summaryrefslogtreecommitdiff
path: root/fs/jbd2/transaction.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/jbd2/transaction.c')
-rw-r--r--fs/jbd2/transaction.c114
1 files changed, 62 insertions, 52 deletions
diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c
index 62e68c5b8ec3..c2125203ef2d 100644
--- a/fs/jbd2/transaction.c
+++ b/fs/jbd2/transaction.c
@@ -107,7 +107,6 @@ static void jbd2_get_transaction(journal_t *journal,
transaction->t_start_time = ktime_get();
transaction->t_tid = journal->j_transaction_sequence++;
transaction->t_expires = jiffies + journal->j_commit_interval;
- spin_lock_init(&transaction->t_handle_lock);
atomic_set(&transaction->t_updates, 0);
atomic_set(&transaction->t_outstanding_credits,
jbd2_descriptor_blocks_per_trans(journal) +
@@ -139,24 +138,21 @@ static void jbd2_get_transaction(journal_t *journal,
/*
* Update transaction's maximum wait time, if debugging is enabled.
*
- * In order for t_max_wait to be reliable, it must be protected by a
- * lock. But doing so will mean that start_this_handle() can not be
- * run in parallel on SMP systems, which limits our scalability. So
- * unless debugging is enabled, we no longer update t_max_wait, which
- * means that maximum wait time reported by the jbd2_run_stats
- * tracepoint will always be zero.
+ * t_max_wait is carefully updated here with use of atomic compare exchange.
+ * Note that there could be multiplre threads trying to do this simultaneously
+ * hence using cmpxchg to avoid any use of locks in this case.
*/
static inline void update_t_max_wait(transaction_t *transaction,
unsigned long ts)
{
#ifdef CONFIG_JBD2_DEBUG
+ unsigned long oldts, newts;
if (jbd2_journal_enable_debug &&
time_after(transaction->t_start, ts)) {
- ts = jbd2_time_diff(ts, transaction->t_start);
- spin_lock(&transaction->t_handle_lock);
- if (ts > transaction->t_max_wait)
- transaction->t_max_wait = ts;
- spin_unlock(&transaction->t_handle_lock);
+ newts = jbd2_time_diff(ts, transaction->t_start);
+ oldts = READ_ONCE(transaction->t_max_wait);
+ while (oldts < newts)
+ oldts = cmpxchg(&transaction->t_max_wait, oldts, newts);
}
#endif
}
@@ -378,7 +374,7 @@ alloc_transaction:
return -ENOMEM;
}
- jbd_debug(3, "New handle %p going live.\n", handle);
+ jbd2_debug(3, "New handle %p going live.\n", handle);
/*
* We need to hold j_state_lock until t_updates has been incremented,
@@ -449,7 +445,7 @@ repeat:
}
/* OK, account for the buffers that this operation expects to
- * use and add the handle to the running transaction.
+ * use and add the handle to the running transaction.
*/
update_t_max_wait(transaction, ts);
handle->h_transaction = transaction;
@@ -458,7 +454,7 @@ repeat:
handle->h_start_jiffies = jiffies;
atomic_inc(&transaction->t_updates);
atomic_inc(&transaction->t_handle_count);
- jbd_debug(4, "Handle %p given %d credits (total %d, free %lu)\n",
+ jbd2_debug(4, "Handle %p given %d credits (total %d, free %lu)\n",
handle, blocks,
atomic_read(&transaction->t_outstanding_credits),
jbd2_log_space_left(journal));
@@ -679,7 +675,7 @@ int jbd2_journal_extend(handle_t *handle, int nblocks, int revoke_records)
/* Don't extend a locked-down transaction! */
if (transaction->t_state != T_RUNNING) {
- jbd_debug(3, "denied handle %p %d blocks: "
+ jbd2_debug(3, "denied handle %p %d blocks: "
"transaction not running\n", handle, nblocks);
goto error_out;
}
@@ -690,15 +686,14 @@ int jbd2_journal_extend(handle_t *handle, int nblocks, int revoke_records)
DIV_ROUND_UP(
handle->h_revoke_credits_requested,
journal->j_revoke_records_per_block);
- spin_lock(&transaction->t_handle_lock);
wanted = atomic_add_return(nblocks,
&transaction->t_outstanding_credits);
if (wanted > journal->j_max_transaction_buffers) {
- jbd_debug(3, "denied handle %p %d blocks: "
+ jbd2_debug(3, "denied handle %p %d blocks: "
"transaction too large\n", handle, nblocks);
atomic_sub(nblocks, &transaction->t_outstanding_credits);
- goto unlock;
+ goto error_out;
}
trace_jbd2_handle_extend(journal->j_fs_dev->bd_dev,
@@ -713,9 +708,7 @@ int jbd2_journal_extend(handle_t *handle, int nblocks, int revoke_records)
handle->h_revoke_credits_requested += revoke_records;
result = 0;
- jbd_debug(3, "extended handle %p by %d\n", handle, nblocks);
-unlock:
- spin_unlock(&transaction->t_handle_lock);
+ jbd2_debug(3, "extended handle %p by %d\n", handle, nblocks);
error_out:
read_unlock(&journal->j_state_lock);
return result;
@@ -803,7 +796,7 @@ int jbd2__journal_restart(handle_t *handle, int nblocks, int revoke_records,
* First unlink the handle from its current transaction, and start the
* commit on that.
*/
- jbd_debug(2, "restarting handle %p\n", handle);
+ jbd2_debug(2, "restarting handle %p\n", handle);
stop_this_handle(handle);
handle->h_transaction = NULL;
@@ -836,6 +829,43 @@ int jbd2_journal_restart(handle_t *handle, int nblocks)
}
EXPORT_SYMBOL(jbd2_journal_restart);
+/*
+ * Waits for any outstanding t_updates to finish.
+ * This is called with write j_state_lock held.
+ */
+void jbd2_journal_wait_updates(journal_t *journal)
+{
+ DEFINE_WAIT(wait);
+
+ while (1) {
+ /*
+ * Note that the running transaction can get freed under us if
+ * this transaction is getting committed in
+ * jbd2_journal_commit_transaction() ->
+ * jbd2_journal_free_transaction(). This can only happen when we
+ * release j_state_lock -> schedule() -> acquire j_state_lock.
+ * Hence we should everytime retrieve new j_running_transaction
+ * value (after j_state_lock release acquire cycle), else it may
+ * lead to use-after-free of old freed transaction.
+ */
+ transaction_t *transaction = journal->j_running_transaction;
+
+ if (!transaction)
+ break;
+
+ prepare_to_wait(&journal->j_wait_updates, &wait,
+ TASK_UNINTERRUPTIBLE);
+ if (!atomic_read(&transaction->t_updates)) {
+ finish_wait(&journal->j_wait_updates, &wait);
+ break;
+ }
+ write_unlock(&journal->j_state_lock);
+ schedule();
+ finish_wait(&journal->j_wait_updates, &wait);
+ write_lock(&journal->j_state_lock);
+ }
+}
+
/**
* jbd2_journal_lock_updates () - establish a transaction barrier.
* @journal: Journal to establish a barrier on.
@@ -848,8 +878,6 @@ EXPORT_SYMBOL(jbd2_journal_restart);
*/
void jbd2_journal_lock_updates(journal_t *journal)
{
- DEFINE_WAIT(wait);
-
jbd2_might_wait_for_commit(journal);
write_lock(&journal->j_state_lock);
@@ -863,27 +891,9 @@ void jbd2_journal_lock_updates(journal_t *journal)
write_lock(&journal->j_state_lock);
}
- /* Wait until there are no running updates */
- while (1) {
- transaction_t *transaction = journal->j_running_transaction;
-
- if (!transaction)
- break;
+ /* Wait until there are no running t_updates */
+ jbd2_journal_wait_updates(journal);
- spin_lock(&transaction->t_handle_lock);
- prepare_to_wait(&journal->j_wait_updates, &wait,
- TASK_UNINTERRUPTIBLE);
- if (!atomic_read(&transaction->t_updates)) {
- spin_unlock(&transaction->t_handle_lock);
- finish_wait(&journal->j_wait_updates, &wait);
- break;
- }
- spin_unlock(&transaction->t_handle_lock);
- write_unlock(&journal->j_state_lock);
- schedule();
- finish_wait(&journal->j_wait_updates, &wait);
- write_lock(&journal->j_state_lock);
- }
write_unlock(&journal->j_state_lock);
/*
@@ -970,7 +980,7 @@ do_get_write_access(handle_t *handle, struct journal_head *jh,
journal = transaction->t_journal;
- jbd_debug(5, "journal_head %p, force_copy %d\n", jh, force_copy);
+ jbd2_debug(5, "journal_head %p, force_copy %d\n", jh, force_copy);
JBUFFER_TRACE(jh, "entry");
repeat:
@@ -1270,7 +1280,7 @@ int jbd2_journal_get_create_access(handle_t *handle, struct buffer_head *bh)
struct journal_head *jh = jbd2_journal_add_journal_head(bh);
int err;
- jbd_debug(5, "journal_head %p\n", jh);
+ jbd2_debug(5, "journal_head %p\n", jh);
err = -EROFS;
if (is_handle_aborted(handle))
goto out;
@@ -1493,7 +1503,7 @@ int jbd2_journal_dirty_metadata(handle_t *handle, struct buffer_head *bh)
* of the running transaction.
*/
jh = bh2jh(bh);
- jbd_debug(5, "journal_head %p\n", jh);
+ jbd2_debug(5, "journal_head %p\n", jh);
JBUFFER_TRACE(jh, "entry");
/*
@@ -1826,7 +1836,7 @@ int jbd2_journal_stop(handle_t *handle)
pid_t pid;
if (--handle->h_ref > 0) {
- jbd_debug(4, "h_ref %d -> %d\n", handle->h_ref + 1,
+ jbd2_debug(4, "h_ref %d -> %d\n", handle->h_ref + 1,
handle->h_ref);
if (is_handle_aborted(handle))
return -EIO;
@@ -1846,7 +1856,7 @@ int jbd2_journal_stop(handle_t *handle)
if (is_handle_aborted(handle))
err = -EIO;
- jbd_debug(4, "Handle %p going down\n", handle);
+ jbd2_debug(4, "Handle %p going down\n", handle);
trace_jbd2_handle_stats(journal->j_fs_dev->bd_dev,
tid, handle->h_type, handle->h_line_no,
jiffies - handle->h_start_jiffies,
@@ -1924,7 +1934,7 @@ int jbd2_journal_stop(handle_t *handle)
* completes the commit thread, it just doesn't write
* anything to disk. */
- jbd_debug(2, "transaction too old, requesting commit for "
+ jbd2_debug(2, "transaction too old, requesting commit for "
"handle %p\n", handle);
/* This is non-blocking */
jbd2_log_start_commit(journal, tid);
@@ -2668,7 +2678,7 @@ static int jbd2_journal_file_inode(handle_t *handle, struct jbd2_inode *jinode,
return -EROFS;
journal = transaction->t_journal;
- jbd_debug(4, "Adding inode %lu, tid:%d\n", jinode->i_vfs_inode->i_ino,
+ jbd2_debug(4, "Adding inode %lu, tid:%d\n", jinode->i_vfs_inode->i_ino,
transaction->t_tid);
spin_lock(&journal->j_list_lock);
generated by cgit v1.2.3 (git 2.0.4) at 2024-05-28 21:37:49 +0000