[<prev] [next>] [day] [month] [year] [list]
Message-id: <013301d16fb1$302a2d20$907e8760$@samsung.com>
Date: Thu, 25 Feb 2016 17:44:01 +0800
From: Chao Yu <chao2.yu@...sung.com>
To: Jaegeuk Kim <jaegeuk@...nel.org>
Cc: linux-f2fs-devel@...ts.sourceforge.net,
linux-kernel@...r.kernel.org
Subject: [PATCH 2/2] f2fs: update inode page under cp_rwsem lock
We should update inode page under protection of cp_rwsem lock, otherwise
new dirty inode pages will generated by ->fsync or ->write_inode during
checkpoint, which should not be allowed.
Signed-off-by: Chao Yu <chao2.yu@...sung.com>
---
fs/f2fs/inode.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index cb269c4..cdfe114 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -295,6 +295,7 @@ retry:
int f2fs_write_inode(struct inode *inode, struct writeback_control *wbc)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+ bool updated;
if (inode->i_ino == F2FS_NODE_INO(sbi) ||
inode->i_ino == F2FS_META_INO(sbi))
@@ -307,7 +308,11 @@ int f2fs_write_inode(struct inode *inode, struct writeback_control *wbc)
* We need to balance fs here to prevent from producing dirty node pages
* during the urgent cleaning time when runing out of free sections.
*/
- if (update_inode_page(inode))
+ f2fs_lock_op(sbi);
+ updated = update_inode_page(inode);
+ f2fs_unlock_op(sbi);
+
+ if (updated)
f2fs_balance_fs(sbi, true);
return 0;
}
--
2.7.0
Powered by blists - more mailing lists