lkml.org 
[lkml]   [2005]   [Oct]   [1]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: [discuss] Re: [PATCH][Fix][Resend] Fix Bug #4959: Page tables corrupted during resume on x86-64 (take 3)
Date
Hi,

On Saturday, 1 of October 2005 03:25, Siddha, Suresh B wrote:
> On Fri, Sep 30, 2005 at 08:51:08AM +0200, Rafael J. Wysocki wrote:
> > On Friday, 30 of September 2005 07:26, Rafael J. Wysocki wrote:
> > > Of course. The .config is attached. Generally, it's a non-SMP box, and commenting
> > > out the zap_low_mappings((0) in setup_arch() makes the box boot again.
> >
> > One more datapoint: The box boots if I move the zap_low_mappings((0)
> > in the following way:
> >
> > --- linux-2.6.14-rc2-git7.orig/arch/x86_64/kernel/setup.c 2005-09-30 07:39:35.000000000 +0200
> > +++ linux-2.6.14-rc2-git7/arch/x86_64/kernel/setup.c 2005-09-30 08:31:20.000000000 +0200
> > @@ -571,8 +571,6 @@
> >
> > init_memory_mapping(0, (end_pfn_map << PAGE_SHIFT));
> >
> > - zap_low_mappings(0);
> > -
> > #ifdef CONFIG_ACPI
> > /*
> > * Initialize the ACPI boot-time table parser (gets the RSDP and SDT).
> > @@ -680,6 +678,8 @@
> > get_smp_config();
> > init_apic_mappings();
> > #endif
> > + zap_low_mappings(0);
> > +
>
> Rafael, I still can't reproduce the issue. All of my EM64T systems are SMP
> and looks like this problem happens only on an UP system. Looks like
> there is some code flow which is still assuming the presence of low mappings.

I think so too.

> Andi, can you please see if you can reproduce this on one of your systems
> similar to Rafael's.
>
> > /*
> > * Request address space for all standard RAM and ROM resources
> >
> > iHowever, if I place the zap_low_mappings((0) before the
> > #define CONFIG_X86_LOCAL_APIC in line 673, it doesn't boot.
> >
> > Certainly init_apic_mappings() is at fault. Could that be a reult of a call to
> > alloc_bootmem_pages()?.
>
> I looked at init_apic_mappings() and didn't give me any clue.
> alloc_bootmem_pages() don't use low direct mappings.

OK, I'll try to narrow this a bit more.

> > And one more: I have to boot with "noapic".
>
> Are you saying "noapic" is the second data point or you need to use "noapic"
> along with moving zap_low_mappings(0)

It could be the second data point. I always boot with "noapic", since otherwise
the sharing of IRQs is not handled properly on my box.

Greetings
Rafael
-
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/

\
 
 \ /
  Last update: 2005-10-01 09:49    [W:0.080 / U:0.404 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site