linux-fbdev-users — Generic mailing list for users of frame buffers

Re: [Linux-fbdev-users] Updated:Error opening framebuffer device/Unknown symbol

[Shane M. <mcd...@gm...>]

Do not be surprised at the lack of response.  You have provided very little
information on your problem, and have not done your homework.

I would strongly suggest purchasing a copy of Linux Kernel Development, 2nd
edition, by Robert Love.  It's an excellent book for giving an introduction
to many aspects of the kernel.  In particular, pages 16 and 17 provide
answers to your questions, which many of the other readers of these lists
have already answered.  A quick read through that book will save you time
when you're confronted with these types of problems.

Shane


On 10/25/07, kaka <sha...@gm...> wrote:
>
> Hi All,
>
> Thanks for the overhelming responses.
> I was able to remove the problem of Unknown symbols by linking the proper
> libraries. Now the problem got reduced to the following messages.
>
> # insmod brcmstfb.ko
> brcmstfb: Unknown symbol printf
> brcmstfb: Unknown symbol malloc
> brcmstfb: Unknown symbol free
> insmod: cannot insert `brcmstfb.ko': Unknown symbol in module (2): No such
> file or directory
> #
>
> for the above problem i had tried to link "libgcc.a " but those symbols
> are also undefined in it also.
> RECAP:
> While running  the cross compiled directFB example on MIPS chip,*
> We tried to install the framebuffer driver(command given above) after
> creating the node fb0.
> APPROACH:
> Actually the code of frambuffer driver consists of usual kernel
> framebuffer code and properitiary graphics lib code.
> The properitiary graphics lib code is using malloc,print and free from <
> stdlib.h> and that is why those symbols are coming undefined.
>
> Could anybody help in this regard?
> Thanks in advance.
>
> kaka
>
>
>
> ---------- Forwarded message ----------
> From: kaka < sha...@gm...>
> Date: Oct 12, 2007 6:33 PM
> Subject: Error opening framebuffer device/Unknown symbol
> register_framebuffer
> To: dir...@di..., dir...@di... .
>
>
>
> > *Hi All,*
> >
> > *While running  the cross compiled directFB example on MIPS chip,*
> >
> > *
> >
> >
> >
> > We tried to install the framebuffer driver(command given at the bottom) and we have already created the node fb0.*
> >
> > *We are getting the following error, *
> >
> >
> > *Can anybody help in this regard ?*
> >
> > *Thanks in Advance.*
> >
> > # ../../cross_directfb/simple_mips
> >
> >      =======================|  DirectFB 1.0.0  |=======================
> >           (c) 2001-2007  The DirectFB Organization (directfb.org)
> >           (c) 2000-2004  Convergence (integrated media) GmbH
> >         ------------------------------------------------------------
> >
> > (*) DirectFB/Core: Single Application Core. (2007-10-05 14:17)
> > (!) Direct/Util: opening '/dev/fb0' failed
> >     --> No such device or address
> > (!) DirectFB/FBDev: Error opening framebuffer device!
> > (!) DirectFB/FBDev: Use 'fbdev' option or set FRAMEBUFFER environment variable.
> > (!) DirectFB/Core: Could not initialize 'system' core!
> >     --> Initialization error!simple.c <96>:
> >         (#) DirectFBError [DirectFBCreate (&dfb)]: Initialization error!
> > #
> >
> > *While running the following command in MIPS chip, we are getting the following error.*
> >
> > # insmod brcmstfb.ko
> > brcmstfb: Unknown symbol unregister_framebuffer
> > brcmstfb: Unknown symbol printf
> > brcmstfb: Unknown symbol malloc
> > brcmstfb: Unknown symbol fb_find_mode
> > brcmstfb: Unknown symbol fb_dealloc_cmap
> > brcmstfb: Unknown symbol fb_alloc_cmap
> > brcmstfb: Unknown symbol framebuffer_release
> > brcmstfb: Unknown symbol free
> > insmod: cannot insert `brcmstfb.ko': Unknown symbol in module (2): No such file or directory
> > #
> > #
> >
> >
> >
> >
>
>
> --
> Thanks & Regards,
> kaka
>
> --
> Thanks & Regards,
> kaka

Re: [Linux-fbdev-users] [directfb-users] Updated:Error opening framebuffer device/Unknown symbol

[gxk <gx...@gm...>]

Do not use functions from libc, like printf, malloc and free.
Find there equivalent for kernel space modules.

kaka wrote:
> Hi All,
>  
> Thanks for the overhelming responses.
> I was able to remove the problem of Unknown symbols by linking the 
> proper libraries. Now the problem got reduced to the following messages.
>  
> # insmod brcmstfb.ko
> brcmstfb: Unknown symbol printf
> brcmstfb: Unknown symbol malloc
> brcmstfb: Unknown symbol free
> insmod: cannot insert `brcmstfb.ko': Unknown symbol in module (2): No 
> such file or directory
> #
>  
> for the above problem i had tried to link "libgcc.a " but those symbols 
> are also undefined in it also.
> RECAP:
> While running  the cross compiled directFB example on MIPS chip,*
> We tried to install the framebuffer driver(command given above) after 
> creating the node fb0.
> APPROACH:
> Actually the code of frambuffer driver consists of usual kernel 
> framebuffer code and properitiary graphics lib code.
> The properitiary graphics lib code is using malloc,print and free from 
> <stdlib.h> and that is why those symbols are coming undefined.
>  
> Could anybody help in this regard?
> Thanks in advance.
>  
> kaka
>  
> 
> 
> ---------- Forwarded message ----------
> From: *kaka* < sha...@gm... <mailto:sha...@gm...>>
> Date: Oct 12, 2007 6:33 PM
> Subject: Error opening framebuffer device/Unknown symbol 
> register_framebuffer
> To: dir...@di... <mailto:dir...@di...>, 
> dir...@di... <mailto:dir...@di...>.
> 
>  
> 
>     *Hi All,*
> 
>     *While running  the cross compiled directFB example on MIPS chip,*
> 
>     *
> 
> 
> 
>     We tried to install the framebuffer driver(command given at the bottom) and we have already created the node fb0.*
> 
>     *We are getting the following error, *
> 
> 
> 
>     *Can anybody help in this regard ?*
> 
>     *Thanks in Advance.*
> 
>     # ../../cross_directfb/simple_mips
>                                                                                                                                  
>          =======================|  DirectFB 1.0.0  |=======================
>               (c) 2001-2007  The DirectFB Organization (directfb.org <http://directfb.org/>)
>               (c) 2000-2004  Convergence (integrated media) GmbH
>             ------------------------------------------------------------
>                                                                                                                                  
>     (*) DirectFB/Core: Single Application Core. (2007-10-05 14:17)
>     (!) Direct/Util: opening '/dev/fb0' failed
>         --> No such device or address
>     (!) DirectFB/FBDev: Error opening framebuffer device!
>     (!) DirectFB/FBDev: Use 'fbdev' option or set FRAMEBUFFER environment variable.
>     (!) DirectFB/Core: Could not initialize 'system' core!
>         --> Initialization error!
>     simple.c <96>:
>             (#) DirectFBError [DirectFBCreate (&dfb)]: Initialization error!
>     #
> 
>     *While running the following command in MIPS chip, we are getting the following error.*
> 
>     # insmod brcmstfb.ko
>     brcmstfb: Unknown symbol unregister_framebuffer
>     brcmstfb: Unknown symbol printf
>     brcmstfb: Unknown symbol malloc
>     brcmstfb: Unknown symbol fb_find_mode
>     brcmstfb: Unknown symbol fb_dealloc_cmap
>     brcmstfb: Unknown symbol fb_alloc_cmap
>     brcmstfb: Unknown symbol framebuffer_release
>     brcmstfb: Unknown symbol free
>     insmod: cannot insert `brcmstfb.ko': Unknown symbol in module (2): No such file or directory
>     #
>     #
> 
>      
> 
> 
> 
> 
> -- 
> Thanks & Regards,
> kaka
> 
> -- 
> Thanks & Regards,
> kaka
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> directfb-users mailing list
> dir...@di...
> http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-users

-- 
     Three things are certain:
     Death, taxes, and lost data.
     Guess which has occurred.

Re: [Linux-fbdev-users] Updated:Error opening framebuffer device/Unknown symbol

[Jakub Z. <dar...@da...>]

On Thu, Oct 25, 2007 at 02:25:40PM +0530, kaka wrote:
> Hi All,
> 
> Thanks for the overhelming responses.
> I was able to remove the problem of Unknown symbols by linking the proper
> libraries. Now the problem got reduced to the following messages.
> 
> # insmod brcmstfb.ko
> brcmstfb: Unknown symbol printf
instead of printf() you should use printk()

sample, printk(KERN_INFO "somethink happen! var1=%d\n", var1);
instead of KERN_INFO you've got defined in <linux/kernel.h> other
message type like: KERN_EMERG, KERN_ERR, KERN_WARNING, KERN_NOTICE,
KERN_DEBUG

see include/linux/kernel.h for details.

> brcmstfb: Unknown symbol malloc
kmalloc() ? kzalloc() ?

> brcmstfb: Unknown symbol free
kfree()

> The properitiary graphics lib code is using malloc,print and free from <
> stdlib.h> and that is why those symbols are coming undefined.
Because you're in kernelspace, not in userspace with (g)libc.

Re: [Linux-fbdev-users] Updated:Error opening framebuffer device/Unknown symbol

[Jan-Benedict G. <jb...@lu...>]

On Thu, 2007-10-25 14:25:40 +0530, kaka <sha...@gm...> wrote:
>=20
> Thanks for the overhelming responses.
> I was able to remove the problem of Unknown symbols by linking the proper
> libraries. Now the problem got reduced to the following messages.
>=20
> # insmod brcmstfb.ko
> brcmstfb: Unknown symbol printf
> brcmstfb: Unknown symbol malloc
> brcmstfb: Unknown symbol free
> insmod: cannot insert `brcmstfb.ko': Unknown symbol in module (2): No such
> file or directory
> #

You cannot use printf(), malloc() and free() in your driver. This is
kernel code, not userland. Kernel has similar functions (printk(),
various memory-allocating functions and kfree()), but keep in mind
that they may behave a bit differently.

> for the above problem i had tried to link "libgcc.a " but those symbols a=
re
> also undefined in it also.

This may not be correct for all cases. libgcc.a is compiled for the
target the compiler is targeted to. In unfortunate cases, there may be
opcodes used in libgcc that are not available in your CPU...

> RECAP:
> While running  the cross compiled directFB example on MIPS chip,*
> We tried to install the framebuffer driver(command given above) after
> creating the node fb0.

You'd post your driver's sources. That way, we'd suggest improvements
to correctly adapt it to Linux's APIs.

> APPROACH:
> Actually the code of frambuffer driver consists of usual kernel framebuff=
er
> code and properitiary graphics lib code.
> The properitiary graphics lib code is using malloc,print and free from <
> stdlib.h> and that is why those symbols are coming undefined.

stdlib.h functionality is not per se available. You either need to
implement the missing parts, or rework the stuff to use the proper
kernel interfaces (which is what I recommend.)

MfG, JBG

--=20
      Jan-Benedict Glaw      jb...@lu...              +49-172-7608481
Signature of:           Ich hatte in letzter Zeit ein bi=C3=9Fchen viel Rea=
litycheck.
the second  :               Langsam m=C3=B6chte ich mal wieder weitertr=C3=
=A4umen k=C3=B6nnen.
                             -- Maximilian Wilhelm (18. Mai 2006, #lug-owl.=
de)

Re: [Linux-fbdev-users] Updated:Error opening framebuffer device/Unknown symbol

[Manuel L. <ma...@ro...>]

On Thu, Oct 25, 2007 at 02:25:40PM +0530, kaka wrote:
> Hi All,
> 
> Thanks for the overhelming responses.
> I was able to remove the problem of Unknown symbols by linking the proper
> libraries. Now the problem got reduced to the following messages.
> 
> # insmod brcmstfb.ko
> brcmstfb: Unknown symbol printf
> brcmstfb: Unknown symbol malloc
> brcmstfb: Unknown symbol free
> insmod: cannot insert `brcmstfb.ko': Unknown symbol in module (2): No such
> file or directory

kernel does not have the above mentioned functions, but you could
go over your code and replace all occurrences of the unknown symbols
with these:
printf -> printk
malloc -> kmalloc
free -> kfree

Pay attention to the different parameters used by some of those.

-- 
 ml.

[Linux-fbdev-users] Updated:Error opening framebuffer device/Unknown symbol

[kaka <sha...@gm...>]

Hi All,

Thanks for the overhelming responses.
I was able to remove the problem of Unknown symbols by linking the proper
libraries. Now the problem got reduced to the following messages.

# insmod brcmstfb.ko
brcmstfb: Unknown symbol printf
brcmstfb: Unknown symbol malloc
brcmstfb: Unknown symbol free
insmod: cannot insert `brcmstfb.ko': Unknown symbol in module (2): No such
file or directory
#

for the above problem i had tried to link "libgcc.a " but those symbols are
also undefined in it also.
RECAP:
While running  the cross compiled directFB example on MIPS chip,*
We tried to install the framebuffer driver(command given above) after
creating the node fb0.
APPROACH:
Actually the code of frambuffer driver consists of usual kernel framebuffer
code and properitiary graphics lib code.
The properitiary graphics lib code is using malloc,print and free from <
stdlib.h> and that is why those symbols are coming undefined.

Could anybody help in this regard?
Thanks in advance.

kaka



---------- Forwarded message ----------
From: kaka <sha...@gm...>
Date: Oct 12, 2007 6:33 PM
Subject: Error opening framebuffer device/Unknown symbol
register_framebuffer
To: dir...@di..., dir...@di... .



> *Hi All,*
>
> *While running  the cross compiled directFB example on MIPS chip,*
>
> *
>
>
> We tried to install the framebuffer driver(command given at the bottom) and we have already created the node fb0.*
>
> *We are getting the following error, *
>
> *Can anybody help in this regard ?*
>
> *Thanks in Advance.*
>
> # ../../cross_directfb/simple_mips
>
>      =======================|  DirectFB 1.0.0  |=======================
>           (c) 2001-2007  The DirectFB Organization (directfb.org)
>           (c) 2000-2004  Convergence (integrated media) GmbH
>         ------------------------------------------------------------
>
> (*) DirectFB/Core: Single Application Core. (2007-10-05 14:17)
> (!) Direct/Util: opening '/dev/fb0' failed
>     --> No such device or address
> (!) DirectFB/FBDev: Error opening framebuffer device!
> (!) DirectFB/FBDev: Use 'fbdev' option or set FRAMEBUFFER environment variable.
> (!) DirectFB/Core: Could not initialize 'system' core!
>     --> Initialization error!simple.c <96>:
>         (#) DirectFBError [DirectFBCreate (&dfb)]: Initialization error!
> #
>
> *While running the following command in MIPS chip, we are getting the following error.*
>
> # insmod brcmstfb.ko
> brcmstfb: Unknown symbol unregister_framebuffer
> brcmstfb: Unknown symbol printf
> brcmstfb: Unknown symbol malloc
> brcmstfb: Unknown symbol fb_find_mode
> brcmstfb: Unknown symbol fb_dealloc_cmap
> brcmstfb: Unknown symbol fb_alloc_cmap
> brcmstfb: Unknown symbol framebuffer_release
> brcmstfb: Unknown symbol free
> insmod: cannot insert `brcmstfb.ko': Unknown symbol in module (2): No such file or directory
> #
> #
>
>
>
>


-- 
Thanks & Regards,
kaka

-- 
Thanks & Regards,
kaka

Re: [Linux-fbdev-users] Switching to my chosen console font when the framebuffer console starts.

[<thu...@xs...>]

From: cga2000 <cg...@op...>
Date: Mon, Oct 22, 2007 at 06:30:30PM -0400
> On Mon, Oct 22, 2007 at 03:20:54PM EDT, thu...@xs... wrote:
> > From: cga2000 <cg...@op...>
> > Date: Sun, Oct 21, 2007 at 08:44:29PM -0400
> > > It looks like I need to:
> > > 
> > > 1. convert the .psf font to a text format (fontforge?)
> > 
> > This is, IMHO, the most difficult step. 
> 
> I remember peeking into font files that looked exactly like the stuff in
> those font*.c files but I can't remember where.
> 
> Surely this particular encoding must correspond to a specific dot
> something format?
> 
> That's why I mentioned "fontforge" hoping there had to be some tool that
> lets you convert from the .psf format to this other format painlessly.
> 
> If not, then it _is_ going to be difficult ..


from font_6x11.c:

#include <linux/font.h>

#define FONTDATAMAX (11*256)

static const unsigned char fontdata_6x11[FONTDATAMAX] = {

<snip>

	/* 35 0x23 '#' */
	0x00, /* 00000000 */
	0x00, /* 00000000 */
	0x28, /* 00 0 000 */
	0x7c, /* 0     00 */
	0x28, /* 00 0 000 */
	0x28, /* 00 0 000 */
	0x7c, /* 0     00 */
	0x28, /* 00 0 000 */
	0x00, /* 00000000 */
	0x00, /* 00000000 */
	0x00, /* 00000000 */

There is no real format, it's just an array full of bits on or off.

> For just this one font I'm not sure a tool that automates the process
> is worth the effort. 
> 
It might be worth it if it can translate from whatever text output
you're starting from to the hexadecimal values you need. I remember I
wrote the comments by hand, then translated that to the hex values I
needed in awk.

Good luck,
Jurriaan

Re: [Linux-fbdev-users] Setting colors on atyfb framebuffer console.

[Geert U. <ge...@li...>]

On Mon, 22 Oct 2007, cga2000 wrote:
> 3 files x 16 entries define a 16-color palette with the three rgb
> channels in separate files.
> 
> So that if the first entry is 0 in all three files, color0 will be
> [rgb:0,0,0] (black) .. if it is 127 in all three files it will be
> [rgb:127,127,127] (grey50..?) etc.
> 
> If the second entry is 255 in the default_red file and 0 in both
> default_green and default_blue .. color1 will be pure red .. 
> 
> And naturally if I want color15 to be pure white the last entry in all
> three files needs to be set to 255.
> 
> Right?

Right!

> In other words the three files represent the three rgb channels and the
> 16 entries the different red, green, blue values for each of the 16
> colors the console is capable of displaying simultaneously.
> 
> But then why does the bash shell display my input commands in light grey
> rather than white?

Normal text is light grey.

> Does this mean that I have 8 x 2 colors and that some kind of a bold
> attribute being on of off as the application decides .. causes only text
> that is highlighted to be displayed in white while "normal" text is
> displayed in grey?

Yes, white is used for highlighted text.

> I don't suppose anyone has come up with a little utility that loops on
> the 3x16 entries and issues escape sequences to display a semi-graphical
> representation of the current color palette? Something that would
> display 8x2 squares visualizing the 16 colors.
> 
> This would make it a lot easier to tell at a glance what the current
> palette looks like .. although I'm sure that with a bit of practice I
> will be able to see it just by glancing at the default_{red|grn|blu}
> files.

I attached such a script I wrote many years ago (last modified in 2004).

However, it no longer displays the highlighted colors correctly on the
console or in a gnome-terminal?

It still works fine in a good old xterm.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ge...@li...

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds

[Linux-fbdev-users] nvidiafb dual cards dual headed : how to ?

[Pierre-Yves P. <py....@wa...>]

 

 

Hi,

 

I'm trying to run dual PCIE cards framebuffered. The aim of is to get 1 VGA
display for console and a 2nd one for the VGA/LCD of my HTPC case.

 

/dev/fb0 exists, but what about /dev/fb1 ? I look for docs nearly everywhere
but could no get infos on it.

 

Here what dmesg gives :

 

nvidiafb_probe START

nvidiafb: Device ID: 10de01df

nvidia #0: Test OK

nvidia #1: Test OK

nvidia #2: bus seems to be busy

nvidiafb 0000:03:00.0: Failed to register I2C bus nvidia #2.

nvidiafb: CRTC0 analog found

nvidiafb: CRTC1 analog not found

nvidiafb: EDID found from BUS1

nvidiafb: CRTC 0 appears to have a CRT attached

nvidiafb: Using CRT on CRTC 0

nvidiafb: MTRR set to ON

nvidia_set_fbinfo START

nvidia_set_fbinfo END

nvidiafb_check_var START

nvidiafb_check_var END

nvidia_save_vga START

nvidia_save_vga END

nvidiafb: PCI nVidia NV1d framebuffer (64MB @ 0xB0000000)

nvidiafb_probe END

nvidiafb_probe START

nvidiafb: Device ID: 10de016a

nvidia #0: Test OK

nvidia #1: Test OK

nvidia #2: SDA stuck low!

nvidiafb 0000:06:00.0: Failed to register I2C bus nvidia #2.

nvidiafb: CRTC0 analog found

nvidiafb: CRTC1 analog found

nvidiafb: EDID found from BUS1

nvidiafb: CRTC 0 appears to have a CRT attached

nvidiafb: Using CRT on CRTC 0

nvidiafb: cannot ioremap FB base

 

 

 

Any help will be really appreciated. I'm loosing my hair .

 

Regards

 

Pierre

 

 

 

-------------------------------------------------
Guess What, THE ...
-------------------------------------------------

 

REAL PORTION of Microsoft Windows code:

 

while (memory_available) {
 eat_major_portion_of_memory (no_real_reason);
 if (feel_like_it)
 make_user_THINK (this_is_an_OS);
 gates_bank_balance++;
;
while (memory_microsofted) {
 system_response;
 cout blow_this_OS_away;
 cout install linux++;

 

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

 

[Linux-fbdev-users] nvidiafb dual cards dual headed : how to ?

[Pierre-Yves P. <py....@wa...>]

Hi,

 

I'm trying to run dual PCIE cards framebuffered. The aim of is to get 1 VGA
display for console and a 2nd one for the VGA/LCD of my HTPC case.

 

/dev/fb0 exists, but what about /dev/fb1 ? I look for docs nearly everywhere
but could no get infos on it.

 

Here what dmesg gives :

 

nvidiafb_probe START

nvidiafb: Device ID: 10de01df

nvidia #0: Test OK

nvidia #1: Test OK

nvidia #2: bus seems to be busy

nvidiafb 0000:03:00.0: Failed to register I2C bus nvidia #2.

nvidiafb: CRTC0 analog found

nvidiafb: CRTC1 analog not found

nvidiafb: EDID found from BUS1

nvidiafb: CRTC 0 appears to have a CRT attached

nvidiafb: Using CRT on CRTC 0

nvidiafb: MTRR set to ON

nvidia_set_fbinfo START

nvidia_set_fbinfo END

nvidiafb_check_var START

nvidiafb_check_var END

nvidia_save_vga START

nvidia_save_vga END

nvidiafb: PCI nVidia NV1d framebuffer (64MB @ 0xB0000000)

nvidiafb_probe END

nvidiafb_probe START

nvidiafb: Device ID: 10de016a

nvidia #0: Test OK

nvidia #1: Test OK

nvidia #2: SDA stuck low!

nvidiafb 0000:06:00.0: Failed to register I2C bus nvidia #2.

nvidiafb: CRTC0 analog found

nvidiafb: CRTC1 analog found

nvidiafb: EDID found from BUS1

nvidiafb: CRTC 0 appears to have a CRT attached

nvidiafb: Using CRT on CRTC 0

nvidiafb: cannot ioremap FB base

 

 

 

Any help will be really appreciated. I'm loosing my hair .

 

Regards

 

Pierre

 

 

 

-------------------------------------------------
Guess What, THE ...
-------------------------------------------------

 

REAL PORTION of Microsoft Windows code:

 

while (memory_available) {
 eat_major_portion_of_memory (no_real_reason);
 if (feel_like_it)
 make_user_THINK (this_is_an_OS);
 gates_bank_balance++;
;
while (memory_microsofted) {
 system_response;
 cout blow_this_OS_away;
 cout install linux++;

 

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

 

Re: [Linux-fbdev-users] Setting colors on atyfb framebuffer console.

[cga2000 <cg...@op...>]

On Mon, Oct 22, 2007 at 10:20:35AM EDT, Antonino A. Daplas wrote:
> On Sun, 2007-10-21 at 20:30 -0400, cga2000 wrote:
> > On Sun, Oct 21, 2007 at 07:31:31PM EDT, Antonino A. Daplas wrote:
> > > On Sun, 2007-10-21 at 08:34 -0400, cga2000 wrote:
> > > > > On Sat, Oct 20, 2007 at 12:06:31PM EDT, Antonino A. Daplas wrote:
> > > > > 

> > > The 16 entries is standard since the linux console is limited to
> > > 16 colors. As to why separate each component (R, G, B) into 3
> > > files, it's because that is the simplest way, otherwise, we have
> > > to add additional code to separate each color into its components
> > > before feeding them to the console layer.

> > OK.
> > 
> > So we have three files for our three fundamental colors.

> > You have probably realized I'm a bit thick by now .. but why 16
> > entries to each file - iow, I would have expected one for the
> > background, one for the foreground .. etc.

> > I'm still a bit lost.
> 
> Do you mean why not one file per attribute? ... ie,
> default_foreground, default_background, default_underline, etc? It is
> simply more complicated since we have to take into consideration how a
> console driver represents an attribute, ie, one driver may use blue
> for underline, another may use green, and in another ignored.

> Color though is treated the same whatever driver you use.

OK .. I think I get it.

3 files x 16 entries define a 16-color palette with the three rgb
channels in separate files.

So that if the first entry is 0 in all three files, color0 will be
[rgb:0,0,0] (black) .. if it is 127 in all three files it will be
[rgb:127,127,127] (grey50..?) etc.

If the second entry is 255 in the default_red file and 0 in both
default_green and default_blue .. color1 will be pure red .. 

And naturally if I want color15 to be pure white the last entry in all
three files needs to be set to 255.

Right?

In other words the three files represent the three rgb channels and the
16 entries the different red, green, blue values for each of the 16
colors the console is capable of displaying simultaneously.

But then why does the bash shell display my input commands in light grey
rather than white?

Does this mean that I have 8 x 2 colors and that some kind of a bold
attribute being on of off as the application decides .. causes only text
that is highlighted to be displayed in white while "normal" text is
displayed in grey?

I don't suppose anyone has come up with a little utility that loops on
the 3x16 entries and issues escape sequences to display a semi-graphical
representation of the current color palette? Something that would
display 8x2 squares visualizing the 16 colors.

This would make it a lot easier to tell at a glance what the current
palette looks like .. although I'm sure that with a bit of practice I
will be able to see it just by glancing at the default_{red|grn|blu}
files.

:-)

Thanks,
cga

Re: [Linux-fbdev-users] Switching to my chosen console font when the framebuffer console starts.

[cga2000 <cg...@op...>]

On Mon, Oct 22, 2007 at 10:27:51AM EDT, Antonino A. Daplas wrote:
> On Sun, 2007-10-21 at 20:44 -0400, cga2000 wrote:
> > On Sun, Oct 21, 2007 at 02:41:38AM EDT, Antonino A. Daplas wrote:
> > > On Sat, 2007-10-20 at 22:10 -0400, cga2000 wrote:
> 
> > Another thing I should consider is (sacrilege!) getting rid of the
> > penguin boot logo.
> > 
> 
> It is sacrilege :-) But you can either compile it out, or if you are
> willing to try Linus's git kernel, we have a patch to do just that,
> disable the boot logo using the option:
> 
> logo.nologo

I think I'll do that.

Five years I have revered the idol and I feel it's time to move on.

Re: [Linux-fbdev-users] Switching to my chosen console font when the framebuffer console starts.

[cga2000 <cg...@op...>]

On Mon, Oct 22, 2007 at 03:20:26AM EDT, Geert Uytterhoeven wrote:
> On Sun, 21 Oct 2007, cga2000 wrote:

[..]

> > It looks like I need to:
> > 
> > 1. convert the .psf font to a text format (fontforge?)
> > 2. wrap it up in a .c file -- same as the other font*.c's
> > 3. signal this new font to "make menuconfig" so it gets compiled in
> > 4. make it the default startup font
> > 
> > Am I correct?
> 
> Yes.

Nice to know I still guess right most of the time.

Devil's in the details, though.

> > Item (4) seems to be the tricky part since I have no recollection of
> > every being asked what default font I would prefer when I configure a
> > kernel.

> The kernel has some heuristics to find a good font.  There's also
> `fbcon=font:...' (cfr. Documentation/fb/fbcon.txt).  And if you build
> in only one font, it'll be chosen for sure ;-)

I will look into this ASAP.

As an aside it's interesting to note that while googling I found
practically no information regarding switching fonts at boot time (a
feature that looks at least marginally useful) and yet at the same time
and though I was not looking for any information on the subject, I ran
into many rather well-written and well-documented articles, mini-howtos,
etc. that explained how you could set up your kernel to display a
"bootsplash" screen instead of your boot messages.

Thanks, 
cga

Re: [Linux-fbdev-users] Switching to my chosen console font when the framebuffer console starts.

[cga2000 <cg...@op...>]

On Mon, Oct 22, 2007 at 03:20:54PM EDT, thu...@xs... wrote:
> From: cga2000 <cg...@op...>
> Date: Sun, Oct 21, 2007 at 08:44:29PM -0400
> > It looks like I need to:
> > 
> > 1. convert the .psf font to a text format (fontforge?)
> 
> This is, IMHO, the most difficult step. 

I remember peeking into font files that looked exactly like the stuff in
those font*.c files but I can't remember where.

Surely this particular encoding must correspond to a specific dot
something format?

That's why I mentioned "fontforge" hoping there had to be some tool that
lets you convert from the .psf format to this other format painlessly.

If not, then it _is_ going to be difficult ..

:-(

> > 2. wrap it up in a .c file -- same as the other font*.c's
> 
> that is a task awk, or perl, or python (etc) should be handy for.

For just this one font I'm not sure a tool that automates the process
is worth the effort. 

Or I am missing something?

> > 3. signal this new font to "make menuconfig" so it gets compiled in
> > 4. make it the default startup font
> 
> if you select it as the 'only' font in step 3, it's automatically the
> default font.

I should have written "make kernel configuration tools aware of this new
font so I can select it at make-menuconfig time"

I would imagine there's some kernel .config boolean that needs to be
created and initialized somehow?

Naturally as Geert also indicated the easiest way to enable the font is
definitely to make it the "only" font .. which is not really such a big
sacrifice .. It has just about all the glyphs I need , so I will not
need to switch to any other font on the fly.

> I did get some adapted fonts into the kernel (which remains my only
> claim to fame), and it really isn't that hard.

Thanks. 

Another good reason I should do it as soon as time allows, hopefully
within the next month or so.

cga.

[Linux-fbdev-users] s2ram with matroxfb, does it work ?

[Gilles H. <ha...@la...>]

Hello,

I am a owner of HP DL530 with a AGP matrox G550 video card.
I have tried all sorts of thing without success, when resuming the 
system freezes with a black screen.

When I stop the framebuffer with /sys/class/graphics/fb0/state before 
suspending, resume works fine. I see that because I use a serial 
console. Then, when I enable it again, the system locks up immediately.

does matroxfb support s2ram ?
I look at the code and found nothing related to suspend/resume stuff...

Thank you

Re: [Linux-fbdev-users] Switching to my chosen console font when the framebuffer console starts.

[<thu...@xs...>]

From: cga2000 <cg...@op...>
Date: Sun, Oct 21, 2007 at 08:44:29PM -0400
> It looks like I need to:
> 
> 1. convert the .psf font to a text format (fontforge?)

This is, IMHO, the most difficult step. 

> 2. wrap it up in a .c file -- same as the other font*.c's

that is a task awk, or perl, or python (etc) should be handy for.

> 3. signal this new font to "make menuconfig" so it gets compiled in
> 4. make it the default startup font

if you select it as the 'only' font in step 3, it's automatically the
default font.

I did get some adapted fonts into the kernel (which remains my only
claim to fame), and it really isn't that hard.

Good luck,
Jurriaan
-- 
Think for yourself and feel the walls
become sand beneath your feet
	Queensryche - Anybody Listening?
Debian (Unstable) GNU/Linux 2.6.23-rc8-mm2 2x2010 bogomips load 0.42
the Jack Vance Integral Edition: http://www.integralarchive.org

Re: [Linux-fbdev-users] Switching to my chosen console font when the framebuffer console starts.

[Antonino A. D. <ad...@gm...>]

On Sun, 2007-10-21 at 20:44 -0400, cga2000 wrote:
> On Sun, Oct 21, 2007 at 02:41:38AM EDT, Antonino A. Daplas wrote:
> > On Sat, 2007-10-20 at 22:10 -0400, cga2000 wrote:

> Another thing I should consider is (sacrilege!) getting rid of the
> penguin boot logo.
> 

It is sacrilege :-) But you can either compile it out, or if you are
willing to try Linus's git kernel, we have a patch to do just that,
disable the boot logo using the option:

logo.nologo

Tony

Re: [Linux-fbdev-users] Setting colors on atyfb framebuffer console.

[Antonino A. D. <ad...@gm...>]

On Sun, 2007-10-21 at 20:30 -0400, cga2000 wrote:
> On Sun, Oct 21, 2007 at 07:31:31PM EDT, Antonino A. Daplas wrote:
> > On Sun, 2007-10-21 at 08:34 -0400, cga2000 wrote:
> > > > On Sat, Oct 20, 2007 at 12:06:31PM EDT, Antonino A. Daplas wrote:
> > > > 
> > 
> > The 16 entries is standard since the linux console is limited to 16
> > colors. As to why separate each component (R, G, B) into 3 files, it's
> > because that is the simplest way, otherwise, we have to add additional
> > code to separate each color into its components before feeding them to
> > the console layer.
> 
> OK.
> 
> So we have three files for our three fundamental colors.
> 
> You have probably realized I'm a bit thick by now .. but why 16 entries
> to each file - iow, I would have expected one for the background, one
> for the foreground .. etc.
> 
> I'm still a bit lost.

Do you mean why not one file per attribute? ... ie, default_foreground,
default_background, default_underline, etc? It is simply more
complicated since we have to take into consideration how a console
driver represents an attribute, ie, one driver may use blue for
underline, another may use green, and in another ignored.

Color though is treated the same whatever driver you use.

Tony

Re: [Linux-fbdev-users] Switching to my chosen console font when the framebuffer console starts.

[Geert U. <ge...@li...>]

On Sun, 21 Oct 2007, cga2000 wrote:
> On Sun, Oct 21, 2007 at 02:41:38AM EDT, Antonino A. Daplas wrote:
> > On Sat, 2007-10-20 at 22:10 -0400, cga2000 wrote:
> > > I use the "Uni3-Terminus12.psf.gz" font on the console and this gives
> > > me 233x87 character cells on a 15" laptop screen.
> > > 
> > > When the boot process initially switches to the framebuffer console, I
> > > am presented with a somewhat larger font.
> > > 
> > > I later switch to my preferred font by having specified the following in
> > > /etc/console-tools/config:
> > > 
> > >     SCREEN_FONT=Uni3-Terminus12x6
> > > 
> > > I was wondering if there is a way I could switch to this font earlier
> > > in the boot process .. ideally as soon as the framebuffer console
> > > starts.
> > 
> > It's a user font, the earliest you can switch to it is when userspace
> > tools kick in.
> > 
> > Or you can port it as a kernel compiled-in font. See
> > drivers/video/console/font*.c
> 
> That would be cool.
> 
> It looks like I need to:
> 
> 1. convert the .psf font to a text format (fontforge?)
> 2. wrap it up in a .c file -- same as the other font*.c's
> 3. signal this new font to "make menuconfig" so it gets compiled in
> 4. make it the default startup font
> 
> Am I correct?

Yes.

> Item (4) seems to be the tricky part since I have no recollection of
> every being asked what default font I would prefer when I configure a
> kernel.

The kernel has some heuristics to find a good font.
There's also `fbcon=font:...' (cfr. Documentation/fb/fbcon.txt).
And if you build in only one font, it'll be chosen for sure ;-)

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ge...@li...

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds

Re: [Linux-fbdev-users] Switching to my chosen console font when the framebuffer console starts.

[cga2000 <cg...@op...>]

On Sun, Oct 21, 2007 at 02:41:38AM EDT, Antonino A. Daplas wrote:
> On Sat, 2007-10-20 at 22:10 -0400, cga2000 wrote:
> > I use the "Uni3-Terminus12.psf.gz" font on the console and this gives
> > me 233x87 character cells on a 15" laptop screen.
> > 
> > When the boot process initially switches to the framebuffer console, I
> > am presented with a somewhat larger font.
> > 
> > I later switch to my preferred font by having specified the following in
> > /etc/console-tools/config:
> > 
> >     SCREEN_FONT=Uni3-Terminus12x6
> > 
> > I was wondering if there is a way I could switch to this font earlier
> > in the boot process .. ideally as soon as the framebuffer console
> > starts.
> 
> It's a user font, the earliest you can switch to it is when userspace
> tools kick in.
> 
> Or you can port it as a kernel compiled-in font. See
> drivers/video/console/font*.c

That would be cool.

It looks like I need to:

1. convert the .psf font to a text format (fontforge?)
2. wrap it up in a .c file -- same as the other font*.c's
3. signal this new font to "make menuconfig" so it gets compiled in
4. make it the default startup font

Am I correct?

Item (4) seems to be the tricky part since I have no recollection of
every being asked what default font I would prefer when I configure a
kernel.

In any event this would be nice since would give me a bit more time to
read the boot messages .. and seeing them in the size my eyes are used
to would make the even faster.

Another thing I should consider is (sacrilege!) getting rid of the
penguin boot logo.

Sorry to bother you with all these trivialities. 

The only good thing about this is that the next maniac that comes along
will be able to google for the answers.

Thanks,
cga

Re: [Linux-fbdev-users] Setting colors on atyfb framebuffer console.

[cga2000 <cg...@op...>]

On Sun, Oct 21, 2007 at 07:31:31PM EDT, Antonino A. Daplas wrote:
> On Sun, 2007-10-21 at 08:34 -0400, cga2000 wrote:
> > > On Sat, Oct 20, 2007 at 12:06:31PM EDT, Antonino A. Daplas wrote:
> > > 
> > > > Each has 16 entries in decimal format standing for the red
> > > > (default_red), green (default_grn) and blue (default_blu) part of the
> > > > color. I'll be using hexadecimal from here on, since that is easier to
> > > > understand.
> > > > 
> > > > So the first entry (index 0) is 0x00 for red, 0x00 for green and 0x00
> > > > for blue. If you combine them as RGB, you get 0x000000 (black). Index 0
> > > > (which is black by default) is the default bacground color. Entry 8
> > > > (index 7), which is RGB 0xaaaaaa (light gray) is the default foreground
> > > > color.
> > > 
> > > Yes, why on earth did they decide on decimal entries?
> 
> It just happened :-). This is probably something you would handle with a
> script, so whether it's in decimal or hex probably does not matter.

Glad you got the implicit smiley.

I'm used to stuff like #808080 being half way or just about between
#000000 and #FFFFFF so I don't find decimal rgb representation very
legible.

[..]

> > > But why three files and 16 entries in each file?
> > > 
> 
> The 16 entries is standard since the linux console is limited to 16
> colors. As to why separate each component (R, G, B) into 3 files, it's
> because that is the simplest way, otherwise, we have to add additional
> code to separate each color into its components before feeding them to
> the console layer.

OK.

So we have three files for our three fundamental colors.

You have probably realized I'm a bit thick by now .. but why 16 entries
to each file - iow, I would have expected one for the background, one
for the foreground .. etc.

I'm still a bit lost.

> > > Is this documented in a man page or anything?
> > >
> 
> Not in so many words. This is a new feature, and it is described in
> Documentation/kernel-parameters.txt.

Well I guess this thread will eventually have all the specs.  Pruning
it a bit, you'd have the a linux/Documentation/fbcon_colors.txt

I do appreciate your help with this.

Cheers,
cga

Re: [Linux-fbdev-users] Setting colors on atyfb framebuffer console.

[Antonino A. D. <ad...@gm...>]

On Sun, 2007-10-21 at 08:34 -0400, cga2000 wrote:
> > On Sat, Oct 20, 2007 at 12:06:31PM EDT, Antonino A. Daplas wrote:
> > 
> > > Each has 16 entries in decimal format standing for the red
> > > (default_red), green (default_grn) and blue (default_blu) part of the
> > > color. I'll be using hexadecimal from here on, since that is easier to
> > > understand.
> > > 
> > > So the first entry (index 0) is 0x00 for red, 0x00 for green and 0x00
> > > for blue. If you combine them as RGB, you get 0x000000 (black). Index 0
> > > (which is black by default) is the default bacground color. Entry 8
> > > (index 7), which is RGB 0xaaaaaa (light gray) is the default foreground
> > > color.
> > 
> > Yes, why on earth did they decide on decimal entries?

It just happened :-). This is probably something you would handle with a
script, so whether it's in decimal or hex probably does not matter.

> > 
> > > If you want to change the background color from black to blue:
> > > 
> > > 1. cat /sys/module/vt/parameters/default_blu > blue.txt
> > > 2. change the first entry from 0 to 255,
> > > 3. echo blue.txt > /sys/module/vt/parameters/default_blu
> > 
> > This should be:
> > 
> > # cp blue.txt /sys/module/vt/parameters/default_blu
> > 
> > .. right?
> >

Yes, my mistake, cp or cat.

>  
> > > 4. Type 'reset' to reset the terminal.
> > 
> > Yes, I now have a blue background.
> > 
> > > 6. You should not have a light gray text on a blue background.
> > 
> > But why three files and 16 entries in each file?
> > 

The 16 entries is standard since the linux console is limited to 16
colors. As to why separate each component (R, G, B) into 3 files, it's
because that is the simplest way, otherwise, we have to add additional
code to separate each color into its components before feeding them to
the console layer.

> > Is this documented in a man page or anything?
> >

Not in so many words. This is a new feature, and it is described in
Documentation/kernel-parameters.txt.

Tony

Re: [Linux-fbdev-users] Setting colors on atyfb framebuffer console.

[cga2000 <cg...@op...>]

> On Sat, Oct 20, 2007 at 12:06:31PM EDT, Antonino A. Daplas wrote:
> 
> > Each has 16 entries in decimal format standing for the red
> > (default_red), green (default_grn) and blue (default_blu) part of the
> > color. I'll be using hexadecimal from here on, since that is easier to
> > understand.
> > 
> > So the first entry (index 0) is 0x00 for red, 0x00 for green and 0x00
> > for blue. If you combine them as RGB, you get 0x000000 (black). Index 0
> > (which is black by default) is the default bacground color. Entry 8
> > (index 7), which is RGB 0xaaaaaa (light gray) is the default foreground
> > color.
> 
> Yes, why on earth did they decide on decimal entries?
> 
> > If you want to change the background color from black to blue:
> > 
> > 1. cat /sys/module/vt/parameters/default_blu > blue.txt
> > 2. change the first entry from 0 to 255,
> > 3. echo blue.txt > /sys/module/vt/parameters/default_blu
> 
> This should be:
> 
> # cp blue.txt /sys/module/vt/parameters/default_blu
> 
> .. right?
> 
> > 4. Type 'reset' to reset the terminal.
> 
> Yes, I now have a blue background.
> 
> > 6. You should not have a light gray text on a blue background.
> 
> But why three files and 16 entries in each file?
> 
> Is this documented in a man page or anything?
> 
> Thanks,
> cga

Re: [Linux-fbdev-users] Switching to my chosen console font when fb console starts.

[Antonino A. D. <ad...@gm...>]

On Sat, 2007-10-20 at 22:10 -0400, cga2000 wrote:
> I use the "Uni3-Terminus12.psf.gz" font on the console and this gives
> me 233x87 character cells on a 15" laptop screen.
> 
> When the boot process initially switches to the framebuffer console, I
> am presented with a somewhat larger font.
> 
> I later switch to my preferred font by having specified the following in
> /etc/console-tools/config:
> 
>     SCREEN_FONT=Uni3-Terminus12x6
> 
> I was wondering if there is a way I could switch to this font earlier
> in the boot process .. ideally as soon as the framebuffer console
> starts.

It's a user font, the earliest you can switch to it is when userspace
tools kick in.

Or you can port it as a kernel compiled-in font. See
drivers/video/console/font*.c

Tony

[Linux-fbdev-users] Switching to my chosen console font when fb console starts.

[cga2000 <cg...@op...>]

I use the "Uni3-Terminus12.psf.gz" font on the console and this gives
me 233x87 character cells on a 15" laptop screen.

When the boot process initially switches to the framebuffer console, I
am presented with a somewhat larger font.

I later switch to my preferred font by having specified the following in
/etc/console-tools/config:

    SCREEN_FONT=Uni3-Terminus12x6

I was wondering if there is a way I could switch to this font earlier
in the boot process .. ideally as soon as the framebuffer console
starts.

Thanks,
cga