[9fans] nVidia GeForce works?

[9fans] nVidia GeForce works?

[Quinn Dunkan]

I'm wondering if Axel ever solved his problem with the cursor offset.  I too
have a nvidia geforce 256 DDR, and at 1600x1200 the cursor image is displayed
512 pixels to the left of where it actually is.  I "solved" it by putting
"scr->offset.x += 512" in nvidiacurload, but that's obviously not addressing
the real problem.

Another interesting artifact is that the top line of the cursor is chopped
off.  But if the cursor is at the top of the screen, you can see the top line
displaced about 48 pixels to the right.

None of these problems occur at 1280x1024.  Also, my nvidiacurload and
nvidiacurmove look like what jmk posted (I'm running the latest update).


thanks

Re: [9fans] nVidia GeForce works?

[Axel Belinfante]

[-- Attachment #1: Type: text/plain, Size: 1242 bytes --]

> I'm wondering if Axel ever solved his problem with the cursor offset.

I did not really solve it; in private mail jmk mentioned that he had it
working succesfully on a Nokia 446Xpro using the 445x entry in vgadb,
so I tried that entry, and the problem went away. I did not look into
it further.
In case anyone is interested, I attach my original monitor description
that triggered the problem. At that time I looked at the driver code,
and noticed that it does something with the 'ehb' mode setting --
without understanding what it did, but wondering whether that's related
to the problem.

Axel.

> I too
> have a nvidia geforce 256 DDR, and at 1600x1200 the cursor image is displayed
> 512 pixels to the left of where it actually is.  I "solved" it by putting
> "scr->offset.x += 512" in nvidiacurload, but that's obviously not addressing
> the real problem.
>
> Another interesting artifact is that the top line of the cursor is chopped
> off.  But if the cursor is at the top of the screen, you can see the top line
> displaced about 48 pixels to the right.
>
> None of these problems occur at 1280x1024.  Also, my nvidiacurload and
> nvidiacurmove look like what jmk posted (I'm running the latest update).


[-- Attachment #2: SM900.VGA --]
[-- Type: text/plain , Size: 388 bytes --]


#
# Samsung SyncMaster 900p
# Horizontal timing:
#	Allowable frequency range: 30-96KHz
# Vertical timing:
#	Allowable frequency range: 50-160Hz
# Video bandwidth:
#	205MHz monitor
#
sm900p
	videobw=200
	alias=multisync
sm900p = 1600x1200						# 93.75Hz, 75.00 KHz
	videobw=205
	defaultclock=202
	shb=1664 ehb=1856 ht=2160
	vrs=1201 vre=1204 vt=1250
	hsync=+ vsync=+

Re: [9fans] nVidia GeForce works?

[Nicholas Waples]

On Mon, 2 Jul 2001, Axel Belinfante wrote:

> > I'm wondering if Axel ever solved his problem with the cursor offset.
>
> I did not really solve it; in private mail jmk mentioned that he had it
> working succesfully on a Nokia 446Xpro using the 445x entry in vgadb,
> so I tried that entry, and the problem went away. I did not look into
> it further.
> In case anyone is interested, I attach my original monitor description
> that triggered the problem. At that time I looked at the driver code,
> and noticed that it does something with the 'ehb' mode setting --
> without understanding what it did, but wondering whether that's related
> to the problem.

 Yes, it was related. My mistake (my first attempt at vga hell). I was
setting the blanking to cover the full overscan, but used 1 bit from the
old value (which added 64 clocks to ehb, which must have caused the 512
pixel offset).

Try http://members.optusnet.com.au/~nwaples/nvidia.c to see if it works
now. I had a quick go with the your monitor definition and it worked for
me, so hopefully....


Nick.

Re: [9fans] nVidia GeForce works?

[Axel Belinfante]

On Thu, 5 Jul 2001 Nicholas Waples replied:
> On Mon, 2 Jul 2001, Axel Belinfante wrote:
>
> > > I'm wondering if Axel ever solved his problem with the cursor offset.
> >
> > I did not really solve it;
[..]
> > In case anyone is interested, I attach my original monitor description
> > that triggered the problem. At that time I looked at the driver code,
> > and noticed that it does something with the 'ehb' mode setting --
> > without understanding what it did, but wondering whether that's related
> > to the problem.
>
>  Yes, it was related. My mistake (my first attempt at vga hell). I was
> setting the blanking to cover the full overscan, but used 1 bit from the
> old value (which added 64 clocks to ehb, which must have caused the 512
> pixel offset).
>
> Try http://members.optusnet.com.au/~nwaples/nvidia.c to see if it works
> now. I had a quick go with the your monitor definition and it worked for
> me, so hopefully....

I saw your message while packing for vacation; I hoped to check your fix
before leaving, but did not make it in time. I'll check and let you know
when I get back. Thanks for your work!

Regards,
Axel,