From: Andreas Beck <becka@rz.uni-duesseldorf.de>
  To  : ggi-develop@eskimo.com
  Date: Mon, 21 Jun 1999 01:50:43 +0200

Re: KGIcon and monosync-monitor

Peter Åstrand <altic@lysator.liu.se> wrote:

> > I'm trying to compile KGIcon for the fixed-frequence monitor HP-A1097C. I
> > have made an entry in ../monitor/monosync/monitors.inc, and it compiles
> > allright. But when I try "insmod kgicon.o", I get:

I'd try generic/generic, as you can then change the setup on the fly, and it
is well tested, but the others should work as well.

> Jun 20 22:06:48 j250 kernel: fbcon-kgi.c:1189: Activating kgicon mode
> 1280x1024#1280x1024[16]. 

Oh - I see you have thought about that catch. You of course need to use a
mode that from the number of lines allows for a 1083 lines mode.

It would also help to use a somewhat bigger range for the frequencies.
2-5% should by easily achieveable, as the physical tolerances in any 
device should be big enough.

> #define MON_VERSION     "0.00 ALPHA"                                            
> static struct kgi_monitor                                                       
> monitor = {                                                                     
>         "Hewlett-Packard",      /* manufacturer         */                      
>         "A1097C",               /* model                */                      
>         GM_ALL,                 /* supported modes      */                      
>         FL_MON_NORMAL,          /* special capabilities */                      
>         { 1280, 1024 },         /* maximum resolution   */                      
>         NULL,                   /* Private data hook    */                      
>         { 380, 280 },           /* picture size [mm]    */                      
>         RGB_DISPLAY,            /* RGB monitor          */                      
>         SYNC_ON_GREEN,          /* Sync-on-Green        */                      
>         { 0, 135000000 },       /* pixel clock range    */                      
> # I'm not sure about the line above. I know that the monitors bandwidth is
> # 135, but can the lower bound be zero?

Yes. That can never be achieved anyway as the other setting defy that. No
problem anyway.

>       { 78125, 78125 },     /* hfreq range          */                      
>       { 72, 72 }            /* vfreq range          */                      

Give those some slack. 76563-79687kHz will allow for 2% deviation, which should
be tolerable. 70-74 Hz should do as well.

> static struct kgi_timing                                                        
> vtiming[VTIMINGS] =             /* vertical timing in lines */                  
> { /* width  bstart  sstart  send  bend  total  polarity & htiming */            
>  {   1024,  1026,   1027,   1030, 1080, 1085,  HPOS | VPOS | 0 }                
> };                                                                              

> # My Xfree-modeline was:
> # Modeline "1280x1024-2"  135.00 1280 1344 1536 1728  1024 1027 1030 1085
> # What about HPOS and VPOS? Where do I get such information?

Looks o.k. The sync polarity shouldn't matter for you anyway, as you want to
do SOG - right ?

> static struct kgi_timing                                                        
> htiming[HTIMINGS] =     /* horizontal timing in ns, descending hfreq */         
> {/* width  bstart sstart send   bend   total  dummy */                          
>  {  9481,  9481,  9955,  11377, 11377, 12800, 0 }                               
> };                                                                              
> # 1/135000000*E9 = 7.4074 ns
> # 1280 * 74074 =~ 9481, etc.  Does it matter if one truncate or round 
> # these values?

Hmm - would have to check, but I'd round up, as the mode checker will
probably round down ...

CU, ANdy

-- 
= Andreas Beck                    |  Email :  <andreas.beck@ggi-project.org> =