On Wed, Feb 12, 2014 at 10:35:56AM -0600, Jim Thompson wrote: > > Can the blobs be reversed so easily? (Too bad about lack of IPv6 > > offloading, but we can live with that for a while, I guess). > > I don’t know. If you’re really curious, you can read this: > http://university.caviumnetworks.com/downloads/Mini_version_of_Prog_Guide_EDU_July_2010.pdf > to find out how to get ahold the real programming guide from Cavium, then > read Chapter 2 “Packet Flow” in same. > > This might give you some ideas as well: > https://hactive.googlecode.com/files/CN50XX-HRM-V0.99E.pdf > Note that this link seems to support the idea that IPv6 processing is > supported by the hardware (see, for example, Sections 7.2.4, 7.5 and 7.7).
Thanks, interesting. > I do know that *I* don’t want to invest a ton of RE effort in a $99 platform > that bears near zero margins, when far, far faster > Intel / AMD platforms that aren’t more than 2-3X the price are just around > the corner. Note slide 17 in this: > https://noppa.aalto.fi/noppa/kurssi/s-38.3310/harjoitustyot/S-38_3310_matias_elo.pdf Are you aware of serius efforts to use GPU (OpenCL) accelerated packet processing in FreeBSD? _______________________________________________ List mailing list [email protected] http://lists.pfsense.org/mailman/listinfo/list
