From fa30dde38aa8628c73a6dded7cb0bba38c27b576 Mon Sep 17 00:00:00 2001 From: Jiufei Xue Date: Thu, 14 Mar 2019 23:19:22 -0400 Subject: ext4: fix NULL pointer dereference while journal is aborted We see the following NULL pointer dereference while running xfstests generic/475: BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 PGD 8000000c84bad067 P4D 8000000c84bad067 PUD c84e62067 PMD 0 Oops: 0000 [#1] SMP PTI CPU: 7 PID: 9886 Comm: fsstress Kdump: loaded Not tainted 5.0.0-rc8 #10 RIP: 0010:ext4_do_update_inode+0x4ec/0x760 ... Call Trace: ? jbd2_journal_get_write_access+0x42/0x50 ? __ext4_journal_get_write_access+0x2c/0x70 ? ext4_truncate+0x186/0x3f0 ext4_mark_iloc_dirty+0x61/0x80 ext4_mark_inode_dirty+0x62/0x1b0 ext4_truncate+0x186/0x3f0 ? unmap_mapping_pages+0x56/0x100 ext4_setattr+0x817/0x8b0 notify_change+0x1df/0x430 do_truncate+0x5e/0x90 ? generic_permission+0x12b/0x1a0 This is triggered because the NULL pointer handle->h_transaction was dereferenced in function ext4_update_inode_fsync_trans(). I found that the h_transaction was set to NULL in jbd2__journal_restart but failed to attached to a new transaction while the journal is aborted. Fix this by checking the handle before updating the inode. Fixes: b436b9bef84d ("ext4: Wait for proper transaction commit on fsync") Signed-off-by: Jiufei Xue Signed-off-by: Theodore Ts'o Reviewed-by: Joseph Qi Cc: stable@kernel.org --- fs/ext4/ext4_jbd2.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext4/ext4_jbd2.h b/fs/ext4/ext4_jbd2.h index 15b6dd733780..df908ef79cce 100644 --- a/fs/ext4/ext4_jbd2.h +++ b/fs/ext4/ext4_jbd2.h @@ -384,7 +384,7 @@ static inline void ext4_update_inode_fsync_trans(handle_t *handle, { struct ext4_inode_info *ei = EXT4_I(inode); - if (ext4_handle_valid(handle)) { + if (ext4_handle_valid(handle) && !is_handle_aborted(handle)) { ei->i_sync_tid = handle->h_transaction->t_tid; if (datasync) ei->i_datasync_tid = handle->h_transaction->t_tid; -- cgit v1.2.3-58-ga151 From 372a03e01853f860560eade508794dd274e9b390 Mon Sep 17 00:00:00 2001 From: Lukas Czerner Date: Thu, 14 Mar 2019 23:20:25 -0400 Subject: ext4: fix data corruption caused by unaligned direct AIO Ext4 needs to serialize unaligned direct AIO because the zeroing of partial blocks of two competing unaligned AIOs can result in data corruption. However it decides not to serialize if the potentially unaligned aio is past i_size with the rationale that no pending writes are possible past i_size. Unfortunately if the i_size is not block aligned and the second unaligned write lands past i_size, but still into the same block, it has the potential of corrupting the previous unaligned write to the same block. This is (very simplified) reproducer from Frank // 41472 = (10 * 4096) + 512 // 37376 = 41472 - 4096 ftruncate(fd, 41472); io_prep_pwrite(iocbs[0], fd, buf[0], 4096, 37376); io_prep_pwrite(iocbs[1], fd, buf[1], 4096, 41472); io_submit(io_ctx, 1, &iocbs[1]); io_submit(io_ctx, 1, &iocbs[2]); io_getevents(io_ctx, 2, 2, events, NULL); Without this patch the 512B range from 40960 up to the start of the second unaligned write (41472) is going to be zeroed overwriting the data written by the first write. This is a data corruption. 00000000 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 * 00009200 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 * 0000a000 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 * 0000a200 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 With this patch the data corruption is avoided because we will recognize the unaligned_aio and wait for the unwritten extent conversion. 00000000 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 * 00009200 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 * 0000a200 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 * 0000b200 Reported-by: Frank Sorenson Signed-off-by: Lukas Czerner Signed-off-by: Theodore Ts'o Fixes: e9e3bcecf44c ("ext4: serialize unaligned asynchronous DIO") Cc: stable@vger.kernel.org --- fs/ext4/file.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext4/file.c b/fs/ext4/file.c index 69d65d49837b..98ec11f69cd4 100644 --- a/fs/ext4/file.c +++ b/fs/ext4/file.c @@ -125,7 +125,7 @@ ext4_unaligned_aio(struct inode *inode, struct iov_iter *from, loff_t pos) struct super_block *sb = inode->i_sb; int blockmask = sb->s_blocksize - 1; - if (pos >= i_size_read(inode)) + if (pos >= ALIGN(i_size_read(inode), sb->s_blocksize)) return 0; if ((pos | iov_iter_alignment(from)) & blockmask) -- cgit v1.2.3-58-ga151 From 1dc1097ff60e4105216da7cd0aa99032b039a994 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 14 Mar 2019 23:46:05 -0400 Subject: ext4: avoid panic during forced reboot When admin calls "reboot -f" - i.e., does a hard system reboot by directly calling reboot(2) - ext4 filesystem mounted with errors=panic can panic the system. This happens because the underlying device gets disabled without unmounting the filesystem and thus some syscall running in parallel to reboot(2) can result in the filesystem getting IO errors. This is somewhat surprising to the users so try improve the behavior by switching to errors=remount-ro behavior when the system is running reboot(2). Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o --- fs/ext4/super.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 6e4cac646345..1a5729d8f9ec 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -430,6 +430,12 @@ static void ext4_journal_commit_callback(journal_t *journal, transaction_t *txn) spin_unlock(&sbi->s_md_lock); } +static bool system_going_down(void) +{ + return system_state == SYSTEM_HALT || system_state == SYSTEM_POWER_OFF + || system_state == SYSTEM_RESTART; +} + /* Deal with the reporting of failure conditions on a filesystem such as * inconsistencies detected or read IO failures. * @@ -460,7 +466,12 @@ static void ext4_handle_error(struct super_block *sb) if (journal) jbd2_journal_abort(journal, -EIO); } - if (test_opt(sb, ERRORS_RO)) { + /* + * We force ERRORS_RO behavior when system is rebooting. Otherwise we + * could panic during 'reboot -f' as the underlying device got already + * disabled. + */ + if (test_opt(sb, ERRORS_RO) || system_going_down()) { ext4_msg(sb, KERN_CRIT, "Remounting filesystem read-only"); /* * Make sure updated value of ->s_mount_flags will be visible @@ -468,8 +479,7 @@ static void ext4_handle_error(struct super_block *sb) */ smp_wmb(); sb->s_flags |= SB_RDONLY; - } - if (test_opt(sb, ERRORS_PANIC)) { + } else if (test_opt(sb, ERRORS_PANIC)) { if (EXT4_SB(sb)->s_journal && !(EXT4_SB(sb)->s_journal->j_flags & JBD2_REC_ERR)) return; -- cgit v1.2.3-58-ga151 From 7cf77140777364d77b2b6e392e7e081a205a08c5 Mon Sep 17 00:00:00 2001 From: Jason Yan Date: Thu, 14 Mar 2019 23:51:13 -0400 Subject: ext4: remove useless ext4_pin_inode() This function is never used from the beginning (and is commented out); let's remove it. Signed-off-by: Jason Yan Signed-off-by: Theodore Ts'o --- fs/ext4/inode.c | 30 ------------------------------ 1 file changed, 30 deletions(-) (limited to 'fs') diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index f84cf62fd290..e5014a6f76e2 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -6082,36 +6082,6 @@ out: return; } -#if 0 -/* - * Bind an inode's backing buffer_head into this transaction, to prevent - * it from being flushed to disk early. Unlike - * ext4_reserve_inode_write, this leaves behind no bh reference and - * returns no iloc structure, so the caller needs to repeat the iloc - * lookup to mark the inode dirty later. - */ -static int ext4_pin_inode(handle_t *handle, struct inode *inode) -{ - struct ext4_iloc iloc; - - int err = 0; - if (handle) { - err = ext4_get_inode_loc(inode, &iloc); - if (!err) { - BUFFER_TRACE(iloc.bh, "get_write_access"); - err = jbd2_journal_get_write_access(handle, iloc.bh); - if (!err) - err = ext4_handle_dirty_metadata(handle, - NULL, - iloc.bh); - brelse(iloc.bh); - } - } - ext4_std_error(inode->i_sb, err); - return err; -} -#endif - int ext4_change_inode_journal_flag(struct inode *inode, int val) { journal_t *journal; -- cgit v1.2.3-58-ga151 From d64264d6218e6892edd832dc3a5a5857c2856c53 Mon Sep 17 00:00:00 2001 From: Lukas Czerner Date: Fri, 15 Mar 2019 00:15:32 -0400 Subject: ext4: add missing brelse() in add_new_gdb_meta_bg() Currently in add_new_gdb_meta_bg() there is a missing brelse of gdb_bh in case ext4_journal_get_write_access() fails. Additionally kvfree() is missing in the same error path. Fix it by moving the ext4_journal_get_write_access() before the ext4 sb update as Ted suggested and release n_group_desc and gdb_bh in case it fails. Fixes: 61a9c11e5e7a ("ext4: add missing brelse() add_new_gdb_meta_bg()'s error path") Signed-off-by: Lukas Czerner Signed-off-by: Theodore Ts'o --- fs/ext4/resize.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c index 3d9b18505c0c..90061c3d048b 100644 --- a/fs/ext4/resize.c +++ b/fs/ext4/resize.c @@ -932,11 +932,18 @@ static int add_new_gdb_meta_bg(struct super_block *sb, memcpy(n_group_desc, o_group_desc, EXT4_SB(sb)->s_gdb_count * sizeof(struct buffer_head *)); n_group_desc[gdb_num] = gdb_bh; + + BUFFER_TRACE(gdb_bh, "get_write_access"); + err = ext4_journal_get_write_access(handle, gdb_bh); + if (err) { + kvfree(n_group_desc); + brelse(gdb_bh); + return err; + } + EXT4_SB(sb)->s_group_desc = n_group_desc; EXT4_SB(sb)->s_gdb_count++; kvfree(o_group_desc); - BUFFER_TRACE(gdb_bh, "get_write_access"); - err = ext4_journal_get_write_access(handle, gdb_bh); return err; } -- cgit v1.2.3-58-ga151 From 6c7328400e0488f7d49e19e02290ba343b6811b2 Mon Sep 17 00:00:00 2001 From: Lukas Czerner Date: Fri, 15 Mar 2019 00:22:28 -0400 Subject: ext4: report real fs size after failed resize Currently when the file system resize using ext4_resize_fs() fails it will report into log that "resized filesystem to ". However this may not be true in the case of failure. Use the current block count as returned by ext4_blocks_count() to report the block count. Additionally, report a warning that "error occurred during file system resize" Signed-off-by: Lukas Czerner Signed-off-by: Theodore Ts'o --- fs/ext4/resize.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c index 90061c3d048b..e7ae26e36c9c 100644 --- a/fs/ext4/resize.c +++ b/fs/ext4/resize.c @@ -2080,6 +2080,10 @@ out: free_flex_gd(flex_gd); if (resize_inode != NULL) iput(resize_inode); - ext4_msg(sb, KERN_INFO, "resized filesystem to %llu", n_blocks_count); + if (err) + ext4_warning(sb, "error (%d) occurred during " + "file system resize", err); + ext4_msg(sb, KERN_INFO, "resized filesystem to %llu", + ext4_blocks_count(es)); return err; } -- cgit v1.2.3-58-ga151 From 674a2b27234d1b7afcb0a9162e81b2e53aeef217 Mon Sep 17 00:00:00 2001 From: "zhangyi (F)" Date: Sat, 23 Mar 2019 11:43:05 -0400 Subject: ext4: brelse all indirect buffer in ext4_ind_remove_space() All indirect buffers get by ext4_find_shared() should be released no mater the branch should be freed or not. But now, we forget to release the lower depth indirect buffers when removing space from the same higher depth indirect block. It will lead to buffer leak and futher more, it may lead to quota information corruption when using old quota, consider the following case. - Create and mount an empty ext4 filesystem without extent and quota features, - quotacheck and enable the user & group quota, - Create some files and write some data to them, and then punch hole to some files of them, it may trigger the buffer leak problem mentioned above. - Disable quota and run quotacheck again, it will create two new aquota files and write the checked quota information to them, which probably may reuse the freed indirect block(the buffer and page cache was not freed) as data block. - Enable quota again, it will invoke vfs_load_quota_inode()->invalidate_bdev() to try to clean unused buffers and pagecache. Unfortunately, because of the buffer of quota data block is still referenced, quota code cannot read the up to date quota info from the device and lead to quota information corruption. This problem can be reproduced by xfstests generic/231 on ext3 file system or ext4 file system without extent and quota features. This patch fix this problem by releasing the missing indirect buffers, in ext4_ind_remove_space(). Reported-by: Hulk Robot Signed-off-by: zhangyi (F) Signed-off-by: Theodore Ts'o Reviewed-by: Jan Kara Cc: stable@kernel.org --- fs/ext4/indirect.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/ext4/indirect.c b/fs/ext4/indirect.c index c2225f0d31b5..02c39524bda2 100644 --- a/fs/ext4/indirect.c +++ b/fs/ext4/indirect.c @@ -1390,10 +1390,14 @@ end_range: partial->p + 1, partial2->p, (chain+n-1) - partial); - BUFFER_TRACE(partial->bh, "call brelse"); - brelse(partial->bh); - BUFFER_TRACE(partial2->bh, "call brelse"); - brelse(partial2->bh); + while (partial > chain) { + BUFFER_TRACE(partial->bh, "call brelse"); + brelse(partial->bh); + } + while (partial2 > chain2) { + BUFFER_TRACE(partial2->bh, "call brelse"); + brelse(partial2->bh); + } return 0; } -- cgit v1.2.3-58-ga151 From 5e86bdda41534e17621d5a071b294943cae4376e Mon Sep 17 00:00:00 2001 From: "zhangyi (F)" Date: Sat, 23 Mar 2019 11:56:01 -0400 Subject: ext4: cleanup bh release code in ext4_ind_remove_space() Currently, we are releasing the indirect buffer where we are done with it in ext4_ind_remove_space(), so we can see the brelse() and BUFFER_TRACE() everywhere. It seems fragile and hard to read, and we may probably forget to release the buffer some day. This patch cleans up the code by putting of the code which releases the buffers to the end of the function. Signed-off-by: zhangyi (F) Signed-off-by: Theodore Ts'o Reviewed-by: Jan Kara --- fs/ext4/indirect.c | 47 ++++++++++++++++++++++------------------------- 1 file changed, 22 insertions(+), 25 deletions(-) (limited to 'fs') diff --git a/fs/ext4/indirect.c b/fs/ext4/indirect.c index 02c39524bda2..2024d3fa5504 100644 --- a/fs/ext4/indirect.c +++ b/fs/ext4/indirect.c @@ -1222,6 +1222,7 @@ int ext4_ind_remove_space(handle_t *handle, struct inode *inode, ext4_lblk_t offsets[4], offsets2[4]; Indirect chain[4], chain2[4]; Indirect *partial, *partial2; + Indirect *p = NULL, *p2 = NULL; ext4_lblk_t max_block; __le32 nr = 0, nr2 = 0; int n = 0, n2 = 0; @@ -1263,7 +1264,7 @@ int ext4_ind_remove_space(handle_t *handle, struct inode *inode, } - partial = ext4_find_shared(inode, n, offsets, chain, &nr); + partial = p = ext4_find_shared(inode, n, offsets, chain, &nr); if (nr) { if (partial == chain) { /* Shared branch grows from the inode */ @@ -1288,13 +1289,11 @@ int ext4_ind_remove_space(handle_t *handle, struct inode *inode, partial->p + 1, (__le32 *)partial->bh->b_data+addr_per_block, (chain+n-1) - partial); - BUFFER_TRACE(partial->bh, "call brelse"); - brelse(partial->bh); partial--; } end_range: - partial2 = ext4_find_shared(inode, n2, offsets2, chain2, &nr2); + partial2 = p2 = ext4_find_shared(inode, n2, offsets2, chain2, &nr2); if (nr2) { if (partial2 == chain2) { /* @@ -1324,16 +1323,14 @@ end_range: (__le32 *)partial2->bh->b_data, partial2->p, (chain2+n2-1) - partial2); - BUFFER_TRACE(partial2->bh, "call brelse"); - brelse(partial2->bh); partial2--; } goto do_indirects; } /* Punch happened within the same level (n == n2) */ - partial = ext4_find_shared(inode, n, offsets, chain, &nr); - partial2 = ext4_find_shared(inode, n2, offsets2, chain2, &nr2); + partial = p = ext4_find_shared(inode, n, offsets, chain, &nr); + partial2 = p2 = ext4_find_shared(inode, n2, offsets2, chain2, &nr2); /* Free top, but only if partial2 isn't its subtree. */ if (nr) { @@ -1390,15 +1387,7 @@ end_range: partial->p + 1, partial2->p, (chain+n-1) - partial); - while (partial > chain) { - BUFFER_TRACE(partial->bh, "call brelse"); - brelse(partial->bh); - } - while (partial2 > chain2) { - BUFFER_TRACE(partial2->bh, "call brelse"); - brelse(partial2->bh); - } - return 0; + goto cleanup; } /* @@ -1413,8 +1402,6 @@ end_range: partial->p + 1, (__le32 *)partial->bh->b_data+addr_per_block, (chain+n-1) - partial); - BUFFER_TRACE(partial->bh, "call brelse"); - brelse(partial->bh); partial--; } if (partial2 > chain2 && depth2 <= depth) { @@ -1422,11 +1409,21 @@ end_range: (__le32 *)partial2->bh->b_data, partial2->p, (chain2+n2-1) - partial2); - BUFFER_TRACE(partial2->bh, "call brelse"); - brelse(partial2->bh); partial2--; } } + +cleanup: + while (p && p > chain) { + BUFFER_TRACE(p->bh, "call brelse"); + brelse(p->bh); + p--; + } + while (p2 && p2 > chain2) { + BUFFER_TRACE(p2->bh, "call brelse"); + brelse(p2->bh); + p2--; + } return 0; do_indirects: @@ -1434,7 +1431,7 @@ do_indirects: switch (offsets[0]) { default: if (++n >= n2) - return 0; + break; nr = i_data[EXT4_IND_BLOCK]; if (nr) { ext4_free_branches(handle, inode, NULL, &nr, &nr+1, 1); @@ -1443,7 +1440,7 @@ do_indirects: /* fall through */ case EXT4_IND_BLOCK: if (++n >= n2) - return 0; + break; nr = i_data[EXT4_DIND_BLOCK]; if (nr) { ext4_free_branches(handle, inode, NULL, &nr, &nr+1, 2); @@ -1452,7 +1449,7 @@ do_indirects: /* fall through */ case EXT4_DIND_BLOCK: if (++n >= n2) - return 0; + break; nr = i_data[EXT4_TIND_BLOCK]; if (nr) { ext4_free_branches(handle, inode, NULL, &nr, &nr+1, 3); @@ -1462,5 +1459,5 @@ do_indirects: case EXT4_TIND_BLOCK: ; } - return 0; + goto cleanup; } -- cgit v1.2.3-58-ga151 From 18915b5873f07e5030e6fb108a050fa7c71c59fb Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Sat, 23 Mar 2019 12:10:29 -0400 Subject: ext4: prohibit fstrim in norecovery mode The ext4 fstrim implementation uses the block bitmaps to find free space that can be discarded. If we haven't replayed the journal, the bitmaps will be stale and we absolutely *cannot* use stale metadata to zap the underlying storage. Signed-off-by: Darrick J. Wong Signed-off-by: Theodore Ts'o --- fs/ext4/ioctl.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'fs') diff --git a/fs/ext4/ioctl.c b/fs/ext4/ioctl.c index eb8ca8d80885..73435444b159 100644 --- a/fs/ext4/ioctl.c +++ b/fs/ext4/ioctl.c @@ -1000,6 +1000,13 @@ resizefs_out: if (!blk_queue_discard(q)) return -EOPNOTSUPP; + /* + * We haven't replayed the journal, so we cannot use our + * block-bitmap-guided storage zapping commands. + */ + if (test_opt(sb, NOLOAD) && ext4_has_feature_journal(sb)) + return -EROFS; + if (copy_from_user(&range, (struct fstrim_range __user *)arg, sizeof(range))) return -EFAULT; -- cgit v1.2.3-58-ga151