On Thu, Jan 10, 2013 at 11:59:48AM +0100, Adrian Byszuk wrote: > I think I won't get the answer this way. Bisecting between > 3.7 and 3.8-rc2 (from linus tree) led me to commit > 63e1ed2364050073770c085021377d7764969b85 titled "Update vt8500 PWM > driver support", which I think is rather nonsense. Also, depending on > bisected version, sometimes I got kernel panic, but sometimes (more > often) whole system would just freeze completely. > > Any other way diagnose this bug?
Hmm, bisection can sometimes lead us in the fields, nasty. Well, you could try bisecting the major releases in between. I.e., test 3.8-rc1, and then test vanilla 3.7. For example, if you still trigger it with -rc1, then no need to bisect the -rc2..-rc1 range. If you still trigger it with 3.7 then not good - in that case you'd need to go further back in time. The best would be if you don't trigger it with 3.7 but trigger it with 3.7.1 :) Then it is easy. Also, git bisect skip can skip ranges of commits. You could try that too but you need to visialize how exactly the tree structure is and I admit, I haven't figured out the details of that yet but Linus has explained this approach a couple of times, you might want to take a look at this http://www.yarchive.net/comp/linux/git_bisect.html or search the net for even better ideas. Also, you said "with 3.8.0-rc2 it looks like it's working correctly - I've tried the unplug/plug sequence more than 5 times and it seems to work fine." Please test that again and repeat that sequence say 15-20 times, to make sure that -rc2 is really ok and not something changed timing and this bug is being only brown-paper-baggy-covered now. I mean, I'm just guessing - you're in front of the machine, you know best what to do to trigger the bug reliably. HTH. -- Regards/Gruss, Boris. Sent from a fat crate under my desk. Formatting is fine. -- -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/