Re: v2.6.27-rc7: x86: #GP on panic?

["Vegard Nossum" <vegard.nossum@gmail.com>]

On Thu, Sep 25, 2008 at 10:53 AM, H. Peter Anvin <hpa@zytor.com> wrote:
> Ingo Molnar wrote:
>>
>> * Vegard Nossum <vegard.nossum@gmail.com> wrote:
>>
>>> Hi,
>>>
>>> With 2.6.27-rc7 on qemu-x86_64, it seems that panic will trigger a
>>> General Protection Fault. I haven't seen it before.
>>
>>> [    4.523641] Code: eb fd 55 48 89 e5 53 51 83 3d 25 e8 78 00 00 75
>>> 1a 31 d2 31 f6 48 c7 c7 e1 9c 01 81 e8 f7 a4 03 00 9c 5b fa e8 94 09
>>> 00 00 53 9d <5a> 5b c9 c3 55 31 c0 48 89 e5 89 04 25 b0 c0 5f ff 65 83
>>> 04 25
>>
>> hm, 0x5a is a simple pop %rdx. A #GP there means the stack segment is
>> bust?
>>
>
> No, that would be #SS (and segments don't really exist in 64-bit mode
> anyway.)  In 32-bit mode it could mean a code segment overrun.
>
> *However*...
>
> [    4.523477] general protection fault: fff2 [1] SMP
>
> There is an error code attached to the #GP, which is supposed to mean that
> somehow a segment selector was involved. This doesn't look like a very valid
> segment selector at all.
>
>> hm:
>>
>>> ffffffff8101a6b9 <disable_local_APIC>
>>> ffffffff81019d25:       53                      push   %rbx
>>> ffffffff81019d26:       9d                      popfq
>>> ffffffff81019d27:       5a                      pop    %rdx
>>
>> so it's preceded by a popfq and on the next instruction we #GP.
>>
>> but the stack and flags state looks good:
>>
>>  [    4.523641] RSP: 0018:ffff880007867d70  EFLAGS: 00000286
>>
>
> My guess is that the popfq enables interrupts, and we try to take an
> interrupt through an IDT entry which isn't set up correctly.

I'm sorry for the false alarm. I discovered that it did not happen on
a clean kernel. My kernel was using this patch.

diff --git a/arch/x86/kernel/cpu/common_64.c b/arch/x86/kernel/cpu/common_64.c
index a11f5d4..abf5bc8 100644
--- a/arch/x86/kernel/cpu/common_64.c
+++ b/arch/x86/kernel/cpu/common_64.c
@@ -261,6 +261,8 @@ void __init early_cpu_init(void)
                 cpu_devs[cvdev->vendor] = cvdev->cpu_dev;
        early_cpu_support_print();
        early_identify_cpu(&boot_cpu_data);
+
+       setup_clear_cpu_cap(X86_FEATURE_PSE);
 }

 /* Do some early cpuid on the boot CPU to get some parameter that are

:-(


Vegard

-- 
"The animistic metaphor of the bug that maliciously sneaked in while
the programmer was not looking is intellectually dishonest as it
disguises that the error is the programmer's own creation."
	-- E. W. Dijkstra, EWD1036