Hi Vladimir, > Any advice on what the micro benchmark should test ?
I will send them to you separately. There shouldn't be any performance impact with software providers only case after your latest change. So, the interesting part is how much is the impact when n2cp is present. I expect this to be insignificant. Best to confirm it though. >> The common case is that there are only software providers and we should >> avoid adding any extra latency for that case. > > Is the reason also that in SCF we assume that SW providers are always > a superset of HW providers in terms of supported mechanisms and key > lengths ? Darren answered this. I will add that all SW providers do key length checks and return CRYPTO_KEY_SIZE_RANGE. It does not matter to a caller of crypto_encrypt() who returned the error. > ... > I removed the check in the SW case from both kcf_get_mech_provider() > and kcf_get_dual_provider(). > > Incremental webrev is here: > > http://cr.opensolaris.org/~vkotal/kcf-keylen_check-6786946.onnv.Krishna/ Looks good. -Krishna
