[ltp] T41: Terminating Fglrx causes Radeon FB to be mangled

Michael Gaber linux-thinkpad@linux-thinkpad.org
Tue, 28 Mar 2006 20:18:01 +0200


Felix E. Klee wrote:
> At Mon, 27 Mar 2006 11:42:39 +0200,
> Michael Gaber wrote:
>>>> X runs fine, and I'm already used to several features of this X
>>>> server, so I don't want to change it.
>>> If there's no solution other than switching back to the "radeon" driver
>>> included in X.org 6.8.2, I'd like to know (chip is a Radeon 9000):
>> you could also switch you console-driver back to vesafb
> 
> Really? As far as I understand the Radeon power drain kernel patch, it's
> necessary to use the radeon driver for the console in order to get the
> power drain problem solved.
> 
> Also, if I had the choice, I'd not take a FB driver for the console at
> all: plain text is much faster.

oh didn't know that, try without fb on console if you want, 
kernel-building is no problem, so just try if it works the way you want 
it to

> 
>>>   Is "rovclock -c 110 -m 110" the same as "aticonfig
>>>   --set-powerstate=1"?
>> the OS driver supports "dynamicclocks" which chooses these speeds on demand
> 
> I recall reading about some issue with dynamicclocks ... here it is:
> 
> NOTE!  Enabling DynamicClocks crashes some models. If the CPU is
> entering one of the lower power states (C3 or lower) during Xorg startup
> the display may stay black. As a workaround disable DynamicClocks in
> Xorg and use Rovclock instead. But it does not scale the clocks to match
> the workload. Update 5/2/2006: Seems fixed in Xorg6.9

well, I'm currently using xorg-7.0 so i didn't notice that

> 
> (Quoted from <URL:http://www.thinkwiki.org/wiki/How_to_make_use_of_Graph
> ics_Chips_Power_Management_features>)
>