diff mbox series

[RFC] ext4: Error handling for ext4_get_inode_loc()

Message ID 20250505140057.671758-1-richard120310@gmail.com
State New
Headers show
Series [RFC] ext4: Error handling for ext4_get_inode_loc() | expand

Commit Message

I Hsin Cheng May 5, 2025, 2 p.m. UTC 
The use of "ext4_get_inode_loc()" inside
"ext4_xattr_inode_dec_ref_all()" lacks of proper error handling. If it
can't find correct "loc" for "parent" inode, the value of "end" can't be
guaranteed as valid.

Link: https://scan5.scan.coverity.com/#/project-view/36179/10063?selectedIssue=1645904
Signed-off-by: I Hsin Cheng <richard120310@gmail.com>
---
I wonder if there's an default value for "end" ? so we can still loop
through certain amount of inode without return directly from the error.
If we use "iloc" from with ext4_get_inode_loc() return an error, it
might corrupt some inodes it shouldn't be able to.

Wish to know what's the correct and thorough error handling here, I'll
make the change if it's possible. Or it's guaranteed that the error
never occurs?

Best regards,
I Hsin Cheng
---
 fs/ext4/xattr.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
diff mbox series

Patch

diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c
index 7ab8f2e8e815..416500ae5248 100644
--- a/fs/ext4/xattr.c
+++ b/fs/ext4/xattr.c
@@ -1169,7 +1169,11 @@  ext4_xattr_inode_dec_ref_all(handle_t *handle, struct inode *parent,
 	if (block_csum)
 		end = (void *)bh->b_data + bh->b_size;
 	else {
-		ext4_get_inode_loc(parent, &iloc);
+		err = ext4_get_inode_loc(parent, &iloc);
+		if (err) {
+			EXT4_ERROR_INODE(parent, "inode loc (error %d)", err);
+			return;
+		}
 		end = (void *)ext4_raw_inode(&iloc) + EXT4_SB(parent->i_sb)->s_inode_size;
 	}