Skip to contentSkip to conversation

Ocfs2: Re-access the journal after ocfs2_insert_extent() in dxdir codes.

PATCH New Tristan Ye 3 participants 5 messages Sep 8, 2010 – Sep 10, 2010 OCFS2

lore
Share & tools
lore

Tristan Ye

author
#0 lore raw

PATCH 1/1 Ocfs2: Re-access the journal after ocfs2_insert_extent() in dxdir codes.

In ocfs2_dx_dir_rebalance(), we need to rejournal_acess the blocks after calling ocfs2_insert_extent() since growing an extent tree may trigger ocfs2_extend_trans(), which makes previous journal_access meaningless.

---
 fs/ocfs2/dir.c |   24 ++++++++++++++++--------
 1 files changed, 16 insertions(+), 8 deletions(-)
fs/ocfs2/dir.c +16 -8
index f04ebcf..c49f6de 100644
@@ -3931,6 +3931,15 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir,
 		goto out_commit;
 	}
 
+	cpos = split_hash;
+	ret = ocfs2_dx_dir_new_cluster(dir, &et, cpos, handle,
+				       data_ac, meta_ac, new_dx_leaves,
+				       num_dx_leaves);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
 	for (i = 0; i < num_dx_leaves; i++) {
 		ret = ocfs2_journal_access_dl(handle, INODE_CACHE(dir),
 					      orig_dx_leaves[i],
@@ -3939,15 +3948,14 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir,
 			mlog_errno(ret);
 			goto out_commit;
 		}
-	}
 
-	cpos = split_hash;
-	ret = ocfs2_dx_dir_new_cluster(dir, &et, cpos, handle,
-				       data_ac, meta_ac, new_dx_leaves,
-				       num_dx_leaves);
-	if (ret) {
-		mlog_errno(ret);
-		goto out_commit;
+		ret = ocfs2_journal_access_dl(handle, INODE_CACHE(dir),
+					      new_dx_leaves[i],
+					      OCFS2_JOURNAL_ACCESS_WRITE);
+		if (ret) {
+			mlog_errno(ret);
+			goto out_commit;
+		}
 	}
 
 	ocfs2_dx_dir_transfer_leaf(dir, split_hash, handle, tmp_dx_leaf,
Signature · 1 line
1.5.5
#1 lore raw

PATCH 1/1 Ocfs2: Re-access the journal after ocfs2_insert_extent() in dxdir codes.

Quoted 19 lines · Tristan Ye

In ocfs2_dx_dir_rebalance(), we need to rejournal_acess the blocks after calling ocfs2_insert_extent() since growing an extent tree may trigger ocfs2_extend_trans(), which makes previous journal_access meaningless.

---
 fs/ocfs2/dir.c |   24 ++++++++++++++++--------
 1 files changed, 16 insertions(+), 8 deletions(-)
fs/ocfs2/dir.c +2 -0
index f04ebcf..c49f6de 100644
@@ -3931,6 +3931,15 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir,
 		goto out_commit;
 	}
 
+	cpos = split_hash;
+	ret = ocfs2_dx_dir_new_cluster(dir, &et, cpos, handle,

fs/ocfs2/dir.c static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir, · line 3945

Quoted 10 lines
fs/ocfs2/dir.c +7 -0
+				       data_ac, meta_ac, new_dx_leaves,
+				       num_dx_leaves);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
 	for (i = 0; i < num_dx_leaves; i++) {
 		ret = ocfs2_journal_access_dl(handle, INODE_CACHE(dir),
 					      orig_dx_leaves[i],

Ok, I see what you're going for here. Don't we still need to add a journal_access_dl call against new_dx_leaves[i] in the for loop though? They will be dirtied again in ocfs2_dx_dir_transfer_leaf.

	--Mark

--
Mark Fasheh

author
#2 lore raw

PATCH 1/1 Ocfs2: Re-access the journal after ocfs2_insert_extent() in dxdir codes.

Quoted 18 lines · Mark Fasheh
Quoted 17 lines · Tristan Ye

In ocfs2_dx_dir_rebalance(), we need to rejournal_acess the blocks after calling ocfs2_insert_extent() since growing an extent tree may trigger ocfs2_extend_trans(), which makes previous journal_access meaningless.

---
 fs/ocfs2/dir.c |   24 ++++++++++++++++--------
 1 files changed, 16 insertions(+), 8 deletions(-)
fs/ocfs2/dir.c +0 -0
index f04ebcf..c49f6de 100644
@@ -3931,6 +3931,15 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir,
 		goto out_commit;
 	}
 
Quoted 16 lines

fs/ocfs2/dir.c static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir, · line 3945

Quoted 12 lines
fs/ocfs2/dir.c +9 -0
+	cpos = split_hash;
+	ret = ocfs2_dx_dir_new_cluster(dir, &et, cpos, handle,
+				       data_ac, meta_ac, new_dx_leaves,
+				       num_dx_leaves);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
 	for (i = 0; i < num_dx_leaves; i++) {
 		ret = ocfs2_journal_access_dl(handle, INODE_CACHE(dir),
 					      orig_dx_leaves[i],

Ok, I see what you're going for here. Don't we still need to add a journal_access_dl call against new_dx_leaves[i] in the for loop though? They will be dirtied again in ocfs2_dx_dir_transfer_leaf.

Didn't I?

This patch already had the codes calling the journal_access_dl against new_dx_leaves.

Look at the second portion of this patch, that is it.

Quoted 4 lines
	--Mark

--
Mark Fasheh

#3 lore raw

PATCH 1/1 Ocfs2: Re-access the journal after ocfs2_insert_extent() in dxdir codes.

Quoted 36 lines · tristan
Quoted 18 lines · Mark Fasheh
Quoted 17 lines · Tristan Ye

In ocfs2_dx_dir_rebalance(), we need to rejournal_acess the blocks after calling ocfs2_insert_extent() since growing an extent tree may trigger ocfs2_extend_trans(), which makes previous journal_access meaningless.

---
 fs/ocfs2/dir.c |   24 ++++++++++++++++--------
 1 files changed, 16 insertions(+), 8 deletions(-)
fs/ocfs2/dir.c +0 -0
index f04ebcf..c49f6de 100644
@@ -3931,6 +3931,15 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir,
 		goto out_commit;
 	}
 
Quoted 16 lines

fs/ocfs2/dir.c static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir, · line 3945

Quoted 12 lines
fs/ocfs2/dir.c +9 -0
+	cpos = split_hash;
+	ret = ocfs2_dx_dir_new_cluster(dir, &et, cpos, handle,
+				       data_ac, meta_ac, new_dx_leaves,
+				       num_dx_leaves);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
 	for (i = 0; i < num_dx_leaves; i++) {
 		ret = ocfs2_journal_access_dl(handle, INODE_CACHE(dir),
 					      orig_dx_leaves[i],

Ok, I see what you're going for here. Don't we still need to add a journal_access_dl call against new_dx_leaves[i] in the for loop though? They will be dirtied again in ocfs2_dx_dir_transfer_leaf.

Quoted 6 lines

Didn't I?

This patch already had the codes calling the journal_access_dl against new_dx_leaves.

Look at the second portion of this patch, that is it.

Ahh yes, you are correct. Sorry about that, not sure how I missed it. Anyway, you can add my ack:

The patch looks good to me.

	--Mark 

--
Mark Fasheh

#4 lore raw

PATCH 1/1 Ocfs2: Re-access the journal after ocfs2_insert_extent() in dxdir codes.

In reply to Tristan Ye
Quoted 5 lines · Tristan Ye

In ocfs2_dx_dir_rebalance(), we need to rejournal_acess the blocks after calling ocfs2_insert_extent() since growing an extent tree may trigger ocfs2_extend_trans(), which makes previous journal_access meaningless.

This patch is now in the fixes branch of ocfs2.git.

Joel

Signature · 9 lines
Life's Little Instruction Book #173

	"Be kinder than necessary."

Joel Becker
Consulting Software Developer
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127