Skip to contentSkip to conversation

Missing spin_lock(&dcache_lock)

PATCH Leonardo Chiquitto 2 participants 2 messages Jan 11, 2010 – Jan 12, 2010 Kernel automounter

lore
Share & tools
lore
#0 lore raw

Hello,

If I compare the 2.6.18 and 2.6.20 kernel trees after applying, respectively, autofs4-2.6.18-v5-update-20090903.patch and autofs4-2.6.20-v5-update-20090903.patch, I can understand most of the differences, but this specific chunk seems strange to me:

fs/autofs4/root.c +1 -0
diff -Nurp linux-2.6.18.8/fs/autofs4/root.c linux-2.6.20.21/fs/autofs4/root.c
@@ -91,6 +91,7 @@ static int autofs4_dir_open(struct inode
 	 * autofs file system so just let the libfs routines handle
 	 * it.
 	 */
+	spin_lock(&dcache_lock);
 	if (!d_mountpoint(dentry) && __simple_empty(dentry)) {
 		spin_unlock(&dcache_lock);
 		return -ENOENT;

Was this line left out by mistake? If yes, can you please consider the patch below?

Thanks,
Leonardo

patches/autofs4-2.6.18-v5-update-20090903.patch +1 -0
Index: autofs/patches/autofs4-2.6.18-v5-update-20090903.patch
===================================================================
@@ -103,6 +103,7 @@
 +	 * autofs file system so just let the libfs routines handle
 +	 * it.
 +	 */
++	spin_lock(&dcache_lock);
 +	if (!d_mountpoint(dentry) && __simple_empty(dentry)) {
  		spin_unlock(&dcache_lock);
 -
patches/autofs4-2.6.19-v5-update-20090903.patch +1 -0
Index: autofs/patches/autofs4-2.6.19-v5-update-20090903.patch
===================================================================
@@ -1041,6 +1041,7 @@
 +	 * autofs file system so just let the libfs routines handle
 +	 * it.
 +	 */
++	spin_lock(&dcache_lock);
 +	if (!d_mountpoint(dentry) && __simple_empty(dentry)) {
  		spin_unlock(&dcache_lock);
 -
maintainer
#1 lore raw
Quoted 6 lines · Leonardo Chiquitto

Hello,

If I compare the 2.6.18 and 2.6.20 kernel trees after applying, respectively, autofs4-2.6.18-v5-update-20090903.patch and autofs4-2.6.20-v5-update-20090903.patch, I can understand most of the differences, but this specific chunk seems strange to me:

Oops.

I'll have a look but, from memory, that lock should be present.

The back porting of changes to older kernels gets harder and more error prone as time passes.

Quoted 48 lines

diff -Nurp linux-2.6.18.8/fs/autofs4/root.c linux-2.6.20.21/fs/autofs4/root.c

fs/autofs4/root.c +1 -0
@@ -91,6 +91,7 @@ static int autofs4_dir_open(struct inode
 	 * autofs file system so just let the libfs routines handle
 	 * it.
 	 */
+	spin_lock(&dcache_lock);
 	if (!d_mountpoint(dentry) && __simple_empty(dentry)) {
 		spin_unlock(&dcache_lock);
 		return -ENOENT;

Was this line left out by mistake? If yes, can you please consider the patch below?

Thanks,
Leonardo

Index: autofs/patches/autofs4-2.6.18-v5-update-20090903.patch
===================================================================

patches/autofs4-2.6.18-v5-update-20090903.patch +1 -0
@@ -103,6 +103,7 @@
 +	 * autofs file system so just let the libfs routines handle
 +	 * it.
 +	 */
++	spin_lock(&dcache_lock);
 +	if (!d_mountpoint(dentry) && __simple_empty(dentry)) {
  		spin_unlock(&dcache_lock);
 -

Index: autofs/patches/autofs4-2.6.19-v5-update-20090903.patch
===================================================================

patches/autofs4-2.6.19-v5-update-20090903.patch +1 -0
@@ -1041,6 +1041,7 @@
 +	 * autofs file system so just let the libfs routines handle
 +	 * it.
 +	 */
++	spin_lock(&dcache_lock);
 +	if (!d_mountpoint(dentry) && __simple_empty(dentry)) {
  		spin_unlock(&dcache_lock);
 -

_______________________________________________
autofs mailing list
[email protected]
http://linux.kernel.org/mailman/listinfo/autofs