David Dawes wrote:
On Tue, Dec 17, 2002 at 01:02:45AM +0000, Keith Whitwell wrote:
David Dawes wrote:
On Mon, Dec 16, 2002 at 12:44:15PM -0500, David Dawes wrote:
On Sat, Dec 14, 2002 at 07:02:50PM +0100, Michel Dänzer wrote:
On Sam, 2002-12-14 at 18:42, Michel Daenzer wrote:
CVSROOT: /cvsroot/dri
Module name: xc
Repository: xc/xc/programs/Xserver/hw/xfree86/os-support/shared/drm/kernel/
Changes by: mdaenzer@sc8-pr-cvs1. 02/12/14 09:42:48
Log message:
merge changes from trunk since Mesa 5.0 merge
Okay, I think that should fix the most pressing issues people think were
fixed by the Mesa 5.0 merge.
Thanks for doing that!
I've taken that and merged it into the XFree86 trunk. I've also merged
the XFree86 trunk back into mesa-4-0-4-branch, but I'm not having much
success in committing it. My first commit attempt aborted, then I
couldn't connect to the CVS repository for a while, then trying again
just now it's failing because there are various locks in the repository
from the first aborted attempt.
Anyway, the pre-merge tag is "mesa-4-0-4-20021214" and when I can finally
finish committing the merge, the post-merge tag will be
"mesa-4-0-4-20021215".
The SourceForge admins cleared the locks, and I've finshed committing
the merge. The part of this merge that I'm least sure of is the 2D
radeon driver.
It'd be a big help if someone could take a look at the merged code and
let me know of any problems. I've put a diff between those before and
after tags at <http://www.xfree86.org/~dawes/mesa404-20021214-15.diff.gz>.
First thing I notice is somebody's disabled dri on radeon 9000's (rv250's).
Added a comment indicating that they didn't know if they were doing the right
thing or not, which is kindof dump -- if you don't know, don't make the
change. Or ask here, I guess.
That looks like a bad merge on my part. There are lots of conflicts
between the two versions, and the XFree86 version of this check got
through when it shouldn't have. I'll fix it.
OK, no worries.
Also the option to turn off the depth buffer (introduced in the twc work) has
been either knobbled or removed. This may not have been deliberate, hard to
tell. This seems to be the second round of removing this -- or alternatly
maybe it was only half merged to xfree in the first place. In any case it
seems to be gone now.
I don't see an option like that on the DRI trunk either. There's a
"NoBackBuffer" option though.
Yes, the old brain wasn't working too well last night. What I was concerned
about was actually Michel's (I think) fix for allocating a depth buffer sized
to an integral number of tiles (we had a similar problem on the i830).
Anyway, it looks like you reinstated this fix too.
I don't know the story behind the removal of RADEONSaveFBDevRegisters() --
seems to be related to keeping pageflipping working over an fbdev call. Can
anyone comment on whether or not this is still necessary?
I expect that got accidentally removed. I'm also not sure if the handling
of gen_int_cntl is correct everywhere.
The attached patch should put back most of what got inadvertently removed.
Thanks, David, it looks good to me now.
Keith
-------------------------------------------------------
This sf.net email is sponsored by:
With Great Power, Comes Great Responsibility
Learn to use your power at OSDN's High Performance Computing Channel
http://hpc.devchannel.org/
_______________________________________________
Dri-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/dri-devel
