Jes Sorensen writes: > >>>>> "Richard" == Richard Gooch <rgooch@ras.ucalgary.ca> writes: > > Richard> Jes Sorensen writes: > >> 2.0.36 reliawhat?? If you are looking for GigE performance you are > >> likely to want SMP and SMP in 2.0.36 is absolutely not realiable. > > Richard> There's a strong perception that 2.0.36 is robust (even > Richard> though SMP isn't). I see people pushing for XYZ driver to be > Richard> added to 2.0.3x for this reason. I assume if Kim is willing > Richard> to back-port the GigE driver to 2.0.36, the option of > Richard> forward-porting some other driver was also considered. > > True, from a driver maintainers point of view I'd just note that I > have no interest in adding 2.0.x compat patches to for instance the > acenic driver. Yeah, but that won't stop people from wanting it :-) > Richard> As for SMP "requirements": perhaps Kim has a 1 GHz Athalon > Richard> and figures that should do nicely :-) > > Hmmm I thought I read a mail from Alan a few minutes ago stating > that 2.2.x didn't even support the Athlon? I can assure you that 2.2.15 works very nicely on an Athalon. I was running it on our server when the 2.3.99 kernels had appalling VM performance. Even MTRR support (the K7 has Intel-compatible MTRRs while the K6 were AMD-specific; the K7 broke an assumption in the MTRR driver which was later fixed in 2.2.x and 2.3.x kernels) works. I think what Alan meant is that support for K7-specific features is not available. So the kernel just treats it as a lowly PIII. Regards, Richard.... Permanent: rgooch@atnf.csiro.au Current: rgooch@ras.ucalgary.ca - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.rutgers.edu
