On Fri, Feb 20, 2009 at 11:38:59PM -0300, Henrique de Moraes Holschuh wrote:
> On Mon, 16 Feb 2009, Marc MERLIN wrote:
> > Were you able to find any config options that don't look good to you?
>
> Sorry, no :( But you have some heavy duty debugging stuff that you should
> disable if you don't use them, like kprobes. I'd also disable cgroup and
> the group scheduler if you don't use them.
I've turned them off and built 2.6.28.7. We'll see...
> > Similarly, I have a nagging but annoying problem where my laptop will
> > occasionally hard deadlock when I plug the power back in while it was
> > running on batteries.
>
> If you are using either nVidia or ATI proprietary drivers (sorry, I forgot
> the details), they react to power management events like AC
> connected/disconnected, and so I suppose a bug there might cause hangs. If
> it is a hard-to-hit race, the hangs would not happen very often.
Yes, I know about them. Luckily I'm now able to run standard kernel ATI dri
and Xorg drivers.
> I don't think the open drivers do this (i.e. they're not advanced enough for
> that yet :p), though. Maybe Intel's do, they're big on the PM stuff. The
They have no hooks, but the open source drivers seem to take less power than
the ATI ones did, even with the PM clock throttling stuff.
> X.org server also wants to listen to acpi events nowadays (and does so,
> probably through HAL), so it gets difficult to know for sure what might be
> reacting to the AC connected event and causing a hang.
Since I have a recent X.org I think it does that internally indeed, and I'm
really wondering if they are linked.
> It could also be the BIOS fighting X.org. Try telling the BIOS to not touch
> the screen brightness level when you plug/unplug AC, and see if that stops
> the hangs...
Ah, good tip. I'll try that.
> > I just upgraded to 2.6.28.4 so we'll see what happens, but if you have
> > recommendations on kernel options, debugging, or other ideas, I'm all ears.
> > (like can I debug on usb serial, or do I need real serial?)
>
> There is a hideous hack to log data to the RTC (will scramble the time,
> though, so it is not good if you cannot reproduce the problem easily as you
> cannot really have that thing enabled all the time) that does work almost
> always.
Ah, yes, I read about that. I'll have to try that to reproduce the battery
hang.
Thanks for the tips,
Marc
--
"A mouse is a device used to point at the xterm you want to type in" - A.S.R.
Microsoft is to operating systems & security ....
.... what McDonalds is to gourmet cooking
Home page: http://marc.merlins.org/
------------------------------------------------------------------------------
Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
-OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise
-Strategies to boost innovation and cut costs with open source participation
-Receive a $600 discount off the registration fee with the source code: SFAD
http://p.sf.net/sfu/XcvMzF8H
_______________________________________________
ibm-acpi-devel mailing list
ibm-acpi-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.sourceforge.net/lists/listinfo/ibm-acpi-devel
