Tomasz Motylewski wrote:
>
> On Thu, 18 May 2000, Andrew Morton wrote:
>
> > Everything looks OK on the NIC. I noticed a comment from Don Becker:
> >
> > > Our clusters typically use SMP machines, and we keep getting
> > > burned by APIC configuration bugs appearing as driver bugs.
> >
> > This supports my suspicion that it's not a 3c59x problem, and that the
> > machine's interrupt routing has got itself screwed up.
>
> I have very similar problem (reported many times, reproducible) with 3c509
> NIC on dual P200 (GA-586DX mobo) in 2.2.13-2.2.15pre19 kernels (I have not
> tried later). See for config http://crds.chemie.unibas.ch/linux/2.2.15/
Well, if I'm right, there is no reason for it to be unique to the
3c905's. It's just more visible and explicable with these NICs because
the tx timeout routine specifically checks the NIC registers for a
posted-but-not-delivered interrupt.
On other NICs it'll just be a mysterious tx timeout. We see reports
like this almost daily...
hmm.. I have a 509 on an SMP machine and I'm doing some 2.3.x tidying
up on it. It looks pretty clean for SMP. Could you please send me one
of your reports? I'll see if I can see anything...
--
-akpm-
-
Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/dmentre/smp-howto/
To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
