Re: [PATCH] atyfb: fix distorted image on PowerMacs

Re: [PATCH] atyfb: fix distorted image on PowerMacs

[Ville Syrjälä]

On Thu, Jan 22, 2009 at 08:22:48PM +0200, Risto Suominen wrote:
> Since the complete re-write in 2.6.10, some PowerMacs (At least PowerMac 5500
> and PowerMac G3 Beige rev A) with ATI Mach64 chip have suffered from unstable
> columns in their framebuffer image. This seems to depend on a value (4) read
> from PLL_EXT_CNTL register, which leads to incorrect DSP config parameters to
> be written to the chip. This patch uses a value calculated by aty_init_pll_ct
> instead, as a starting point.

AFAICS it should be the right thing to do on other systems too. I have
a good collection of mach64 pci cards and laptops so I'll try to find
some time next week to test that theory on x86.

Also cc:ing fbdev-devel...

> Signed-off-by: Risto Suominen <Risto.Suominen@gmail.com>
> ---
> The testing is done on kernel version 2.6.18.
> 
> --- a/drivers/video/aty/mach64_ct.c.org	2006-09-20 06:42:06.000000000 +0300
> +++ b/drivers/video/aty/mach64_ct.c	2009-01-22 14:28:13.000000000 +0200
> @@ -7,6 +7,7 @@
>  #include <linux/delay.h>
>  #include <asm/io.h>
>  #include <video/mach64.h>
> +#include <asm/machdep.h>
>  #include "atyfb.h"
> 
>  #undef DEBUG
> @@ -536,6 +537,14 @@ static int __init aty_init_pll_ct(const
>  	pll->ct.xclk_post_div_real = postdividers[xpost_div];
>  	pll->ct.mclk_fb_div = q * pll->ct.xclk_post_div_real / 8;
> 
> +#ifdef CONFIG_PPC
> +	if (machine_is(powermac)) {
> +		/* Override PLL_EXT_CNTL & 0x07. */
> +		pll->ct.xclk_post_div = xpost_div;
> +		pll->ct.xclk_ref_div = 1;
> +	}
> +#endif
> +
>  #ifdef DEBUG
>  	pllmclk = (1000000 * pll->ct.mclk_fb_mult * pll->ct.mclk_fb_div) /
>  			(par->ref_clk_per * pll->ct.pll_ref_div);

-- 
Ville Syrjälä
syrjala@sci.fi
http://www.sci.fi/~syrjala/

Re: [Linux-fbdev-devel] [PATCH] atyfb: fix distorted image on PowerMacs

[Ville Syrjälä]

On Thu, Jan 22, 2009 at 10:12:04PM +0200, Ville Syrjälä wrote:
> On Thu, Jan 22, 2009 at 08:22:48PM +0200, Risto Suominen wrote:
> > Since the complete re-write in 2.6.10, some PowerMacs (At least PowerMac 5500
> > and PowerMac G3 Beige rev A) with ATI Mach64 chip have suffered from unstable
> > columns in their framebuffer image. This seems to depend on a value (4) read
> > from PLL_EXT_CNTL register, which leads to incorrect DSP config parameters to
> > be written to the chip. This patch uses a value calculated by aty_init_pll_ct
> > instead, as a starting point.
> 
> AFAICS it should be the right thing to do on other systems too. I have
> a good collection of mach64 pci cards and laptops so I'll try to find
> some time next week to test that theory on x86.

Well, the experiment failed on the first Mobility I tried. The FIFO
calculations aren't obvious from the code and there's no mention how
to actually calculate them in the programming guide. So I guess special
casing powermacs is the best way forward.

-- 
Ville Syrjälä
syrjala@sci.fi
http://www.sci.fi/~syrjala/

Re: [Linux-fbdev-devel] [PATCH] atyfb: fix distorted image on PowerMacs

[Benjamin Herrenschmidt]

On Tue, 2009-02-03 at 20:56 +0200, Ville Syrj채l채 wrote:
> On Thu, Jan 22, 2009 at 10:12:04PM +0200, Ville Syrj채l채 wrote:
> > On Thu, Jan 22, 2009 at 08:22:48PM +0200, Risto Suominen wrote:
> > > Since the complete re-write in 2.6.10, some PowerMacs (At least PowerMac 5500
> > > and PowerMac G3 Beige rev A) with ATI Mach64 chip have suffered from unstable
> > > columns in their framebuffer image. This seems to depend on a value (4) read
> > > from PLL_EXT_CNTL register, which leads to incorrect DSP config parameters to
> > > be written to the chip. This patch uses a value calculated by aty_init_pll_ct
> > > instead, as a starting point.
> > 
> > AFAICS it should be the right thing to do on other systems too. I have
> > a good collection of mach64 pci cards and laptops so I'll try to find
> > some time next week to test that theory on x86.
> 
> Well, the experiment failed on the first Mobility I tried. The FIFO
> calculations aren't obvious from the code and there's no mention how
> to actually calculate them in the programming guide. So I guess special
> casing powermacs is the best way forward.

There's one powermac with a M1 (the first clamshell ibook). I don't have
such a machine however, it would be useful if somebody who did could
test if the patch caused a regression there...

It's possible that we should instead base the code on the chip variant.

Ben.