more mostly unreviewed fixes

[Dave Airlie]

On Mon, Sep 17, 2012 at 1:47 PM, Keith Packard <keithp at keithp.com> wrote:
> Dave Airlie <airlied at gmail.com> writes:
>
>>       dri2: invalidate drawable after sharing pixmap
>
> Merged.
>    f0bad69..22746df  master -> master
>
>>       xf86: return NULL for compat output if no outputs.
>
> I don't like this patch -- reading through the function, there's just
> one little bit at the bottom which isn't doing the right thing, and it
> will *still* do the wrong thing if you delete outputs from a screen.
>
> Here's an (untested) patch which fixes that test instead:
>
And it works, so I've stuck it in the new pull.

>>       xf86/platform: scan pci after probing devices
>
> Ick. This seems to assume that the device will remain powered up after
> get_drm_info closes it again? If that's true, then this seems like it
> should be a fine kludge to me.

Its true for now, I'm not really sure the best way to do this I expect
it needs libpciaccess enhancments so we can update the device info for
a single device, so we can in the probe, open the device, update pci
info, close it.

> Can you get rid of the gratuitous '&' in front of
> xf86PlatformDeviceProbe for me though?

Done.

Thanks,
Dave.