Framebuffer test kernel

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

Hello,

I've uploaded a very preliminary copy of a GENERIC mac68k
kernel[*]  
with framebuffer+wscons support to:

	ftp://ftp.NetBSD.org/pub/NetBSD/misc/jmmv/netbsd-mac68k-fb.
gz

This will hopefully work on all machines with integrated
video; i.e.,  
those that previously used 'intvid at obio?'.  It *might*
also work  
with those using NuBus cards, but that is completely
untested.  That  
does not mean that you should not test it in them!

Note that this indeed supports a color console!  Make sure
to change  
your monitor configuration in the Booter to 256 colors
(recommended)  
or "millions of colors".  The "thousands of
colors" configuration  
will *not* work yet.  Of course, 1-bit will also work.

Machines with a valkyrie card will hopefully have correct
colors on  
the console.  Others may not: e.g. you may get a pink on
white  
console as happened here.  I'd like to know which these are
so, if  
that's your case, please attach the output of dmesg. 
Obviously, if  
it does not boot at all, it'd be interesting to know which
is your  
machine, your video card and how it hang.

I have not committed the code yet because it's still not
polished.

Enjoy 

*: This kernel is built from NetBSD-current, using a
softfloat  
toolchain (but that should not change the kernel, AFAIK). 
It is a  
modified version of the current GENERIC *without* macfb,
which is  
replaced with genfb and valkyriefb.  I've also disabled zs
because  
there is still a problem left in the code WRT console
output.

- --
Julio M. Merino Vidal <jmmv84gmail.com>


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

iD8DBQFG1xxwuIm9UEGtViURAoUPAJ9B3ilpENiMJgG2GoWcQgE7C7xpRwCg
wwiK
vBwISQHoC3c4KU014w/PCys=
=6KvR
-----END PGP SIGNATURE-----

Hi,

On 8/30/07, Julio M. Merino Vidal <jmmv84gmail.com> wrote:
> This will hopefully work on all machines with
integrated video; i.e.,
> those that previously used 'intvid at obio?'.  It
*might* also work
> with those using NuBus cards, but that is completely
untested.  That
> does not mean that you should not test it in them!
>
> Note that this indeed supports a color console!  Make
sure to change
> your monitor configuration in the Booter to 256 colors
(recommended)
> or "millions of colors".  The "thousands
of colors" configuration
> will *not* work yet.  Of course, 1-bit will also work.

I tested the kernel on a Macintosh Quadra 700, internal
video (no
NuBus video card), with both 256 colors & monochrome.

When booting in 256 color mode, the screen gets odd colors
(light
yellow on white) and then the kernel panics after attaching
wsdisplay0:

wsdisplay0 at genfb0 kdbmux 1 [...]
trap type 0, code = 0xa5, v=0x4947818
kernel program counter = 0x1f17d8
kernel: Bus error trap
[...]

(I couldn't get the first part of the dmesg in this setup.)

When booting in 1-bit mode, the screen gets white-on-white
colors, and
the effect is the same.
These are the kernel messages when booting with intvid0 and
macfb0:

intvid0 at obio0  f9001000: DAFB video subsystem, monitor
sense 1
intvid0: 640 x 480, monochrome
macfb0 at intvid0


     -Marco

On Aug 31, 2007, at 1:25 PM, Marco Trillo wrote:

> Hi,
>
> On 8/30/07, Julio M. Merino Vidal <jmmv84gmail.com> wrote:
>> This will hopefully work on all machines with
integrated video; i.e.,
>> those that previously used 'intvid at obio?'.  It
*might* also work
>> with those using NuBus cards, but that is
completely untested.  That
>> does not mean that you should not test it in them!
>>
>> Note that this indeed supports a color console! 
Make sure to change
>> your monitor configuration in the Booter to 256
colors (recommended)
>> or "millions of colors".  The
"thousands of colors" configuration
>> will *not* work yet.  Of course, 1-bit will also
work.
>
> I tested the kernel on a Macintosh Quadra 700, internal
video (no
> NuBus video card), with both 256 colors &
monochrome.
>
> When booting in 256 color mode, the screen gets odd
colors (light
> yellow on white)

I somewhat expected that, and I'm afraid this problem will
appear  
with all video cards.  I see that Linux has code to handle
the  
palette in the DAFB card, so I can look at it to fix this
issue (but  
haven't done that yet).

> and then the kernel panics after attaching
> wsdisplay0:
>
> wsdisplay0 at genfb0 kdbmux 1 [...]
> trap type 0, code = 0xa5, v=0x4947818
> kernel program counter = 0x1f17d8
> kernel: Bus error trap
> [...]

Bleh, that was due to a bug in the match routine of
valkyriefb: it  
assumed it could always attach!  Should be fixed now.  Could
you  
redownload the kernel and try again please?

Thank you very much!

-- 
Julio M. Merino Vidal <jmmv84gmail.com>

Hi,

On 8/31/07, Julio M. Merino Vidal <jmmv84gmail.com> wrote:
> On Aug 31, 2007, at 1:25 PM, Marco Trillo wrote:
> > I tested the kernel on a Macintosh Quadra 700,
internal video (no
> > NuBus video card), with both 256 colors &
monochrome.
> >
> > When booting in 256 color mode, the screen gets
odd colors (light
> > yellow on white)
>
> I somewhat expected that, and I'm afraid this problem
will appear
> with all video cards.  I see that Linux has code to
handle the
> palette in the DAFB card, so I can look at it to fix
this issue (but
> haven't done that yet).
>
> > and then the kernel panics after attaching
> > wsdisplay0:
> >
> > wsdisplay0 at genfb0 kdbmux 1 [...]
> > trap type 0, code = 0xa5, v=0x4947818
> > kernel program counter = 0x1f17d8
> > kernel: Bus error trap
> > [...]
>
> Bleh, that was due to a bug in the match routine of
valkyriefb: it
> assumed it could always attach!  Should be fixed now. 
Could you
> redownload the kernel and try again please?

I've downloaded the new kernel and booted it, but the Booter
gets
stuck after displaying the "So I sez to him"
message and quitting the
Finder -- the machine appears to hang. Mac OS extensions
were off, and
both 256 and 1-bit color modes were tried. I also tried to
enable the
"extra kernel debug messages" option but the
kernel doesn't output
anything nor the framebuffer appears to be initialized.

This may be related to something odd I noticed with the old
kernel:
when booting it, no messages could be read from the screen
before the
kernel panics. The first line that could be read in the
framebuffer
was the "wsdisplay0 at ..." message I posted
before. It appears to
'jump' driectly from the Booter to this screen, the earlier
messages
are not seen in the framebuffer while they are written.

 - Marco.

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

Hi,

So here we go.  I have just uploaded another kernel at the
same  
address as before:

>
	ftp://ftp.NetBSD.org/pub/NetBSD/misc/jmmv/netbsd-mac68k-fb.
gz

This one should let you run the Xmac68k server as before. 
It also  
has some changes in the genfbNuBus attachment routines,
although I'm  
still afraid it will not work in these cases.  1bpp still
does not  
work as expected.

I feel this is close to being finished (and it's taking
longer than I  
wanted).  I hope to be able to publish some preliminary
patches for  
review during this week.

Any testing will be highly appreciated 

Thanks,


On Aug 30, 2007, at 9:37 PM, Julio M. Merino Vidal wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hello,
>
> I've uploaded a very preliminary copy of a GENERIC
mac68k kernel[*]  
> with framebuffer+wscons support to:
>
>
	ftp://ftp.NetBSD.org/pub/NetBSD/misc/jmmv/netbsd-mac68k-fb.
gz
>
> This will hopefully work on all machines with
integrated video;  
> i.e., those that previously used 'intvid at obio?'.  It
*might*  
> also work with those using NuBus cards, but that is
completely  
> untested.  That does not mean that you should not test
it in them!
>
> Note that this indeed supports a color console!  Make
sure to  
> change your monitor configuration in the Booter to 256
colors  
> (recommended) or "millions of colors".  The
"thousands of colors"  
> configuration will *not* work yet.  Of course, 1-bit
will also work.
>
> Machines with a valkyrie card will hopefully have
correct colors on  
> the console.  Others may not: e.g. you may get a pink
on white  
> console as happened here.  I'd like to know which these
are so, if  
> that's your case, please attach the output of dmesg. 
Obviously, if  
> it does not boot at all, it'd be interesting to know
which is your  
> machine, your video card and how it hang.
>
> I have not committed the code yet because it's still
not polished.
>
> Enjoy 
>
> *: This kernel is built from NetBSD-current, using a
softfloat  
> toolchain (but that should not change the kernel,
AFAIK).  It is a  
> modified version of the current GENERIC *without*
macfb, which is  
> replaced with genfb and valkyriefb.  I've also disabled
zs because  
> there is still a problem left in the code WRT console
output.
>
> - --
> Julio M. Merino Vidal <jmmv84gmail.com>
>
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.7 (Darwin)
>
>
iD8DBQFG1xxwuIm9UEGtViURAoUPAJ9B3ilpENiMJgG2GoWcQgE7C7xpRwCg
wwiK
> vBwISQHoC3c4KU014w/PCys=
> =6KvR
> -----END PGP SIGNATURE-----

- --
Julio M. Merino Vidal <jmmv84gmail.com>


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

iD8DBQFG2vp8uIm9UEGtViURAq6jAKC3THMr3Naz681GBUqfXsUpTaGX+wCe
PsxR
zWVoiaG0yUYuYOFX4z2hFM8=
=rPUY
-----END PGP SIGNATURE-----

Anything new on the fb development?

Just curious,
Tim
-- 
Tim & Alethea
christtrek.org

On 26/11/2007, at 23:50, Tim & Alethea Larson wrote:

> Anything new on the fb development?

No, sorry.  I got busy with other stuff and could not work
more on  
that.  I still have the files around though, and I hope to
be able to  
get back to them, at most, during the christmas season.

Cheers,

-- 
Julio M. Merino Vidal <jmerinoac.upc.edu>