Re: [Dri-devel] Re: Kind of success! (r128 on PPC (Re: LinuxPPC X Server))

On Mon, Aug 07, 2000 at 09:08:01AM +0200, Michel D?nzer wrote:

Benjamin Herrenschmidt wrote:

quoted

The Pismo allows that by scaling the LCD, but I don't think code for that
is implemented in the XFree r128 driver (I didn't check lately).

No it doesn't work, neither with aty128fb. Would that be hard to implement?

If it's anything like the mach 64, it shouldn't be hard to implement
at all.  you just have to program 3 registers.

Speaking of this, I'm having trouble with the stretching setup for my
wallstreet.  It works fine in console mode, but something in the X
server is messing with the chip setup in a way to screw up the
rendering.  After playing with it I was able to get proper output:

console settings:
    mclk = 99.844107 MHz
    vclk = 44.702930 MHz
    dsp_loop_latency = 10
    dsp_precision = 5
    dsp_xclks_per_row = 1143 => 17.859375
    dsp_on = 69 => 34.500000
    dsp_off = 1113 => 556.500000

X (xf4.0.1) settings:
    mclk = 99.844107 MHz
    vclk = 44.702930 MHz
    dsp_loop_latency = 10
    dsp_precision = 4
    dsp_xclks_per_row = 1143 => 8.929688
    dsp_on = 103 => 25.750000
    dsp_off = 1119 => 279.750000

X after tweaks to get it working properly:
    mclk = 99.844107 MHz
    vclk = 59.603906 MHz
    dsp_loop_latency = 10
    dsp_precision = 4
    dsp_xclks_per_row = 1143 => 8.929688
    dsp_on = 131 => 32.750000
    dsp_off = 1986 => 496.500000

using geerts handy ati debugging application, I set the vclk post
divider to 3 (was 4), and changed the high and low watermarks back to
what the console settings were, and things worked great in X now.  Of
course, if I switch back to a VC and back to X the old (bad) settings
come back.

Where's the proper place to fix this?  I assume it's not in the fbcon
interface, because text mode is correctly displayed.

Thoughts?

-- 
Josh
6B21489A | GnuPG ID/Fingerprint | huber@mclx.com |
61F0 6138 BE7B FEBF A223  E9D1 BFE1 2065 6B21 489A

On Mon, 7 Aug 2000, Josh Huber wrote:

On Mon, Aug 07, 2000 at 09:08:01AM +0200, Michel D?nzer wrote:

quoted

Benjamin Herrenschmidt wrote:

quoted

The Pismo allows that by scaling the LCD, but I don't think code for that
is implemented in the XFree r128 driver (I didn't check lately).

No it doesn't work, neither with aty128fb. Would that be hard to implement?

If it's anything like the mach 64, it shouldn't be hard to implement
at all.  you just have to program 3 registers.

Speaking of this, I'm having trouble with the stretching setup for my
wallstreet.  It works fine in console mode, but something in the X
server is messing with the chip setup in a way to screw up the
rendering.  After playing with it I was able to get proper output:

console settings:
    mclk = 99.844107 MHz
    vclk = 44.702930 MHz
    dsp_loop_latency = 10
    dsp_precision = 5
    dsp_xclks_per_row = 1143 => 17.859375
    dsp_on = 69 => 34.500000
    dsp_off = 1113 => 556.500000

X (xf4.0.1) settings:
    mclk = 99.844107 MHz
    vclk = 44.702930 MHz
    dsp_loop_latency = 10
    dsp_precision = 4
    dsp_xclks_per_row = 1143 => 8.929688
    dsp_on = 103 => 25.750000
    dsp_off = 1119 => 279.750000

Is this XF4.0.1 in `fbdev' mode? If yes, it's a bug in the X server. Setting
the DSP registers is part of the video mode initialization and thus is the
sole responsability of the frame buffer device, when running in `fbdev' mode.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/