I am looking for the matroxfb updates patch that went from -ac to
2.4.21-preX? Can somebody give me a clue or a pointer?
Anuradha
--
Debian GNU/Linux (kernel 2.4.20)
panic("Foooooooood fight!");
-- In the kernel source aha1542.c, after detecting a bad segment list
On Wed, Jan 29, 2003 at 08:06:39AM +0600, Anuradha Ratnaweera wrote:
>
> I am looking for the matroxfb updates patch that went from -ac to
> 2.4.21-preX? Can somebody give me a clue or a pointer?
Easiest is probably doing diff between 2.4.21-pre4 and 2.4.20, and
then taking only changes in include/linux/matroxfb.h,
drivers/video/matrox and drivers/video/Config.in... Though I'm
not sure why you just do not upgrade to 2.4.21-pre4.
You can find older version of these changes at
ftp://platan.vc.cvut.cz/pub/linux/matrox-latest, but please note
that 2.4.21-pre4 is more uptodate than 2.4.x patches at the URL
above. 2.5.59 patch at URL above is uptodate, but it is not going to
the kernel, because it is incompatible with other fbdev drivers.
Petr Vandrovec
On Wed, Jan 29, 2003 at 06:31:59AM +0100, Petr Vandrovec wrote:
>
> Though I'm not sure why you just do not upgrade to 2.4.21-pre4.
-pre3 and -pre4 don't build matroxfb_g450 and matroxfb_crtc2 as modules.
I have FB_MATROX_G450 set to "m", so these modules don't get added to
obj-m. The "ifeq"s in the Makefile now check only for the value "y" of
this symbol, not for "m".
Anuradha
--
Debian GNU/Linux (kernel 2.4.20)
Tis man's perdition to be safe, when for the truth he ought to die.
On Wed, Jan 29, 2003 at 01:36:29PM +0600, Anuradha Ratnaweera wrote:
> On Wed, Jan 29, 2003 at 06:31:59AM +0100, Petr Vandrovec wrote:
> >
> > Though I'm not sure why you just do not upgrade to 2.4.21-pre4.
>
> -pre3 and -pre4 don't build matroxfb_g450 and matroxfb_crtc2 as modules.
> I have FB_MATROX_G450 set to "m", so these modules don't get added to
> obj-m. The "ifeq"s in the Makefile now check only for the value "y" of
> this symbol, not for "m".
You did not run 'make oldconfig', did you? G450 cannot be set to module
anymore, either you build core with gX50 support, or without. By default
people use secondary output on g550 and they were complaining that they see
nothing. So you do not have choice to screw up things now.
Petr
On Wed, Jan 29, 2003 at 12:04:20AM -0800, Petr Vandrovec wrote:
>
> On Wed, Jan 29, 2003 at 01:36:29PM +0600, Anuradha Ratnaweera wrote:
> >
> > -pre3 and -pre4 don't build matroxfb_g450 and matroxfb_crtc2 as
> > modules. I have FB_MATROX_G450 set to "m", so these modules don't
> > get added to obj-m. The "ifeq"s in the Makefile now check only for
> > the value "y" of this symbol, not for "m".
>
> You did not run 'make oldconfig', did you?
I did. (I use make-kpkg on Debian, so it does another superfluous make
oldconfig, too). But I tried many other variations before looking at
the Makefile and Config.in itself.
> By default people use secondary output on g550 and they were
> complaining that they see nothing.
Then isn't it then sane to make FB_MATROX_G450 to "bool" and not
"tristate", because selecting "m" does _nothing_ !
> So you do not have choice to screw up things now.
Only for 550 owners, though :-(
BTW, now I _have_ g450 built into the kernel (see signature below) and
both monitors display the same screen. /dev/fb1 gets registered, and
ttys get moved to it with con2fb, but not to any of the monitors. Both
monitors still seem to act as /dev/fb0. The moved tty no longer exist.
Didn't have time to do more research on this.
Thanks.
Anuradha
--
Debian GNU/Linux (kernel 2.4.21-pre4)
It is amazing how complete is the delusion that beauty is goodness.
On Wed, Jan 29, 2003 at 02:22:26PM +0600, Anuradha Ratnaweera wrote:
> On Wed, Jan 29, 2003 at 12:04:20AM -0800, Petr Vandrovec wrote:
> >
> > On Wed, Jan 29, 2003 at 01:36:29PM +0600, Anuradha Ratnaweera wrote:
> > >
> > > -pre3 and -pre4 don't build matroxfb_g450 and matroxfb_crtc2 as
> > > modules. I have FB_MATROX_G450 set to "m", so these modules don't
> > > get added to obj-m. The "ifeq"s in the Makefile now check only for
> > > the value "y" of this symbol, not for "m".
> >
> > You did not run 'make oldconfig', did you?
>
> I did. (I use make-kpkg on Debian, so it does another superfluous make
> oldconfig, too). But I tried many other variations before looking at
> the Makefile and Config.in itself.
Oops. See below, Config.in patch disappeared when patch was travelling
from Alan to Marcelo.
> > By default people use secondary output on g550 and they were
> > complaining that they see nothing.
>
> Then isn't it then sane to make FB_MATROX_G450 to "bool" and not
> "tristate", because selecting "m" does _nothing_ !
Yes, it is...
> > So you do not have choice to screw up things now.
>
> Only for 550 owners, though :-(
>
> BTW, now I _have_ g450 built into the kernel (see signature below) and
> both monitors display the same screen. /dev/fb1 gets registered, and
> ttys get moved to it with con2fb, but not to any of the monitors. Both
> monitors still seem to act as /dev/fb0. The moved tty no longer exist.
What about using matroxset to set things up? It was always mandatory to
use matroxset to set fb<->outputs relation, and now just default changed
because of there is more people who want same picture on both outputs
than people who want different outputs... And with introducing DVI support
there was no other sane alternative.
Marcelo,
Alan missed this chunk when sending you matroxfb updates and I did not
notice. Please apply...
Petr
--- linux-2.4.21-pre4.dist/drivers/video/Config.in 2002-11-29 00:53:15.000000000 +0100
+++ linux-2.4.21-pre4/drivers/video/Config.in 2003-01-29 09:30:30.000000000 +0100
@@ -124,14 +124,20 @@
if [ "$CONFIG_FB_MATROX" != "n" ]; then
bool ' Millennium I/II support' CONFIG_FB_MATROX_MILLENIUM
bool ' Mystique support' CONFIG_FB_MATROX_MYSTIQUE
- bool ' G100/G200/G400/G450/G550 support' CONFIG_FB_MATROX_G100
+ bool ' G100/G200/G400/G450/G550 support' CONFIG_FB_MATROX_G450
+ if [ "$CONFIG_FB_MATROX_G450" = "n" ]; then
+ bool ' G100/G200/G400 support' CONFIG_FB_MATROX_G100A
+ fi
+ if [ "$CONFIG_FB_MATROX_G450" = "y" -o "$CONFIG_FB_MATROX_G100A" = "y" ]; then
+ define_bool CONFIG_FB_MATROX_G100 y
+ fi
if [ "$CONFIG_I2C" != "n" ]; then
dep_tristate ' Matrox I2C support' CONFIG_FB_MATROX_I2C $CONFIG_FB_MATROX $CONFIG_I2C_ALGOBIT
if [ "$CONFIG_FB_MATROX_G100" = "y" ]; then
dep_tristate ' G400 second head support' CONFIG_FB_MATROX_MAVEN $CONFIG_FB_MATROX_I2C
fi
fi
- dep_tristate ' G450/G550 second head support (mandatory for G550)' CONFIG_FB_MATROX_G450 $CONFIG_FB_MATROX_G100
+ dep_tristate ' Matrox /proc interface' CONFIG_FB_MATROX_PROC $CONFIG_FB_MATROX
bool ' Multihead support' CONFIG_FB_MATROX_MULTIHEAD
fi
tristate ' ATI Mach64 display support (EXPERIMENTAL)' CONFIG_FB_ATY
On Wed, Jan 29, 2003 at 12:37:52AM -0800, Petr Vandrovec wrote:
>
> On Wed, Jan 29, 2003 at 02:22:26PM +0600, Anuradha Ratnaweera wrote:
>
> > both monitors display the same screen. /dev/fb1 gets registered, and
> > ttys get moved to it with con2fb, but not to any of the monitors. Both
> > monitors still seem to act as /dev/fb0. The moved tty no longer exist.
>
> What about using matroxset to set things up?
Config.in patch applied.
# ./matroxset -m 00001111
# ioctl failed: Device or resource busy
Have tried other variations, too.
# ls /proc/drivers/mga/
fb0
# cat /proc/drivers/mga/fb0/bios
BIOS: 1.2.21
Output: 0x00
TVOut: no
PINS: found
Info: c02a0a8c
Howerver, fb1 gets registered (yes as the secondary head of fb0).
# dmesg | grep matrox
Kernel command line: BOOT_IMAGE=linux ro root=306 video=matrox:vesa:0x192,font:SUN12x22,nopan
matroxfb: Matrox G450 detected
matroxfb: MTRR's turned on
matroxfb: 1152x864x16bpp (virtual: 1152x864)
matroxfb: framebuffer at 0x42000000, mapped to 0xe0805000, size 33554432
matroxfb_crtc2: secondary head of fb0 was registered as fb1
Anuradha
--
Debian GNU/Linux (kernel 2.4.21-pre4)
Adore, v.:
To venerate expectantly.
-- Ambrose Bierce, "The Devil's Dictionary"
On Wed, Jan 29, 2003 at 03:16:08PM +0600, Anuradha Ratnaweera wrote:
>
> # ./matroxset -m 00001111
> # ioctl failed: Device or resource busy
Got it working. Sorry about the wrong use of matroxfb.
Here is my /etc/init.d/movefb:
#!/bin/sh
/sbin/matroxset -f /dev/fb1 -m 0
/sbin/matroxset -f /dev/fb0 -m 1
/sbin/matroxset -f /dev/fb1 -m 2
/sbin/con2fb /dev/fb1 /dev/tty6
Thanks, Petr.
Anuradha
--
Debian GNU/Linux (kernel 2.4.21-pre4)
According to the obituary notices, a mean and unimportant person never dies.