From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ingo Molnar Subject: Re: [Bug #11035] System hangs on 2.6.26-rc8 Date: Sat, 19 Jul 2008 00:25:12 +0200 Message-ID: <20080718222512.GD31073@elte.hu> References: <4878D259.7050403@r000n.net> <487CA8E5.8020208@r000n.net> <20080718071121.GB6875@elte.hu> <19f34abd0807180028y2be61c06ueab5fcc305956f80@mail.gmail.com> Mime-Version: 1.0 Return-path: Content-Disposition: inline In-Reply-To: <19f34abd0807180028y2be61c06ueab5fcc305956f80-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org> Sender: kernel-testers-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Vegard Nossum Cc: Roman Mindalev , "Rafael J. Wysocki" , Linux Kernel Mailing List , Kernel Testers List , Thomas Gleixner * Vegard Nossum wrote: > BTW, I didn't consider that argument (of continuing as far as > possible) before, but it's a good one; if we don't crash completely, > the user can still copy the log we have a better report of it. I guess > kerneloops.org is currently missing out a great deal of reports which > all shut down the machine immediately without a chance to go into the > log. yes. There are two techniques to improve the 'yield' of kerneloops.org: 1) make a better job of getting the logs off the box 2) make a better job of not crashing the box when we can do better. For example lockdep tries very hard to never crash the box. It is a feature that warns about a chance of a lockup, not about a lockup itself - so crashing the box at the point of the bug detection is counter-productive. The same applies to DEBUG_PAGEALLOC as well: technically nobody (but the buggy code itself) is hurt by accessing already freed data. So we could try and let it run. (Btw., this might be a way to share a mechanism between kmemcheck and DEBUG_PAGEALLOC, and make kmemcheck more useful to the general kernel as a whole.) Ingo