Re: [PATCH 2/2] Support for VIA PadLock crypto engine

[Jari Ruusu]

Michal Ludvig wrote:
> My padlock driver can be used for anything that uses CryptoAPI and in fact
> it speeds things a lot (see a simple disk-based benchmark at
> http://www.logix.cz/michal/dl/padlock.xp).

Cryptoapi version of AES is slowest implementation of AES that I know of.
For speed tests, please compare against more modern implementation.

Below is one old AES128 speed test that I ran on my 300 MHz test box:

KERNEL      IMPLEMENTATION  MODE                WRITE MiB/s     READ MiB/s
2.6.1       cryptoloop      single-key           5.21            4.08
2.6.1       loop-AES        single-key           9.52            7.56
2.6.1       loop-AES        multi-key(MD5 IV)    7.67            6.35
2.4.22aa1   loop-AES        single-key          10.55           10.16
2.4.22aa1   loop-AES        multi-key(MD5 IV)    8.75            8.13

The cryptoloop implementation is busted in more than one way, so it is
useless for security needs:

http://marc.theaimsgroup.com/?l=linux-kernel&m=107419912024246&w=2
http://marc.theaimsgroup.com/?l=linux-kernel&m=107719798631935&w=2

-- 
Jari Ruusu  1024R/3A220F51 5B 4B F9 BB D3 3F 52 E9  DB 1D EB E3 24 0E A9 DD
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/