sha1/botan stuff (was Re: [Monotone-devel] updates to net.venge.monotone.experiment.performance)

[Nathaniel Smith]

On Wed, Aug 09, 2006 at 10:42:37PM -0700, Eric Anderson wrote:
> Needs more discussion:
> 
> e1a721eb1b1bf8d64229419ac1f73bda0a855590: stop zeroing in Botan::gzip
> 
> Remove zeroing of memory used by Botan to do compression. 1.06x
> reduction in client time, 1.02x in server time.  I don't think for
> monotone's usage there is any security risk introduced by this change,
> but in needs to be thought through a little.  I'd expect that more
> improvements could be done in this manner as Botan zeros both on free
> and allocate, but my one attempt to remove more of the zeroing caused
> failures.

Sounds plausible enough.

> -----------
> 997a677db676734acc0d098979d2a9cee8765ec9: libcrypto ssl linking
> 
> Enable optional compilation with openssl libcrypto for the optimized
> SHA1 hash.  Likely to be obsoleted by getting the fast assembly code
> from libcrypto used in Botan.  Depending on how long that's expected
> to take, it may be worth merging this patch now (it's disabled by
> default) and letting people enable it if they want.  It was a
> relatively substantial improvement at the time of the measurements,
> probably a whole lot more after all of the other improvements have
> been applied.

I'm nervous about landing it on mainline, even as optional, because of
the potential legal issues -- e.g., possibly debian would have to rip
it out again before distributing our standard packages, etc. etc.

Unfortunately, the status of getting fast SHA1 is:
  arm, ppc: License-compatible implementations are in git, with a
    trivial API to drop in.  Getting these working is just a matter of
    build system pieces and a few lines of trivial Botan wrapper code.
    Someone send in a patch!  (More instructions available on
    request.)
  x86-64, ia64: Author of openssl code is willing in principle to
    relicense, though nothing has happened.  Given that happens, just
    need build system changes plus a slightly less trivial Botan
    wrapper, except that I think Jack already wrote the needed wrapper
    in an earlier email, so again, no problem.
  x86: Apparently the openssl code does, after all, have a copyright
    interest from the original SSLeay authors, who work at RSA Labs
    these days and refuse to give up the advertising clause on
    anything.

So everything's great, unless you're one of those fringe weirdos that
uses an x86.

Umm... anyone want to test out their asm chops?

-- Nathaniel

-- 
"The problem...is that sets have a very limited range of
activities -- they can't carry pianos, for example, nor drink
beer."