3.2.45-rc1 review patch. If anyone has any objections, please let me know.
------------------
From: Theodore Ts'o <tytso@xxxxxxx>
commit d76a3a77113db020d9bb1e894822869410450bd9 upstream.
In the case where an inode has a very stale transaction id (tid) in
i_datasync_tid or i_sync_tid, it's possible that after a very large
(2**31) number of transactions, that the tid number space might wrap,
causing tid_geq()'s calculations to fail.
Commit deeeaf13 "jbd2: fix fsync() tid wraparound bug", later modified
by commit e7b04ac0 "jbd2: don't wake kjournald unnecessarily",
attempted to fix this problem, but it only avoided kjournald spinning
forever by fixing the logic in jbd2_log_start_commit().
Unfortunately, in the codepaths in fs/ext4/fsync.c and fs/ext4/inode.c
that might call jbd2_log_start_commit() with a stale tid, those
functions will subsequently call jbd2_log_wait_commit() with the same
stale tid, and then wait for a very long time. To fix this, we
replace the calls to jbd2_log_start_commit() and
jbd2_log_wait_commit() with a call to a new function,
jbd2_complete_transaction(), which will correctly handle stale tid's.
As a bonus, jbd2_complete_transaction() will avoid locking
j_state_lock for writing unless a commit needs to be started. This
should have a small (but probably not measurable) improvement for
ext4's scalability.
Signed-off-by: "Theodore Ts'o" <tytso@xxxxxxx>
Reported-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
Reported-by: George Barnett <gbarnett@xxxxxxxxxxxxx>
[bwh: Backported to 3.2: adjust context]
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
---
fs/ext4/fsync.c | 3 +--
fs/ext4/inode.c | 3 +--
fs/jbd2/journal.c | 31 +++++++++++++++++++++++++++++++
include/linux/jbd2.h | 1 +
4 files changed, 34 insertions(+), 4 deletions(-)
--- a/fs/ext4/fsync.c
+++ b/fs/ext4/fsync.c
@@ -260,8 +260,7 @@ int ext4_sync_file(struct file *file, lo
if (journal->j_flags & JBD2_BARRIER &&
!jbd2_trans_will_send_data_barrier(journal, commit_tid))
needs_barrier = true;
- jbd2_log_start_commit(journal, commit_tid);
- ret = jbd2_log_wait_commit(journal, commit_tid);
+ ret = jbd2_complete_transaction(journal, commit_tid);
if (needs_barrier)
blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL);
out:
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -147,8 +147,7 @@ void ext4_evict_inode(struct inode *inod
journal_t *journal = EXT4_SB(inode->i_sb)->s_journal;
tid_t commit_tid = EXT4_I(inode)->i_datasync_tid;
- jbd2_log_start_commit(journal, commit_tid);
- jbd2_log_wait_commit(journal, commit_tid);
+ jbd2_complete_transaction(journal, commit_tid);
filemap_write_and_wait(&inode->i_data);
}
truncate_inode_pages(&inode->i_data, 0);
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -663,6 +663,37 @@ int jbd2_log_wait_commit(journal_t *jour
}
/*
+ * When this function returns the transaction corresponding to tid
+ * will be completed. If the transaction has currently running, start
+ * committing that transaction before waiting for it to complete. If
+ * the transaction id is stale, it is by definition already completed,
+ * so just return SUCCESS.
+ */
+int jbd2_complete_transaction(journal_t *journal, tid_t tid)
+{
+ int need_to_wait = 1;
+
+ read_lock(&journal->j_state_lock);
+ if (journal->j_running_transaction &&
+ journal->j_running_transaction->t_tid == tid) {
+ if (journal->j_commit_request != tid) {
+ /* transaction not yet started, so request it */
+ read_unlock(&journal->j_state_lock);
+ jbd2_log_start_commit(journal, tid);
+ goto wait_commit;
+ }
+ } else if (!(journal->j_committing_transaction &&
+ journal->j_committing_transaction->t_tid == tid))
+ need_to_wait = 0;
+ read_unlock(&journal->j_state_lock);
+ if (!need_to_wait)
+ return 0;
+wait_commit:
+ return jbd2_log_wait_commit(journal, tid);
+}
+EXPORT_SYMBOL(jbd2_complete_transaction);
+
+/*
* Log buffer allocation routines:
*/
--- a/include/linux/jbd2.h
+++ b/include/linux/jbd2.h
@@ -1165,6 +1165,7 @@ int __jbd2_log_start_commit(journal_t *j
int jbd2_journal_start_commit(journal_t *journal, tid_t *tid);
int jbd2_journal_force_commit_nested(journal_t *journal);
int jbd2_log_wait_commit(journal_t *journal, tid_t tid);
+int jbd2_complete_transaction(journal_t *journal, tid_t tid);
int jbd2_log_do_checkpoint(journal_t *journal);
int jbd2_trans_will_send_data_barrier(journal_t *journal, tid_t tid);
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html