[ltp] T41p : APM & ATI fglrx

[Arndt Twickel]

These are my experiences with the T41p and the binary ATI drivers
(Bios 3.05a, Suse 9.1, XFree86 4.3.99):

- Suspend to disk (APM) works (stable with 3.7.6, unreliable with
3.9.0)

- Supend to ram (APM + ACPI) as well as suspend to disk (ACPI) do not
work, that means the machine powers down o.k. but upon awakening the
machine immediately freezes. I played around with most of the
XF86config options (internal/external agpart etc.), installed
different kernels and fglrx-drivers but without luck.

- Blanking the screen freezes the machine (it used to work before but
i'm not sure if this was due to the Bios Version or to the fglrx
version).

- 3.7.6 gives me way better 3D performance compared to the newer 3.9.0
(2200 fps compared to 1800 fps with glxgears) and apart from the
suspend and blanking stuff it runs very stable.

Now I don't really understand why some people on this list are able to
do all the suspend stuff with their thinkpad and the fglrx driver
(e.g. Didrik Pinte with an R50 - see his post earlier in this thread
and in another thread in may).  Does it depend on the graphics card,
on the video driver, on the bios, on the x setup or on something else?
Someone on this list (can't find the post anymore) said that for him
it was crucial to do an xrefresh in the resume script. I tried that
but it didn't work for me. Actually I haven't heard of anyone getting
the supend to ram to work with a t41p and the ati driver.

If you depend on 3D acceleration it really sucks not being able to use
the suspend to ram function since suspending to disk (apm) almost
takes 1:30 minutes! I contacted IBM (support hotline) and ATI to
report the error and recommend that anyone experiencing the same
problem does so as well.  Maybe if enough people complain they will
provide a fix. ATI's feedback page can be found here:
http://apps.ati.com/linuxDfeedback/index.asp 
(Unfortunately ATI doesn't seem to be very interested in communicating
with the linux community and is not putting much effort in developing
the drivers - see the numerous discussions in this forum:
http://www.rage3d.com/board/forumdisplay.php?f=61)

Now about the power consumption: Volker Braun found that using the
power saving functions of the graphics chip (in his case using an
experimental patch for the radeon driver) significantly prolongs
battery life:
http://www.sas.upenn.edu/~vbraun/computing/T41/power.html

Well I guess this is all I can contribute and I hope someone will come
up with a solution to the t41p + fglrx suspend problem in the near
future!

Arndt