Pārlūkot izejas kodu

fs/9p: Make sure we properly instantiate dentry.

For lookup if we get ENOENT error from the server we still
instantiate the dentry. We need to make sure we have dentry
operations set in that case so that a later dput on the dentry
does the expected. Without the patch we get the below error

#ln  -sf abc abclink
ln: creating symbolic link `abclink': No such file or directory

Now on the host do
$ touch abclink

Guest now gives ENOENT error.
# ls
ls: cannot access abclink: No such file or directory

Debugged-by:Gautham R Shenoy <ego@in.ibm.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Aneesh Kumar K.V 15 gadi atpakaļ
vecāks
revīzija
85e0df240e
1 mainītis faili ar 4 papildinājumiem un 3 dzēšanām
  1. 4 3
      fs/9p/vfs_inode.c

+ 4 - 3
fs/9p/vfs_inode.c

@@ -697,8 +697,8 @@ static struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
 	if (IS_ERR(fid)) {
 	if (IS_ERR(fid)) {
 		result = PTR_ERR(fid);
 		result = PTR_ERR(fid);
 		if (result == -ENOENT) {
 		if (result == -ENOENT) {
-			d_add(dentry, NULL);
-			return NULL;
+			inode = NULL;
+			goto inst_out;
 		}
 		}
 
 
 		return ERR_PTR(result);
 		return ERR_PTR(result);
@@ -715,7 +715,8 @@ static struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
 	if (result < 0)
 	if (result < 0)
 		goto error;
 		goto error;
 
 
-	if ((fid->qid.version) && (v9ses->cache))
+inst_out:
+	if (v9ses->cache)
 		dentry->d_op = &v9fs_cached_dentry_operations;
 		dentry->d_op = &v9fs_cached_dentry_operations;
 	else
 	else
 		dentry->d_op = &v9fs_dentry_operations;
 		dentry->d_op = &v9fs_dentry_operations;