Suspend regression from -12 to -13 (includes -14 as well)

[Tim Hull]

Hi,

I've been testing Gutsy on my MacBook (Core Duo) since before the beta
release, and I recently ran into issues with suspend with a recent kernel
update.
While suspend/resume worked fine for me from early Tribe releases up to and
including the 2.6.22-12 kernel, with the 2.6.22-13 kernel I have begun
experiencing random difficulty resuming from suspend-to-RAM.  Basically, I
can suspend fine, and most of the time I can resume fine, but about one out
of every 4 or 5
resume attempts is unsuccessful - i.e. the video never POSTs and the system
is unresponsive despite the fans running.  I have reported this on launchpad
(bug #151016).  Furthermore, a user has reported a similar problem in bug
#127101 which may be related to this.

I was initially unsure of the cause of this - at first I suspected X - but
now that I have a working preliminary fix to my X-server's random crashing
problem (bug #127101, for anyone who is interested in random X-server
crashes), it is evident that the problem was in fact introduced by the
2.6.22-13 kernel and persists with -14.

Anyway, I did look at the changelog for 2.6.22-13, and came across this
change which may have some relevance here...

  * clockevents: remove the suspend/resume workaround^Wthinko

Also, there is some other ACPI-related changes in there and a sync with the
upstream kernel - any which could have introduced the bug.

Has anybody seen this bug, or have any idea as to what specifically
introduced it?  Is there any additional information I can provide to help
resolve this?  I apologize if this list isn't the best place to bring this
up, though it is the one central kernel-related contact I could find.  I
hope this can get resolved before the release - either by a reversion of
some patches or a fix...

Tim
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/kernel-team/attachments/20071011/94477f0d/attachment.html>