Radeon FB in 2.6.0-test4 fails for me

Radeon FB in 2.6.0-test4 fails for me

[Peter Chubb]

Hi,
	On a Clevo laptop with a Radeon Mobility M7 LW (AGP), enabling
	the Radeon frame buffer results in an unusable LCD display (as
	if the horizontal and/or vertical sync were incorrect -- lots
	of skewed diagonal lines)

The kernel log shows this:

 radeonfb_pci_register BEGIN
 radeonfb: ref_clk=2700, ref_div=12, xclk=16600 from BIOS
 radeonfb: probed DDR SGRAM 65536k videoram
 radeon_get_moninfo: bios 4 scratch = 1000004
 radeonfb: panel ID string: Samsung LTN150P1-L02
 radeonfb: detected DFP panel size from BIOS
 radeonfb: probed DDR SGRAM 65536k videoram
 radeon_get_moninfo: bios 4 scratch = 1000004
 radeonfb: panel ID string: Samsung LTN150P1-L02
 radeonfb: detected DFP panel size from BIOS: 1400x1050
 radeonfb: ATI Radeon M7 LW DDR SGRAM 64 MB
 radeonfb: DVI port LCD monitor connected
 radeonfb: CRT port no monitor connected
 radeonfb_pci_register END
 hStart = 1440, hEnd = 1552, hTotal = 1688
 vStart = 1050, vEnd = 1053, vTotal = 1063
 h_total_disp = 0xae00d2^I   hsync_strt_wid = 0x8e059a
 v_total_disp = 0x4190426^I   vsync_strt_wid = 0x830419
 post div = 0x2
 fb_div = 0x60
 ppll_div_3 = 0x10060
 ron = 1792, roff = 22036
 vclk_freq = 10800, per = 787
 Console: switching to colour frame buffer device 175x65


With FBDEV disabled in X, I see this in Xfree86.0.log:
(WW) RADEON(0): LCD: Using default hsync range of 28.00-33.00kHz
(WW) RADEON(0): LCD: using default vrefresh range of 43.00-72.00Hz
(II) RADEON(0): Clock range:  12.00 to 350.00 MHz
...
(**) RADEON(0): *Mode "1400x1050": 108.0 MHz (scaled from 0.0 MHz), 64.0 kHz, 60.2 Hz
(II) RADEON(0): Modeline "1400x1050"  108.00  1400 34208 34320 1688  1050 1050 1053 1063

Re: Radeon FB in 2.6.0-test4 fails for me

[Ani Joshi]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 1827 bytes --]


try the attatched patch (its against test4).


ani

On Tue, 26 Aug 2003, Peter Chubb wrote:

>
> Hi,
> 	On a Clevo laptop with a Radeon Mobility M7 LW (AGP), enabling
> 	the Radeon frame buffer results in an unusable LCD display (as
> 	if the horizontal and/or vertical sync were incorrect -- lots
> 	of skewed diagonal lines)
>
> The kernel log shows this:
>
>  radeonfb_pci_register BEGIN
>  radeonfb: ref_clk=2700, ref_div=12, xclk=16600 from BIOS
>  radeonfb: probed DDR SGRAM 65536k videoram
>  radeon_get_moninfo: bios 4 scratch = 1000004
>  radeonfb: panel ID string: Samsung LTN150P1-L02
>  radeonfb: detected DFP panel size from BIOS
>  radeonfb: probed DDR SGRAM 65536k videoram
>  radeon_get_moninfo: bios 4 scratch = 1000004
>  radeonfb: panel ID string: Samsung LTN150P1-L02
>  radeonfb: detected DFP panel size from BIOS: 1400x1050
>  radeonfb: ATI Radeon M7 LW DDR SGRAM 64 MB
>  radeonfb: DVI port LCD monitor connected
>  radeonfb: CRT port no monitor connected
>  radeonfb_pci_register END
>  hStart = 1440, hEnd = 1552, hTotal = 1688
>  vStart = 1050, vEnd = 1053, vTotal = 1063
>  h_total_disp = 0xae00d2^I   hsync_strt_wid = 0x8e059a
>  v_total_disp = 0x4190426^I   vsync_strt_wid = 0x830419
>  post div = 0x2
>  fb_div = 0x60
>  ppll_div_3 = 0x10060
>  ron = 1792, roff = 22036
>  vclk_freq = 10800, per = 787
>  Console: switching to colour frame buffer device 175x65
>
>
> With FBDEV disabled in X, I see this in Xfree86.0.log:
> (WW) RADEON(0): LCD: Using default hsync range of 28.00-33.00kHz
> (WW) RADEON(0): LCD: using default vrefresh range of 43.00-72.00Hz
> (II) RADEON(0): Clock range:  12.00 to 350.00 MHz
> ...
> (**) RADEON(0): *Mode "1400x1050": 108.0 MHz (scaled from 0.0 MHz), 64.0 kHz, 60.2 Hz
> (II) RADEON(0): Modeline "1400x1050"  108.00  1400 34208 34320 1688  1050 1050 1053 1063
>

[-- Attachment #2: Type: TEXT/PLAIN, Size: 583 bytes --]

diff -uNr linux-2.6.0-test4.orig/drivers/video/radeonfb.c linux-2.6.0-test4/drivers/video/radeonfb.c
--- linux-2.6.0-test4.orig/drivers/video/radeonfb.c	2003-08-22 17:01:33.000000000 -0700
+++ linux-2.6.0-test4/drivers/video/radeonfb.c	2003-08-26 10:48:03.000000000 -0700
@@ -2090,7 +2090,7 @@
 	
 	}
 	/* Update fix */
-        info->fix.line_length = rinfo->pitch*64;
+        info->fix.line_length = mode->xres_virtual*(mode->bits_per_pixel/8);
         info->fix.visual = rinfo->depth == 8 ? FB_VISUAL_PSEUDOCOLOR : FB_VISUAL_DIRECTCOLOR;
 
 #ifdef CONFIG_BOOTX_TEXT

Re: Radeon FB in 2.6.0-test4 fails for me

[Peter Chubb]

>>>>> "Ani" == Ani Joshi <ajoshi@unixbox.com> writes:

Ani> try the attatched patch (its against test4).

The patch fixes the text-mode console, but X still cannot use the
frame buffer (the display with UseFB is skewed, and something appears
to go wrong so that all disc activity stops, and the keyboard stops
responding -- as if interrupts were disabled)

Peter C

Re: Radeon FB in 2.6.0-test4 fails for me

[Peter Osterlund]

Ani Joshi <ajoshi@unixbox.com> writes:

> try the attatched patch (its against test4).

Thanks, this patch makes the frame buffer usable on my computer. I
suggest you send it to Linus.

> On Tue, 26 Aug 2003, Peter Chubb wrote:
> 
> >
> > Hi,
> > 	On a Clevo laptop with a Radeon Mobility M7 LW (AGP), enabling
> > 	the Radeon frame buffer results in an unusable LCD display (as
> > 	if the horizontal and/or vertical sync were incorrect -- lots
> > 	of skewed diagonal lines)
> 
> diff -uNr linux-2.6.0-test4.orig/drivers/video/radeonfb.c linux-2.6.0-test4/drivers/video/radeonfb.c
> --- linux-2.6.0-test4.orig/drivers/video/radeonfb.c	2003-08-22 17:01:33.000000000 -0700
> +++ linux-2.6.0-test4/drivers/video/radeonfb.c	2003-08-26 10:48:03.000000000 -0700
> @@ -2090,7 +2090,7 @@
>  	
>  	}
>  	/* Update fix */
> -        info->fix.line_length = rinfo->pitch*64;
> +        info->fix.line_length = mode->xres_virtual*(mode->bits_per_pixel/8);
>          info->fix.visual = rinfo->depth == 8 ? FB_VISUAL_PSEUDOCOLOR : FB_VISUAL_DIRECTCOLOR;
>  
>  #ifdef CONFIG_BOOTX_TEXT

-- 
Peter Osterlund - petero2@telia.com
http://w1.894.telia.com/~u89404340