* xen_tlb_flush_all mfn validation panic
@ 2009-12-24 15:06 Kieran Mansley
2009-12-24 15:32 ` Keir Fraser
0 siblings, 1 reply; 4+ messages in thread
From: Kieran Mansley @ 2009-12-24 15:06 UTC (permalink / raw)
To: xen-devel
I'm seeing the following when trying to boot a guest domain with the
current xen-unstable:
(XEN) mm.c:2695:d0 Bad gmfn_to_mfn
------------[ cut here ]------------
kernel BUG at /home/kjm/xen/linux-2.6.18-xen.hg/arch/i386/mm/hypervisor.c:123!
invalid opcode: 0000 [#1]
SMP
Modules linked in: video thermal processor fan container button battery asus_acd
CPU: 0
EIP: 0061:[<c011d145>] Not tainted VLI
EFLAGS: 00010282 (2.6.18.8-xen0 #1)
EIP is at xen_tlb_flush_all+0x35/0x40
eax: fffffff2 ebx: deadbeef ecx: deadbeef edx: deadbeef
esi: deadbeef edi: c5350dc8 ebp: cf877dcc esp: cf877db8
ds: 007b es: 007b ss: 0069
Process xenwatch (pid: 9, ti=cf876000 task=cf855030 task.ti=cf876000)
Stack: 0000000a ffffffff c5388000 d11ba000 05350067 cf877e80 c0166610 00000000
00000000 c01310fa cf877dfc 00000000 00000000 00000001 00000000 0001bf06
cf877e24 8f0560fc 00000150 00000067 00000000 05350000 00000000 00000063
Call Trace:
[<c0106221>] show_stack_log_lvl+0xb1/0xe0
[<c0106486>] show_registers+0x1b6/0x240
[<c0106641>] die+0x131/0x330
[<c01068d8>] do_trap+0x98/0xe0
[<c0107170>] do_invalid_op+0xa0/0xb0
[<c0105d5b>] error_code+0x2b/0x30
[<c0166610>] unmap_vm_area+0x330/0x370
[<c0166694>] __remove_vm_area+0x44/0x60
[<c01666cb>] remove_vm_area+0x1b/0x30
[<c0316c44>] free_vm_area+0x14/0x30
[<c032648e>] netif_disconnect+0xce/0x180
[<c0325c3f>] frontend_changed+0x26f/0x630
[<c0314e55>] otherend_changed+0xa5/0xb0
[<c0312c71>] xenwatch_handle_callback+0x21/0x60
[<c0313a15>] xenwatch_thread+0x135/0x150
[<c013c266>] kthread+0x106/0x110
[<c0103005>] kernel_thread_helper+0x5/0x10
Code: 83 ec 14 89 5d f8 8d 5d ec 89 75 fc be f0 7f 00 00 c7 45 ec 0a 00 00 00 e
EIP: [<c011d145>] xen_tlb_flush_all+0x35/0x40 SS:ESP 0069:cf877db8
The initial "bad gmfn" message seems to be due to an attempt to validate
the mfn argument to the mmuext_op. It's not surprising to me that this
fails as xen_tlb_flush_all doesn't supply an mfn argument - it's not
needed for the MMUEXT_TLB_FLUSH_ALL command. Is there something that is
supposed to mean that the mfn argument should be valid even when not
explicitly set by the caller, or is the validation of the mfn for all
mmuext_op commands a bug?
Thanks
Kieran
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: xen_tlb_flush_all mfn validation panic
2009-12-24 15:06 xen_tlb_flush_all mfn validation panic Kieran Mansley
@ 2009-12-24 15:32 ` Keir Fraser
2009-12-24 16:01 ` Keir Fraser
0 siblings, 1 reply; 4+ messages in thread
From: Keir Fraser @ 2009-12-24 15:32 UTC (permalink / raw)
To: Kieran Mansley, xen-devel@lists.xensource.com
On 24/12/2009 15:06, "Kieran Mansley" <kmansley@solarflare.com> wrote:
> The initial "bad gmfn" message seems to be due to an attempt to validate
> the mfn argument to the mmuext_op. It's not surprising to me that this
> fails as xen_tlb_flush_all doesn't supply an mfn argument - it's not
> needed for the MMUEXT_TLB_FLUSH_ALL command. Is there something that is
> supposed to mean that the mfn argument should be valid even when not
> explicitly set by the caller, or is the validation of the mfn for all
> mmuext_op commands a bug?
The explicit check for INVALID_MFN is new, and that is obviously a bug. I
will sort out a suitable fix for this.
-- Keir
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: xen_tlb_flush_all mfn validation panic
2009-12-24 15:32 ` Keir Fraser
@ 2009-12-24 16:01 ` Keir Fraser
2009-12-24 16:13 ` Kieran Mansley
0 siblings, 1 reply; 4+ messages in thread
From: Keir Fraser @ 2009-12-24 16:01 UTC (permalink / raw)
To: Kieran Mansley, xen-devel@lists.xensource.com
On 24/12/2009 15:32, "Keir Fraser" <keir.fraser@eu.citrix.com> wrote:
> On 24/12/2009 15:06, "Kieran Mansley" <kmansley@solarflare.com> wrote:
>
>> The initial "bad gmfn" message seems to be due to an attempt to validate
>> the mfn argument to the mmuext_op. It's not surprising to me that this
>> fails as xen_tlb_flush_all doesn't supply an mfn argument - it's not
>> needed for the MMUEXT_TLB_FLUSH_ALL command. Is there something that is
>> supposed to mean that the mfn argument should be valid even when not
>> explicitly set by the caller, or is the validation of the mfn for all
>> mmuext_op commands a bug?
>
> The explicit check for INVALID_MFN is new, and that is obviously a bug. I
> will sort out a suitable fix for this.
Should be fixed by c/s 20723.
-- Keir
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: xen_tlb_flush_all mfn validation panic
2009-12-24 16:01 ` Keir Fraser
@ 2009-12-24 16:13 ` Kieran Mansley
0 siblings, 0 replies; 4+ messages in thread
From: Kieran Mansley @ 2009-12-24 16:13 UTC (permalink / raw)
To: Keir Fraser; +Cc: xen-devel@lists.xensource.com
On Thu, 2009-12-24 at 16:01 +0000, Keir Fraser wrote:
> On 24/12/2009 15:32, "Keir Fraser" <keir.fraser@eu.citrix.com> wrote:
>
> > On 24/12/2009 15:06, "Kieran Mansley" <kmansley@solarflare.com> wrote:
> >
> >> The initial "bad gmfn" message seems to be due to an attempt to validate
> >> the mfn argument to the mmuext_op. It's not surprising to me that this
> >> fails as xen_tlb_flush_all doesn't supply an mfn argument - it's not
> >> needed for the MMUEXT_TLB_FLUSH_ALL command. Is there something that is
> >> supposed to mean that the mfn argument should be valid even when not
> >> explicitly set by the caller, or is the validation of the mfn for all
> >> mmuext_op commands a bug?
> >
> > The explicit check for INVALID_MFN is new, and that is obviously a bug. I
> > will sort out a suitable fix for this.
>
> Should be fixed by c/s 20723.
Great, thanks. I'll retest after Christmas.
Kieran
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2009-12-24 16:13 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2009-12-24 15:06 xen_tlb_flush_all mfn validation panic Kieran Mansley
2009-12-24 15:32 ` Keir Fraser
2009-12-24 16:01 ` Keir Fraser
2009-12-24 16:13 ` Kieran Mansley
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.