Kernel BUG at mm/rmap.c:491

Kernel BUG at mm/rmap.c:491

[Chen, Kenneth W]

Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.

- Ken


Bad page state at free_hot_cold_page (in process 'sh', page ffff81000482dde8)
flags:0x8000000000000000 mapping:0000000000000000 mapcount:1 count:0
Backtrace:

Call Trace:<ffffffff8014ef8c>{bad_page+115} <ffffffff8014f71f>{free_hot_cold_page+120}
       <ffffffff8014f7d3>{__pagevec_free+33} <ffffffff80154ecc>{release_pages+369}
       <ffffffff80154fc4>{__pagevec_lru_add_active+230} <ffffffff8015b14f>{__handle_mm_fault+680}
       <ffffffff801b0f03>{journal_stop+487} <ffffffff80396970>{do_page_fault+992}
       <ffffffff8013a3bf>{do_sigaction+110} <ffffffff8013a849>{sys_rt_sigaction+132}
       <ffffffff8010e249>{error_exit+0} 
Trying to fix it up, but a reboot is needed
Bad page state at free_hot_cold_page (in process 'sh', page ffff8100049d0f78)
flags:0x8000000000000000 mapping:0000000000000000 mapcount:1 count:0
Backtrace:

Call Trace:<ffffffff8014ef8c>{bad_page+115} <ffffffff8014f71f>{free_hot_cold_page+120}
       <ffffffff8014f7d3>{__pagevec_free+33} <ffffffff80154ecc>{release_pages+369}
       <ffffffff80154fc4>{__pagevec_lru_add_active+230} <ffffffff8015b14f>{__handle_mm_fault+680}
       <ffffffff801b0f03>{journal_stop+487} <ffffffff80396970>{do_page_fault+992}
       <ffffffff8013a3bf>{do_sigaction+110} <ffffffff8013a849>{sys_rt_sigaction+132}
       <ffffffff8010e249>{error_exit+0} 
Trying to fix it up, but a reboot is needed
Bad page state at free_hot_cold_page (in process 'sh', page ffff8100049d0f40)
flags:0x8000000000000004 mapping:0000000000000000 mapcount:1 count:0
Backtrace:

Call Trace:<ffffffff8014ef8c>{bad_page+115} <ffffffff8014f71f>{free_hot_cold_page+120}
       <ffffffff8014f7d3>{__pagevec_free+33} <ffffffff80154ecc>{release_pages+369}
       <ffffffff80154fc4>{__pagevec_lru_add_active+230} <ffffffff8015b14f>{__handle_mm_fault+680}
       <ffffffff801b0f03>{journal_stop+487} <ffffffff80396970>{do_page_fault+992}
       <ffffffff8013a3bf>{do_sigaction+110} <ffffffff8013a849>{sys_rt_sigaction+132}
       <ffffffff8010e249>{error_exit+0} 
Trying to fix it up, but a reboot is needed
sh[16651]: segfault at 0000000000000028 rip 000000000042c830 rsp 00007fffffa8bad8 error 4
Kernel BUG at mm/swap.c:218
invalid operand: 0000 [1] SMP 
CPU 1 
Modules linked in:
Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
RIP: 0010:[<ffffffff80154db0>] <ffffffff80154db0>{release_pages+85}
RSP: 0018:ffff81011569dca8  EFLAGS: 00010257
RAX: 0000000000000000 RBX: ffff8100049d0fb0 RCX: 0000000000000000
RDX: ffff8100049d0fb0 RSI: 0000000000000010 RDI: ffff810004f38298
RBP: 0000000000000003 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000001 R12: 0000000000000000
R13: 0000000000000010 R14: ffff810004f38298 R15: ffff81011569dca8
FS:  00002aaaaaac7b00(0000) GS:ffffffff804e1880(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 000000390128f070 CR3: 00000001034a4000 CR4: 00000000000006e0
Process cc1 (pid: 16500, threadinfo ffff81011569c000, task ffff810116cd03c0)
Stack: 0000000000000000 0000000000000000 ffff810004bdc960 ffff810004ceea70 
       ffff810004aae9d8 ffff810004b2dc70 ffff810004b8f1f0 ffff810004a46730 
       ffff810004d5f090 ffff8100048c5298 
Call Trace:<ffffffff80162597>{free_pages_and_swap_cache+116} <ffffffff80159aa6>{unmap_vma
s+1484}
       <ffffffff8015e98a>{exit_mmap+121} <ffffffff8012c7b0>{mmput+38}
       <ffffffff80130c10>{do_exit+462} <ffffffff8020cbf3>{__up_write+34}
       <ffffffff8013169f>{sys_exit_group+0} <ffffffff8010d50e>{system_call+126}
       

Code: 0f 0b 68 ca c9 3b 80 c2 da 00 f0 83 43 08 ff 0f 98 c0 84 c0 
RIP <ffffffff80154db0>{release_pages+85} RSP <ffff81011569dca8>
 ----------- [cut here ] --------- [please bite here ] ---------
Fixing recursive fault but reboot is needed!
Kernel BUG at mm/rmap.c:491
invalid operand: 0000 [2] SMP 
CPU 3 
Modules linked in:
Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
RIP: 0010:[<ffffffff80160b33>] <ffffffff80160b33>{page_remove_rmap+21}
RSP: 0000:ffff810104a89bf0  EFLAGS: 00010286
RAX: 00000000ffffffff RBX: ffff81010466d648 RCX: 000000010466d000
RDX: ffff810001000000 RSI: 0000000000000001 RDI: ffff8100049d0f40
RBP: 00000000006c9000 R08: ffff810104a89d00 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000030001 R12: 80000001084d8067
R13: ffff8100049d0f40 R14: ffff810004f38280 R15: ffff810104a89cb8
FS:  0000000000000000(0000) GS:ffffffff804e1980(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000028 CR3: 0000000113261000 CR4: 00000000000006e0
Process sh (pid: 16651, threadinfo ffff810104a88000, task ffff810119681890)
Stack: ffffffff80159a52 ffff810004f38280 00000000006dcfff 00000000006dcfff 
       00000000006dcfff ffff8100048f67e8 00000000fffffff2 ffff81010e1a9c40 
       00000000006dd000 ffff81010a638018 
Call Trace:<ffffffff80159a52>{unmap_vmas+1400} <ffffffff8015e98a>{exit_mmap+121}
       <ffffffff8012c7b0>{mmput+38} <ffffffff80130c10>{do_exit+462}
       <ffffffff80137ff7>{__dequeue_signal+435} <ffffffff8013169f>{sys_exit_group+0}
       <ffffffff801399fd>{get_signal_to_deliver+1163} <ffffffff80138406>{specific_send_sig_info+168}
       <ffffffff8010c9ed>{do_signal+109} <ffffffff80182dc9>{mntput_no_expire+28}
       <ffffffff80182dc9>{mntput_no_expire+28} <ffffffff8016896e>{sys_access+247}
       <ffffffff80172263>{sys_newstat+33} <ffffffff8010db78>{retint_signal+61}
       

Code: 0f 0b 68 b1 ca 3b 80 c2 eb 01 48 c7 c6 ff ff ff ff bf 20 00 
RIP <ffffffff80160b33>{page_remove_rmap+21} RSP <ffff810104a89bf0>
 <1>Fixing recursive fault but reboot is needed!


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Con Kolivas]

Chen, Kenneth W writes:

> Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.

> Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3

> Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3

                       ^^^^^^^^^^

Please try to reproduce it without proprietary binary modules linked in.

Con

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

RE: Kernel BUG at mm/rmap.c:491

[Chen, Kenneth W]

Con Kolivas wrote on Wednesday, November 23, 2005 3:24 PM
> Chen, Kenneth W writes:
> 
> > Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
> > 
> > Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
> > 
> > Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
> 
>                        ^^^^^^^^^^
> 
> Please try to reproduce it without proprietary binary modules linked in.


???, I'm not using any modules at all.

[albat]$ /sbin/lsmod
Module                  Size  Used by
[albat]$ 


Also, isn't it 'P' indicate proprietary module, not 'G'?
line 159: kernel/panic.c:

        snprintf(buf, sizeof(buf), "Tainted: %c%c%c%c%c%c",
                tainted & TAINT_PROPRIETARY_MODULE ? 'P' : 'G',

- Ken

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Alistair John Strachan]

On Wednesday 23 November 2005 23:24, Con Kolivas wrote:
> Chen, Kenneth W writes:
> > Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
> >
> > Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
> >
> > Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
>
>                        ^^^^^^^^^^
>
> Please try to reproduce it without proprietary binary modules linked in.

AFAIK "G" means all loaded modules are GPL, P is for proprietary modules.

FWIW, I reported a similarly located BUG() in this file (line 487 in 2.6.14) a 
couple of weeks ago. I believe there is a problem lurking somewhere.

-- 
Cheers,
Alistair.

'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Con Kolivas]

Chen, Kenneth W writes:

> Con Kolivas wrote on Wednesday, November 23, 2005 3:24 PM
>> Chen, Kenneth W writes:
>> 
>> > Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
>> > 
>> > Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
>> > 
>> > Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
>> 
>>                        ^^^^^^^^^^
>> 
>> Please try to reproduce it without proprietary binary modules linked in.
> 
> 
> ???, I'm not using any modules at all.
> 
> [albat]$ /sbin/lsmod
> Module                  Size  Used by
> [albat]$ 
> 
> 
> Also, isn't it 'P' indicate proprietary module, not 'G'?
> line 159: kernel/panic.c:
> 
>         snprintf(buf, sizeof(buf), "Tainted: %c%c%c%c%c%c",
>                 tainted & TAINT_PROPRIETARY_MODULE ? 'P' : 'G',

Sorry it's not proprietary module indeed. But what is tainting it?

Con

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

RE: Kernel BUG at mm/rmap.c:491

[Randy.Dunlap]

On Wed, 23 Nov 2005, Chen, Kenneth W wrote:

> Con Kolivas wrote on Wednesday, November 23, 2005 3:24 PM
> > Chen, Kenneth W writes:
> >
> > > Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
> > >
> > > Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
> > >
> > > Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
> >
> >                        ^^^^^^^^^^
> >
> > Please try to reproduce it without proprietary binary modules linked in.
>
>
> ???, I'm not using any modules at all.
>
> [albat]$ /sbin/lsmod
> Module                  Size  Used by
> [albat]$
>
>
> Also, isn't it 'P' indicate proprietary module, not 'G'?

Yes.  It's the 'B' that is tainting in this case:
TAINT_BAD_PAGE is set.

> line 159: kernel/panic.c:
>
>         snprintf(buf, sizeof(buf), "Tainted: %c%c%c%c%c%c",
>                 tainted & TAINT_PROPRIETARY_MODULE ? 'P' : 'G',

-- 
~Randy

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Alistair John Strachan]

On Wednesday 23 November 2005 23:35, Con Kolivas wrote:
> Chen, Kenneth W writes:
> > Con Kolivas wrote on Wednesday, November 23, 2005 3:24 PM
> >
> >> Chen, Kenneth W writes:
> >> > Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
> >> >
> >> > Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
> >> >
> >> > Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
> >>
> >>                        ^^^^^^^^^^
> >>
> >> Please try to reproduce it without proprietary binary modules linked in.
> >
> > ???, I'm not using any modules at all.
> >
> > [albat]$ /sbin/lsmod
> > Module                  Size  Used by
> > [albat]$
> >
> >
> > Also, isn't it 'P' indicate proprietary module, not 'G'?
> > line 159: kernel/panic.c:
> >
> >         snprintf(buf, sizeof(buf), "Tainted: %c%c%c%c%c%c",
> >                 tainted & TAINT_PROPRIETARY_MODULE ? 'P' : 'G',
>
> Sorry it's not proprietary module indeed. But what is tainting it?

Probably a prior oops or some other marked error condition.

-- 
Cheers,
Alistair.

'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

RE: Kernel BUG at mm/rmap.c:491

[Randy.Dunlap]

On Wed, 23 Nov 2005, Randy.Dunlap wrote:

> On Wed, 23 Nov 2005, Chen, Kenneth W wrote:
>
> > Con Kolivas wrote on Wednesday, November 23, 2005 3:24 PM
> > > Chen, Kenneth W writes:
> > >
> > > > Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
> > > >
> > > > Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
> > > >
> > > > Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
> > >
> > >                        ^^^^^^^^^^
> > >
> > > Please try to reproduce it without proprietary binary modules linked in.
> >
> >
> > ???, I'm not using any modules at all.
> >
> > [albat]$ /sbin/lsmod
> > Module                  Size  Used by
> > [albat]$
> >
> >
> > Also, isn't it 'P' indicate proprietary module, not 'G'?
>
> Yes.  It's the 'B' that is tainting in this case:
> TAINT_BAD_PAGE is set.

in only one place:
./mm/page_alloc.c:148:  add_taint(TAINT_BAD_PAGE);


> > line 159: kernel/panic.c:
> >
> >         snprintf(buf, sizeof(buf), "Tainted: %c%c%c%c%c%c",
> >                 tainted & TAINT_PROPRIETARY_MODULE ? 'P' : 'G',
>
>

-- 
~Randy

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Con Kolivas]

[-- Attachment #1: Type: text/plain, Size: 1165 bytes --]

Alistair John Strachan writes:

> On Wednesday 23 November 2005 23:35, Con Kolivas wrote:
>> Chen, Kenneth W writes:
>> > Con Kolivas wrote on Wednesday, November 23, 2005 3:24 PM
>> >
>> >> Chen, Kenneth W writes:
>> >> > Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
>> >> >
>> >> > Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
>> >> >
>> >> > Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
>> >>
>> >>                        ^^^^^^^^^^
>> >>
>> >> Please try to reproduce it without proprietary binary modules linked in.
>> >
>> > ???, I'm not using any modules at all.
>> >
>> > [albat]$ /sbin/lsmod
>> > Module                  Size  Used by
>> > [albat]$
>> >
>> >
>> > Also, isn't it 'P' indicate proprietary module, not 'G'?
>> > line 159: kernel/panic.c:
>> >
>> >         snprintf(buf, sizeof(buf), "Tainted: %c%c%c%c%c%c",
>> >                 tainted & TAINT_PROPRIETARY_MODULE ? 'P' : 'G',
>>
>> Sorry it's not proprietary module indeed. But what is tainting it?
> 
> Probably a prior oops or some other marked error condition.

My humble apologies! Force of habit when seeing tainted message which 
comes up so often :(

Con


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Kernel BUG at mm/rmap.c:491

[Dave Jones]

On Thu, Nov 24, 2005 at 10:35:46AM +1100, Con Kolivas wrote:
 > Chen, Kenneth W writes:
 > 
 > >Con Kolivas wrote on Wednesday, November 23, 2005 3:24 PM
 > >>Chen, Kenneth W writes:
 > >>
 > >>> Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
 > >>> 
 > >>> Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
 > >>> 
 > >>> Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
 > >
 > >Also, isn't it 'P' indicate proprietary module, not 'G'?
 > >line 159: kernel/panic.c:
 > >
 > >        snprintf(buf, sizeof(buf), "Tainted: %c%c%c%c%c%c",
 > >                tainted & TAINT_PROPRIETARY_MODULE ? 'P' : 'G',
 > 
 > Sorry it's not proprietary module indeed. But what is tainting it?
 
'B' = bad page ;)

		Dave

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Dave Jones]

On Wed, Nov 23, 2005 at 11:35:15PM +0000, Alistair John Strachan wrote:
 > On Wednesday 23 November 2005 23:24, Con Kolivas wrote:
 > > Chen, Kenneth W writes:
 > > > Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
 > > >
 > > > Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
 > > >
 > > > Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
 > >
 > >                        ^^^^^^^^^^
 > >
 > > Please try to reproduce it without proprietary binary modules linked in.
 > 
 > AFAIK "G" means all loaded modules are GPL, P is for proprietary modules.


The 'G' seems to confuse a hell of a lot of people.
(I've been asked about it when people got machine checks a lot over
 the last few months).

Would anyone object to changing it to conform to the style of
the other taint flags ? Ie, change it to ' ' ?

		Dave

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Lee Revell]

On Wed, 2005-11-23 at 23:40 -0500, Dave Jones wrote:
> The 'G' seems to confuse a hell of a lot of people.
> (I've been asked about it when people got machine checks a lot over
>  the last few months).
> 
> Would anyone object to changing it to conform to the style of
> the other taint flags ? Ie, change it to ' ' ? 

While you're at it why not print a big loud warning that says not to
post the Oops to LKML, and instructing the user to reproduce with a
clean kernel, if the P flag is set?  Presumably the reason for the terse
output is to get the maximum possible debug information on the screen,
but we don't care about stack traces for tainted kernels anyway.  

Something must need fixing, as the volume of tainted Oops reports shows
no sign of diminishing, and the users aren't getting any less pissy when
you tell them to come back with a clean bug report.

Lee 

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Alistair John Strachan]

On Thursday 24 November 2005 04:40, Dave Jones wrote:
> On Wed, Nov 23, 2005 at 11:35:15PM +0000, Alistair John Strachan wrote:
>  > On Wednesday 23 November 2005 23:24, Con Kolivas wrote:
>  > > Chen, Kenneth W writes:
>  > > > Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
>  > > >
>  > > > Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
>  > > >
>  > > > Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
>  > >
>  > >                        ^^^^^^^^^^
>  > >
>  > > Please try to reproduce it without proprietary binary modules linked
>  > > in.
>  >
>  > AFAIK "G" means all loaded modules are GPL, P is for proprietary
>  > modules.
>
> The 'G' seems to confuse a hell of a lot of people.
> (I've been asked about it when people got machine checks a lot over
>  the last few months).
>
> Would anyone object to changing it to conform to the style of
> the other taint flags ? Ie, change it to ' ' ?

I don't understand the reasons for making the tainted string all the same 
length anyway. Why not just remove all the extra spaces?

Unless you know what you're looking for, I can assure you that:

Tainted: G    B SOMEOTHERTEXT

Is not intuitively readable (which text does B belong to?).

Tainted:    B SOMEOTHERTEXT

Is better, but still not very good. Why not drop the spaces?

3rd party parsing purposes?

-- 
Cheers,
Alistair.

'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Hugh Dickins]

On Wed, 23 Nov 2005, Dave Jones wrote:
> On Wed, Nov 23, 2005 at 11:35:15PM +0000, Alistair John Strachan wrote:
>  > On Wednesday 23 November 2005 23:24, Con Kolivas wrote:
>  > > Chen, Kenneth W writes:
>  > > > Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
>  > > >
>  > > > Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
>  > > >
>  > > > Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
>  > >
>  > >                        ^^^^^^^^^^
>  > >
>  > > Please try to reproduce it without proprietary binary modules linked in.
>  > 
>  > AFAIK "G" means all loaded modules are GPL, P is for proprietary modules.
> 
> The 'G' seems to confuse a hell of a lot of people.
> (I've been asked about it when people got machine checks a lot over
>  the last few months).
> 
> Would anyone object to changing it to conform to the style of
> the other taint flags ? Ie, change it to ' ' ?

Please, please do: it's insane as is.  But I've CC'ed Keith,
we sometimes find the kernel does things so to suit ksymoops.

Hugh

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Hugh Dickins]

On Wed, 23 Nov 2005, Chen, Kenneth W wrote:
> Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
> 
> Bad page state at free_hot_cold_page (in process 'sh', page ffff81000482dde8)
> flags:0x8000000000000000 mapping:0000000000000000 mapcount:1 count:0
> Bad page state at free_hot_cold_page (in process 'sh', page ffff8100049d0f78)
> flags:0x8000000000000000 mapping:0000000000000000 mapcount:1 count:0
> Bad page state at free_hot_cold_page (in process 'sh', page ffff8100049d0f40)
> flags:0x8000000000000004 mapping:0000000000000000 mapcount:1 count:0
> Kernel BUG at mm/swap.c:218
> Kernel BUG at mm/rmap.c:491

Neither mm/rmap.c (page_remove_rmap) nor mm/swap.c (put_page_testzero)
BUG is interesting in this case, they're just side-effects of trying to
recover from the preceding "Bad page state"s.

Which are interesting.  Not at all the same case as the many recently
reported while we were fixing up PageReserved removal cases; though
yours will probably be related.

It could conceivably be an effect of a DRM pci_alloc_consistent issue
which Dave Airlie spotted yesterday; but not a typical case of it,
and I'm probably only thinking of that one because it's uppermost.

Please send your .config (I hope it's tailored somewhat to your machine,
rather than an allyesconfig or the like?) and bootup dmesg, in case they
help to narrow the search.  You were just running straight 2.6.15-rc2,
no additional patches?  Doing anything interesting just before this
happened?

Thanks,
Hugh

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Lee Revell]

On Thu, 2005-11-24 at 07:50 +0000, Hugh Dickins wrote:
> But I've CC'ed Keith,
> we sometimes find the kernel does things so to suit ksymoops. 

Um, unless someone has been merging Documentation patches without
reading them, ksymoops shouldn't be used with 2.6 anyway.

Lee

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Zwane Mwaikambo]

On Thu, 24 Nov 2005, Lee Revell wrote:

> On Wed, 2005-11-23 at 23:40 -0500, Dave Jones wrote:
> > The 'G' seems to confuse a hell of a lot of people.
> > (I've been asked about it when people got machine checks a lot over
> >  the last few months).
> > 
> > Would anyone object to changing it to conform to the style of
> > the other taint flags ? Ie, change it to ' ' ? 
> 
> While you're at it why not print a big loud warning that says not to
> post the Oops to LKML, and instructing the user to reproduce with a

I don't think wasting precious screen real estate on warnings is a good 
idea. The oops may also be of use, there have been occassions where the 
only oops output had a proprietary bit set. The person handling the bug 
report should be the one making the decision as to whether to repost a new 
oops.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Dave Jones]

On Thu, Nov 24, 2005 at 06:33:12AM -0500, Lee Revell wrote:
 > On Thu, 2005-11-24 at 07:50 +0000, Hugh Dickins wrote:
 > > But I've CC'ed Keith,
 > > we sometimes find the kernel does things so to suit ksymoops. 
 > 
 > Um, unless someone has been merging Documentation patches without
 > reading them, ksymoops shouldn't be used with 2.6 anyway.

It is occasionally still useful for decoding Code: lines back
to assembly.

		Dave

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Keith Owens]

On Thu, 24 Nov 2005 07:50:49 +0000 (GMT), 
Hugh Dickins <hugh@veritas.com> wrote:
>On Wed, 23 Nov 2005, Dave Jones wrote:
>> On Wed, Nov 23, 2005 at 11:35:15PM +0000, Alistair John Strachan wrote:
>>  > On Wednesday 23 November 2005 23:24, Con Kolivas wrote:
>>  > > Chen, Kenneth W writes:
>>  > > > Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
>>  > > >
>>  > > > Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
>>  > > >
>>  > > > Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
>>  > >
>>  > >                        ^^^^^^^^^^
>>  > >
>>  > > Please try to reproduce it without proprietary binary modules linked in.
>>  > 
>>  > AFAIK "G" means all loaded modules are GPL, P is for proprietary modules.
>> 
>> The 'G' seems to confuse a hell of a lot of people.
>> (I've been asked about it when people got machine checks a lot over
>>  the last few months).
>> 
>> Would anyone object to changing it to conform to the style of
>> the other taint flags ? Ie, change it to ' ' ?
>
>Please, please do: it's insane as is.  But I've CC'ed Keith,
>we sometimes find the kernel does things so to suit ksymoops.

'G' is not one of mine, I find it annoying as well.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Con Kolivas]

On Fri, 25 Nov 2005 10:47, Keith Owens wrote:
> On Thu, 24 Nov 2005 07:50:49 +0000 (GMT),
>
> Hugh Dickins <hugh@veritas.com> wrote:
> >On Wed, 23 Nov 2005, Dave Jones wrote:
> >> On Wed, Nov 23, 2005 at 11:35:15PM +0000, Alistair John Strachan wrote:
> >>  > On Wednesday 23 November 2005 23:24, Con Kolivas wrote:
> >>  > > Chen, Kenneth W writes:
> >>  > > > Has people seen this BUG_ON before?  On 2.6.15-rc2, x86-64.
> >>  > > >
> >>  > > > Pid: 16500, comm: cc1 Tainted: G    B 2.6.15-rc2 #3
> >>  > > >
> >>  > > > Pid: 16651, comm: sh Tainted: G    B 2.6.15-rc2 #3
> >>  > >
> >>  > >                        ^^^^^^^^^^
> >>  > >
> >>  > > Please try to reproduce it without proprietary binary modules
> >>  > > linked in.
> >>  >
> >>  > AFAIK "G" means all loaded modules are GPL, P is for proprietary
> >>  > modules.
> >>
> >> The 'G' seems to confuse a hell of a lot of people.
> >> (I've been asked about it when people got machine checks a lot over
> >>  the last few months).
> >>
> >> Would anyone object to changing it to conform to the style of
> >> the other taint flags ? Ie, change it to ' ' ?
> >
> >Please, please do: it's insane as is.  But I've CC'ed Keith,
> >we sometimes find the kernel does things so to suit ksymoops.
>
> 'G' is not one of mine, I find it annoying as well.

Would anyone object to changing it so that tainted only means Proprietary 
taint and use a different keyword for GPL tainting such as "Corrupted"?

Con

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Hugh Dickins]

On Fri, 25 Nov 2005, Con Kolivas wrote:
> On Fri, 25 Nov 2005 10:47, Keith Owens wrote:
> > On Thu, 24 Nov 2005 07:50:49 +0000 (GMT),
> > Hugh Dickins <hugh@veritas.com> wrote:
> > >On Wed, 23 Nov 2005, Dave Jones wrote:
> > >>
> > >> The 'G' seems to confuse a hell of a lot of people.
> > >> (I've been asked about it when people got machine checks a lot over
> > >>  the last few months).
> > >>
> > >> Would anyone object to changing it to conform to the style of
> > >> the other taint flags ? Ie, change it to ' ' ?
> > >
> > >Please, please do: it's insane as is.  But I've CC'ed Keith,
> > >we sometimes find the kernel does things so to suit ksymoops.
> >
> > 'G' is not one of mine, I find it annoying as well.
> 
> Would anyone object to changing it so that tainted only means Proprietary 
> taint and use a different keyword for GPL tainting such as "Corrupted"?

I don't see the point.  The system is in a dubious state, tainted is
the word we've been using for that, the flags indicate what's suspect,
why play with the wording further?  But replace 'G' by ' ' certainly.

Hugh

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: Kernel BUG at mm/rmap.c:491

[Con Kolivas]

On Fri, 25 Nov 2005 21:43, Hugh Dickins wrote:
> On Fri, 25 Nov 2005, Con Kolivas wrote:
> > Would anyone object to changing it so that tainted only means Proprietary
> > taint and use a different keyword for GPL tainting such as "Corrupted"?
>
> I don't see the point.  The system is in a dubious state, tainted is
> the word we've been using for that, the flags indicate what's suspect,
> why play with the wording further?

I was simply thinking of us confused users. No good reason otherwise.

Cheers,
Con

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>