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>

1 files changed, 2 insertions, 2 deletions

diff --git a/fs/ecryptfs/crypto.c b/fs/ecryptfs/crypto.c
index b8d5c8091024..f48c4987a15c 100644
--- a/fs/ecryptfs/crypto.c
+++ b/fs/ecryptfs/crypto.c
@@ -1568,11 +1568,11 @@ out:
 }
 
 int ecryptfs_read_and_validate_xattr_region(char *page_virt,
-                                            struct dentry *ecryptfs_dentry)
+                                            struct inode *inode)
 {
         int rc;
 
-        rc = ecryptfs_read_xattr_region(page_virt, ecryptfs_dentry->d_inode);
+        rc = ecryptfs_read_xattr_region(page_virt, inode);
         if (rc)
                 goto out;
         if (!contains_ecryptfs_marker(page_virt + ECRYPTFS_FILE_SIZE_BYTES)) {