1 files changed, 271 insertions, 202 deletions

diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index 0f32b445582a..3d72565ec6e8 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -65,21 +65,11 @@
  *
  * Fast Commit Ineligibility
  * -------------------------
- * Not all operations are supported by fast commits today (e.g extended
- * attributes). Fast commit ineligibility is marked by calling one of the
- * two following functions:
- *
- * - ext4_fc_mark_ineligible(): This makes next fast commit operation to fall
- *   back to full commit. This is useful in case of transient errors.
  *
- * - ext4_fc_start_ineligible() and ext4_fc_stop_ineligible() - This makes all
- *   the fast commits happening between ext4_fc_start_ineligible() and
- *   ext4_fc_stop_ineligible() and one fast commit after the call to
- *   ext4_fc_stop_ineligible() to fall back to full commits. It is important to
- *   make one more fast commit to fall back to full commit after stop call so
- *   that it guaranteed that the fast commit ineligible operation contained
- *   within ext4_fc_start_ineligible() and ext4_fc_stop_ineligible() is
- *   followed by at least 1 full commit.
+ * Not all operations are supported by fast commits today (e.g extended
+ * attributes). Fast commit ineligibility is marked by calling
+ * ext4_fc_mark_ineligible(): This makes next fast commit operation to fall back
+ * to full commit.
  *
  * Atomicity of commits
  * --------------------
@@ -166,15 +156,13 @@
  *    fast commit recovery even if that area is invalidated by later full
  *    commits.
  *
- * 1) Make fast commit atomic updates more fine grained. Today, a fast commit
- *    eligible update must be protected within ext4_fc_start_update() and
- *    ext4_fc_stop_update(). These routines are called at much higher
- *    routines. This can be made more fine grained by combining with
- *    ext4_journal_start().
+ * 1) Fast commit's commit path locks the entire file system during fast
+ *    commit. This has significant performance penalty. Instead of that, we
+ *    should use ext4_fc_start/stop_update functions to start inode level
+ *    updates from ext4_journal_start/stop. Once we do that we can drop file
+ *    system locking during commit path.
  *
- * 2) Same above for ext4_fc_start_ineligible() and ext4_fc_stop_ineligible()
- *
- * 3) Handle more ineligible cases.
+ * 2) Handle more ineligible cases.
  */
 
 #include <trace/events/ext4.h>
@@ -211,6 +199,7 @@ void ext4_fc_init_inode(struct inode *inode)
 	ext4_fc_reset_inode(inode);
 	ext4_clear_inode_state(inode, EXT4_STATE_FC_COMMITTING);
 	INIT_LIST_HEAD(&ei->i_fc_list);
+	INIT_LIST_HEAD(&ei->i_fc_dilist);
 	init_waitqueue_head(&ei->i_fc_wait);
 	atomic_set(&ei->i_fc_updates, 0);
 }
@@ -291,6 +280,8 @@ void ext4_fc_stop_update(struct inode *inode)
 void ext4_fc_del(struct inode *inode)
 {
 	struct ext4_inode_info *ei = EXT4_I(inode);
+	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
+	struct ext4_fc_dentry_update *fc_dentry;
 
 	if (!test_opt2(inode->i_sb, JOURNAL_FAST_COMMIT) ||
 	    (EXT4_SB(inode->i_sb)->s_mount_state & EXT4_FC_REPLAY))
@@ -298,7 +289,7 @@ void ext4_fc_del(struct inode *inode)
 
 restart:
 	spin_lock(&EXT4_SB(inode->i_sb)->s_fc_lock);
-	if (list_empty(&ei->i_fc_list)) {
+	if (list_empty(&ei->i_fc_list) && list_empty(&ei->i_fc_dilist)) {
 		spin_unlock(&EXT4_SB(inode->i_sb)->s_fc_lock);
 		return;
 	}
@@ -307,63 +298,64 @@ restart:
 		ext4_fc_wait_committing_inode(inode);
 		goto restart;
 	}
-	list_del_init(&ei->i_fc_list);
-	spin_unlock(&EXT4_SB(inode->i_sb)->s_fc_lock);
-}
 
-/*
- * Mark file system as fast commit ineligible. This means that next commit
- * operation would result in a full jbd2 commit.
- */
-void ext4_fc_mark_ineligible(struct super_block *sb, int reason)
-{
-	struct ext4_sb_info *sbi = EXT4_SB(sb);
+	if (!list_empty(&ei->i_fc_list))
+		list_del_init(&ei->i_fc_list);
 
-	if (!test_opt2(sb, JOURNAL_FAST_COMMIT) ||
-	    (EXT4_SB(sb)->s_mount_state & EXT4_FC_REPLAY))
+	/*
+	 * Since this inode is getting removed, let's also remove all FC
+	 * dentry create references, since it is not needed to log it anyways.
+	 */
+	if (list_empty(&ei->i_fc_dilist)) {
+		spin_unlock(&sbi->s_fc_lock);
 		return;
+	}
 
-	ext4_set_mount_flag(sb, EXT4_MF_FC_INELIGIBLE);
-	WARN_ON(reason >= EXT4_FC_REASON_MAX);
-	sbi->s_fc_stats.fc_ineligible_reason_count[reason]++;
-}
+	fc_dentry = list_first_entry(&ei->i_fc_dilist, struct ext4_fc_dentry_update, fcd_dilist);
+	WARN_ON(fc_dentry->fcd_op != EXT4_FC_TAG_CREAT);
+	list_del_init(&fc_dentry->fcd_list);
+	list_del_init(&fc_dentry->fcd_dilist);
 
-/*
- * Start a fast commit ineligible update. Any commits that happen while
- * such an operation is in progress fall back to full commits.
- */
-void ext4_fc_start_ineligible(struct super_block *sb, int reason)
-{
-	struct ext4_sb_info *sbi = EXT4_SB(sb);
+	WARN_ON(!list_empty(&ei->i_fc_dilist));
+	spin_unlock(&sbi->s_fc_lock);
 
-	if (!test_opt2(sb, JOURNAL_FAST_COMMIT) ||
-	    (EXT4_SB(sb)->s_mount_state & EXT4_FC_REPLAY))
-		return;
+	if (fc_dentry->fcd_name.name &&
+		fc_dentry->fcd_name.len > DNAME_INLINE_LEN)
+		kfree(fc_dentry->fcd_name.name);
+	kmem_cache_free(ext4_fc_dentry_cachep, fc_dentry);
 
-	WARN_ON(reason >= EXT4_FC_REASON_MAX);
-	sbi->s_fc_stats.fc_ineligible_reason_count[reason]++;
-	atomic_inc(&sbi->s_fc_ineligible_updates);
+	return;
 }
 
 /*
- * Stop a fast commit ineligible update. We set EXT4_MF_FC_INELIGIBLE flag here
- * to ensure that after stopping the ineligible update, at least one full
- * commit takes place.
+ * Mark file system as fast commit ineligible, and record latest
+ * ineligible transaction tid. This means until the recorded
+ * transaction, commit operation would result in a full jbd2 commit.
  */
-void ext4_fc_stop_ineligible(struct super_block *sb)
+void ext4_fc_mark_ineligible(struct super_block *sb, int reason, handle_t *handle)
 {
+	struct ext4_sb_info *sbi = EXT4_SB(sb);
+	tid_t tid;
+
 	if (!test_opt2(sb, JOURNAL_FAST_COMMIT) ||
 	    (EXT4_SB(sb)->s_mount_state & EXT4_FC_REPLAY))
 		return;
 
 	ext4_set_mount_flag(sb, EXT4_MF_FC_INELIGIBLE);
-	atomic_dec(&EXT4_SB(sb)->s_fc_ineligible_updates);
-}
-
-static inline int ext4_fc_is_ineligible(struct super_block *sb)
-{
-	return (ext4_test_mount_flag(sb, EXT4_MF_FC_INELIGIBLE) ||
-		atomic_read(&EXT4_SB(sb)->s_fc_ineligible_updates));
+	if (handle && !IS_ERR(handle))
+		tid = handle->h_transaction->t_tid;
+	else {
+		read_lock(&sbi->s_journal->j_state_lock);
+		tid = sbi->s_journal->j_running_transaction ?
+				sbi->s_journal->j_running_transaction->t_tid : 0;
+		read_unlock(&sbi->s_journal->j_state_lock);
+	}
+	spin_lock(&sbi->s_fc_lock);
+	if (sbi->s_fc_ineligible_tid < tid)
+		sbi->s_fc_ineligible_tid = tid;
+	spin_unlock(&sbi->s_fc_lock);
+	WARN_ON(reason >= EXT4_FC_REASON_MAX);
+	sbi->s_fc_stats.fc_ineligible_reason_count[reason]++;
 }
 
 /*
@@ -387,13 +379,6 @@ static int ext4_fc_track_template(
 	tid_t tid = 0;
 	int ret;
 
-	if (!test_opt2(inode->i_sb, JOURNAL_FAST_COMMIT) ||
-	    (sbi->s_mount_state & EXT4_FC_REPLAY))
-		return -EOPNOTSUPP;
-
-	if (ext4_fc_is_ineligible(inode->i_sb))
-		return -EINVAL;
-
 	tid = handle->h_transaction->t_tid;
 	mutex_lock(&ei->i_fc_lock);
 	if (tid == ei->i_sync_tid) {
@@ -411,7 +396,8 @@ static int ext4_fc_track_template(
 	spin_lock(&sbi->s_fc_lock);
 	if (list_empty(&EXT4_I(inode)->i_fc_list))
 		list_add_tail(&EXT4_I(inode)->i_fc_list,
-				(ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_COMMITTING)) ?
+				(sbi->s_journal->j_flags & JBD2_FULL_COMMIT_ONGOING ||
+				 sbi->s_journal->j_flags & JBD2_FAST_COMMIT_ONGOING) ?
 				&sbi->s_fc_q[FC_Q_STAGING] :
 				&sbi->s_fc_q[FC_Q_MAIN]);
 	spin_unlock(&sbi->s_fc_lock);
@@ -437,7 +423,7 @@ static int __track_dentry_update(struct inode *inode, void *arg, bool update)
 	mutex_unlock(&ei->i_fc_lock);
 	node = kmem_cache_alloc(ext4_fc_dentry_cachep, GFP_NOFS);
 	if (!node) {
-		ext4_fc_mark_ineligible(inode->i_sb, EXT4_FC_REASON_NOMEM);
+		ext4_fc_mark_ineligible(inode->i_sb, EXT4_FC_REASON_NOMEM, NULL);
 		mutex_lock(&ei->i_fc_lock);
 		return -ENOMEM;
 	}
@@ -450,7 +436,7 @@ static int __track_dentry_update(struct inode *inode, void *arg, bool update)
 		if (!node->fcd_name.name) {
 			kmem_cache_free(ext4_fc_dentry_cachep, node);
 			ext4_fc_mark_ineligible(inode->i_sb,
-				EXT4_FC_REASON_NOMEM);
+				EXT4_FC_REASON_NOMEM, NULL);
 			mutex_lock(&ei->i_fc_lock);
 			return -ENOMEM;
 		}
@@ -462,13 +448,28 @@ static int __track_dentry_update(struct inode *inode, void *arg, bool update)
 		node->fcd_name.name = node->fcd_iname;
 	}
 	node->fcd_name.len = dentry->d_name.len;
-
+	INIT_LIST_HEAD(&node->fcd_dilist);
 	spin_lock(&sbi->s_fc_lock);
-	if (ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_COMMITTING))
+	if (sbi->s_journal->j_flags & JBD2_FULL_COMMIT_ONGOING ||
+		sbi->s_journal->j_flags & JBD2_FAST_COMMIT_ONGOING)
 		list_add_tail(&node->fcd_list,
 				&sbi->s_fc_dentry_q[FC_Q_STAGING]);
 	else
 		list_add_tail(&node->fcd_list, &sbi->s_fc_dentry_q[FC_Q_MAIN]);
+
+	/*
+	 * This helps us keep a track of all fc_dentry updates which is part of
+	 * this ext4 inode. So in case the inode is getting unlinked, before
+	 * even we get a chance to fsync, we could remove all fc_dentry
+	 * references while evicting the inode in ext4_fc_del().
+	 * Also with this, we don't need to loop over all the inodes in
+	 * sbi->s_fc_q to get the corresponding inode in
+	 * ext4_fc_commit_dentry_updates().
+	 */
+	if (dentry_update->op == EXT4_FC_TAG_CREAT) {
+		WARN_ON(!list_empty(&ei->i_fc_dilist));
+		list_add_tail(&node->fcd_dilist, &ei->i_fc_dilist);
+	}
 	spin_unlock(&sbi->s_fc_lock);
 	mutex_lock(&ei->i_fc_lock);
 
@@ -486,12 +487,22 @@ void __ext4_fc_track_unlink(handle_t *handle,
 
 	ret = ext4_fc_track_template(handle, inode, __track_dentry_update,
 					(void *)&args, 0);
-	trace_ext4_fc_track_unlink(inode, dentry, ret);
+	trace_ext4_fc_track_unlink(handle, inode, dentry, ret);
 }
 
 void ext4_fc_track_unlink(handle_t *handle, struct dentry *dentry)
 {
-	__ext4_fc_track_unlink(handle, d_inode(dentry), dentry);
+	struct inode *inode = d_inode(dentry);
+	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
+
+	if (!test_opt2(inode->i_sb, JOURNAL_FAST_COMMIT) ||
+	    (sbi->s_mount_state & EXT4_FC_REPLAY))
+		return;
+
+	if (ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_INELIGIBLE))
+		return;
+
+	__ext4_fc_track_unlink(handle, inode, dentry);
 }
 
 void __ext4_fc_track_link(handle_t *handle,
@@ -505,12 +516,22 @@ void __ext4_fc_track_link(handle_t *handle,
 
 	ret = ext4_fc_track_template(handle, inode, __track_dentry_update,
 					(void *)&args, 0);
-	trace_ext4_fc_track_link(inode, dentry, ret);
+	trace_ext4_fc_track_link(handle, inode, dentry, ret);
 }
 
 void ext4_fc_track_link(handle_t *handle, struct dentry *dentry)
 {
-	__ext4_fc_track_link(handle, d_inode(dentry), dentry);
+	struct inode *inode = d_inode(dentry);
+	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
+
+	if (!test_opt2(inode->i_sb, JOURNAL_FAST_COMMIT) ||
+	    (sbi->s_mount_state & EXT4_FC_REPLAY))
+		return;
+
+	if (ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_INELIGIBLE))
+		return;
+
+	__ext4_fc_track_link(handle, inode, dentry);
 }
 
 void __ext4_fc_track_create(handle_t *handle, struct inode *inode,
@@ -524,12 +545,22 @@ void __ext4_fc_track_create(handle_t *handle, struct inode *inode,
 
 	ret = ext4_fc_track_template(handle, inode, __track_dentry_update,
 					(void *)&args, 0);
-	trace_ext4_fc_track_create(inode, dentry, ret);
+	trace_ext4_fc_track_create(handle, inode, dentry, ret);
 }
 
 void ext4_fc_track_create(handle_t *handle, struct dentry *dentry)
 {
-	__ext4_fc_track_create(handle, d_inode(dentry), dentry);
+	struct inode *inode = d_inode(dentry);
+	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
+
+	if (!test_opt2(inode->i_sb, JOURNAL_FAST_COMMIT) ||
+	    (sbi->s_mount_state & EXT4_FC_REPLAY))
+		return;
+
+	if (ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_INELIGIBLE))
+		return;
+
+	__ext4_fc_track_create(handle, inode, dentry);
 }
 
 /* __track_fn for inode tracking */
@@ -545,6 +576,7 @@ static int __track_inode(struct inode *inode, void *arg, bool update)
 
 void ext4_fc_track_inode(handle_t *handle, struct inode *inode)
 {
+	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
 	int ret;
 
 	if (S_ISDIR(inode->i_mode))
@@ -552,12 +584,19 @@ void ext4_fc_track_inode(handle_t *handle, struct inode *inode)
 
 	if (ext4_should_journal_data(inode)) {
 		ext4_fc_mark_ineligible(inode->i_sb,
-					EXT4_FC_REASON_INODE_JOURNAL_DATA);
+					EXT4_FC_REASON_INODE_JOURNAL_DATA, handle);
 		return;
 	}
 
+	if (!test_opt2(inode->i_sb, JOURNAL_FAST_COMMIT) ||
+	    (sbi->s_mount_state & EXT4_FC_REPLAY))
+		return;
+
+	if (ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_INELIGIBLE))
+		return;
+
 	ret = ext4_fc_track_template(handle, inode, __track_inode, NULL, 1);
-	trace_ext4_fc_track_inode(inode, ret);
+	trace_ext4_fc_track_inode(handle, inode, ret);
 }
 
 struct __track_range_args {
@@ -595,18 +634,26 @@ static int __track_range(struct inode *inode, void *arg, bool update)
 void ext4_fc_track_range(handle_t *handle, struct inode *inode, ext4_lblk_t start,
 			 ext4_lblk_t end)
 {
+	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
 	struct __track_range_args args;
 	int ret;
 
 	if (S_ISDIR(inode->i_mode))
 		return;
 
+	if (!test_opt2(inode->i_sb, JOURNAL_FAST_COMMIT) ||
+	    (sbi->s_mount_state & EXT4_FC_REPLAY))
+		return;
+
+	if (ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_INELIGIBLE))
+		return;
+
 	args.start = start;
 	args.end = end;
 
 	ret = ext4_fc_track_template(handle, inode,  __track_range, &args, 1);
 
-	trace_ext4_fc_track_range(inode, start, end, ret);
+	trace_ext4_fc_track_range(handle, inode, start, end, ret);
 }
 
 static void ext4_fc_submit_bh(struct super_block *sb, bool is_tail)
@@ -796,7 +843,6 @@ static bool ext4_fc_add_dentry_tlv(struct super_block *sb, u32 *crc,
 	ext4_fc_memcpy(sb, dst, &fcd, sizeof(fcd), crc);
 	dst += sizeof(fcd);
 	ext4_fc_memcpy(sb, dst, fc_dentry->fcd_name.name, dlen, crc);
-	dst += dlen;
 
 	return true;
 }
@@ -930,7 +976,6 @@ static int ext4_fc_submit_inode_data_all(journal_t *journal)
 	int ret = 0;
 
 	spin_lock(&sbi->s_fc_lock);
-	ext4_set_mount_flag(sb, EXT4_MF_FC_COMMITTING);
 	list_for_each_entry(ei, &sbi->s_fc_q[FC_Q_MAIN], i_fc_list) {
 		ext4_set_inode_state(&ei->vfs_inode, EXT4_STATE_FC_COMMITTING);
 		while (atomic_read(&ei->i_fc_updates)) {
@@ -990,7 +1035,7 @@ __releases(&sbi->s_fc_lock)
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
 	struct ext4_fc_dentry_update *fc_dentry, *fc_dentry_n;
 	struct inode *inode;
-	struct ext4_inode_info *ei, *ei_n;
+	struct ext4_inode_info *ei;
 	int ret;
 
 	if (list_empty(&sbi->s_fc_dentry_q[FC_Q_MAIN]))
@@ -1006,21 +1051,16 @@ __releases(&sbi->s_fc_lock)
 			spin_lock(&sbi->s_fc_lock);
 			continue;
 		}
-
-		inode = NULL;
-		list_for_each_entry_safe(ei, ei_n, &sbi->s_fc_q[FC_Q_MAIN],
-					 i_fc_list) {
-			if (ei->vfs_inode.i_ino == fc_dentry->fcd_ino) {
-				inode = &ei->vfs_inode;
-				break;
-			}
-		}
 		/*
-		 * If we don't find inode in our list, then it was deleted,
-		 * in which case, we don't need to record it's create tag.
+		 * With fcd_dilist we need not loop in sbi->s_fc_q to get the
+		 * corresponding inode pointer
 		 */
-		if (!inode)
-			continue;
+		WARN_ON(list_empty(&fc_dentry->fcd_dilist));
+		ei = list_first_entry(&fc_dentry->fcd_dilist,
+				struct ext4_inode_info, i_fc_dilist);
+		inode = &ei->vfs_inode;
+		WARN_ON(inode->i_ino != fc_dentry->fcd_ino);
+
 		spin_unlock(&sbi->s_fc_lock);
 
 		/*
@@ -1123,6 +1163,33 @@ out:
 	return ret;
 }
 
+static void ext4_fc_update_stats(struct super_block *sb, int status,
+				 u64 commit_time, int nblks, tid_t commit_tid)
+{
+	struct ext4_fc_stats *stats = &EXT4_SB(sb)->s_fc_stats;
+
+	jbd_debug(1, "Fast commit ended with status = %d for tid %u",
+			status, commit_tid);
+	if (status == EXT4_FC_STATUS_OK) {
+		stats->fc_num_commits++;
+		stats->fc_numblks += nblks;
+		if (likely(stats->s_fc_avg_commit_time))
+			stats->s_fc_avg_commit_time =
+				(commit_time +
+				 stats->s_fc_avg_commit_time * 3) / 4;
+		else
+			stats->s_fc_avg_commit_time = commit_time;
+	} else if (status == EXT4_FC_STATUS_FAILED ||
+		   status == EXT4_FC_STATUS_INELIGIBLE) {
+		if (status == EXT4_FC_STATUS_FAILED)
+			stats->fc_failed_commits++;
+		stats->fc_ineligible_commits++;
+	} else {
+		stats->fc_skipped_commits++;
+	}
+	trace_ext4_fc_commit_stop(sb, nblks, status, commit_tid);
+}
+
 /*
  * The main commit entry point. Performs a fast commit for transaction
  * commit_tid if needed. If it's not possible to perform a fast commit
@@ -1135,18 +1202,15 @@ int ext4_fc_commit(journal_t *journal, tid_t commit_tid)
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
 	int nblks = 0, ret, bsize = journal->j_blocksize;
 	int subtid = atomic_read(&sbi->s_fc_subtid);
-	int reason = EXT4_FC_REASON_OK, fc_bufs_before = 0;
+	int status = EXT4_FC_STATUS_OK, fc_bufs_before = 0;
 	ktime_t start_time, commit_time;
 
-	trace_ext4_fc_commit_start(sb);
+	if (!test_opt2(sb, JOURNAL_FAST_COMMIT))
+		return jbd2_complete_transaction(journal, commit_tid);
 
-	start_time = ktime_get();
+	trace_ext4_fc_commit_start(sb, commit_tid);
 
-	if (!test_opt2(sb, JOURNAL_FAST_COMMIT) ||
-		(ext4_fc_is_ineligible(sb))) {
-		reason = EXT4_FC_REASON_INELIGIBLE;
-		goto out;
-	}
+	start_time = ktime_get();
 
 restart_fc:
 	ret = jbd2_fc_begin_commit(journal, commit_tid);
@@ -1155,74 +1219,61 @@ restart_fc:
 		if (atomic_read(&sbi->s_fc_subtid) <= subtid &&
 			commit_tid > journal->j_commit_sequence)
 			goto restart_fc;
-		reason = EXT4_FC_REASON_ALREADY_COMMITTED;
-		goto out;
+		ext4_fc_update_stats(sb, EXT4_FC_STATUS_SKIPPED, 0, 0,
+				commit_tid);
+		return 0;
 	} else if (ret) {
-		sbi->s_fc_stats.fc_ineligible_reason_count[EXT4_FC_COMMIT_FAILED]++;
-		reason = EXT4_FC_REASON_FC_START_FAILED;
-		goto out;
+		/*
+		 * Commit couldn't start. Just update stats and perform a
+		 * full commit.
+		 */
+		ext4_fc_update_stats(sb, EXT4_FC_STATUS_FAILED, 0, 0,
+				commit_tid);
+		return jbd2_complete_transaction(journal, commit_tid);
+	}
+
+	/*
+	 * After establishing journal barrier via jbd2_fc_begin_commit(), check
+	 * if we are fast commit ineligible.
+	 */
+	if (ext4_test_mount_flag(sb, EXT4_MF_FC_INELIGIBLE)) {
+		status = EXT4_FC_STATUS_INELIGIBLE;
+		goto fallback;
 	}
 
 	fc_bufs_before = (sbi->s_fc_bytes + bsize - 1) / bsize;
 	ret = ext4_fc_perform_commit(journal);
 	if (ret < 0) {
-		sbi->s_fc_stats.fc_ineligible_reason_count[EXT4_FC_COMMIT_FAILED]++;
-		reason = EXT4_FC_REASON_FC_FAILED;
-		goto out;
+		status = EXT4_FC_STATUS_FAILED;
+		goto fallback;
 	}
 	nblks = (sbi->s_fc_bytes + bsize - 1) / bsize - fc_bufs_before;
 	ret = jbd2_fc_wait_bufs(journal, nblks);
 	if (ret < 0) {
-		sbi->s_fc_stats.fc_ineligible_reason_count[EXT4_FC_COMMIT_FAILED]++;
-		reason = EXT4_FC_REASON_FC_FAILED;
-		goto out;
+		status = EXT4_FC_STATUS_FAILED;
+		goto fallback;
 	}
 	atomic_inc(&sbi->s_fc_subtid);
-	jbd2_fc_end_commit(journal);
-out:
-	/* Has any ineligible update happened since we started? */
-	if (reason == EXT4_FC_REASON_OK && ext4_fc_is_ineligible(sb)) {
-		sbi->s_fc_stats.fc_ineligible_reason_count[EXT4_FC_COMMIT_FAILED]++;
-		reason = EXT4_FC_REASON_INELIGIBLE;
-	}
-
-	spin_lock(&sbi->s_fc_lock);
-	if (reason != EXT4_FC_REASON_OK &&
-		reason != EXT4_FC_REASON_ALREADY_COMMITTED) {
-		sbi->s_fc_stats.fc_ineligible_commits++;
-	} else {
-		sbi->s_fc_stats.fc_num_commits++;
-		sbi->s_fc_stats.fc_numblks += nblks;
-	}
-	spin_unlock(&sbi->s_fc_lock);
-	nblks = (reason == EXT4_FC_REASON_OK) ? nblks : 0;
-	trace_ext4_fc_commit_stop(sb, nblks, reason);
-	commit_time = ktime_to_ns(ktime_sub(ktime_get(), start_time));
+	ret = jbd2_fc_end_commit(journal);
 	/*
-	 * weight the commit time higher than the average time so we don't
-	 * react too strongly to vast changes in the commit time
+	 * weight the commit time higher than the average time so we
+	 * don't react too strongly to vast changes in the commit time
 	 */
-	if (likely(sbi->s_fc_avg_commit_time))
-		sbi->s_fc_avg_commit_time = (commit_time +
-				sbi->s_fc_avg_commit_time * 3) / 4;
-	else
-		sbi->s_fc_avg_commit_time = commit_time;
-	jbd_debug(1,
-		"Fast commit ended with blks = %d, reason = %d, subtid - %d",
-		nblks, reason, subtid);
-	if (reason == EXT4_FC_REASON_FC_FAILED)
-		return jbd2_fc_end_commit_fallback(journal);
-	if (reason == EXT4_FC_REASON_FC_START_FAILED ||
-		reason == EXT4_FC_REASON_INELIGIBLE)
-		return jbd2_complete_transaction(journal, commit_tid);
-	return 0;
+	commit_time = ktime_to_ns(ktime_sub(ktime_get(), start_time));
+	ext4_fc_update_stats(sb, status, commit_time, nblks, commit_tid);
+	return ret;
+
+fallback:
+	ret = jbd2_fc_end_commit_fallback(journal);
+	ext4_fc_update_stats(sb, status, 0, 0, commit_tid);
+	return ret;
 }
 
 /*
  * Fast commit cleanup routine. This is called after every fast commit and
  * full commit. full is true if we are called after a full commit.
  */
-static void ext4_fc_cleanup(journal_t *journal, int full)
+static void ext4_fc_cleanup(journal_t *journal, int full, tid_t tid)
 {
 	struct super_block *sb = journal->j_private;
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
@@ -1232,6 +1283,7 @@ static void ext4_fc_cleanup(journal_t *journal, int full)
 	if (full && sbi->s_fc_bh)
 		sbi->s_fc_bh = NULL;
 
+	trace_ext4_fc_cleanup(journal, full, tid);
 	jbd2_fc_release_bufs(journal);
 
 	spin_lock(&sbi->s_fc_lock);
@@ -1240,7 +1292,8 @@ static void ext4_fc_cleanup(journal_t *journal, int full)
 		list_del_init(&iter->i_fc_list);
 		ext4_clear_inode_state(&iter->vfs_inode,
 				       EXT4_STATE_FC_COMMITTING);
-		ext4_fc_reset_inode(&iter->vfs_inode);
+		if (iter->i_sync_tid <= tid)
+			ext4_fc_reset_inode(&iter->vfs_inode);
 		/* Make sure EXT4_STATE_FC_COMMITTING bit is clear */
 		smp_mb();
 #if (BITS_PER_LONG < 64)
@@ -1255,6 +1308,7 @@ static void ext4_fc_cleanup(journal_t *journal, int full)
 					     struct ext4_fc_dentry_update,
 					     fcd_list);
 		list_del_init(&fc_dentry->fcd_list);
+		list_del_init(&fc_dentry->fcd_dilist);
 		spin_unlock(&sbi->s_fc_lock);
 
 		if (fc_dentry->fcd_name.name &&
@@ -1269,8 +1323,10 @@ static void ext4_fc_cleanup(journal_t *journal, int full)
 	list_splice_init(&sbi->s_fc_q[FC_Q_STAGING],
 				&sbi->s_fc_q[FC_Q_MAIN]);
 
-	ext4_clear_mount_flag(sb, EXT4_MF_FC_COMMITTING);
-	ext4_clear_mount_flag(sb, EXT4_MF_FC_INELIGIBLE);
+	if (tid >= sbi->s_fc_ineligible_tid) {
+		sbi->s_fc_ineligible_tid = 0;
+		ext4_clear_mount_flag(sb, EXT4_MF_FC_INELIGIBLE);
+	}
 
 	if (full)
 		sbi->s_fc_bytes = 0;
@@ -1435,14 +1491,15 @@ static int ext4_fc_record_modified_inode(struct super_block *sb, int ino)
 		if (state->fc_modified_inodes[i] == ino)
 			return 0;
 	if (state->fc_modified_inodes_used == state->fc_modified_inodes_size) {
-		state->fc_modified_inodes_size +=
-			EXT4_FC_REPLAY_REALLOC_INCREMENT;
 		state->fc_modified_inodes = krealloc(
-					state->fc_modified_inodes, sizeof(int) *
-					state->fc_modified_inodes_size,
-					GFP_KERNEL);
+				state->fc_modified_inodes,
+				sizeof(int) * (state->fc_modified_inodes_size +
+				EXT4_FC_REPLAY_REALLOC_INCREMENT),
+				GFP_KERNEL);
 		if (!state->fc_modified_inodes)
 			return -ENOMEM;
+		state->fc_modified_inodes_size +=
+			EXT4_FC_REPLAY_REALLOC_INCREMENT;
 	}
 	state->fc_modified_inodes[state->fc_modified_inodes_used++] = ino;
 	return 0;
@@ -1474,7 +1531,9 @@ static int ext4_fc_replay_inode(struct super_block *sb, struct ext4_fc_tl *tl,
 	}
 	inode = NULL;
 
-	ext4_fc_record_modified_inode(sb, ino);
+	ret = ext4_fc_record_modified_inode(sb, ino);
+	if (ret)
+		goto out;
 
 	raw_fc_inode = (struct ext4_inode *)
 		(val + offsetof(struct ext4_fc_inode, fc_raw_inode));
@@ -1606,16 +1665,23 @@ out:
 }
 
 /*
- * Record physical disk regions which are in use as per fast commit area. Our
- * simple replay phase allocator excludes these regions from allocation.
+ * Record physical disk regions which are in use as per fast commit area,
+ * and used by inodes during replay phase. Our simple replay phase
+ * allocator excludes these regions from allocation.
  */
-static int ext4_fc_record_regions(struct super_block *sb, int ino,
-		ext4_lblk_t lblk, ext4_fsblk_t pblk, int len)
+int ext4_fc_record_regions(struct super_block *sb, int ino,
+		ext4_lblk_t lblk, ext4_fsblk_t pblk, int len, int replay)
 {
 	struct ext4_fc_replay_state *state;
 	struct ext4_fc_alloc_region *region;
 
 	state = &EXT4_SB(sb)->s_fc_replay_state;
+	/*
+	 * during replay phase, the fc_regions_valid may not same as
+	 * fc_regions_used, update it when do new additions.
+	 */
+	if (replay && state->fc_regions_used != state->fc_regions_valid)
+		state->fc_regions_used = state->fc_regions_valid;
 	if (state->fc_regions_used == state->fc_regions_size) {
 		state->fc_regions_size +=
 			EXT4_FC_REPLAY_REALLOC_INCREMENT;
@@ -1633,6 +1699,9 @@ static int ext4_fc_record_regions(struct super_block *sb, int ino,
 	region->pblk = pblk;
 	region->len = len;
 
+	if (replay)
+		state->fc_regions_valid++;
+
 	return 0;
 }
 
@@ -1664,6 +1733,8 @@ static int ext4_fc_replay_add_range(struct super_block *sb,
 	}
 
 	ret = ext4_fc_record_modified_inode(sb, inode->i_ino);
+	if (ret)
+		goto out;
 
 	start = le32_to_cpu(ex->ee_block);
 	start_pblk = ext4_ext_pblock(ex);
@@ -1681,18 +1752,14 @@ static int ext4_fc_replay_add_range(struct super_block *sb,
 		map.m_pblk = 0;
 		ret = ext4_map_blocks(NULL, inode, &map, 0);
 
-		if (ret < 0) {
-			iput(inode);
-			return 0;
-		}
+		if (ret < 0)
+			goto out;
 
 		if (ret == 0) {
 			/* Range is not mapped */
 			path = ext4_find_extent(inode, cur, NULL, 0);
-			if (IS_ERR(path)) {
-				iput(inode);
-				return 0;
-			}
+			if (IS_ERR(path))
+				goto out;
 			memset(&newex, 0, sizeof(newex));
 			newex.ee_block = cpu_to_le32(cur);
 			ext4_ext_store_pblock(
@@ -1706,10 +1773,8 @@ static int ext4_fc_replay_add_range(struct super_block *sb,
 			up_write((&EXT4_I(inode)->i_data_sem));
 			ext4_ext_drop_refs(path);
 			kfree(path);
-			if (ret) {
-				iput(inode);
-				return 0;
-			}
+			if (ret)
+				goto out;
 			goto next;
 		}
 
@@ -1722,10 +1787,8 @@ static int ext4_fc_replay_add_range(struct super_block *sb,
 			ret = ext4_ext_replay_update_ex(inode, cur, map.m_len,
 					ext4_ext_is_unwritten(ex),
 					start_pblk + cur - start);
-			if (ret) {
-				iput(inode);
-				return 0;
-			}
+			if (ret)
+				goto out;
 			/*
 			 * Mark the old blocks as free since they aren't used
 			 * anymore. We maintain an array of all the modified
@@ -1745,10 +1808,8 @@ static int ext4_fc_replay_add_range(struct super_block *sb,
 			ext4_ext_is_unwritten(ex), map.m_pblk);
 		ret = ext4_ext_replay_update_ex(inode, cur, map.m_len,
 					ext4_ext_is_unwritten(ex), map.m_pblk);
-		if (ret) {
-			iput(inode);
-			return 0;
-		}
+		if (ret)
+			goto out;
 		/*
 		 * We may have split the extent tree while toggling the state.
 		 * Try to shrink the extent tree now.
@@ -1760,6 +1821,7 @@ next:
 	}
 	ext4_ext_replay_shrink_inode(inode, i_size_read(inode) >>
 					sb->s_blocksize_bits);
+out:
 	iput(inode);
 	return 0;
 }
@@ -1789,6 +1851,8 @@ ext4_fc_replay_del_range(struct super_block *sb, struct ext4_fc_tl *tl,
 	}
 
 	ret = ext4_fc_record_modified_inode(sb, inode->i_ino);
+	if (ret)
+		goto out;
 
 	jbd_debug(1, "DEL_RANGE, inode %ld, lblk %d, len %d\n",
 			inode->i_ino, le32_to_cpu(lrange.fc_lblk),
@@ -1798,10 +1862,8 @@ ext4_fc_replay_del_range(struct super_block *sb, struct ext4_fc_tl *tl,
 		map.m_len = remaining;
 
 		ret = ext4_map_blocks(NULL, inode, &map, 0);
-		if (ret < 0) {
-			iput(inode);
-			return 0;
-		}
+		if (ret < 0)
+			goto out;
 		if (ret > 0) {
 			remaining -= ret;
 			cur += ret;
@@ -1812,16 +1874,18 @@ ext4_fc_replay_del_range(struct super_block *sb, struct ext4_fc_tl *tl,
 		}
 	}
 
-	ret = ext4_punch_hole(inode,
-		le32_to_cpu(lrange.fc_lblk) << sb->s_blocksize_bits,
-		le32_to_cpu(lrange.fc_len) <<  sb->s_blocksize_bits);
+	down_write(&EXT4_I(inode)->i_data_sem);
+	ret = ext4_ext_remove_space(inode, le32_to_cpu(lrange.fc_lblk),
+				le32_to_cpu(lrange.fc_lblk) +
+				le32_to_cpu(lrange.fc_len) - 1);
+	up_write(&EXT4_I(inode)->i_data_sem);
 	if (ret)
-		jbd_debug(1, "ext4_punch_hole returned %d", ret);
+		goto out;
 	ext4_ext_replay_shrink_inode(inode,
 		i_size_read(inode) >> sb->s_blocksize_bits);
 	ext4_mark_inode_dirty(NULL, inode);
+out:
 	iput(inode);
-
 	return 0;
 }
 
@@ -1892,8 +1956,8 @@ bool ext4_fc_replay_check_excluded(struct super_block *sb, ext4_fsblk_t blk)
 		if (state->fc_regions[i].ino == 0 ||
 			state->fc_regions[i].len == 0)
 			continue;
-		if (blk >= state->fc_regions[i].pblk &&
-		    blk < state->fc_regions[i].pblk + state->fc_regions[i].len)
+		if (in_range(blk, state->fc_regions[i].pblk,
+					state->fc_regions[i].len))
 			return true;
 	}
 	return false;
@@ -1977,7 +2041,7 @@ static int ext4_fc_replay_scan(journal_t *journal,
 			ret = ext4_fc_record_regions(sb,
 				le32_to_cpu(ext.fc_ino),
 				le32_to_cpu(ex->ee_block), ext4_ext_pblock(ex),
-				ext4_ext_get_actual_len(ex));
+				ext4_ext_get_actual_len(ex), 0);
 			if (ret < 0)
 				break;
 			ret = JBD2_FC_REPLAY_CONTINUE;
@@ -2173,7 +2237,7 @@ int ext4_fc_info_show(struct seq_file *seq, void *v)
 		"fc stats:\n%ld commits\n%ld ineligible\n%ld numblks\n%lluus avg_commit_time\n",
 		   stats->fc_num_commits, stats->fc_ineligible_commits,
 		   stats->fc_numblks,
-		   div_u64(sbi->s_fc_avg_commit_time, 1000));
+		   div_u64(stats->s_fc_avg_commit_time, 1000));
 	seq_puts(seq, "Ineligible reasons:\n");
 	for (i = 0; i < EXT4_FC_REASON_MAX; i++)
 		seq_printf(seq, "\"%s\":\t%d\n", fc_ineligible_reasons[i],
@@ -2192,3 +2256,8 @@ int __init ext4_fc_init_dentry_cache(void)
 
 	return 0;
 }
+
+void ext4_fc_destroy_dentry_cache(void)
+{
+	kmem_cache_destroy(ext4_fc_dentry_cachep);
+}