Re: [OFFTOPIC] UltraSparc MMU problems

Richard Gooch (rgooch@atnf.csiro.au)
Tue, 27 Oct 1998 22:21:25 +1100


David S. Miller writes:
> Date: Tue, 27 Oct 1998 21:35:16 +1100
> From: Richard Gooch <rgooch@atnf.csiro.au>
>
> David S. Miller writes:
> > Solaris's TLB miss strategy is pretty hopeless...
>
> What does the OS have to do with it? Once the TLB's are set up, why
> does the OS need to intervene?
>
> If the OS uses a stupid scheme to replace a missing TLB entry and uses
> inefficient tables as backing store for mappings, it will thrash the
> L2 cache.
>
> Note that Solaris and some other svr4 folks swap page tables out to
> disk when memory is tight... when they get tapped back in you might
> not get the nice cache patterns you had before...

Ah, OK. So while doing the renderings I didn't hear the disc rattle,
it still matters that the page tables were swapped out.

> BTW: we have FORE systems ATM cards in these machines, so that
> probably rules out Linux anyway :-(
>
> We have half-way done drivers for those, but there is still a lot of
> work to do. Actually if I remember correctly, these cards are very
> similar to the one's Werner Almesberger wrote drivers for in the
> Linux/ATM project. His were PCI but the chipsets are the same as
> the SBUS ones on Sparc systems.

Yeah, my own machine has an ENI card and that works quite well :-)
The Linux ATM code looks good. Unfortunately I won't be able to take
advantage of this because by the time Linux drivers for the SBus FORE
cards is ready, I'll have left this place.

Anyway, your feeling is this isn't a hardware problem, it's an OS
problem?

Regards,

Richard....

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/