RE: 2.6.19-rc4-mm2

Hi, Andrew -

I messed around with trying to recreate the workaround for the 24bpp
Radeon stuff for chips that have the problem (like yours does) all
afternoon yesterday, but I never did get it to work. I can't spend any
more time on this right now (boss is after me!) but I will try to do
some more poking around later or next week. Do you think I should submit
a patch that at least enables 24bpp for the chip I have where it
definitely does work? I'm sure that is overly restrictive, but I don't
know which ones work and which are broken at this point. I tried going
to the ATI web site (we have an OEM agreement with them, so I can get
more data that way) but their documents are no better now that they are
part of AMD than they ever were, and I didn't find anything useful.
Grumble! Actually, as more people use linux and Xfree86, they may have
to do a bit better with the specs. One can hope!

If you get a chance before I do, you might look at what Xfree86 does -
they might have the workaround.

/Charlotte

PS - You have your reply-to email address set to
ajwade@alumni.uwaterloo.ca
which evidently is stale, 'cause replying to that bounces.

-----Original Message-----
From: Andrew Wade [mailto:andrew.j.wade@gmail.com]
Sent: Tuesday, November 07, 2006 10:58 AM
To: Richardson, Charlotte
Cc: Andrew Morton; linux-kernel@vger.kernel.org; Kimball Murray;

linux-

fbdev-devel@lists.sourceforge.net
Subject: Re: 2.6.19-rc4-mm2

Hello Charlotte,

quoted

(Sorry, I don't know what timezone you're in, but I went home,

cooked

quoted

supper, ate supper, did two loads of laundry, slept for about seven
hours, ate breakfast, did another load of laundry, and voted, and

now

quoted

I'm back!)

I'm EST (GMT-5). But the hours I'm online are somewhat erratic.

...

quoted

If I can't repro it with this chip, if you want to mess around with

it

quoted

on yours, here's what I think we had to do... I believe the trick

was

quoted

to use 16bpp mode as far as what mode you write to the chip, and

then

quoted

double all the x coordinate values for things like offset, width,

and

quoted

pitch. You would have to do that to the accelerated routines also.

I'd be happy to mess around with the driver, but I won't have much
time to do so until tomorrow. I'll let you know if I find anything,
and of course I'll be happy to test patches.

-ajw

Hello Charlotte,

     I have just noticed that the garbled X-server mode is different
than what I thought it was. rinfo->depth is 24, but rinfo->bpp is 32.
(measured in radeon_write_mode). I haven't yet thought through the
implications of this. I've been unable to get the display correct, but
I've been trying to fix the wrong mode.
     The console mode is ->depth==8, ->bpp==8, and presumably the same
when it's garbled, but I've been unable to reproduce the problem since
instrumenting the kernel.

On Wednesday 08 November 2006 09:01, Richardson, Charlotte wrote:

... Do you think I should submit
a patch that at least enables 24bpp for the chip I have where it
definitely does work? I'm sure that is overly restrictive, but I don't
know which ones work and which are broken at this point.

That sounds like a good idea to me, but I'm not very familiar with the
programming practices here.

...

If you get a chance before I do, you might look at what Xfree86 does -
they might have the workaround.

It looks like the X.org radeon driver doesn't support 24bpp.

-ajw