[[PATCH] fix an endianness bug in ext4_encrypted_zeroout()

[[PATCH] fix an endianness bug in ext4_encrypted_zeroout()

[Al Viro]

ex->ee_block is not host-endian (note that accesses of other fields
of *ex right next to that line go through the helpers that do proper
conversion from little-endian to host-endian; it might make sense
to add similar for ->ee_block to avoid reintroducing that kind of
bugs...)

Cc: stable@vger.kernel.org # v4.1+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---

diff --git a/fs/ext4/crypto.c b/fs/ext4/crypto.c
index af06830..1a08350 100644
--- a/fs/ext4/crypto.c
+++ b/fs/ext4/crypto.c
@@ -389,7 +389,7 @@ int ext4_encrypted_zeroout(struct inode *inode, struct ext4_extent *ex)
 	struct ext4_crypto_ctx	*ctx;
 	struct page		*ciphertext_page = NULL;
 	struct bio		*bio;
-	ext4_lblk_t		lblk = ex->ee_block;
+	ext4_lblk_t		lblk = le32_to_cpu(ex->ee_block);
 	ext4_fsblk_t		pblk = ext4_ext_pblock(ex);
 	unsigned int		len = ext4_ext_get_actual_len(ex);
 	int			ret, err = 0;

Re: [[PATCH] fix an endianness bug in ext4_encrypted_zeroout()

[Theodore Ts'o]

On Thu, Nov 26, 2015 at 04:04:34PM +0000, Al Viro wrote:
> ex->ee_block is not host-endian (note that accesses of other fields
> of *ex right next to that line go through the helpers that do proper
> conversion from little-endian to host-endian; it might make sense
> to add similar for ->ee_block to avoid reintroducing that kind of
> bugs...)
> 
> Cc: stable@vger.kernel.org # v4.1+
> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

Thanks for catching this!   Applied.

							- Ted