Re: 48GB NUMA-Q boots, with major IO-APIC hassles

Martin J. Bligh (mbligh@aracnet.com)
Wed, 15 Jan 2003 07:24:40 -0800


> (2) MAX_IO_APIC's got clobbered in the subarch cleanups.
> -- CONFIG_X86_NUMA was removed, use CONFIG_X86_NUMAQ
> -- this is greppable, folks...

That wasn't the subarch cleanups that removed it, please be careful
what you're saying. I plead not guilty to that one.

> (4) PCI bridges get misnumbered children.
> -- Brew up a PCI hook for giving child buses their bus numbers.
> -- Basically, fwd port mbligh's fix for 2.4.x more cleanly.
> -- Okay, not IO-APIC-related, but it annoys me greatly.
> -- ink is at least trying to steer me in the right direction here.

Additional PCI-PCI bridges (eg starfire cards) have never been supported
in non-boot quads. It's not impossible, but don't be suprised if it
doesn't work.

> (5) Booting with notsc panic()'s.
> -- Remove tsc_disable assignment in the __setup() call.
> -- I'd be much obliged if the SMP TSC issues were at long
> -- last conclusively dealt with. Not IO-APIC-related either,
> -- but also very annoying.

You don't have PIT support compiled in, and you turned off TSC support,
leaving yourself with no timer. There's a patch in my tree to force on
PIT support for NUMA-Q.

M.

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