Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx> wrote: > > [!] Note that the net/sunrpc/auth_gss/ implementation gets a pair of > > ciphers, one non-CTS and one CTS, using the former to do all the aligned > > blocks and the latter to do the last two blocks if they aren't also > > aligned. It may be necessary to do this here too for performance reasons - > > but there are considerations both ways: > > The CTS template will take any hardware accelerated CBC implementation > and turn it into CTS. > > So there is no reason to do the CTS/CBC thing by hand at all. Glad to hear it. I'm just reporting what net/sunrpc/ does now. My suspicion is that this is from before a lot of cpu crypto-based optimisations were made available in the crypto layer. David
