[ltp] T40, ACPI, and sleeping

[Michael Keyl]

On Thursday 26 May 2005 06:15, Chris Frost wrote:
> On Wed, May 25, 2005 at 07:29:22PM -0700, Chris Frost wrote:
> > The lcd's backlight is off during sleep. I am not using the patch at
> > http://bugme.osdl.org/show_bug.cgi?id=3022 as I am not using the
> > kernel fb (I assume this only applies if one is).
>
> In case one does need to use the radeon fb and the patch to enable
> sleeping, I turned this on and applied the above patch. A five minute
> suspend to ram using the thinkwiki test script yielded:
>
> Wed May 25 20:58:40 PDT 2005
> before: 39570 mWh
> after: 39370 mWh
> diff: -200 mWh
> seconds: 337 sec
> result: -2136 mW
> Your model seems to be affected.

As far as I have understood, the power drain problem is caused (at least on 
some models) by the fact that the radeon card is not switched off before the 
machine enters sleep mode. This is something which has to be done by the 
kernel and not by the X-server, and the natural place is somewhere within the 
radeon_fb driver. The patch you have mentioned adds such code (or more 
previsely enables some power- mac code also on intel machines). Hence, if the 
radeon card is really the culprit on your machine you need the patch to solve 
the problem and the radeon_fb driver has to be active during s2r. The savest 
method to guarantee this is to compile the radeon_fb driver into the kernel  
(not as a module) and to add video=radeonfb:force_sleep as a kernel 
parameter. If this does not help, the power drain is probably caused by a 
different component of your system (but I doubt that the graphics card is 
completely innocent if its a radeon). 

Michael Keyl