> -----Original Message----- > From: Eric Biggers <ebiggers@xxxxxxxxxx> > Sent: Sunday, September 15, 2019 10:21 PM > To: Pascal Van Leeuwen <pvanleeuwen@xxxxxxxxxxxxxx> > Cc: Pascal van Leeuwen <pascalvanl@xxxxxxxxx>; linux-crypto@xxxxxxxxxxxxxxx; > antoine.tenart@xxxxxxxxxxx; herbert@xxxxxxxxxxxxxxxxxxx; davem@xxxxxxxxxxxxx > Subject: Re: [PATCH 4/7] crypto: testmgr - Added testvectors for the ofb(sm4) & cfb(sm4) > skciphers > > On Wed, Sep 11, 2019 at 07:34:31PM +0000, Pascal Van Leeuwen wrote: > > > -----Original Message----- > > > From: Eric Biggers <ebiggers@xxxxxxxxxx> > > > Sent: Wednesday, September 11, 2019 6:06 PM > > > To: Pascal van Leeuwen <pascalvanl@xxxxxxxxx> > > > Cc: linux-crypto@xxxxxxxxxxxxxxx; antoine.tenart@xxxxxxxxxxx; herbert@xxxxxxxxxxxxxxxxxxx; > > > davem@xxxxxxxxxxxxx; Pascal Van Leeuwen <pvanleeuwen@xxxxxxxxxxxxxx> > > > Subject: Re: [PATCH 4/7] crypto: testmgr - Added testvectors for the ofb(sm4) & cfb(sm4) > > > skciphers > > > > > > On Wed, Sep 11, 2019 at 12:38:21PM +0200, Pascal van Leeuwen wrote: > > > > Added testvectors for the ofb(sm4) and cfb(sm4) skcipher algorithms > > > > > > > > > > What is the use case for these algorithms? Who/what is going to use them? > > > > > > - Eric > > > > > SM4 is a Chinese replacement for 128 bit AES, which is mandatory to be used for many > > Chinese use cases. So they would use these whereever you would normally use ofb(aes) > > or cfb(aes). Frankly, I'm not aware of any practicle use cases for these feedback > > modes, but we've been supporting them for decades and apparently the Crypto API > > supports them for AES as well. So they must be useful for something ... > > > > The obvious advantage over CBC mode was that they only require the encrypt part of > > the cipher, but that holds for the (newer) CTR mode as well. So, my guess would be > > some legacy uses cases from before the time CTR mode and AEAD's became popular. > > > > Maybe someone remembers why these were added for AES in the first place? > > > > So if you have no idea why they should be added, why are you adding them? > Because our hardware supports these modes and I added support for this to the inside-secure driver, so it made sense to be able to test that as well. That's the primary reason for adding them, otherwise I would not have bothered. > - Eric Regards, Pascal van Leeuwen Silicon IP Architect, Multi-Protocol Engines @ Verimatrix www.insidesecure.com
