Re: [PATCH v2 2/2] random: use BLAKE2s instead of SHA1 in extraction

[Geert Uytterhoeven]

Hi Jason,

On Tue, Jan 11, 2022 at 1:50 PM Jason A. Donenfeld <Jason@xxxxxxxxx> wrote:
> On Tue, Jan 11, 2022 at 1:28 PM Jason A. Donenfeld <Jason@xxxxxxxxx> wrote:
> > If you're really quite concerned about m68k code size, I can probably

It's not just m68k. There exist ARM SoCs with 8 MiB builtin SRAM that
are used in products running Linux.

> > do some things to reduce that. For example, blake2s256_hmac is only
> > used by wireguard and it could probably be made local there. And with
> > some trivial loop re-rolling, I can shave off another 2300 bytes. And
> > I bet I can find a few other things too. The question is: how
> > important is this to you?
>
> And with another trick (see below), another extra 1000 bytes or so
> shaved off. Aside from moving blake2s256_hmac, I'm not really super
> enthusiastic about making these changes, but depending on how important
> this is to you, maybe we can make something work. There are probably
> additional possibilities too with the code.

Cool, much more than 1000 bytes:

add/remove: 1/0 grow/shrink: 0/1 up/down: 160/-4032 (-3872)
Function                                     old     new   delta
blake2s_sigma                                  -     160    +160
blake2s_compress_generic                    4448     416   -4032
Total: Before=4227876, After=4224004, chg -0.09%

I don't know what the impact is on performance, and if the compiler
might do a good job unrolling this again when performance matters
(i.e. if CONFIG_CC_OPTIMIZE_FOR_SIZE is not set).

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds