Top 9 kernel oopses/warnings for the week of December 29th, 2007

Top 9 kernel oopses/warnings for the week of December 29th, 2007

[Arjan van de Ven]

The http://www.kerneloops.org website collects kernel oops and
warning reports from various mailing lists and bugzillas as well as
with a client users can install to auto-submit oopses.
Below is a top 10 list of the oopses collected in the last 7 days.
(Reports prior to 2.6.23 have been omitted in collecting the top 10)

It has been a quiet week due to the holidays, only 55 oops traces
have been collected.

Rank 1: sysctl_head_finish
	sysctl table check failed
	Reported 7 times
	Only reported for the proprietary madwifi driver
	More info: http://www.kerneloops.org/search.php?search=sysctl_head_finish
	
Rank 2: __ieee80211_rx
	Warning at net/mac80211/rx.c:1672
	Reported 5 times
	Only seen in 2.6.24-rc6 so far
	Johannes has diagnosed this as a driver bug in the iwlwifi drivers
	More info: http://www.kerneloops.org/search.php?search=__ieee80211_rx
	
Rank 3: rb_erase
	Null pointer dereference
	Reported 4 times
	Only seen in 2.6.24-rc6 so far
	(Note: the reported removed the oops end marker manually; may be duplicate reports)
	More info: http://www.kerneloops.org/guilty.php?version=2.6.24-rc6&guilty=rb_erase&start=1574400&end=1574400

Rank 4: remove_proc_entry
	Warning at fs/proc/generic.c:736
	Reported 4 times for 2.6.24-rc4 and rc5 (Mandrake)
	Seems to be a module unload issue
	More info: http://www.kerneloops.org/search.php?search=remove_proc_entry
	
Rank 5: aes_encrypt
	Invalid kernel pointer oops
	Reported 2 times for 2.6.23.11
	More info: http://www.kerneloops.org/search.php?search=aes_encrypt
	
Rank 6: cascade
	BUG statement at kernel/timer.c:606
	Reported 2 times for 2.6.23 (Gentoo)
	More info: http://www.kerneloops.org/search.php?search=cascade
	
Rank 7: kref_put
	Kernel paging request
	Reported 2 times, only for 2.6.24-rc6-mm1
	Diagnosed as a bug in the md-subsystem, patch available
	More info: http://www.kerneloops.org/search.php?search=kref_put
	More info: http://article.gmane.org/gmane.linux.kernel/618056
	
Rank 8: device_release
	Warning at drivers/base/core.c:107
	Reported 2 times last week, 6 incidents total
	Seems to be a bug in the MTD physmap driver and the hgafb driver
	More info: http://article.gmane.org/gmane.linux.kernel/618076
	More info: http://article.gmane.org/gmane.linux.kernel/618076
	
Rank 9: mutex_lock
	Null pointer dereference
	Reported 2 times, of which one was a duplicate report
	Seems to be related to running PPP over a bluetooth link
	More info: http://www.kerneloops.org/guilty.php?version=2.6.24-rc6&guilty=mutex_lock&start=1574400&end=1574400
	

kerneloops.org news:
* As of 2.6.24-rc6, oopses have a unique identifier which makes it possible to filter out duplicate
   reports of (and replies to) the same oops. Unfortunately some reporters remove this line from
   their reports to lkml.

Re: Top 9 kernel oopses/warnings for the week of December 29th, 2007

[Ingo Molnar]

* Arjan van de Ven <arjan@linux.intel.com> wrote:

> kerneloops.org news:
>
> * As of 2.6.24-rc6, oopses have a unique identifier which makes it 
>   possible to filter out duplicate reports of (and replies to) the 
>   same oops. Unfortunately some reporters remove this line from their 
>   reports to lkml.

i'd not mind printing out the bootup-ID as a regular part of the oops, 
as long as it's put into an unintrusive place, preferably without adding 
a new line. (just use an empty spot somewhere)

for example:

Call Trace: (ID afbf0104e2079481)
 [<ffffffff805079f6>] sysdev_driver_unregister+0x5b/0x91
 [<ffffffff806f2ef4>] cpufreq_register_driver+0x15a/0x19b
 [<ffffffff80cb0586>] powernowk8_init+0x86/0x94

	Ingo

Re: Top 9 kernel oopses/warnings for the week of December 29th, 2007

[Linus Torvalds]

On Sat, 29 Dec 2007, Arjan van de Ven wrote:
>
> It has been a quiet week due to the holidays, only 55 oops traces
> have been collected.

This would be more useful if it was more readable. As it is, you seem to 
have some formatting errors in your automation, where the things are 
incorrectly grouped:

> Rank 1: sysctl_head_finish
> 	sysctl table check failed
> 	Reported 7 times
> 	Only reported for the proprietary madwifi driver
> 	More info: http://www.kerneloops.org/search.php?search=sysctl_head_finish
> 	Rank 2: __ieee80211_rx
> 	Warning at net/mac80211/rx.c:1672

ie notice how that "Rank 2" got reported as if it was part of the "Rank 1" 
issue. 

		Linus

Re: Top 9 kernel oopses/warnings for the week of December 29th, 2007

[Arjan van de Ven]

Linus Torvalds wrote:
> 
> On Sat, 29 Dec 2007, Arjan van de Ven wrote:
>> It has been a quiet week due to the holidays, only 55 oops traces
>> have been collected.
> 
> This would be more useful if it was more readable. As it is, you seem to 
> have some formatting errors in your automation, where the things are 
> incorrectly grouped:

this is manual made; I'll blame thunderbird for now

let me see if I can send it again without corruption

> 
>> Rank 1: sysctl_head_finish
>> 	sysctl table check failed
>> 	Reported 7 times
>> 	Only reported for the proprietary madwifi driver
>> 	More info: http://www.kerneloops.org/search.php?search=sysctl_head_finish
>> 	Rank 2: __ieee80211_rx
>> 	Warning at net/mac80211/rx.c:1672
> 
> ie notice how that "Rank 2" got reported as if it was part of the "Rank 1" 
> issue. 
> 
> 		Linus

Re: Top 9 kernel oopses/warnings for the week of December 29th, 2007

[Arjan van de Ven]

Linus Torvalds wrote:
> 
> On Sat, 29 Dec 2007, Arjan van de Ven wrote:
>> It has been a quiet week due to the holidays, only 55 oops traces
>> have been collected.
> 
> This would be more useful if it was more readable. As it is, you seem to 
> have some formatting errors in your automation, where the things are 
> incorrectly grouped:
> 
>> Rank 1: sysctl_head_finish
>> 	sysctl table check failed
>> 	Reported 7 times
>> 	Only reported for the proprietary madwifi driver
>> 	More info: http://www.kerneloops.org/search.php?search=sysctl_head_finish
>> 	Rank 2: __ieee80211_rx
>> 	Warning at net/mac80211/rx.c:1672
> 
> ie notice how that "Rank 2" got reported as if it was part of the "Rank 1" 
> issue. 
> 


hmmm.. the copy in my Sent folder looks fine, as does the one in the lkml archive:
http://lkml.org/lkml/2007/12/29/41

This is distinctly weird.

Re: Top 9 kernel oopses/warnings for the week of December 29th, 2007

[Linus Torvalds]

On Sat, 29 Dec 2007, Arjan van de Ven wrote:
> 
> hmmm.. the copy in my Sent folder looks fine, as does the one in the lkml
> archive:
> http://lkml.org/lkml/2007/12/29/41
> 
> This is distinctly weird.

Ahh, it seems to be a alpine bug. Probably brought on by alpine trying to 
highlight the web addresses.

It doesn't always happen - between Rank 3 and Rank 4, you have an empty 
line, and alpine reacted correctly to that one, but the "empty" line 
between Rank 1 and Rank 2 actually contained a single TAB, and that 
apparently made alpine really confused.

So never mind. I'll make a bug-report on alpine, it wasn't a bug in your 
email.

		Linus

Re: Top 9 kernel oopses/warnings for the week of December 29th, 2007

[Martin Josefsson]

On Sat, 29 Dec 2007, Linus Torvalds wrote:

> Ahh, it seems to be a alpine bug. Probably brought on by alpine trying to
> highlight the web addresses.
>
> It doesn't always happen - between Rank 3 and Rank 4, you have an empty
> line, and alpine reacted correctly to that one, but the "empty" line
> between Rank 1 and Rank 2 actually contained a single TAB, and that
> apparently made alpine really confused.
>
> So never mind. I'll make a bug-report on alpine, it wasn't a bug in your
> email.

FYI, exactly the same thing happens with pine.

/Martin

Re: Top 9 kernel oopses/warnings for the week of December 29th, 2007

[Geert Uytterhoeven]

On Sat, 29 Dec 2007, Linus Torvalds wrote:
> On Sat, 29 Dec 2007, Arjan van de Ven wrote:
> > hmmm.. the copy in my Sent folder looks fine, as does the one in the lkml
> > archive:
> > http://lkml.org/lkml/2007/12/29/41
> > 
> > This is distinctly weird.
> 
> Ahh, it seems to be a alpine bug. Probably brought on by alpine trying to 
> highlight the web addresses.
> 
> It doesn't always happen - between Rank 3 and Rank 4, you have an empty 
> line, and alpine reacted correctly to that one, but the "empty" line 
> between Rank 1 and Rank 2 actually contained a single TAB, and that 
> apparently made alpine really confused.
> 
> So never mind. I'll make a bug-report on alpine, it wasn't a bug in your 
> email.

The bug is also present in the original branch, called pine 4.64 ;-)

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds

Re: Top 9 kernel oopses/warnings for the week of December 29th, 2007

[Björn Steinbrink]

On 2007.12.29 11:18:18 -0800, Linus Torvalds wrote:
> 
> 
> On Sat, 29 Dec 2007, Arjan van de Ven wrote:
> >
> > It has been a quiet week due to the holidays, only 55 oops traces
> > have been collected.
> 
> This would be more useful if it was more readable. As it is, you seem to 
> have some formatting errors in your automation, where the things are 
> incorrectly grouped:
> 
> > Rank 1: sysctl_head_finish
> > 	sysctl table check failed
> > 	Reported 7 times
> > 	Only reported for the proprietary madwifi driver
> > 	More info: http://www.kerneloops.org/search.php?search=sysctl_head_finish
> > 	Rank 2: __ieee80211_rx
> > 	Warning at net/mac80211/rx.c:1672
> 
> ie notice how that "Rank 2" got reported as if it was part of the "Rank 1" 
> issue. 

Shows up correctly here. Random guess is that your MUA treats tabs like
spaces when it sees format=flowed. The lines between the different ranks
weren't empty but had a single tab. And for whatever reason (I don't see
any trailing spaces), Arjan's mail had a format=flowed header.

AFAIK, only SPACE CRLF is valid for line-continuation with
format=flowed, not TAB CRLF.  So the above formatting would be the
result of a pointless format=flowed header meeting a MUA that was too
eager searching for line continuations.

Björn