Re: Top 10 kernel oopses for the week ending January 5th, 2008

[Linus Torvalds <torvalds@linux-foundation.org>]

On Tue, 8 Jan 2008, Arjan van de Ven wrote:
> 
> I've made life easier for those using the www.kerneloops.org website;
> at least for x86 oopses the website now does this for you and shows
> the decoded Code: line in the raw oops data:
> 
> http://www.kerneloops.org/raw.php?rawid=2716

Cool.

One thing I wonder about - could you separate out the bug-ons and warnings 
from the oopses? They really are different issues, and an oops with 
register information etc is very different from a BUG() with line numbers, 
which in turn is very different from a WARN_ON().

Right now, it says

	Oopses reported for kernel 2.6.24-rc7


	7 oopses reported

	hfsplus_releasepage	3
	enqueue_task		1
	lock_acquire		1
	__hfs_brec_find		1
	__ieee80211_rx		1

and in fact three of those five entries are really WARN_ON's. It would be 
nicer if it would look more along the lines of

	Backtraces reported for kernel 2.6.24-rc7


	4 oopses reported

	hfsplus_releasepage     3
	__hfs_brec_find         1


	3 warnings repored

	enqueue_task            1
	lock_acquire            1
	__ieee80211_rx          1

because those things really don't have the same kind of impact at all, and 
tend to be very different to debug (a "BUG_ON()" is perhaps somewhat 
closer to an oops, but a WARN_ON() is definitely in a class of its own).

On that "Code:" side, it seems there is still some problem with oops 
parsing. See for example:

	http://www.kerneloops.org/raw.php?rawid=1521&msgid=http://mid.gmane.org/20071017154655.GA13394@elte.hu

and notice how the Code: never made it into the raw message (and thus 
there is also no instruction disassembly).

			Linus