[ltp] T40 Powe Management Problem?

Alex Deucher linux-thinkpad@linux-thinkpad.org
Thu, 28 Jun 2007 09:26:11 -0500


On 6/27/07, sjk <sjk@cupacoffee.net> wrote:
> Henrique de Moraes Holschuh wrote:
> > Hi sjk!
> >
> > On Wed, 27 Jun 2007, sjk wrote:
> >
> >> Marius Gedminas wrote:
> >>
> >>> I assume this is true by looking at Ubuntu's /etc/acpi/sleep.sh, which
> >>> ignores ACPI events when it finds one of the other above-mentioned
> >>> daemons running.  Before it did that, suspending the laptop used to
> >>> result in nice sleep-resume-sleep again cycles when both acpid and
> >>> gnome-power-manager both handled the event.
> >>>
> >>>> This is really annoying as once the machine does into hibernate, I can't
> >>>> get it to come out without rebooting or forcing suspend to disk and then
> >>>> recover from that -- which hoses X.
> >>> Doesn't "hibernate" mean "suspend to disk"?  I think you mean "sleep".
> >>>
> >>> Marius Gedminas
> >> Perhaps it is sleep. . . the screen goes blank and I can't recover so I
> >> am not sure what it is. I run pretty light -- windowmaker and acpid. I
> >> have no acpi events or actions defined for the lid. If I kill acpid the
> >> machine still halts when the lid closes -- I did just notice that
> >> [kacpid] is still running -- leaving me to believe it is something in
> >> the kernel. . . .
> >
> > The kernel does not react to lid events.  But go into the bios and tell it
> > to not try to suspend the machine when the lid closes, that MIGHT make some
> > difference to the ACPI events it generates when nothing is processing
> > hotkeys.
> >
> > And run an lsof to track down everything that is listening to
> > /proc/acpi/event.
> >
>
> I disabled that function in the bios and it made no difference. I am
> beginning to wonder if it's not a problem with X recovering. I can
> suspend to disk when after the lid is closed and then opened and
> occasionally recover from the suspend. I discovered today that once the
> lid is reopened the keyboard is active -- but the display won't 'un-blank'

It could be that the bios has blanked the panel behind X's back, and
as such X doesn't realize that the hardware state has been altered.
Although if you disabled screen blanking in the bios... If regular X
dpms works fine, that I suspect the bios is doing something.

Alex