From d796c52ef0b71a988364f6109aeb63d79c5b116b Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Sun, 5 Aug 2012 19:04:57 -0400 Subject: ext4: make sure the journal sb is written in ext4_clear_journal_err() After we transfer set the EXT4_ERROR_FS bit in the file system superblock, it's not enough to call jbd2_journal_clear_err() to clear the error indication from journal superblock --- we need to call jbd2_journal_update_sb_errno() as well. Otherwise, when the root file system is mounted read-only, the journal is replayed, and the error indicator is transferred to the superblock --- but the s_errno field in the jbd2 superblock is left set (since although we cleared it in memory, we never flushed it out to disk). This can end up confusing e2fsck. We should make e2fsck more robust in this case, but the kernel shouldn't be leaving things in this confused state, either. Signed-off-by: "Theodore Ts'o" Cc: stable@kernel.org --- fs/ext4/super.c | 1 + fs/jbd2/journal.c | 3 ++- include/linux/jbd2.h | 1 + 3 files changed, 4 insertions(+), 1 deletion(-) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index d76ec8277d3f..ccc4bcad5616 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -4430,6 +4430,7 @@ static void ext4_clear_journal_err(struct super_block *sb, ext4_commit_super(sb, 1); jbd2_journal_clear_err(journal); + jbd2_journal_update_sb_errno(journal); } } diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index e9a3c4c85594..bd23f2ebaa67 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -1377,7 +1377,7 @@ static void jbd2_mark_journal_empty(journal_t *journal) * Update a journal's errno. Write updated superblock to disk waiting for IO * to complete. */ -static void jbd2_journal_update_sb_errno(journal_t *journal) +void jbd2_journal_update_sb_errno(journal_t *journal) { journal_superblock_t *sb = journal->j_superblock; @@ -1390,6 +1390,7 @@ static void jbd2_journal_update_sb_errno(journal_t *journal) jbd2_write_superblock(journal, WRITE_SYNC); } +EXPORT_SYMBOL(jbd2_journal_update_sb_errno); /* * Read the superblock for a given journal, performing initial diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h index f334c7fab967..3efc43f3f162 100644 --- a/include/linux/jbd2.h +++ b/include/linux/jbd2.h @@ -1125,6 +1125,7 @@ extern int jbd2_journal_destroy (journal_t *); extern int jbd2_journal_recover (journal_t *journal); extern int jbd2_journal_wipe (journal_t *, int); extern int jbd2_journal_skip_recovery (journal_t *); +extern void jbd2_journal_update_sb_errno(journal_t *); extern void jbd2_journal_update_sb_log_tail (journal_t *, tid_t, unsigned long, int); extern void __jbd2_journal_abort_hard (journal_t *); -- cgit v1.2.3-55-g7522 From 7e731bc9a12339f344cddf82166b82633d99dd86 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Sun, 5 Aug 2012 23:28:16 -0400 Subject: ext4: avoid kmemcheck complaint from reading uninitialized memory Commit 03179fe923 introduced a kmemcheck complaint in ext4_da_get_block_prep() because we save and restore ei->i_da_metadata_calc_last_lblock even though it is left uninitialized in the case where i_da_metadata_calc_len is zero. This doesn't hurt anything, but silencing the kmemcheck complaint makes it easier for people to find real bugs. Addresses https://bugzilla.kernel.org/show_bug.cgi?id=45631 (which is marked as a regression). Signed-off-by: "Theodore Ts'o" Cc: stable@vger.kernel.org --- fs/ext4/super.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index ccc4bcad5616..56bcaec9149c 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -959,6 +959,7 @@ static struct inode *ext4_alloc_inode(struct super_block *sb) ei->i_reserved_meta_blocks = 0; ei->i_allocated_meta_blocks = 0; ei->i_da_metadata_calc_len = 0; + ei->i_da_metadata_calc_last_lblock = 0; spin_lock_init(&(ei->i_block_reservation_lock)); #ifdef CONFIG_QUOTA ei->i_reserved_quota = 0; -- cgit v1.2.3-55-g7522 From 7a4c5de27efa4c2ecca87af0a3deea63446367e2 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Fri, 10 Aug 2012 13:57:52 -0400 Subject: ext4: don't call ext4_error while block group is locked While in ext4_validate_block_bitmap(), if an block allocation bitmap is found to be invalid, we call ext4_error() while the block group is still locked. This causes ext4_commit_super() to call a function which might sleep while in an atomic context. There's no need to keep the block group locked at this point, so hoist the ext4_error() call up to ext4_validate_block_bitmap() and release the block group spinlock before calling ext4_error(). The reported stack trace can be found at: http://article.gmane.org/gmane.comp.file-systems.ext4/33731 Reported-by: Dave Jones Signed-off-by: "Theodore Ts'o" Cc: stable@vger.kernel.org --- fs/ext4/balloc.c | 62 +++++++++++++++++++++++++++++++++----------------------- fs/ext4/bitmap.c | 1 - 2 files changed, 37 insertions(+), 26 deletions(-) diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c index d23b31ca9d7a..1b5089067d01 100644 --- a/fs/ext4/balloc.c +++ b/fs/ext4/balloc.c @@ -280,14 +280,18 @@ struct ext4_group_desc * ext4_get_group_desc(struct super_block *sb, return desc; } -static int ext4_valid_block_bitmap(struct super_block *sb, - struct ext4_group_desc *desc, - unsigned int block_group, - struct buffer_head *bh) +/* + * Return the block number which was discovered to be invalid, or 0 if + * the block bitmap is valid. + */ +static ext4_fsblk_t ext4_valid_block_bitmap(struct super_block *sb, + struct ext4_group_desc *desc, + unsigned int block_group, + struct buffer_head *bh) { ext4_grpblk_t offset; ext4_grpblk_t next_zero_bit; - ext4_fsblk_t bitmap_blk; + ext4_fsblk_t blk; ext4_fsblk_t group_first_block; if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_FLEX_BG)) { @@ -297,37 +301,33 @@ static int ext4_valid_block_bitmap(struct super_block *sb, * or it has to also read the block group where the bitmaps * are located to verify they are set. */ - return 1; + return 0; } group_first_block = ext4_group_first_block_no(sb, block_group); /* check whether block bitmap block number is set */ - bitmap_blk = ext4_block_bitmap(sb, desc); - offset = bitmap_blk - group_first_block; + blk = ext4_block_bitmap(sb, desc); + offset = blk - group_first_block; if (!ext4_test_bit(offset, bh->b_data)) /* bad block bitmap */ - goto err_out; + return blk; /* check whether the inode bitmap block number is set */ - bitmap_blk = ext4_inode_bitmap(sb, desc); - offset = bitmap_blk - group_first_block; + blk = ext4_inode_bitmap(sb, desc); + offset = blk - group_first_block; if (!ext4_test_bit(offset, bh->b_data)) /* bad block bitmap */ - goto err_out; + return blk; /* check whether the inode table block number is set */ - bitmap_blk = ext4_inode_table(sb, desc); - offset = bitmap_blk - group_first_block; + blk = ext4_inode_table(sb, desc); + offset = blk - group_first_block; next_zero_bit = ext4_find_next_zero_bit(bh->b_data, offset + EXT4_SB(sb)->s_itb_per_group, offset); - if (next_zero_bit >= offset + EXT4_SB(sb)->s_itb_per_group) - /* good bitmap for inode tables */ - return 1; - -err_out: - ext4_error(sb, "Invalid block bitmap - block_group = %d, block = %llu", - block_group, bitmap_blk); + if (next_zero_bit < offset + EXT4_SB(sb)->s_itb_per_group) + /* bad bitmap for inode tables */ + return blk; return 0; } @@ -336,14 +336,26 @@ void ext4_validate_block_bitmap(struct super_block *sb, unsigned int block_group, struct buffer_head *bh) { + ext4_fsblk_t blk; + if (buffer_verified(bh)) return; ext4_lock_group(sb, block_group); - if (ext4_valid_block_bitmap(sb, desc, block_group, bh) && - ext4_block_bitmap_csum_verify(sb, block_group, desc, bh, - EXT4_BLOCKS_PER_GROUP(sb) / 8)) - set_buffer_verified(bh); + blk = ext4_valid_block_bitmap(sb, desc, block_group, bh); + if (unlikely(blk != 0)) { + ext4_unlock_group(sb, block_group); + ext4_error(sb, "bg %u: block %llu: invalid block bitmap", + block_group, blk); + return; + } + if (unlikely(!ext4_block_bitmap_csum_verify(sb, block_group, + desc, bh, EXT4_BLOCKS_PER_GROUP(sb) / 8))) { + ext4_unlock_group(sb, block_group); + ext4_error(sb, "bg %u: bad block bitmap checksum", block_group); + return; + } + set_buffer_verified(bh); ext4_unlock_group(sb, block_group); } diff --git a/fs/ext4/bitmap.c b/fs/ext4/bitmap.c index f8716eab9995..5c2d1813ebe9 100644 --- a/fs/ext4/bitmap.c +++ b/fs/ext4/bitmap.c @@ -79,7 +79,6 @@ int ext4_block_bitmap_csum_verify(struct super_block *sb, ext4_group_t group, if (provided == calculated) return 1; - ext4_error(sb, "Bad block bitmap checksum: block_group = %u", group); return 0; } -- cgit v1.2.3-55-g7522 From 0548bbb85337e532ca2ed697c3e9b227ff2ed4b4 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Thu, 16 Aug 2012 11:59:04 -0400 Subject: ext4: fix long mount times on very big file systems Commit 8aeb00ff85a: "ext4: fix overhead calculation used by ext4_statfs()" introduced a O(n**2) calculation which makes very large file systems take forever to mount. Fix this with an optimization for non-bigalloc file systems. (For bigalloc file systems the overhead needs to be set in the the superblock.) Signed-off-by: "Theodore Ts'o" Cc: stable@vger.kernel.org --- fs/ext4/super.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 56bcaec9149c..598498904035 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -3120,6 +3120,10 @@ static int count_overhead(struct super_block *sb, ext4_group_t grp, ext4_group_t i, ngroups = ext4_get_groups_count(sb); int s, j, count = 0; + if (!EXT4_HAS_RO_COMPAT_FEATURE(sb, EXT4_FEATURE_RO_COMPAT_BIGALLOC)) + return (ext4_bg_has_super(sb, grp) + ext4_bg_num_gdb(sb, grp) + + sbi->s_itb_per_group + 2); + first_block = le32_to_cpu(sbi->s_es->s_first_data_block) + (grp * EXT4_BLOCKS_PER_GROUP(sb)); last_block = first_block + EXT4_BLOCKS_PER_GROUP(sb) - 1; -- cgit v1.2.3-55-g7522 From 89a4e48f8479f8145eca9698f39fe188c982212f Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Fri, 17 Aug 2012 08:54:52 -0400 Subject: ext4: fix kernel BUG on large-scale rm -rf commands Commit 968dee7722: "ext4: fix hole punch failure when depth is greater than 0" introduced a regression in v3.5.1/v3.6-rc1 which caused kernel crashes when users ran run "rm -rf" on large directory hierarchy on ext4 filesystems on RAID devices: BUG: unable to handle kernel NULL pointer dereference at 0000000000000028 Process rm (pid: 18229, threadinfo ffff8801276bc000, task ffff880123631710) Call Trace: [] ? __ext4_handle_dirty_metadata+0x83/0x110 [] ext4_ext_truncate+0x193/0x1d0 [] ? ext4_mark_inode_dirty+0x7f/0x1f0 [] ext4_truncate+0xf5/0x100 [] ext4_evict_inode+0x461/0x490 [] evict+0xa2/0x1a0 [] iput+0x103/0x1f0 [] do_unlinkat+0x154/0x1c0 [] ? sys_newfstatat+0x2a/0x40 [] sys_unlinkat+0x1b/0x50 [] system_call_fastpath+0x16/0x1b Code: 8b 4d 20 0f b7 41 02 48 8d 04 40 48 8d 04 81 49 89 45 18 0f b7 49 02 48 83 c1 01 49 89 4d 00 e9 ae f8 ff ff 0f 1f 00 49 8b 45 28 <48> 8b 40 28 49 89 45 20 e9 85 f8 ff ff 0f 1f 80 00 00 00 RIP [] ext4_ext_remove_space+0xa34/0xdf0 This could be reproduced as follows: The problem in commit 968dee7722 was that caused the variable 'i' to be left uninitialized if the truncate required more space than was available in the journal. This resulted in the function ext4_ext_truncate_extend_restart() returning -EAGAIN, which caused ext4_ext_remove_space() to restart the truncate operation after starting a new jbd2 handle. Reported-by: Maciej Żenczykowski Reported-by: Marti Raudsepp Tested-by: Fengguang Wu Signed-off-by: "Theodore Ts'o" Cc: stable@vger.kernel.org --- fs/ext4/extents.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index cd0c7ed06772..aabbb3f53683 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -2662,6 +2662,7 @@ cont: } path[0].p_depth = depth; path[0].p_hdr = ext_inode_hdr(inode); + i = 0; if (ext4_ext_check(inode, path[0].p_hdr, depth)) { err = -EIO; -- cgit v1.2.3-55-g7522