[PATCH] xf86: return NULL for compat output if no outputs.

Dave Airlie airlied at gmail.com
Mon Sep 17 01:32:58 PDT 2012


On Mon, Sep 17, 2012 at 4:47 PM, Chris Wilson <chris at chris-wilson.co.uk> wrote:
> On Mon, 17 Sep 2012 09:48:13 +1000, Dave Airlie <airlied at gmail.com> wrote:
>> On Mon, Sep 10, 2012 at 5:11 PM, Chris Wilson <chris at chris-wilson.co.uk> wrote:
>> > On Mon, 10 Sep 2012 11:30:01 +1000, Dave Airlie <airlied at gmail.com> wrote:
>> >> From: Dave Airlie <airlied at redhat.com>
>> >>
>> >> With outputless GPUs showing up we crash here if there are not outputs
>> >> try and recover with a bit of grace.
>> >
>> > With this fixed, are we then in a position where all modes can be added
>> > later upon hotplug detection?
>>
>> It doesn't really have anything to do with hotplug detection so I've no idea.
>>
>> This is purely for GPUs which have 0 outputs ever and will never have
>> any. I've no idea what would block GPUs that grow outputs.
>
> I'm referring to the code in every driver to abort PreInit if
> (!scrn->modes) which I presumed was a limitation of the core. Are we in
> a position to be discard those checks and the drivers start up with no
> outputs connected?

Oh for non-gpu outputs probably not not something I was concerned with
yet. For radeon we check modes and if we are a gpu object.

Of course for GPU switch and everyone becomes a GPU object I'll
probably start hitting this, but kms always gives us some modes I
think anyways.

Dave.


More information about the xorg-devel mailing list