Darren J Moffat wrote: >Mark Powers wrote: > > >>Folks, >> >>So far, Garrett D'Amore has posted review comments - thanks Garrett! >>I still need one more reviewer. >> >>6534628 /dev/crypto can use in-place when calling crypto_encrypt API >> >>The updated webrev is >>http://cr.grommit.com/~mcpowers/6534628/ >> >> > >General comment, this causes a Cap-Eye Install flag day, but bfu is >okay. If this is backproted to S10 then the same patch must include >drv/crypto and pkcs11_kernel binaries since this changes the API and ABI >between them. > >I like the approach of pkcs11_kernel telling drv/crypto this using a new >ce_flags field. > >The one thing I don't really like is that INPLACE_MECHANISM is a >hardcoded macro. I wish there was something we could do to make it not >so hardcoded. My concern is that was we add new modes and mechanisms we >will forget to update this macro and won't get the benefits of this fix. > However I can't think of a better way of doing it at the moment so I'm >okay with you integrating as is. > > > I agree - but I'm already wondering why RC4 is not included as well as AES CCM and even AES CTR?
thanks, -gary
