[Bug 75629] [r128] Changes needed for RandR support

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=75629

--- Comment #25 from Alex Deucher <agd5f at yahoo.com> ---
(In reply to comment #24)
> (In reply to comment #23)
> > > Detecting the valid modes with EDID would be nice, but I think this solves
> > > the most important problem. The driver now knows whether a monitor is
> > > plugged in and hence whether to drive Crtc2. If you agree, I will send this
> > > patch to the list.
> > 
> > Does it actually change when you have a monitor vs. not?  If probing works,
> > you are probably pretty close to getting full i2c working.
> 
> Yes. And you are right, i2c was almost working. I got it working by trying
> all 16 combinations for the bits in a loop. The valid pair was
> (R128_GPIO_MONID_*_3, R128_GPIO_MONID_*_1) instead of (R128_GPIO_MONID_*_3,
> R128_GPIO_MONID_*_0).

Was it the same register (R128_GPIO_MONID) as well?

> 
> Should I just change the *_0 bits to *_1 or should I put that inside a
> powerpc check? I think the old code might've been for Apple Sense.

I suspect the pin mapping is different for DVI vs. VGA.  Take the
RADEONI2CBusRec from radeon and set the masks and registers differently for VGA
vs. DVI.  Then you can use the appropriate registers/masks for each connector
when you want to probe the monitor.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.x.org/archives/xorg-driver-ati/attachments/20140528/a243368a/attachment.html>