From 736603ab297506f4396cb5af592004499950fcfd Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Fri, 11 Jul 2008 19:27:31 -0400 Subject: jbd2: Add commit time into the commit block Carlo Wood has demonstrated that it's possible to recover deleted files from the journal. Something that will make this easier is if we can put the time of the commit into commit block. Signed-off-by: "Theodore Ts'o" --- fs/jbd2/commit.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs/jbd2/commit.c') diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index a2ed72f7ceee..92b6ac3df8ab 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -112,6 +112,7 @@ static int journal_submit_commit_record(journal_t *journal, struct buffer_head *bh; int ret; int barrier_done = 0; + struct timespec now = current_kernel_time(); if (is_journal_aborted(journal)) return 0; @@ -126,6 +127,8 @@ static int journal_submit_commit_record(journal_t *journal, tmp->h_magic = cpu_to_be32(JBD2_MAGIC_NUMBER); tmp->h_blocktype = cpu_to_be32(JBD2_COMMIT_BLOCK); tmp->h_sequence = cpu_to_be32(commit_transaction->t_tid); + tmp->h_commit_sec = cpu_to_be64(now.tv_sec); + tmp->h_commit_nsec = cpu_to_be32(now.tv_nsec); if (JBD2_HAS_COMPAT_FEATURE(journal, JBD2_FEATURE_COMPAT_CHECKSUM)) { -- cgit v1.2.3 From c851ed540173736e60d48b53b91a16ea5c903896 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 11 Jul 2008 19:27:31 -0400 Subject: jbd2: Implement data=ordered mode handling via inodes This patch adds necessary framework into JBD2 to be able to track inodes with each transaction and write-out their dirty data during transaction commit time. This new ordered mode brings all sorts of advantages such as possibility to get rid of journal heads and buffer heads for data buffers in ordered mode, better ordering of writes on transaction commit, simplification of some JBD code, no more anonymous pages when truncate of data being committed happens. Also with this new ordered mode, delayed allocation on ordered mode is much simpler. Signed-off-by: Jan Kara --- fs/jbd2/commit.c | 90 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 90 insertions(+) (limited to 'fs/jbd2/commit.c') diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index 92b6ac3df8ab..3ca107b5c86b 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -355,6 +355,81 @@ write_out_data: journal_do_submit_data(wbuf, bufs); } +/* + * Submit all the data buffers of inode associated with the transaction to + * disk. + * + * We are in a committing transaction. Therefore no new inode can be added to + * our inode list. We use JI_COMMIT_RUNNING flag to protect inode we currently + * operate on from being released while we write out pages. + */ +static int journal_submit_inode_data_buffers(journal_t *journal, + transaction_t *commit_transaction) +{ + struct jbd2_inode *jinode; + int err, ret = 0; + struct address_space *mapping; + + spin_lock(&journal->j_list_lock); + list_for_each_entry(jinode, &commit_transaction->t_inode_list, i_list) { + mapping = jinode->i_vfs_inode->i_mapping; + jinode->i_flags |= JI_COMMIT_RUNNING; + spin_unlock(&journal->j_list_lock); + err = filemap_fdatawrite_range(mapping, 0, + i_size_read(jinode->i_vfs_inode)); + if (!ret) + ret = err; + spin_lock(&journal->j_list_lock); + J_ASSERT(jinode->i_transaction == commit_transaction); + jinode->i_flags &= ~JI_COMMIT_RUNNING; + wake_up_bit(&jinode->i_flags, __JI_COMMIT_RUNNING); + } + spin_unlock(&journal->j_list_lock); + return ret; +} + +/* + * Wait for data submitted for writeout, refile inodes to proper + * transaction if needed. + * + */ +static int journal_finish_inode_data_buffers(journal_t *journal, + transaction_t *commit_transaction) +{ + struct jbd2_inode *jinode, *next_i; + int err, ret = 0; + + /* For locking, see the comment in journal_submit_inode_data_buffers() */ + spin_lock(&journal->j_list_lock); + list_for_each_entry(jinode, &commit_transaction->t_inode_list, i_list) { + jinode->i_flags |= JI_COMMIT_RUNNING; + spin_unlock(&journal->j_list_lock); + err = filemap_fdatawait(jinode->i_vfs_inode->i_mapping); + if (!ret) + ret = err; + spin_lock(&journal->j_list_lock); + jinode->i_flags &= ~JI_COMMIT_RUNNING; + wake_up_bit(&jinode->i_flags, __JI_COMMIT_RUNNING); + } + + /* Now refile inode to proper lists */ + list_for_each_entry_safe(jinode, next_i, + &commit_transaction->t_inode_list, i_list) { + list_del(&jinode->i_list); + if (jinode->i_next_transaction) { + jinode->i_transaction = jinode->i_next_transaction; + jinode->i_next_transaction = NULL; + list_add(&jinode->i_list, + &jinode->i_transaction->t_inode_list); + } else { + jinode->i_transaction = NULL; + } + } + spin_unlock(&journal->j_list_lock); + + return ret; +} + static __u32 jbd2_checksum_data(__u32 crc32_sum, struct buffer_head *bh) { struct page *page = bh->b_page; @@ -529,6 +604,9 @@ void jbd2_journal_commit_transaction(journal_t *journal) */ err = 0; journal_submit_data_buffers(journal, commit_transaction); + err = journal_submit_inode_data_buffers(journal, commit_transaction); + if (err) + jbd2_journal_abort(journal, err); /* * Wait for all previously submitted IO to complete if commit @@ -760,6 +838,17 @@ start_journal_io: __jbd2_journal_abort_hard(journal); } + /* + * This is the right place to wait for data buffers both for ASYNC + * and !ASYNC commit. If commit is ASYNC, we need to wait only after + * the commit block went to disk (which happens above). If commit is + * SYNC, we need to wait for data buffers before we start writing + * commit block, which happens below in such setting. + */ + err = journal_finish_inode_data_buffers(journal, commit_transaction); + if (err) + jbd2_journal_abort(journal, err); + /* Lo and behold: we have just managed to send a transaction to the log. Before we can commit it, wait for the IO so far to complete. Control buffers being written are on the @@ -880,6 +969,7 @@ wait_for_iobuf: jbd_debug(3, "JBD: commit phase 7\n"); J_ASSERT(commit_transaction->t_sync_datalist == NULL); + J_ASSERT(list_empty(&commit_transaction->t_inode_list)); J_ASSERT(commit_transaction->t_buffers == NULL); J_ASSERT(commit_transaction->t_checkpoint_list == NULL); J_ASSERT(commit_transaction->t_iobuf_list == NULL); -- cgit v1.2.3 From 87c89c232c8f7b3820c33c3b9bc803e9358027da Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 11 Jul 2008 19:27:31 -0400 Subject: jbd2: Remove data=ordered mode support using jbd buffer heads Signed-off-by: Jan Kara --- fs/jbd2/commit.c | 221 ++----------------------------------------------------- 1 file changed, 8 insertions(+), 213 deletions(-) (limited to 'fs/jbd2/commit.c') diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index 3ca107b5c86b..483183d15ed5 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -37,8 +37,8 @@ static void journal_end_buffer_io_sync(struct buffer_head *bh, int uptodate) } /* - * When an ext3-ordered file is truncated, it is possible that many pages are - * not sucessfully freed, because they are attached to a committing transaction. + * When an ext4 file is truncated, it is possible that some pages are not + * successfully freed, because they are attached to a committing transaction. * After the transaction commits, these pages are left on the LRU, with no * ->mapping, and with attached buffers. These pages are trivially reclaimable * by the VM, but their apparent absence upsets the VM accounting, and it makes @@ -79,21 +79,6 @@ nope: __brelse(bh); } -/* - * Try to acquire jbd_lock_bh_state() against the buffer, when j_list_lock is - * held. For ranking reasons we must trylock. If we lose, schedule away and - * return 0. j_list_lock is dropped in this case. - */ -static int inverted_lock(journal_t *journal, struct buffer_head *bh) -{ - if (!jbd_trylock_bh_state(bh)) { - spin_unlock(&journal->j_list_lock); - schedule(); - return 0; - } - return 1; -} - /* * Done it all: now submit the commit record. We should have * cleaned up our previous buffers by now, so if we are in abort @@ -199,162 +184,6 @@ static int journal_wait_on_commit_record(struct buffer_head *bh) return ret; } -/* - * Wait for all submitted IO to complete. - */ -static int journal_wait_on_locked_list(journal_t *journal, - transaction_t *commit_transaction) -{ - int ret = 0; - struct journal_head *jh; - - while (commit_transaction->t_locked_list) { - struct buffer_head *bh; - - jh = commit_transaction->t_locked_list->b_tprev; - bh = jh2bh(jh); - get_bh(bh); - if (buffer_locked(bh)) { - spin_unlock(&journal->j_list_lock); - wait_on_buffer(bh); - if (unlikely(!buffer_uptodate(bh))) - ret = -EIO; - spin_lock(&journal->j_list_lock); - } - if (!inverted_lock(journal, bh)) { - put_bh(bh); - spin_lock(&journal->j_list_lock); - continue; - } - if (buffer_jbd(bh) && jh->b_jlist == BJ_Locked) { - __jbd2_journal_unfile_buffer(jh); - jbd_unlock_bh_state(bh); - jbd2_journal_remove_journal_head(bh); - put_bh(bh); - } else { - jbd_unlock_bh_state(bh); - } - put_bh(bh); - cond_resched_lock(&journal->j_list_lock); - } - return ret; - } - -static void journal_do_submit_data(struct buffer_head **wbuf, int bufs) -{ - int i; - - for (i = 0; i < bufs; i++) { - wbuf[i]->b_end_io = end_buffer_write_sync; - /* We use-up our safety reference in submit_bh() */ - submit_bh(WRITE, wbuf[i]); - } -} - -/* - * Submit all the data buffers to disk - */ -static void journal_submit_data_buffers(journal_t *journal, - transaction_t *commit_transaction) -{ - struct journal_head *jh; - struct buffer_head *bh; - int locked; - int bufs = 0; - struct buffer_head **wbuf = journal->j_wbuf; - - /* - * Whenever we unlock the journal and sleep, things can get added - * onto ->t_sync_datalist, so we have to keep looping back to - * write_out_data until we *know* that the list is empty. - * - * Cleanup any flushed data buffers from the data list. Even in - * abort mode, we want to flush this out as soon as possible. - */ -write_out_data: - cond_resched(); - spin_lock(&journal->j_list_lock); - - while (commit_transaction->t_sync_datalist) { - jh = commit_transaction->t_sync_datalist; - bh = jh2bh(jh); - locked = 0; - - /* Get reference just to make sure buffer does not disappear - * when we are forced to drop various locks */ - get_bh(bh); - /* If the buffer is dirty, we need to submit IO and hence - * we need the buffer lock. We try to lock the buffer without - * blocking. If we fail, we need to drop j_list_lock and do - * blocking lock_buffer(). - */ - if (buffer_dirty(bh)) { - if (test_set_buffer_locked(bh)) { - BUFFER_TRACE(bh, "needs blocking lock"); - spin_unlock(&journal->j_list_lock); - /* Write out all data to prevent deadlocks */ - journal_do_submit_data(wbuf, bufs); - bufs = 0; - lock_buffer(bh); - spin_lock(&journal->j_list_lock); - } - locked = 1; - } - /* We have to get bh_state lock. Again out of order, sigh. */ - if (!inverted_lock(journal, bh)) { - jbd_lock_bh_state(bh); - spin_lock(&journal->j_list_lock); - } - /* Someone already cleaned up the buffer? */ - if (!buffer_jbd(bh) - || jh->b_transaction != commit_transaction - || jh->b_jlist != BJ_SyncData) { - jbd_unlock_bh_state(bh); - if (locked) - unlock_buffer(bh); - BUFFER_TRACE(bh, "already cleaned up"); - put_bh(bh); - continue; - } - if (locked && test_clear_buffer_dirty(bh)) { - BUFFER_TRACE(bh, "needs writeout, adding to array"); - wbuf[bufs++] = bh; - __jbd2_journal_file_buffer(jh, commit_transaction, - BJ_Locked); - jbd_unlock_bh_state(bh); - if (bufs == journal->j_wbufsize) { - spin_unlock(&journal->j_list_lock); - journal_do_submit_data(wbuf, bufs); - bufs = 0; - goto write_out_data; - } - } else if (!locked && buffer_locked(bh)) { - __jbd2_journal_file_buffer(jh, commit_transaction, - BJ_Locked); - jbd_unlock_bh_state(bh); - put_bh(bh); - } else { - BUFFER_TRACE(bh, "writeout complete: unfile"); - __jbd2_journal_unfile_buffer(jh); - jbd_unlock_bh_state(bh); - if (locked) - unlock_buffer(bh); - jbd2_journal_remove_journal_head(bh); - /* Once for our safety reference, once for - * jbd2_journal_remove_journal_head() */ - put_bh(bh); - put_bh(bh); - } - - if (need_resched() || spin_needbreak(&journal->j_list_lock)) { - spin_unlock(&journal->j_list_lock); - goto write_out_data; - } - } - spin_unlock(&journal->j_list_lock); - journal_do_submit_data(wbuf, bufs); -} - /* * Submit all the data buffers of inode associated with the transaction to * disk. @@ -602,24 +431,7 @@ void jbd2_journal_commit_transaction(journal_t *journal) * Now start flushing things to disk, in the order they appear * on the transaction lists. Data blocks go first. */ - err = 0; - journal_submit_data_buffers(journal, commit_transaction); err = journal_submit_inode_data_buffers(journal, commit_transaction); - if (err) - jbd2_journal_abort(journal, err); - - /* - * Wait for all previously submitted IO to complete if commit - * record is to be written synchronously. - */ - spin_lock(&journal->j_list_lock); - if (!JBD2_HAS_INCOMPAT_FEATURE(journal, - JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT)) - err = journal_wait_on_locked_list(journal, - commit_transaction); - - spin_unlock(&journal->j_list_lock); - if (err) jbd2_journal_abort(journal, err); @@ -627,16 +439,6 @@ void jbd2_journal_commit_transaction(journal_t *journal) jbd_debug(3, "JBD: commit phase 2\n"); - /* - * If we found any dirty or locked buffers, then we should have - * looped back up to the write_out_data label. If there weren't - * any then journal_clean_data_list should have wiped the list - * clean by now, so check that it is in fact empty. - */ - J_ASSERT (commit_transaction->t_sync_datalist == NULL); - - jbd_debug (3, "JBD: commit phase 3\n"); - /* * Way to go: we have now written out all of the data for a * transaction! Now comes the tricky part: we need to write out @@ -655,6 +457,7 @@ void jbd2_journal_commit_transaction(journal_t *journal) J_ASSERT(commit_transaction->t_nr_buffers <= commit_transaction->t_outstanding_credits); + err = 0; descriptor = NULL; bufs = 0; while (commit_transaction->t_buffers) { @@ -829,13 +632,6 @@ start_journal_io: &cbh, crc32_sum); if (err) __jbd2_journal_abort_hard(journal); - - spin_lock(&journal->j_list_lock); - err = journal_wait_on_locked_list(journal, - commit_transaction); - spin_unlock(&journal->j_list_lock); - if (err) - __jbd2_journal_abort_hard(journal); } /* @@ -860,7 +656,7 @@ start_journal_io: so we incur less scheduling load. */ - jbd_debug(3, "JBD: commit phase 4\n"); + jbd_debug(3, "JBD: commit phase 3\n"); /* * akpm: these are BJ_IO, and j_list_lock is not needed. @@ -919,7 +715,7 @@ wait_for_iobuf: J_ASSERT (commit_transaction->t_shadow_list == NULL); - jbd_debug(3, "JBD: commit phase 5\n"); + jbd_debug(3, "JBD: commit phase 4\n"); /* Here we wait for the revoke record and descriptor record buffers */ wait_for_ctlbuf: @@ -946,7 +742,7 @@ wait_for_iobuf: /* AKPM: bforget here */ } - jbd_debug(3, "JBD: commit phase 6\n"); + jbd_debug(3, "JBD: commit phase 5\n"); if (!JBD2_HAS_INCOMPAT_FEATURE(journal, JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT)) { @@ -966,9 +762,8 @@ wait_for_iobuf: transaction can be removed from any checkpoint list it was on before. */ - jbd_debug(3, "JBD: commit phase 7\n"); + jbd_debug(3, "JBD: commit phase 6\n"); - J_ASSERT(commit_transaction->t_sync_datalist == NULL); J_ASSERT(list_empty(&commit_transaction->t_inode_list)); J_ASSERT(commit_transaction->t_buffers == NULL); J_ASSERT(commit_transaction->t_checkpoint_list == NULL); @@ -1090,7 +885,7 @@ restart_loop: /* Done with this transaction! */ - jbd_debug(3, "JBD: commit phase 8\n"); + jbd_debug(3, "JBD: commit phase 7\n"); J_ASSERT(commit_transaction->t_state == T_COMMIT); -- cgit v1.2.3 From cd1aac32923a9c8adcc0ae85e33c1ca0c5855838 Mon Sep 17 00:00:00 2001 From: "Aneesh Kumar K.V" Date: Fri, 11 Jul 2008 19:27:31 -0400 Subject: ext4: Add ordered mode support for delalloc This provides a new ordered mode implementation which gets rid of using buffer heads to enforce the ordering between metadata change with the related data chage. Instead, in the new ordering mode, it keeps track of all of the inodes touched by each transaction on a list, and when that transaction is committed, it flushes all of the dirty pages for those inodes. In addition, the new ordered mode reverses the lock ordering of the page lock and transaction lock, which provides easier support for delayed allocation. Signed-off-by: Aneesh Kumar K.V Signed-off-by: Mingming Cao Signed-off-by: "Theodore Ts'o" --- fs/jbd2/commit.c | 38 +++++++++++++++++++++++++++++++++----- 1 file changed, 33 insertions(+), 5 deletions(-) (limited to 'fs/jbd2/commit.c') diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index 483183d15ed5..f8b3be873226 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -22,6 +22,8 @@ #include #include #include +#include +#include /* * Default IO end handler for temporary BJ_IO buffer_heads. @@ -184,6 +186,27 @@ static int journal_wait_on_commit_record(struct buffer_head *bh) return ret; } +/* + * write the filemap data using writepage() address_space_operations. + * We don't do block allocation here even for delalloc. We don't + * use writepages() because with dealyed allocation we may be doing + * block allocation in writepages(). + */ +static int journal_submit_inode_data_buffers(struct address_space *mapping) +{ + int ret; + struct writeback_control wbc = { + .sync_mode = WB_SYNC_ALL, + .nr_to_write = mapping->nrpages * 2, + .range_start = 0, + .range_end = i_size_read(mapping->host), + .for_writepages = 1, + }; + + ret = generic_writepages(mapping, &wbc); + return ret; +} + /* * Submit all the data buffers of inode associated with the transaction to * disk. @@ -192,7 +215,7 @@ static int journal_wait_on_commit_record(struct buffer_head *bh) * our inode list. We use JI_COMMIT_RUNNING flag to protect inode we currently * operate on from being released while we write out pages. */ -static int journal_submit_inode_data_buffers(journal_t *journal, +static int journal_submit_data_buffers(journal_t *journal, transaction_t *commit_transaction) { struct jbd2_inode *jinode; @@ -204,8 +227,13 @@ static int journal_submit_inode_data_buffers(journal_t *journal, mapping = jinode->i_vfs_inode->i_mapping; jinode->i_flags |= JI_COMMIT_RUNNING; spin_unlock(&journal->j_list_lock); - err = filemap_fdatawrite_range(mapping, 0, - i_size_read(jinode->i_vfs_inode)); + /* + * submit the inode data buffers. We use writepage + * instead of writepages. Because writepages can do + * block allocation with delalloc. We need to write + * only allocated blocks here. + */ + err = journal_submit_inode_data_buffers(mapping); if (!ret) ret = err; spin_lock(&journal->j_list_lock); @@ -228,7 +256,7 @@ static int journal_finish_inode_data_buffers(journal_t *journal, struct jbd2_inode *jinode, *next_i; int err, ret = 0; - /* For locking, see the comment in journal_submit_inode_data_buffers() */ + /* For locking, see the comment in journal_submit_data_buffers() */ spin_lock(&journal->j_list_lock); list_for_each_entry(jinode, &commit_transaction->t_inode_list, i_list) { jinode->i_flags |= JI_COMMIT_RUNNING; @@ -431,7 +459,7 @@ void jbd2_journal_commit_transaction(journal_t *journal) * Now start flushing things to disk, in the order they appear * on the transaction lists. Data blocks go first. */ - err = journal_submit_inode_data_buffers(journal, commit_transaction); + err = journal_submit_data_buffers(journal, commit_transaction); if (err) jbd2_journal_abort(journal, err); -- cgit v1.2.3 From e9e34f4e8f42177c66754fec1edfd35e70c18f99 Mon Sep 17 00:00:00 2001 From: Hidehiro Kawai Date: Thu, 31 Jul 2008 22:26:04 -0400 Subject: jbd2: don't abort if flushing file data failed In ordered mode, the current jbd2 aborts the journal if a file data buffer has an error. But this behavior is unintended, and we found that it has been adopted accidentally. This patch undoes it and just calls printk() instead of aborting the journal. Unlike a similar patch for ext3/jbd, file data buffers are written via generic_writepages(). But we also need to set AS_EIO into their mappings because wait_on_page_writeback_range() clears AS_EIO before a user process sees it. Signed-off-by: Hidehiro Kawai Signed-off-by: "Theodore Ts'o" --- fs/jbd2/commit.c | 24 ++++++++++++++++++++---- 1 file changed, 20 insertions(+), 4 deletions(-) (limited to 'fs/jbd2/commit.c') diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index f8b3be873226..adf0395f318e 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -262,8 +262,18 @@ static int journal_finish_inode_data_buffers(journal_t *journal, jinode->i_flags |= JI_COMMIT_RUNNING; spin_unlock(&journal->j_list_lock); err = filemap_fdatawait(jinode->i_vfs_inode->i_mapping); - if (!ret) - ret = err; + if (err) { + /* + * Because AS_EIO is cleared by + * wait_on_page_writeback_range(), set it again so + * that user process can get -EIO from fsync(). + */ + set_bit(AS_EIO, + &jinode->i_vfs_inode->i_mapping->flags); + + if (!ret) + ret = err; + } spin_lock(&journal->j_list_lock); jinode->i_flags &= ~JI_COMMIT_RUNNING; wake_up_bit(&jinode->i_flags, __JI_COMMIT_RUNNING); @@ -670,8 +680,14 @@ start_journal_io: * commit block, which happens below in such setting. */ err = journal_finish_inode_data_buffers(journal, commit_transaction); - if (err) - jbd2_journal_abort(journal, err); + if (err) { + char b[BDEVNAME_SIZE]; + + printk(KERN_WARNING + "JBD2: Detected IO errors while flushing file data " + "on %s\n", bdevname(journal->j_fs_dev, b)); + err = 0; + } /* Lo and behold: we have just managed to send a transaction to the log. Before we can commit it, wait for the IO so far to -- cgit v1.2.3 From 529ae9aaa08378cfe2a4350bded76f32cc8ff0ce Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Sat, 2 Aug 2008 12:01:03 +0200 Subject: mm: rename page trylock Converting page lock to new locking bitops requires a change of page flag operation naming, so we might as well convert it to something nicer (!TestSetPageLocked_Lock => trylock_page, SetPageLocked => set_page_locked). This also facilitates lockdeping of page lock. Signed-off-by: Nick Piggin Acked-by: KOSAKI Motohiro Acked-by: Peter Zijlstra Acked-by: Andrew Morton Acked-by: Benjamin Herrenschmidt Signed-off-by: Linus Torvalds --- fs/jbd2/commit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/jbd2/commit.c') diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index adf0395f318e..f2ad061e95ec 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -67,7 +67,7 @@ static void release_buffer_page(struct buffer_head *bh) goto nope; /* OK, it's a truncated page */ - if (TestSetPageLocked(page)) + if (!trylock_page(page)) goto nope; page_cache_get(page); -- cgit v1.2.3 From 05496769e5da83ce22ed97345afd9c7b71d6bd24 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 16 Sep 2008 14:36:17 -0400 Subject: jbd2: clean up how the journal device name is printed Calculate the journal device name once and stash it away in the journal_s structure. This avoids needing to call bdevname() everywhere and reduces stack usage by not needing to allocate an on-stack buffer. In addition, we eliminate the '/' that can appear in device names (e.g. "cciss/c0d0p9" --- see kernel bugzilla #11321) that can cause problems when creating proc directory names, and include the inode number to support ocfs2 which creates multiple journals with different inode numbers. Signed-off-by: "Theodore Ts'o" --- fs/jbd2/commit.c | 11 +++-------- 1 file changed, 3 insertions(+), 8 deletions(-) (limited to 'fs/jbd2/commit.c') diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index f2ad061e95ec..b091e5378fe0 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -147,12 +147,9 @@ static int journal_submit_commit_record(journal_t *journal, * to remember if we sent a barrier request */ if (ret == -EOPNOTSUPP && barrier_done) { - char b[BDEVNAME_SIZE]; - printk(KERN_WARNING - "JBD: barrier-based sync failed on %s - " - "disabling barriers\n", - bdevname(journal->j_dev, b)); + "JBD: barrier-based sync failed on %s - " + "disabling barriers\n", journal->j_devname); spin_lock(&journal->j_state_lock); journal->j_flags &= ~JBD2_BARRIER; spin_unlock(&journal->j_state_lock); @@ -681,11 +678,9 @@ start_journal_io: */ err = journal_finish_inode_data_buffers(journal, commit_transaction); if (err) { - char b[BDEVNAME_SIZE]; - printk(KERN_WARNING "JBD2: Detected IO errors while flushing file data " - "on %s\n", bdevname(journal->j_fs_dev, b)); + "on %s\n", journal->j_devname); err = 0; } -- cgit v1.2.3 From ede86cc473defab74d778aeac14b19f43129d4d1 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Sun, 5 Oct 2008 20:50:06 -0400 Subject: ext4: Add debugging markers that can be used by systemtap This debugging markers are designed to debug problems such as the random filesystem latency problems reported by Arjan. Signed-off-by: "Theodore Ts'o" --- fs/jbd2/commit.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'fs/jbd2/commit.c') diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index b091e5378fe0..e91f051a9859 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -16,6 +16,7 @@ #include #include #include +#include #include #include #include @@ -368,6 +369,8 @@ void jbd2_journal_commit_transaction(journal_t *journal) commit_transaction = journal->j_running_transaction; J_ASSERT(commit_transaction->t_state == T_RUNNING); + trace_mark(jbd2_start_commit, "dev %s transaction %d", + journal->j_devname, commit_transaction->t_tid); jbd_debug(1, "JBD: starting commit of transaction %d\n", commit_transaction->t_tid); @@ -985,6 +988,9 @@ restart_loop: } spin_unlock(&journal->j_list_lock); + trace_mark(jbd2_end_commit, "dev %s transaction %d head %d", + journal->j_devname, commit_transaction->t_tid, + journal->j_tail_sequence); jbd_debug(1, "JBD: commit %d complete, head %d\n", journal->j_commit_sequence, journal->j_tail_sequence); -- cgit v1.2.3 From 45a90bfd90c1215bf824c0f705b409723f52361b Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Mon, 6 Oct 2008 12:04:02 -0400 Subject: jbd2: Fix buffer head leak when writing the commit block Also make sure the buffer heads are marked clean before submitting bh for writing. The previous code was marking the buffer head dirty, which would have forced an unneeded write (and seek) to the journal for no good reason. Signed-off-by: "Theodore Ts'o" --- fs/jbd2/commit.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'fs/jbd2/commit.c') diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index e91f051a9859..0d3814a35ed1 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -127,8 +127,7 @@ static int journal_submit_commit_record(journal_t *journal, JBUFFER_TRACE(descriptor, "submit commit block"); lock_buffer(bh); - get_bh(bh); - set_buffer_dirty(bh); + clear_buffer_dirty(bh); set_buffer_uptodate(bh); bh->b_end_io = journal_end_buffer_io_sync; @@ -158,7 +157,7 @@ static int journal_submit_commit_record(journal_t *journal, /* And try again, without the barrier */ lock_buffer(bh); set_buffer_uptodate(bh); - set_buffer_dirty(bh); + clear_buffer_dirty(bh); ret = submit_bh(WRITE, bh); } *cbh = bh; -- cgit v1.2.3 From 77e841de8abac4755cc83ca224fdf71418d65380 Mon Sep 17 00:00:00 2001 From: Hidehiro Kawai Date: Sun, 12 Oct 2008 16:39:16 -0400 Subject: jbd2: abort when failed to log metadata buffers If we failed to write metadata buffers to the journal space and succeeded to write the commit record, stale data can be written back to the filesystem as metadata in the recovery phase. To avoid this, when we failed to write out metadata buffers, abort the journal before writing the commit record. We can also avoid this kind of corruption by using the journal checksum feature because it can detect invalid metadata blocks in the journal and avoid them from being replayed. So we don't need to care about asynchronous commit record writeout with a checksum. Signed-off-by: Hidehiro Kawai Signed-off-by: Theodore Ts'o --- fs/jbd2/commit.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs/jbd2/commit.c') diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index 0d3814a35ed1..78e4da934121 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -783,6 +783,9 @@ wait_for_iobuf: /* AKPM: bforget here */ } + if (err) + jbd2_journal_abort(journal, err); + jbd_debug(3, "JBD: commit phase 5\n"); if (!JBD2_HAS_INCOMPAT_FEATURE(journal, -- cgit v1.2.3 From 7ad7445f60fe4d46c4c9d2a9463db180d2a3b270 Mon Sep 17 00:00:00 2001 From: Hidehiro Kawai Date: Fri, 10 Oct 2008 20:29:31 -0400 Subject: jbd2: don't dirty original metadata buffer on abort Currently, original metadata buffers are dirtied when they are unfiled whether the journal has aborted or not. Eventually these buffers will be written-back to the filesystem by pdflush. This means some metadata buffers are written to the filesystem without journaling if the journal aborts. So if both journal abort and system crash happen at the same time, the filesystem would become inconsistent state. Additionally, replaying journaled metadata can overwrite the latest metadata on the filesystem partly. Because, if the journal gets aborted, journaled metadata are preserved and replayed during the next mount not to lose uncheckpointed metadata. This would also break the consistency of the filesystem. This patch prevents original metadata buffers from being dirtied on abort by clearing BH_JBDDirty flag from those buffers. Thus, no metadata buffers are written to the filesystem without journaling. Signed-off-by: Hidehiro Kawai Signed-off-by: Theodore Ts'o --- fs/jbd2/commit.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs/jbd2/commit.c') diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index 78e4da934121..849f10496cea 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -504,9 +504,10 @@ void jbd2_journal_commit_transaction(journal_t *journal) jh = commit_transaction->t_buffers; /* If we're in abort mode, we just un-journal the buffer and - release it for background writing. */ + release it. */ if (is_journal_aborted(journal)) { + clear_buffer_jbddirty(jh2bh(jh)); JBUFFER_TRACE(jh, "journal is aborting: refile"); jbd2_journal_refile_buffer(journal, jh); /* If that was the last one, we need to clean up @@ -884,6 +885,8 @@ restart_loop: if (buffer_jbddirty(bh)) { JBUFFER_TRACE(jh, "add to new checkpointing trans"); __jbd2_journal_insert_checkpoint(jh, commit_transaction); + if (is_journal_aborted(journal)) + clear_buffer_jbddirty(bh); JBUFFER_TRACE(jh, "refile for checkpoint writeback"); __jbd2_journal_refile_buffer(jh); jbd_unlock_bh_state(bh); -- cgit v1.2.3 From 5bf5683a33f3584da6eced480967c4f7e11515a8 Mon Sep 17 00:00:00 2001 From: Hidehiro Kawai Date: Fri, 10 Oct 2008 22:12:43 -0400 Subject: ext4: add an option to control error handling on file data If the journal doesn't abort when it gets an IO error in file data blocks, the file data corruption will spread silently. Because most of applications and commands do buffered writes without fsync(), they don't notice the IO error. It's scary for mission critical systems. On the other hand, if the journal aborts whenever it gets an IO error in file data blocks, the system will easily become inoperable. So this patch introduces a filesystem option to determine whether it aborts the journal or just call printk() when it gets an IO error in file data. If you mount an ext4 fs with data_err=abort option, it aborts on file data write error. If you mount it with data_err=ignore, it doesn't abort, just call printk(). data_err=ignore is the default. Here is the corresponding patch of the ext3 version: http://kerneltrap.org/mailarchive/linux-kernel/2008/9/9/3239374 Signed-off-by: Hidehiro Kawai Signed-off-by: Theodore Ts'o --- fs/jbd2/commit.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs/jbd2/commit.c') diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index 849f10496cea..0abe02c4242a 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -684,6 +684,8 @@ start_journal_io: printk(KERN_WARNING "JBD2: Detected IO errors while flushing file data " "on %s\n", journal->j_devname); + if (journal->j_flags & JBD2_ABORT_ON_SYNCDATA_ERR) + jbd2_journal_abort(journal, err); err = 0; } -- cgit v1.2.3