On Fri, Jul 25, 2008 at 2:33 PM, Mike Snitzer [off-list ref] wrote:

On Fri, Jul 25, 2008 at 3:02 AM,  [off-list ref] wrote:

quoted

From: "Duane Griffin" <redacted>

If the orphan node list includes valid, untruncatable nodes with nlink > 0
the ext3_orphan_cleanup loop which attempts to delete them will not do so,
causing it to loop forever. Fix by checking for such nodes in the
ext3_orphan_get function.

This patch fixes the second case (image hdb.20000009.softlockup.gz)
reported in http://bugzilla.kernel.org/show_bug.cgi?id=10882.

[akpm@linux-foundation.org: coding-style fixes]
[akpm@linux-foundation.org: printk warning fix]
Signed-off-by: Duane Griffin <redacted>
Cc: <redacted>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

quoted

diff -puN fs/ext3/inode.c~ext3-handle-corrupted-orphan-list-at-mount fs/ext3/inode.c

--- a/fs/ext3/inode.c~ext3-handle-corrupted-orphan-list-at-mount
+++ a/fs/ext3/inode.c

@@ -2253,6 +2253,19 @@ static void ext3_free_branches(handle_t
       }
 }

+int ext3_can_truncate(struct inode *inode)
+{
+       if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
+               return 0;
+       if (S_ISREG(inode->i_mode))
+               return 1;
+       if (S_ISDIR(inode->i_mode))
+               return 1;
+       if (S_ISLNK(inode->i_mode))
+               return !ext3_inode_is_fast_symlink(inode);
+       return 0;
+}
+
 /*
 * ext3_truncate()
 *

@@ -2297,12 +2310,7 @@ void ext3_truncate(struct inode *inode)
       unsigned blocksize = inode->i_sb->s_blocksize;
       struct page *page;

-       if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
-           S_ISLNK(inode->i_mode)))
-               return;
-       if (ext3_inode_is_fast_symlink(inode))
-               return;
-       if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
+       if (!ext3_can_truncate(inode))
               return;

I may be missing something here but doesn't the above change the logic
that was used in ext3_truncate for the S_ISDIR and S_ISLNK cases?

Before ext3_truncate would be short-circuited if S_ISDIR or S_ISLNK,
now it won't... is that intended?

Gah, I mistakenly read it as only being !S_ISDIR (I missed the fact
that it was !S_ISREG, !S_ISDIR, or !S_ISLNK). And after Duane's change
that logic is maintained... and more readable!

Sorry for the noise.