Re: 2.3.43-pre[25] do not work with HIGHMEM=64GB

From: Manfred Spraul (manfreds@colorfullife.com)
Date: Wed Feb 09 2000 - 15:42:09 EST


Ingo Molnar wrote:
>
> On Wed, 9 Feb 2000, Manfred Spraul wrote:
>
> > > Unable to handle kernel NULL pointer dereference at virtual addres
> > > s 00000088
> > > *pde = 00104001
>
> > Strange. Ingo reported the same problem. [...]
>
> this one is definitely fixed in pre5. (with directed APIC broadcasts it's
> not possible to hit an uninitialized CPU)
>

I would prefer if we explicitly initialize cpu_tlbstate during cpu_init:
it's cleaner, and it prevents any ordering problems within switch_mm()
[there is a tiny race, around 30 instructions]

Do you have any further crashes due to the tlb changes? I was running
with my debug patches without any special messages.

Oh. Are you using gcc-2.95? I forgot to mark the cpu_tlbstate array as
volatile [arch/i386/kernel/smp.c and include/asm/pgalloc.h], could you
change that?

--
	Manfred

- 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/



This archive was generated by hypermail 2b29 : Tue Feb 15 2000 - 21:00:16 EST