On Fri, May 05, 2006 at 11:59:19AM -0400, Michael Richardson ([EMAIL
PROTECTED]) wrote:
> Evgeniy> Actually that is how acrypto works.
>
> I'd like to merge, but we'd also like to keep a BSD copyright on the
> core components.
acrypto heavily uses linux specific internals. As far as I recall, I
posted comparison of OCF and acrypto design some time ago in
linux-crypto@/[EMAIL PROTECTED]
If BSD folks decide to change OCF they will rewrite it from scratch
anyway.
And current linux native IPsec was not designed to be asynchronous.
Herbert Xu (linux crypto maintainer, in Cc:) proposed to create similar
to netfilter queueing, but it will hurt performance for synchronous-only
case which is also supported by acrypto, so it requires _heavy_
modifications to xfrm engine (current sw-only benchmark shows the same
speed for native ipsec and acrypto+async ipsec (it is even slightly
faster with HT enabled) in esp4 mode).
Network folks will never merge it :)
And there is ongoing asynchronous work on top of existing synchronous
software cryptoapi, which was discussed in linux-crypto@, although there
are a lot of reefs there...
--
Evgeniy Polyakov
-
To unsubscribe from this list: send the line "unsubscribe linux-crypto" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
