[PATCH] AB-BA deadlock in drop_caches sysctl

[PATCH] AB-BA deadlock in drop_caches sysctl

[Denis V. Lunev]

There is a AB-BA deadlock regarding drop_caches sysctl. Here are the code
paths:

drop_pagecache
  spin_lock(&inode_lock);
  invalidate_mapping_pages
    try_to_release_page
      ext3_releasepage
        journal_try_to_free_buffers
          __journal_try_to_free_buffer
	    spin_lock(&journal->j_list_lock);

journal_commit_transaction
  spin_lock(&journal->j_list_lock);
  __journal_remove_checkpoint
    __journal_refile_buffer
      __journal_unfile_buffer
        __journal_temp_unlink_buffer
	  __set_page_dirty_nobuffers
	    __mark_inode_dirt
	    spin_lock(&inode_lock);

The patch tries to address the issue - it drops inode_lock before digging into
invalidate_inode_pages. This seems sane as inode hold should not gone from the
list and should not change its place.

Signed-off-by: Denis V. Lunev <den@openvz.org>

--- ./fs/drop_caches.c.marker	2006-09-20 07:42:06.000000000 +0400
+++ ./fs/drop_caches.c	2007-12-03 15:43:44.000000000 +0300
@@ -14,15 +14,27 @@ int sysctl_drop_caches;
 
 static void drop_pagecache_sb(struct super_block *sb)
 {
-	struct inode *inode;
+	struct inode *inode, *old;
 
+	old = NULL;
 	spin_lock(&inode_lock);
 	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
 		if (inode->i_state & (I_FREEING|I_WILL_FREE))
 			continue;
+		__iget(inode);
+		spin_unlock(&inode_lock);
+
 		invalidate_inode_pages(inode->i_mapping);
+		if (old != NULL)
+			iput(old);
+		old = inode;
+
+		spin_lock(&inode_lock);
 	}
 	spin_unlock(&inode_lock);
+
+	if (old != NULL)
+		iput(old);
 }
 
 void drop_pagecache(void)