Re: [PATCH net-next] crypto: af_alg/hash: Fix recvmsg() after sendmsg(MSG_MORE)

[Herbert Xu]

On Thu, Jun 15, 2023 at 12:27:53AM +0100, David Howells wrote:
>     
> If an AF_ALG socket bound to a hashing algorithm is sent a zero-length
> message with MSG_MORE set and then recvmsg() is called without first
> sending another message without MSG_MORE set to end the operation, an oops
> will occur because the crypto context and result doesn't now get set up in
> advance because hash_sendmsg() now defers that as long as possible in the
> hope that it can use crypto_ahash_digest() - and then because the message
> is zero-length, it the data wrangling loop is skipped.
> 
> Fix this by always making a pass of the loop, even in the case that no data
> is provided to the sendmsg().
> 
> Fix also extract_iter_to_sg() to handle a zero-length iterator by returning
> 0 immediately.
> 
> Whilst we're at it, remove the code to create a kvmalloc'd scatterlist if
> we get more than ALG_MAX_PAGES - this shouldn't happen.

I don't think this is right.  If it's a zero-length message with
MSG_MORE set, it should be ignored until a recvmsg(2) call is made.

In any case, this patch doesn't fix all the syzbot reports.

We need to think about reverting this change if it can't be fixed
in time.

Thanks,
-- 
Email: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt