Re: [PATCH] crypto: caam - properly set IV after {en,de}crypt

[Gilad Ben-Yossef]

Hi,

On Thu, Jun 29, 2017 at 1:19 PM, Horia GeantÄ <horia.geanta@xxxxxxx> wrote:
> On 6/28/2017 4:42 PM, Horia GeantÄ wrote:
>> On 6/28/2017 4:27 PM, David Gstir wrote:
>>> Certain cipher modes like CTS expect the IV (req->info) of
>>> ablkcipher_request (or equivalently req->iv of skcipher_request) to
>>> contain the last ciphertext block when the {en,de}crypt operation is done.
>>> This is currently not the case for the CAAM driver which in turn breaks
>>> e.g. cts(cbc(aes)) when the CAAM driver is enabled.
>>>
>>> This patch fixes the CAAM driver to properly set the IV after the
>>> {en,de}crypt operation of ablkcipher finishes.
>>>
>>> This issue was revealed by the changes in the SW CTS mode in commit
>>> 0605c41cc53ca ("crypto: cts - Convert to skcipher")
>>>
>>> Cc: <stable@xxxxxxxxxxxxxxx> # 4.8+
>>> Signed-off-by: David Gstir <david@xxxxxxxxxxxxx>
>> Reviewed-by: Horia GeantÄ <horia.geanta@xxxxxxx>
>>
> Btw, instead of updating the IV in SW, CAAM engine could be programmed
> to do it - by saving the Context Register of the AES accelerator.
>
> Unfortunately this would require changes in quite a few places: shared
> descriptor, HW S/G generation logic, IV dma (un)mapping and maybe others.
>
> So it's better to have this fix now (which, considering size, is
> appropriate for -stable) and later, if needed, offload IV updating in HW.
>

My apologies for reviving this thread from the dead, but doesn't the patch fail
for in-place decryption since we are copying from req->dst after
the operation is done, and therefore it no longer contains the ciphertext?

I'm asking since I ran into a similar issue in the ccree driver and thought
to deploy a similar fix but could not convince myself why this works.

Thanks!
Gilad

-- 
Gilad Ben-Yossef
Chief Coffee Drinker

"If you take a class in large-scale robotics, can you end up in a
situation where the homework eats your dog?"
 -- Jean-Baptiste Queru