Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <20180911214737.GA81235@gmail.com> <CAHmME9rFUruF-VN1pmU-k5nFsb9ppAPhPpW-5Ho9dKL2HCg4kA@mail.gmail.com>
 <20180911233015.GD11474@lunn.ch> <20180911.165739.2032677219588723041.davem@davemloft.net>
 <CALCETrXrJPcs3h9CtF50N4k2AGx3qzspJd_UW3t+KGYOj7+p=Q@mail.gmail.com>
In-Reply-To: <CALCETrXrJPcs3h9CtF50N4k2AGx3qzspJd_UW3t+KGYOj7+p=Q@mail.gmail.com>
From:   "Jason A. Donenfeld" <Jason@zx2c4.com>
Date:   Mon, 17 Sep 2018 07:07:42 +0200
Message-ID: <CAHmME9ohGPTX-fkhP6zaWdx3SbkiqO5kWkWE-bDk=pm-nUqYvw@mail.gmail.com>
Subject: Re: [PATCH net-next v3 02/17] zinc: introduce minimal cryptography library
To:     Andrew Lutomirski <luto@kernel.org>
Cc:     David Miller <davem@davemloft.net>, Andrew Lunn <andrew@lunn.ch>,
        Eric Biggers <ebiggers@kernel.org>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Ard Biesheuvel <ard.biesheuvel@linaro.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Netdev <netdev@vger.kernel.org>, Samuel Neves <sneves@dei.uc.pt>,
        Jean-Philippe Aumasson <jeanphilippe.aumasson@gmail.com>,
        Linux Crypto Mailing List <linux-crypto@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hey Andy,

Thanks a lot for your feedback.

On Mon, Sep 17, 2018 at 6:09 AM Andy Lutomirski <luto@kernel.org> wrote:
> 1. Zinc conflates the addition of a new API with the replacement of
> some algorithm implementations.  This is problematic.  Look at the
> recent benchmarks of ipsec before and after this series.  Apparently
> big packets get faster and small packets get slower.  It would be
> really nice to bisect the series to narrow down *where* the regression
> came from, but, as currently structured, you can't.
>
> The right way to do this is to rearrange the series.  First, the new
> Zinc APIs should be added, and they should be backed with the
> *existing* crypto code.  (If the code needs to be moved or copied to a
> new location, so be it.  The patch will be messy because somehow the
> Zinc API is going to have to dispatch to the arch-specific code, and
> the way that the crypto API handles it is not exactly friendly to this
> type of use.  So be it.)  Then another patch should switch the crypto
> API to use the Zinc interface.  That patch, *by itself*, can be
> benchmarked.  If it causes a regression for small ipsec packets, then
> it can be tracked down relatively easily.  Once this is all done, the
> actual crypto implementation can be changed, and that changed can be
> reviewed on its own merits.

That ipsec regression was less related to the implementation and more
related to calling kernel_fpu_begin() unnecessarily, something I've
now fixed. So I'm not sure that's such a good example. However, I can
try to implement Zinc over the existing assembly (Martin's and Ard's),
first, as you've described. This will be a pretty large amount of
work, but if you think it's worth it for the commit history, then I'll
do it.

> 2. The new Zinc crypto implementations look like they're brand new.  I
> realize that they have some history, some of them are derived from
> OpenSSL, etc, but none of this is really apparent in the patches
> themselves.

The whole point of going with these is that they _aren't_ brand new,
yet they are very fast. Eyeballs and fuzzer hours are important, and
AndyP's seems to get the most eyeballs and fuzzer hours, generally.

> it would be nice if
> the patches made it more clear how the code differs from its origin.
> At the very least, though, if the replacement of the crypto code were,
> as above, a patch that just replaced the crypto code, it would be much
> easier to review and benchmark intelligently.

You seem to have replied to the v3 thread, not the v4 thread. I've
already started to include lots of detail about the origins of the
code and [any] important differences in v4, and I'll continue to add
more detail for v5.

On <https://git.zx2c4.com/linux-dev/log/>, have a look at AndyP's x86_64 ones:
- zinc: ChaCha20 x86_64 implementation
- zinc: Poly1305 x86_64 implementation
For the arm/arm64 ones, the changes were even more trivial, so much so
that at Ard's urging, I included a cleaned-up diff inside the commit
message:
- zinc: ChaCha20 ARM and ARM64 implementations
- zinc: Poly1305 ARM and ARM64 implementations

How's that level of detail looking to you?

Thanks again for the review.

Regards,
Jason