Stephan,
On Wed, 9 Nov 2016, Stephan Mueller wrote:
Am Montag, 31. Oktober 2016, 16:18:32 CET schrieb Mat Martineau:
Hi Mat,
My main concern is getting the semantics correct and consistent in a
single patch series. It would be a big problem to explain that AF_ALG AEAD
read and write works one way in 4.x, another way in 4.y, and some
different way in 4.z.
I do have a patch now available that exactly does what you suggest. See the
patch attached. It works with the following exception.
In the case of sendpage and using an in-place cipher operation, the patch
breaks as follows. When the caller sends the same buffer for a sendpage
operation, the cipher operation now will write the ciphertext to the beginning
of the buffer where the AAD used to be. The subsequent tag calculation will
now use the data it finds where the AAD is expected. As the cipher operation
has already replaced the AAD with the ciphertext, the tag calculation will
take the ciphertext as AAD and thus calculate a wrong tag.
Thus, the only way to avoid that would be to duplicate the AAD into an
internal buffer. But that would defeat the entire purpose of sendpage.
The use case for an "in place" operation would be to have the ciphertext
overwrite the plaintext, correct? If the src and dst overlap, does it make
sense to require the plaintext and ciphertext to be in exactly the same
location?
The patch, however, works with sendmsg as well as sendpage when the src and
dst buffers are different.
Thanks - I tested your patch and found that it works as expected with
sendmsg.
Regards,
--
Mat Martineau
Intel OTC
--
To unsubscribe from this list: send the line "unsubscribe linux-crypto" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
