eCryptfs: Fix new inode race condition

Only unlock and d_add() new inodes after the plaintext inode size has
been read from the lower filesystem. This fixes a race condition that
was sometimes seen during a multi-job kernel build in an eCryptfs mount.

https://bugzilla.kernel.org/show_bug.cgi?id=36002

Signed-off-by: Tyler Hicks <tyhicks@linux.vnet.ibm.com>
Reported-by: David <david@unsolicited.net>
Tested-by: David <david@unsolicited.net>
This commit is contained in:
Tyler Hicks
2011-05-24 03:49:02 -05:00
parent 5ccf92037c
commit 3b06b3ebf4
5 changed files with 30 additions and 28 deletions

View File

@@ -191,7 +191,7 @@ static int ecryptfs_open(struct inode *inode, struct file *file)
| ECRYPTFS_ENCRYPTED);
}
mutex_unlock(&crypt_stat->cs_mutex);
rc = ecryptfs_get_lower_file(ecryptfs_dentry);
rc = ecryptfs_get_lower_file(ecryptfs_dentry, inode);
if (rc) {
printk(KERN_ERR "%s: Error attempting to initialize "
"the lower file for the dentry with name "