From 50f689af019b19f9b9a39be782c21b6f52b1615a Mon Sep 17 00:00:00 2001 From: Zhu Yanhai Date: Mon, 4 Apr 2011 12:58:12 -0400 Subject: jbd2: move bdget out of critical section bdget() should not be called when we hold spinlocks since it might sleep. Reviewed-by: Jan Kara Signed-off-by: Zhu Yanhai Signed-off-by: "Theodore Ts'o" --- fs/jbd2/journal.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index 90407b8fece..33dd3ef0ccd 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -2413,10 +2413,12 @@ const char *jbd2_dev_to_name(dev_t device) new_dev = kmalloc(sizeof(struct devname_cache), GFP_KERNEL); if (!new_dev) return "NODEV-ALLOCFAILURE"; /* Something non-NULL */ + bd = bdget(device); spin_lock(&devname_cache_lock); if (devcache[i]) { if (devcache[i]->device == device) { kfree(new_dev); + bdput(bd); ret = devcache[i]->devname; spin_unlock(&devname_cache_lock); return ret; @@ -2425,7 +2427,6 @@ const char *jbd2_dev_to_name(dev_t device) } devcache[i] = new_dev; devcache[i]->device = device; - bd = bdget(device); if (bd) { bdevname(bd, devcache[i]->devname); bdput(bd); -- cgit v1.2.3-18-g5258 From 21f976975cbecbdaf23ceeacc1cab2b1c05a028e Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 4 Apr 2011 15:33:39 -0400 Subject: ext4: remove unnecessary [cm]time update of quota file It is not necessary to update [cm]time of quota file on each quota file write and it wastes journal space and IO throughput with inode writes. So just remove the updating from ext4_quota_write() and only update times when quotas are being turned off. Userspace cannot get anything reliable from quota files while they are used by the kernel anyway. Signed-off-by: Jan Kara Signed-off-by: "Theodore Ts'o" --- fs/ext4/ext4_jbd2.h | 4 ++-- fs/ext4/super.c | 16 ++++++++++++++-- 2 files changed, 16 insertions(+), 4 deletions(-) diff --git a/fs/ext4/ext4_jbd2.h b/fs/ext4/ext4_jbd2.h index e25e99bf7ee..d0f53538a57 100644 --- a/fs/ext4/ext4_jbd2.h +++ b/fs/ext4/ext4_jbd2.h @@ -86,8 +86,8 @@ #ifdef CONFIG_QUOTA /* Amount of blocks needed for quota update - we know that the structure was - * allocated so we need to update only inode+data */ -#define EXT4_QUOTA_TRANS_BLOCKS(sb) (test_opt(sb, QUOTA) ? 2 : 0) + * allocated so we need to update only data block */ +#define EXT4_QUOTA_TRANS_BLOCKS(sb) (test_opt(sb, QUOTA) ? 1 : 0) /* Amount of blocks needed for quota insert/delete - we do some block writes * but inode, sb and group updates are done only once */ #define EXT4_QUOTA_INIT_BLOCKS(sb) (test_opt(sb, QUOTA) ? (DQUOT_INIT_ALLOC*\ diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 22546ad7f0a..35ff9fef68b 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -4614,11 +4614,24 @@ static int ext4_quota_on(struct super_block *sb, int type, int format_id, static int ext4_quota_off(struct super_block *sb, int type) { + struct inode *inode = sb_dqopt(sb)->files[type]; + handle_t *handle; + /* Force all delayed allocation blocks to be allocated. * Caller already holds s_umount sem */ if (test_opt(sb, DELALLOC)) sync_filesystem(sb); + /* Update modification times of quota files when userspace can + * start looking at them */ + handle = ext4_journal_start(inode, 1); + if (IS_ERR(handle)) + goto out; + inode->i_mtime = inode->i_ctime = CURRENT_TIME; + ext4_mark_inode_dirty(handle, inode); + ext4_journal_stop(handle); + +out: return dquot_quota_off(sb, type); } @@ -4714,9 +4727,8 @@ out: if (inode->i_size < off + len) { i_size_write(inode, off + len); EXT4_I(inode)->i_disksize = inode->i_size; + ext4_mark_inode_dirty(handle, inode); } - inode->i_mtime = inode->i_ctime = CURRENT_TIME; - ext4_mark_inode_dirty(handle, inode); mutex_unlock(&inode->i_mutex); return len; } -- cgit v1.2.3-18-g5258 From 5b41395fcc0265fc9f193aef9df39ce49d64677c Mon Sep 17 00:00:00 2001 From: Yongqiang Yang Date: Mon, 4 Apr 2011 15:40:24 -0400 Subject: ext4: fix credits computing for indirect mapped files When writing a contiguous set of blocks, two indirect blocks could be needed depending on how the blocks are aligned, so we need to increase the number of credits needed by one. [ Also fixed a another bug which could further underestimate the number of journal credits needed by 1; the code was using integer division instead of DIV_ROUND_UP() -- tytso] Signed-off-by: Yongqiang Yang Signed-off-by: "Theodore Ts'o" Cc: stable@kernel.org --- fs/ext4/inode.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 1a86282b902..7d11e02ad01 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -5398,13 +5398,12 @@ static int ext4_indirect_trans_blocks(struct inode *inode, int nrblocks, /* if nrblocks are contiguous */ if (chunk) { /* - * With N contiguous data blocks, it need at most - * N/EXT4_ADDR_PER_BLOCK(inode->i_sb) indirect blocks - * 2 dindirect blocks - * 1 tindirect block + * With N contiguous data blocks, we need at most + * N/EXT4_ADDR_PER_BLOCK(inode->i_sb) + 1 indirect blocks, + * 2 dindirect blocks, and 1 tindirect block */ - indirects = nrblocks / EXT4_ADDR_PER_BLOCK(inode->i_sb); - return indirects + 3; + return DIV_ROUND_UP(nrblocks, + EXT4_ADDR_PER_BLOCK(inode->i_sb)) + 4; } /* * if nrblocks are not contiguous, worse case, each block touch -- cgit v1.2.3-18-g5258 From 46e4690bbd9a4f8d9e7c4f34e34b48f703ad47e0 Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Mon, 4 Apr 2011 16:00:49 -0400 Subject: ext4: fix a double free in ext4_register_li_request In ext4_register_li_request, we malloc a ext4_li_request and inserts it into ext4_li_info->li_request_list. In case of any error later, we free it in the end. But if we have some error in ext4_run_lazyinit_thread, the whole li_request_list will be dropped and freed in it. So we will double free this ext4_li_request. This patch just sets elr to NULL after it is inserted to the list so that the latter kfree won't double free it. Signed-off-by: Tao Ma Reviewed-by: Lukas Czerner Signed-off-by: "Theodore Ts'o" Cc: stable@kernel.org --- fs/ext4/super.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 35ff9fef68b..35bd0206bbb 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -2975,6 +2975,12 @@ static int ext4_register_li_request(struct super_block *sb, mutex_unlock(&ext4_li_info->li_list_mtx); sbi->s_li_request = elr; + /* + * set elr to NULL here since it has been inserted to + * the request_list and the removal and free of it is + * handled by ext4_clear_request_list from now on. + */ + elr = NULL; if (!(ext4_li_info->li_state & EXT4_LAZYINIT_RUNNING)) { ret = ext4_run_lazyinit_thread(); -- cgit v1.2.3-18-g5258 From 6cba611e600ded15f642552ce6b5f7ee243bacf0 Mon Sep 17 00:00:00 2001 From: Zhang Huan Date: Tue, 5 Apr 2011 19:16:20 -0400 Subject: jbd2: fix potential memory leak on transaction commit There is potential memory leak of journal head in function jbd2_journal_commit_transaction. The problem is that JBD2 will not reclaim the journal head of commit record if error occurs or journal is abotred. I use the following script to reproduce this issue, on a RHEL6 system. I found it very easy to reproduce with async commit enabled. mount /dev/sdb /mnt -o journal_checksum,journal_async_commit touch /mnt/xxx echo offline > /sys/block/sdb/device/state sync umount /mnt rmmod ext4 rmmod jbd2 Removal of the jbd2 module will make slab complaining that "cache `jbd2_journal_head': can't free all objects". Signed-off-by: Zhang Huan Signed-off-by: "Theodore Ts'o" --- fs/jbd2/commit.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index fa36d7662b2..b98e4c1eecf 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -105,6 +105,8 @@ static int journal_submit_commit_record(journal_t *journal, int ret; struct timespec now = current_kernel_time(); + *cbh = NULL; + if (is_journal_aborted(journal)) return 0; @@ -806,7 +808,7 @@ wait_for_iobuf: if (err) __jbd2_journal_abort_hard(journal); } - if (!err && !is_journal_aborted(journal)) + if (cbh) err = journal_wait_on_commit_record(journal, cbh); if (JBD2_HAS_INCOMPAT_FEATURE(journal, JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT) && -- cgit v1.2.3-18-g5258 From 0449641130f5652b344ef6fa39fa019d7e94660a Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Tue, 5 Apr 2011 19:55:28 -0400 Subject: ext4: init timer earlier to avoid a kernel panic in __save_error_info During mount, when we fail to open journal inode or root inode, the __save_error_info will mod_timer. But actually s_err_report isn't initialized yet and the kernel oops. The detailed information can be found https://bugzilla.kernel.org/show_bug.cgi?id=32082. The best way is to check whether the timer s_err_report is initialized or not. But it seems that in include/linux/timer.h, we can't find a good function to check the status of this timer, so this patch just move the initializtion of s_err_report earlier so that we can avoid the kernel panic. The corresponding del_timer is also added in the error path. Reported-by: Sami Liedes Signed-off-by: Tao Ma Signed-off-by: "Theodore Ts'o" --- fs/ext4/super.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 35bd0206bbb..551cb8e2110 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -3391,6 +3391,10 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) get_random_bytes(&sbi->s_next_generation, sizeof(u32)); spin_lock_init(&sbi->s_next_gen_lock); + init_timer(&sbi->s_err_report); + sbi->s_err_report.function = print_daily_error_info; + sbi->s_err_report.data = (unsigned long) sb; + err = percpu_counter_init(&sbi->s_freeblocks_counter, ext4_count_free_blocks(sb)); if (!err) { @@ -3652,9 +3656,6 @@ no_journal: "Opts: %s%s%s", descr, sbi->s_es->s_mount_opts, *sbi->s_es->s_mount_opts ? "; " : "", orig_data); - init_timer(&sbi->s_err_report); - sbi->s_err_report.function = print_daily_error_info; - sbi->s_err_report.data = (unsigned long) sb; if (es->s_error_count) mod_timer(&sbi->s_err_report, jiffies + 300*HZ); /* 5 minutes */ @@ -3678,6 +3679,7 @@ failed_mount_wq: sbi->s_journal = NULL; } failed_mount3: + del_timer(&sbi->s_err_report); if (sbi->s_flex_groups) { if (is_vmalloc_addr(sbi->s_flex_groups)) vfree(sbi->s_flex_groups); -- cgit v1.2.3-18-g5258 From 0893ed458b4b1d7c7667ca7ffb8b11febe7e7e6c Mon Sep 17 00:00:00 2001 From: Curt Wohlgemuth Date: Sun, 10 Apr 2011 22:05:31 -0400 Subject: ext4: sync the directory inode in ext4_sync_parent() ext4 has taken the stance that, in the absence of a journal, when an fsync/fdatasync of an inode is done, the parent directory should be sync'ed if this inode entry is new. ext4_sync_parent(), which implements this, does indeed sync the dirent pages for parent directories, but it does not sync the directory *inode*. This patch fixes this. Also now return error status from ext4_sync_parent(). I tested this using a power fail test, which panics a machine running a file server getting requests from a client. Without this patch, on about every other test run, the server is missing many, many files that had been synced. With this patch, on > 6 runs, I see zero files being lost. Google-Bug-Id: 4179519 Signed-off-by: Curt Wohlgemuth Signed-off-by: "Theodore Ts'o" --- fs/ext4/fsync.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) diff --git a/fs/ext4/fsync.c b/fs/ext4/fsync.c index 7f74019d6d7..b1f9b5fc93e 100644 --- a/fs/ext4/fsync.c +++ b/fs/ext4/fsync.c @@ -125,9 +125,11 @@ extern int ext4_flush_completed_IO(struct inode *inode) * the parent directory's parent as well, and so on recursively, if * they are also freshly created. */ -static void ext4_sync_parent(struct inode *inode) +static int ext4_sync_parent(struct inode *inode) { + struct writeback_control wbc; struct dentry *dentry = NULL; + int ret = 0; while (inode && ext4_test_inode_state(inode, EXT4_STATE_NEWENTRY)) { ext4_clear_inode_state(inode, EXT4_STATE_NEWENTRY); @@ -136,8 +138,17 @@ static void ext4_sync_parent(struct inode *inode) if (!dentry || !dentry->d_parent || !dentry->d_parent->d_inode) break; inode = dentry->d_parent->d_inode; - sync_mapping_buffers(inode->i_mapping); + ret = sync_mapping_buffers(inode->i_mapping); + if (ret) + break; + memset(&wbc, 0, sizeof(wbc)); + wbc.sync_mode = WB_SYNC_ALL; + wbc.nr_to_write = 0; /* only write out the inode */ + ret = sync_inode(inode, &wbc); + if (ret) + break; } + return ret; } /* @@ -176,7 +187,7 @@ int ext4_sync_file(struct file *file, int datasync) if (!journal) { ret = generic_file_fsync(file, datasync); if (!ret && !list_empty(&inode->i_dentry)) - ext4_sync_parent(inode); + ret = ext4_sync_parent(inode); goto out; } -- cgit v1.2.3-18-g5258 From be4f27d324e8ddd57cc0d4d604fe85ee0425cba9 Mon Sep 17 00:00:00 2001 From: Yongqiang Yang Date: Sun, 10 Apr 2011 22:06:07 -0400 Subject: ext4: allow an active handle to be started when freezing ext4_journal_start_sb() should not prevent an active handle from being started due to s_frozen. Otherwise, deadlock is easy to happen, below is a situation. ================================================ freeze | truncate ================================================ | ext4_ext_truncate() freeze_super() | starts a handle sets s_frozen | | ext4_ext_truncate() | holds i_data_sem ext4_freeze() | waits for updates | | ext4_free_blocks() | calls dquot_free_block() | | dquot_free_blocks() | calls ext4_dirty_inode() | | ext4_dirty_inode() | trys to start an active | handle | | block due to s_frozen ================================================ Signed-off-by: Yongqiang Yang Signed-off-by: "Theodore Ts'o" Reported-by: Amir Goldstein Reviewed-by: Jan Kara Reviewed-by: Andreas Dilger --- fs/ext4/super.c | 44 +++++++++++++++++++++++++++++++++----------- 1 file changed, 33 insertions(+), 11 deletions(-) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 551cb8e2110..7b636ceb878 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -242,27 +242,44 @@ static void ext4_put_nojournal(handle_t *handle) * journal_end calls result in the superblock being marked dirty, so * that sync() will call the filesystem's write_super callback if * appropriate. + * + * To avoid j_barrier hold in userspace when a user calls freeze(), + * ext4 prevents a new handle from being started by s_frozen, which + * is in an upper layer. */ handle_t *ext4_journal_start_sb(struct super_block *sb, int nblocks) { journal_t *journal; + handle_t *handle; if (sb->s_flags & MS_RDONLY) return ERR_PTR(-EROFS); - vfs_check_frozen(sb, SB_FREEZE_TRANS); - /* Special case here: if the journal has aborted behind our - * backs (eg. EIO in the commit thread), then we still need to - * take the FS itself readonly cleanly. */ journal = EXT4_SB(sb)->s_journal; - if (journal) { - if (is_journal_aborted(journal)) { - ext4_abort(sb, "Detected aborted journal"); - return ERR_PTR(-EROFS); - } - return jbd2_journal_start(journal, nblocks); + handle = ext4_journal_current_handle(); + + /* + * If a handle has been started, it should be allowed to + * finish, otherwise deadlock could happen between freeze + * and others(e.g. truncate) due to the restart of the + * journal handle if the filesystem is forzen and active + * handles are not stopped. + */ + if (!handle) + vfs_check_frozen(sb, SB_FREEZE_TRANS); + + if (!journal) + return ext4_get_nojournal(); + /* + * Special case here: if the journal has aborted behind our + * backs (eg. EIO in the commit thread), then we still need to + * take the FS itself readonly cleanly. + */ + if (is_journal_aborted(journal)) { + ext4_abort(sb, "Detected aborted journal"); + return ERR_PTR(-EROFS); } - return ext4_get_nojournal(); + return jbd2_journal_start(journal, nblocks); } /* @@ -4146,6 +4163,11 @@ static int ext4_sync_fs(struct super_block *sb, int wait) /* * LVM calls this function before a (read-only) snapshot is created. This * gives us a chance to flush the journal completely and mark the fs clean. + * + * Note that only this function cannot bring a filesystem to be in a clean + * state independently, because ext4 prevents a new handle from being started + * by @sb->s_frozen, which stays in an upper layer. It thus needs help from + * the upper layer. */ static int ext4_freeze(struct super_block *sb) { -- cgit v1.2.3-18-g5258 From f80da1e70f1ffec3825aa0a1d0801f4896e002b6 Mon Sep 17 00:00:00 2001 From: Kazuya Mio Date: Sun, 10 Apr 2011 22:06:36 -0400 Subject: ext4: Allow indirect-block file to grow the file size to max file size We can create 4402345721856 byte file with indirect block mapping. However, if we grow an indirect-block file to the size with ftruncate(), we can see an ext4 warning. The following patch fixes this problem. How to reproduce: # dd if=/dev/zero of=/mnt/mp1/hoge bs=1 count=0 seek=4402345721856 0+0 records in 0+0 records out 0 bytes (0 B) copied, 0.000221428 s, 0.0 kB/s # tail -n 1 /var/log/messages Nov 25 15:10:27 test kernel: EXT4-fs warning (device sda8): ext4_block_to_path:345: block 1074791436 > max in inode 12 Signed-off-by: Kazuya Mio Signed-off-by: "Theodore Ts'o" --- fs/ext4/inode.c | 23 +++++++++++++++++------ 1 file changed, 17 insertions(+), 6 deletions(-) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 7d11e02ad01..5560f78690a 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -4429,8 +4429,8 @@ void ext4_truncate(struct inode *inode) Indirect chain[4]; Indirect *partial; __le32 nr = 0; - int n; - ext4_lblk_t last_block; + int n = 0; + ext4_lblk_t last_block, max_block; unsigned blocksize = inode->i_sb->s_blocksize; trace_ext4_truncate_enter(inode); @@ -4455,14 +4455,18 @@ void ext4_truncate(struct inode *inode) last_block = (inode->i_size + blocksize-1) >> EXT4_BLOCK_SIZE_BITS(inode->i_sb); + max_block = (EXT4_SB(inode->i_sb)->s_bitmap_maxbytes + blocksize-1) + >> EXT4_BLOCK_SIZE_BITS(inode->i_sb); if (inode->i_size & (blocksize - 1)) if (ext4_block_truncate_page(handle, mapping, inode->i_size)) goto out_stop; - n = ext4_block_to_path(inode, last_block, offsets, NULL); - if (n == 0) - goto out_stop; /* error */ + if (last_block != max_block) { + n = ext4_block_to_path(inode, last_block, offsets, NULL); + if (n == 0) + goto out_stop; /* error */ + } /* * OK. This truncate is going to happen. We add the inode to the @@ -4493,7 +4497,13 @@ void ext4_truncate(struct inode *inode) */ ei->i_disksize = inode->i_size; - if (n == 1) { /* direct blocks */ + if (last_block == max_block) { + /* + * It is unnecessary to free any data blocks if last_block is + * equal to the indirect block limit. + */ + goto out_unlock; + } else if (n == 1) { /* direct blocks */ ext4_free_data(handle, inode, NULL, i_data+offsets[0], i_data + EXT4_NDIR_BLOCKS); goto do_indirects; @@ -4553,6 +4563,7 @@ do_indirects: ; } +out_unlock: up_write(&ei->i_data_sem); inode->i_mtime = inode->i_ctime = ext4_current_time(inode); ext4_mark_inode_dirty(handle, inode); -- cgit v1.2.3-18-g5258 From c8205636029fc869278c55b7336053b3e7ae3ef4 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Sun, 10 Apr 2011 22:30:07 -0400 Subject: ext4: fix data corruption regression by reverting commit 6de9843dab3f Revert commit 6de9843dab3f2a1d4d66d80aa9e5782f80977d20, since it caused a data corruption regression with BitTorrent downloads. Thanks to Damien for discovering and bisecting to find the problem commit. https://bugzilla.kernel.org/show_bug.cgi?id=32972 Reported-by: Damien Grassart Signed-off-by: "Theodore Ts'o" --- fs/ext4/inode.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 5560f78690a..9c8cf811d93 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2502,6 +2502,7 @@ static int ext4_da_get_block_prep(struct inode *inode, sector_t iblock, * for partial write. */ set_buffer_new(bh); + set_buffer_mapped(bh); } return 0; } -- cgit v1.2.3-18-g5258