Am Dienstag, 26. Mai 2020, 05:07:15 CEST schrieb Bhat, Jayalakshmi Manjunath: Hi Jayalakshmi, > Hi Stephen, > > I to add the backend support using libkcapi APIs to exercise Kernel CAVP. > Can you please confirm if my understanding is correct? You would need to implement an equivalent to backend_openssl.c or backend_nettle.c > > Regards, > Jaya > > From: linux-crypto-owner@xxxxxxxxxxxxxxx > <linux-crypto-owner@xxxxxxxxxxxxxxx> On Behalf Of Stephan Müller Sent: > Sunday, May 24, 2020 12:14 AM > To: Bhat, Jayalakshmi Manjunath <jayalakshmi.bhat@xxxxxx>; Ard Biesheuvel > <ardb@xxxxxxxxxx> Cc: linux-crypto@xxxxxxxxxxxxxxx > Subject: Re: Monte Carlo Test (MCT) for AES > > Am Samstag, 23. Mai 2020, 00:11:35 CEST schrieb Ard Biesheuvel: > > Hi Ard, > > > (+ Stephan) > > > > On Fri, 22 May 2020 at 05:20, Bhat, Jayalakshmi Manjunath > > > > <mailto:jayalakshmi.bhat@xxxxxx> wrote: > > > Hi All, > > > > > > We are using libkcapi for CAVS vectors verification on our Linux kernel. > > > Our Linux kernel version is 4.14. Monte Carlo Test (MCT) for SHA worked > > > fine using libkcapi. We are trying to perform Monte Carlo Test (MCT) for > > > AES using libkcapi. We not able to get the result successfully. Is it > > > possible to use libkcapi to achieve AES MCT? > > Yes, it is possible. I have the ACVP testing implemented completely for AES > (ECB, CBC, CFB8, CFB128, CTR, XTS, GCM internal and external IV generation, > CCM), TDES (ECB, CTR, CBC), SHA, HMAC, CMAC (AES and TDES). I did not yet > try TDES CFB8 and CFB64 through, but it should work out of the box. > > AES-KW is the only one that cannot be tested through libkcapi as AF_ALG has > one shortcoming preventing this test. > > The testing is implemented with [1] but the libkcapi test backend is not > public. The public code in [1] already implements the MCT. So, if you want > to use [1], all you need to implement is a libkcapi backend that just > invokes the ciphers as defined by the API in [1]. > > [1] https://github.com/smuellerDD/acvpparser > > Ciao > Stephan Ciao Stephan