Re: radeonfb vs. Mac Mini

[Michael <macallan%netbsd.org@localhost>]

On Wed, 6 Sep 2017 14:44:18 +0300
Valery Ushakov <uwe%stderr.spb.ru@localhost> wrote:

> On Tue, Sep 05, 2017 at 23:04:21 -0400, Michael wrote:
> 
> > On Wed, 6 Sep 2017 04:31:13 +0300
> > Valery Ushakov <uwe%stderr.spb.ru@localhost> wrote:
> >   
> > > On Thu, Aug 31, 2017 at 14:10:47 -0400, Michael wrote:
> > >   
> > > > I committed a few changes to radeonfb lately which make it work
> > > > properly on my Mini, with both DVI and VGA output.
> > > > My Mini is the 'silent upgrade' model, PowerMac10,2 as per Open
> > > > Firmware. The DVI port is wired slightly different on first
> > > > generation Minis ( PowerMac10,1 ), I *think* I set it up
> > > > correctly but I have no way to test this.    
> > > 
> > > I have tested it on the early mini with DVI and, unfortunately, it
> > > doesn't seem to work.  The screen is black.
> > > 
> > > radeonfb0 at pci0 dev 16 function 0found EDID property...
> > > : ATI Technologies Radeon 9200 5962 (rev. 0x01)
> > > radeonfb0: Video BIOS not present
> > > radeonfb0: No video BIOS, using default clocks
> > > radeonfb0: refclk = 27.000 MHz, refdiv = 12 minpll = 125000,
> > > maxpll = 400000 radeonfb0: 64 MB aperture at 0x98000000, 64 KB
> > > registers at 0x90000000 radeonfb0: display 0: initial virtual
> > > resolution 1920x1200 at 8 bpp radeonfb0: using 32 MB per display
> > > radeonfb0: port 0: physical 1920x1200 60Hz
> > > radeonfb0: port 1: physical 1920x1200 60Hz
> > > wsdisplay0 at radeonfb0 kbdmux 1: console (fb, vt100 emulation)  
> > 
> > Looks ok so far...
> >   
> > > display-type            4c434400 ........ ........ ........
> > > LCD.  
> > 
> > ... ok, so it's a DVI monitor. Any way you could test with a VGA
> > adaptor?  
> 
> Right, that's with DVI.  I've just tested current GENERIC with VGA +
> the original adapter and it works (but VGA always worked, iirc, didn't
> it?).

Kinda - DDC didn't since all Minis have the monitor port's DDC pins
wired to something odd, not the pins normally associated with VGA or
DVI outputs. The DAC/PLL/whatever wiring for VGA output looks mostly
normal.

> > Also, does X work (with an updated radeon driver)?  Doesn't matter
> > if it's on top of genfb or radeonfb, it should Just Work without any
> > config file hackery.  I tried to replicate their Mini workarounds in
> > radeonfb, I'd like to know if the original works before looking into
> > my copy.  
> 
> This mini currently runs -7 for some tests and I'm a bit reluctant to
> upgrade it.  Any way I can test new X without complete upgrade?  If
> xsrc on 7 is not that diverged, a patch will do (you gonna pull it up
> anyway :).

My last commit to xf86-video-radeon will probably just apply or be
trivial to port to -7. All you need is the driver.

> > > PS: Please, can this "found EDID property" in copy_disp_props() be
> > > fixed?  
> > 
> > Will do, that should be Very Verbose Debug Output(tm)...  
> 
> Thanks!

Done. downgraded some aprint_verbose() to aprint_debug() where they
belong.

have fun
Michael