE(ip->i_mount); + xfs_acl = kzalloc(len, GFP_KERNEL); if (!xfs_acl) return ERR_PTR(-ENOMEM); @@ -153,7 +157,7 @@ xfs_get_acl(struct inode *inode, int type) goto out; } - acl = xfs_acl_from_disk(xfs_acl); + acl = xfs_acl_from_disk(xfs_acl, XFS_ACL_MAX_ENTRIES(ip->i_mount)); if (IS_ERR(acl)) goto out; @@ -189,16 +193,17 @@ xfs_set_acl(struct inode *inode, int type, struct posix_acl *acl) if (acl) { struct xfs_acl *xfs_acl; - int len; + int len = XFS_ACL_MAX_SIZE(ip->i_mount); - xfs_acl = kzalloc(sizeof(struct xfs_acl), GFP_KERNEL); + xfs_acl = kzalloc(len, GFP_KERNEL); if (!xfs_acl) return -ENOMEM; xfs_acl_to_disk(xfs_acl, acl); - len = sizeof(struct xfs_acl) - - (sizeof(struct xfs_acl_entry) * - (XFS_ACL_MAX_ENTRIES - acl->a_count)); + + /* subtract away the unused acl entries */ + len -= sizeof(struct xfs_acl_entry) * + (XFS_ACL_MAX_ENTRIES(ip->i_mount) - acl->a_count); error = -xfs_attr_set(ip, ea_name, (unsigned char *)xfs_acl, len, ATTR_ROOT); @@ -243,7 +248,7 @@ xfs_set_mode(struct inode *inode, umode_t mode) static int xfs_acl_exists(struct inode *inode, unsigned char *name) { - int len = sizeof(struct xfs_acl); + int len = XFS_ACL_MAX_SIZE(XFS_M(inode->i_sb)); return (xfs_attr_get(XFS_I(inode), name, NULL, &len, ATTR_ROOT|ATTR_KERNOVAL) == 0); @@ -379,7 +384,7 @@ xfs_xattr_acl_set(struct dentry *dentry, const char *name, goto out_release; error = -EINVAL; - if (acl->a_count > XFS_ACL_MAX_ENTRIES) + if (acl->a_count > XFS_ACL_MAX_ENTRIES(XFS_M(inode->i_sb))) goto out_release; if (type == ACL_TYPE_ACCESS) { diff --git a/fs/xfs/xfs_acl.h b/fs/xfs/xfs_acl.h index 39632d941354..4016a567b83c 100644 --- a/fs/xfs/xfs_acl.h +++ b/fs/xfs/xfs_acl.h @@ -22,19 +22,36 @@ struct inode; struct posix_acl; struct xfs_inode; -#define XFS_ACL_MAX_ENTRIES 25 #define XFS_ACL_NOT_PRESENT (-1) /* On-disk XFS access control list structure */ -struct xfs_acl { - __be32 acl_cnt; - struct xfs_acl_entry { - __be32 ae_tag; - __be32 ae_id; - __be16 ae_perm; - } acl_entry[XFS_ACL_MAX_ENTRIES]; +struct xfs_acl_entry { + __be32 ae_tag; + __be32 ae_id; + __be16 ae_perm; + __be16 ae_pad; /* fill the implicit hole in the structure */ }; +struct xfs_acl { + __be32 acl_cnt; + struct xfs_acl_entry acl_entry[0]; +}; + +/* + * The number of ACL entries allowed is defined by the on-disk format. + * For v4 superblocks, that is limited to 25 entries. For v5 superblocks, it is + * limited only by the maximum size of the xattr that stores the information. + */ +#define XFS_ACL_MAX_ENTRIES(mp) \ + (xfs_sb_version_hascrc(&mp->m_sb) \ + ? (XATTR_SIZE_MAX - sizeof(struct xfs_acl)) / \ + sizeof(struct xfs_acl_entry) \ + : 25) + +#define XFS_ACL_MAX_SIZE(mp) \ + (sizeof(struct xfs_acl) + \ + sizeof(struct xfs_acl_entry) * XFS_ACL_MAX_ENTRIES((mp))) + /* On-disk XFS extended attribute names */ #define SGI_ACL_FILE (unsigned char *)"SGI_ACL_FILE" #define SGI_ACL_DEFAULT (unsigned char *)"SGI_ACL_DEFAULT" From 8a1fd2950e1fe267e11fc8c85dcaa6b023b51b60 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 12 Jun 2013 12:19:07 +1000 Subject: [PATCH 30/61] xfs: fix implicit padding in directory and attr CRC formats Michael L. Semon has been testing CRC patches on a 32 bit system and been seeing assert failures in the directory code from xfs/080. Thanks to Michael's heroic efforts with printk debugging, we found that the problem was that the last free space being left in the directory structure was too small to fit a unused tag structure and it was being corrupted and attempting to log a region out of bounds. Hence the assert failure looked something like: ..... #5 calling xfs_dir2_data_log_unused() 36 32 #1 4092 4095 4096 #2 8182 8183 4096 XFS: Assertion failed: first <= last && last < BBTOB(bp->b_length), file: fs/xfs/xfs_trans_buf.c, line: 568 Where #1 showed the first region of the dup being logged (i.e. the last 4 bytes of a directory buffer) and #2 shows the corrupt values being calculated from the length of the dup entry which overflowed the size of the buffer. It turns out that the problem was not in the logging code, nor in the freespace handling code. It is an initial condition bug that only shows up on 32 bit systems. When a new buffer is initialised, where's the freespace that is set up: [ 172.316249] calling xfs_dir2_leaf_addname() from xfs_dir_createname() [ 172.316346] #9 calling xfs_dir2_data_log_unused() [ 172.316351] #1 calling xfs_trans_log_buf() 60 63 4096 [ 172.316353] #2 calling xfs_trans_log_buf() 4094 4095 4096 Note the offset of the first region being logged? It's 60 bytes into the buffer. Once I saw that, I pretty much knew that the bug was going to be caused by this. Essentially, all direct entries are rounded to 8 bytes in length, and all entries start with an 8 byte alignment. This means that we can decode inplace as variables are naturally aligned. With the directory data supposedly starting on a 8 byte boundary, and all entries padded to 8 bytes, the minimum freespace in a directory block is supposed to be 8 bytes, which is large enough to fit a unused data entry structure (6 bytes in size). The fact we only have 4 bytes of free space indicates a directory data block alignment problem. And what do you know - there's an implicit hole in the directory data block header for the CRC format, which means the header is 60 byte on 32 bit intel systems and 64 bytes on 64 bit systems. Needs padding. And while looking at the structures, I found the same problem in the attr leaf header. Fix them both. Note that this only affects 32 bit systems with CRCs enabled. Everything else is just fine. Note that CRC enabled filesystems created before this fix on such systems will not be readable with this fix applied. Reported-by: Michael L. Semon Debugged-by: Michael L. Semon Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_attr_leaf.h | 1 + fs/xfs/xfs_dir2_format.h | 5 +++-- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/fs/xfs/xfs_attr_leaf.h b/fs/xfs/xfs_attr_leaf.h index f9d7846097e2..444a7704596c 100644 --- a/fs/xfs/xfs_attr_leaf.h +++ b/fs/xfs/xfs_attr_leaf.h @@ -128,6 +128,7 @@ struct xfs_attr3_leaf_hdr { __u8 holes; __u8 pad1; struct xfs_attr_leaf_map freemap[XFS_ATTR_LEAF_MAPSIZE]; + __be32 pad2; /* 64 bit alignment */ }; #define XFS_ATTR3_LEAF_CRC_OFF (offsetof(struct xfs_attr3_leaf_hdr, info.crc)) diff --git a/fs/xfs/xfs_dir2_format.h b/fs/xfs/xfs_dir2_format.h index 995f1f505a52..7826782b8d78 100644 --- a/fs/xfs/xfs_dir2_format.h +++ b/fs/xfs/xfs_dir2_format.h @@ -266,6 +266,7 @@ struct xfs_dir3_blk_hdr { struct xfs_dir3_data_hdr { struct xfs_dir3_blk_hdr hdr; xfs_dir2_data_free_t best_free[XFS_DIR2_DATA_FD_COUNT]; + __be32 pad; /* 64 bit alignment */ }; #define XFS_DIR3_DATA_CRC_OFF offsetof(struct xfs_dir3_data_hdr, hdr.crc) @@ -477,7 +478,7 @@ struct xfs_dir3_leaf_hdr { struct xfs_da3_blkinfo info; /* header for da routines */ __be16 count; /* count of entries */ __be16 stale; /* count of stale entries */ - __be32 pad; + __be32 pad; /* 64 bit alignment */ }; struct xfs_dir3_icleaf_hdr { @@ -715,7 +716,7 @@ struct xfs_dir3_free_hdr { __be32 firstdb; /* db of first entry */ __be32 nvalid; /* count of valid entries */ __be32 nused; /* count of used entries */ - __be32 pad; /* 64 bit alignment. */ + __be32 pad; /* 64 bit alignment */ }; struct xfs_dir3_free { From ade1335afef556df6538eb02e8c0dc91fbd9cc37 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 12 Jun 2013 12:19:08 +1000 Subject: [PATCH 31/61] xfs: ensure btree root split sets blkno correctly For CRC enabled filesystems, the BMBT is rooted in an inode, so it passes through a different code path on root splits than the freespace and inode btrees. This is much less traversed by xfstests than the other trees. When testing on a 1k block size filesystem, I've been seeing ASSERT failures in generic/234 like: XFS: Assertion failed: cur->bc_btnum != XFS_BTNUM_BMAP || cur->bc_private.b.allocated == 0, file: fs/xfs/xfs_btree.c, line: 317 which are generally preceded by a lblock check failure. I noticed this in the bmbt stats: $ pminfo -f xfs.btree.block_map xfs.btree.block_map.lookup value 39135 xfs.btree.block_map.compare value 268432 xfs.btree.block_map.insrec value 15786 xfs.btree.block_map.delrec value 13884 xfs.btree.block_map.newroot value 2 xfs.btree.block_map.killroot value 0 ..... Very little coverage of root splits and merges. Indeed, on a 4k filesystem, block_map.newroot and block_map.killroot are both zero. i.e. the code is not exercised at all, and it's the only generic btree infrastructure operation that is not exercised by a default run of xfstests. Turns out that on a 1k filesystem, generic/234 accounts for one of those two root splits, and that is somewhat of a smoking gun. In fact, it's the same problem we saw in the directory/attr code where headers are memcpy()d from one block to another without updating the self describing metadata. Simple fix - when copying the header out of the root block, make sure the block number is updated correctly. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_btree.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/fs/xfs/xfs_btree.c b/fs/xfs/xfs_btree.c index 8804b8a3c310..0903960410a2 100644 --- a/fs/xfs/xfs_btree.c +++ b/fs/xfs/xfs_btree.c @@ -2544,7 +2544,17 @@ xfs_btree_new_iroot( if (error) goto error0; + /* + * we can't just memcpy() the root in for CRC enabled btree blocks. + * In that case have to also ensure the blkno remains correct + */ memcpy(cblock, block, xfs_btree_block_len(cur)); + if (cur->bc_flags & XFS_BTREE_CRC_BLOCKS) { + if (cur->bc_flags & XFS_BTREE_LONG_PTRS) + cblock->bb_u.l.bb_blkno = cpu_to_be64(cbp->b_bn); + else + cblock->bb_u.s.bb_blkno = cpu_to_be64(cbp->b_bn); + } be16_add_cpu(&block->bb_level, 1); xfs_btree_set_numrecs(block, 1); From 9222a9cf86c0d64ffbedf567412b55da18763aa3 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 12 Jun 2013 12:19:06 +1000 Subject: [PATCH 32/61] xfs: don't shutdown log recovery on validation errors Unfortunately, we cannot guarantee that items logged multiple times and replayed by log recovery do not take objects back in time. When they are taken back in time, the go into an intermediate state which is corrupt, and hence verification that occurs on this intermediate state causes log recovery to abort with a corruption shutdown. Instead of causing a shutdown and unmountable filesystem, don't verify post-recovery items before they are written to disk. This is less than optimal, but there is no way to detect this issue for non-CRC filesystems If log recovery successfully completes, this will be undone and the object will be consistent by subsequent transactions that are replayed, so in most cases we don't need to take drastic action. For CRC enabled filesystems, leave the verifiers in place - we need to call them to recalculate the CRCs on the objects anyway. This recovery problem can be solved for such filesystems - we have a LSN stamped in all metadata at writeback time that we can to determine whether the item should be replayed or not. This is a separate piece of work, so is not addressed by this patch. Signed-off-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_log_recover.c | 19 +++++++++++++++++-- 1 file changed, 17 insertions(+), 2 deletions(-) diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 45a85ff84da1..7cf5e4eafe28 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -1845,7 +1845,13 @@ xlog_recover_do_inode_buffer( xfs_agino_t *buffer_nextp; trace_xfs_log_recover_buf_inode_buf(mp->m_log, buf_f); - bp->b_ops = &xfs_inode_buf_ops; + + /* + * Post recovery validation only works properly on CRC enabled + * filesystems. + */ + if (xfs_sb_version_hascrc(&mp->m_sb)) + bp->b_ops = &xfs_inode_buf_ops; inodes_per_buf = BBTOB(bp->b_io_length) >> mp->m_sb.sb_inodelog; for (i = 0; i < inodes_per_buf; i++) { @@ -2205,7 +2211,16 @@ xlog_recover_do_reg_buffer( /* Shouldn't be any more regions */ ASSERT(i == item->ri_total); - xlog_recovery_validate_buf_type(mp, bp, buf_f); + /* + * We can only do post recovery validation on items on CRC enabled + * fielsystems as we need to know when the buffer was written to be able + * to determine if we should have replayed the item. If we replay old + * metadata over a newer buffer, then it will enter a temporarily + * inconsistent state resulting in verification failures. Hence for now + * just avoid the verification stage for non-crc filesystems + */ + if (xfs_sb_version_hascrc(&mp->m_sb)) + xlog_recovery_validate_buf_type(mp, bp, buf_f); } /* From b2410e92b70507e054deb6cdb605ee1d61fc7c95 Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Thu, 13 Jun 2013 19:37:50 -0700 Subject: [PATCH 33/61] xfs: Convert use of typedef ctl_table to struct ctl_table This typedef is unnecessary and should just be removed. Signed-off-by: Joe Perches Acked-by: Dave Chinner Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_sysctl.c | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-) diff --git a/fs/xfs/xfs_sysctl.c b/fs/xfs/xfs_sysctl.c index 2801b5ce6cdb..1743b9f8e23d 100644 --- a/fs/xfs/xfs_sysctl.c +++ b/fs/xfs/xfs_sysctl.c @@ -25,11 +25,11 @@ static struct ctl_table_header *xfs_table_header; #ifdef CONFIG_PROC_FS STATIC int xfs_stats_clear_proc_handler( - ctl_table *ctl, - int write, - void __user *buffer, - size_t *lenp, - loff_t *ppos) + struct ctl_table *ctl, + int write, + void __user *buffer, + size_t *lenp, + loff_t *ppos) { int c, ret, *valp = ctl->data; __uint32_t vn_active; @@ -55,11 +55,11 @@ xfs_stats_clear_proc_handler( STATIC int xfs_panic_mask_proc_handler( - ctl_table *ctl, - int write, - void __user *buffer, - size_t *lenp, - loff_t *ppos) + struct ctl_table *ctl, + int write, + void __user *buffer, + size_t *lenp, + loff_t *ppos) { int ret, *valp = ctl->data; @@ -74,7 +74,7 @@ xfs_panic_mask_proc_handler( } #endif /* CONFIG_PROC_FS */ -static ctl_table xfs_table[] = { +static struct ctl_table xfs_table[] = { { .procname = "irix_sgid_inherit", .data = &xfs_params.sgid_inherit.val, @@ -227,7 +227,7 @@ static ctl_table xfs_table[] = { {} }; -static ctl_table xfs_dir_table[] = { +static struct ctl_table xfs_dir_table[] = { { .procname = "xfs", .mode = 0555, @@ -236,7 +236,7 @@ static ctl_table xfs_dir_table[] = { {} }; -static ctl_table xfs_root_table[] = { +static struct ctl_table xfs_root_table[] = { { .procname = "fs", .mode = 0555, From 897366f0e4bdc320ade81c3c4035977c99de9aad Mon Sep 17 00:00:00 2001 From: Jie Liu Date: Tue, 14 May 2013 22:50:21 +0800 Subject: [PATCH 34/61] xfs: Remove redundant error variable from xfs_growfs_data_private() Commit eab4e633 "xfs: uncached buffer reads need to return an error". Remove redundant error variable, using the function level error variable to store bp->b_error instead. Signed-off-by: Jie Liu Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_fsops.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index 3c3644ea825b..614eb0cc3608 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -176,7 +176,7 @@ xfs_growfs_data_private( if (!bp) return EIO; if (bp->b_error) { - int error = bp->b_error; + error = bp->b_error; xfs_buf_relse(bp); return error; } From 34d7f603b93be0ac38c2464219f5a2795d76c8ee Mon Sep 17 00:00:00 2001 From: Jie Liu Date: Thu, 2 May 2013 19:27:53 +0800 Subject: [PATCH 35/61] xfs: Don't keep silent if sunit/swidth can not be changed via mount As per the mount man page, sunit and swidth can be changed via mount options. For XFS, on the face of it, those options seems works if the specified alignments is properly, e.g. # mount -o sunit=4096,swidth=8192 /dev/sdb1 /mnt # mount | grep sdb1 /dev/sdb1 on /mnt type xfs (rw,sunit=4096,swidth=8192) However, neither sunit nor swidth is shown from the xfs_info output. # xfs_info /mnt meta-data=/dev/sdb1 isize=256 agcount=4, agsize=262144 blks = sectsz=512 attr=2 data = bsize=4096 blocks=1048576, imaxpct=25 = sunit=0 swidth=0 blks ^^^^^^^^^^^^^^^^^^^^^^^^^^ naming =version 2 bsize=4096 ascii-ci=0 log =internal bsize=4096 blocks=2560, version=2 = sectsz=512 sunit=0 blks, lazy-count=1 realtime =none extsz=4096 blocks=0, rtextents=0 The reason is that the alignment can only be changed if the relevant super block is already configured with alignments, otherwise, the given value is silently ignored. With this fix, the attempt to mount a storage without strip alignment setup on a super block will get an error with a warning in syslog to indicate the true cause, e.g. # mount -o sunit=4096,swidth=8192 /dev/sdb1 /mnt mount: wrong fs type, bad option, bad superblock on /dev/sdb1, missing codepage or helper program, or other error In some cases useful info is found in syslog - try dmesg | tail or so ....... XFS (sdb1): cannot change alignment: superblock does not support data alignment Signed-off-by: Jie Liu Cc: Mark Tinguely Cc: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_mount.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index e8e310c05097..6a19434eba2a 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -1039,6 +1039,10 @@ xfs_update_alignment(xfs_mount_t *mp) sbp->sb_width = mp->m_swidth; mp->m_update_flags |= XFS_SB_WIDTH; } + } else { + xfs_warn(mp, + "cannot change alignment: superblock does not support data alignment"); + return XFS_ERROR(EINVAL); } } else if ((mp->m_flags & XFS_MOUNT_NOALIGN) != XFS_MOUNT_NOALIGN && xfs_sb_version_hasdalign(&mp->m_sb)) { From 1ebdf3611c8968e7202c47c2dcb2d36986c44cb0 Mon Sep 17 00:00:00 2001 From: Jeff Liu Date: Thu, 2 May 2013 19:23:20 +0800 Subject: [PATCH 36/61] xfs: Remove struct xfs_chash from xfs_mount Remove struct xfs_chash from struct xfs_mount as there is no user of it nowadays. Signed-off-by: Jie Liu Reviewed-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_mount.h | 2 -- 1 file changed, 2 deletions(-) diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index b004cecdfb04..de7abf677abe 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -192,8 +192,6 @@ typedef struct xfs_mount { xfs_dablk_t m_dirleafblk; /* blockno of dir non-data v2 */ xfs_dablk_t m_dirfreeblk; /* blockno of dirfreeindex v2 */ uint m_chsize; /* size of next field */ - struct xfs_chash *m_chash; /* fs private inode per-cluster - * hash table */ atomic_t m_active_trans; /* number trans frozen */ #ifdef HAVE_PERCPU_SB xfs_icsb_cnts_t __percpu *m_sb_cnts; /* per-cpu superblock counters */ From 725eb1eb2ae88c200466fec34bcf1fbce4b8eca3 Mon Sep 17 00:00:00 2001 From: Mark Tinguely Date: Mon, 17 Jun 2013 15:35:57 -0500 Subject: [PATCH 37/61] xfs: fix the symbolic link assert in xfs_ifree Adding an extended attribute to a symbolic link can force that link to an remote extent. xfs_inactive() incorrectly assumes that any symbolic link small enough to be in the inode core is incore, resulting in the remote extent to not be removed. xfs_ifree() will assert on presence of this leaked remote extent. Signed-off-by: Mark Tinguely Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_symlink.c | 48 +++++++++++++++++++++++++++++++++++++++++-- fs/xfs/xfs_symlink.h | 2 +- fs/xfs/xfs_trace.h | 1 + fs/xfs/xfs_vnodeops.c | 15 +++----------- 4 files changed, 51 insertions(+), 15 deletions(-) diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c index 195a403e1522..738c04be2019 100644 --- a/fs/xfs/xfs_symlink.c +++ b/fs/xfs/xfs_symlink.c @@ -585,7 +585,7 @@ xfs_symlink( /* * Free a symlink that has blocks associated with it. */ -int +STATIC int xfs_inactive_symlink_rmt( xfs_inode_t *ip, xfs_trans_t **tpp) @@ -606,7 +606,7 @@ xfs_inactive_symlink_rmt( tp = *tpp; mp = ip->i_mount; - ASSERT(ip->i_d.di_size > XFS_IFORK_DSIZE(ip)); + ASSERT(ip->i_df.if_flags & XFS_IFEXTENTS); /* * We're freeing a symlink that has some * blocks allocated to it. Free the @@ -720,3 +720,47 @@ xfs_inactive_symlink_rmt( error0: return error; } + +/* + * xfs_inactive_symlink - free a symlink + */ +int +xfs_inactive_symlink( + struct xfs_inode *ip, + struct xfs_trans **tp) +{ + struct xfs_mount *mp = ip->i_mount; + int pathlen; + + trace_xfs_inactive_symlink(ip); + + ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + + if (XFS_FORCED_SHUTDOWN(mp)) + return XFS_ERROR(EIO); + + /* + * Zero length symlinks _can_ exist. + */ + pathlen = (int)ip->i_d.di_size; + if (!pathlen) + return 0; + + if (pathlen < 0 || pathlen > MAXPATHLEN) { + xfs_alert(mp, "%s: inode (0x%llx) bad symlink length (%d)", + __func__, (unsigned long long)ip->i_ino, pathlen); + ASSERT(0); + return XFS_ERROR(EFSCORRUPTED); + } + + if (ip->i_df.if_flags & XFS_IFINLINE) { + if (ip->i_df.if_bytes > 0) + xfs_idata_realloc(ip, -(ip->i_df.if_bytes), + XFS_DATA_FORK); + ASSERT(ip->i_df.if_bytes == 0); + return 0; + } + + /* remove the remote symlink */ + return xfs_inactive_symlink_rmt(ip, tp); +} diff --git a/fs/xfs/xfs_symlink.h b/fs/xfs/xfs_symlink.h index b39398d2097c..374394880c01 100644 --- a/fs/xfs/xfs_symlink.h +++ b/fs/xfs/xfs_symlink.h @@ -60,7 +60,7 @@ extern const struct xfs_buf_ops xfs_symlink_buf_ops; int xfs_symlink(struct xfs_inode *dp, struct xfs_name *link_name, const char *target_path, umode_t mode, struct xfs_inode **ipp); int xfs_readlink(struct xfs_inode *ip, char *link); -int xfs_inactive_symlink_rmt(struct xfs_inode *ip, struct xfs_trans **tpp); +int xfs_inactive_symlink(struct xfs_inode *ip, struct xfs_trans **tpp); #endif /* __KERNEL__ */ #endif /* __XFS_SYMLINK_H */ diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index aa4db3307d36..e31867270077 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -571,6 +571,7 @@ DEFINE_INODE_EVENT(xfs_iget_miss); DEFINE_INODE_EVENT(xfs_getattr); DEFINE_INODE_EVENT(xfs_setattr); DEFINE_INODE_EVENT(xfs_readlink); +DEFINE_INODE_EVENT(xfs_inactive_symlink); DEFINE_INODE_EVENT(xfs_alloc_file_space); DEFINE_INODE_EVENT(xfs_free_file_space); DEFINE_INODE_EVENT(xfs_readdir); diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 0176bb21f09a..42c0ef288aeb 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -322,18 +322,9 @@ xfs_inactive( xfs_trans_ijoin(tp, ip, 0); if (S_ISLNK(ip->i_d.di_mode)) { - /* - * Zero length symlinks _can_ exist. - */ - if (ip->i_d.di_size > XFS_IFORK_DSIZE(ip)) { - error = xfs_inactive_symlink_rmt(ip, &tp); - if (error) - goto out_cancel; - } else if (ip->i_df.if_bytes > 0) { - xfs_idata_realloc(ip, -(ip->i_df.if_bytes), - XFS_DATA_FORK); - ASSERT(ip->i_df.if_bytes == 0); - } + error = xfs_inactive_symlink(ip, &tp); + if (error) + goto out_cancel; } else if (truncate) { ip->i_d.di_size = 0; xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); From 635c4d0bd94512de3aafa7ae8f45c1dce0fc1ffe Mon Sep 17 00:00:00 2001 From: Jie Liu Date: Thu, 6 Jun 2013 06:24:01 -0700 Subject: [PATCH 38/61] xfs: return FIEMAP_EXTENT_UNKNOWN for delayed allocation extent For FIEMAP ioctl(2), if an extent is in delayed allocation state, we need to return the FIEMAP_EXTENT_UNKNOWN flag except the FIEMAP_EXTENT_DELALLOC because its data location is unknown. Signed-off-by: Jie Liu Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_iops.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index ca9ecaa81112..c69bbc493cb0 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -987,7 +987,8 @@ xfs_fiemap_format( if (bmv->bmv_oflags & BMV_OF_PREALLOC) fiemap_flags |= FIEMAP_EXTENT_UNWRITTEN; else if (bmv->bmv_oflags & BMV_OF_DELALLOC) { - fiemap_flags |= FIEMAP_EXTENT_DELALLOC; + fiemap_flags |= (FIEMAP_EXTENT_DELALLOC | + FIEMAP_EXTENT_UNKNOWN); physical = 0; /* no block yet */ } if (bmv->bmv_oflags & BMV_OF_LAST) From 2fb8b5027dbde32a45edf5f3d7ee082be9261d93 Mon Sep 17 00:00:00 2001 From: Jie Liu Date: Fri, 3 May 2013 15:41:19 +0800 Subject: [PATCH 39/61] xfs: Remove two dead transaction log reservaion macros Upstream commit 5b292ae3a951a58e32119d73c7ac8f5bec7395a3 xfs: make use of xfs_calc_buf_res() in xfs_trans.c Beginning from above commit, neither XFS_ALLOCFREE_LOG_RES() nor XFS_DIROP_LOG_RES() is used by those routines for calculating transaction space reservations, so it's safe to remove them now. Also, with a slightly update for the relevant comments to reflect the ideas of why those log count numbers should be. Signed-off-by: Jie Liu Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_trans.h | 11 +++-------- 1 file changed, 3 insertions(+), 8 deletions(-) diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index a44dba5b2cdb..6d526569820c 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -210,23 +210,18 @@ struct xfs_log_item_desc { /* * Per-extent log reservation for the allocation btree changes * involved in freeing or allocating an extent. - * 2 trees * (2 blocks/level * max depth - 1) * block size + * 2 trees * (2 blocks/level * max depth - 1) */ -#define XFS_ALLOCFREE_LOG_RES(mp,nx) \ - ((nx) * (2 * XFS_FSB_TO_B((mp), 2 * XFS_AG_MAXLEVELS(mp) - 1))) #define XFS_ALLOCFREE_LOG_COUNT(mp,nx) \ ((nx) * (2 * (2 * XFS_AG_MAXLEVELS(mp) - 1))) /* * Per-directory log reservation for any directory change. - * dir blocks: (1 btree block per level + data block + free block) * dblock size - * bmap btree: (levels + 2) * max depth * block size + * dir blocks: (1 btree block per level + data block + free block) + * bmap btree: (levels + 2) * max depth * v2 directory blocks can be fragmented below the dirblksize down to the fsb * size, so account for that in the DAENTER macros. */ -#define XFS_DIROP_LOG_RES(mp) \ - (XFS_FSB_TO_B(mp, XFS_DAENTER_BLOCKS(mp, XFS_DATA_FORK)) + \ - (XFS_FSB_TO_B(mp, XFS_DAENTER_BMAPS(mp, XFS_DATA_FORK) + 1))) #define XFS_DIROP_LOG_COUNT(mp) \ (XFS_DAENTER_BLOCKS(mp, XFS_DATA_FORK) + \ XFS_DAENTER_BMAPS(mp, XFS_DATA_FORK) + 1) From 39a45d8463d98ea57347b871641136be64b216a9 Mon Sep 17 00:00:00 2001 From: Jie Liu Date: Thu, 2 May 2013 19:27:47 +0800 Subject: [PATCH 40/61] xfs: Remove XFS_MOUNT_RETERR XFS_MOUNT_RETERR is going to be set at xfs_parseargs() if mp->m_dalign is enabled, so any time we enter "if (mp->m_dalign)" branch in xfs_update_alignment(), XFS_MOUNT_RETERR is set and so we always be emitting a warning and returning an error. Hence, we can remove it and get rid of a couple of redundant check up against it at xfs_upate_alignment(). Thanks Dave Chinner for the suggestions of simplify the code in xfs_parseargs(). Signed-off-by: Jie Liu Cc: Dave Chinner Cc: Mark Tinguely Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_mount.c | 37 +++++++++++-------------------------- fs/xfs/xfs_mount.h | 2 -- fs/xfs/xfs_super.c | 11 +++-------- 3 files changed, 14 insertions(+), 36 deletions(-) diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 6a19434eba2a..2978bb4cc72a 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -987,42 +987,27 @@ xfs_update_alignment(xfs_mount_t *mp) */ if ((BBTOB(mp->m_dalign) & mp->m_blockmask) || (BBTOB(mp->m_swidth) & mp->m_blockmask)) { - if (mp->m_flags & XFS_MOUNT_RETERR) { - xfs_warn(mp, "alignment check failed: " - "(sunit/swidth vs. blocksize)"); - return XFS_ERROR(EINVAL); - } - mp->m_dalign = mp->m_swidth = 0; + xfs_warn(mp, + "alignment check failed: sunit/swidth vs. blocksize(%d)", + sbp->sb_blocksize); + return XFS_ERROR(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)) { - if (mp->m_flags & XFS_MOUNT_RETERR) { - xfs_warn(mp, "alignment check failed: " - "(sunit/swidth vs. ag size)"); - return XFS_ERROR(EINVAL); - } xfs_warn(mp, - "stripe alignment turned off: sunit(%d)/swidth(%d) " - "incompatible with agsize(%d)", - mp->m_dalign, mp->m_swidth, - sbp->sb_agblocks); - - mp->m_dalign = 0; - mp->m_swidth = 0; + "alignment check failed: sunit/swidth vs. agsize(%d)", + sbp->sb_agblocks); + return XFS_ERROR(EINVAL); } else if (mp->m_dalign) { mp->m_swidth = XFS_BB_TO_FSBT(mp, mp->m_swidth); } else { - if (mp->m_flags & XFS_MOUNT_RETERR) { - xfs_warn(mp, "alignment check failed: " - "sunit(%d) less than bsize(%d)", - mp->m_dalign, - mp->m_blockmask +1); - return XFS_ERROR(EINVAL); - } - mp->m_swidth = 0; + xfs_warn(mp, + "alignment check failed: sunit(%d) less than bsize(%d)", + mp->m_dalign, sbp->sb_blocksize); + return XFS_ERROR(EINVAL); } } diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index de7abf677abe..4e374d4a9189 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -227,8 +227,6 @@ typedef struct xfs_mount { operations, typically for disk errors in metadata */ #define XFS_MOUNT_DISCARD (1ULL << 5) /* discard unused blocks */ -#define XFS_MOUNT_RETERR (1ULL << 6) /* return alignment errors to - user */ #define XFS_MOUNT_NOALIGN (1ULL << 7) /* turn off stripe alignment allocations */ #define XFS_MOUNT_ATTR2 (1ULL << 8) /* allow use of attr2 format */ diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 3033ba5e9762..14924099bcb8 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -439,20 +439,15 @@ xfs_parseargs( } done: - if (!(mp->m_flags & XFS_MOUNT_NOALIGN)) { + if (dsunit && !(mp->m_flags & XFS_MOUNT_NOALIGN)) { /* * At this point the superblock has not been read * in, therefore we do not know the block size. * Before the mount call ends we will convert * these to FSBs. */ - if (dsunit) { - mp->m_dalign = dsunit; - mp->m_flags |= XFS_MOUNT_RETERR; - } - - if (dswidth) - mp->m_swidth = dswidth; + mp->m_dalign = dsunit; + mp->m_swidth = dswidth; } if (mp->m_logbufs != -1 && From 427d9fe233ccec3a293be4bcf07f9bac12463a99 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Tue, 27 Mar 2012 15:40:26 -0500 Subject: [PATCH 41/61] xfs: check on-disk (not incore) btree root size in dfrag.c xfs_swap_extents_check_format() contains checks to make sure that original and the temporary files during defrag are compatible; Gabriel VLASIU ran into a case where xfs_fsr returned EINVAL because the tests found the btree root to be of size 120, while the fork offset was only 104; IOW, they overlapped. However, this is just due to an error in the xfs_swap_extents_check_format() tests, because it is checking the in-memory btree root size against the on-disk fork offset. We should be checking the on-disk sizes in both cases. This patch adds a new macro to calculate this size, and uses it in the tests. With this change, the filesystem image provided by Gabriel allows for proper file degragmentation. Reported-by: Gabriel VLASIU Signed-off-by: Eric Sandeen Reviewed-by: Dave Chinner Signed-off-by: Ben Myers --- fs/xfs/xfs_bmap_btree.h | 2 ++ fs/xfs/xfs_dfrag.c | 8 +++++--- 2 files changed, 7 insertions(+), 3 deletions(-) diff --git a/fs/xfs/xfs_bmap_btree.h b/fs/xfs/xfs_bmap_btree.h index 70c43d9f72c1..1b726d626941 100644 --- a/fs/xfs/xfs_bmap_btree.h +++ b/fs/xfs/xfs_bmap_btree.h @@ -196,6 +196,8 @@ typedef __be64 xfs_bmbt_ptr_t, xfs_bmdr_ptr_t; #define XFS_BMDR_SPACE_CALC(nrecs) \ (int)(sizeof(xfs_bmdr_block_t) + \ ((nrecs) * (sizeof(xfs_bmbt_key_t) + sizeof(xfs_bmbt_ptr_t)))) +#define XFS_BMAP_BMDR_SPACE(bb) \ + (XFS_BMDR_SPACE_CALC(be16_to_cpu((bb)->bb_numrecs))) /* * Maximum number of bmap btree levels. diff --git a/fs/xfs/xfs_dfrag.c b/fs/xfs/xfs_dfrag.c index c407e1ccff43..e36445ceaf80 100644 --- a/fs/xfs/xfs_dfrag.c +++ b/fs/xfs/xfs_dfrag.c @@ -24,6 +24,9 @@ #include "xfs_ag.h" #include "xfs_mount.h" #include "xfs_bmap_btree.h" +#include "xfs_alloc_btree.h" +#include "xfs_ialloc_btree.h" +#include "xfs_btree.h" #include "xfs_dinode.h" #include "xfs_inode.h" #include "xfs_inode_item.h" @@ -182,7 +185,7 @@ xfs_swap_extents_check_format( */ if (tip->i_d.di_format == XFS_DINODE_FMT_BTREE) { if (XFS_IFORK_BOFF(ip) && - tip->i_df.if_broot_bytes > XFS_IFORK_BOFF(ip)) + XFS_BMAP_BMDR_SPACE(tip->i_df.if_broot) > XFS_IFORK_BOFF(ip)) return EINVAL; if (XFS_IFORK_NEXTENTS(tip, XFS_DATA_FORK) <= XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK)) @@ -192,9 +195,8 @@ xfs_swap_extents_check_format( /* Reciprocal target->temp btree format checks */ if (ip->i_d.di_format == XFS_DINODE_FMT_BTREE) { if (XFS_IFORK_BOFF(tip) && - ip->i_df.if_broot_bytes > XFS_IFORK_BOFF(tip)) + XFS_BMAP_BMDR_SPACE(ip->i_df.if_broot) > XFS_IFORK_BOFF(tip)) return EINVAL; - if (XFS_IFORK_NEXTENTS(ip, XFS_DATA_FORK) <= XFS_IFORK_MAXEXT(tip, XFS_DATA_FORK)) return EINVAL; From 43df2ee659d3d854d17686f8f38fc1071b25774b Mon Sep 17 00:00:00 2001 From: Jie Liu Date: Mon, 24 Jun 2013 13:37:23 +0800 Subject: [PATCH 42/61] xfs: Remove the left function variable from xfs_ialloc_get_rec() This patch clean out the left function variable as it is useless to xfs_ialloc_get_rec(). Signed-off-by: Jie Liu Reviewed-by: Dave Chinner Signed-off-by: Ben Myers --- fs/xfs/xfs_ialloc.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index c8f5ae1debf2..4345c53c729c 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -615,8 +615,7 @@ xfs_ialloc_get_rec( struct xfs_btree_cur *cur, xfs_agino_t agino, xfs_inobt_rec_incore_t *rec, - int *done, - int left) + int *done) { int error; int i; @@ -724,12 +723,12 @@ xfs_dialloc_ag( pag->pagl_leftrec != NULLAGINO && pag->pagl_rightrec != NULLAGINO) { error = xfs_ialloc_get_rec(tcur, pag->pagl_leftrec, - &trec, &doneleft, 1); + &trec, &doneleft); if (error) goto error1; error = xfs_ialloc_get_rec(cur, pag->pagl_rightrec, - &rec, &doneright, 0); + &rec, &doneright); if (error) goto error1; } else { From 80a4049813a2ae0977d8e5db78e711c7f21c420b Mon Sep 17 00:00:00 2001 From: Jie Liu Date: Mon, 24 Jun 2013 13:38:34 +0800 Subject: [PATCH 43/61] xfs: Remove dead function prototype xfs_sync_inode_grab() Remove dead function prototype xfs_sync_inode_grab() from xfs_icache.h. Signed-off-by: Jie Liu Reviewed-by: Dave Chinner Signed-off-by: Ben Myers --- fs/xfs/xfs_icache.h | 1 - 1 file changed, 1 deletion(-) diff --git a/fs/xfs/xfs_icache.h b/fs/xfs/xfs_icache.h index e0f138c70a2f..a01afbb3909a 100644 --- a/fs/xfs/xfs_icache.h +++ b/fs/xfs/xfs_icache.h @@ -40,7 +40,6 @@ void xfs_inode_clear_eofblocks_tag(struct xfs_inode *ip); int xfs_icache_free_eofblocks(struct xfs_mount *, struct xfs_eofblocks *); void xfs_eofblocks_worker(struct work_struct *); -int xfs_sync_inode_grab(struct xfs_inode *ip); int xfs_inode_ag_iterator(struct xfs_mount *mp, int (*execute)(struct xfs_inode *ip, struct xfs_perag *pag, int flags, void *args), From cbb2864aa48977205c76291ba5a23331393b2578 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Jun 2013 16:04:46 +1000 Subject: [PATCH 44/61] xfs: add pluging for bulkstat readahead I was running some tests on bulkstat on CRC enabled filesystems when I noticed that all the IO being issued was 8k in size, regardless of the fact taht we are issuing sequential 8k buffers for inodes clusters. The IO size should be 16k for 256 byte inodes, and 32k for 512 byte inodes, but this wasn't happening. blktrace showed that there was an explict plug and unplug happening around each readahead IO from _xfs_buf_ioapply, and the unplug was causing the IO to be issued immediately. Hence no opportunity was being given to the elevator to merge adjacent readahead requests and dispatch them as a single IO. Add plugging around the inode chunk readahead dispatch loop in bulkstat to ensure that we don't unplug the queue between adjacent inode buffer readahead IOs and so we get fewer, larger IO requests hitting the storage subsystem for bulkstat. Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_itable.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c index 2ea7d402188d..06d004d85bf4 100644 --- a/fs/xfs/xfs_itable.c +++ b/fs/xfs/xfs_itable.c @@ -383,11 +383,13 @@ xfs_bulkstat( * Also start read-ahead now for this chunk. */ if (r.ir_freecount < XFS_INODES_PER_CHUNK) { + struct blk_plug plug; /* * Loop over all clusters in the next chunk. * Do a readahead if there are any allocated * inodes in that cluster. */ + blk_start_plug(&plug); agbno = XFS_AGINO_TO_AGBNO(mp, r.ir_startino); for (chunkidx = 0; chunkidx < XFS_INODES_PER_CHUNK; @@ -399,6 +401,7 @@ xfs_bulkstat( agbno, nbcluster, &xfs_inode_buf_ops); } + blk_finish_plug(&plug); irbp->ir_startino = r.ir_startino; irbp->ir_freecount = r.ir_freecount; irbp->ir_free = r.ir_free; From 34eefc06a06f496b92c3267a0601129a932c7174 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Jun 2013 16:04:47 +1000 Subject: [PATCH 45/61] xfs: plug directory buffer readahead Similar to bulkstat inode chunk readahead, we need to plug directory data buffer readahead during getdents to ensure that we can merge adjacent readahead requests and sort out of order requests optimally before they are dispatched. This improves the readahead efficiency and reduces the IO load it generates as the IO patterns are significantly better for both contiguous and fragmented directories. Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_dir2_leaf.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/fs/xfs/xfs_dir2_leaf.c b/fs/xfs/xfs_dir2_leaf.c index da71a1819d78..bb22aac9fc01 100644 --- a/fs/xfs/xfs_dir2_leaf.c +++ b/fs/xfs/xfs_dir2_leaf.c @@ -1108,6 +1108,7 @@ xfs_dir2_leaf_readbuf( struct xfs_mount *mp = dp->i_mount; struct xfs_buf *bp = *bpp; struct xfs_bmbt_irec *map = mip->map; + struct blk_plug plug; int error = 0; int length; int i; @@ -1236,6 +1237,7 @@ xfs_dir2_leaf_readbuf( /* * Do we need more readahead? */ + blk_start_plug(&plug); for (mip->ra_index = mip->ra_offset = i = 0; mip->ra_want > mip->ra_current && i < mip->map_blocks; i += mp->m_dirblkfsbs) { @@ -1287,6 +1289,7 @@ xfs_dir2_leaf_readbuf( } } } + blk_finish_plug(&plug); out: *bpp = bp; From 133eeb1747c33b6d75483c074b27d4e5e02286dc Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Jun 2013 16:04:48 +1000 Subject: [PATCH 46/61] xfs: don't use speculative prealloc for small files Dedicated small file workloads have been seeing significant free space fragmentation causing premature inode allocation failure when large inode sizes are in use. A particular test case showed that a workload that runs to a real ENOSPC on 256 byte inodes would fail inode allocation with ENOSPC about about 80% full with 512 byte inodes, and at about 50% full with 1024 byte inodes. The same workload, when run with -o allocsize=4096 on 1024 byte inodes would run to being 100% full before giving ENOSPC. That is, no freespace fragmentation at all. The issue was caused by the specific IO pattern the application had - the framework it was using did not support direct IO, and so it was emulating it by using fadvise(DONT_NEED). The result was that the data was getting written back before the speculative prealloc had been trimmed from memory by the close(), and so small single block files were being allocated with 2 blocks, and then having one truncated away. The result was lots of small 4k free space extents, and hence each new 8k allocation would take another 8k from contiguous free space and turn it into 4k of allocated space and 4k of free space. Hence inode allocation, which requires contiguous, aligned allocation of 16k (256 byte inodes), 32k (512 byte inodes) or 64k (1024 byte inodes) can fail to find sufficiently large freespace and hence fail while there is still lots of free space available. There's a simple fix for this, and one that has precendence in the allocator code already - don't do speculative allocation unless the size of the file is larger than a certain size. In this case, that size is the minimum default preallocation size: mp->m_writeio_blocks. And to keep with the concept of being nice to people when the files are still relatively small, cap the prealloc to mp->m_writeio_blocks until the file goes over a stripe unit is size, at which point we'll fall back to the current behaviour based on the last extent size. This will effectively turn off speculative prealloc for very small files, keep preallocation low for small files, and behave as it currently does for any file larger than a stripe unit. This completely avoids the freespace fragmentation problem this particular IO pattern was causing. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_iomap.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index 8f8aaee7f379..6a7096422295 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -283,6 +283,15 @@ xfs_iomap_eof_want_preallocate( if (offset + count <= XFS_ISIZE(ip)) return 0; + /* + * If the file is smaller than the minimum prealloc and we are using + * dynamic preallocation, don't do any preallocation at all as it is + * likely this is the only write to the file that is going to be done. + */ + if (!(mp->m_flags & XFS_MOUNT_DFLT_IOSIZE) && + XFS_ISIZE(ip) < XFS_FSB_TO_B(mp, mp->m_writeio_blocks)) + return 0; + /* * If there are any real blocks past eof, then don't * do any speculative allocation. @@ -345,6 +354,10 @@ xfs_iomap_eof_prealloc_initial_size( if (mp->m_flags & XFS_MOUNT_DFLT_IOSIZE) return 0; + /* If the file is small, then use the minimum prealloc */ + if (XFS_ISIZE(ip) < XFS_FSB_TO_B(mp, mp->m_dalign)) + return 0; + /* * As we write multiple pages, the offset will always align to the * start of a page and hence point to a hole at EOF. i.e. if the size is From cca9f93a52d2ead50b5da59ca83d5f469ee4be5f Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Jun 2013 16:04:49 +1000 Subject: [PATCH 47/61] xfs: don't do IO when creating an new inode When we are allocating a new inode, we read the inode cluster off disk to increment the generation number. We are already using a random generation number for newly allocated inodes, so if we are not using the ikeep mode, we can just generate a new generation number when we initialise the newly allocated inode. This avoids the need for reading the inode buffer during inode creation. This will speed up allocation of inodes in cold, partially allocated clusters as they will no longer need to be read from disk during allocation. It will also reduce the CPU overhead of inode allocation by not having the process the buffer read, even on cache hits. Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_inode.c | 36 ++++++++++++++++++++++++++++-------- 1 file changed, 28 insertions(+), 8 deletions(-) diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 7f7be5f98f52..d1f76da6f8bf 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1028,6 +1028,11 @@ xfs_dinode_calc_crc( /* * Read the disk inode attributes into the in-core inode structure. + * + * If we are initialising a new inode and we are not utilising the + * XFS_MOUNT_IKEEP inode cluster mode, we can simple build the new inode core + * with a random generation number. If we are keeping inodes around, we need to + * read the inode cluster to get the existing generation number off disk. */ int xfs_iread( @@ -1047,6 +1052,22 @@ xfs_iread( if (error) return error; + /* shortcut IO on inode allocation if possible */ + if ((iget_flags & XFS_IGET_CREATE) && + !(mp->m_flags & XFS_MOUNT_IKEEP)) { + /* initialise the on-disk inode core */ + memset(&ip->i_d, 0, sizeof(ip->i_d)); + ip->i_d.di_magic = XFS_DINODE_MAGIC; + ip->i_d.di_gen = prandom_u32(); + if (xfs_sb_version_hascrc(&mp->m_sb)) { + ip->i_d.di_version = 3; + ip->i_d.di_ino = ip->i_ino; + uuid_copy(&ip->i_d.di_uuid, &mp->m_sb.sb_uuid); + } else + ip->i_d.di_version = 2; + return 0; + } + /* * Get pointers to the on-disk inode and the buffer containing it. */ @@ -1133,17 +1154,16 @@ xfs_iread( xfs_buf_set_ref(bp, XFS_INO_REF); /* - * Use xfs_trans_brelse() to release the buffer containing the - * on-disk inode, because it was acquired with xfs_trans_read_buf() - * in xfs_imap_to_bp() above. If tp is NULL, this is just a normal + * Use xfs_trans_brelse() to release the buffer containing the on-disk + * inode, because it was acquired with xfs_trans_read_buf() in + * xfs_imap_to_bp() above. If tp is NULL, this is just a normal * brelse(). If we're within a transaction, then xfs_trans_brelse() * will only release the buffer if it is not dirty within the * transaction. It will be OK to release the buffer in this case, - * because inodes on disk are never destroyed and we will be - * locking the new in-core inode before putting it in the hash - * table where other processes can find it. Thus we don't have - * to worry about the inode being changed just because we released - * the buffer. + * because inodes on disk are never destroyed and we will be locking the + * new in-core inode before putting it in the cache where other + * processes can find it. Thus we don't have to worry about the inode + * being changed just because we released the buffer. */ out_brelse: xfs_trans_brelse(tp, bp); From 1baaed8fa955ab0d23aab24477dae566ed6a105b Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Jun 2013 16:04:50 +1000 Subject: [PATCH 48/61] xfs: xfs_ifree doesn't need to modify the inode buffer Long ago, bulkstat used to read inodes directly from the backing buffer for speed. This had the unfortunate problem of being cache incoherent with unlinks, and so xfs_ifree() had to mark the inode as free directly in the backing buffer. bulkstat was changed some time ago to use inode cache coherent lookups, and so will never see unlinked inodes in it's lookups. Hence xfs_ifree() does not need to touch the inode backing buffer anymore. Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_inode.c | 32 ++++---------------------------- 1 file changed, 4 insertions(+), 28 deletions(-) diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index d1f76da6f8bf..9ecfe1e559fc 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -2048,8 +2048,6 @@ xfs_ifree( int error; int delete; xfs_ino_t first_ino; - xfs_dinode_t *dip; - xfs_buf_t *ibp; ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); ASSERT(ip->i_d.di_nlink == 0); @@ -2062,14 +2060,13 @@ xfs_ifree( * Pull the on-disk inode from the AGI unlinked list. */ error = xfs_iunlink_remove(tp, ip); - if (error != 0) { + if (error) return error; - } error = xfs_difree(tp, ip->i_ino, flist, &delete, &first_ino); - if (error != 0) { + if (error) return error; - } + ip->i_d.di_mode = 0; /* mark incore inode as free */ ip->i_d.di_flags = 0; ip->i_d.di_dmevmask = 0; @@ -2081,31 +2078,10 @@ xfs_ifree( * by reincarnations of this inode. */ ip->i_d.di_gen++; - xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - error = xfs_imap_to_bp(ip->i_mount, tp, &ip->i_imap, &dip, &ibp, - 0, 0); - if (error) - return error; - - /* - * Clear the on-disk di_mode. This is to prevent xfs_bulkstat - * from picking up this inode when it is reclaimed (its incore state - * initialzed but not flushed to disk yet). The in-core di_mode is - * already cleared and a corresponding transaction logged. - * The hack here just synchronizes the in-core to on-disk - * di_mode value in advance before the actual inode sync to disk. - * This is OK because the inode is already unlinked and would never - * change its di_mode again for this inode generation. - * This is a temporary hack that would require a proper fix - * in the future. - */ - dip->di_mode = 0; - - if (delete) { + if (delete) error = xfs_ifree_cluster(ip, tp, first_ino); - } return error; } From fd63875cc4cd60b9e5c609c24d75eaaad3e6d1c4 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Jun 2013 16:04:51 +1000 Subject: [PATCH 49/61] xfs: Introduce ordered log vector support And "ordered log vector" is a log vector that is used for tracking a log item through the CIL and into the AIL as part of the log checkpointing. These ordered log vectors are special in that they are not written to to journal in any way, and are not accounted to the checkpoint being written. The reason for this behaviour is to allow operations to attach items to transactions and have them follow the normal transactional lifecycle without actually having to write them to the journal. This allows logging of items that track high level logical changes and writing them to the log, while the physical items being modified pass through into the AIL and pin the tail of the log (and therefore the logical item in the log) until all the modified items are physically written to disk. IOWs, it allows us to write metadata without physically logging every individual change but still maintain the full transactional integrity guarantees we currently have w.r.t. crash recovery. This change modifies some of the CIL item insertion loops, as ordered log vectors introduce some new constraints as they don't track any data. One advantage of this change is that it combines two log vector chain walks into a single pass, so there is less overhead in the transaction commit pass as well. It also kills some unused code in the log vector walk loop when committing the CIL. Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_log.c | 22 ++++++++++--- fs/xfs/xfs_log.h | 2 ++ fs/xfs/xfs_log_cil.c | 75 ++++++++++++++++++++++++++++++-------------- 3 files changed, 71 insertions(+), 28 deletions(-) diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index b345a7c85153..d852a2b3e1fd 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -1963,6 +1963,10 @@ xlog_write_calc_vec_length( headers++; for (lv = log_vector; lv; lv = lv->lv_next) { + /* we don't write ordered log vectors */ + if (lv->lv_buf_len == XFS_LOG_VEC_ORDERED) + continue; + headers += lv->lv_niovecs; for (i = 0; i < lv->lv_niovecs; i++) { @@ -2216,7 +2220,7 @@ xlog_write( index = 0; lv = log_vector; vecp = lv->lv_iovecp; - while (lv && index < lv->lv_niovecs) { + while (lv && (!lv->lv_niovecs || index < lv->lv_niovecs)) { void *ptr; int log_offset; @@ -2236,13 +2240,22 @@ xlog_write( * This loop writes out as many regions as can fit in the amount * of space which was allocated by xlog_state_get_iclog_space(). */ - while (lv && index < lv->lv_niovecs) { - struct xfs_log_iovec *reg = &vecp[index]; + while (lv && (!lv->lv_niovecs || index < lv->lv_niovecs)) { + struct xfs_log_iovec *reg; struct xlog_op_header *ophdr; int start_rec_copy; int copy_len; int copy_off; + bool ordered = false; + /* ordered log vectors have no regions to write */ + if (lv->lv_buf_len == XFS_LOG_VEC_ORDERED) { + ASSERT(lv->lv_niovecs == 0); + ordered = true; + goto next_lv; + } + + reg = &vecp[index]; ASSERT(reg->i_len % sizeof(__int32_t) == 0); ASSERT((unsigned long)ptr % sizeof(__int32_t) == 0); @@ -2302,12 +2315,13 @@ xlog_write( break; if (++index == lv->lv_niovecs) { +next_lv: lv = lv->lv_next; index = 0; if (lv) vecp = lv->lv_iovecp; } - if (record_cnt == 0) { + if (record_cnt == 0 && ordered == false) { if (!lv) return 0; break; diff --git a/fs/xfs/xfs_log.h b/fs/xfs/xfs_log.h index 5caee96059df..b20918c554aa 100644 --- a/fs/xfs/xfs_log.h +++ b/fs/xfs/xfs_log.h @@ -105,6 +105,8 @@ struct xfs_log_vec { int lv_buf_len; /* size of formatted buffer */ }; +#define XFS_LOG_VEC_ORDERED (-1) + /* * Structure used to pass callback function and the function's argument * to the log manager. diff --git a/fs/xfs/xfs_log_cil.c b/fs/xfs/xfs_log_cil.c index d0833b54e55d..02b9cf3f8252 100644 --- a/fs/xfs/xfs_log_cil.c +++ b/fs/xfs/xfs_log_cil.c @@ -127,6 +127,7 @@ xlog_cil_prepare_log_vecs( int index; int len = 0; uint niovecs; + bool ordered = false; /* Skip items which aren't dirty in this transaction. */ if (!(lidp->lid_flags & XFS_LID_DIRTY)) @@ -137,14 +138,30 @@ xlog_cil_prepare_log_vecs( if (!niovecs) continue; + /* + * Ordered items need to be tracked but we do not wish to write + * them. We need a logvec to track the object, but we do not + * need an iovec or buffer to be allocated for copying data. + */ + if (niovecs == XFS_LOG_VEC_ORDERED) { + ordered = true; + niovecs = 0; + } + new_lv = kmem_zalloc(sizeof(*new_lv) + niovecs * sizeof(struct xfs_log_iovec), KM_SLEEP|KM_NOFS); + new_lv->lv_item = lidp->lid_item; + new_lv->lv_niovecs = niovecs; + if (ordered) { + /* track as an ordered logvec */ + new_lv->lv_buf_len = XFS_LOG_VEC_ORDERED; + goto next; + } + /* The allocated iovec region lies beyond the log vector. */ new_lv->lv_iovecp = (struct xfs_log_iovec *)&new_lv[1]; - new_lv->lv_niovecs = niovecs; - new_lv->lv_item = lidp->lid_item; /* build the vector array and calculate it's length */ IOP_FORMAT(new_lv->lv_item, new_lv->lv_iovecp); @@ -165,6 +182,7 @@ xlog_cil_prepare_log_vecs( } ASSERT(ptr == new_lv->lv_buf + new_lv->lv_buf_len); +next: if (!ret_lv) ret_lv = new_lv; else @@ -191,8 +209,18 @@ xfs_cil_prepare_item( if (old) { /* existing lv on log item, space used is a delta */ - ASSERT(!list_empty(&lv->lv_item->li_cil)); - ASSERT(old->lv_buf && old->lv_buf_len && old->lv_niovecs); + ASSERT((old->lv_buf && old->lv_buf_len && old->lv_niovecs) || + old->lv_buf_len == XFS_LOG_VEC_ORDERED); + + /* + * If the new item is ordered, keep the old one that is already + * tracking dirty or ordered regions + */ + if (lv->lv_buf_len == XFS_LOG_VEC_ORDERED) { + ASSERT(!lv->lv_buf); + kmem_free(lv); + return; + } *len += lv->lv_buf_len - old->lv_buf_len; *diff_iovecs += lv->lv_niovecs - old->lv_niovecs; @@ -201,10 +229,11 @@ xfs_cil_prepare_item( } else { /* new lv, must pin the log item */ ASSERT(!lv->lv_item->li_lv); - ASSERT(list_empty(&lv->lv_item->li_cil)); - *len += lv->lv_buf_len; - *diff_iovecs += lv->lv_niovecs; + if (lv->lv_buf_len != XFS_LOG_VEC_ORDERED) { + *len += lv->lv_buf_len; + *diff_iovecs += lv->lv_niovecs; + } IOP_PIN(lv->lv_item); } @@ -259,18 +288,24 @@ xlog_cil_insert_items( * We can do this safely because the context can't checkpoint until we * are done so it doesn't matter exactly how we update the CIL. */ - for (lv = log_vector; lv; lv = lv->lv_next) + spin_lock(&cil->xc_cil_lock); + for (lv = log_vector; lv; ) { + struct xfs_log_vec *next = lv->lv_next; + + ASSERT(lv->lv_item->li_lv || list_empty(&lv->lv_item->li_cil)); + lv->lv_next = NULL; + + /* + * xfs_cil_prepare_item() may free the lv, so move the item on + * the CIL first. + */ + list_move_tail(&lv->lv_item->li_cil, &cil->xc_cil); xfs_cil_prepare_item(log, lv, &len, &diff_iovecs); + lv = next; + } /* account for space used by new iovec headers */ len += diff_iovecs * sizeof(xlog_op_header_t); - - spin_lock(&cil->xc_cil_lock); - - /* move the items to the tail of the CIL */ - for (lv = log_vector; lv; lv = lv->lv_next) - list_move_tail(&lv->lv_item->li_cil, &cil->xc_cil); - ctx->nvecs += diff_iovecs; /* @@ -381,9 +416,7 @@ xlog_cil_push( struct xfs_cil_ctx *new_ctx; struct xlog_in_core *commit_iclog; struct xlog_ticket *tic; - int num_lv; int num_iovecs; - int len; int error = 0; struct xfs_trans_header thdr; struct xfs_log_iovec lhdr; @@ -428,12 +461,9 @@ xlog_cil_push( * side which is currently locked out by the flush lock. */ lv = NULL; - num_lv = 0; num_iovecs = 0; - len = 0; while (!list_empty(&cil->xc_cil)) { struct xfs_log_item *item; - int i; item = list_first_entry(&cil->xc_cil, struct xfs_log_item, li_cil); @@ -444,11 +474,7 @@ xlog_cil_push( lv->lv_next = item->li_lv; lv = item->li_lv; item->li_lv = NULL; - - num_lv++; num_iovecs += lv->lv_niovecs; - for (i = 0; i < lv->lv_niovecs; i++) - len += lv->lv_iovecp[i].i_len; } /* @@ -701,6 +727,7 @@ xfs_log_commit_cil( if (commit_lsn) *commit_lsn = log->l_cilp->xc_ctx->sequence; + /* xlog_cil_insert_items() destroys log_vector list */ xlog_cil_insert_items(log, log_vector, tp->t_ticket); /* check we didn't blow the reservation */ From 5f6bed76c0c85cb4d04885a5de00b629deee550b Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Jun 2013 16:04:52 +1000 Subject: [PATCH 50/61] xfs: Introduce an ordered buffer item If we have a buffer that we have modified but we do not wish to physically log in a transaction (e.g. we've logged a logical change), we still need to ensure that transactional integrity is maintained. Hence we must not move the tail of the log past the transaction that the buffer is associated with before the buffer is written to disk. This means these special buffers still need to be included in the transaction and added to the AIL just like a normal buffer, but we do not want the modifications to the buffer written into the transaction. IOWs, what we want is an "ordered buffer" that maintains the same transactional life cycle as a physically logged buffer, just without the transcribing of the modifications to the log. Hence we need to flag the buffer as an "ordered buffer" to avoid including it in vector size calculations or formatting during the transaction. Once the transaction is committed, the buffer appears for all intents to be the same as a physically logged buffer as it transitions through the log and AIL. Relogging will also work just fine for such an ordered buffer - the logical transaction will be replayed before the subsequent modifications that relog the buffer, so everything will be reconstructed correctly by recovery. Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_buf_item.c | 75 +++++++++++++++++++++++++++--------------- fs/xfs/xfs_buf_item.h | 4 ++- fs/xfs/xfs_trace.h | 4 +++ fs/xfs/xfs_trans.h | 1 + fs/xfs/xfs_trans_buf.c | 34 +++++++++++++++++-- 5 files changed, 87 insertions(+), 31 deletions(-) diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index 4ec431777048..61f68768ee84 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -140,6 +140,16 @@ xfs_buf_item_size( ASSERT(bip->bli_flags & XFS_BLI_LOGGED); + if (bip->bli_flags & XFS_BLI_ORDERED) { + /* + * The buffer has been logged just to order it. + * It is not being included in the transaction + * commit, so no vectors are used at all. + */ + trace_xfs_buf_item_size_ordered(bip); + return XFS_LOG_VEC_ORDERED; + } + /* * the vector count is based on the number of buffer vectors we have * dirty bits in. This will only be greater than one when we have a @@ -212,6 +222,7 @@ xfs_buf_item_format_segment( goto out; } + /* * Fill in an iovec for each set of contiguous chunks. */ @@ -311,6 +322,16 @@ xfs_buf_item_format( bip->bli_flags &= ~XFS_BLI_INODE_BUF; } + if ((bip->bli_flags & (XFS_BLI_ORDERED|XFS_BLI_STALE)) == + XFS_BLI_ORDERED) { + /* + * The buffer has been logged just to order it. It is not being + * included in the transaction commit, so don't format it. + */ + trace_xfs_buf_item_format_ordered(bip); + return; + } + for (i = 0; i < bip->bli_format_count; i++) { vecp = xfs_buf_item_format_segment(bip, vecp, offset, &bip->bli_formats[i]); @@ -340,6 +361,7 @@ xfs_buf_item_pin( ASSERT(atomic_read(&bip->bli_refcount) > 0); ASSERT((bip->bli_flags & XFS_BLI_LOGGED) || + (bip->bli_flags & XFS_BLI_ORDERED) || (bip->bli_flags & XFS_BLI_STALE)); trace_xfs_buf_item_pin(bip); @@ -512,8 +534,9 @@ xfs_buf_item_unlock( { struct xfs_buf_log_item *bip = BUF_ITEM(lip); struct xfs_buf *bp = bip->bli_buf; - int aborted, clean, i; - uint hold; + bool clean; + bool aborted; + int flags; /* Clear the buffer's association with this transaction. */ bp->b_transp = NULL; @@ -524,23 +547,21 @@ xfs_buf_item_unlock( * (cancelled) buffers at unpin time, but we'll never go through the * pin/unpin cycle if we abort inside commit. */ - aborted = (lip->li_flags & XFS_LI_ABORTED) != 0; - + aborted = (lip->li_flags & XFS_LI_ABORTED) ? true : false; /* - * Before possibly freeing the buf item, determine if we should - * release the buffer at the end of this routine. + * Before possibly freeing the buf item, copy the per-transaction state + * so we can reference it safely later after clearing it from the + * buffer log item. */ - hold = bip->bli_flags & XFS_BLI_HOLD; - - /* Clear the per transaction state. */ - bip->bli_flags &= ~(XFS_BLI_LOGGED | XFS_BLI_HOLD); + flags = bip->bli_flags; + bip->bli_flags &= ~(XFS_BLI_LOGGED | XFS_BLI_HOLD | XFS_BLI_ORDERED); /* * If the buf item is marked stale, then don't do anything. We'll * unlock the buffer and free the buf item when the buffer is unpinned * for the last time. */ - if (bip->bli_flags & XFS_BLI_STALE) { + if (flags & XFS_BLI_STALE) { trace_xfs_buf_item_unlock_stale(bip); ASSERT(bip->__bli_format.blf_flags & XFS_BLF_CANCEL); if (!aborted) { @@ -557,13 +578,19 @@ xfs_buf_item_unlock( * be the only reference to the buf item, so we free it anyway * regardless of whether it is dirty or not. A dirty abort implies a * shutdown, anyway. + * + * Ordered buffers are dirty but may have no recorded changes, so ensure + * we only release clean items here. */ - clean = 1; - for (i = 0; i < bip->bli_format_count; i++) { - if (!xfs_bitmap_empty(bip->bli_formats[i].blf_data_map, - bip->bli_formats[i].blf_map_size)) { - clean = 0; - break; + clean = (flags & XFS_BLI_DIRTY) ? false : true; + if (clean) { + int i; + for (i = 0; i < bip->bli_format_count; i++) { + if (!xfs_bitmap_empty(bip->bli_formats[i].blf_data_map, + bip->bli_formats[i].blf_map_size)) { + clean = false; + break; + } } } if (clean) @@ -576,7 +603,7 @@ xfs_buf_item_unlock( } else atomic_dec(&bip->bli_refcount); - if (!hold) + if (!(flags & XFS_BLI_HOLD)) xfs_buf_relse(bp); } @@ -841,12 +868,6 @@ xfs_buf_item_log( uint end; struct xfs_buf *bp = bip->bli_buf; - /* - * Mark the item as having some dirty data for - * quick reference in xfs_buf_item_dirty. - */ - bip->bli_flags |= XFS_BLI_DIRTY; - /* * walk each buffer segment and mark them dirty appropriately. */ @@ -873,7 +894,7 @@ xfs_buf_item_log( /* - * Return 1 if the buffer has some data that has been logged (at any + * Return 1 if the buffer has been logged or ordered in a transaction (at any * point, not just the current transaction) and 0 if not. */ uint @@ -907,11 +928,11 @@ void xfs_buf_item_relse( xfs_buf_t *bp) { - xfs_buf_log_item_t *bip; + xfs_buf_log_item_t *bip = bp->b_fspriv; trace_xfs_buf_item_relse(bp, _RET_IP_); + ASSERT(!(bip->bli_item.li_flags & XFS_LI_IN_AIL)); - bip = bp->b_fspriv; bp->b_fspriv = bip->bli_item.li_bio_list; if (bp->b_fspriv == NULL) bp->b_iodone = NULL; diff --git a/fs/xfs/xfs_buf_item.h b/fs/xfs/xfs_buf_item.h index 2573d2a75fc8..0f1c247dc680 100644 --- a/fs/xfs/xfs_buf_item.h +++ b/fs/xfs/xfs_buf_item.h @@ -120,6 +120,7 @@ xfs_blft_from_flags(struct xfs_buf_log_format *blf) #define XFS_BLI_INODE_ALLOC_BUF 0x10 #define XFS_BLI_STALE_INODE 0x20 #define XFS_BLI_INODE_BUF 0x40 +#define XFS_BLI_ORDERED 0x80 #define XFS_BLI_FLAGS \ { XFS_BLI_HOLD, "HOLD" }, \ @@ -128,7 +129,8 @@ xfs_blft_from_flags(struct xfs_buf_log_format *blf) { XFS_BLI_LOGGED, "LOGGED" }, \ { XFS_BLI_INODE_ALLOC_BUF, "INODE_ALLOC" }, \ { XFS_BLI_STALE_INODE, "STALE_INODE" }, \ - { XFS_BLI_INODE_BUF, "INODE_BUF" } + { XFS_BLI_INODE_BUF, "INODE_BUF" }, \ + { XFS_BLI_ORDERED, "ORDERED" } #ifdef __KERNEL__ diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index e31867270077..ee8b3a3b5d65 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -486,9 +486,12 @@ DEFINE_EVENT(xfs_buf_item_class, name, \ TP_PROTO(struct xfs_buf_log_item *bip), \ TP_ARGS(bip)) DEFINE_BUF_ITEM_EVENT(xfs_buf_item_size); +DEFINE_BUF_ITEM_EVENT(xfs_buf_item_size_ordered); DEFINE_BUF_ITEM_EVENT(xfs_buf_item_size_stale); DEFINE_BUF_ITEM_EVENT(xfs_buf_item_format); +DEFINE_BUF_ITEM_EVENT(xfs_buf_item_format_ordered); DEFINE_BUF_ITEM_EVENT(xfs_buf_item_format_stale); +DEFINE_BUF_ITEM_EVENT(xfs_buf_item_ordered); DEFINE_BUF_ITEM_EVENT(xfs_buf_item_pin); DEFINE_BUF_ITEM_EVENT(xfs_buf_item_unpin); DEFINE_BUF_ITEM_EVENT(xfs_buf_item_unpin_stale); @@ -508,6 +511,7 @@ DEFINE_BUF_ITEM_EVENT(xfs_trans_bjoin); DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold); DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold_release); DEFINE_BUF_ITEM_EVENT(xfs_trans_binval); +DEFINE_BUF_ITEM_EVENT(xfs_trans_buf_ordered); DECLARE_EVENT_CLASS(xfs_lock_class, TP_PROTO(struct xfs_inode *ip, unsigned lock_flags, diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index 6d526569820c..822570ec605a 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -498,6 +498,7 @@ void xfs_trans_bhold_release(xfs_trans_t *, struct xfs_buf *); void xfs_trans_binval(xfs_trans_t *, struct xfs_buf *); void xfs_trans_inode_buf(xfs_trans_t *, struct xfs_buf *); void xfs_trans_stale_inode_buf(xfs_trans_t *, struct xfs_buf *); +void xfs_trans_ordered_buf(xfs_trans_t *, struct xfs_buf *); void xfs_trans_dquot_buf(xfs_trans_t *, struct xfs_buf *, uint); void xfs_trans_inode_alloc_buf(xfs_trans_t *, struct xfs_buf *); void xfs_trans_ichgtime(struct xfs_trans *, struct xfs_inode *, int); diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c index 73a5fa457e16..aa5a04b844d6 100644 --- a/fs/xfs/xfs_trans_buf.c +++ b/fs/xfs/xfs_trans_buf.c @@ -397,7 +397,6 @@ shutdown_abort: return XFS_ERROR(EIO); } - /* * Release the buffer bp which was previously acquired with one of the * xfs_trans_... buffer allocation routines if the buffer has not @@ -603,8 +602,14 @@ xfs_trans_log_buf(xfs_trans_t *tp, tp->t_flags |= XFS_TRANS_DIRTY; bip->bli_item.li_desc->lid_flags |= XFS_LID_DIRTY; - bip->bli_flags |= XFS_BLI_LOGGED; - xfs_buf_item_log(bip, first, last); + + /* + * If we have an ordered buffer we are not logging any dirty range but + * it still needs to be marked dirty and that it has been logged. + */ + bip->bli_flags |= XFS_BLI_DIRTY | XFS_BLI_LOGGED; + if (!(bip->bli_flags & XFS_BLI_ORDERED)) + xfs_buf_item_log(bip, first, last); } @@ -756,6 +761,29 @@ xfs_trans_inode_alloc_buf( xfs_trans_buf_set_type(tp, bp, XFS_BLFT_DINO_BUF); } +/* + * Mark the buffer as ordered for this transaction. This means + * that the contents of the buffer are not recorded in the transaction + * but it is tracked in the AIL as though it was. This allows us + * to record logical changes in transactions rather than the physical + * changes we make to the buffer without changing writeback ordering + * constraints of metadata buffers. + */ +void +xfs_trans_ordered_buf( + struct xfs_trans *tp, + struct xfs_buf *bp) +{ + struct xfs_buf_log_item *bip = bp->b_fspriv; + + ASSERT(bp->b_transp == tp); + ASSERT(bip != NULL); + ASSERT(atomic_read(&bip->bli_refcount) > 0); + + bip->bli_flags |= XFS_BLI_ORDERED; + trace_xfs_buf_item_ordered(bip); +} + /* * Set the type of the buffer for log recovery so that it can correctly identify * and hence attach the correct buffer ops to the buffer after replay. From 3ebe7d2d73179c4874aee4f32e043eb5acd9fa0f Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Jun 2013 16:04:53 +1000 Subject: [PATCH 51/61] xfs: Inode create log items Introduce the inode create log item type for logical inode create logging. Instead of logging the changes in buffers, pass the range to be initialised through the log by a new transaction type. This reduces the amount of log space required to record initialisation during allocation from about 128 bytes per inode to a small fixed amount per inode extent to be initialised. This requires a new log item type to track it through the log and the AIL. This is a relatively simple item - most callbacks are noops as this item has the same life cycle as the transaction. Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/Makefile | 1 + fs/xfs/xfs_icreate_item.c | 195 ++++++++++++++++++++++++++++++++++++++ fs/xfs/xfs_icreate_item.h | 52 ++++++++++ fs/xfs/xfs_log.h | 3 +- fs/xfs/xfs_super.c | 8 ++ fs/xfs/xfs_trans.h | 4 +- 6 files changed, 261 insertions(+), 2 deletions(-) create mode 100644 fs/xfs/xfs_icreate_item.c create mode 100644 fs/xfs/xfs_icreate_item.h diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile index 6313b69b6644..4a4508023a3c 100644 --- a/fs/xfs/Makefile +++ b/fs/xfs/Makefile @@ -71,6 +71,7 @@ xfs-y += xfs_alloc.o \ xfs_dir2_sf.o \ xfs_ialloc.o \ xfs_ialloc_btree.o \ + xfs_icreate_item.o \ xfs_inode.o \ xfs_log_recover.o \ xfs_mount.o \ diff --git a/fs/xfs/xfs_icreate_item.c b/fs/xfs/xfs_icreate_item.c new file mode 100644 index 000000000000..7716a4e7375e --- /dev/null +++ b/fs/xfs/xfs_icreate_item.c @@ -0,0 +1,195 @@ +/* + * Copyright (c) 2008-2010, 2013 Dave Chinner + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_types.h" +#include "xfs_bit.h" +#include "xfs_log.h" +#include "xfs_inum.h" +#include "xfs_trans.h" +#include "xfs_buf_item.h" +#include "xfs_sb.h" +#include "xfs_ag.h" +#include "xfs_dir2.h" +#include "xfs_mount.h" +#include "xfs_trans_priv.h" +#include "xfs_bmap_btree.h" +#include "xfs_alloc_btree.h" +#include "xfs_ialloc_btree.h" +#include "xfs_attr_sf.h" +#include "xfs_dinode.h" +#include "xfs_inode.h" +#include "xfs_inode_item.h" +#include "xfs_btree.h" +#include "xfs_ialloc.h" +#include "xfs_error.h" +#include "xfs_icreate_item.h" + +kmem_zone_t *xfs_icreate_zone; /* inode create item zone */ + +static inline struct xfs_icreate_item *ICR_ITEM(struct xfs_log_item *lip) +{ + return container_of(lip, struct xfs_icreate_item, ic_item); +} + +/* + * This returns the number of iovecs needed to log the given inode item. + * + * We only need one iovec for the icreate log structure. + */ +STATIC uint +xfs_icreate_item_size( + struct xfs_log_item *lip) +{ + return 1; +} + +/* + * This is called to fill in the vector of log iovecs for the + * given inode create log item. + */ +STATIC void +xfs_icreate_item_format( + struct xfs_log_item *lip, + struct xfs_log_iovec *log_vector) +{ + struct xfs_icreate_item *icp = ICR_ITEM(lip); + + log_vector->i_addr = (xfs_caddr_t)&icp->ic_format; + log_vector->i_len = sizeof(struct xfs_icreate_log); + log_vector->i_type = XLOG_REG_TYPE_ICREATE; +} + + +/* Pinning has no meaning for the create item, so just return. */ +STATIC void +xfs_icreate_item_pin( + struct xfs_log_item *lip) +{ +} + + +/* pinning has no meaning for the create item, so just return. */ +STATIC void +xfs_icreate_item_unpin( + struct xfs_log_item *lip, + int remove) +{ +} + +STATIC void +xfs_icreate_item_unlock( + struct xfs_log_item *lip) +{ + struct xfs_icreate_item *icp = ICR_ITEM(lip); + + if (icp->ic_item.li_flags & XFS_LI_ABORTED) + kmem_zone_free(xfs_icreate_zone, icp); + return; +} + +/* + * Because we have ordered buffers being tracked in the AIL for the inode + * creation, we don't need the create item after this. Hence we can free + * the log item and return -1 to tell the caller we're done with the item. + */ +STATIC xfs_lsn_t +xfs_icreate_item_committed( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ + struct xfs_icreate_item *icp = ICR_ITEM(lip); + + kmem_zone_free(xfs_icreate_zone, icp); + return (xfs_lsn_t)-1; +} + +/* item can never get into the AIL */ +STATIC uint +xfs_icreate_item_push( + struct xfs_log_item *lip, + struct list_head *buffer_list) +{ + ASSERT(0); + return XFS_ITEM_SUCCESS; +} + +/* Ordered buffers do the dependency tracking here, so this does nothing. */ +STATIC void +xfs_icreate_item_committing( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ +} + +/* + * This is the ops vector shared by all buf log items. + */ +static struct xfs_item_ops xfs_icreate_item_ops = { + .iop_size = xfs_icreate_item_size, + .iop_format = xfs_icreate_item_format, + .iop_pin = xfs_icreate_item_pin, + .iop_unpin = xfs_icreate_item_unpin, + .iop_push = xfs_icreate_item_push, + .iop_unlock = xfs_icreate_item_unlock, + .iop_committed = xfs_icreate_item_committed, + .iop_committing = xfs_icreate_item_committing, +}; + + +/* + * Initialize the inode log item for a newly allocated (in-core) inode. + * + * Inode extents can only reside within an AG. Hence specify the starting + * block for the inode chunk by offset within an AG as well as the + * length of the allocated extent. + * + * This joins the item to the transaction and marks it dirty so + * that we don't need a separate call to do this, nor does the + * caller need to know anything about the icreate item. + */ +void +xfs_icreate_log( + struct xfs_trans *tp, + xfs_agnumber_t agno, + xfs_agblock_t agbno, + unsigned int count, + unsigned int inode_size, + xfs_agblock_t length, + unsigned int generation) +{ + struct xfs_icreate_item *icp; + + icp = kmem_zone_zalloc(xfs_icreate_zone, KM_SLEEP); + + xfs_log_item_init(tp->t_mountp, &icp->ic_item, XFS_LI_ICREATE, + &xfs_icreate_item_ops); + + icp->ic_format.icl_type = XFS_LI_ICREATE; + icp->ic_format.icl_size = 1; /* single vector */ + icp->ic_format.icl_ag = cpu_to_be32(agno); + icp->ic_format.icl_agbno = cpu_to_be32(agbno); + icp->ic_format.icl_count = cpu_to_be32(count); + icp->ic_format.icl_isize = cpu_to_be32(inode_size); + icp->ic_format.icl_length = cpu_to_be32(length); + icp->ic_format.icl_gen = cpu_to_be32(generation); + + xfs_trans_add_item(tp, &icp->ic_item); + tp->t_flags |= XFS_TRANS_DIRTY; + icp->ic_item.li_desc->lid_flags |= XFS_LID_DIRTY; +} diff --git a/fs/xfs/xfs_icreate_item.h b/fs/xfs/xfs_icreate_item.h new file mode 100644 index 000000000000..88ba8aa0bc41 --- /dev/null +++ b/fs/xfs/xfs_icreate_item.h @@ -0,0 +1,52 @@ +/* + * Copyright (c) 2008-2010, Dave Chinner + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ +#ifndef XFS_ICREATE_ITEM_H +#define XFS_ICREATE_ITEM_H 1 + +/* + * on disk log item structure + * + * Log recovery assumes the first two entries are the type and size and they fit + * in 32 bits. Also in host order (ugh) so they have to be 32 bit aligned so + * decoding can be done correctly. + */ +struct xfs_icreate_log { + __uint16_t icl_type; /* type of log format structure */ + __uint16_t icl_size; /* size of log format structure */ + __be32 icl_ag; /* ag being allocated in */ + __be32 icl_agbno; /* start block of inode range */ + __be32 icl_count; /* number of inodes to initialise */ + __be32 icl_isize; /* size of inodes */ + __be32 icl_length; /* length of extent to initialise */ + __be32 icl_gen; /* inode generation number to use */ +}; + +/* in memory log item structure */ +struct xfs_icreate_item { + struct xfs_log_item ic_item; + struct xfs_icreate_log ic_format; +}; + +extern kmem_zone_t *xfs_icreate_zone; /* inode create item zone */ + +void xfs_icreate_log(struct xfs_trans *tp, xfs_agnumber_t agno, + xfs_agblock_t agbno, unsigned int count, + unsigned int inode_size, xfs_agblock_t length, + unsigned int generation); + +#endif /* XFS_ICREATE_ITEM_H */ diff --git a/fs/xfs/xfs_log.h b/fs/xfs/xfs_log.h index b20918c554aa..fb630e496c12 100644 --- a/fs/xfs/xfs_log.h +++ b/fs/xfs/xfs_log.h @@ -88,7 +88,8 @@ static inline xfs_lsn_t _lsn_cmp(xfs_lsn_t lsn1, xfs_lsn_t lsn2) #define XLOG_REG_TYPE_UNMOUNT 17 #define XLOG_REG_TYPE_COMMIT 18 #define XLOG_REG_TYPE_TRANSHDR 19 -#define XLOG_REG_TYPE_MAX 19 +#define XLOG_REG_TYPE_ICREATE 20 +#define XLOG_REG_TYPE_MAX 20 typedef struct xfs_log_iovec { void *i_addr; /* beginning address of region */ diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 14924099bcb8..30ef68f8a390 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -51,6 +51,7 @@ #include "xfs_inode_item.h" #include "xfs_icache.h" #include "xfs_trace.h" +#include "xfs_icreate_item.h" #include #include @@ -1650,9 +1651,15 @@ xfs_init_zones(void) KM_ZONE_SPREAD, NULL); if (!xfs_ili_zone) goto out_destroy_inode_zone; + xfs_icreate_zone = kmem_zone_init(sizeof(struct xfs_icreate_item), + "xfs_icr"); + if (!xfs_icreate_zone) + goto out_destroy_ili_zone; return 0; + out_destroy_ili_zone: + kmem_zone_destroy(xfs_ili_zone); out_destroy_inode_zone: kmem_zone_destroy(xfs_inode_zone); out_destroy_efi_zone: @@ -1691,6 +1698,7 @@ xfs_destroy_zones(void) * destroy caches. */ rcu_barrier(); + kmem_zone_destroy(xfs_icreate_zone); kmem_zone_destroy(xfs_ili_zone); kmem_zone_destroy(xfs_inode_zone); kmem_zone_destroy(xfs_efi_zone); diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index 822570ec605a..2b4946393e30 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -48,6 +48,7 @@ typedef struct xfs_trans_header { #define XFS_LI_BUF 0x123c /* v2 bufs, variable sized inode bufs */ #define XFS_LI_DQUOT 0x123d #define XFS_LI_QUOTAOFF 0x123e +#define XFS_LI_ICREATE 0x123f #define XFS_LI_TYPE_DESC \ { XFS_LI_EFI, "XFS_LI_EFI" }, \ @@ -107,7 +108,8 @@ typedef struct xfs_trans_header { #define XFS_TRANS_SWAPEXT 40 #define XFS_TRANS_SB_COUNT 41 #define XFS_TRANS_CHECKPOINT 42 -#define XFS_TRANS_TYPE_MAX 42 +#define XFS_TRANS_ICREATE 43 +#define XFS_TRANS_TYPE_MAX 43 /* new transaction types need to be reflected in xfs_logprint(8) */ #define XFS_TRANS_TYPES \ From b8402b4729495ac719a3f532c2e33ac653b222a8 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Jun 2013 16:04:54 +1000 Subject: [PATCH 52/61] xfs: Inode create transaction reservations Define the log and space transaction sizes. Factor the current create log reservation macro into the two logical halves and reuse one half for the new icreate transactions. The icreate transaction is transparent to all the high level create code - the pre-calculated reservations will correctly set the reservations dependent on whether the filesystem supports the icreate transaction. Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_trans.c | 126 +++++++++++++++++++++++++++++---------------- 1 file changed, 81 insertions(+), 45 deletions(-) diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c index 2fd7c1ff1d21..35a229981354 100644 --- a/fs/xfs/xfs_trans.c +++ b/fs/xfs/xfs_trans.c @@ -234,37 +234,10 @@ xfs_calc_remove_reservation( } /* - * For symlink we can modify: - * the parent directory inode: inode size - * the new inode: inode size - * the inode btree entry: 1 block - * the directory btree: (max depth + v2) * dir block size - * the directory inode's bmap btree: (max depth + v2) * block size - * the blocks for the symlink: 1 kB - * Or in the first xact we allocate some inodes giving: - * the agi and agf of the ag getting the new inodes: 2 * sectorsize - * the inode blocks allocated: XFS_IALLOC_BLOCKS * blocksize - * the inode btree: max depth * blocksize - * the allocation btrees: 2 trees * (2 * max depth - 1) * block size + * For create, break it in to the two cases that the transaction + * covers. We start with the modify case - allocation done by modification + * of the state of existing inodes - and the allocation case. */ -STATIC uint -xfs_calc_symlink_reservation( - struct xfs_mount *mp) -{ - return XFS_DQUOT_LOGRES(mp) + - MAX((xfs_calc_buf_res(2, mp->m_sb.sb_inodesize) + - xfs_calc_buf_res(1, XFS_FSB_TO_B(mp, 1)) + - xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp), - XFS_FSB_TO_B(mp, 1)) + - xfs_calc_buf_res(1, 1024)), - (xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) + - xfs_calc_buf_res(XFS_IALLOC_BLOCKS(mp), - XFS_FSB_TO_B(mp, 1)) + - xfs_calc_buf_res(mp->m_in_maxlevels, - XFS_FSB_TO_B(mp, 1)) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), - XFS_FSB_TO_B(mp, 1)))); -} /* * For create we can modify: @@ -274,7 +247,19 @@ xfs_calc_symlink_reservation( * the superblock for the nlink flag: sector size * the directory btree: (max depth + v2) * dir block size * the directory inode's bmap btree: (max depth + v2) * block size - * Or in the first xact we allocate some inodes giving: + */ +STATIC uint +xfs_calc_create_resv_modify( + struct xfs_mount *mp) +{ + return xfs_calc_buf_res(2, mp->m_sb.sb_inodesize) + + xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) + + (uint)XFS_FSB_TO_B(mp, 1) + + xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp), XFS_FSB_TO_B(mp, 1)); +} + +/* + * For create we can allocate some inodes giving: * the agi and agf of the ag getting the new inodes: 2 * sectorsize * the superblock for the nlink flag: sector size * the inode blocks allocated: XFS_IALLOC_BLOCKS * blocksize @@ -282,23 +267,60 @@ xfs_calc_symlink_reservation( * the allocation btrees: 2 trees * (max depth - 1) * block size */ STATIC uint -xfs_calc_create_reservation( +xfs_calc_create_resv_alloc( + struct xfs_mount *mp) +{ + return xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) + + mp->m_sb.sb_sectsize + + xfs_calc_buf_res(XFS_IALLOC_BLOCKS(mp), XFS_FSB_TO_B(mp, 1)) + + xfs_calc_buf_res(mp->m_in_maxlevels, XFS_FSB_TO_B(mp, 1)) + + xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), + XFS_FSB_TO_B(mp, 1)); +} + +STATIC uint +__xfs_calc_create_reservation( struct xfs_mount *mp) { return XFS_DQUOT_LOGRES(mp) + - MAX((xfs_calc_buf_res(2, mp->m_sb.sb_inodesize) + - xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) + - (uint)XFS_FSB_TO_B(mp, 1) + - xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp), - XFS_FSB_TO_B(mp, 1))), - (xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) + - mp->m_sb.sb_sectsize + - xfs_calc_buf_res(XFS_IALLOC_BLOCKS(mp), - XFS_FSB_TO_B(mp, 1)) + - xfs_calc_buf_res(mp->m_in_maxlevels, - XFS_FSB_TO_B(mp, 1)) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), - XFS_FSB_TO_B(mp, 1)))); + MAX(xfs_calc_create_resv_alloc(mp), + xfs_calc_create_resv_modify(mp)); +} + +/* + * For icreate we can allocate some inodes giving: + * the agi and agf of the ag getting the new inodes: 2 * sectorsize + * the superblock for the nlink flag: sector size + * the inode btree: max depth * blocksize + * the allocation btrees: 2 trees * (max depth - 1) * block size + */ +STATIC uint +xfs_calc_icreate_resv_alloc( + struct xfs_mount *mp) +{ + return xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) + + mp->m_sb.sb_sectsize + + xfs_calc_buf_res(mp->m_in_maxlevels, XFS_FSB_TO_B(mp, 1)) + + xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), + XFS_FSB_TO_B(mp, 1)); +} + +STATIC uint +xfs_calc_icreate_reservation(xfs_mount_t *mp) +{ + return XFS_DQUOT_LOGRES(mp) + + MAX(xfs_calc_icreate_resv_alloc(mp), + xfs_calc_create_resv_modify(mp)); +} + +STATIC uint +xfs_calc_create_reservation( + struct xfs_mount *mp) +{ + if (xfs_sb_version_hascrc(&mp->m_sb)) + return xfs_calc_icreate_reservation(mp); + return __xfs_calc_create_reservation(mp); + } /* @@ -311,6 +333,20 @@ xfs_calc_mkdir_reservation( return xfs_calc_create_reservation(mp); } + +/* + * Making a new symplink is the same as creating a new file, but + * with the added blocks for remote symlink data which can be up to 1kB in + * length (MAXPATHLEN). + */ +STATIC uint +xfs_calc_symlink_reservation( + struct xfs_mount *mp) +{ + return xfs_calc_create_reservation(mp) + + xfs_calc_buf_res(1, MAXPATHLEN); +} + /* * In freeing an inode we can modify: * the inode being freed: inode size From 28c8e41af693e4b5cd2d68218f144cf40ce15781 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Jun 2013 16:04:55 +1000 Subject: [PATCH 53/61] xfs: Inode create item recovery When we find a icreate transaction, we need to get and initialise the buffers in the range that has been passed. Extract and verify the information in the item record, then loop over the range initialising and issuing the buffer writes delayed. Support an arbitrary size range to initialise so that in future when we allocate inodes in much larger chunks all kernels that understand this transaction can still recover them. Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_ialloc.c | 37 +++++++++---- fs/xfs/xfs_ialloc.h | 8 +++ fs/xfs/xfs_log_recover.c | 114 +++++++++++++++++++++++++++++++++++++-- 3 files changed, 145 insertions(+), 14 deletions(-) diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index 4345c53c729c..4411565b718b 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -150,12 +150,16 @@ xfs_check_agi_freecount( #endif /* - * Initialise a new set of inodes. + * Initialise a new set of inodes. When called without a transaction context + * (e.g. from recovery) we initiate a delayed write of the inode buffers rather + * than logging them (which in a transaction context puts them into the AIL + * for writeback rather than the xfsbufd queue). */ STATIC int xfs_ialloc_inode_init( struct xfs_mount *mp, struct xfs_trans *tp, + struct list_head *buffer_list, xfs_agnumber_t agno, xfs_agblock_t agbno, xfs_agblock_t length, @@ -247,18 +251,33 @@ xfs_ialloc_inode_init( ino++; uuid_copy(&free->di_uuid, &mp->m_sb.sb_uuid); xfs_dinode_calc_crc(mp, free); - } else { + } else if (tp) { /* just log the inode core */ xfs_trans_log_buf(tp, fbuf, ioffset, ioffset + isize - 1); } } - if (version == 3) { - /* need to log the entire buffer */ - xfs_trans_log_buf(tp, fbuf, 0, - BBTOB(fbuf->b_length) - 1); + + if (tp) { + /* + * Mark the buffer as an inode allocation buffer so it + * sticks in AIL at the point of this allocation + * transaction. This ensures the they are on disk before + * the tail of the log can be moved past this + * transaction (i.e. by preventing relogging from moving + * it forward in the log). + */ + xfs_trans_inode_alloc_buf(tp, fbuf); + if (version == 3) { + /* need to log the entire buffer */ + xfs_trans_log_buf(tp, fbuf, 0, + BBTOB(fbuf->b_length) - 1); + } + } else { + fbuf->b_flags |= XBF_DONE; + xfs_buf_delwri_queue(fbuf, buffer_list); + xfs_buf_relse(fbuf); } - xfs_trans_inode_alloc_buf(tp, fbuf); } return 0; } @@ -303,7 +322,7 @@ xfs_ialloc_ag_alloc( * First try to allocate inodes contiguous with the last-allocated * chunk of inodes. If the filesystem is striped, this will fill * an entire stripe unit with inodes. - */ + */ agi = XFS_BUF_TO_AGI(agbp); newino = be32_to_cpu(agi->agi_newino); agno = be32_to_cpu(agi->agi_seqno); @@ -402,7 +421,7 @@ xfs_ialloc_ag_alloc( * rather than a linear progression to prevent the next generation * number from being easily guessable. */ - error = xfs_ialloc_inode_init(args.mp, tp, agno, args.agbno, + error = xfs_ialloc_inode_init(args.mp, tp, NULL, agno, args.agbno, args.len, prandom_u32()); if (error) diff --git a/fs/xfs/xfs_ialloc.h b/fs/xfs/xfs_ialloc.h index c8da3df271e6..68c07320f096 100644 --- a/fs/xfs/xfs_ialloc.h +++ b/fs/xfs/xfs_ialloc.h @@ -150,6 +150,14 @@ int xfs_inobt_lookup(struct xfs_btree_cur *cur, xfs_agino_t ino, int xfs_inobt_get_rec(struct xfs_btree_cur *cur, xfs_inobt_rec_incore_t *rec, int *stat); +/* + * Inode chunk initialisation routine + */ +int xfs_ialloc_inode_init(struct xfs_mount *mp, struct xfs_trans *tp, + struct list_head *buffer_list, + xfs_agnumber_t agno, xfs_agblock_t agbno, + xfs_agblock_t length, unsigned int gen); + extern const struct xfs_buf_ops xfs_agi_buf_ops; #endif /* __XFS_IALLOC_H__ */ diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 7cf5e4eafe28..6fcc910a50b9 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -45,6 +45,7 @@ #include "xfs_cksum.h" #include "xfs_trace.h" #include "xfs_icache.h" +#include "xfs_icreate_item.h" /* Need all the magic numbers and buffer ops structures from these headers */ #include "xfs_symlink.h" @@ -1617,7 +1618,10 @@ xlog_recover_add_to_trans( * form the cancelled buffer table. Hence they have tobe done last. * * 3. Inode allocation buffers must be replayed before inode items that - * read the buffer and replay changes into it. + * read the buffer and replay changes into it. For filesystems using the + * ICREATE transactions, this means XFS_LI_ICREATE objects need to get + * treated the same as inode allocation buffers as they create and + * initialise the buffers directly. * * 4. Inode unlink buffers must be replayed after inode items are replayed. * This ensures that inodes are completely flushed to the inode buffer @@ -1632,10 +1636,17 @@ xlog_recover_add_to_trans( * from all the other buffers and move them to last. * * Hence, 4 lists, in order from head to tail: - * - buffer_list for all buffers except cancelled/inode unlink buffers - * - item_list for all non-buffer items - * - inode_buffer_list for inode unlink buffers - * - cancel_list for the cancelled buffers + * - buffer_list for all buffers except cancelled/inode unlink buffers + * - item_list for all non-buffer items + * - inode_buffer_list for inode unlink buffers + * - cancel_list for the cancelled buffers + * + * Note that we add objects to the tail of the lists so that first-to-last + * ordering is preserved within the lists. Adding objects to the head of the + * list means when we traverse from the head we walk them in last-to-first + * order. For cancelled buffers and inode unlink buffers this doesn't matter, + * but for all other items there may be specific ordering that we need to + * preserve. */ STATIC int xlog_recover_reorder_trans( @@ -1655,6 +1666,9 @@ xlog_recover_reorder_trans( xfs_buf_log_format_t *buf_f = item->ri_buf[0].i_addr; switch (ITEM_TYPE(item)) { + case XFS_LI_ICREATE: + list_move_tail(&item->ri_list, &buffer_list); + break; case XFS_LI_BUF: if (buf_f->blf_flags & XFS_BLF_CANCEL) { trace_xfs_log_recover_item_reorder_head(log, @@ -2981,6 +2995,93 @@ xlog_recover_efd_pass2( return 0; } +/* + * This routine is called when an inode create format structure is found in a + * committed transaction in the log. It's purpose is to initialise the inodes + * being allocated on disk. This requires us to get inode cluster buffers that + * match the range to be intialised, stamped with inode templates and written + * by delayed write so that subsequent modifications will hit the cached buffer + * and only need writing out at the end of recovery. + */ +STATIC int +xlog_recover_do_icreate_pass2( + struct xlog *log, + struct list_head *buffer_list, + xlog_recover_item_t *item) +{ + struct xfs_mount *mp = log->l_mp; + struct xfs_icreate_log *icl; + xfs_agnumber_t agno; + xfs_agblock_t agbno; + unsigned int count; + unsigned int isize; + xfs_agblock_t length; + + icl = (struct xfs_icreate_log *)item->ri_buf[0].i_addr; + if (icl->icl_type != XFS_LI_ICREATE) { + xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad type"); + return EINVAL; + } + + if (icl->icl_size != 1) { + xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad icl size"); + return EINVAL; + } + + agno = be32_to_cpu(icl->icl_ag); + if (agno >= mp->m_sb.sb_agcount) { + xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad agno"); + return EINVAL; + } + agbno = be32_to_cpu(icl->icl_agbno); + if (!agbno || agbno == NULLAGBLOCK || agbno >= mp->m_sb.sb_agblocks) { + xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad agbno"); + return EINVAL; + } + isize = be32_to_cpu(icl->icl_isize); + if (isize != mp->m_sb.sb_inodesize) { + xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad isize"); + return EINVAL; + } + count = be32_to_cpu(icl->icl_count); + if (!count) { + xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad count"); + return EINVAL; + } + length = be32_to_cpu(icl->icl_length); + if (!length || length >= mp->m_sb.sb_agblocks) { + xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad length"); + return EINVAL; + } + + /* existing allocation is fixed value */ + ASSERT(count == XFS_IALLOC_INODES(mp)); + ASSERT(length == XFS_IALLOC_BLOCKS(mp)); + if (count != XFS_IALLOC_INODES(mp) || + length != XFS_IALLOC_BLOCKS(mp)) { + xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad count 2"); + return EINVAL; + } + + /* + * Inode buffers can be freed. Do not replay the inode initialisation as + * we could be overwriting something written after this inode buffer was + * cancelled. + * + * XXX: we need to iterate all buffers and only init those that are not + * cancelled. I think that a more fine grained factoring of + * xfs_ialloc_inode_init may be appropriate here to enable this to be + * done easily. + */ + if (xlog_check_buffer_cancelled(log, + XFS_AGB_TO_DADDR(mp, agno, agbno), length, 0)) + return 0; + + xfs_ialloc_inode_init(mp, NULL, buffer_list, agno, agbno, length, + be32_to_cpu(icl->icl_gen)); + return 0; +} + /* * Free up any resources allocated by the transaction * @@ -3023,6 +3124,7 @@ xlog_recover_commit_pass1( case XFS_LI_EFI: case XFS_LI_EFD: case XFS_LI_DQUOT: + case XFS_LI_ICREATE: /* nothing to do in pass 1 */ return 0; default: @@ -3053,6 +3155,8 @@ xlog_recover_commit_pass2( return xlog_recover_efd_pass2(log, item); case XFS_LI_DQUOT: return xlog_recover_dquot_pass2(log, buffer_list, item); + case XFS_LI_ICREATE: + return xlog_recover_do_icreate_pass2(log, buffer_list, item); case XFS_LI_QUOTAOFF: /* nothing to do in pass2 */ return 0; From ddf6ad01434e72bfc8423e1619abdaa0af9394a8 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Jun 2013 16:04:56 +1000 Subject: [PATCH 54/61] xfs: Use inode create transaction Replace the use of buffer based logging of inode initialisation, uses the new logical form to describe the range to be initialised in recovery. We continue to "log" the inode buffers to push them into the AIL and ensure that the inode create transaction is not removed from the log before the inode buffers are written to disk. Update the transaction identifier and reservations to match the changed implementation. Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_buf_item.c | 12 ++++++++++-- fs/xfs/xfs_ialloc.c | 32 +++++++++++++++++++++++--------- 2 files changed, 33 insertions(+), 11 deletions(-) diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index 61f68768ee84..bfc4e0c26fd3 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -310,13 +310,21 @@ xfs_buf_item_format( /* * If it is an inode buffer, transfer the in-memory state to the - * format flags and clear the in-memory state. We do not transfer + * format flags and clear the in-memory state. + * + * For buffer based inode allocation, we do not transfer * this state if the inode buffer allocation has not yet been committed * to the log as setting the XFS_BLI_INODE_BUF flag will prevent * correct replay of the inode allocation. + * + * For icreate item based inode allocation, the buffers aren't written + * to the journal during allocation, and hence we should always tag the + * buffer as an inode buffer so that the correct unlinked list replay + * occurs during recovery. */ if (bip->bli_flags & XFS_BLI_INODE_BUF) { - if (!((bip->bli_flags & XFS_BLI_INODE_ALLOC_BUF) && + if (xfs_sb_version_hascrc(&lip->li_mountp->m_sb) || + !((bip->bli_flags & XFS_BLI_INODE_ALLOC_BUF) && xfs_log_item_in_current_chkpt(lip))) bip->__bli_format.blf_flags |= XFS_BLF_INODE_BUF; bip->bli_flags &= ~XFS_BLI_INODE_BUF; diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index 4411565b718b..7a0c17d7ec09 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -38,6 +38,7 @@ #include "xfs_bmap.h" #include "xfs_cksum.h" #include "xfs_buf_item.h" +#include "xfs_icreate_item.h" /* @@ -155,7 +156,7 @@ xfs_check_agi_freecount( * than logging them (which in a transaction context puts them into the AIL * for writeback rather than the xfsbufd queue). */ -STATIC int +int xfs_ialloc_inode_init( struct xfs_mount *mp, struct xfs_trans *tp, @@ -212,6 +213,18 @@ xfs_ialloc_inode_init( version = 3; ino = XFS_AGINO_TO_INO(mp, agno, XFS_OFFBNO_TO_AGINO(mp, agbno, 0)); + + /* + * log the initialisation that is about to take place as an + * logical operation. This means the transaction does not + * need to log the physical changes to the inode buffers as log + * recovery will know what initialisation is actually needed. + * Hence we only need to log the buffers as "ordered" buffers so + * they track in the AIL as if they were physically logged. + */ + if (tp) + xfs_icreate_log(tp, agno, agbno, XFS_IALLOC_INODES(mp), + mp->m_sb.sb_inodesize, length, gen); } else if (xfs_sb_version_hasnlink(&mp->m_sb)) version = 2; else @@ -227,13 +240,8 @@ xfs_ialloc_inode_init( XBF_UNMAPPED); if (!fbuf) return ENOMEM; - /* - * Initialize all inodes in this buffer and then log them. - * - * XXX: It would be much better if we had just one transaction - * to log a whole cluster of inodes instead of all the - * individual transactions causing a lot of log traffic. - */ + + /* Initialize the inode buffers and log them appropriately. */ fbuf->b_ops = &xfs_inode_buf_ops; xfs_buf_zero(fbuf, 0, BBTOB(fbuf->b_length)); for (i = 0; i < ninodes; i++) { @@ -269,7 +277,13 @@ xfs_ialloc_inode_init( */ xfs_trans_inode_alloc_buf(tp, fbuf); if (version == 3) { - /* need to log the entire buffer */ + /* + * Mark the buffer as ordered so that they are + * not physically logged in the transaction but + * still tracked in the AIL as part of the + * transaction and pin the log appropriately. + */ + xfs_trans_ordered_buf(tp, fbuf); xfs_trans_log_buf(tp, fbuf, 0, BBTOB(fbuf->b_length) - 1); } From dc037ad7d24f3711e431a45c053b5d425995e9e4 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Jun 2013 16:04:59 +1000 Subject: [PATCH 55/61] xfs: implement inode change count For CRC enabled filesystems, add support for the monotonic inode version change counter that is needed by protocols like NFSv4 for determining if the inode has changed in any way at all between two unrelated operations on the inode. This bumps the change count the first time an inode is dirtied in a transaction. Since all modifications to the inode are logged, this will catch all changes that are made to the inode, including timestamp updates that occur during data writes. Signed-off-by: Dave Chinner Reviewed-by: Mark Tinguely Reviewed-by: Chandra Seetharaman Signed-off-by: Ben Myers --- fs/xfs/xfs_super.c | 4 ++++ fs/xfs/xfs_trans_inode.c | 11 +++++++++++ 2 files changed, 15 insertions(+) diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 30ef68f8a390..e5e8b5ee8bfa 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -1477,6 +1477,10 @@ xfs_fs_fill_super( sb->s_time_gran = 1; set_posix_acl_flag(sb); + /* version 5 superblocks support inode version counters. */ + if (XFS_SB_VERSION_NUM(&mp->m_sb) == XFS_SB_VERSION_5) + sb->s_flags |= MS_I_VERSION; + error = xfs_mountfs(mp); if (error) goto out_filestream_unmount; diff --git a/fs/xfs/xfs_trans_inode.c b/fs/xfs/xfs_trans_inode.c index ac6d567704db..53dfe46f3680 100644 --- a/fs/xfs/xfs_trans_inode.c +++ b/fs/xfs/xfs_trans_inode.c @@ -112,6 +112,17 @@ xfs_trans_log_inode( ASSERT(ip->i_itemp != NULL); ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + /* + * First time we log the inode in a transaction, bump the inode change + * counter if it is configured for this to occur. + */ + if (!(ip->i_itemp->ili_item.li_desc->lid_flags & XFS_LID_DIRTY) && + IS_I_VERSION(VFS_I(ip))) { + inode_inc_iversion(VFS_I(ip)); + ip->i_d.di_changecount = VFS_I(ip)->i_version; + flags |= XFS_ILOG_CORE; + } + tp->t_flags |= XFS_TRANS_DIRTY; ip->i_itemp->ili_item.li_desc->lid_flags |= XFS_LID_DIRTY; From 9cad19d2cb57a2c32887a303b516d74254aa4b1c Mon Sep 17 00:00:00 2001 From: Chandra Seetharaman Date: Thu, 27 Jun 2013 17:25:04 -0500 Subject: [PATCH 56/61] xfs: Define a new function xfs_is_quota_inode() In preparation for combined pquota/gquota support, define a new function to check if the given inode is a quota inode. Signed-off-by: Chandra Seetharaman Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_itable.c | 2 +- fs/xfs/xfs_qm.c | 8 +++----- fs/xfs/xfs_sb.h | 6 ++++++ fs/xfs/xfs_trans_dquot.c | 6 ++---- 4 files changed, 12 insertions(+), 10 deletions(-) diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c index 06d004d85bf4..bc92c5306a17 100644 --- a/fs/xfs/xfs_itable.c +++ b/fs/xfs/xfs_itable.c @@ -43,7 +43,7 @@ xfs_internal_inum( { return (ino == mp->m_sb.sb_rbmino || ino == mp->m_sb.sb_rsumino || (xfs_sb_version_hasquota(&mp->m_sb) && - (ino == mp->m_sb.sb_uquotino || ino == mp->m_sb.sb_gquotino))); + xfs_is_quota_inode(&mp->m_sb, ino))); } /* diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index b75c9bb6e71e..28a5e8ae3829 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -489,8 +489,7 @@ xfs_qm_need_dqattach( return false; if (!XFS_NOT_DQATTACHED(mp, ip)) return false; - if (ip->i_ino == mp->m_sb.sb_uquotino || - ip->i_ino == mp->m_sb.sb_gquotino) + if (xfs_is_quota_inode(&mp->m_sb, ip->i_ino)) return false; return true; } @@ -606,8 +605,7 @@ xfs_qm_dqdetach( trace_xfs_dquot_dqdetach(ip); - ASSERT(ip->i_ino != ip->i_mount->m_sb.sb_uquotino); - ASSERT(ip->i_ino != ip->i_mount->m_sb.sb_gquotino); + ASSERT(!xfs_is_quota_inode(&ip->i_mount->m_sb, ip->i_ino)); if (ip->i_udquot) { xfs_qm_dqrele(ip->i_udquot); ip->i_udquot = NULL; @@ -1152,7 +1150,7 @@ xfs_qm_dqusage_adjust( * rootino must have its resources accounted for, not so with the quota * inodes. */ - if (ino == mp->m_sb.sb_uquotino || ino == mp->m_sb.sb_gquotino) { + if (xfs_is_quota_inode(&mp->m_sb, ino)) { *res = BULKSTAT_RV_NOTHING; return XFS_ERROR(EINVAL); } diff --git a/fs/xfs/xfs_sb.h b/fs/xfs/xfs_sb.h index 2de58a85833c..78f9e70b80c7 100644 --- a/fs/xfs/xfs_sb.h +++ b/fs/xfs/xfs_sb.h @@ -618,6 +618,12 @@ xfs_sb_has_incompat_log_feature( return (sbp->sb_features_log_incompat & feature) != 0; } +static inline bool +xfs_is_quota_inode(struct xfs_sb *sbp, xfs_ino_t ino) +{ + return (ino == sbp->sb_uquotino || ino == sbp->sb_gquotino); +} + /* * end of superblock version macros */ diff --git a/fs/xfs/xfs_trans_dquot.c b/fs/xfs/xfs_trans_dquot.c index fec75d023703..3fa369c75a4d 100644 --- a/fs/xfs/xfs_trans_dquot.c +++ b/fs/xfs/xfs_trans_dquot.c @@ -157,8 +157,7 @@ xfs_trans_mod_dquot_byino( if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp) || - ip->i_ino == mp->m_sb.sb_uquotino || - ip->i_ino == mp->m_sb.sb_gquotino) + xfs_is_quota_inode(&mp->m_sb, ip->i_ino)) return; if (tp->t_dqinfo == NULL) @@ -816,8 +815,7 @@ xfs_trans_reserve_quota_nblks( if (XFS_IS_PQUOTA_ON(mp)) flags |= XFS_QMOPT_ENOSPC; - ASSERT(ip->i_ino != mp->m_sb.sb_uquotino); - ASSERT(ip->i_ino != mp->m_sb.sb_gquotino); + ASSERT(!xfs_is_quota_inode(&mp->m_sb, ip->i_ino)); ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); ASSERT((flags & ~(XFS_QMOPT_FORCE_RES | XFS_QMOPT_ENOSPC)) == From 329e0875286984df9053d410df83f839f85bea6e Mon Sep 17 00:00:00 2001 From: Chandra Seetharaman Date: Thu, 27 Jun 2013 17:25:05 -0500 Subject: [PATCH 57/61] xfs: Replace macro XFS_DQUOT_TREE with a function In preparation for combined pquota/gquota support, for the sake of readability, change the macro to an inline function. Signed-off-by: Chandra Seetharaman Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_dquot.c | 2 +- fs/xfs/xfs_qm.c | 6 +++--- fs/xfs/xfs_qm.h | 22 ++++++++++++++++------ 3 files changed, 20 insertions(+), 10 deletions(-) diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index 044e97a33c8d..09af322653a2 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -804,7 +804,7 @@ xfs_qm_dqget( xfs_dquot_t **O_dqpp) /* OUT : locked incore dquot */ { struct xfs_quotainfo *qi = mp->m_quotainfo; - struct radix_tree_root *tree = XFS_DQUOT_TREE(qi, type); + struct radix_tree_root *tree = xfs_dquot_tree(qi, type); struct xfs_dquot *dqp; int error; diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index 28a5e8ae3829..8e707d33e106 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -70,7 +70,7 @@ xfs_qm_dquot_walk( void *data) { struct xfs_quotainfo *qi = mp->m_quotainfo; - struct radix_tree_root *tree = XFS_DQUOT_TREE(qi, type); + struct radix_tree_root *tree = xfs_dquot_tree(qi, type); uint32_t next_index; int last_error = 0; int skipped; @@ -189,7 +189,7 @@ xfs_qm_dqpurge( xfs_dqfunlock(dqp); xfs_dqunlock(dqp); - radix_tree_delete(XFS_DQUOT_TREE(qi, dqp->q_core.d_flags), + radix_tree_delete(xfs_dquot_tree(qi, dqp->q_core.d_flags), be32_to_cpu(dqp->q_core.d_id)); qi->qi_dquots--; @@ -1471,7 +1471,7 @@ xfs_qm_dqfree_one( struct xfs_quotainfo *qi = mp->m_quotainfo; mutex_lock(&qi->qi_tree_lock); - radix_tree_delete(XFS_DQUOT_TREE(qi, dqp->q_core.d_flags), + radix_tree_delete(xfs_dquot_tree(qi, dqp->q_core.d_flags), be32_to_cpu(dqp->q_core.d_id)); qi->qi_dquots--; diff --git a/fs/xfs/xfs_qm.h b/fs/xfs/xfs_qm.h index 5d16a6e6900f..96568c270670 100644 --- a/fs/xfs/xfs_qm.h +++ b/fs/xfs/xfs_qm.h @@ -69,12 +69,22 @@ typedef struct xfs_quotainfo { struct shrinker qi_shrinker; } xfs_quotainfo_t; -#define XFS_DQUOT_TREE(qi, type) \ - ((type & XFS_DQ_USER) ? \ - &((qi)->qi_uquota_tree) : \ - &((qi)->qi_gquota_tree)) - - +static inline struct radix_tree_root * +xfs_dquot_tree( + struct xfs_quotainfo *qi, + int type) +{ + switch (type) { + case XFS_DQ_USER: + return &qi->qi_uquota_tree; + case XFS_DQ_GROUP: + case XFS_DQ_PROJ: + return &qi->qi_gquota_tree; + default: + ASSERT(0); + } + return NULL; +} extern int xfs_qm_calc_dquots_per_chunk(struct xfs_mount *mp, unsigned int nbblks); extern void xfs_trans_mod_dquot(xfs_trans_t *, xfs_dquot_t *, uint, long); From 995961c4510460d9eef9b5ae46789aa2315545fe Mon Sep 17 00:00:00 2001 From: Chandra Seetharaman Date: Thu, 27 Jun 2013 17:25:06 -0500 Subject: [PATCH 58/61] xfs: Replace macro XFS_DQ_TO_QIP with a function In preparation for combined pquota/gquota support, for the sake of readability, change the macro to an inline function. Signed-off-by: Chandra Seetharaman Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_dquot.c | 2 +- fs/xfs/xfs_dquot.h | 4 ---- fs/xfs/xfs_qm.h | 16 ++++++++++++++++ 3 files changed, 17 insertions(+), 5 deletions(-) diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index 09af322653a2..7d184dead99a 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -573,7 +573,7 @@ xfs_qm_dqtobp( xfs_bmbt_irec_t map; int nmaps = 1, error; xfs_buf_t *bp; - xfs_inode_t *quotip = XFS_DQ_TO_QIP(dqp); + struct xfs_inode *quotip = xfs_dq_to_quota_inode(dqp); xfs_mount_t *mp = dqp->q_mount; xfs_dqid_t id = be32_to_cpu(dqp->q_core.d_id); xfs_trans_t *tp = (tpp ? *tpp : NULL); diff --git a/fs/xfs/xfs_dquot.h b/fs/xfs/xfs_dquot.h index 4f0ebfc43cc9..b596626249b8 100644 --- a/fs/xfs/xfs_dquot.h +++ b/fs/xfs/xfs_dquot.h @@ -143,10 +143,6 @@ static inline xfs_dquot_t *xfs_inode_dquot(struct xfs_inode *ip, int type) #define XFS_QM_ISUDQ(dqp) ((dqp)->dq_flags & XFS_DQ_USER) #define XFS_QM_ISPDQ(dqp) ((dqp)->dq_flags & XFS_DQ_PROJ) #define XFS_QM_ISGDQ(dqp) ((dqp)->dq_flags & XFS_DQ_GROUP) -#define XFS_DQ_TO_QINF(dqp) ((dqp)->q_mount->m_quotainfo) -#define XFS_DQ_TO_QIP(dqp) (XFS_QM_ISUDQ(dqp) ? \ - XFS_DQ_TO_QINF(dqp)->qi_uquotaip : \ - XFS_DQ_TO_QINF(dqp)->qi_gquotaip) extern int xfs_qm_dqread(struct xfs_mount *, xfs_dqid_t, uint, uint, struct xfs_dquot **); diff --git a/fs/xfs/xfs_qm.h b/fs/xfs/xfs_qm.h index 96568c270670..051e43a5e20b 100644 --- a/fs/xfs/xfs_qm.h +++ b/fs/xfs/xfs_qm.h @@ -85,6 +85,22 @@ xfs_dquot_tree( } return NULL; } + +static inline struct xfs_inode * +xfs_dq_to_quota_inode(struct xfs_dquot *dqp) +{ + switch (dqp->dq_flags & XFS_DQ_ALLTYPES) { + case XFS_DQ_USER: + return dqp->q_mount->m_quotainfo->qi_uquotaip; + case XFS_DQ_GROUP: + case XFS_DQ_PROJ: + return dqp->q_mount->m_quotainfo->qi_gquotaip; + default: + ASSERT(0); + } + return NULL; +} + extern int xfs_qm_calc_dquots_per_chunk(struct xfs_mount *mp, unsigned int nbblks); extern void xfs_trans_mod_dquot(xfs_trans_t *, xfs_dquot_t *, uint, long); From 113a56835d938d5cf9b4599053da7afb80d6f710 Mon Sep 17 00:00:00 2001 From: Chandra Seetharaman Date: Thu, 27 Jun 2013 17:25:07 -0500 Subject: [PATCH 59/61] xfs: Code cleanup and removal of some typedef usage In preparation for combined pquota/gquota support, for the sake of readability, do some code cleanup surrounding the affected code. Signed-off-by: Chandra Seetharaman Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_dquot.c | 12 ++-- fs/xfs/xfs_icache.c | 3 +- fs/xfs/xfs_qm.c | 152 ++++++++++++++++++++------------------- fs/xfs/xfs_qm.h | 37 +++++----- fs/xfs/xfs_qm_syscalls.c | 8 +-- fs/xfs/xfs_symlink.c | 3 +- fs/xfs/xfs_trans_dquot.c | 46 ++++++------ 7 files changed, 135 insertions(+), 126 deletions(-) diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index 7d184dead99a..f01012de06d0 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -570,13 +570,13 @@ xfs_qm_dqtobp( xfs_buf_t **O_bpp, uint flags) { - xfs_bmbt_irec_t map; - int nmaps = 1, error; - xfs_buf_t *bp; + struct xfs_bmbt_irec map; + int nmaps = 1, error; + struct xfs_buf *bp; struct xfs_inode *quotip = xfs_dq_to_quota_inode(dqp); - xfs_mount_t *mp = dqp->q_mount; - xfs_dqid_t id = be32_to_cpu(dqp->q_core.d_id); - xfs_trans_t *tp = (tpp ? *tpp : NULL); + struct xfs_mount *mp = dqp->q_mount; + xfs_dqid_t id = be32_to_cpu(dqp->q_core.d_id); + struct xfs_trans *tp = (tpp ? *tpp : NULL); dqp->q_fileoffset = (xfs_fileoff_t)id / mp->m_quotainfo->qi_dqperchunk; diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c index 96e344e3e927..9560dc1f15a9 100644 --- a/fs/xfs/xfs_icache.c +++ b/fs/xfs/xfs_icache.c @@ -335,7 +335,8 @@ xfs_iget_cache_miss( iflags = XFS_INEW; if (flags & XFS_IGET_DONTCACHE) iflags |= XFS_IDONTCACHE; - ip->i_udquot = ip->i_gdquot = NULL; + ip->i_udquot = NULL; + ip->i_gdquot = NULL; xfs_iflags_set(ip, iflags); /* insert the new inode */ diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index 8e707d33e106..cf09aa845d8c 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -1260,19 +1260,20 @@ int xfs_qm_quotacheck( xfs_mount_t *mp) { - int done, count, error, error2; - xfs_ino_t lastino; - size_t structsz; - xfs_inode_t *uip, *gip; - uint flags; - LIST_HEAD (buffer_list); + int done, count, error, error2; + xfs_ino_t lastino; + size_t structsz; + uint flags; + LIST_HEAD (buffer_list); + struct xfs_inode *uip = mp->m_quotainfo->qi_uquotaip; + struct xfs_inode *gip = mp->m_quotainfo->qi_gquotaip; count = INT_MAX; structsz = 1; lastino = 0; flags = 0; - ASSERT(mp->m_quotainfo->qi_uquotaip || mp->m_quotainfo->qi_gquotaip); + ASSERT(uip || gip); ASSERT(XFS_IS_QUOTA_RUNNING(mp)); xfs_notice(mp, "Quotacheck needed: Please wait."); @@ -1282,7 +1283,6 @@ xfs_qm_quotacheck( * their counters to zero. We need a clean slate. * We don't log our changes till later. */ - uip = mp->m_quotainfo->qi_uquotaip; if (uip) { error = xfs_qm_dqiterate(mp, uip, XFS_QMOPT_UQUOTA, &buffer_list); @@ -1291,7 +1291,6 @@ xfs_qm_quotacheck( flags |= XFS_UQUOTA_CHKD; } - gip = mp->m_quotainfo->qi_gquotaip; if (gip) { error = xfs_qm_dqiterate(mp, gip, XFS_IS_GQUOTA_ON(mp) ? XFS_QMOPT_GQUOTA : XFS_QMOPT_PQUOTA, @@ -1393,15 +1392,13 @@ STATIC int xfs_qm_init_quotainos( xfs_mount_t *mp) { - xfs_inode_t *uip, *gip; - int error; - __int64_t sbflags; - uint flags; + struct xfs_inode *uip = NULL; + struct xfs_inode *gip = NULL; + int error; + __int64_t sbflags = 0; + uint flags = 0; ASSERT(mp->m_quotainfo); - uip = gip = NULL; - sbflags = 0; - flags = 0; /* * Get the uquota and gquota inodes @@ -1410,19 +1407,18 @@ xfs_qm_init_quotainos( if (XFS_IS_UQUOTA_ON(mp) && mp->m_sb.sb_uquotino != NULLFSINO) { ASSERT(mp->m_sb.sb_uquotino > 0); - if ((error = xfs_iget(mp, NULL, mp->m_sb.sb_uquotino, - 0, 0, &uip))) + error = xfs_iget(mp, NULL, mp->m_sb.sb_uquotino, + 0, 0, &uip); + if (error) return XFS_ERROR(error); } if (XFS_IS_OQUOTA_ON(mp) && mp->m_sb.sb_gquotino != NULLFSINO) { ASSERT(mp->m_sb.sb_gquotino > 0); - if ((error = xfs_iget(mp, NULL, mp->m_sb.sb_gquotino, - 0, 0, &gip))) { - if (uip) - IRELE(uip); - return XFS_ERROR(error); - } + error = xfs_iget(mp, NULL, mp->m_sb.sb_gquotino, + 0, 0, &gip); + if (error) + goto error_rele; } } else { flags |= XFS_QMOPT_SBVERSION; @@ -1437,10 +1433,11 @@ xfs_qm_init_quotainos( * temporarily switch to read-write to do this. */ if (XFS_IS_UQUOTA_ON(mp) && uip == NULL) { - if ((error = xfs_qm_qino_alloc(mp, &uip, + error = xfs_qm_qino_alloc(mp, &uip, sbflags | XFS_SB_UQUOTINO, - flags | XFS_QMOPT_UQUOTA))) - return XFS_ERROR(error); + flags | XFS_QMOPT_UQUOTA); + if (error) + goto error_rele; flags &= ~XFS_QMOPT_SBVERSION; } @@ -1449,18 +1446,21 @@ xfs_qm_init_quotainos( XFS_QMOPT_GQUOTA : XFS_QMOPT_PQUOTA); error = xfs_qm_qino_alloc(mp, &gip, sbflags | XFS_SB_GQUOTINO, flags); - if (error) { - if (uip) - IRELE(uip); - - return XFS_ERROR(error); - } + if (error) + goto error_rele; } mp->m_quotainfo->qi_uquotaip = uip; mp->m_quotainfo->qi_gquotaip = gip; return 0; + +error_rele: + if (uip) + IRELE(uip); + if (gip) + IRELE(gip); + return XFS_ERROR(error); } STATIC void @@ -1657,7 +1657,8 @@ xfs_qm_vop_dqalloc( struct xfs_dquot **O_gdqpp) { struct xfs_mount *mp = ip->i_mount; - struct xfs_dquot *uq, *gq; + struct xfs_dquot *uq = NULL; + struct xfs_dquot *gq = NULL; int error; uint lockflags; @@ -1682,7 +1683,6 @@ xfs_qm_vop_dqalloc( } } - uq = gq = NULL; if ((flags & XFS_QMOPT_UQUOTA) && XFS_IS_UQUOTA_ON(mp)) { if (ip->i_d.di_uid != uid) { /* @@ -1695,11 +1695,12 @@ xfs_qm_vop_dqalloc( * holding ilock. */ xfs_iunlock(ip, lockflags); - if ((error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t) uid, + error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t) uid, XFS_DQ_USER, XFS_QMOPT_DQALLOC | XFS_QMOPT_DOWARN, - &uq))) { + &uq); + if (error) { ASSERT(error != ENOENT); return error; } @@ -1721,15 +1722,14 @@ xfs_qm_vop_dqalloc( if ((flags & XFS_QMOPT_GQUOTA) && XFS_IS_GQUOTA_ON(mp)) { if (ip->i_d.di_gid != gid) { xfs_iunlock(ip, lockflags); - if ((error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)gid, + error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)gid, XFS_DQ_GROUP, XFS_QMOPT_DQALLOC | XFS_QMOPT_DOWARN, - &gq))) { - if (uq) - xfs_qm_dqrele(uq); + &gq); + if (error) { ASSERT(error != ENOENT); - return error; + goto error_rele; } xfs_dqunlock(gq); lockflags = XFS_ILOCK_SHARED; @@ -1741,15 +1741,14 @@ xfs_qm_vop_dqalloc( } else if ((flags & XFS_QMOPT_PQUOTA) && XFS_IS_PQUOTA_ON(mp)) { if (xfs_get_projid(ip) != prid) { xfs_iunlock(ip, lockflags); - if ((error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)prid, + error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)prid, XFS_DQ_PROJ, XFS_QMOPT_DQALLOC | XFS_QMOPT_DOWARN, - &gq))) { - if (uq) - xfs_qm_dqrele(uq); + &gq); + if (error) { ASSERT(error != ENOENT); - return (error); + goto error_rele; } xfs_dqunlock(gq); lockflags = XFS_ILOCK_SHARED; @@ -1772,6 +1771,11 @@ xfs_qm_vop_dqalloc( else if (gq) xfs_qm_dqrele(gq); return 0; + +error_rele: + if (uq) + xfs_qm_dqrele(uq); + return error; } /* @@ -1819,29 +1823,31 @@ xfs_qm_vop_chown( */ int xfs_qm_vop_chown_reserve( - xfs_trans_t *tp, - xfs_inode_t *ip, - xfs_dquot_t *udqp, - xfs_dquot_t *gdqp, - uint flags) + struct xfs_trans *tp, + struct xfs_inode *ip, + struct xfs_dquot *udqp, + struct xfs_dquot *gdqp, + uint flags) { - xfs_mount_t *mp = ip->i_mount; - uint delblks, blkflags, prjflags = 0; - xfs_dquot_t *unresudq, *unresgdq, *delblksudq, *delblksgdq; - int error; + struct xfs_mount *mp = ip->i_mount; + uint delblks, blkflags, prjflags = 0; + struct xfs_dquot *udq_unres = NULL; + struct xfs_dquot *gdq_unres = NULL; + struct xfs_dquot *udq_delblks = NULL; + struct xfs_dquot *gdq_delblks = NULL; + int error; ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)); ASSERT(XFS_IS_QUOTA_RUNNING(mp)); delblks = ip->i_delayed_blks; - delblksudq = delblksgdq = unresudq = unresgdq = NULL; blkflags = XFS_IS_REALTIME_INODE(ip) ? XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS; if (XFS_IS_UQUOTA_ON(mp) && udqp && ip->i_d.di_uid != (uid_t)be32_to_cpu(udqp->q_core.d_id)) { - delblksudq = udqp; + udq_delblks = udqp; /* * If there are delayed allocation blocks, then we have to * unreserve those from the old dquot, and add them to the @@ -1849,7 +1855,7 @@ xfs_qm_vop_chown_reserve( */ if (delblks) { ASSERT(ip->i_udquot); - unresudq = ip->i_udquot; + udq_unres = ip->i_udquot; } } if (XFS_IS_OQUOTA_ON(ip->i_mount) && gdqp) { @@ -1860,18 +1866,19 @@ xfs_qm_vop_chown_reserve( if (prjflags || (XFS_IS_GQUOTA_ON(ip->i_mount) && ip->i_d.di_gid != be32_to_cpu(gdqp->q_core.d_id))) { - delblksgdq = gdqp; + gdq_delblks = gdqp; if (delblks) { ASSERT(ip->i_gdquot); - unresgdq = ip->i_gdquot; + gdq_unres = ip->i_gdquot; } } } - if ((error = xfs_trans_reserve_quota_bydquots(tp, ip->i_mount, - delblksudq, delblksgdq, ip->i_d.di_nblocks, 1, - flags | blkflags | prjflags))) - return (error); + error = xfs_trans_reserve_quota_bydquots(tp, ip->i_mount, + udq_delblks, gdq_delblks, ip->i_d.di_nblocks, 1, + flags | blkflags | prjflags); + if (error) + return error; /* * Do the delayed blks reservations/unreservations now. Since, these @@ -1883,14 +1890,15 @@ xfs_qm_vop_chown_reserve( /* * Do the reservations first. Unreservation can't fail. */ - ASSERT(delblksudq || delblksgdq); - ASSERT(unresudq || unresgdq); - if ((error = xfs_trans_reserve_quota_bydquots(NULL, ip->i_mount, - delblksudq, delblksgdq, (xfs_qcnt_t)delblks, 0, - flags | blkflags | prjflags))) - return (error); + ASSERT(udq_delblks || gdq_delblks); + ASSERT(udq_unres || gdq_unres); + error = xfs_trans_reserve_quota_bydquots(NULL, ip->i_mount, + udq_delblks, gdq_delblks, (xfs_qcnt_t)delblks, 0, + flags | blkflags | prjflags); + if (error) + return error; xfs_trans_reserve_quota_bydquots(NULL, ip->i_mount, - unresudq, unresgdq, -((xfs_qcnt_t)delblks), 0, + udq_unres, gdq_unres, -((xfs_qcnt_t)delblks), 0, blkflags); } diff --git a/fs/xfs/xfs_qm.h b/fs/xfs/xfs_qm.h index 051e43a5e20b..978bbb1a9e77 100644 --- a/fs/xfs/xfs_qm.h +++ b/fs/xfs/xfs_qm.h @@ -103,11 +103,13 @@ xfs_dq_to_quota_inode(struct xfs_dquot *dqp) extern int xfs_qm_calc_dquots_per_chunk(struct xfs_mount *mp, unsigned int nbblks); -extern void xfs_trans_mod_dquot(xfs_trans_t *, xfs_dquot_t *, uint, long); -extern int xfs_trans_reserve_quota_bydquots(xfs_trans_t *, xfs_mount_t *, - xfs_dquot_t *, xfs_dquot_t *, long, long, uint); -extern void xfs_trans_dqjoin(xfs_trans_t *, xfs_dquot_t *); -extern void xfs_trans_log_dquot(xfs_trans_t *, xfs_dquot_t *); +extern void xfs_trans_mod_dquot(struct xfs_trans *, + struct xfs_dquot *, uint, long); +extern int xfs_trans_reserve_quota_bydquots(struct xfs_trans *, + struct xfs_mount *, struct xfs_dquot *, + struct xfs_dquot *, long, long, uint); +extern void xfs_trans_dqjoin(struct xfs_trans *, struct xfs_dquot *); +extern void xfs_trans_log_dquot(struct xfs_trans *, struct xfs_dquot *); /* * We keep the usr and grp dquots separately so that locking will be easier @@ -132,22 +134,23 @@ typedef struct xfs_dquot_acct { #define XFS_QM_IWARNLIMIT 5 #define XFS_QM_RTBWARNLIMIT 5 -extern void xfs_qm_destroy_quotainfo(xfs_mount_t *); -extern int xfs_qm_quotacheck(xfs_mount_t *); -extern int xfs_qm_write_sb_changes(xfs_mount_t *, __int64_t); +extern void xfs_qm_destroy_quotainfo(struct xfs_mount *); +extern int xfs_qm_quotacheck(struct xfs_mount *); +extern int xfs_qm_write_sb_changes(struct xfs_mount *, __int64_t); /* dquot stuff */ -extern void xfs_qm_dqpurge_all(xfs_mount_t *, uint); -extern void xfs_qm_dqrele_all_inodes(xfs_mount_t *, uint); +extern void xfs_qm_dqpurge_all(struct xfs_mount *, uint); +extern void xfs_qm_dqrele_all_inodes(struct xfs_mount *, uint); /* quota ops */ -extern int xfs_qm_scall_trunc_qfiles(xfs_mount_t *, uint); -extern int xfs_qm_scall_getquota(xfs_mount_t *, xfs_dqid_t, uint, - fs_disk_quota_t *); +extern int xfs_qm_scall_trunc_qfiles(struct xfs_mount *, uint); +extern int xfs_qm_scall_getquota(struct xfs_mount *, xfs_dqid_t, + uint, struct fs_disk_quota *); extern int xfs_qm_scall_setqlim(struct xfs_mount *, xfs_dqid_t, uint, - fs_disk_quota_t *); -extern int xfs_qm_scall_getqstat(xfs_mount_t *, fs_quota_stat_t *); -extern int xfs_qm_scall_quotaon(xfs_mount_t *, uint); -extern int xfs_qm_scall_quotaoff(xfs_mount_t *, uint); + struct fs_disk_quota *); +extern int xfs_qm_scall_getqstat(struct xfs_mount *, + struct fs_quota_stat *); +extern int xfs_qm_scall_quotaon(struct xfs_mount *, uint); +extern int xfs_qm_scall_quotaoff(struct xfs_mount *, uint); #endif /* __XFS_QM_H__ */ diff --git a/fs/xfs/xfs_qm_syscalls.c b/fs/xfs/xfs_qm_syscalls.c index 6cdf6ffc36a1..b03b2abd1e13 100644 --- a/fs/xfs/xfs_qm_syscalls.c +++ b/fs/xfs/xfs_qm_syscalls.c @@ -407,11 +407,11 @@ xfs_qm_scall_getqstat( struct fs_quota_stat *out) { struct xfs_quotainfo *q = mp->m_quotainfo; - struct xfs_inode *uip, *gip; - bool tempuqip, tempgqip; + struct xfs_inode *uip = NULL; + struct xfs_inode *gip = NULL; + bool tempuqip = false; + bool tempgqip = false; - uip = gip = NULL; - tempuqip = tempgqip = false; memset(out, 0, sizeof(fs_quota_stat_t)); out->qs_version = FS_QSTAT_VERSION; diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c index 738c04be2019..e830fb56e27f 100644 --- a/fs/xfs/xfs_symlink.c +++ b/fs/xfs/xfs_symlink.c @@ -358,7 +358,8 @@ xfs_symlink( int n; xfs_buf_t *bp; prid_t prid; - struct xfs_dquot *udqp, *gdqp; + struct xfs_dquot *udqp = NULL; + struct xfs_dquot *gdqp = NULL; uint resblks; *ipp = NULL; diff --git a/fs/xfs/xfs_trans_dquot.c b/fs/xfs/xfs_trans_dquot.c index 3fa369c75a4d..45b3e2d0773a 100644 --- a/fs/xfs/xfs_trans_dquot.c +++ b/fs/xfs/xfs_trans_dquot.c @@ -169,13 +169,13 @@ xfs_trans_mod_dquot_byino( (void) xfs_trans_mod_dquot(tp, ip->i_gdquot, field, delta); } -STATIC xfs_dqtrx_t * +STATIC struct xfs_dqtrx * xfs_trans_get_dqtrx( - xfs_trans_t *tp, - xfs_dquot_t *dqp) + struct xfs_trans *tp, + struct xfs_dquot *dqp) { - int i; - xfs_dqtrx_t *qa; + int i; + struct xfs_dqtrx *qa; qa = XFS_QM_ISUDQ(dqp) ? tp->t_dqinfo->dqa_usrdquots : tp->t_dqinfo->dqa_grpdquots; @@ -747,15 +747,15 @@ error_return: */ int xfs_trans_reserve_quota_bydquots( - xfs_trans_t *tp, - xfs_mount_t *mp, - xfs_dquot_t *udqp, - xfs_dquot_t *gdqp, - long nblks, - long ninos, - uint flags) + struct xfs_trans *tp, + struct xfs_mount *mp, + struct xfs_dquot *udqp, + struct xfs_dquot *gdqp, + long nblks, + long ninos, + uint flags) { - int resvd = 0, error; + int error; if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp)) return 0; @@ -770,28 +770,24 @@ xfs_trans_reserve_quota_bydquots( (flags & ~XFS_QMOPT_ENOSPC)); if (error) return error; - resvd = 1; } if (gdqp) { error = xfs_trans_dqresv(tp, mp, gdqp, nblks, ninos, flags); - if (error) { - /* - * can't do it, so backout previous reservation - */ - if (resvd) { - flags |= XFS_QMOPT_FORCE_RES; - xfs_trans_dqresv(tp, mp, udqp, - -nblks, -ninos, flags); - } - return error; - } + if (error) + goto unwind_usr; } /* * Didn't change anything critical, so, no need to log */ return 0; + +unwind_usr: + flags |= XFS_QMOPT_FORCE_RES; + if (udqp) + xfs_trans_dqresv(tp, mp, udqp, -nblks, -ninos, flags); + return error; } From 0e6436d99e730f9384f3c11d24a5efb788885fc6 Mon Sep 17 00:00:00 2001 From: Chandra Seetharaman Date: Thu, 27 Jun 2013 17:25:09 -0500 Subject: [PATCH 60/61] xfs: Change xfs_dquot_acct to be a 2-dimensional array In preparation for combined pquota/gquota support, for the sake of readability, change xfs_dquot_acct to be a 2-dimensional array. Signed-off-by: Chandra Seetharaman Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_qm.h | 12 ++++++++---- fs/xfs/xfs_trans_dquot.c | 31 ++++++++++++------------------- 2 files changed, 20 insertions(+), 23 deletions(-) diff --git a/fs/xfs/xfs_qm.h b/fs/xfs/xfs_qm.h index 978bbb1a9e77..bdb4f8b95207 100644 --- a/fs/xfs/xfs_qm.h +++ b/fs/xfs/xfs_qm.h @@ -116,11 +116,15 @@ extern void xfs_trans_log_dquot(struct xfs_trans *, struct xfs_dquot *); * to do at commit time. All transactions that we know of at this point * affect no more than two dquots of one type. Hence, the TRANS_MAXDQS value. */ +enum { + XFS_QM_TRANS_USR = 0, + XFS_QM_TRANS_GRP, + XFS_QM_TRANS_DQTYPES +}; #define XFS_QM_TRANS_MAXDQS 2 -typedef struct xfs_dquot_acct { - xfs_dqtrx_t dqa_usrdquots[XFS_QM_TRANS_MAXDQS]; - xfs_dqtrx_t dqa_grpdquots[XFS_QM_TRANS_MAXDQS]; -} xfs_dquot_acct_t; +struct xfs_dquot_acct { + struct xfs_dqtrx dqs[XFS_QM_TRANS_DQTYPES][XFS_QM_TRANS_MAXDQS]; +}; /* * Users are allowed to have a usage exceeding their softlimit for diff --git a/fs/xfs/xfs_trans_dquot.c b/fs/xfs/xfs_trans_dquot.c index 45b3e2d0773a..7ea7485f9fb8 100644 --- a/fs/xfs/xfs_trans_dquot.c +++ b/fs/xfs/xfs_trans_dquot.c @@ -103,8 +103,6 @@ xfs_trans_dup_dqinfo( return; xfs_trans_alloc_dqinfo(ntp); - oqa = otp->t_dqinfo->dqa_usrdquots; - nqa = ntp->t_dqinfo->dqa_usrdquots; /* * Because the quota blk reservation is carried forward, @@ -113,7 +111,9 @@ xfs_trans_dup_dqinfo( if(otp->t_flags & XFS_TRANS_DQ_DIRTY) ntp->t_flags |= XFS_TRANS_DQ_DIRTY; - for (j = 0; j < 2; j++) { + for (j = 0; j < XFS_QM_TRANS_DQTYPES; j++) { + oqa = otp->t_dqinfo->dqs[j]; + nqa = ntp->t_dqinfo->dqs[j]; for (i = 0; i < XFS_QM_TRANS_MAXDQS; i++) { if (oqa[i].qt_dquot == NULL) break; @@ -138,8 +138,6 @@ xfs_trans_dup_dqinfo( oq->qt_ino_res = oq->qt_ino_res_used; } - oqa = otp->t_dqinfo->dqa_grpdquots; - nqa = ntp->t_dqinfo->dqa_grpdquots; } } @@ -177,8 +175,10 @@ xfs_trans_get_dqtrx( int i; struct xfs_dqtrx *qa; - qa = XFS_QM_ISUDQ(dqp) ? - tp->t_dqinfo->dqa_usrdquots : tp->t_dqinfo->dqa_grpdquots; + if (XFS_QM_ISUDQ(dqp)) + qa = tp->t_dqinfo->dqs[XFS_QM_TRANS_USR]; + else + qa = tp->t_dqinfo->dqs[XFS_QM_TRANS_GRP]; for (i = 0; i < XFS_QM_TRANS_MAXDQS; i++) { if (qa[i].qt_dquot == NULL || @@ -338,12 +338,10 @@ xfs_trans_apply_dquot_deltas( return; ASSERT(tp->t_dqinfo); - qa = tp->t_dqinfo->dqa_usrdquots; - for (j = 0; j < 2; j++) { - if (qa[0].qt_dquot == NULL) { - qa = tp->t_dqinfo->dqa_grpdquots; + for (j = 0; j < XFS_QM_TRANS_DQTYPES; j++) { + qa = tp->t_dqinfo->dqs[j]; + if (qa[0].qt_dquot == NULL) continue; - } /* * Lock all of the dquots and join them to the transaction. @@ -494,10 +492,6 @@ xfs_trans_apply_dquot_deltas( ASSERT(dqp->q_res_rtbcount >= be64_to_cpu(dqp->q_core.d_rtbcount)); } - /* - * Do the group quotas next - */ - qa = tp->t_dqinfo->dqa_grpdquots; } } @@ -520,9 +514,9 @@ xfs_trans_unreserve_and_mod_dquots( if (!tp->t_dqinfo || !(tp->t_flags & XFS_TRANS_DQ_DIRTY)) return; - qa = tp->t_dqinfo->dqa_usrdquots; + for (j = 0; j < XFS_QM_TRANS_DQTYPES; j++) { + qa = tp->t_dqinfo->dqs[j]; - for (j = 0; j < 2; j++) { for (i = 0; i < XFS_QM_TRANS_MAXDQS; i++) { qtrx = &qa[i]; /* @@ -564,7 +558,6 @@ xfs_trans_unreserve_and_mod_dquots( xfs_dqunlock(dqp); } - qa = tp->t_dqinfo->dqa_grpdquots; } } From 83e782e1a1cc0159888e58e14dfc8f3289663338 Mon Sep 17 00:00:00 2001 From: Chandra Seetharaman Date: Thu, 27 Jun 2013 17:25:10 -0500 Subject: [PATCH 61/61] xfs: Remove incore use of XFS_OQUOTA_ENFD and XFS_OQUOTA_CHKD Remove all incore use of XFS_OQUOTA_ENFD and XFS_OQUOTA_CHKD. Instead, start using XFS_GQUOTA_.* XFS_PQUOTA_.* counterparts for GQUOTA and PQUOTA respectively. On-disk copy still uses XFS_OQUOTA_ENFD and XFS_OQUOTA_CHKD. Read and write of the superblock does the conversion from *OQUOTA* to *[PG]QUOTA*. Signed-off-by: Chandra Seetharaman Reviewed-by: Ben Myers Signed-off-by: Ben Myers --- fs/xfs/xfs_mount.c | 51 ++++++++++++++++++++++++++++++++++++++++ fs/xfs/xfs_qm.c | 9 ++++--- fs/xfs/xfs_qm_syscalls.c | 43 +++++++++++++++++---------------- fs/xfs/xfs_quota.h | 47 ++++++++++++++++++++++-------------- fs/xfs/xfs_quotaops.c | 6 +++-- fs/xfs/xfs_super.c | 16 ++++++------- fs/xfs/xfs_trans_dquot.c | 4 ++-- 7 files changed, 123 insertions(+), 53 deletions(-) diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 2978bb4cc72a..2b0ba3581656 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -336,6 +336,14 @@ xfs_mount_validate_sb( return XFS_ERROR(EWRONGFS); } + if ((sbp->sb_qflags & (XFS_OQUOTA_ENFD | XFS_OQUOTA_CHKD)) && + (sbp->sb_qflags & (XFS_PQUOTA_ENFD | XFS_GQUOTA_ENFD | + XFS_PQUOTA_CHKD | XFS_GQUOTA_CHKD))) { + xfs_notice(mp, +"Super block has XFS_OQUOTA bits along with XFS_PQUOTA and/or XFS_GQUOTA bits.\n"); + return XFS_ERROR(EFSCORRUPTED); + } + /* * Version 5 superblock feature mask validation. Reject combinations the * kernel cannot support up front before checking anything else. For @@ -561,6 +569,18 @@ out_unwind: return error; } +static void +xfs_sb_quota_from_disk(struct xfs_sb *sbp) +{ + if (sbp->sb_qflags & XFS_OQUOTA_ENFD) + sbp->sb_qflags |= (sbp->sb_qflags & XFS_PQUOTA_ACCT) ? + XFS_PQUOTA_ENFD : XFS_GQUOTA_ENFD; + if (sbp->sb_qflags & XFS_OQUOTA_CHKD) + sbp->sb_qflags |= (sbp->sb_qflags & XFS_PQUOTA_ACCT) ? + XFS_PQUOTA_CHKD : XFS_GQUOTA_CHKD; + sbp->sb_qflags &= ~(XFS_OQUOTA_ENFD | XFS_OQUOTA_CHKD); +} + void xfs_sb_from_disk( struct xfs_sb *to, @@ -622,6 +642,35 @@ xfs_sb_from_disk( to->sb_lsn = be64_to_cpu(from->sb_lsn); } +static inline void +xfs_sb_quota_to_disk( + xfs_dsb_t *to, + xfs_sb_t *from, + __int64_t *fields) +{ + __uint16_t qflags = from->sb_qflags; + + if (*fields & XFS_SB_QFLAGS) { + /* + * The in-core version of sb_qflags do not have + * XFS_OQUOTA_* flags, whereas the on-disk version + * does. So, convert incore XFS_{PG}QUOTA_* flags + * to on-disk XFS_OQUOTA_* flags. + */ + qflags &= ~(XFS_PQUOTA_ENFD | XFS_PQUOTA_CHKD | + XFS_GQUOTA_ENFD | XFS_GQUOTA_CHKD); + + if (from->sb_qflags & + (XFS_PQUOTA_ENFD | XFS_GQUOTA_ENFD)) + qflags |= XFS_OQUOTA_ENFD; + if (from->sb_qflags & + (XFS_PQUOTA_CHKD | XFS_GQUOTA_CHKD)) + qflags |= XFS_OQUOTA_CHKD; + to->sb_qflags = cpu_to_be16(qflags); + *fields &= ~XFS_SB_QFLAGS; + } +} + /* * Copy in core superblock to ondisk one. * @@ -643,6 +692,7 @@ xfs_sb_to_disk( if (!fields) return; + xfs_sb_quota_to_disk(to, from, &fields); while (fields) { f = (xfs_sb_field_t)xfs_lowbit64((__uint64_t)fields); first = xfs_sb_info[f].offset; @@ -835,6 +885,7 @@ reread: */ xfs_sb_from_disk(&mp->m_sb, XFS_BUF_TO_SBP(bp)); + xfs_sb_quota_from_disk(&mp->m_sb); /* * We must be able to do sector-sized and sector-aligned IO. */ diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index cf09aa845d8c..7a3e007b49f4 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -299,8 +299,10 @@ xfs_qm_mount_quotas( */ if (!XFS_IS_UQUOTA_ON(mp)) mp->m_qflags &= ~XFS_UQUOTA_CHKD; - if (!(XFS_IS_GQUOTA_ON(mp) || XFS_IS_PQUOTA_ON(mp))) - mp->m_qflags &= ~XFS_OQUOTA_CHKD; + if (!XFS_IS_GQUOTA_ON(mp)) + mp->m_qflags &= ~XFS_GQUOTA_CHKD; + if (!XFS_IS_PQUOTA_ON(mp)) + mp->m_qflags &= ~XFS_PQUOTA_CHKD; write_changes: /* @@ -1297,7 +1299,8 @@ xfs_qm_quotacheck( &buffer_list); if (error) goto error_return; - flags |= XFS_OQUOTA_CHKD; + flags |= XFS_IS_GQUOTA_ON(mp) ? + XFS_GQUOTA_CHKD : XFS_PQUOTA_CHKD; } do { diff --git a/fs/xfs/xfs_qm_syscalls.c b/fs/xfs/xfs_qm_syscalls.c index b03b2abd1e13..a08801ae24e2 100644 --- a/fs/xfs/xfs_qm_syscalls.c +++ b/fs/xfs/xfs_qm_syscalls.c @@ -117,11 +117,11 @@ xfs_qm_scall_quotaoff( } if (flags & XFS_GQUOTA_ACCT) { dqtype |= XFS_QMOPT_GQUOTA; - flags |= (XFS_OQUOTA_CHKD | XFS_OQUOTA_ENFD); + flags |= (XFS_GQUOTA_CHKD | XFS_GQUOTA_ENFD); inactivate_flags |= XFS_GQUOTA_ACTIVE; } else if (flags & XFS_PQUOTA_ACCT) { dqtype |= XFS_QMOPT_PQUOTA; - flags |= (XFS_OQUOTA_CHKD | XFS_OQUOTA_ENFD); + flags |= (XFS_PQUOTA_CHKD | XFS_PQUOTA_ENFD); inactivate_flags |= XFS_PQUOTA_ACTIVE; } @@ -335,14 +335,14 @@ xfs_qm_scall_quotaon( * quota acct on ondisk without m_qflags' knowing. */ if (((flags & XFS_UQUOTA_ACCT) == 0 && - (mp->m_sb.sb_qflags & XFS_UQUOTA_ACCT) == 0 && - (flags & XFS_UQUOTA_ENFD)) - || + (mp->m_sb.sb_qflags & XFS_UQUOTA_ACCT) == 0 && + (flags & XFS_UQUOTA_ENFD)) || + ((flags & XFS_GQUOTA_ACCT) == 0 && + (mp->m_sb.sb_qflags & XFS_GQUOTA_ACCT) == 0 && + (flags & XFS_GQUOTA_ENFD)) || ((flags & XFS_PQUOTA_ACCT) == 0 && - (mp->m_sb.sb_qflags & XFS_PQUOTA_ACCT) == 0 && - (flags & XFS_GQUOTA_ACCT) == 0 && - (mp->m_sb.sb_qflags & XFS_GQUOTA_ACCT) == 0 && - (flags & XFS_OQUOTA_ENFD))) { + (mp->m_sb.sb_qflags & XFS_PQUOTA_ACCT) == 0 && + (flags & XFS_PQUOTA_ENFD))) { xfs_debug(mp, "%s: Can't enforce without acct, flags=%x sbflags=%x\n", __func__, flags, mp->m_sb.sb_qflags); @@ -776,9 +776,12 @@ xfs_qm_scall_getquota( * gets turned off. No need to confuse the user level code, * so return zeroes in that case. */ - if ((!XFS_IS_UQUOTA_ENFORCED(mp) && dqp->q_core.d_flags == XFS_DQ_USER) || - (!XFS_IS_OQUOTA_ENFORCED(mp) && - (dqp->q_core.d_flags & (XFS_DQ_PROJ | XFS_DQ_GROUP)))) { + if ((!XFS_IS_UQUOTA_ENFORCED(mp) && + dqp->q_core.d_flags == XFS_DQ_USER) || + (!XFS_IS_GQUOTA_ENFORCED(mp) && + dqp->q_core.d_flags == XFS_DQ_GROUP) || + (!XFS_IS_PQUOTA_ENFORCED(mp) && + dqp->q_core.d_flags == XFS_DQ_PROJ)) { dst->d_btimer = 0; dst->d_itimer = 0; dst->d_rtbtimer = 0; @@ -786,8 +789,8 @@ xfs_qm_scall_getquota( #ifdef DEBUG if (((XFS_IS_UQUOTA_ENFORCED(mp) && dst->d_flags == FS_USER_QUOTA) || - (XFS_IS_OQUOTA_ENFORCED(mp) && - (dst->d_flags & (FS_PROJ_QUOTA | FS_GROUP_QUOTA)))) && + (XFS_IS_GQUOTA_ENFORCED(mp) && dst->d_flags == FS_GROUP_QUOTA) || + (XFS_IS_PQUOTA_ENFORCED(mp) && dst->d_flags == FS_PROJ_QUOTA)) && dst->d_id != 0) { if ((dst->d_bcount > dst->d_blk_softlimit) && (dst->d_blk_softlimit > 0)) { @@ -833,16 +836,16 @@ xfs_qm_export_flags( uflags = 0; if (flags & XFS_UQUOTA_ACCT) uflags |= FS_QUOTA_UDQ_ACCT; - if (flags & XFS_PQUOTA_ACCT) - uflags |= FS_QUOTA_PDQ_ACCT; if (flags & XFS_GQUOTA_ACCT) uflags |= FS_QUOTA_GDQ_ACCT; + if (flags & XFS_PQUOTA_ACCT) + uflags |= FS_QUOTA_PDQ_ACCT; if (flags & XFS_UQUOTA_ENFD) uflags |= FS_QUOTA_UDQ_ENFD; - if (flags & (XFS_OQUOTA_ENFD)) { - uflags |= (flags & XFS_GQUOTA_ACCT) ? - FS_QUOTA_GDQ_ENFD : FS_QUOTA_PDQ_ENFD; - } + if (flags & XFS_GQUOTA_ENFD) + uflags |= FS_QUOTA_GDQ_ENFD; + if (flags & XFS_PQUOTA_ENFD) + uflags |= FS_QUOTA_PDQ_ENFD; return (uflags); } diff --git a/fs/xfs/xfs_quota.h b/fs/xfs/xfs_quota.h index c38068f26c55..c3483bab9cde 100644 --- a/fs/xfs/xfs_quota.h +++ b/fs/xfs/xfs_quota.h @@ -160,31 +160,43 @@ typedef struct xfs_qoff_logformat { #define XFS_OQUOTA_CHKD 0x0020 /* quotacheck run on other (grp/prj) quotas */ #define XFS_GQUOTA_ACCT 0x0040 /* group quota accounting ON */ +/* + * Conversion to and from the combined OQUOTA flag (if necessary) + * is done only in xfs_sb_qflags_to_disk() and xfs_sb_qflags_from_disk() + */ +#define XFS_GQUOTA_ENFD 0x0080 /* group quota limits enforced */ +#define XFS_GQUOTA_CHKD 0x0100 /* quotacheck run on group quotas */ +#define XFS_PQUOTA_ENFD 0x0200 /* project quota limits enforced */ +#define XFS_PQUOTA_CHKD 0x0400 /* quotacheck run on project quotas */ + /* * Quota Accounting/Enforcement flags */ #define XFS_ALL_QUOTA_ACCT \ (XFS_UQUOTA_ACCT | XFS_GQUOTA_ACCT | XFS_PQUOTA_ACCT) -#define XFS_ALL_QUOTA_ENFD (XFS_UQUOTA_ENFD | XFS_OQUOTA_ENFD) -#define XFS_ALL_QUOTA_CHKD (XFS_UQUOTA_CHKD | XFS_OQUOTA_CHKD) +#define XFS_ALL_QUOTA_ENFD \ + (XFS_UQUOTA_ENFD | XFS_GQUOTA_ENFD | XFS_PQUOTA_ENFD) +#define XFS_ALL_QUOTA_CHKD \ + (XFS_UQUOTA_CHKD | XFS_GQUOTA_CHKD | XFS_PQUOTA_CHKD) #define XFS_IS_QUOTA_RUNNING(mp) ((mp)->m_qflags & XFS_ALL_QUOTA_ACCT) #define XFS_IS_UQUOTA_RUNNING(mp) ((mp)->m_qflags & XFS_UQUOTA_ACCT) #define XFS_IS_PQUOTA_RUNNING(mp) ((mp)->m_qflags & XFS_PQUOTA_ACCT) #define XFS_IS_GQUOTA_RUNNING(mp) ((mp)->m_qflags & XFS_GQUOTA_ACCT) #define XFS_IS_UQUOTA_ENFORCED(mp) ((mp)->m_qflags & XFS_UQUOTA_ENFD) -#define XFS_IS_OQUOTA_ENFORCED(mp) ((mp)->m_qflags & XFS_OQUOTA_ENFD) +#define XFS_IS_GQUOTA_ENFORCED(mp) ((mp)->m_qflags & XFS_GQUOTA_ENFD) +#define XFS_IS_PQUOTA_ENFORCED(mp) ((mp)->m_qflags & XFS_PQUOTA_ENFD) /* * Incore only flags for quotaoff - these bits get cleared when quota(s) * are in the process of getting turned off. These flags are in m_qflags but * never in sb_qflags. */ -#define XFS_UQUOTA_ACTIVE 0x0100 /* uquotas are being turned off */ -#define XFS_PQUOTA_ACTIVE 0x0200 /* pquotas are being turned off */ -#define XFS_GQUOTA_ACTIVE 0x0400 /* gquotas are being turned off */ +#define XFS_UQUOTA_ACTIVE 0x1000 /* uquotas are being turned off */ +#define XFS_GQUOTA_ACTIVE 0x2000 /* gquotas are being turned off */ +#define XFS_PQUOTA_ACTIVE 0x4000 /* pquotas are being turned off */ #define XFS_ALL_QUOTA_ACTIVE \ - (XFS_UQUOTA_ACTIVE | XFS_PQUOTA_ACTIVE | XFS_GQUOTA_ACTIVE) + (XFS_UQUOTA_ACTIVE | XFS_GQUOTA_ACTIVE | XFS_PQUOTA_ACTIVE) /* * Checking XFS_IS_*QUOTA_ON() while holding any inode lock guarantees @@ -268,24 +280,23 @@ typedef struct xfs_qoff_logformat { ((XFS_IS_UQUOTA_ON(mp) && \ (mp->m_sb.sb_qflags & XFS_UQUOTA_CHKD) == 0) || \ (XFS_IS_GQUOTA_ON(mp) && \ - ((mp->m_sb.sb_qflags & XFS_OQUOTA_CHKD) == 0 || \ - (mp->m_sb.sb_qflags & XFS_PQUOTA_ACCT))) || \ + (mp->m_sb.sb_qflags & XFS_GQUOTA_CHKD) == 0) || \ (XFS_IS_PQUOTA_ON(mp) && \ - ((mp->m_sb.sb_qflags & XFS_OQUOTA_CHKD) == 0 || \ - (mp->m_sb.sb_qflags & XFS_GQUOTA_ACCT)))) + (mp->m_sb.sb_qflags & XFS_PQUOTA_CHKD) == 0)) #define XFS_MOUNT_QUOTA_SET1 (XFS_UQUOTA_ACCT|XFS_UQUOTA_ENFD|\ - XFS_UQUOTA_CHKD|XFS_PQUOTA_ACCT|\ - XFS_OQUOTA_ENFD|XFS_OQUOTA_CHKD) + XFS_UQUOTA_CHKD|XFS_GQUOTA_ACCT|\ + XFS_GQUOTA_ENFD|XFS_GQUOTA_CHKD) #define XFS_MOUNT_QUOTA_SET2 (XFS_UQUOTA_ACCT|XFS_UQUOTA_ENFD|\ - XFS_UQUOTA_CHKD|XFS_GQUOTA_ACCT|\ - XFS_OQUOTA_ENFD|XFS_OQUOTA_CHKD) + XFS_UQUOTA_CHKD|XFS_PQUOTA_ACCT|\ + XFS_PQUOTA_ENFD|XFS_PQUOTA_CHKD) #define XFS_MOUNT_QUOTA_ALL (XFS_UQUOTA_ACCT|XFS_UQUOTA_ENFD|\ - XFS_UQUOTA_CHKD|XFS_PQUOTA_ACCT|\ - XFS_OQUOTA_ENFD|XFS_OQUOTA_CHKD|\ - XFS_GQUOTA_ACCT) + XFS_UQUOTA_CHKD|XFS_GQUOTA_ACCT|\ + XFS_GQUOTA_ENFD|XFS_GQUOTA_CHKD|\ + XFS_PQUOTA_ACCT|XFS_PQUOTA_ENFD|\ + XFS_PQUOTA_CHKD) /* diff --git a/fs/xfs/xfs_quotaops.c b/fs/xfs/xfs_quotaops.c index 71926d630527..20e30f93b0c7 100644 --- a/fs/xfs/xfs_quotaops.c +++ b/fs/xfs/xfs_quotaops.c @@ -75,8 +75,10 @@ xfs_fs_set_xstate( flags |= XFS_GQUOTA_ACCT; if (uflags & FS_QUOTA_UDQ_ENFD) flags |= XFS_UQUOTA_ENFD; - if (uflags & (FS_QUOTA_PDQ_ENFD|FS_QUOTA_GDQ_ENFD)) - flags |= XFS_OQUOTA_ENFD; + if (uflags & FS_QUOTA_GDQ_ENFD) + flags |= XFS_GQUOTA_ENFD; + if (uflags & FS_QUOTA_PDQ_ENFD) + flags |= XFS_PQUOTA_ENFD; switch (op) { case Q_XQUOTAON: diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index e5e8b5ee8bfa..1d68ffcdeaa7 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -360,17 +360,17 @@ xfs_parseargs( } else if (!strcmp(this_char, MNTOPT_PQUOTA) || !strcmp(this_char, MNTOPT_PRJQUOTA)) { mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE | - XFS_OQUOTA_ENFD); + XFS_PQUOTA_ENFD); } else if (!strcmp(this_char, MNTOPT_PQUOTANOENF)) { mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE); - mp->m_qflags &= ~XFS_OQUOTA_ENFD; + mp->m_qflags &= ~XFS_PQUOTA_ENFD; } else if (!strcmp(this_char, MNTOPT_GQUOTA) || !strcmp(this_char, MNTOPT_GRPQUOTA)) { mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE | - XFS_OQUOTA_ENFD); + XFS_GQUOTA_ENFD); } else if (!strcmp(this_char, MNTOPT_GQUOTANOENF)) { mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE); - mp->m_qflags &= ~XFS_OQUOTA_ENFD; + mp->m_qflags &= ~XFS_GQUOTA_ENFD; } else if (!strcmp(this_char, MNTOPT_DELAYLOG)) { xfs_warn(mp, "delaylog is the default now, option is deprecated."); @@ -559,12 +559,12 @@ xfs_showargs( /* Either project or group quotas can be active, not both */ if (mp->m_qflags & XFS_PQUOTA_ACCT) { - if (mp->m_qflags & XFS_OQUOTA_ENFD) + if (mp->m_qflags & XFS_PQUOTA_ENFD) seq_puts(m, "," MNTOPT_PRJQUOTA); else seq_puts(m, "," MNTOPT_PQUOTANOENF); } else if (mp->m_qflags & XFS_GQUOTA_ACCT) { - if (mp->m_qflags & XFS_OQUOTA_ENFD) + if (mp->m_qflags & XFS_GQUOTA_ENFD) seq_puts(m, "," MNTOPT_GRPQUOTA); else seq_puts(m, "," MNTOPT_GQUOTANOENF); @@ -1132,8 +1132,8 @@ xfs_fs_statfs( spin_unlock(&mp->m_sb_lock); if ((ip->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) && - ((mp->m_qflags & (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))) == - (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD)) + ((mp->m_qflags & (XFS_PQUOTA_ACCT|XFS_PQUOTA_ENFD))) == + (XFS_PQUOTA_ACCT|XFS_PQUOTA_ENFD)) xfs_qm_statvfs(ip, statp); return 0; } diff --git a/fs/xfs/xfs_trans_dquot.c b/fs/xfs/xfs_trans_dquot.c index 7ea7485f9fb8..3ba64d540168 100644 --- a/fs/xfs/xfs_trans_dquot.c +++ b/fs/xfs/xfs_trans_dquot.c @@ -632,8 +632,8 @@ xfs_trans_dqresv( if ((flags & XFS_QMOPT_FORCE_RES) == 0 && dqp->q_core.d_id && ((XFS_IS_UQUOTA_ENFORCED(dqp->q_mount) && XFS_QM_ISUDQ(dqp)) || - (XFS_IS_OQUOTA_ENFORCED(dqp->q_mount) && - (XFS_QM_ISPDQ(dqp) || XFS_QM_ISGDQ(dqp))))) { + (XFS_IS_GQUOTA_ENFORCED(dqp->q_mount) && XFS_QM_ISGDQ(dqp)) || + (XFS_IS_PQUOTA_ENFORCED(dqp->q_mount) && XFS_QM_ISPDQ(dqp)))) { if (nblks > 0) { /* * dquot is locked already. See if we'd go over the