Nothing newer as https://wiki.freebsd.org/Graphics/Intel-GPU-Matrix?

[Milan Obuch]

On Fri, 17 May 2019 11:59:16 +0200
Milan Obuch <freebsd-x11 at dino.sk> wrote:

> On Fri, 17 May 2019 11:49:36 +0200
> Niclas Zeising <zeising+freebsd at daemonic.se> wrote:
> 
> > On 2019-05-17 11:36, Milan Obuch wrote:  
> > > On Thu, 16 May 2019 11:35:13 +0200
> > > Niclas Zeising <zeising+freebsd at daemonic.se> wrote:
> > >     
> > >> On 2019-05-16 11:12, Niclas Zeising wrote:    
> > >>> On 2019-05-16 11:05, Milan Obuch wrote:    
> > >>>> Hi,
> > >>>>
> > >>>> this page is last edited end of year 2018. Is there anything
> > >>>> newer? Of particular interest for me are Bay Trail CPUs I am
> > >>>> trying to run X on, but no success yet. Well, vesa video works,
> > >>>> kind of, but I need Full HD resolution which seems not possible
> > >>>> with vesa.
> > >>>>
> > >>>> If anybody has something to share, maybe some work-in-progress
> > >>>> for test, it would bee great...
> > >>>>       
> > >>>
> > >>> Hi!
> > >>> Graphics driver versions hans't been updated since the end of
> > >>> 2018, so that is the newest version.  There is ongoing work to
> > >>> update the driver to current, but that is not in base yet.
> > >>> Regards    
> > >>
> > >> Hi!
> > >> The bay trail SoCs could be supported from a graphics perspective
> > >> even if they aren't listed.  I suggest you install drm-kmod and
> > >> try. You'll see if it attaches to the hardware.
> > >>    
> > > 
> > > Hi,
> > > 
> > > so I did full reinstall with 12.0-STABLE snapshot, used port
> > > drm-fbsd12.0-kmod-4.16.g20190430, kldload /boot/modules/i915kms.ko
> > > and it works - kind of. Just one trouble, I can't see anything on
> > > screen connected to VGA output, just mouse X cursor, because other
> > > output is active before VGA and all three xterms are there, left
> > > of my visible screen (I can move mouse cursor to the left, off the
> > > screen).
> > > 
> > > grep EDID Xorg.0.log
> > > 
> > > [   715.398] (II) intel(0): EDID for output eDP1
> > > [   715.398] (II) intel(0): EDID Version: 1.4
> > > [   715.399] (II) intel(0): EDID (in hex):
> > > [   715.427] (II) intel(0): EDID for output VGA1
> > > [   715.428] (II) intel(0): EDID Version: 1.3
> > > [   715.429] (II) intel(0): EDID (in hex):
> > > [   715.429] (II) intel(0): EDID for output DP1
> > > [   715.430] (II) intel(0): EDID for output HDMI1
> > > 
> > > So I think I need somehow disable eDP1 output and it should be OK,
> > > but how? When i915kms module is kldloaded, there are some tunables
> > > mentioned, like kern.vt.fb.modes.VGA-1 - could it be used for
> > > it?    
> > 
> > Hi!
> > If you have started X (which I assume you have from your
> > description), try fiddling with xrandr.  Just running xrandr without
> > any arguments should show all screens connected to your computer.
> > You can then disable certain screens with
> > xrandr --output [name-of-output] --off
> > The name of the output is listed in the xrandr output from the first
> > run. Is this a dual or multi screen set up?
> >   
> 
> I did, and yes, I tried already xrandr. It gives me just what I need.
> My question was if it is possible to do it before starting X, but
> calling xrandr from xinitrc is not a big deal for me. I just checked,
> works satisfactorily.
> 
> There is one strange thing, though not X related - when i915kms module
> is kldloaded, screen switched to higher resolution, but strangely, not
> full screen could be used, it looks like it scrolls up in the middle
> of physical screen. Any idea on this?
> 
> Regards,
> Milan
>

Addendum:

in dmesg I see following after kldload i915kms:

VT: Replacing driver "efifb" with new "fb".
start FB_INFO:
type=11 height=768 width=1366 depth=32
cmsize=16 size=8294400
pbase=0xc01a0000 vbase=0xfffff800c01a0000
name=drmn0 flags=0x0 stride=7680 bpp=32
cmap[0]=0 cmap[1]=7f0000 cmap[2]=7f00 cmap[3]=c4a000
end FB_INFO

It looks like again here, some data are read from eDP1 port, at least
resolution, at this is used for VGA1 port as well, albeit in this case
it is wrong. So I need some kind of override here, too.

Milan