From patchwork Wed Apr 11 19:58:36 2012 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jan Kara X-Patchwork-Id: 151865 Return-Path: X-Original-To: patchwork-incoming@ozlabs.org Delivered-To: patchwork-incoming@ozlabs.org Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by ozlabs.org (Postfix) with ESMTP id D3BF6B7074 for ; Thu, 12 Apr 2012 05:58:50 +1000 (EST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933005Ab2DKT6s (ORCPT ); Wed, 11 Apr 2012 15:58:48 -0400 Received: from cantor2.suse.de ([195.135.220.15]:57641 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933049Ab2DKT6p (ORCPT ); Wed, 11 Apr 2012 15:58:45 -0400 Received: from relay2.suse.de (unknown [195.135.220.254]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by mx2.suse.de (Postfix) with ESMTP id D2DB390182 for ; Wed, 11 Apr 2012 21:58:44 +0200 (CEST) Received: by quack.suse.cz (Postfix, from userid 1000) id 7E8AE2046F; Wed, 11 Apr 2012 21:58:43 +0200 (CEST) From: Jan Kara To: linux-ext4@vger.kernel.org Cc: Jan Kara Subject: [PATCH 1/3] jbd: Split updating of journal superblock and marking journal empty Date: Wed, 11 Apr 2012 21:58:36 +0200 Message-Id: <1334174318-11735-2-git-send-email-jack@suse.cz> X-Mailer: git-send-email 1.7.1 In-Reply-To: <1334174318-11735-1-git-send-email-jack@suse.cz> References: <1334174318-11735-1-git-send-email-jack@suse.cz> Sender: linux-ext4-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-ext4@vger.kernel.org There are three case of updating journal superblock. In the first case, we want to mark journal as empty (setting s_sequence to 0), in the second case we want to update log tail, in the third case we want to update s_errno. Split these cases into separate functions. It makes the code slightly more straightforward and later patches will make the distinction even more important. Signed-off-by: Jan Kara --- fs/jbd/checkpoint.c | 2 +- fs/jbd/commit.c | 2 +- fs/jbd/journal.c | 164 ++++++++++++++++++++++++++------------------ include/linux/jbd.h | 2 +- include/trace/events/jbd.h | 12 +-- 5 files changed, 104 insertions(+), 78 deletions(-) diff --git a/fs/jbd/checkpoint.c b/fs/jbd/checkpoint.c index 05f0754..80c85f3 100644 --- a/fs/jbd/checkpoint.c +++ b/fs/jbd/checkpoint.c @@ -540,7 +540,7 @@ int cleanup_journal_tail(journal_t *journal) journal->j_tail = blocknr; spin_unlock(&journal->j_state_lock); if (!(journal->j_flags & JFS_ABORT)) - journal_update_superblock(journal, 1); + journal_update_sb_log_tail(journal); return 0; } diff --git a/fs/jbd/commit.c b/fs/jbd/commit.c index 9d31e6a..dba9cfd 100644 --- a/fs/jbd/commit.c +++ b/fs/jbd/commit.c @@ -308,7 +308,7 @@ void journal_commit_transaction(journal_t *journal) /* Do we need to erase the effects of a prior journal_flush? */ if (journal->j_flags & JFS_FLUSHED) { jbd_debug(3, "super block updated\n"); - journal_update_superblock(journal, 1); + journal_update_sb_log_tail(journal); } else { jbd_debug(3, "superblock not updated\n"); } diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c index 2047fd7..44c104a 100644 --- a/fs/jbd/journal.c +++ b/fs/jbd/journal.c @@ -923,8 +923,22 @@ static int journal_reset(journal_t *journal) journal->j_max_transaction_buffers = journal->j_maxlen / 4; - /* Add the dynamic fields and write it to disk. */ - journal_update_superblock(journal, 1); + /* + * As a special case, if the on-disk copy is already marked as needing + * no recovery (s_start == 0), then we can safely defer the superblock + * update until the next commit by setting JFS_FLUSHED. This avoids + * attempting a write to a potential-readonly device. + */ + if (sb->s_start == 0) { + jbd_debug(1,"JBD: Skipping superblock update on recovered sb " + "(start %u, seq %d, errno %d)\n", + journal->j_tail, journal->j_tail_sequence, + journal->j_errno); + journal->j_flags |= JFS_FLUSHED; + } else { + /* Add the dynamic fields and write it to disk. */ + journal_update_sb_log_tail(journal); + } return journal_start_thread(journal); } @@ -1001,35 +1015,11 @@ int journal_create(journal_t *journal) return journal_reset(journal); } -/** - * void journal_update_superblock() - Update journal sb on disk. - * @journal: The journal to update. - * @wait: Set to '0' if you don't want to wait for IO completion. - * - * Update a journal's dynamic superblock fields and write it to disk, - * optionally waiting for the IO to complete. - */ -void journal_update_superblock(journal_t *journal, int wait) +static void journal_write_superblock(journal_t *journal) { - journal_superblock_t *sb = journal->j_superblock; struct buffer_head *bh = journal->j_sb_buffer; - /* - * As a special case, if the on-disk copy is already marked as needing - * no recovery (s_start == 0) and there are no outstanding transactions - * in the filesystem, then we can safely defer the superblock update - * until the next commit by setting JFS_FLUSHED. This avoids - * attempting a write to a potential-readonly device. - */ - if (sb->s_start == 0 && journal->j_tail_sequence == - journal->j_transaction_sequence) { - jbd_debug(1,"JBD: Skipping superblock update on recovered sb " - "(start %u, seq %d, errno %d)\n", - journal->j_tail, journal->j_tail_sequence, - journal->j_errno); - goto out; - } - + trace_journal_write_superblock(journal); if (buffer_write_io_error(bh)) { char b[BDEVNAME_SIZE]; /* @@ -1047,44 +1037,94 @@ void journal_update_superblock(journal_t *journal, int wait) set_buffer_uptodate(bh); } + BUFFER_TRACE(bh, "marking dirty"); + mark_buffer_dirty(bh); + sync_dirty_buffer(bh); + if (buffer_write_io_error(bh)) { + char b[BDEVNAME_SIZE]; + printk(KERN_ERR "JBD: I/O error detected " + "when updating journal superblock for %s.\n", + journal_dev_name(journal, b)); + clear_buffer_write_io_error(bh); + set_buffer_uptodate(bh); + } +} + +/** + * journal_update_sb_log_tail() - Update log tail in journal sb on disk. + * @journal: The journal to update. + * + * Update a journal's superblock information about log tail and write it to + * disk, waiting for the IO to complete. + */ +void journal_update_sb_log_tail(journal_t *journal) +{ + journal_superblock_t *sb = journal->j_superblock; + spin_lock(&journal->j_state_lock); jbd_debug(1,"JBD: updating superblock (start %u, seq %d, errno %d)\n", journal->j_tail, journal->j_tail_sequence, journal->j_errno); sb->s_sequence = cpu_to_be32(journal->j_tail_sequence); sb->s_start = cpu_to_be32(journal->j_tail); - sb->s_errno = cpu_to_be32(journal->j_errno); spin_unlock(&journal->j_state_lock); - BUFFER_TRACE(bh, "marking dirty"); - mark_buffer_dirty(bh); - if (wait) { - sync_dirty_buffer(bh); - if (buffer_write_io_error(bh)) { - char b[BDEVNAME_SIZE]; - printk(KERN_ERR "JBD: I/O error detected " - "when updating journal superblock for %s.\n", - journal_dev_name(journal, b)); - clear_buffer_write_io_error(bh); - set_buffer_uptodate(bh); - } - } else - write_dirty_buffer(bh, WRITE); + journal_write_superblock(journal); - trace_jbd_update_superblock_end(journal, wait); -out: - /* If we have just flushed the log (by marking s_start==0), then - * any future commit will have to be careful to update the - * superblock again to re-record the true start of the log. */ + /* Log is no longer empty */ + spin_lock(&journal->j_state_lock); + WARN_ON(!sb->s_sequence); + journal->j_flags &= ~JFS_FLUSHED; + spin_unlock(&journal->j_state_lock); +} + +/** + * mark_journal_empty() - Mark on disk journal as empty. + * @journal: The journal to update. + * + * Update a journal's dynamic superblock fields to show that journal is empty. + * Write updated superblock to disk waiting for IO to complete. + */ +static void mark_journal_empty(journal_t *journal) +{ + journal_superblock_t *sb = journal->j_superblock; spin_lock(&journal->j_state_lock); - if (sb->s_start) - journal->j_flags &= ~JFS_FLUSHED; - else - journal->j_flags |= JFS_FLUSHED; + jbd_debug(1, "JBD: Marking journal as empty (seq %d)\n", + journal->j_tail_sequence); + + sb->s_sequence = cpu_to_be32(journal->j_tail_sequence); + sb->s_start = cpu_to_be32(0); + spin_unlock(&journal->j_state_lock); + + journal_write_superblock(journal); + + spin_lock(&journal->j_state_lock); + /* Log is empty */ + journal->j_flags |= JFS_FLUSHED; spin_unlock(&journal->j_state_lock); } +/** + * journal_update_sb_errno() - Update error in the journal. + * @journal: The journal to update. + * + * Update a journal's errno. Write updated superblock to disk waiting for IO + * to complete. + */ +static void journal_update_sb_errno(journal_t *journal) +{ + journal_superblock_t *sb = journal->j_superblock; + + spin_lock(&journal->j_state_lock); + jbd_debug(1, "JBD: updating superblock error (errno %d)\n", + journal->j_errno); + sb->s_errno = cpu_to_be32(journal->j_errno); + spin_unlock(&journal->j_state_lock); + + journal_write_superblock(journal); +} + /* * Read the superblock for a given journal, performing initial * validation of the format. @@ -1268,14 +1308,11 @@ int journal_destroy(journal_t *journal) if (journal->j_sb_buffer) { if (!is_journal_aborted(journal)) { - /* We can now mark the journal as empty. */ - journal->j_tail = 0; journal->j_tail_sequence = ++journal->j_transaction_sequence; - journal_update_superblock(journal, 1); - } else { + mark_journal_empty(journal); + } else err = -EIO; - } brelse(journal->j_sb_buffer); } @@ -1457,7 +1494,6 @@ int journal_flush(journal_t *journal) { int err = 0; transaction_t *transaction = NULL; - unsigned int old_tail; spin_lock(&journal->j_state_lock); @@ -1499,14 +1535,8 @@ int journal_flush(journal_t *journal) * the magic code for a fully-recovered superblock. Any future * commits of data to the journal will restore the current * s_start value. */ + mark_journal_empty(journal); spin_lock(&journal->j_state_lock); - old_tail = journal->j_tail; - journal->j_tail = 0; - spin_unlock(&journal->j_state_lock); - journal_update_superblock(journal, 1); - spin_lock(&journal->j_state_lock); - journal->j_tail = old_tail; - J_ASSERT(!journal->j_running_transaction); J_ASSERT(!journal->j_committing_transaction); J_ASSERT(!journal->j_checkpoint_transactions); @@ -1547,7 +1577,7 @@ int journal_wipe(journal_t *journal, int write) err = journal_skip_recovery(journal); if (write) - journal_update_superblock(journal, 1); + mark_journal_empty(journal); no_recovery: return err; @@ -1615,7 +1645,7 @@ static void __journal_abort_soft (journal_t *journal, int errno) __journal_abort_hard(journal); if (errno) - journal_update_superblock(journal, 1); + journal_update_sb_errno(journal); } /** diff --git a/include/linux/jbd.h b/include/linux/jbd.h index f265682..9716d37 100644 --- a/include/linux/jbd.h +++ b/include/linux/jbd.h @@ -864,7 +864,7 @@ extern int journal_destroy (journal_t *); extern int journal_recover (journal_t *journal); extern int journal_wipe (journal_t *, int); extern int journal_skip_recovery (journal_t *); -extern void journal_update_superblock (journal_t *, int); +extern void journal_update_sb_log_tail (journal_t *); extern void journal_abort (journal_t *, int); extern int journal_errno (journal_t *); extern void journal_ack_err (journal_t *); diff --git a/include/trace/events/jbd.h b/include/trace/events/jbd.h index 9305e1b..d9658a9 100644 --- a/include/trace/events/jbd.h +++ b/include/trace/events/jbd.h @@ -169,24 +169,20 @@ TRACE_EVENT(jbd_cleanup_journal_tail, __entry->block_nr, __entry->freed) ); -TRACE_EVENT(jbd_update_superblock_end, - TP_PROTO(journal_t *journal, int wait), +TRACE_EVENT(journal_write_superblock, + TP_PROTO(journal_t *journal), - TP_ARGS(journal, wait), + TP_ARGS(journal), TP_STRUCT__entry( __field( dev_t, dev ) - __field( int, wait ) ), TP_fast_assign( __entry->dev = journal->j_fs_dev->bd_dev; - __entry->wait = wait; ), - TP_printk("dev %d,%d wait %d", - MAJOR(__entry->dev), MINOR(__entry->dev), - __entry->wait) + TP_printk("dev %d,%d", MAJOR(__entry->dev), MINOR(__entry->dev)) ); #endif /* _TRACE_JBD_H */