[9fans] Shared memory graphic on cpu server

[9fans] Shared memory graphic on cpu server

[Matthias Teege]

Moin,

my new cpu/auth server is up now. Very nice. The only problem is the
graphic on the console which didnt work. The computer I use for the
cpu server has a shared memory graphic and I don't know, if plan9
support somthing like that.

I set monitor=vga and vgasize=640x480x8 but the picture is not very
usefull. ;-) Anyhow I can work with the cpu server. Is plan9
supporting shared memory graphic?

Many thanks
Matthias

--
Matthias Teege -- http://www.mteege.de
make world not war
PGP-Key auf Anfrage

Re: [9fans] Shared memory graphic on cpu server

[Russ Cox]

What kind of video card is it?

Re: [9fans] Shared memory graphic on cpu server

[Matthias Teege]

"Russ Cox" <rsc@mit.edu> writes:

> What kind of video card is it?

S3 Graphics ProSavage8

http://us.shuttle.com/specs_access.asp?pro_id=285
http://www.via.com.tw/en/ProSavage%20Chipsets/km266.jsp

Matthias
--
Matthias Teege -- http://www.mteege.de
make world not war
PGP-Key auf Anfrage

Re: [9fans] Shared memory graphic on cpu server

[Russ Cox]

It's likely that the driver doesn't require significant
work to get it running on the ProSavage8 (the Savage
series seem to all be very similar), but it would take
a little.  If you're game, you might look at the X driver
to see how the ProSavage8 is handled differently from the
cards that our Savage driver supports, and then apply
those differences.

Russ

Re: [9fans] Shared memory graphic on cpu server

[jmk]

On Fri Sep 26 09:06:46 EDT 2003, matthias@mteege.de wrote:
> "Russ Cox" <rsc@mit.edu> writes:
>
> > What kind of video card is it?
>
> S3 Graphics ProSavage8
>
> http://us.shuttle.com/specs_access.asp?pro_id=285
> http://www.via.com.tw/en/ProSavage%20Chipsets/km266.jsp
>
> Matthias
> --
> Matthias Teege -- http://www.mteege.de
> make world not war
> PGP-Key auf Anfrage

Kenji Okamoto did a driver for the ProSavage8 in August and it
is included in /lib/vgadb, /sys/src/cmd/aux/vga on sources.
Perhaps you have a variant not understood by that code, what is
the PCI VID/DID of the chip?

Re: [9fans] Shared memory graphic on cpu server

[Matthias Teege]

"Russ Cox" <rsc@mit.edu> writes:

> It's likely that the driver doesn't require significant
> work to get it running on the ProSavage8 (the Savage
> series seem to all be very similar), but it would take
> a little.  If you're game, you might look at the X driver
> to see how the ProSavage8 is handled differently from the
> cards that our Savage driver supports, and then apply
> those differences.

I'm not a great programmer but it maybe interesting to read the
code. How can I make sure, that plan9 uses the correct driver?

Many thanks
Matthias

--
Matthias Teege -- http://www.mteege.de
make world not war
PGP-Key auf Anfrage

Re: [9fans] Shared memory graphic on cpu server

[Matthias Teege]

jmk@plan9.bell-labs.com writes:

> Perhaps you have a variant not understood by that code, what is
> the PCI VID/DID of the chip?

5333/8D04 !?

Matthias

--
Matthias Teege -- http://www.mteege.de
make world not war
PGP-Key auf Anfrage

Re: [9fans] Shared memory graphic on cpu server

[Russ Cox]

> I'm not a great programmer but it maybe interesting to read the
> code. How can I make sure, that plan9 uses the correct driver?

Actually it may be a bit easier than that, as jmk said.

Run
	pci | grep '	03'
(That's a tab not a space before the 03.)
It will print a line like
	0.3.0:	03.00.00 5333/8811   9 0:f8000000 67108864
the 5333/8811 is the vid/did (vendor and device id)
of the PCI device.  That's the line for the video card
because field 2 is the type, and 03.xx.xx is video cards.
The vid will be 5333 (S3), but the did will be different.
Note it.

Then look in /lib/vgadb for the line
	vid=0x5333 did=0x8D04		# ProSavage8 in P4M266 chipset
and add a line like it except with your did.

Finally, edit /sys/src/cmd/aux/vga/virge.c and everywhere
there is a "case 0x8D04:" add a "case your-did:" as well.
Then
	cd /sys/src/cmd/aux/vga; mk install
and give it a try.

Russ

Re: [9fans] Shared memory graphic on cpu server

[jmk]

On Fri Sep 26 10:36:42 EDT 2003, matthias@mteege.de wrote:
> jmk@plan9.bell-labs.com writes:
>
> > Perhaps you have a variant not understood by that code, what is
> > the PCI VID/DID of the chip?
>
> 5333/8D04 !?
>
> Matthias
>
> --
> Matthias Teege -- http://www.mteege.de
> make world not war
> PGP-Key auf Anfrage

OK. There should be code for that, done by Kenji.
Check you have up to date /lib/vgadb and aux/vga.

what is the output of 'aux/vga -p >somefile; cat somefile'?

Re: [9fans] Shared memory graphic on cpu server

[Matthias Teege]

jmk@plan9.bell-labs.com writes:

> OK. There should be code for that, done by Kenji.
> Check you have up to date /lib/vgadb and aux/vga.

Yes, I've seen the entry in /lib/vgadb after send my last email.

> what is the output of 'aux/vga -p >somefile; cat somefile'?

vga->attr: vid=0x5333
vga->attr: did=0x8D04
vga misc             E3
vga feature          00
vga sequencer        03 01 0F 00 0A
vga crt              5F 4F 52 9F 53 1F 0B 3E - 00 40 00 00 00 00 FF 00
                     EB 2D DF 50 60 EB EC A3 - FF
vga graphics         00 00 00 00 00 50 05 0F - FF
vga attribute        00 01 02 03 04 05 06 07 - 08 09 0A 0B 0C 0D 0E 0F
                     41 FF 0F 00 00
vga virtual         0 0
vga panning         off
vga clock[1] f      135000000
vga clock[1] d i m          0        0       -      127
vga clock[1] n p q r       31        0       -        0        3
vga apz                     0
vga linear                  1
virge Crt30          E1 8D 40 28 10 04 BF DB - 5B BD 14 59 00 00
virge Crt40          01 03 00 00 00 01 00 00 - 00 00 00 FF 01 2C 02 01
virge Crt50          40 00 00 08 E8 00 00 00 - 99 00 00 00 00 00 40 00
virge Crt60          00 00 00 00 00 04 88 00 - FF 00 00 01 00 FD 00 EE
virge hde           280   640
virge shb           298   664
virge ehb           29F   671
virge ht            320   800
virge vde           1E0   480
virge vrs           1EB   491
virge vre           1ED   493
virge vt            20D   525
virge Crt70          10 C0 07 1F 1F 1F 00 1F - 01 01 1F 1F 17 17 17 17
                     00 97 12 41 F6 10 00 00 - 00 00 00 00 00 00 00 00
                     80 50 00 00 00 00 00 00 - 00
virge Seq08          06 00 0A 00 0C 00 0E 0F
virge Seq10          00 00 00 00 00 02 2B 00 - 40 02 02 02 00 00 0A 00
                     20 21 22 23 24 25 26 04 - 0C 00 2A 02 03 00 2E 2F
                     10 06 41 63 00 00 01 52 - 5D 00 3A 3B 3C 00 3E 00
                     40 41 42 43 44 45 46 47 - 48 49 4A 4B 4C 4D 4E 00
virge Crt2D          8D 04 00
virge dclk m n r     14318180        0       -        0        0
virge mclk m n r     14318180        0       -        0        0


Matthias

--
Matthias Teege -- http://www.mteege.de
make world not war
PGP-Key auf Anfrage

Re: [9fans] Shared memory graphic on cpu server

[Russ Cox]

You don't have an up-to-date aux/vga.
cd /sys/src/cmd/aux/vga; mk install.

The binary is out-of-date.  I have been
working on rebuilding the out-of-date
binaries all morning.

Russ

Re: [9fans] Shared memory graphic on cpu server

[Matthias Teege]

"Russ Cox" <rsc@mit.edu> writes:

> You don't have an up-to-date aux/vga.
> cd /sys/src/cmd/aux/vga; mk install.

Thanks, works now. Only the cursor is a big fat quadrat, but for
the console it maybe ok.

Many thanks
Matthias

--
Matthias Teege -- http://www.mteege.de
make world not war
PGP-Key auf Anfrage

Re: [9fans] Shared memory graphic on cpu server

[Russ Cox]

It looks like there's no kernel support for the ProSavage8.
You'll need that for the cursor.  You could probably add it
to /sys/src/9/pc/vgasavage.c easily enough -- just do what
was done for PROSAVAGEK and see if it works.  (If so, send
us a new source file!)

Russ

Re: [9fans] Shared memory graphic on cpu server

[jmk]

On Fri Sep 26 12:07:54 EDT 2003, rsc@mit.edu wrote:
> It looks like there's no kernel support for the ProSavage8.
> You'll need that for the cursor.  You could probably add it
> to /sys/src/9/pc/vgasavage.c easily enough -- just do what
> was done for PROSAVAGEK and see if it works.  (If so, send
> us a new source file!)
>
> Russ

Failing that, you will have to wait for Kenji to comment on
the cursor problem (it's already the weekend in Japan).

Re: [9fans] Shared memory graphic on cpu server

[Charles Forsyth]

i often leave cpu consoles in cga mode or make them use a uart

Re: [9fans] Shared memory graphic on cpu server

[Matthias Teege]

jmk@plan9.bell-labs.com writes:

>> to /sys/src/9/pc/vgasavage.c easily enough -- just do what
>> was done for PROSAVAGEK and see if it works.  (If so, send
>> us a new source file!)

> Failing that, you will have to wait for Kenji to comment on
> the cursor problem (it's already the weekend in Japan).

Yes, it failed and i will wait for a comment from kenji.

Thanks, Matthias

--
Matthias Teege -- http://www.mteege.de
make world not war
PGP-Key auf Anfrage

Re: [9fans] Shared memory graphic on cpu server

[Matthias Teege]

Charles Forsyth <forsyth@caldo.demon.co.uk> writes:

> i often leave cpu consoles in cga mode or make them use a uart

Yes, this was adequate for a cpu server but in this special case it
would be nice to have full graphic support because the shuttle
boxes are nice and cheap terminals to. To much CPU power but
compact.

Matthias

--
Matthias Teege -- http://www.mteege.de
make world not war
PGP-Key auf Anfrage

Re: [9fans] Shared memory graphic on cpu server

[Matthias Teege]

Matthias Teege <matthias@mteege.de> writes:


>> Failing that, you will have to wait for Kenji to comment on
>> the cursor problem (it's already the weekend in Japan).

> Yes, it failed and i will wait for a comment from kenji.

Now it works. To my surprise I got it. Where do I drop the new
sources?

Matthias

Re: [9fans] Shared memory graphic on cpu server

[Russ Cox]

you can mail them to 9trouble@plan9.bell-labs.com
and we'll put them in the distribution.

thanks!

Re: [9fans] Shared memory graphic on cpu server

[okamoto]

> Matthias Teege <matthias@mteege.de> writes:
>
>
>>> Failing that, you will have to wait for Kenji to comment on
>>> the cursor problem (it's already the weekend in Japan).
>
>> Yes, it failed and i will wait for a comment from kenji.
>
> Now it works. To my surprise I got it. Where do I drop the new
> sources?

Does this mean, we have two different card with same ID?

Kenji

Re: [9fans] Shared memory graphic on cpu server

[Matthias Teege]

okamoto@granite.cias.osakafu-u.ac.jp writes:

> Does this mean, we have two different card with same ID?

No, the ID was in /lib/vgadb but there was no support in the
kernel. I copy the code from PROSAVAGEK and it works.

Matthias

--
Matthias Teege -- http://www.mteege.de
make world not war
PGP-Key auf Anfrage