Re: Bug reports on the psb-gfx driver

Re: Bug reports on the psb-gfx driver

[Kristoffer Eriksson]

Håvar Nøvik skrev 2012-04-17 12:53:
> Hi,
>
Greetings,
> I'm experiencing a bug with the driver after a kernel upgrade (3.2 -> 
> 3.3). Every once in a while the screen is blacking out for a second or 
> two.
>
> Anyways, I was wondering where I could look at bug reports or report a 
> new one if I can't find a similar?
>
Ive done a fresh build and Im seeing that same issue that you are 
describing. In a frequency of about 5-15mins it blackens and then returns
within 1-2 seconds. Could you give us a dmesg output? I will do some 
bugtracking once I get back to my laptop.
Alan, any thoughts?

PS: Thanks for the great work on a GMA 500 driver!
>
Alan is the one to thank. :)
> ----
> Regards Håvar Nøvik

Re: Bug reports on the psb-gfx driver

[Alan Cox]

> That frequency seems about the same as the one I'm experiences. All though
> some times it's more often, but I can't say I've done anything special at
> the same time.
> 
> 3.3.2-1-ARCH
> Intel(R) Atom(TM) CPU Z520 @ 1.33GHz
> 
> Attachment: dmesg ouput
> 

What X display server are you using (fbdev or other ?). Also I see a few
writes to the brightness in the log - did you fiddle with the brightness
by hand at all ?


Either way I suspect the only way to find this is to verify a
configuration it occurs on, go back a kernel, verify it doesn't happen
there and then build a kernel to the point in the git tree before the
gma500 patches were changed (to check its not caused by something else
such as an ACPI change), then try and find which one caused it.

That will be tedious but unfortunately I've got no documentation or other
good ways to chase this down.

Alan

Re: Bug reports on the psb-gfx driver

[Kristoffer Eriksson]

Greetings,

Please try 3.4 and see if the issue remains. Currently running 3.4-RC3 
and cant see it happening at all.

Best wishes
Kristoffer

Håvar Nøvik skrev 2012-04-17 12:53:
> Hi,
>
> I'm experiencing a bug with the driver after a kernel upgrade (3.2 -> 
> 3.3). Every once in a while the screen is blacking out for a second or 
> two.
>
> Anyways, I was wondering where I could look at bug reports or report a 
> new one if I can't find a similar?
>
> PS: Thanks for the great work on a GMA 500 driver!
>
> ----
> Regards Håvar Nøvik

Re: Bug reports on the psb-gfx driver

[Alan Cox]

On Sun, 22 Apr 2012 11:46:27 +0200
Kristoffer Eriksson <kristoffer.ericson@gmail.com> wrote:

> Greetings,
> 
> Please try 3.4 and see if the issue remains. Currently running
> 3.4-RC3 and cant see it happening at all.

Interesting. Guillaume Clements also root caused it on his system and
sent a patch to dri-devel. That is also probably worth trying.

Alan

Re: Bug reports on the psb-gfx driver

[Kristoffer Eriksson]

Alan Cox skrev 2012-04-18 23:21:
>> That frequency seems about the same as the one I'm experiences. All though
>> some times it's more often, but I can't say I've done anything special at
>> the same time.
>>
>> 3.3.2-1-ARCH
>> Intel(R) Atom(TM) CPU Z520 @ 1.33GHz
>>
>> Attachment: dmesg ouput
>>
> What X display server are you using (fbdev or other ?). Also I see a few
> writes to the brightness in the log - did you fiddle with the brightness
> by hand at all ?
Alan,

Without looking at the code I can say that those brightness commands are 
not related to this
issue since Ive been having them since the first working kernel version.
Havent had time to look at the code but Ive understod them to happen 
simply when you
press a key to return from backlight suspend. Something like the default 
brightness setting that gets invoked
after return from suspend.

Håvar see
[ 7659.177126] gma500 0000:00:02.0: Backlight lvds set brightness 186a186a

while I see
[....] gma500 0000:00:02.0: Backlight lvds set brightness 7a127a12



>
> Either way I suspect the only way to find this is to verify a
> configuration it occurs on, go back a kernel, verify it doesn't happen
> there and then build a kernel to the point in the git tree before the
> gma500 patches were changed (to check its not caused by something else
> such as an ACPI change), then try and find which one caused it.
Will do further fresh builds to see if I can reproduce the blanking 
issue and also
test out the patch.


> That will be tedious but unfortunately I've got no documentation or other
> good ways to chase this down.
>
> Alan
>