Re: Display goes off when framebuffer enabled and Kernel Panic with fbset

["Emil Velikov" <emil.velikov-/E1597aS9LT10XsdtD+oqA@public.gmane.org>]

Roger, looking at your logs it appears that you are still running an old  
version of nouveau

Make sure that you are running the latest git as it includes a few AGP  
related bug-fixes
Note that nouveau is NOT only a kernel. For more info on how to download,  
compile and install take a look at the wiki [1] [2]

[1] http://nouveau.freedesktop.org/wiki/InstallNouveau
[2] http://nouveau.freedesktop.org/wiki/GentooEbuilds

On Fri, 10 Dec 2010 10:43:52 -0000, Roger <rogerx.oss-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:

> On Thu, Dec 09, 2010 at 04:38:07PM +0100, Francisco Jerez wrote:
>> Roger <rogerx.oss-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> writes:
>>
>>> I think I've found the problem.
>>>
>>> # echo "option nouveau noagp=1" >> /etc/modprobe.d/nouveau.conf
>>>
>>> So far, it looks like things have stopped freezing.
>>
>> Do you get the same kind of AGP problems with the other nvidia cards you
>> said you had access to? Have you tried to set a saner AGP aperture size
>> (e.g. 64MB) in your BIOS settings? Does it make any difference?
> No real difference between AGP memory reservations.  Aside from the  
> anomoly I
> saw on the first freeze with 64MB AGP, saw my last framebuffer screen on  
> a good
> halt of the system -- done sometime last night?
>
> Again, I got one working with noagp=0 and then one frozen.  Attaching  
> both
> complete logs using debug settings for the drm/nouveau modules.
>
>
>> The output from "lspci -vvv" (executed as root before having loaded the
>> nouveau kernel module) and *full* kernel logs are likely to be useful
>> too.
> Attaching lscpi -vvv as well.
>
> (Hopefully the mailing list doesn't strip them.)