ATI Victoria on Malta

ATI Victoria on Malta

[Andrew Thornton]

Has anyone managed to get the ATI Victoria graphics card working on the MIPS
Malta development board with Linux? This is the card that MIPS specify for
use with the Malta board with another OS (which I shall refrain from
naming).

There does seem to be some code for the ATI Rage XL hardware (as used on the
card) in v2.4.6 of the kernel.

Regards,

Andrew Thornton
Insignia Solutions

Re: ATI Victoria on Malta

[James Simmons]

> There does seem to be some code for the ATI Rage XL hardware (as used on the
> card) in v2.4.6 of the kernel.

I don't have this board but give the atyfb framebuffer a try. 

Re: ATI Victoria on Malta

[Andrew Thornton]

>> There does seem to be some code for the ATI Rage XL hardware (as used on
the
>> card) in v2.4.6 of the kernel.
>
>I don't have this board but give the atyfb framebuffer a try.

This is what I tried. Sadly it locks up at various places and I don't have a
logic analyser or any documentation.

Andrew Thornton

Re: ATI Victoria on Malta

[James Simmons]

> >> There does seem to be some code for the ATI Rage XL hardware (as used on
> the
> >> card) in v2.4.6 of the kernel.
> >
> >I don't have this board but give the atyfb framebuffer a try.
> 
> This is what I tried. Sadly it locks up at various places and I don't have a
> logic analyser or any documentation.

Turn on debugging in atyfb.c and post the results.

/*
 * Debug flags.
 */
#undef DEBUG

Re: ATI Victoria on Malta

[Andrew Thornton]

>Turn on debugging in atyfb.c and post the results.
>
>/*
> * Debug flags.
> */
>#undef DEBUG

OK. I'm afraid I haven't got that much time to spare on this, which is why I
asked if anyone else had managed this!

What I've got is linux-2.4.3.mips-src-01.00.tar.gz (from ftp.mips.com)
patched to make the FPU emulator work reliably (taken from the mail list),
aty.h and atyfb.c from linux-2.4.6.tar.gz (from ftp.kernel.org) with DEBUG
defined. The kernel is configured for a little endian Malta board, virtual
terminal, framebuffer, ATI Mach64, PS2 keyboard (I had to change the MIPS
config.in for this), GDB debugging.

The relevant console output is:

atyfb: 3D RAGE (XL) [0x4752 rev 0x65] 512K SGRAM, 14.31818 MHz XTAL, 230 MHz
PLL
, 120 Mhz MCLK
BUS_CNTL DAC_CNTL MEM_CNTL EXT_MEM_CNTL CRTC_GEN_CNTL DSP_CONFIG DSP_ON_OFF
80000001 06010000 00085838 00000081     04000000      00000000   00000000
PLL ac ac 24 df f6 04 00 fd 8e 9e 65 05 00 00 00 00

It hangs here.

Adding printk()'s, it seems that one of the places it hangs is in
wait_for_fifo() reading FIFO_STAT.

Andrew Thornton

Re: ATI Victoria on Malta

[Geert Uytterhoeven]

On Tue, 24 Jul 2001, Andrew Thornton wrote:
> >Turn on debugging in atyfb.c and post the results.
> >
> >/*
> > * Debug flags.
> > */
> >#undef DEBUG
> 
> OK. I'm afraid I haven't got that much time to spare on this, which is why I
> asked if anyone else had managed this!
> 
> What I've got is linux-2.4.3.mips-src-01.00.tar.gz (from ftp.mips.com)
> patched to make the FPU emulator work reliably (taken from the mail list),
> aty.h and atyfb.c from linux-2.4.6.tar.gz (from ftp.kernel.org) with DEBUG
> defined. The kernel is configured for a little endian Malta board, virtual
> terminal, framebuffer, ATI Mach64, PS2 keyboard (I had to change the MIPS
> config.in for this), GDB debugging.
> 
> The relevant console output is:
> 
> atyfb: 3D RAGE (XL) [0x4752 rev 0x65] 512K SGRAM, 14.31818 MHz XTAL, 230 MHz
> PLL
> , 120 Mhz MCLK
> BUS_CNTL DAC_CNTL MEM_CNTL EXT_MEM_CNTL CRTC_GEN_CNTL DSP_CONFIG DSP_ON_OFF
> 80000001 06010000 00085838 00000081     04000000      00000000   00000000
> PLL ac ac 24 df f6 04 00 fd 8e 9e 65 05 00 00 00 00
> 
> It hangs here.
> 
> Adding printk()'s, it seems that one of the places it hangs is in
> wait_for_fifo() reading FIFO_STAT.

Using `atydebug' (from tools in CVS module atyfb at
http://www.sourceforge.net/projects/linux-fbdev/), the PLL debug values mean:

| tux$ ./atydebug ac ac 24 df f6 04 00 fd 8e 9e 65 05 00 00 00 00
| PLL rate = 417.901480 MHz (guessed)
| bad MCLK post divider 5
| VCLK0 = 414.623821 MHz
| VCLK1 = 232.713765 MHz
| VCLK2 = 86.311678 MHz
| VCLK3 = 165.521763 MHz
| tux$ 

Which looks a bit odd. The same for the 512 K SGRAM.

So I guess the Malta firmware hasn't initialized the RAGE XL yet. And atyfb
requires an initialized chip.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven ------------- Sony Software Development Center Europe (SDCE)
Geert.Uytterhoeven@sonycom.com ------------------- Sint-Stevens-Woluwestraat 55
Voice +32-2-7248626 Fax +32-2-7262686 ---------------- B-1130 Brussels, Belgium

Re: ATI Victoria on Malta

[Andrew Thornton]

Geert,

>Using `atydebug' (from tools in CVS module atyfb at
>http://www.sourceforge.net/projects/linux-fbdev/), the PLL debug values
mean:
>
>| tux$ ./atydebug ac ac 24 df f6 04 00 fd 8e 9e 65 05 00 00 00 00
>| PLL rate = 417.901480 MHz (guessed)
>| bad MCLK post divider 5
>| VCLK0 = 414.623821 MHz
>| VCLK1 = 232.713765 MHz
>| VCLK2 = 86.311678 MHz
>| VCLK3 = 165.521763 MHz
>| tux$
>
>Which looks a bit odd. The same for the 512 K SGRAM.
>
>So I guess the Malta firmware hasn't initialized the RAGE XL yet. And atyfb
>requires an initialized chip.

I guess this is not surprising because the Malta firmware isn't a PC BIOS.

Andrew Thornton

Re: ATI Victoria on Malta

[Geert Uytterhoeven]

On Tue, 24 Jul 2001, Andrew Thornton wrote:
> >Using `atydebug' (from tools in CVS module atyfb at
> >http://www.sourceforge.net/projects/linux-fbdev/), the PLL debug values
> mean:
> >
> >| tux$ ./atydebug ac ac 24 df f6 04 00 fd 8e 9e 65 05 00 00 00 00
> >| PLL rate = 417.901480 MHz (guessed)
> >| bad MCLK post divider 5
> >| VCLK0 = 414.623821 MHz
> >| VCLK1 = 232.713765 MHz
> >| VCLK2 = 86.311678 MHz
> >| VCLK3 = 165.521763 MHz
> >| tux$
> >
> >Which looks a bit odd. The same for the 512 K SGRAM.
> >
> >So I guess the Malta firmware hasn't initialized the RAGE XL yet. And atyfb
> >requires an initialized chip.
> 
> I guess this is not surprising because the Malta firmware isn't a PC BIOS.

If the RAGE XL is the officially supported video board for the Malta, I
wouldn't have been surprised if its firmware would have contained code to
initialize the RAGE XL. But unfortunately this doesn't seem to be the case.

Next question: is there sample code available (e.g. with the `supported' OS for
the Malta) to initialize the RAGE XL?

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven ------------- Sony Software Development Center Europe (SDCE)
Geert.Uytterhoeven@sonycom.com ------------------- Sint-Stevens-Woluwestraat 55
Voice +32-2-7248626 Fax +32-2-7262686 ---------------- B-1130 Brussels, Belgium

Re: ATI Victoria on Malta

[Andrew Thornton]

Geert,

>> I guess this is not surprising because the Malta firmware isn't a PC
BIOS.
>
>If the RAGE XL is the officially supported video board for the Malta, I
>wouldn't have been surprised if its firmware would have contained code to
>initialize the RAGE XL. But unfortunately this doesn't seem to be the case.

I don't know if YAMON supports anything other than a serial terminal for its
console.

>Next question: is there sample code available (e.g. with the `supported' OS
for
>the Malta) to initialize the RAGE XL?

It has been pointed out to me that there is some source code, but this
originates from a non-suitable source for use in Linux. Need I say more?

Andrew Thornton

Re: ATI Victoria on Malta

[Ralf Baechle]

On Tue, Jul 24, 2001 at 02:24:16PM +0100, Andrew Thornton wrote:

> OK. I'm afraid I haven't got that much time to spare on this, which is why I
> asked if anyone else had managed this!
> 
> What I've got is linux-2.4.3.mips-src-01.00.tar.gz (from ftp.mips.com)
> patched to make the FPU emulator work reliably (taken from the mail list),

Sorry to destroy your illusions but we've got still a bunch of rather
tricky bugs in the fp emulation code.

  Ralf

Re: ATI Victoria on Malta

[Andrew Thornton]

Ralf,

>> What I've got is linux-2.4.3.mips-src-01.00.tar.gz (from ftp.mips.com)
>> patched to make the FPU emulator work reliably (taken from the mail
list),
>
>Sorry to destroy your illusions but we've got still a bunch of rather
>tricky bugs in the fp emulation code.

OK then, patched to make the FPU emulator work more reliably.

Andrew Thornton

Re: ATI Victoria on Malta

[Kevin D. Kissell]

>
> > OK. I'm afraid I haven't got that much time to spare on this, which is
why I
> > asked if anyone else had managed this!
> >
> > What I've got is linux-2.4.3.mips-src-01.00.tar.gz (from ftp.mips.com)
> > patched to make the FPU emulator work reliably (taken from the mail
list),
>
> Sorry to destroy your illusions but we've got still a bunch of rather
> tricky bugs in the fp emulation code.

Note, however, that all known bugs have to deal with catching
and handling signals during certain emulation sequences - if
you're doing "straight-line" FP computation, you're probably OK.  ;-)

            Kevin K.

Re: ATI Victoria on Malta

[Carsten Langgaard]

Andrew Thornton wrote:

> Geert,
>
> >> I guess this is not surprising because the Malta firmware isn't a PC
> BIOS.
> >
> >If the RAGE XL is the officially supported video board for the Malta, I
> >wouldn't have been surprised if its firmware would have contained code to
> >initialize the RAGE XL. But unfortunately this doesn't seem to be the case.
>
> I don't know if YAMON supports anything other than a serial terminal for its
> console.

YAMON doesn't initialize the RAGE XL as it is a PCI plug-in card.
The video driver in the 'supported' OS initialize the card.

>
> >Next question: is there sample code available (e.g. with the `supported' OS
> for
> >the Malta) to initialize the RAGE XL?
>
> It has been pointed out to me that there is some source code, but this
> originates from a non-suitable source for use in Linux. Need I say more?

I'm afraid we can't provide you with the sources, you need to contact the
OS vendor yourself.

>
> Andrew Thornton

--
_    _ ____  ___   Carsten Langgaard   Mailto:carstenl@mips.com
|\  /|||___)(___   MIPS Denmark        Direct: +45 4486 5527
| \/ |||    ____)  Lautrupvang 4B      Switch: +45 4486 5555
  TECHNOLOGIES     2750 Ballerup       Fax...: +45 4486 5556
                   Denmark             http://www.mips.com

Re: ATI Victoria on Malta

[James Simmons]

> > It has been pointed out to me that there is some source code, but this
> > originates from a non-suitable source for use in Linux. Need I say more?
> 
> I'm afraid we can't provide you with the sources, you need to contact the
> OS vendor yourself.

I really wish ATI would open up about intializing their cards.