The attached patch allows vesafb to work on PPC. vesafb does surprisingly
little with the BIOS (which I thought was the whole point of VESA). With some
ifdef's I'm able to use it as sort of a VGA offb -- the firmware initializes
the card, and the driver just uses it and never mucks with it.
This patch allows me to get an 800x600x8 console on the Thinkpad 850, and
640x480x8 on the Carolina machines. More interestingly it allows X to work on
the Thinkpad, which currently doesn't work and has never worked in >16 colors.
It should also work on the 7020 (Sven?), which uses the Weitek 9100 card and
currently is driverless.
This patch is only for PReP, which is probably the only place it's really
useful (since CHRP and PMac have offb), but would be easy enough to add to.
Comments?
-Hollis
From: Michel Lanners <hidden> Date: 2002-01-24 04:42:46
On 23 Jan, this message from hollis@austin.ibm.com echoed through cyberspace:
The attached patch allows vesafb to work on PPC. vesafb does surprisingly
little with the BIOS (which I thought was the whole point of VESA). With some
ifdef's I'm able to use it as sort of a VGA offb -- the firmware initializes
the card, and the driver just uses it and never mucks with it.
Wouldn't this be a good start to use generic PC-type grafic cards on PPC
machines in general? Or do you lack initialization if nobody (neither
firmware nor OS) can talk to the card 'natively'?
Cheers
Michel
-------------------------------------------------------------------------
Michel Lanners | " Read Philosophy. Study Art.
23, Rue Paul Henkes | Ask Questions. Make Mistakes.
L-1710 Luxembourg |
email mlan@cpu.lu |
http://www.cpu.lu/~mlan | Learn Always. "
** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/
The attached patch allows vesafb to work on PPC. vesafb does surprisingly
Nice!
little with the BIOS (which I thought was the whole point of VESA). With some
ifdef's I'm able to use it as sort of a VGA offb -- the firmware initializes
the card, and the driver just uses it and never mucks with it.
Well, vesafb assumes the VESA BIOS initialized the card to some linear graphics
mode _before_ the real kernel code was started. Once the kernel is started, no
VESA calls are made, and the video mode is fixed.
So in theory (and you proved the practical side on PReP) it can be used on any
platform that initializes a more-or-less VGA card to a linear graphics mode.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
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
** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/
On Wed, Jan 23, 2002 at 01:20:10PM -0600, hollis@austin.ibm.com wrote:
The attached patch allows vesafb to work on PPC. vesafb does surprisingly
little with the BIOS (which I thought was the whole point of VESA). With some
ifdef's I'm able to use it as sort of a VGA offb -- the firmware initializes
the card, and the driver just uses it and never mucks with it.
This looks good to me. My only comment is you should probably use
CONFIG_MTRR around the mtrr bits instead of __i386__ (which is what
other drivers do). Geert, since you're still overall fb maintainer
(yes?) would you be willing to sign off on these changes to be sent to
Marcelo for 2.4.19-pre1?
--
Tom Rini (TR1265)
http://gate.crashing.org/~trini/
** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/
On Thu, Jan 24, 2002 at 05:42:46AM +0100, Michel Lanners wrote:
On 23 Jan, this message from hollis@austin.ibm.com echoed through cyberspace:
quoted
The attached patch allows vesafb to work on PPC. vesafb does surprisingly
little with the BIOS (which I thought was the whole point of VESA). With
some ifdef's I'm able to use it as sort of a VGA offb -- the firmware
initializes the card, and the driver just uses it and never mucks with it.
Wouldn't this be a good start to use generic PC-type grafic cards on PPC
machines in general? Or do you lack initialization if nobody (neither
firmware nor OS) can talk to the card 'natively'?
Yes, we would lack initialization on Power Mac (for example). PReP firmware
does proper setup for a very limited range of chips, which is why this is
useful at all.
-Hollis
** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/
On Wed, Jan 23, 2002 at 01:20:10PM -0600, hollis@austin.ibm.com wrote:
quoted
The attached patch allows vesafb to work on PPC. vesafb does surprisingly
little with the BIOS (which I thought was the whole point of VESA). With some
ifdef's I'm able to use it as sort of a VGA offb -- the firmware initializes
the card, and the driver just uses it and never mucks with it.
This looks good to me. My only comment is you should probably use
CONFIG_MTRR around the mtrr bits instead of __i386__ (which is what
Indeed.
other drivers do). Geert, since you're still overall fb maintainer
What's in a name...
(yes?) would you be willing to sign off on these changes to be sent to
Marcelo for 2.4.19-pre1?
Just send them to Marcelo, CC me, and say it's OK for me to include them.
Or better, ask Gerd Knorr to send them in (please CC me)?
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
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
** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/
From: Gabriel Paubert <hidden> Date: 2002-01-25 09:28:12
On Thu, 24 Jan 2002 hollis@austin.ibm.com wrote:
On Thu, Jan 24, 2002 at 05:42:46AM +0100, Michel Lanners wrote:
quoted
On 23 Jan, this message from hollis@austin.ibm.com echoed through cyberspace:
quoted
The attached patch allows vesafb to work on PPC. vesafb does surprisingly
little with the BIOS (which I thought was the whole point of VESA). With
some ifdef's I'm able to use it as sort of a VGA offb -- the firmware
initializes the card, and the driver just uses it and never mucks with it.
Wouldn't this be a good start to use generic PC-type grafic cards on PPC
machines in general? Or do you lack initialization if nobody (neither
firmware nor OS) can talk to the card 'natively'?
Yes, we would lack initialization on Power Mac (for example). PReP firmware
does proper setup for a very limited range of chips, which is why this is
useful at all.
I remind you that I am using since 3 years or so an x86 emulator which
should allow to initialize any board with has an x86 ROM BIOS, provided
the host bridge can be configured to generate VGA memory accesses
(0xa0000-0xbffff), so I think most MACs are out. It is in my bootloader
but I have no time to work on it right now, some people have expressed
interest in including it in the official source tree but I don't believe
it will happen soon.
Regards,
Gabriel.
** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/
From: Leigh Brown <hidden> Date: 2002-01-25 13:59:28
Hi Hollis,
The attached patch allows vesafb to work on PPC. vesafb does
surprisingly little with the BIOS (which I thought was the whole point
of VESA). With some ifdef's I'm able to use it as sort of a VGA offb --
the firmware initializes the card, and the driver just uses it and
never mucks with it.
This patch allows me to get an 800x600x8 console on the Thinkpad 850,
and 640x480x8 on the Carolina machines. More interestingly it allows X
to work on the Thinkpad, which currently doesn't work and has never
worked in >16 colors. It should also work on the 7020 (Sven?), which
uses the Weitek 9100 card and currently is driverless.
This patch is only for PReP, which is probably the only place it's
really useful (since CHRP and PMac have offb), but would be easy enough
to add to.
Comments?
Well, I thought I'd give it a go on my 7043-140. I think that it
would help those people with the proprietary IBM graphics cards
(GXT550P, GXT800P that sort of thing, IIRC), as well as being at least
as good as the S3Triofb frame buffer driver (which can't switch video
modes any way).
Unsurprisingly it didn't work. That is, unsurprisingly because nothing
ever goes to plan on these boxes :-)
It seems there is no residual data entry for the graphics card. I
enabled the code to dump the residual data and it was conspicuous by
its absence. In fact, everything else seemed to be there (6 PCI devices
and 17 ISA devices).
Any ideas?
Cheers,
Leigh.
** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/