atyfb and 8500+ATI RagePro

atyfb and 8500+ATI RagePro

[Mike Ladwig]

Greetings!

I've been having major problems using the atyfb driver with my 8500, and
finally have a clue or two to report.

If I try booting the 2.2.6 kernel from the new beta R5 with an unchecked
"No video driver" and "video=atyfb:vmode:19,cmode:16", I get a frozen
machine and a colored static rectangle on my display.  If I use the same
command line, but specify a lower mclk value (most all lower ones work),
the system will boot, but I will get the same types of colored static rectangles.

Recently, I came across a post my Hugy Caley about a "hacked" atyfb
driver from Abe White.  I thought I'd give it a try.  Using the
parameter line, "video=atyfb:vmode:19,cmode:16", I actually get a
somewhat reasonable display, although the text of the console lines are
repeated several times on each  line; every ~25 characters, the line
starts over.  Also, the kernel Hugh sent me didn't boot all the way, not
finding my boot device - probably unrelated to the video issue. The
atyfb driver prints out (a close reconstruction from the garbled video):

/bandit, ATY,XCLAIMVRPro atyfb: 3D RagePro(BGA, PCI) 0x479 rev 0x7c 12M
SGRAM 230 MhzPLL, 100 Mhz MCLK

Anyway, I'm not sure where to go from here.  I'm a programmer and
willing to jump in, but not knowledgable about device driver debugging.

Any help or advice would be greatly appreciated!

Thanks,
mike.

[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]

Re: atyfb and 8500+ATI RagePro

[Geert Uytterhoeven]

	Hi Mike,

> I've been having major problems using the atyfb driver with my 8500, and
> finally have a clue or two to report.
> 
> If I try booting the 2.2.6 kernel from the new beta R5 with an unchecked
> "No video driver" and "video=atyfb:vmode:19,cmode:16", I get a frozen
> machine and a colored static rectangle on my display.  If I use the same
> command line, but specify a lower mclk value (most all lower ones work),
> the system will boot, but I will get the same types of colored static rectangles.
> 
> Recently, I came across a post my Hugy Caley about a "hacked" atyfb
> driver from Abe White.  I thought I'd give it a try.  Using the
> parameter line, "video=atyfb:vmode:19,cmode:16", I actually get a
> somewhat reasonable display, although the text of the console lines are
> repeated several times on each  line; every ~25 characters, the line
> starts over.  Also, the kernel Hugh sent me didn't boot all the way, not

Usually such repeats are caused by incorrect DSP values for the display FIFO.

> finding my boot device - probably unrelated to the video issue. The
> atyfb driver prints out (a close reconstruction from the garbled video):
> 
> /bandit, ATY,XCLAIMVRPro atyfb: 3D RagePro(BGA, PCI) 0x479 rev 0x7c 12M
> SGRAM 230 MhzPLL, 100 Mhz MCLK
> 
> Anyway, I'm not sure where to go from here.  I'm a programmer and
> willing to jump in, but not knowledgable about device driver debugging.
> 
> Any help or advice would be greatly appreciated!

Sorry, no more clues :-(

Greetings,

						Geert

--
Geert Uytterhoeven                     Geert.Uytterhoeven@cs.kuleuven.ac.be
Wavelets, Linux/{m68k~Amiga,PPC~CHRP}  http://www.cs.kuleuven.ac.be/~geert/
Department of Computer Science -- Katholieke Universiteit Leuven -- Belgium



[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]

Re: atyfb and 8500+ATI RagePro

[Mike Ladwig]

I wrote:
> > Recently, I came across a post my Hugy Caley about a "hacked" atyfb
> > driver from Abe White.  I thought I'd give it a try.  Using the
> > parameter line, "video=atyfb:vmode:19,cmode:16", I actually get a
> > somewhat reasonable display

Since writing this, I have found the single key line to the Abe White
patch
that helps me.  I noticed that in the 104 version of atyfb that
RagePro LT stuff similar to the hack was added in, so I thought I'd
focus on
what wasn't added.

If I add an immediate 'return' to aty_set_pll_ct, my
system will boot with BootX parameter "video=atyfb:vmode:19,cmode:16". 
I believe
the atyfb driver is at least somewhat working here, because I can run
Xconfigurator and get an FB_Dev X server up, so long as I pick a display
mode close to
vmode:19,cmode:16.  X doesn't work outside of Xconfigurator, but I hope
that's
a different issue.

The relevent part of dmesg:

MacOS display is /bandit/ATY,XCLAIMVRPro
atyfb: 3D RAGE PRO (BGA, PCI) [0x4749 rev 0x7c] 12M SGRAM, 230 MHz PLL,
100 Mhz
MCLK
Console: switching to colour frame buffer device 160x60
fb0: ATY Mach64 frame buffer device on /bandit/ATY,XCLAIMVRPro
Monitor sense value = 0x73f, using video mode 6 and color mode 0.
fb1: control display adapter

[root@zuul mike]# /usr/sbin/fbset -i

mode "name"
    # D: 125.644 MHz, H: 74.788 kHz, V: 74.788 Hz
    geometry 1280 960 1280 4908 16
    timings 7959 224 32 36 1 144 3
endmode

Frame buffer device information:
    Name        : ATY Mach64
    Address     : 0x81800000
    Size        : 12578816
    Type        : PACKED PIXELS
    Visual      : DIRECTCOLOR
    XPanStep    : 8
    YPanStep    : 1
    YWrapStep   : 0
    LineLength  : 2560
    MMIO Address: 0x817ff800
    MMIO Size   : 2048
    Accelerator : ATI Mach64GT

[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]

Re: atyfb and 8500+ATI RagePro

[Geert Uytterhoeven]

On Thu, 6 May 1999, Mike Ladwig wrote:
> I wrote:
> > > Recently, I came across a post my Hugy Caley about a "hacked" atyfb
> > > driver from Abe White.  I thought I'd give it a try.  Using the
> > > parameter line, "video=atyfb:vmode:19,cmode:16", I actually get a
> > > somewhat reasonable display
> 
> Since writing this, I have found the single key line to the Abe White
> patch
> that helps me.  I noticed that in the 104 version of atyfb that
> RagePro LT stuff similar to the hack was added in, so I thought I'd
> focus on
> what wasn't added.
> 
> If I add an immediate 'return' to aty_set_pll_ct, my
> system will boot with BootX parameter "video=atyfb:vmode:19,cmode:16". 

Hence the problem lies within the programming of the pixel and memory clocks.

> I believe
> the atyfb driver is at least somewhat working here, because I can run
> Xconfigurator and get an FB_Dev X server up, so long as I pick a display
> mode close to
> vmode:19,cmode:16.  X doesn't work outside of Xconfigurator, but I hope

Indeed, then you cannot change the pixel clock so you must choose modes that
are close enough.

Greetings,

						Geert

--
Geert Uytterhoeven                     Geert.Uytterhoeven@cs.kuleuven.ac.be
Wavelets, Linux/{m68k~Amiga,PPC~CHRP}  http://www.cs.kuleuven.ac.be/~geert/
Department of Computer Science -- Katholieke Universiteit Leuven -- Belgium


[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]

Re: atyfb and 8500+ATI RagePro

[Hugh Caley]

As an FYI, 29498928 does work with my PB G3 Series 300mhz; I can boot
with video=atyfb:vmode:14,cmode:32,mclk:83.

Sometimes, though, the console speed is really slow, but if I boot and
exit X, either Xpmac or XF68_FBDev, it becomes faster.  Also, I am
getting the little vertical snow band mentioned below when running X.

I'm using a 2.2.8 kernel I built from CVS source from openprojects
yesterday.

Hugh


"Timothy A. Seufert" wrote:
>
> At 4:48 PM +0200 5/6/99, Geert Uytterhoeven wrote:
> >Hence the problem lies within the programming of the pixel and memory
clocks.
>
> Could it be the same reference clock problem I was having?  (I need to
sit
> down and write some code to parameterize the reference clock.)
>
> Mike, could you try the following change to atyfb.c? (without any
other
> patches applied):
>
> Find where ref_clk_per is defined.  I don't have the source in front
of me
> at the moment, but it will look something like:
>
> unsigned long int ref_clk_per = 1000000000000UL/14318180;
>
> This line sets atyfb's notion of the frequency of the card's reference

> clock (from which all others are derived) to 14.318180 MHz.  Please
try
> changing 14318180 to 29498928 and 28636360 to get 29.499 and 28.636
MHz
> respectively.
>
> I discovered that my Rage Pro card has a 29.498928 MHz reference
clock, and
> had some of the same symptoms as you (can't boot unless mclk is set
low).
> Once I set the reference clock to 29.499 MHz atyfb started working a
lot
> better (though I do get some snow in a thin vertical band near the
left
> side of the screen whenever video memory is being written).

I received similar advice from Geert yesterday.  Indeed it helps with my

problem,
as I too have a card with a reference clock at 29.498 Mhz.  I also get
the same small
amount of snow down the penguin centerline during video activity.

Since the initial success, I have noticed that with actual use, I have
various
problems (thin colored line at boot which goes away with the console,
blanking and of the console display, etc.) so I think more remains to be

done.

I duplicated the DEBUG dump block so that I get a before and after
picture of the
card setup on boot:

MacOS display is /bandit/ATY,XCLAIMVRPro
atyfb: 3D RAGE PRO (BGA, PCI) [0x4749 rev 0x7c] 12M SGRAM, 230 MHz PLL,
100 Mhz
MCLK
BUS_CNTL DAC_CNTL MEM_CNTL EXT_MEM_CNTL CRTC_GEN_CNTL DSP_CONFIG
DSP_ON_OFF
7b33a040 87010182 00651a7b 75130c01     03000200      0038064a
0118061f
PLL ad d5 44 64 e4 13 80 91 8e 9e 29 01 a6 1b 00 00
BUS_CNTL DAC_CNTL MEM_CNTL EXT_MEM_CNTL CRTC_GEN_CNTL DSP_CONFIG
DSP_ON_OFF
7b23a040 87010182 00751a7b 75130c01     03000200      003a0660
063500c0
PLL ad d5 41 14 dc 03 00 8a 8e 9e 29 01 a6 1b 00 00
Console: switching to colour frame buffer device 160x60
fb0: ATY Mach64 frame buffer device on /bandit/ATY,XCLAIMVRPro
Monitor sense value = 0x73f, using video mode 6 and color mode 0.
fb1: control display adapter

mike.



--
"Brain-eating mutants are bad for business..." - Battle Angel
Hugh "Batmensch" Caley


[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]