Skip to contentSkip to conversation

ext[34] orphan list check on destroy_inode

PATCH RESEND In review Vasily Averin 2 participants 2 messages Jun 5, 2007 ext4 file system

lore
Share & tools
lore

Vasily Averin

author
#0 lore raw

PATCH RESEND 1/1 ext[34] orphan list check on destroy_inode

Customers claims to ext3-related errors, investigation showed that ext3 orphan list has been corrupted and have the reference to non-ext3 inode. The following debug helps to understand the reasons of this issue.

Signed-off-by Vasily Averin <[email protected]>
fs/ext3/super.c +7 -0
index 6e30629..e986376 100644
@@ -459,6 +459,13 @@ static struct inode *ext3_alloc_inode(struct super_block *sb)
 
 static void ext3_destroy_inode(struct inode *inode)
 {
+	if (!list_empty(&(EXT3_I(inode)->i_orphan))) {
+		printk("EXT3 Inode %p: orphan list check failed!\n",
+			EXT3_I(inode));
+		print_hex_dump(KERN_INFO, DUMP_PREFIX_ADDRESS,
+				EXT3_I(inode), sizeof(struct ext3_inode_info));
+		dump_stack();
+	}
 	kmem_cache_free(ext3_inode_cachep, EXT3_I(inode));
 }
fs/ext4/super.c +7 -0
index cb9afdd..935420d 100644
@@ -510,6 +510,13 @@ static struct inode *ext4_alloc_inode(struct super_block *sb)
 
 static void ext4_destroy_inode(struct inode *inode)
 {
+	if (!list_empty(&(EXT4_I(inode)->i_orphan))) {
+		printk("EXT4 Inode %p: orphan list check failed!\n",
+			EXT4_I(inode));
+		print_hex_dump(KERN_INFO, DUMP_PREFIX_ADDRESS,
+				EXT4_I(inode), sizeof(struct ext4_inode_info));
+		dump_stack();
+	}
 	kmem_cache_free(ext4_inode_cachep, EXT4_I(inode));
 }
#1 lore raw
Quoted 3 lines · Vasily Averin

Customers claims to ext3-related errors, investigation showed that ext3 orphan list has been corrupted and have the reference to non-ext3 inode. The following debug helps to understand the reasons of this issue.

This looks like it might be related to the -as far as I recall- unfixed issue we've noticed some time ago: http://osdir.com/ml/file-systems/2004-02/msg00033.html

I have'nt heard any reports of a reproduction on modern kernels on 390 lately, our issue seems to have disappeared.

so long,
Carsten