[Intel-gfx] [PATCH] i915: Modify for pineview clock source

Mark Hayter mdhayter at chromium.org
Thu Dec 16 06:11:07 CET 2010


[Again using correct address to get to list, sorry about that]

>
Chris,

>
I am happy to use the ssc frequency setting code and expand beyond pineview
(although I can't really test those configurations), but Im not sure what to
do about the high_speed flag. This would normally come from the BDB general
features ssc_freq flag. Since the code is to set defaults that will be used
in the absence of VBIOS/VBT would it be acceptable to just set high_speed to
1 when I call intel_bios_ssc_frequency to configure the speed, so that for
pineview it matches the original code? This would be correct for the
machines we know that do not have VBIOS and is otherwise as arbitrary as
using 0.

>
Thanks

> Mark

>
On Wed, Dec 15, 2010 at 6:52 AM, Chris Wilson <chris at chris-wilson.co.uk>wrote:

>
On Tue, 14 Dec 2010 11:40:30 -0800, bfreed at chromium.org wrote:
>
>> > The i915 driver normally assumes the video bios has configured several
>
>> > of the LVDS panel registers, and it just inherits the values. If the
>
>> > vbios has not run, several of these will need to be setup.
>
>>
> I'm having problems applying these using git-am. Notably your base objects
>
>> are not in my tree.
>
>>
> I'm dubious about flipping the SSC bit just for PineView. On recent
>
>> platforms, I think using an SSC reference is prevalent.
>
>>
> Instead of hardcoding 100Mhz as the clock, I'd prefer to see those SSC
>
>> frequencies split out of the general definitions and shared with this
>
>> initialiser. Something like:
>
>> int intel_bios_ssc_frequency(struct drm_device *dev, bool high_speed)
>
>> {
>
>>   switch(INTEL_INFO(dev)->gen) {
>
>>   case 2: return high_speed ? 66 : 48;
>
>>   case 3:
>
>>   case 4: return high_speed ? 100 : 96;
>
>>   default: return high_speed ? 120 : 100;
>
>>   }
>
>> }
>
>> -Chris
>
> --
> Chris Wilson, Intel Open Source Technology Centre
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/intel-gfx/attachments/20101215/0e7698a8/attachment.html>


More information about the Intel-gfx mailing list