From: James A Simmons <jsimmons@acsu.buffalo.edu>
  To  : ggi-develop@eskimo.com
  Date: Mon, 7 Sep 1998 17:12:53 -0400 (EDT)

Re: generic monitor test result

On Mon, 7 Sep 1998, Marcus Sundberg wrote:

> James A Simmons wrote:
> > 
> > I managed to get it to work with some patching. The diffs are below. The
> > reason I disabled the kgi_monitor commands is one to make it compile.
> 
> What was it that prevented it from compiling?

Good thing I keep originals.
------------------------------------------------------------------------

make[1]: Entering directory `/home/jsimmons/degas/kgicon/kgi/chipset'

Chipset driver: Matrox mga2064...

make[2]: Entering directory `/home/jsimmons/degas/kgicon/kgi/chipset/Matrox'
make[2]: `../chipdrv.o' is up to date.
make[2]: Leaving directory `/home/jsimmons/degas/kgicon/kgi/chipset/Matrox'
make[1]: Leaving directory `/home/jsimmons/degas/kgicon/kgi/chipset'
make[1]: Entering directory `/home/jsimmons/degas/kgicon/kgi/ramdac'

Ramdac driver:  TI tvp3026...

make[2]: Entering directory `/home/jsimmons/degas/kgicon/kgi/ramdac/TI'
make[2]: `../dacdrv.o' is up to date.
make[2]: Leaving directory `/home/jsimmons/degas/kgicon/kgi/ramdac/TI'
make[1]: Leaving directory `/home/jsimmons/degas/kgicon/kgi/ramdac'
make[1]: Entering directory `/home/jsimmons/degas/kgicon/kgi/clock'

Clock driver:   prog tvp3026...

make[2]: Entering directory `/home/jsimmons/degas/kgicon/kgi/clock/prog'
make[2]: `../clockdrv.o' is up to date.
make[2]: Leaving directory `/home/jsimmons/degas/kgicon/kgi/clock/prog'
make[1]: Leaving directory `/home/jsimmons/degas/kgicon/kgi/clock'
make[1]: Entering directory `/home/jsimmons/degas/kgicon/kgi/accel'

Accel driver: Matrox mga2064...

make[2]: Entering directory `/home/jsimmons/degas/kgicon/kgi/accel/Matrox'
make[2]: `../accel.o' is up to date.
make[2]: Leaving directory `/home/jsimmons/degas/kgicon/kgi/accel/Matrox'
make[1]: Leaving directory `/home/jsimmons/degas/kgicon/kgi/accel'
make[1]: Entering directory `/home/jsimmons/degas/kgicon/kgi/monitor'

Monitor driver: generic generic...

make[2]: Entering directory `/home/jsimmons/degas/kgicon/kgi/monitor/generic'
cc -I /usr/src/linux/include -I /home/jsimmons/degas/kgicon/kgi/include -I /usr/src/linux/include -I /usr/src/linux/drivers/video -Wall -DDRIVER_MONITOR -Wstrict-prototypes  -O2 -fomit-frame-pointer -pipe  -m486 -DMODULE -DMODVERSIONS -include /usr/src/li
nux/include/linux/modversions.h -D__KERNEL__ -DEXPORT_SYMTAB -c generic.c -o generic.o
generic.c:26: warning: braces around scalar initializer for `monitor.synctype'
generic.c:26: warning: excess elements in scalar initializer after `monitor.synctype'
generic.c:28: parse error before `montype'
generic.c:28: `montyp_timelist' undeclared here (not in a function)
generic.c:28: warning: data definition has no type or storage class
generic.c: In function `kgim_monitor_init':
generic.c:562: `montyp_monosync' undeclared (first use this function)
generic.c:562: (Each undeclared identifier is reported only once
generic.c:562: for each function it appears in.)
generic.c:565: `montyp_timelist' undeclared (first use this function)
generic.c:568: `montyp_multisync' undeclared (first use this function)
generic.c:571: warning: unreachable code at beginning of switch statement
generic.c: In function `kgim_monitor_command':
generic.c:608: `MONITOR_SETTYPE' undeclared (first use this function)
generic.c:615: `montyp_monosync' undeclared (first use this function)
generic.c:619: `montyp_timelist' undeclared (first use this function)
generic.c:623: `montyp_multisync' undeclared (first use this function)
generic.c:626: warning: unreachable code at beginning of switch statement
generic.c:632: `MONITOR_GETTYPE' undeclared (first use this function)
make[2]: *** [generic.o] Error 1
make[2]: Leaving directory `/home/jsimmons/degas/kgicon/kgi/monitor/generic'
make[1]: *** [mondrv.o] Error 2
make[1]: Leaving directory `/home/jsimmons/degas/kgicon/kgi/monitor'
make: *** [monitor/mondrv.o] Error 2

> 
> > Two I want to see all of kgicon except fbcon-kgi.c OS indepent. Using suser
> > makes the monitor driver dependent on the OS.
> 
> suser() has always been used in the monitor drivers, that's nothing
> specific to the generic monitor driver.
> The current KGI API is full of bad things, but as Steffen's new KGI
> will soon be ready there's no point in messing to much with the 
> current one.

Okay. Agree. 

> 
> > Also I'm tring to see if we
> > could use sysctl to work the monitor (this should be in fbcon-kgi.c).
> 
> sysctl is not a replacement for ioctl(), it's just a complement for
> things that should be easily tuneable.
> 
Yes. I was just think of a way not to use suser. I thought of this problem
too. What if you have XF86Config and say XF68FB_DEV tries to use the
monitor ioctl calls to use the modeline configuration. Now you have to set 
the Server as root again to use certain ioctl calls. Please correct me if
I'm thinking wrong. 


> >   Now the results. When I did a insmod kgicon my display shrunk in the
> > horizontal direction and was half way down the monitor. It also was
> > jumping up and down on the monitor.
> 
> Does the monitor setting you used work when compiled into one of
> the old drivers?

No I didn't change the settings in default.h. Perhaps I should try setting
them to the same ones in timelist/monitors.c with __VGA_ defined. The
timelist monitor worked before.