BUG: Bad page map in process

BUG: Bad page map in process

[Sean MacLennan]

With 2.6.32 I sometimes get lots and lots of "Bad page map" errors as
shown below. I believe these started in 2.6.32-rc8 or possibly
2.6.32-rc7. Pika just switched to 2.6.31 so I have been concentrating
on that release, and not really testing the 2.6.32 stream.

I wish I could give more info, but I don't really even have the time to
write this email :( I am hoping somebody will have a eureka moment.

Cheers,
   Sean


[  147.410448] BUG: Bad page map in process udevd  pte:ffffffffffffffff pmd:cd883000
[  147.417989] addr:48003000 vm_flags:08000875 anon_vma:(null) mapping:cf4150ac index:3
[  147.425837] vma->vm_ops->fault: filemap_fault+0x0/0x410
[  147.431101] vma->vm_file->f_op->mmap: nfs_file_mmap+0x0/0x94
[  147.436795] Call Trace:
[  147.439268] [cf2dbd80] [c00065a4] show_stack+0x48/0x168 (unreliable)
[  147.445708] [cf2dbdb0] [c0064ac8] print_bad_pte+0x14c/0x204
[  147.451337] [cf2dbde0] [c0064cd4] vm_normal_page+0x90/0x98
[  147.456875] [cf2dbdf0] [c0064ee8] unmap_vmas+0x20c/0x638
[  147.462239] [cf2dbe80] [c0069194] exit_mmap+0xc8/0x170
[  147.467441] [cf2dbea0] [c001caa4] mmput+0x50/0xf4
[  147.472195] [cf2dbeb0] [c0020808] exit_mm+0x100/0x138
[  147.477299] [cf2dbee0] [c00211b0] do_exit+0xb0/0x580
[  147.482315] [cf2dbf20] [c00216c4] do_group_exit+0x44/0x9c
[  147.487766] [cf2dbf30] [c0021730] sys_exit_group+0x14/0x28
[  147.493308] [cf2dbf40] [c000dde4] ret_from_syscall+0x0/0x3c

Re: BUG: Bad page map in process

[Jake Magee]

Have you tried the following patches?  I'm not sure if they have made
it into the newer releases.
http://www.mail-archive.com/linuxppc-dev@lists.ozlabs.org/msg37188.html

Also for reference... http://patchwork.ozlabs.org/patch/34047/


On Thu, Dec 3, 2009 at 1:18 PM, Sean MacLennan <smaclennan@pikatech.com> wr=
ote:
> With 2.6.32 I sometimes get lots and lots of "Bad page map" errors as
> shown below. I believe these started in 2.6.32-rc8 or possibly
> 2.6.32-rc7. Pika just switched to 2.6.31 so I have been concentrating
> on that release, and not really testing the 2.6.32 stream.
>
> I wish I could give more info, but I don't really even have the time to
> write this email :( I am hoping somebody will have a eureka moment.
>
> Cheers,
> =A0 Sean
>
>
> [ =A0147.410448] BUG: Bad page map in process udevd =A0pte:ffffffffffffff=
ff pmd:cd883000
> [ =A0147.417989] addr:48003000 vm_flags:08000875 anon_vma:(null) mapping:=
cf4150ac index:3
> [ =A0147.425837] vma->vm_ops->fault: filemap_fault+0x0/0x410
> [ =A0147.431101] vma->vm_file->f_op->mmap: nfs_file_mmap+0x0/0x94
> [ =A0147.436795] Call Trace:
> [ =A0147.439268] [cf2dbd80] [c00065a4] show_stack+0x48/0x168 (unreliable)
> [ =A0147.445708] [cf2dbdb0] [c0064ac8] print_bad_pte+0x14c/0x204
> [ =A0147.451337] [cf2dbde0] [c0064cd4] vm_normal_page+0x90/0x98
> [ =A0147.456875] [cf2dbdf0] [c0064ee8] unmap_vmas+0x20c/0x638
> [ =A0147.462239] [cf2dbe80] [c0069194] exit_mmap+0xc8/0x170
> [ =A0147.467441] [cf2dbea0] [c001caa4] mmput+0x50/0xf4
> [ =A0147.472195] [cf2dbeb0] [c0020808] exit_mm+0x100/0x138
> [ =A0147.477299] [cf2dbee0] [c00211b0] do_exit+0xb0/0x580
> [ =A0147.482315] [cf2dbf20] [c00216c4] do_group_exit+0x44/0x9c
> [ =A0147.487766] [cf2dbf30] [c0021730] sys_exit_group+0x14/0x28
> [ =A0147.493308] [cf2dbf40] [c000dde4] ret_from_syscall+0x0/0x3c
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@lists.ozlabs.org
> https://lists.ozlabs.org/listinfo/linuxppc-dev
>

Re: BUG: Bad page map in process

[Sean MacLennan]

On Thu, 3 Dec 2009 14:06:34 -0600
Jake Magee <jakemagee@gmail.com> wrote:

> Have you tried the following patches?  I'm not sure if they have made
> it into the newer releases.
> http://www.mail-archive.com/linuxppc-dev@lists.ozlabs.org/msg37188.html
> 
> Also for reference... http://patchwork.ozlabs.org/patch/34047/

It looks like that patch made it into 2.6.32.

Cheers,
   Sean

Re: BUG: Bad page map in process

[Benjamin Herrenschmidt]

On Thu, 2009-12-03 at 14:18 -0500, Sean MacLennan wrote:
> With 2.6.32 I sometimes get lots and lots of "Bad page map" errors as
> shown below. I believe these started in 2.6.32-rc8 or possibly
> 2.6.32-rc7. Pika just switched to 2.6.31 so I have been concentrating
> on that release, and not really testing the 2.6.32 stream.
> 
> I wish I could give more info, but I don't really even have the time to
> write this email :( I am hoping somebody will have a eureka moment.

That's a concern. Remind me what CPU / MMU type is Pika using ?

Cheers,
Ben.

> Cheers,
>    Sean
> 
> 
> [  147.410448] BUG: Bad page map in process udevd  pte:ffffffffffffffff pmd:cd883000
> [  147.417989] addr:48003000 vm_flags:08000875 anon_vma:(null) mapping:cf4150ac index:3
> [  147.425837] vma->vm_ops->fault: filemap_fault+0x0/0x410
> [  147.431101] vma->vm_file->f_op->mmap: nfs_file_mmap+0x0/0x94
> [  147.436795] Call Trace:
> [  147.439268] [cf2dbd80] [c00065a4] show_stack+0x48/0x168 (unreliable)
> [  147.445708] [cf2dbdb0] [c0064ac8] print_bad_pte+0x14c/0x204
> [  147.451337] [cf2dbde0] [c0064cd4] vm_normal_page+0x90/0x98
> [  147.456875] [cf2dbdf0] [c0064ee8] unmap_vmas+0x20c/0x638
> [  147.462239] [cf2dbe80] [c0069194] exit_mmap+0xc8/0x170
> [  147.467441] [cf2dbea0] [c001caa4] mmput+0x50/0xf4
> [  147.472195] [cf2dbeb0] [c0020808] exit_mm+0x100/0x138
> [  147.477299] [cf2dbee0] [c00211b0] do_exit+0xb0/0x580
> [  147.482315] [cf2dbf20] [c00216c4] do_group_exit+0x44/0x9c
> [  147.487766] [cf2dbf30] [c0021730] sys_exit_group+0x14/0x28
> [  147.493308] [cf2dbf40] [c000dde4] ret_from_syscall+0x0/0x3c
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@lists.ozlabs.org
> https://lists.ozlabs.org/listinfo/linuxppc-dev

Re: BUG: Bad page map in process

[Josh Boyer]

On Fri, Dec 04, 2009 at 08:08:27PM +1100, Benjamin Herrenschmidt wrote:
>On Thu, 2009-12-03 at 14:18 -0500, Sean MacLennan wrote:
>> With 2.6.32 I sometimes get lots and lots of "Bad page map" errors as
>> shown below. I believe these started in 2.6.32-rc8 or possibly
>> 2.6.32-rc7. Pika just switched to 2.6.31 so I have been concentrating
>> on that release, and not really testing the 2.6.32 stream.
>> 
>> I wish I could give more info, but I don't really even have the time to
>> write this email :( I am hoping somebody will have a eureka moment.
>
>That's a concern. Remind me what CPU / MMU type is Pika using ?

440EP

josh

Re: BUG: Bad page map in process

[Benjamin Herrenschmidt]

On Fri, 2009-12-04 at 06:00 -0500, Josh Boyer wrote:
> On Fri, Dec 04, 2009 at 08:08:27PM +1100, Benjamin Herrenschmidt wrote:
> >On Thu, 2009-12-03 at 14:18 -0500, Sean MacLennan wrote:
> >> With 2.6.32 I sometimes get lots and lots of "Bad page map" errors as
> >> shown below. I believe these started in 2.6.32-rc8 or possibly
> >> 2.6.32-rc7. Pika just switched to 2.6.31 so I have been concentrating
> >> on that release, and not really testing the 2.6.32 stream.
> >> 
> >> I wish I could give more info, but I don't really even have the time to
> >> write this email :( I am hoping somebody will have a eureka moment.
> >
> >That's a concern. Remind me what CPU / MMU type is Pika using ?
> 
> 440EP

Ok, I'll have a look next week. In the meantime, Sean, can you give me a
hint of what you do to trigger it ? Boot time ? specific workload ?

Cheers,
Ben.

Re: BUG: Bad page map in process

[Sean MacLennan]

On Fri, 04 Dec 2009 22:18:55 +1100
Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote:

> Ok, I'll have a look next week. In the meantime, Sean, can you give
> me a hint of what you do to trigger it ? Boot time ? specific
> workload ?

It hasn't happened that often, but I have been focusing on 2.6.31.

So far, it has always happened well after the boot, but not hours. I
haven't worked out a specific workload.

Cheers,
   Sean

Re: BUG: Bad page map in process

[Sean MacLennan]

On Fri, 04 Dec 2009 22:18:55 +1100
Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote:

> Ok, I'll have a look next week. In the meantime, Sean, can you give
> me a hint of what you do to trigger it ? Boot time ? specific
> workload ?

I have not been able to reproduce the problem :( I haven't had a lot of
time to run tests... but every once in a while I have been firing off
test runs.

Maybe writing this email will get it to crash ;)

Cheers,
   Sean