[FFmpeg-devel] [RFC] Public API for RC4 and DES

Michael Niedermayer michaelni
Mon Feb 2 17:08:46 CET 2009


On Mon, Feb 02, 2009 at 05:05:08PM +0100, Reimar D?ffinger wrote:
> On Mon, Feb 02, 2009 at 03:58:44PM +0100, Michael Niedermayer wrote:
> > On Mon, Feb 02, 2009 at 01:38:43PM +0000, M?ns Rullg?rd wrote:
> > > Using the same API for all the encryption types makes sense.
> > 
> > yes and no
> >
> > if it slows 2 of 3 down by a significant amount it wouldnt be that ideal
> 
> The slowdown depends mostly on the usage, I think performance-wise the current
> design of the AES routines would be quite bad if you were only decoding
> a single block each time.
> On the other hand, the current DES API means every user has to do the
> endianness conversion, CBC and other stuff themselves, and makes it hard
> to support 3DES.
> I can't see how this change should noticeable affect RC4 speed, and for
> the current uses we have DES speed is basically irrelevant and the
> implementation is not optimized for absolutely highest speed anyway.

would there be a problem with having both the low level 1 block and highlevel
APIs public?


> I would not mind a better API, but I doubt there is much choice (except
> making the struct structure public).

i dont mind making it public if consensus is that its better and that we
wont need to change it anytime soon

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Complexity theory is the science of finding the exact solution to an
approximation. Benchmarking OTOH is finding an approximation of the exact
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20090202/4af5ede1/attachment.pgp>



More information about the ffmpeg-devel mailing list