Re: [directfb-users] Framebuffer woes...

Re: [directfb-users] Framebuffer woes...

[Antonino A. Daplas]

Nathanael D. Noblet wrote:
> Hello,
> 
>     Obviously this isn't *directly* a DirectFB problem I'm experience,
> however I'm hoping someone here might have a solution to my problem.
> 
> I have a VIA EPIA board. It has a built in Trident graphics adapter. The
> problem I'm having is that when I try to get it to run, the lcd OSD
> throws up an error message about 'Out Of Range', sometimes blanking the
> screen, sometimes not.
> 
> I have developed a DirectFB application running on a RHEL 3 (kernel 2.4)
> clone machine with the vesafb framebuffer. It works well on that.
> However when I try to boot my 2.6 kernel with the tridentfb kernel I get
> the out of range. Now I tried setting it with identical fbset settings
> same as the vesafb. For example, when vesafb (2.6 OR 2.4 kernel) fbset
> -i gives
> 
> mode "1024x768-76"
>     # D: 78.653 MHz, H: 59.949 kHz, V: 75.694 Hz
>     geometry 1024 768 1024 768 16
>     timings 12714 128 32 16 4 128 4
>     rgba 5/11,6/5,5/0,0/0
> endmode
> 
> Frame buffer device information:
>     Name        : VESA VGA
>     Address     : 0xd9800000
>     Size        : 6291456
>     Type        : PACKED PIXELS
>     Visual      : TRUECOLOR
>     XPanStep    : 0
>     YPanStep    : 0
>     YWrapStep   : 0
>     LineLength  : 2048
>     Accelerator : No
> 
> The only difference I see when using the tridentfb is "YPanStep : 1."
> 
> I've also tried with epiafb found at http://epiafb.sourceforge.net/
> It loads up, but also gives the out of range errors. I even took the
> lines from
> 
> http://sourceforge.net/forum/forum.php?thread_id=756152&forum_id=221920
> 
> 1024x768x24: 
> fbset -t 10582 208 48 36 1 96 3 -g 1024 768 1024 768 24 -hsync high
> -vsync high 
> 800x600x24: 
> fbset -t 17762 152 32 27 1 64 3 -g 800 600 800 600 24 -hsync high -vsync
> high 
> 640x480x24: 
> fbset -t 27778 80 56 25 1 56 3 -g 640 480 640 480 24 -hsync low -vsync
> low 
> 
> All of those (regardless of depth (8,16,24) all produce 'out of range'
> problems.
> 
> I read
> http://www.ibiblio.org/pub/Linux/docs/HOWTO/other-formats/html_single/Framebuffer-HOWTO.html#s18
> 
> and converted the two 1024x768 modelines
> 75.0 1024 1048 1184 1328 768 771 777 806 -hsync -vsync
> 78.8 1024 1040 1136 1312 768 769 772 800 +hsync +vsync
> Which became
> 75: 13333 144 24 136 29 3 6
> 78: 12690 179 16 96 28 1 3
> 
> X Windows on the RHEL 3 kernel 2.4 system is where those came from, and
> works fine.
> 
> Can ANYONE give me any insight into what I should be doing to fix
> this... pleeeeaaaaase? I'm at the end of my rope with this one.
> 

Try booting with video=tridentfb:fp:1024x768@60.  Also post dmesg, lspci -v
and fbset -i.

I'm CC'ing fbdev-devel.

Tony


-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl

Re: Re: [directfb-users] Framebuffer woes...

[Knut Petersen]

Hi Nathanael,

>>I have a VIA EPIA board. It has a built in Trident graphics adapter. The
>>problem I'm having is that when I try to get it to run, the lcd OSD
>>throws up an error message about 'Out Of Range', sometimes blanking the
>>screen, sometimes not.
>>
>>    
>>

What kind of Epia? I suppose Epia 800 or Epia 5000. If that´s the case, 
disable
tridentfb and enable the cyblafb driver included in the 2.6.14-rc kernels.

>>mode "1024x768-76"
>>    # D: 78.653 MHz, H: 59.949 kHz, V: 75.694 Hz
>>    geometry 1024 768 1024 768 16
>>    timings 12714 128 32 16 4 128 4
>>    rgba 5/11,6/5,5/0,0/0
>>endmode
>>
>>    
>>
Add
       video=cyblafb:1024x768,bpp=16,ref=76
to the  kernel command line. If that does not help immediately, use
       video=cyblafb:1024x768,bpp=16,ref=76,verbosity=2 vga=0x317
and send complete dmesg output after booting as well as lspci -vv output.

There will be a lot of debugging output, be sure to use a big enough
buffer size for kernel messages.

>>All of those (regardless of depth (8,16,24) all produce 'out of range'
>>problems.
>>
>>    
>>
tridentfb timing calcualtion was broken in 2.4 kernels and is broken for 
2.6 kernels,
at least for the cyberblade/i1.

cu,
 Knut


-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl

Re: Re: [directfb-users] Framebuffer woes...

[Nathanael D. Noblet]

On Wed, 2005-10-19 at 08:26 +0200, Knut Petersen wrote:
> Hi Nathanael,
> 
> >>I have a VIA EPIA board. It has a built in Trident graphics adapter. The
> >>problem I'm having is that when I try to get it to run, the lcd OSD
> >>throws up an error message about 'Out Of Range', sometimes blanking the
> >>screen, sometimes not.
> >>
> >>    
> >>
> 
> What kind of Epia? I suppose Epia 800 or Epia 5000. If that´s the case, 
> disable
> tridentfb and enable the cyblafb driver included in the 2.6.14-rc kernels.
> 
> >>mode "1024x768-76"
> >>    # D: 78.653 MHz, H: 59.949 kHz, V: 75.694 Hz
> >>    geometry 1024 768 1024 768 16
> >>    timings 12714 128 32 16 4 128 4
> >>    rgba 5/11,6/5,5/0,0/0
> >>endmode
> >>
> >>    
> >>
> Add
>        video=cyblafb:1024x768,bpp=16,ref=76
> to the  kernel command line. If that does not help immediately, use
>        video=cyblafb:1024x768,bpp=16,ref=76,verbosity=2 vga=0x317
> and send complete dmesg output after booting as well as lspci -vv output.
> 
> There will be a lot of debugging output, be sure to use a big enough
> buffer size for kernel messages.
> 
> >>All of those (regardless of depth (8,16,24) all produce 'out of range'
> >>problems.
> >>
> >>    
> >>
> tridentfb timing calcualtion was broken in 2.4 kernels and is broken for 
> 2.6 kernels,
> at least for the cyberblade/i1.

Well thank you SOO much. I'm not done yet, but those messages are gone.
I will continue testing & working on this, if I run into more problems
or bugs I'll post any feedback/information I have.

-- 
Nathanael D. Noblet
Gnat Solutions, Inc
http://www.gnat.ca/
T 250.385.4613
C 250.383.4613




-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl