Re: [Dri-devel] Mach64 Success Report:

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 454-5900

I would love if this got merged in.  I don't see any reason people should have 
to apply a seperate patch to watch videos.  If the people working on this are 
gonna lag (no offense) you should take action.  DRI is ahead of the game  :-P

-James

On Thursday 17 April 2003 05:09 pm, Leif Delgass wrote:
> Attached is a version of your stipped down patch against
> mach64-0-0-6-branch that I cleaned up a bit.  I ran it through indent with
> the following options: '-nut -i4 -di4 -ip4 -cli4 -bl -bli0 -nbad -bls
> -psl'.  I then did a little manual cleanup on the function declarations
> (couldn't find options to get the right style).  I removed the BIOS stuff,
> which is unnecessary as you pointed out.  I also removed a few XVideo
> atoms that didn't seem necessary.  I'd like to do some more comparision
> between this and the r128 and radeon code, but it's looking pretty good.
> I've tested it and it's working fine.
>
> I'm actually quite tempted to commit this to the mach64 branch.  It's
> self-contained and could easily be removed later if Marc comes up with his
> own version.
>
> --Leif
>
> On Thu, 17 Apr 2003, James Jones wrote:
> > OK, I have a patch against xf86 CVS and an updated one against the dri. 
> > (Took out some BIOS fields in the ati private structure that weren't
> > used.  I haven't cleaned up atixv.c code yet.  Any advice on a quick way
> > to do so?  My girlfriend's already mad at me for staring at this stuff
> > for so long and I have homework to do.
> >
> > I also don't see any chance of anything in the XFree86 CVS compiling on
> > my machine right now.  I know they don't gurantee it to build, but I
> > figured I could at least get the ati drivers to build/install.  No luck
> > there, so I don't actually know if this works.
> >
> > -James
> >
> > On Wednesday 16 April 2003 10:54 am, Leif Delgass wrote:
> > > On Wed, 16 Apr 2003, James Jones wrote:
> > > > I couldn't agree more with the formatting comment.  I have no idea
> > > > what the XFree coding standards are, but this surely doesn't comply
> > > > with them. I simply copied the code nearly verbatim from your patch. 
> > > > My Vim tab settings and stuff might be the culprit.  ;-)
> > >
> > > I use emacs most of the time. :) Most of the ati* files uses spaces for
> > > indenting instead of tabs.  In the DRI driver, I try to remember to use
> > > the same spacing and formatting as Marc's code when editing files that
> > > are already part of the driver, since I've seen Marc commit
> > > whitespace-ony changes in the XFree86 CVS logs, iirc.  The new files
> > > for DRI support (atidri.c, etc.), most of which Gareth Hughes created,
> > > have a different style (closer to the r128/radeon driver code).  I
> > > hadn't made an attempt at reformatting most of the GATOS code in
> > > atixv.c, but I think it's needed before submitting a patch to XFree86
> > > to make it more readable and consistent with existing code.
> > >
> > > > I looked back over the patch, and it looks like there's only 1 change
> > > > that even applies to DRI at all, so it should apply very cleanly to
> > > > XFree86 CVS. I've never looked at the actual Xfree86 CVS trunk, so
> > > > I'm just speculating here.  I'll check it out and see what I can do.
> > >
> > > I think that's right, it should be pretty simple. The changes to
> > > atidri.c wouldn't be needed, obviously.  There are only a few minor
> > > changes to files other than atixv.[ch], which are nearly empty in
> > > XFree86 CVS.
> > >
> > > > -James
> > > >
> > > > On Wednesday 16 April 2003 10:05 am, Leif Delgass wrote:
> > > > > This is great, good work!  It looks like it could use a run through
> > > > > indent, but I'll definitely take a look at it and compare it to
> > > > > what I've got.  If we could do a few things -- make it apply
> > > > > cleanly to the XFree86 CVS trunk, clean up the formatting a bit
> > > > > (preferably to match the coding style of the existing XFree86
> > > > > driver code), and test it -- I think it would be good to attach it
> > > > > to XFree86 Bugzilla #139 (a better summary for this meta-bug would
> > > > > be something like "XVideo PutImage not supported in XFree86 mach64
> > > > > driver").  It might be a good idea to first post the proposed patch
> > > > > against XFree86 CVS to the gatos-devel list to get feedback, and
> > > > > then attach it to the bug.  I know Marc LaFrance is still working
> > > > > on the GATOS "merge"  (he mentioned it in a thread on the XFree86
> > > > > devel list), and this could be helpful for that work.
> > > > >
> > > > > --Leif
> > > > >
> > > > > On Wed, 16 Apr 2003, James Jones wrote:
> > > > > > I found the problem and fixed it.  I mangled a few values while
> > > > > > reformatting some code so I could read it better.  Everything
> > > > > > appears to work now, I've tested resizing, clipping against all
> > > > > > sides of windows, ran gears while watching videos, etc.  xvinfo
> > > > > > reports healthy values.  I was using your patch before, I haven't
> > > > > > compared them side to side but everything seems to be working the
> > > > > > same.
> > > > > >
> > > > > > I removed all TV tuner/TV out code.  All I left in was the
> > > > > > functions required to make PutImage work.
> > > > > >
> > > > > > I'm attaching the patch here, its fairly small, I hope no one
> > > > > > minds. If someone has time to try it out and let me know how if
> > > > > > it works alright for people besides me, that would be keen.
> > > > > >
> > > > > > It can be applied to a fresh mach64-0-0-6-branch checkout, just
> > > > > > like Leif's patch, from the
> > > > > > xc/programs/Xserver/hw/xfree86/drivers/ati
> > > > > > directory use the following command:
> > > > > > gunzip -c mach64-xv-putimage-only-patch.diff.gz | patch -p1
> > > > > >
> > > > > > -James
> > > > > >
> > > > > > On Wednesday 16 April 2003 07:54 am, Leif Delgass wrote:
> > > > > > > I've posted new mach64 DRI+XVideo binaries for XFree86 4.3.0 on
> > > > > > > http://www.retinalburn.net/linux/dri_xv.html .  You might want
> > > > > > > to try those and compare the results.  The latest source patch
> > > > > > > was made against the current branch, so it should apply
> > > > > > > cleanly.  What is it that you want to remove from the patch? 
> > > > > > > The theatre* files shouldn't be needed, since they are only
> > > > > > > used by the r128 and radeon GATOS drivers, which I haven't
> > > > > > > merged or included in the patch.
> > > > > > >
> > > > > > > --Leif
> > > > > > >
> > > > > > > On Wed, 16 Apr 2003, James Jones wrote:
> > > > > > > > Well, I got a clean copy of the mach64-0-0-6-branch, read
> > > > > > > > through your quick XV+TVin/out patch, and started moving
> > > > > > > > chunks of it over to the clean dri code tonight.  I got it
> > > > > > > > running, but all the videos I play are sorta sliced in half. 
> > > > > > > > Hopefully I can get this fixed tomorrow.  I also think I can
> > > > > > > > cut a little of what I moved over out, as I'm not very
> > > > > > > > familiar with the code and had some trouble telling what did
> > > > > > > > what at first.
> > > > > > > >
> > > > > > > > With a little luck I should be able to get it all working and
> > > > > > > > make a patch soon.
> > > > > > > >
> > > > > > > > -James
> >
> > -------------------------------------------------------
> > This sf.net email is sponsored by:ThinkGeek
> > Welcome to geek heaven.
> > http://thinkgeek.com/sf
> > _______________________________________________
> > Dri-devel mailing list
> > Dri...@li...
> > https://lists.sourceforge.net/lists/listinfo/dri-devel