On Fri, 2011-03-04 at 14:27 +0100, Frédéric Boiteux wrote: > Le Thu, 3 Mar 2011 17:46:38 +0100, > Frédéric Boiteux <fboiteux@calistel.com> a écrit : > > > Hi, > > > > It was a false assertion, I did get hibernation resuming fail later > > with this commit reversed :-(( Indeed, the bug is quite difficult to > > verify, I've done 2/3 hibernation/resume loops before stating a > > bissecting kernel was good, perhaps it was not enough ! > > I'll retry, starting from last 'good' kernel before this. > > [ Sorry for my previous e-mail, I didn't finished it ] > > I've done more tests, and finally got this commit as faulty for > hibernation resuming : > > ee7181e2b0625ffaa8a7b3dd19dd0c1684cc2d70 is first bad commit > commit ee7181e2b0625ffaa8a7b3dd19dd0c1684cc2d70 > Author: Debian kernel team <debian-kernel@lists.debian.org> > Date: Thu Feb 17 22:22:25 2011 +0000 > > Apply "2.6.32.28.patch" Thanks for the confirmation. [...] > I've tried at least 5 hibernation / resume cycles with a 2.6.32-30 > version without this commit (with some parts of patch not reversed > cleanly, but seemed not a problem) , waiting for some time unplugged, > and it seems to work. > > Perhaps we could bisect further inside this multi-purpose patch, but > I don't know how to do it. You can find all the individual patches as a quilt series in: git://git.kernel.org/pub/scm/linux/kernel/git/longterm/longterm-queue-2.6.32.git subdirectory releases/2.6.32.28. Since quilt doesn't have a bisect command you would have to have to keep track of good and bad positions in the series yourself. I would guess that it's most likely one of: releases/2.6.32.28/pm-hibernate-fix-pm_post_-notification-with-user-space-suspend.patch releases/2.6.32.28/x86-enable-the-intr-remap-fault-handling-after-local-apic-setup.patch so you could try reverting those before going through the whole series. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse.
Attachment:
signature.asc
Description: This is a digitally signed message part