Re: genfb not working on E250 with Raptor GFX card

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

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

On Jan 13, 2008, at 15:54, Peter Bex wrote:

> On Sun, Jan 13, 2008 at 02:50:56PM -0500, Michael Lorenz wrote:
> > Sounds like genfb and the video hardware disagree about colour depth
> > for some reason.
> >
> > > I have no idea how to debug the genfb issue.  I tried reverting the
> > > changes I found in the CVS changelog (especially after reading last
> > > month's announcement by macallan), but none of these fixed the
> > > problem.
> >
> > These changes were workarounds for bugs in my Raptor's firmware, I
> > wouldn't be surprised if yours needs more than that. To see what's
> > going on I need:
> > - - ofctl -p output
>
> Unfortunately I can't do that, because I can't get a working NetBSD
> on the system.

No way to hook up a serial console?

> If you can tell me how to get the useful information directly from  
> OFW, I'll be happy to type them in manually, though.

No need to, your ten finger copy of genfb's output is all I need.

> > - - dmesg output
>
> The info from dmesg you need I can tell you as well (see also below).
> Do you really need all of dmesg?  It would require me to copy it  
> manually
> too..

Only genfb's output...

> I enabled GENFB_DEBUG and got the following output just before the
> top part of the screen went black:
>
> genfb0 at pci0 dev 2 function 0membase: 1ff80000000 fbpa: 1ff80800000
> : Texas Instruments TVP4020 Permedia 2
> <?xml version="1.0" encoding="UTF-8"?>
> <!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN"
> "http://www.apple.com/DTDs/PropertyList-1.0.dtd";>
> <plist version="1.0">
> <dict>
>         <key>OFpnode</key>
>         <integer>-267800264</integer>
>         <key>address</key>
>         <integer>0x800000</integer>
>         <key>cmap_callback</key>
>         <integer>0x18a9850</integer>
>         <key>depth</key>
>         <integer>0x8</integer>
>         <key>height</key>
>         <integer>0x400</integer>
>         <key>instance_handle</key>
>         <integer>0xfff69ce8</integer>
>         <key>is_console</key>
>         <true/>
>         <key>linebytes</key>
>         <integer>0x500</integer>
>         <key>width</key>
>         <integer>0x500</integer>
> </dict>
> </plist>

And here is the problem - the card claims to run in 8bit colour and  
genfb believes it. So, this isn't genfb's fault, it's the firmware  
that's lying.

> Nope, I'm not sure.  Can you get anything useful out of the genfb line
> above?

Yes, it's a Permedia2, that's all I need.

You might be able to work around this - just have a look at the words  
defined on the Raptor's node. Several of them should be video modes,  
like r1152x900x76x8. Change your output-device variable to something  
like screen:r1152x900x76x32 ( pick any suitable mode, the x32 at the  
end is important ), reset-all and have a look at the depth property.  
If it changed to 0x20 genfb should do The Right Thing.

have fun
Michael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iQEVAwUBR4p+ScpnzkX8Yg2nAQI/Jgf+KuDFG3Eqw4Bz+V5mAOxtZg4lEwGfxVOl
OxnXoyOGPvPLw5VCS987KWC3AWT2P3jLvIsl49pWe29HISoFjTPoHvnTJPEYBGUf
5QiGrujKn8bFGh8Fq1e/3lT4oOqlrGuw07u9C6IbaThIWocT2BtYad1n32g0vb0Y
QknNnoVsVHfb7WkpETbx9TTRTeO7VjkMMIEnrK6I7DwPwVEXzwgnpJ/Axk09muVs
Ixsmq6I1F4fBO3FSTQk3d0NP2/0oYQdDxSBe6/ORObqx6tQd3iCRAim2iYa6KmeR
MtOcy7HE3/bZzZE/AjsN7b/vF2jxJkDrWOTcuLHYDNSt8eMQsLDSgg==
=pPwC
-----END PGP SIGNATURE-----