linuxconsole-dev

Re: secondary graphic card has no memory

[Aivils <ai...@un...>]

On Friday 24 September 2004 09:17, ludovic pollet wrote:
> Hello...
>=20
> Le vendredi 24 Septembre 2004 00:46, Johannes Gajdosik a =E9crit :
> > The result in both cases is the same: according to
> > lspci -v the secondary display adapter has no memory.
> > Anyway it did not crash my computer.
>=20
> That is strange... Maybe the normal way of configuring pci devices doesn'=
t=20
> work on amd64 kernel.=20
> That was my problem (due to the hackvideo thing), and in this case, X rep=
orts=20
> no errors but silently fails to set the pci configuration...

Hack cannot be a hack if Your X fails with screamy cry.
Agree , usage of hackvideo is a bit lightminded, but may make
life easily for John Doe. During development my X uses pci configuration
very, very rarely, which featrure seems is device driver depended.

> Can you try the same setpci command with the swithch -H1 ? It asks setpci=
 to=20
> bypass the kernel when accessing the pci bus.
>=20
> > Up to now only one configuration has some promising results:
> > MGA is primary, NVidia secondary, i386 kernel:
> > In this case I can start the first XServer on the NVidia (vt7),
> > and then a second Xserver on the MGA (vt17). Both Xservers
> > have the prefbusid patch.
>=20
> That looks like a virtual console problem...=20
> If the MGA is primay, I suppose it is used for the vga console display. I=
n=20
> this case you should start the MGA x server on vt7, and NVidia on vt17.
> Appart from the keyboard problem, is the first server really dead ?=20
> (ie does it respond to mouse click ... ) If not, you can try killing the =
X=20
> server with the SIGUSR1 signal.
>=20
> And are you sure that the prefbusid is correctly set ? I think it is deci=
mal,=20
> while lspci reports hexadecimal bus id...

Wow! Bug.

Aivils

Re: secondary graphic card has no memory

[ludovic p. <pl...@nn...>]

Hello...

Le vendredi 24 Septembre 2004 00:46, Johannes Gajdosik a =E9crit :
> The result in both cases is the same: according to
> lspci -v the secondary display adapter has no memory.
> Anyway it did not crash my computer.

That is strange... Maybe the normal way of configuring pci devices doesn'=
t=20
work on amd64 kernel.=20
That was my problem (due to the hackvideo thing), and in this case, X rep=
orts=20
no errors but silently fails to set the pci configuration...
Can you try the same setpci command with the swithch -H1 ? It asks setpci=
 to=20
bypass the kernel when accessing the pci bus.

> Up to now only one configuration has some promising results:
> MGA is primary, NVidia secondary, i386 kernel:
> In this case I can start the first XServer on the NVidia (vt7),
> and then a second Xserver on the MGA (vt17). Both Xservers
> have the prefbusid patch.

That looks like a virtual console problem...=20
If the MGA is primay, I suppose it is used for the vga console display. I=
n=20
this case you should start the MGA x server on vt7, and NVidia on vt17.
Appart from the keyboard problem, is the first server really dead ?=20
(ie does it respond to mouse click ... ) If not, you can try killing the =
X=20
server with the SIGUSR1 signal.

And are you sure that the prefbusid is correctly set ? I think it is deci=
mal,=20
while lspci reports hexadecimal bus id...

Ludovic Pollet
--=20
If it happens once, it's a bug.
If it happens twice, it's a feature.
If it happens more than twice, it's a design philosophy.

Re: Re: secondary graphic card has no memory

[Johannes G. <joh...@gm...>]

Ludovic Pollet wrote:
>Maybe you can enable your mga manually with lspci/setpci :
>the idea is to use the config set by the i386 kernel,
>on the amd64 one...
>
>You can create a dump of your matrox pci configuration on your
>i386 kernel, assuming the busid of the mga is 00:0a.a
>       lspci -x -s 00:0a.0 > mga.pci
>
>You will need to transform it into a setpci command line,
>by piping it into this perl script :
>
>#!/usr/bin/perl
>$line=<STDIN>;
>$i = 0;
>while($line = <STDIN>) {
>        $line=~s/^[0-9a-fA-F][0-9a-fA-F]://;
>        while($line=~m/^ ([0-9a-fA-F][0-9a-fA-F])(.*)$/) {
>                $line = $2;
>                print sprintf(" %x.b=$1",$i);
>                $i++;
>        }
>}
>print "\n";
>
>It will give you something looking like :
>  0.b=2b 1.b=10 2.b=1a 3.b=05 4.b=87  ...
>
>Then on the amd64 kernel, issue the setpci command with
>the registers values :
>       setpci -s XX:XX.X 0.b=2b 1.b=10 2.b=1a 3.b=05 4.b=87
>
>If it works, your mga should be enabled in lspci...
>
>Beware that if the memory/IRQ ressources used on i386
>are not available on x86-64,
>it will certainly just crash your box.

Thanks a lot, I tried in both cases:
1) when the MGA is the primary display adapter I performed
above procedure with the NVidia
2) when the NVidia is the primary display adapter I performed
above procedure with the MGA.
The result in both cases is the same: according to
lspci -v the secondary display adapter has no memory.
Anyway it did not crash my computer.

Up to now only one configuration has some promising results:
MGA is primary, NVidia secondary, i386 kernel:
In this case I can start the first XServer on the NVidia (vt7),
and then a second Xserver on the MGA (vt17). Both Xservers
have the prefbusid patch.
After starting the second XServer I still see an image on the
NVidia, I can move the curser, but the keyboard is dead
(no reaction whatever key I press, no leds).
After stopping the second XServer with <Ctrl-Alt-Backspace>
on the second keyboard I must reboot the computer.

When I use x86_64 kernel the secondary graphic card
(MGA or NVidia, whatever) has no memory.
When I use i386 kernel with MGA as secondary
graphic card the memory on the MGA is disabled and I cannot
start an XServer on the MGA.

Johannes Gajdosik

Dual DRI success, matrox G550 AGP + Radeon 9200SE PCI

[Helge H. <hel...@ai...>]

I finally got my radeon PCI card.
I can now use 3D acceleration on two screens, with
two simultaneous 3D games.

I use ruby 2.6.8.1 and X from debian testing,
with no other patches than the ruby-2.6 patch.


Remaining problems:
The G550 hangs when doing resolution change+3D
(i.e. starting a game that want a different resolution)
while using 24-bit color.  
This force me to use 16-bit color which is ok for games
but unfortunately not for photo editing.

The Radeon 9200SE pci card is ok with 24-bit color,
but not very fast at 3D.

It does 394FPS glxgears, or
59FPS glxgears with glxgears enlarged to
the full 1280x1024. 
It gets 24fps in pinball.

The matrox G550 @ AGP 1X gets twice those numbers,
so I wonder if this all I can expect from a pci radeon?
Or are there any tweaks?

Tuxracer is noticeably choppy at 1280x1024 on the radeon, and smooth
on the G550 at the same resolution.  Of course I can
lower the resolution to fix this. :-/

Helge Hafting

Re: secondary graphic card has no memory

[ludovic p. <pl...@nn...>]

Hello...

I think you misunderstood me : Do not activate the hack video stuff.  It=20
prevents X from activating the mga mmio and bios (it was the problem in m=
y=20
case)

Maybe you can enable your mga manually with lspci/setpci : the idea is to=
 use=20
the config set by the i386 kernel, on the amd64 one...

You can create a dump of your matrox pci configuration on your i386 kerne=
l,=20
assuming the busid of the mga is 00:0a.a
	lspci -x -s 00:0a.0 > mga.pci

You will need to transform it into a setpci command line, by piping it in=
to=20
this perl script :

#!/usr/bin/perl
$line=3D<STDIN>;
$i =3D 0;
while($line =3D <STDIN>) {
        $line=3D~s/^[0-9a-fA-F][0-9a-fA-F]://;
        while($line=3D~m/^ ([0-9a-fA-F][0-9a-fA-F])(.*)$/) {
                $line =3D $2;
                print sprintf(" %x.b=3D$1",$i);
                $i++;
        }
}
print "\n";

It will give you something looking like :  0.b=3D2b 1.b=3D10 2.b=3D1a 3.b=
=3D05=20
4.b=3D87  ...

Then on the amd64 kernel, issue the setpci command with the registers val=
ues :
	setpci -s XX:XX.X 0.b=3D2b 1.b=3D10 2.b=3D1a 3.b=3D05 4.b=3D87

If it works, your mga should be enabled in lspci...

Beware that if the memory/IRQ ressources used on i386 are not available o=
n=20
x86-64, it will certainly just crash your box.

Ludovic Pollet


Le mercredi 22 Septembre 2004 23:35, Johannes Gajdosik a =E9crit :
> Dear list members,
>
> thank you very much for your replies!
> Because Vojtech Pavlik wrote that it has to do
> with x86-64 mainboards I installed on the same hardware
> (AMD64+K8S8XMainboard+NVidiaFX5200+MGAMillenium)
> 2.4.27 kernel for i386 architecture with Jean-Daniel Paugets
> patch. lspci -v showed:
>
> 00:0a.0 VGA compatible controller: Matrox Graphics, Inc. MGA 2064W
> [Millennium] (rev 01) (prog-if 00 [VGA])
>         Flags: stepping, medium devsel, IRQ 5
>         Memory at 40000000 (32-bit, non-prefetchable) [disabled]
> [size=3D16K]
>         Memory at 40800000 (32-bit, prefetchable) [disabled] [size=3D8M=
]
>         Expansion ROM at ffff0000 [disabled] [size=3D64K]
> ...
> 01:00.0 VGA compatible controller: nVidia Corporation NV34 [GeForce FX
> 5200] (rev a1) (prog-if 00 [VGA])
>         Flags: bus master, 66Mhz, medium devsel, latency 248, IRQ 11
>         Memory at fd000000 (32-bit, non-prefetchable) [size=3D16M]
>         Memory at f0000000 (32-bit, prefetchable) [size=3D128M]
>         Expansion ROM at feae0000 [disabled] [size=3D128K]
>         Capabilities: [60] Power Management version 2
>         Capabilities: [44] AGP version 3.0
>
> I wrote
> echo 1 > /proc/bus/pci/hackvideo
> and started the second XServer from
> an xterm on the first XServer:
> X :1 -layout MGA vt17
> which gave me the error
> (EE) MGA(0): Cannot read V_BIOS
>
> I must point out that I used XFree86 4.0.3.1 from
> SuSE 9.0 distribution without prefbusid-patch.
>
> Nothing was displayed on the second monitor and I
> had to kill -9 the second XServer from another xterm,
> it did not react on <Ctrl-Alt-Backspace> from the second
> keyboard.
>
> When I stoped all XServers and started only one Xserver with
> X -layout MGA
> from the console both monitors went blank, the leds on my
> keyboard did not work any more and I had to turn the power off.
>
> Yours,
> Johannes Gajdosik
>
> -----------------------------
>
> Ludovic Pollet wrote:
> > Does lspci returns the same before and after you start the X server ?
>
> Yes, it does.
> ----------------------------
>
> Petr Vandrovec wrote:
> > > When I start X on the MGA I get the error
> > > (WW) ****INVALID MEM ALLOCATION**** b: 0xffffc000 e: 0xffffc000
> > > correcting
> > > (EE) MGA(0): No valid MMIO address in PCI config space
> > > (EE) Screen(s) found, but none have a usable configuration.
> >
> > If you'll build matroxfb into kernel, it should force kernel to
> > allocate
> > MMIO regions for Matrox card, and it will also take a care about
> > initializing
> > matrox's hardware during bootup (X do not contain that code).
>
> Im sorry but make menuconfig tells me:
> Support for frame buffer devices (NOT IMPLEMENTED)
> with the 2.4.27 bruby kernel.
> But I will try 2.6.8.1 later.
>
> ------------
>
> Aivils wrote:
> >> I have kernel 2.4.27 for AMD64 (xf86_64)  and use Jean-Daniel .
> >> Paugets
> >> patch which I have slightly modified by replacing "i386" with
> >> "x86_64"
> >> where it seemed appropriate to me.
> >
> > --- arch/x86_64/kernel/setup.c~ 2004-08-19 14:40:13.000000000 +0300
> > +++ arch/x86_64/kernel/setup.c  2004-09-22 09:46:59.952898432 +0300
> > @@ -577,7 +577,7 @@ void __init setup_arch(char **cmdline_p)
> >         iommu_hole_init();
> >  #endif
> >
> > -#ifdef CONFIG_VT
> > +#if 0
> >  #if defined(CONFIG_VGA_CONSOLE)
> >         conswitchp =3D &vga_con;
> >  #elif defined(CONFIG_DUMMY_CONSOLE)
>
> Thanks very much for the patch. I had to remove the whole block
> in order to make it compile.
>
> >> I had to hack a little on IA32
> >> support, but finally the result was quite ok for me.
>
> I wanted to ask you about arch/x86_64/ia32/ia32_ioctl.c.
> I did the following:
>
>
> ----------------------------------------------------
> static int do_fontx_ioctl(unsigned int fd, int cmd, struct
> consolefontdesc32 *user_cfd, struct file *file)
> {
>        struct consolefontdesc cfdarg;
>        struct console_font_op op;
>        int i, perm;
>
>        perm =3D vt_check(file);
>        if (perm < 0) return perm;
>
>        if (copy_from_user(&cfdarg, user_cfd,
>                sizeof(struct consolefontdesc32)))
>                return -EFAULT;
>
>        cfdarg.chardata =3D (unsigned char *)
>          A(((struct consolefontdesc32 *)&cfdarg)->chardata);
>
>        switch (cmd) {
>        case PIO_FONTX:
> /*gaj
>                if (!perm)
> */
>                        return -EPERM;
> /*
>                op.op =3D KD_FONT_OP_SET;
>                op.flags =3D 0;
>                op.width =3D 8;
>                op.height =3D cfdarg.charheight;
>                op.charcount =3D cfdarg.charcount;
>                op.data =3D cfdarg.chardata;
>                return con_font_op(fg_console, &op);
> */
>        case GIO_FONTX:
> /* gaj
>                if (!cfdarg.chardata)
> */
>                        return 0;
> /*
>                op.op =3D KD_FONT_OP_GET;
>                op.flags =3D 0;
>                op.width =3D 8;
>                op.height =3D cfdarg.charheight;
>                op.charcount =3D cfdarg.charcount;
>                op.data =3D cfdarg.chardata;
>                i =3D con_font_op(fg_console, &op);
>                if (i)
>                        return i;
>                cfdarg.charheight =3D op.height;
>                cfdarg.charcount =3D op.charcount;
>                ((struct consolefontdesc32 *)&cfdarg)->chardata =3D
>                   (unsigned long)cfdarg.chardata;
>                if (copy_to_user(user_cfd, &cfdarg,
>                   sizeof(struct consolefontdesc32)))
>                        return -EFAULT;
>                return 0;
> */
>        }
>        return -EINVAL;
> }
>
> static int do_kdfontop_ioctl(unsigned int fd, unsigned int cmd,
>                  struct console_font_op32 *fontop, struct file *file)
> {
>        struct console_font_op op;
>        int perm =3D vt_check(file), i;
>        struct vt_struct *vt;
>
>        if (perm < 0) return perm;
>
>        if (copy_from_user(&op, (void *) fontop,
>               sizeof(struct console_font_op32)))
>                return -EFAULT;
>        if (!perm && op.op !=3D KD_FONT_OP_GET)
>                return -EPERM;
>        op.data =3D (unsigned char *)
>               A(((struct console_font_op32 *)&op)->data);
>        op.flags |=3D KD_FONT_FLAG_OLD;
>        vt =3D (struct vt_struct *)((struct tty_struct *)
>           file->private_data)->driver_data;
>
> /*gaj
>        i =3D con_font_op(vt->vc_num, &op);
> */
>        i =3D con_font_op(vt->fg_console->vc_num, &op);
>
>        if (i) return i;
>        ((struct console_font_op32 *)&op)->data =3D
>            (unsigned long)op.data;
>        if (copy_to_user((void *) fontop, &op,
>           sizeof(struct console_font_op32)))
>                return -EFAULT;
>        return 0;
> }
>
> static int do_unimap_ioctl(unsigned int fd, unsigned int cmd,
>           struct unimapdesc32 *user_ud, struct file *file)
> {
>        struct unimapdesc32 tmp;
>        int perm =3D vt_check(file);
>
>        if (perm < 0) return perm;
>        if (copy_from_user(&tmp, user_ud, sizeof tmp))
>                return -EFAULT;
>        switch (cmd) {
>        case PIO_UNIMAP:
> /* gaj
>                if (!perm)
> */
>                 return -EPERM;
> /*
>                return con_set_unimap(fg_console, tmp.entry_ct,
>                  (struct unipair *)A(tmp.entries));
> */
>        case GIO_UNIMAP:
>                return -EPERM;
> /* gaj
>                return con_get_unimap(fg_console, tmp.entry_ct,
>                 &(user_ud->entry_ct),
>                (struct unipair *)A(tmp.entries));
> */
>        }
>        return 0;
> }
>
> --------------------------------------------------------------------
> I do not care much about fonts on the console but you certainly
> know a better solution.
>
> Also thank you for your hint using 2.6.8 kernel, I will try that.
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
> Project Admins to receive an Apple iPod Mini FREE for your judgement on
> who ports your project to Linux PPC the best. Sponsored by IBM.
> Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php
> _______________________________________________
> Linuxconsole-dev mailing list
> Lin...@li...
> https://lists.sourceforge.net/lists/listinfo/linuxconsole-dev

--=20
Yes, we will be going to OSI, Mars, and Pluto, but not necessarily in
that order.
		-- Jeffrey Honig

Re: secondary graphic card has no memory

[Johannes G. <joh...@gm...>]

Dear list members,

thank you very much for your replies!
Because Vojtech Pavlik wrote that it has to do
with x86-64 mainboards I installed on the same hardware
(AMD64+K8S8XMainboard+NVidiaFX5200+MGAMillenium)
2.4.27 kernel for i386 architecture with Jean-Daniel Paugets
patch. lspci -v showed:

00:0a.0 VGA compatible controller: Matrox Graphics, Inc. MGA 2064W  
[Millennium] (rev 01) (prog-if 00 [VGA])
        Flags: stepping, medium devsel, IRQ 5
        Memory at 40000000 (32-bit, non-prefetchable) [disabled]  
[size=16K]
        Memory at 40800000 (32-bit, prefetchable) [disabled] [size=8M]
        Expansion ROM at ffff0000 [disabled] [size=64K]
...
01:00.0 VGA compatible controller: nVidia Corporation NV34 [GeForce FX  
5200] (rev a1) (prog-if 00 [VGA])
        Flags: bus master, 66Mhz, medium devsel, latency 248, IRQ 11
        Memory at fd000000 (32-bit, non-prefetchable) [size=16M]
        Memory at f0000000 (32-bit, prefetchable) [size=128M]
        Expansion ROM at feae0000 [disabled] [size=128K]
        Capabilities: [60] Power Management version 2
        Capabilities: [44] AGP version 3.0

I wrote
echo 1 > /proc/bus/pci/hackvideo
and started the second XServer from
an xterm on the first XServer:
X :1 -layout MGA vt17
which gave me the error
(EE) MGA(0): Cannot read V_BIOS

I must point out that I used XFree86 4.0.3.1 from
SuSE 9.0 distribution without prefbusid-patch.

Nothing was displayed on the second monitor and I
had to kill -9 the second XServer from another xterm,
it did not react on <Ctrl-Alt-Backspace> from the second
keyboard.

When I stoped all XServers and started only one Xserver with
X -layout MGA
from the console both monitors went blank, the leds on my
keyboard did not work any more and I had to turn the power off.

Yours,
Johannes Gajdosik

-----------------------------

Ludovic Pollet wrote:
> Does lspci returns the same before and after you start the X server ?

Yes, it does.
----------------------------

Petr Vandrovec wrote:
> > When I start X on the MGA I get the error
> > (WW) ****INVALID MEM ALLOCATION**** b: 0xffffc000 e: 0xffffc000
> > correcting
> > (EE) MGA(0): No valid MMIO address in PCI config space
> > (EE) Screen(s) found, but none have a usable configuration.
> 
> If you'll build matroxfb into kernel, it should force kernel to
> allocate
> MMIO regions for Matrox card, and it will also take a care about
> initializing
> matrox's hardware during bootup (X do not contain that code).

Im sorry but make menuconfig tells me:
Support for frame buffer devices (NOT IMPLEMENTED)
with the 2.4.27 bruby kernel.
But I will try 2.6.8.1 later.

------------

Aivils wrote:
>> I have kernel 2.4.27 for AMD64 (xf86_64)  and use Jean-Daniel . 
>> Paugets
>> patch which I have slightly modified by replacing "i386" with
>> "x86_64"
>> where it seemed appropriate to me.
> 
> --- arch/x86_64/kernel/setup.c~ 2004-08-19 14:40:13.000000000 +0300
> +++ arch/x86_64/kernel/setup.c  2004-09-22 09:46:59.952898432 +0300
> @@ -577,7 +577,7 @@ void __init setup_arch(char **cmdline_p)
>         iommu_hole_init();
>  #endif
> 
> -#ifdef CONFIG_VT
> +#if 0
>  #if defined(CONFIG_VGA_CONSOLE)
>         conswitchp = &vga_con;
>  #elif defined(CONFIG_DUMMY_CONSOLE)

Thanks very much for the patch. I had to remove the whole block
in order to make it compile.

>> I had to hack a little on IA32
>> support, but finally the result was quite ok for me.

I wanted to ask you about arch/x86_64/ia32/ia32_ioctl.c.
I did the following:


----------------------------------------------------
static int do_fontx_ioctl(unsigned int fd, int cmd, struct  
consolefontdesc32 *user_cfd, struct file *file)
{
       struct consolefontdesc cfdarg;
       struct console_font_op op;
       int i, perm;

       perm = vt_check(file);
       if (perm < 0) return perm;

       if (copy_from_user(&cfdarg, user_cfd,
               sizeof(struct consolefontdesc32)))
               return -EFAULT;

       cfdarg.chardata = (unsigned char *)
         A(((struct consolefontdesc32 *)&cfdarg)->chardata);

       switch (cmd) {
       case PIO_FONTX:
/*gaj
               if (!perm)
*/
                       return -EPERM;
/*
               op.op = KD_FONT_OP_SET;
               op.flags = 0;
               op.width = 8;
               op.height = cfdarg.charheight;
               op.charcount = cfdarg.charcount;
               op.data = cfdarg.chardata;
               return con_font_op(fg_console, &op);
*/
       case GIO_FONTX:
/* gaj
               if (!cfdarg.chardata)
*/
                       return 0;
/*
               op.op = KD_FONT_OP_GET;
               op.flags = 0;
               op.width = 8;
               op.height = cfdarg.charheight;
               op.charcount = cfdarg.charcount;
               op.data = cfdarg.chardata;
               i = con_font_op(fg_console, &op);
               if (i)
                       return i;
               cfdarg.charheight = op.height;
               cfdarg.charcount = op.charcount;
               ((struct consolefontdesc32 *)&cfdarg)->chardata =
                  (unsigned long)cfdarg.chardata;
               if (copy_to_user(user_cfd, &cfdarg,
                  sizeof(struct consolefontdesc32)))
                       return -EFAULT;
               return 0;
*/
       }
       return -EINVAL;
}

static int do_kdfontop_ioctl(unsigned int fd, unsigned int cmd,
                 struct console_font_op32 *fontop, struct file *file)
{
       struct console_font_op op;
       int perm = vt_check(file), i;
       struct vt_struct *vt;

       if (perm < 0) return perm;

       if (copy_from_user(&op, (void *) fontop,
              sizeof(struct console_font_op32)))
               return -EFAULT;
       if (!perm && op.op != KD_FONT_OP_GET)
               return -EPERM;
       op.data = (unsigned char *)
              A(((struct console_font_op32 *)&op)->data);
       op.flags |= KD_FONT_FLAG_OLD;
       vt = (struct vt_struct *)((struct tty_struct *)
          file->private_data)->driver_data;

/*gaj
       i = con_font_op(vt->vc_num, &op);
*/
       i = con_font_op(vt->fg_console->vc_num, &op);

       if (i) return i;
       ((struct console_font_op32 *)&op)->data =
           (unsigned long)op.data;
       if (copy_to_user((void *) fontop, &op,
          sizeof(struct console_font_op32)))
               return -EFAULT;
       return 0;
}

static int do_unimap_ioctl(unsigned int fd, unsigned int cmd,
          struct unimapdesc32 *user_ud, struct file *file)
{
       struct unimapdesc32 tmp;
       int perm = vt_check(file);

       if (perm < 0) return perm;
       if (copy_from_user(&tmp, user_ud, sizeof tmp))
               return -EFAULT;
       switch (cmd) {
       case PIO_UNIMAP:
/* gaj
               if (!perm)
*/
                return -EPERM;
/*
               return con_set_unimap(fg_console, tmp.entry_ct,
                 (struct unipair *)A(tmp.entries));
*/
       case GIO_UNIMAP:
               return -EPERM;
/* gaj
               return con_get_unimap(fg_console, tmp.entry_ct,
                &(user_ud->entry_ct),
               (struct unipair *)A(tmp.entries));
*/
       }
       return 0;
}

--------------------------------------------------------------------
I do not care much about fonts on the console but you certainly
know a better solution.

Also thank you for your hint using 2.6.8 kernel, I will try that.

Graphic card mem allocation

[James v. Z. <ja...@dv...>]

In have seen this same error on an Intel Desktop mainboard using only
nVidia cards. The fix for me was to use the nvidia driver on one card
and the xfree nv driver on the second.

The same configuration on every other board I've tried worked fine.

J

On Wed, 2004-09-22 at 08:53, Johannes Gajdosik wrote:

> 
> When I start X on the MGA I get the error
> (WW) ****INVALID MEM ALLOCATION**** b: 0xffffc000 e: 0xffffc000  
> correcting
> (EE) MGA(0): No valid MMIO address in PCI config space
> (EE) Screen(s) found, but none have a usable configuration.

Re: secondary graphic card has no memory

[Vojtech P. <vo...@su...>]

On Wed, Sep 22, 2004 at 10:07:14AM +0300, Aivils wrote:

> > lspci -v shows
> > ....
> > 00:0a.0 VGA compatible controller: Matrox Graphics, Inc. MGA 2064W  
> > [Millennium]
> > (rev 01) (prog-if 00 [VGA])
> >         Flags: stepping, medium devsel, IRQ 18
> >         Memory at <ignored> (32-bit, non-prefetchable) [disabled]
> >         Memory at <ignored> (32-bit, prefetchable) [disabled]
> > ....
> > 01:00.0 VGA compatible controller: nVidia Corporation NV34 [GeForce FX  
> > 5200] (rev a1) (prog-if 00 [VGA])
> >         Flags: bus master, 66Mhz, medium devsel, latency 248, IRQ 16
> >         Memory at fd000000 (32-bit, non-prefetchable) [size=16M]
> >         Memory at f0000000 (32-bit, prefetchable) [size=128M]
> >         Expansion ROM at feae0000 [disabled] [size=128K]
> >         Capabilities: [60] Power Management version 2
> >         Capabilities: [44] AGP version 3.0
> 
> This may interest hairspliters in  the Linux kernel mailing list.
> Cause is not linux-ruby and not XFree86.
> 1) You can try 2.6.8
> 2) Drop milenium into waste box.

It's a mainboard BIOS issue. Happens on many x86-64 mainboards with two
VGA cards installed - only the first one gets resources assigned. I'm
not yet sure why the kernel can't fix that, but at the moment it can't.

> I heard about DIP swiches on the old matrox video cards.
> In one state of dip switch video card is primary. Only when
> dip switch is in the right state, then video card works as secondary.
> But Your card is relative new.
 
It happens with my MGA G400 PCI as well.

-- 
Vojtech Pavlik
SuSE Labs, SuSE CR

Re: secondary graphic card has no memory

[ludovic p. <pl...@nn...>]

Hello...

I had a similar problem with a MGA mystique board, with a 2.4.25 (i386)=20
kernel : the board was not correctly activated in PCI. And X failled to=20
activate it.

My problem was that I was using the hackvideo stuff. I don't think you ne=
ed it=20
with a patched XFree. (cat /proc/bus/pci/hackvideo )

Do you manage to have a running Xserver driving both cards, without any=20
prefbusid ?

Does lspci returns the same before and after you start the X server ?

Ludovic Pollet

Le mercredi 22 Septembre 2004 00:53, Johannes Gajdosik a =E9crit :
> Dear list members,
> Yitzhak Bar Geva has told me I can ask in this list
> for help on problems setting up a multiuser system.
>
> I have kernel 2.4.27 for AMD64 (xf86_64) and use Jean-Daniel Paugets
> patch which I have slightly modified by replacing "i386" with "x86_64"
> where it seemed appropriate to me. I had to hack a little on IA32
> support, but finally the result was quite ok for me.
> I can now attach a second PS2 keyboard and start a second XServer
> on the DUMB console which uses the second keyboard but shows up
> on my primary graphic card (NVidia FX 5200).
>
> Now for the second XServer.
> Yitzhak Bar Geva has told me I cannot use the second head of the NVidia
> for the second XServer, so I installed my old MGA Millenium.
> I use XFree86 4.3.0.1 with xf86-430-prefbusid3.diff.
>
> When I start X on the MGA I get the error
> (WW) ****INVALID MEM ALLOCATION**** b: 0xffffc000 e: 0xffffc000
> correcting
> (EE) MGA(0): No valid MMIO address in PCI config space
> (EE) Screen(s) found, but none have a usable configuration.
>
> lspci -v shows
> ...
> 00:0a.0 VGA compatible controller: Matrox Graphics, Inc. MGA 2064W
> [Millennium]
> (rev 01) (prog-if 00 [VGA])
>         Flags: stepping, medium devsel, IRQ 18
>         Memory at <ignored> (32-bit, non-prefetchable) [disabled]
>         Memory at <ignored> (32-bit, prefetchable) [disabled]
> ...
> 01:00.0 VGA compatible controller: nVidia Corporation NV34 [GeForce FX
> 5200] (rev a1) (prog-if 00 [VGA])
>         Flags: bus master, 66Mhz, medium devsel, latency 248, IRQ 16
>         Memory at fd000000 (32-bit, non-prefetchable) [size=3D16M]
>         Memory at f0000000 (32-bit, prefetchable) [size=3D128M]
>         Expansion ROM at feae0000 [disabled] [size=3D128K]
>         Capabilities: [60] Power Management version 2
>         Capabilities: [44] AGP version 3.0
>
> which suggests to me that the memory on the MGA cannot be used.
> When I make the MGA the primary graphic adapter (BIOS setup)
> it is the other way round:
> The MGA works ok, I can start an XServer on it, but the NVidia has no
> memory and I cannot start an XServer on the NVidia with the same error
> as described above.
>
> I have tried two XFree86 patches I have found on the Internet:
> 1) http://www.redhat.com/archives/axp-list/2001-February/msg00349.html
> 2) http://www.redhat.com/archives/axp-list/2001-February/msg00348.html
> The second patch does not apply cleanly, perhaps it interferes with
> xf86-430-prefbusid3.diff, but I have tried my best to apply it in a
> meaningful way. The resulting XServer works exactly as the unpatched
> XServer: my problem is not fixed.
>
> I would be very glad If you could give me some help on what is going
> wrong. Yours,
>
> Johannes Gajdosik
>
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
> Project Admins to receive an Apple iPod Mini FREE for your judgement on
> who ports your project to Linux PPC the best. Sponsored by IBM.
> Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php
> _______________________________________________
> Linuxconsole-dev mailing list
> Lin...@li...
> https://lists.sourceforge.net/lists/listinfo/linuxconsole-dev

--=20
God made machine language; all the rest is the work of man.

Re: secondary graphic card has no memory

[Aivils <ai...@un...>]

On Wednesday 22 September 2004 01:53, Johannes Gajdosik wrote:
> Dear list members,
> Yitzhak Bar Geva has told me I can ask in this list
> for help on problems setting up a multiuser system.
> 
> I have kernel 2.4.27 for AMD64 (xf86_64) and use Jean-Daniel Paugets
> patch which I have slightly modified by replacing "i386" with "x86_64"
> where it seemed appropriate to me.

--- arch/x86_64/kernel/setup.c~ 2004-08-19 14:40:13.000000000 +0300
+++ arch/x86_64/kernel/setup.c  2004-09-22 09:46:59.952898432 +0300
@@ -577,7 +577,7 @@ void __init setup_arch(char **cmdline_p)
        iommu_hole_init();
 #endif

-#ifdef CONFIG_VT
+#if 0
 #if defined(CONFIG_VGA_CONSOLE)
        conswitchp = &vga_con;
 #elif defined(CONFIG_DUMMY_CONSOLE)

> I had to hack a little on IA32 
> support, but finally the result was quite ok for me.
> I can now attach a second PS2 keyboard and start a second XServer
> on the DUMB console which uses the second keyboard but shows up
> on my primary graphic card (NVidia FX 5200).
> 
> Now for the second XServer.
> Yitzhak Bar Geva has told me I cannot use the second head of the NVidia
> for the second XServer, so I installed my old MGA Millenium.
> I use XFree86 4.3.0.1 with xf86-430-prefbusid3.diff.
> 
> When I start X on the MGA I get the error
> (WW) ****INVALID MEM ALLOCATION**** b: 0xffffc000 e: 0xffffc000  
> correcting
> (EE) MGA(0): No valid MMIO address in PCI config space
> (EE) Screen(s) found, but none have a usable configuration.
> 
> lspci -v shows
> ....
> 00:0a.0 VGA compatible controller: Matrox Graphics, Inc. MGA 2064W  
> [Millennium]
> (rev 01) (prog-if 00 [VGA])
>         Flags: stepping, medium devsel, IRQ 18
>         Memory at <ignored> (32-bit, non-prefetchable) [disabled]
>         Memory at <ignored> (32-bit, prefetchable) [disabled]
> ....
> 01:00.0 VGA compatible controller: nVidia Corporation NV34 [GeForce FX  
> 5200] (rev a1) (prog-if 00 [VGA])
>         Flags: bus master, 66Mhz, medium devsel, latency 248, IRQ 16
>         Memory at fd000000 (32-bit, non-prefetchable) [size=16M]
>         Memory at f0000000 (32-bit, prefetchable) [size=128M]
>         Expansion ROM at feae0000 [disabled] [size=128K]
>         Capabilities: [60] Power Management version 2
>         Capabilities: [44] AGP version 3.0

This may interest hairspliters in  the Linux kernel mailing list.
Cause is not linux-ruby and not XFree86.
1) You can try 2.6.8
2) Drop milenium into waste box.

> which suggests to me that the memory on the MGA cannot be used.
> When I make the MGA the primary graphic adapter (BIOS setup)
> it is the other way round:
> The MGA works ok, I can start an XServer on it, but the NVidia has no
> memory and I cannot start an XServer on the NVidia with the same error
> as described above.
> 
> I have tried two XFree86 patches I have found on the Internet:
> 1) http://www.redhat.com/archives/axp-list/2001-February/msg00349.html
> 2) http://www.redhat.com/archives/axp-list/2001-February/msg00348.html
> The second patch does not apply cleanly, perhaps it interferes with
> xf86-430-prefbusid3.diff, but I have tried my best to apply it in a
> meaningful way. The resulting XServer works exactly as the unpatched
> XServer: my problem is not fixed.

I heard about DIP swiches on the old matrox video cards.
In one state of dip switch video card is primary. Only when
dip switch is in the right state, then video card works as secondary.
But Your card is relative new.

Aivils

secondary graphic card has no memory

[Johannes G. <joh...@gm...>]

Dear list members,
Yitzhak Bar Geva has told me I can ask in this list
for help on problems setting up a multiuser system.

I have kernel 2.4.27 for AMD64 (xf86_64) and use Jean-Daniel Paugets
patch which I have slightly modified by replacing "i386" with "x86_64"
where it seemed appropriate to me. I had to hack a little on IA32
support, but finally the result was quite ok for me.
I can now attach a second PS2 keyboard and start a second XServer
on the DUMB console which uses the second keyboard but shows up
on my primary graphic card (NVidia FX 5200).

Now for the second XServer.
Yitzhak Bar Geva has told me I cannot use the second head of the NVidia
for the second XServer, so I installed my old MGA Millenium.
I use XFree86 4.3.0.1 with xf86-430-prefbusid3.diff.

When I start X on the MGA I get the error
(WW) ****INVALID MEM ALLOCATION**** b: 0xffffc000 e: 0xffffc000  
correcting
(EE) MGA(0): No valid MMIO address in PCI config space
(EE) Screen(s) found, but none have a usable configuration.

lspci -v shows
...
00:0a.0 VGA compatible controller: Matrox Graphics, Inc. MGA 2064W  
[Millennium]
(rev 01) (prog-if 00 [VGA])
        Flags: stepping, medium devsel, IRQ 18
        Memory at <ignored> (32-bit, non-prefetchable) [disabled]
        Memory at <ignored> (32-bit, prefetchable) [disabled]
...
01:00.0 VGA compatible controller: nVidia Corporation NV34 [GeForce FX  
5200] (rev a1) (prog-if 00 [VGA])
        Flags: bus master, 66Mhz, medium devsel, latency 248, IRQ 16
        Memory at fd000000 (32-bit, non-prefetchable) [size=16M]
        Memory at f0000000 (32-bit, prefetchable) [size=128M]
        Expansion ROM at feae0000 [disabled] [size=128K]
        Capabilities: [60] Power Management version 2
        Capabilities: [44] AGP version 3.0

which suggests to me that the memory on the MGA cannot be used.
When I make the MGA the primary graphic adapter (BIOS setup)
it is the other way round:
The MGA works ok, I can start an XServer on it, but the NVidia has no
memory and I cannot start an XServer on the NVidia with the same error
as described above.

I have tried two XFree86 patches I have found on the Internet:
1) http://www.redhat.com/archives/axp-list/2001-February/msg00349.html
2) http://www.redhat.com/archives/axp-list/2001-February/msg00348.html
The second patch does not apply cleanly, perhaps it interferes with
xf86-430-prefbusid3.diff, but I have tried my best to apply it in a
meaningful way. The resulting XServer works exactly as the unpatched
XServer: my problem is not fixed.

I would be very glad If you could give me some help on what is going
wrong. Yours,

Johannes Gajdosik

Re: framebuffer oddities, cont'd...

[Helge H. <hel...@hi...>]

Aivils wrote:

>On Friday 17 September 2004 23:04, Hugo Vanwoerkom wrote:
>  
>
>>--- Hugo Vanwoerkom <hvw...@ya...> wrote:
>>
>>    
>>
>>>I now see that I ought to try nVidia riva in the
>>>kernel without the sun font or vesafb driver and see
>>>if he changes the number of framebuffers.
>>>
>>>      
>>>
>>Result:
>>Only catting to /dev/fb0 produces green speckles on
>>the screen. Anything else gets "no such device".
>>    
>>
> IMHO rivafb does not work for me.
>vesafb-tng at least have stable maintainer, i choose it.
>Lone fb driver take over VGA console and so use
>same vc:1-16. For second X  please use same vc:17
>That mean - no changes, just compile fbcon into
>kernel.
>I am not sure about fbcon and _second_ X server
>compatibility, because each X try to block video
>output.
>  
>
2.6.7 fbcon seems to work fine with dual X servers,
at least on the matrox G550.

Helge Hafting

Re: Is modified hotplug input.agent & X.org vs. XFree

[Svetoslav S. <sv...@gm...>]

> Thanks Aivils. Also, can the X.org supplied with
> Mandrake 10.1 be used for Ruby without a patch? Or do
> we have to patch it like you did with XFree?
> Yitzhak

it's already patched :-)

if you are using "SingleCard" - no changes needed
if you are using "PrefBusID" - change it to "IsolateDevice"

( 
and see the man page :-)
)


> --- Aivils <ai...@un...> wrote:
> 
> > On Saturday 18 September 2004 21:43, yitzhak bar
> > geva wrote:
> > > Is Aivils' modified input.agent still required for
> > > 2.6.8.1, or is it now sufficient to use
> > input.agent
> > > supplied with the latest Mandrake Linux?
> >  Latest Mandrake Linux do not know how to work
> > /proc/bus/console/*
> > because Linus-tree kernel do not contain this.

yes it's needed,
but i guess you could use the older input.agent(from 10.0)

best,

svetljo

-- 
+++ GMX DSL Premiumtarife 3 Monate gratis* + WLAN-Router 0,- EUR* +++
Clevere DSL-Nutzer wechseln jetzt zu GMX: http://www.gmx.net/de/go/dsl

Re: Is modified hotplug input.agent & X.org vs. XFree

[yitzhak b. g. <yit...@ya...>]

Thanks Aivils. Also, can the X.org supplied with
Mandrake 10.1 be used for Ruby without a patch? Or do
we have to patch it like you did with XFree?
Yitzhak
--- Aivils <ai...@un...> wrote:

> On Saturday 18 September 2004 21:43, yitzhak bar
> geva wrote:
> > Is Aivils' modified input.agent still required for
> > 2.6.8.1, or is it now sufficient to use
> input.agent
> > supplied with the latest Mandrake Linux?
>  Latest Mandrake Linux do not know how to work
> /proc/bus/console/*
> because Linus-tree kernel do not contain this.
> 
> Aivils
> 

Re: framebuffer oddities, cont'd...

[Aivils <ai...@un...>]

On Friday 17 September 2004 23:04, Hugo Vanwoerkom wrote:
> 
> --- Hugo Vanwoerkom <hvw...@ya...> wrote:
> 
> > 
> > I now see that I ought to try nVidia riva in the
> > kernel without the sun font or vesafb driver and see
> > if he changes the number of framebuffers.
> > 
> 
> Result:
> Only catting to /dev/fb0 produces green speckles on
> the screen. Anything else gets "no such device".
 IMHO rivafb does not work for me.
vesafb-tng at least have stable maintainer, i choose it.
Lone fb driver take over VGA console and so use
same vc:1-16. For second X  please use same vc:17
That mean - no changes, just compile fbcon into
kernel.
I am not sure about fbcon and _second_ X server
compatibility, because each X try to block video
output.

Aivils

Re: Is modified hotplug input.agent still needed

[Aivils <ai...@un...>]

On Saturday 18 September 2004 21:43, yitzhak bar geva wrote:
> Is Aivils' modified input.agent still required for
> 2.6.8.1, or is it now sufficient to use input.agent
> supplied with the latest Mandrake Linux?
 Latest Mandrake Linux do not know how to work
/proc/bus/console/*
because Linus-tree kernel do not contain this.

Aivils

Re: framebuffer oddities, cont'd...

[Helge H. <hel...@hi...>]

Hugo Vanwoerkom wrote:

>--- Helge Hafting <hel...@hi...> wrote:
>
>  
>
>>You don't really assign framebuffers yourself.  It
>>usually
>>happens automatically, if you have framebuffer
>>drivers for
>>your cards.
>>
>>I have a matrox G550 dualhead card, and my kernel
>>is compiled with two framebuffers for it, one for
>>each head.
>>    
>>
>
>I see that option for "matrox" and "dualhead". But I
>have a nVidia TNT2 on AGP and a nVidia MX-440 on PCI.
>  
>
Then you need framebuffer drivers for those nvidia cards,
if such drivers exists.  (Or do without framebuffers.)

>Like I said, I patched in the vesafb-tng patch. That
>  
>
I don't know what kind of driver this is, vesa
is probably some kind of generic framebuffer driver?
Such drivers tend to be unaccelearated (ok but slow)
and usually only works with the first graphichs card
in the system.  The second one tend to go unused.
A driver made specifically for one kind of card have a better
chance for working for a secondary card too.

>gives you on the kernel config a default resolution
>which I set to 1280x1024-32@85 AND I compile in
>SUN12x22 fonts, but nothing else. Then I get a good
>screen. The "only" thing is on boot, after the initial
>vga stuff, the fbcon starts up with a big white blob
>that moves up and disappears.
>
>Now when I disable vesafb and enable nVidia Riva,
>everything turns to s**t. The screen becomes black
>with tiny little green things above (that's the part
>where the penguin shows) and then afterwards the
>characters become huge (because of the selection of
>the sun12x22 font).
>
>  
>
>>Therefore, /dev/fb0 is the first head and /dev/fb1
>>is the second.
>>cat something > /dev/fb0 cause garbage to appear on
>>the
>>first screen, while cat somthing > /dev/fb1 cause
>>garbage on
>>the other screen. Using /dev/fb2 gives me ENODEV
>>because
>>there is no third framebuffer at this time. There
>>will be when I get my radeon pci card though.
>>
>>
>>    
>>
>>>I do "cat <something> > /dev/fb0 or /dev/fb1 and it
>>>always ends up on the primary console...
>>> 
>>>
>>>      
>>>
>>Very strange.  Only one of them should refer to
>>the primary console.  Have you tried higher numbers,
>>such as fb2, fb3, and so on?
>>
>>Try "ls -l /dev/fb*" and see
>>if the device numbers are correct.  You should
>>get something like this:
>>crw-rw----  1 root video 29, 0 2004-08-16 15:33
>>/dev/fb0
>>crw-rw----  1 root video 29, 1 2004-08-16 15:33
>>/dev/fb1
>>crw-rw----  1 root video 29, 2 2004-08-16 15:33
>>/dev/fb2
>>    
>>
>
>I get:
>
>crw-rw----  1 root video 29,   0 Sep 16 07:07 /dev/fb0
>crw--w----  1 root video 29,   1 Mar 14  2002
>/dev/fb0autodetect
>crw--w----  1 root video 29,   0 Mar 14  2002
>/dev/fb0current
>crw-rw----  1 root video 29,   1 Sep 16 07:07 /dev/fb1
>crw--w----  1 root video 29,  33 Mar 14  2002
>/dev/fb1autodetect
>crw--w----  1 root video 29,  32 Mar 14  2002
>/dev/fb1current
>crw-rw----  1 root video 29,   2 Sep 16 07:07 /dev/fb2
>crw--w----  1 root video 29,  65 Mar 14  2002
>/dev/fb2autodetect
>crw--w----  1 root video 29,  64 Mar 14  2002
>/dev/fb2current
>crw-rw----  1 root video 29,   3 Sep 16 07:07 /dev/fb3
>crw--w----  1 root video 29,  97 Mar 14  2002
>/dev/fb3autodetect
>crw--w----  1 root video 29,  96 Mar 14  2002
>/dev/fb3current
>crw-rw----  1 root video 29,   4 Sep 16 07:07 /dev/fb4
>crw--w----  1 root video 29, 129 Mar 14  2002
>/dev/fb4autodetect
>crw--w----  1 root video 29, 128 Mar 14  2002
>/dev/fb4current
>crw-rw----  1 root video 29,   5 Sep 16 07:07 /dev/fb5
>crw--w----  1 root video 29, 161 Mar 14  2002
>/dev/fb5autodetect
>crw--w----  1 root video 29, 160 Mar 14  2002
>/dev/fb5current
>crw-rw----  1 root video 29,   6 Sep 16 07:07 /dev/fb6
>crw--w----  1 root video 29, 193 Mar 14  2002
>/dev/fb6autodetect
>crw--w----  1 root video 29, 192 Mar 14  2002
>/dev/fb6current
>crw-rw----  1 root video 29,   7 Sep 16 07:07 /dev/fb7
>crw--w----  1 root video 29, 225 Mar 14  2002
>/dev/fb7autodetect
>crw--w----  1 root video 29, 224 Mar 14  2002
>/dev/fb7current
>
>But only catting to /dev/fb0 has any effect.
>
><snip>
>
>I now see that I ought to try nVidia riva in the
>kernel without the sun font or vesafb driver and see
>if he changes the number of framebuffers.
>
>Aivils uses the vesafb-tng patch and if he reads this
>maybe he can say whether that gets him framebuffers
>for all his cards OK. See if he replies.
>
>All in all: 
>1. Obviously framebuffers are much slower than direct
>drivers (like SVGATextMode)
>  
>
SVGATextMode uses the vga text console.  It is fast because an
entire character is written by writing two bytes only. (A character byte
and a color byte.)  Downsides:
* Can't display graphichs on it, obviously
* Limited to 256 or 512 simultaneous different characters on screen, so
   chinese text is tricky at best
* No multicolored fonts
* Font limitations, a character must be 8 or 9 pixels wide.

A framebuffer have to paint each character as graphichs. That is always
more than 2 bytes:
Best case: 1-bit color depth (monochrome) 8x8 font, -> 8 byte per character.
Bad case:  12x22 font, 24-bit color: 792 byte per character!

It doesn't have to be slow though, some framebuffer drivers use the
on-chip graphichs accelerator.  (Same as X uses) They may then match
vgacon for speed. :-) I don't think any VESA thing can use the 
accelerator though.


Other message:

>Result:
>Only catting to /dev/fb0 produces green speckles on
>the screen. Anything else gets "no such device".
>  
>
This shows that the framebuffer driver for your first screen is fine,
and you have no framebuffer driver loaded for the other screen.

If the _same_ driver is supposed to be able to drive both cards, then
you might need to tell it by giving it the right parameters.
I don't know your cards but they have different names,
perhaps they need different drivers?  One way of testing
is to temporarily remove the AGP card and see if the
driver then is capable of driving the PCI card as /dev/fb0.

If so, look at documentation (or source) and try to find
parameters that makes the driver drive two cards.

Helge Hafting

Ruby on laptops, cont'd

[yitzhak b. g. <yit...@ya...>]

This might sound futuristic at present, but the
opportunities are tremendous. Presently, we are
limited by the PCI bus in terms of available slots,
throughput and the fact PCI is only on desktop
computers. Also, the installation procedure requires
opening up the computer, which  limits the acceptance
of Ruby. Laptops and mobile devices are coming down in
price and are widely available. Furthermore, they have
plenty of unused processing power which could easily
be used by additional concurrent users. We should
consider how to output video streams via USB or other
channels such that legacy CRTs, keyboards and mice can
be connected to laptops, etc.
Yitzhak Bar Geva

Is modified hotplug input.agent still needed

[yitzhak b. g. <yit...@ya...>]

Is Aivils' modified input.agent still required for
2.6.8.1, or is it now sufficient to use input.agent
supplied with the latest Mandrake Linux?
Yitzhak Bar Geva

Re: framebuffer oddities, cont'd...

[Hugo V. <hvw...@ya...>]

--- Hugo Vanwoerkom <hvw...@ya...> wrote:

> 
> I now see that I ought to try nVidia riva in the
> kernel without the sun font or vesafb driver and see
> if he changes the number of framebuffers.
> 

Result:
Only catting to /dev/fb0 produces green speckles on
the screen. Anything else gets "no such device".

Hugo.


		
_______________________________
Do you Yahoo!?
Declare Yourself - Register online to vote today!
http://vote.yahoo.com

Getting 2.6.7 patch right

[Hugo V. <hvw...@ya...>]

Hi!

Helge mentioned that he got the 2.6.7 ruby patch from
cvs.
I got it from:
http://sourceforge.net/mailarchive/message.php?msg_id=9284799
and got an error compiling (that I fixed).

I have trouble generating the patch from the cvs tree.
It is huge (3.9mb), which cannot be. Anyone care to
share the right diff options? I thought -Naur would do
it, but not so...

Hugo.


		
__________________________________
Do you Yahoo!?
Yahoo! Mail - 50x more storage than other providers!
http://promotions.yahoo.com/new_mail

Re: framebuffer oddities, cont'd...

[Hugo V. <hvw...@ya...>]

--- Helge Hafting <hel...@hi...> wrote:

> You don't really assign framebuffers yourself.  It
> usually
> happens automatically, if you have framebuffer
> drivers for
> your cards.
> 
> I have a matrox G550 dualhead card, and my kernel
> is compiled with two framebuffers for it, one for
> each head.

I see that option for "matrox" and "dualhead". But I
have a nVidia TNT2 on AGP and a nVidia MX-440 on PCI.
Like I said, I patched in the vesafb-tng patch. That
gives you on the kernel config a default resolution
which I set to 1280x1024-32@85 AND I compile in
SUN12x22 fonts, but nothing else. Then I get a good
screen. The "only" thing is on boot, after the initial
vga stuff, the fbcon starts up with a big white blob
that moves up and disappears.

Now when I disable vesafb and enable nVidia Riva,
everything turns to s**t. The screen becomes black
with tiny little green things above (that's the part
where the penguin shows) and then afterwards the
characters become huge (because of the selection of
the sun12x22 font).

> Therefore, /dev/fb0 is the first head and /dev/fb1
> is the second.
> cat something > /dev/fb0 cause garbage to appear on
> the
> first screen, while cat somthing > /dev/fb1 cause
> garbage on
> the other screen. Using /dev/fb2 gives me ENODEV
> because
> there is no third framebuffer at this time. There
> will be when I get my radeon pci card though.
> 
> 
> >I do "cat <something> > /dev/fb0 or /dev/fb1 and it
> >always ends up on the primary console...
> >  
> >
> Very strange.  Only one of them should refer to
> the primary console.  Have you tried higher numbers,
> such as fb2, fb3, and so on?
> 
> Try "ls -l /dev/fb*" and see
> if the device numbers are correct.  You should
> get something like this:
> crw-rw----  1 root video 29, 0 2004-08-16 15:33
> /dev/fb0
> crw-rw----  1 root video 29, 1 2004-08-16 15:33
> /dev/fb1
> crw-rw----  1 root video 29, 2 2004-08-16 15:33
> /dev/fb2

I get:

crw-rw----  1 root video 29,   0 Sep 16 07:07 /dev/fb0
crw--w----  1 root video 29,   1 Mar 14  2002
/dev/fb0autodetect
crw--w----  1 root video 29,   0 Mar 14  2002
/dev/fb0current
crw-rw----  1 root video 29,   1 Sep 16 07:07 /dev/fb1
crw--w----  1 root video 29,  33 Mar 14  2002
/dev/fb1autodetect
crw--w----  1 root video 29,  32 Mar 14  2002
/dev/fb1current
crw-rw----  1 root video 29,   2 Sep 16 07:07 /dev/fb2
crw--w----  1 root video 29,  65 Mar 14  2002
/dev/fb2autodetect
crw--w----  1 root video 29,  64 Mar 14  2002
/dev/fb2current
crw-rw----  1 root video 29,   3 Sep 16 07:07 /dev/fb3
crw--w----  1 root video 29,  97 Mar 14  2002
/dev/fb3autodetect
crw--w----  1 root video 29,  96 Mar 14  2002
/dev/fb3current
crw-rw----  1 root video 29,   4 Sep 16 07:07 /dev/fb4
crw--w----  1 root video 29, 129 Mar 14  2002
/dev/fb4autodetect
crw--w----  1 root video 29, 128 Mar 14  2002
/dev/fb4current
crw-rw----  1 root video 29,   5 Sep 16 07:07 /dev/fb5
crw--w----  1 root video 29, 161 Mar 14  2002
/dev/fb5autodetect
crw--w----  1 root video 29, 160 Mar 14  2002
/dev/fb5current
crw-rw----  1 root video 29,   6 Sep 16 07:07 /dev/fb6
crw--w----  1 root video 29, 193 Mar 14  2002
/dev/fb6autodetect
crw--w----  1 root video 29, 192 Mar 14  2002
/dev/fb6current
crw-rw----  1 root video 29,   7 Sep 16 07:07 /dev/fb7
crw--w----  1 root video 29, 225 Mar 14  2002
/dev/fb7autodetect
crw--w----  1 root video 29, 224 Mar 14  2002
/dev/fb7current

But only catting to /dev/fb0 has any effect.

<snip>

I now see that I ought to try nVidia riva in the
kernel without the sun font or vesafb driver and see
if he changes the number of framebuffers.

Aivils uses the vesafb-tng patch and if he reads this
maybe he can say whether that gets him framebuffers
for all his cards OK. See if he replies.

All in all: 
1. Obviously framebuffers are much slower than direct
drivers (like SVGATextMode)
2. But the consoles look very nice. I now have exactly
what I have on 2.4.26 (X on PCI vc's + X on AGP).
3. And gpm pastes again... no small thing.
4. And netled uses the keyboard led again, "I think"
with less trouble (I maybe wrong).
5. And no garbage on starting the second X.
6. And not having to be in X on AGP when X on PCI
restarts.

Hugo






		
_______________________________
Do you Yahoo!?
Declare Yourself - Register online to vote today!
http://vote.yahoo.com

Re: keyboard bound to [dummy device]

[Helge H. <hel...@hi...>]

Hugo Vanwoerkom wrote:

>Hi!
>
>Still have a question:
>
>--- Helge Hafting <hel...@hi...> wrote:
>
><snip>
>
>  
>
>>If you have two framebuffers, one for each card,
>>then I recommend
>>using framebuffer console and no dumbcon for ruby
>>2.6.7.
>>
>>    
>>
>
><snip>
>
>I don't understand that. Dense. :-( How do you assign
>framebuffers to a card? And how many framebuffers do I
>have?
>  
>
You don't really assign framebuffers yourself.  It usually
happens automatically, if you have framebuffer drivers for
your cards.

I have a matrox G550 dualhead card, and my kernel
is compiled with two framebuffers for it, one for each head.
Therefore, /dev/fb0 is the first head and /dev/fb1 is the second.
cat something > /dev/fb0 cause garbage to appear on the
first screen, while cat somthing > /dev/fb1 cause garbage on
the other screen. Using /dev/fb2 gives me ENODEV because
there is no third framebuffer at this time. There
will be when I get my radeon pci card though.


>I do "cat <something> > /dev/fb0 or /dev/fb1 and it
>always ends up on the primary console...
>  
>
Very strange.  Only one of them should refer to
the primary console.  Have you tried higher numbers,
such as fb2, fb3, and so on?

Try "ls -l /dev/fb*" and see
if the device numbers are correct.  You should
get something like this:
crw-rw----  1 root video 29, 0 2004-08-16 15:33 /dev/fb0
crw-rw----  1 root video 29, 1 2004-08-16 15:33 /dev/fb1
crw-rw----  1 root video 29, 2 2004-08-16 15:33 /dev/fb2

The date may be different, but the major number should be 29
and the minors should go from 0 and up.  The devices must have
distinct minor numbers.


You can assign keyboards to consoles - once your framebuffer
oddities are sorted out.
Look at the files in /proc/bus/console
For a 2-keyboard machine with one vgacon and one dummycon I have:

cat /proc/bus/console/00/display_desc
VGA+
cat /proc/bus/console/00/keyboard
isa0060/serio0/input0

cat /proc/bus/console/01/display_desc
dummy device
cat /proc/bus/console/01/keyboard    
isa0060/serio1/input0

These are changeable.  I can swap the two keyboards like this, for example:
echo isa060/serio1/input0 > /proc/bus/console/00/keyboard
echo isa060/serio0/input0 > /proc/bus/console/01/keyboard

(serio1 is the ps2 kbd connector, serio0 is the ps2 mouse connector,
where I have another keyboard instead. My mice are serial mice.)

Helge Hafting

Re: keyboard bound to [dummy device]

[Hugo V. <hvw...@ya...>]

Hi!

Still have a question:

--- Helge Hafting <hel...@hi...> wrote:

<snip>

> If you have two framebuffers, one for each card,
> then I recommend
> using framebuffer console and no dumbcon for ruby
> 2.6.7.
> 

<snip>

I don't understand that. Dense. :-( How do you assign
framebuffers to a card? And how many framebuffers do I
have?
I do "cat <something> > /dev/fb0 or /dev/fb1 and it
always ends up on the primary console...



Hugo.



		
_______________________________
Do you Yahoo!?
Declare Yourself - Register online to vote today!
http://vote.yahoo.com

Re: keyboard bound to [dummy device]

[Helge H. <hel...@hi...>]

Hugo Vanwoerkom wrote:

>Hi List!!!
>
>With Helge's help I got 2.6.7 + framebuffers +vc's
>running. Looks great.
>
>Have 2 X's running framebuffer vc's also on primary
>display.
>
>However: 2nd X on PCI card shows the login screen but
>the keyboard doesn't work. (It does on 2.4.x)
>  
>
You usually have to tell X to use the appropriate vt. If you don't,
X grabs the next free vt which is likely bound to the first keyboard.
(Std. linux setup runs gettys at vt1 through vt6, and first X on vt7.
Another X will then grab vt8, which is bound to the first keyboard.
Press ctrl+alt+f8 and the first keyboard brobably works on the
second display.   This is not what you want.

A command line for starting X, taken from /etc/X11/xdm/Xservers:
/usr/X11R6/bin/X vt7 -dpi 100 -nolisten tcp -deferglyphs 16
Note the explicit "vt7" in there.
Ruby 2.6 allocate vt1 through vt16 to the first keyboard,
and vt17 through vt32 to the second.

So, specify "vt17" for the second xserver, and it will use
the second keyboard.  You may also want to
specify a different xf86config file, but I guess you
already figured that out, considering that X came up at all.

Don't specify "vt17" if you're running a getty there, if so,
specify vt18 or whatever higher number you have that
isn't in use.  Note that console switching can be troublesome
on the second console, so I recommendrunning only X on it,
at least in the beginning.

>Dmesg shows this:
>keyboard: [AT Raw Set 2 keyboard] bound to [dummy
>device] vc:17-17
>keyboard: [AT Translated Set 2 keyboard] bound to
>[VESA VGA] vc:1-16
>
>Those messages come out of vt_proc.c, which...
>has Aivils' name on it!
>
>Aivils, what am I doing wrong? Why doesn't he say "AT
>Translated Set 2 keyboard" and VESA VGA for vc:17-17?
>
>Thanks! Getting closer!
>
Are you using a "dumbcon=1" parameter on the kernel command line?
That is ok with X, but not with fbcon.  (vt17 works either way, but
a vt17 bound to dumbcon isn't displayed anywhere.  So it is
unusable for gettys and other console uses.)
If you have two framebuffers, one for each card, then I recommend
using framebuffer console and no dumbcon for ruby 2.6.7.

"dumbcon=1" is the solution for those without fbcon, because
plain vgacon will only grab one keyboard and something (i.e. dumbcon)
mus grab the second keyboard and assign VT numbers to it.

Helge Hafting