Luc Verhaegen.
Luc Verhaegen wrote:
sb/via/k8t890: add vga textmode code for k8m890 chrome igp.
Add initialisation for the VIA Chrome 9 IGP on the k8m890 through native code and through the general vga infrastructure i committed a month or two ago. Add videoram_size option for k8m890 and the Asus M2V-MX SE.
Now the Asus M2V-MX SE will magically come up with a working standard VGA 80x25 textmode.
Many thanks to the people who worked hard on the Asus M2V-MX SE, and all of its components; this vga bringup was a breeze thanks to your hard work for this excellently supported board. And separate thanks to Rudolf Marek for spurring me on and for providing a register dump.
Signed-off-by: Luc Verhaegen libv@skynet.be
Acked-by: Peter Stuge peter@stuge.se
On Thu, Jul 23, 2009 at 03:18:49PM +0200, Peter Stuge wrote:
Luc Verhaegen wrote:
sb/via/k8t890: add vga textmode code for k8m890 chrome igp.
Add initialisation for the VIA Chrome 9 IGP on the k8m890 through native code and through the general vga infrastructure i committed a month or two ago. Add videoram_size option for k8m890 and the Asus M2V-MX SE.
Now the Asus M2V-MX SE will magically come up with a working standard VGA 80x25 textmode.
Many thanks to the people who worked hard on the Asus M2V-MX SE, and all of its components; this vga bringup was a breeze thanks to your hard work for this excellently supported board. And separate thanks to Rudolf Marek for spurring me on and for providing a register dump.
Signed-off-by: Luc Verhaegen libv@skynet.be
Acked-by: Peter Stuge peter@stuge.se
Thanks: -> r4465
Luc Verhaegen.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi Luc,
looks fine. Please check the vga_init.c as we discussed on IRC. Also, maybe you can develop simple patch that will write the memsize and speedgrade to scratch registers.
Thanks,
Rudolf
On Thu, Jul 23, 2009 at 05:29:01PM +0200, Rudolf Marek wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi Luc,
looks fine. Please check the vga_init.c as we discussed on IRC.
For that, we need to figure out what these two registers really do depend on and what their effect is.
Also, maybe you can develop simple patch that will write the memsize and speedgrade to scratch registers.
Tell you what; i'll go off for two days, and develop the solid PLL calculation routine needed to bring up the k8m890 in my unichrome driver.
Luc Verhaegen.
On Thu, Jul 23, 2009 at 02:20:53PM +0200, Luc Verhaegen wrote:
Add initialisation for the VIA Chrome 9 IGP on the k8m890 through native code and through the general vga infrastructure i committed a month or two ago. Add videoram_size option for k8m890 and the Asus M2V-MX SE.
Now the Asus M2V-MX SE will magically come up with a working standard VGA 80x25 textmode.
Hi Luc,
With your code, will the SeaBIOS vgabios (make out/vgabios.bin) run on the board?
-Kevin
On Sun, Jul 26, 2009 at 02:21:34PM -0400, Kevin O'Connor wrote:
On Thu, Jul 23, 2009 at 02:20:53PM +0200, Luc Verhaegen wrote:
Add initialisation for the VIA Chrome 9 IGP on the k8m890 through native code and through the general vga infrastructure i committed a month or two ago. Add videoram_size option for k8m890 and the Asus M2V-MX SE.
Now the Asus M2V-MX SE will magically come up with a working standard VGA 80x25 textmode.
Hi Luc,
With your code, will the SeaBIOS vgabios (make out/vgabios.bin) run on the board?
-Kevin
No; a general VGA bios is a broken concept to begin with.
What you should know is that VGA is _not_ a standard, and one should never, ever try to change the mode without more detailed hw knowledge. It is only a standard through the int10 and vbe interfaces, where the manufacturers vga bios does all the vga and hw specific stuff for you.
What such a general bios should do is find out, one way or another, what mode has been set and how to stuff things into the framebuffer.
In this case: mode 3h: 80x25 with 8x16 fonts, with the framebuffer living at 0xB8000 and with vga standard cursor and fb offset handling.
So a general int10 bios should find this out and then claim to only support this and absolutely nothing else.
Everything else requires a whole bunch of hw specific code, and while you could spend the rest of your days porting graphics drivers from fb or X to this code, i think there are much better ways to spend ones time.
Luc Verhaegen.
On Mon, Jul 27, 2009 at 04:06:45PM +0200, Luc Verhaegen wrote:
No; a general VGA bios is a broken concept to begin with.
What you should know is that VGA is _not_ a standard, and one should never, ever try to change the mode without more detailed hw knowledge. It is only a standard through the int10 and vbe interfaces, where the manufacturers vga bios does all the vga and hw specific stuff for you.
What such a general bios should do is find out, one way or another, what mode has been set and how to stuff things into the framebuffer.
In this case: mode 3h: 80x25 with 8x16 fonts, with the framebuffer living at 0xB8000 and with vga standard cursor and fb offset handling.
So a general int10 bios should find this out and then claim to only support this and absolutely nothing else.
Everything else requires a whole bunch of hw specific code, and while you could spend the rest of your days porting graphics drivers from fb or X to this code, i think there are much better ways to spend ones time.
Luc Verhaegen.
Let me further quantise this:
It would be great if there was int10 infrastructure which would do just the above. * Seabios can show us its menu. * Grub will then give us its menu too. * The linux kernel will no longer think that it is talking to CGA and will use the correct fontsize while setting the cursor (so that we no longer get a floating cursor).
Luc Verhaegen.
-
Kevin O'Connor -
Luc Verhaegen -
Peter Stuge -
Rudolf Marek