On Fri, Feb 6, 2009 at 8:12 PM, Alan Coopersmith
<Alan.Coopersmith at sun.com> wrote:
> Martin Bochnig wrote:
>> Did you sometimes mention this before?
>> Then I can trash ati-drm, i doubt 2 different versions are needed.
>> or did i miss it, because it came on another alias (e.g. driver-discuss)?
>
> I found out on a phone call with them on Wednesday - I haven't seen them
> post their plans, and would suggest you talk to them to find out what exactly
> their ATI DRI plans are (I don't know if it's a full update to current
> community
> DRI or just fixing the current code back to working with the new interfaces).
>
> --
> -Alan Coopersmith- alan.coopersmith at sun.com
> Sun Microsystems, Inc. - X Window System Engineering
>
>
Alan, ok. Then it is not your fault.
Well, the problem with upgrading DRM: I bet it has to do with moving
some functionality into (Linux- and BSD-) kernel on 2007-07-21
01:16:42 (GMT) , which is not yet in the Solaris kernel. More
precisely with this global patch (affecting all vendors, not only ati)
:
"""Replace DRM_IOCTL_ARGS with (dev, data, file_priv) and remove DRM_DEVICE.
The data is now in kernel space, copied in/out as appropriate
according to the This results in DRM_COPY_{TO,FROM}_USER going away,
and error paths to deal with those failures. This also means that
XFree86 4.2.0 support for i810 DRM is lost. """
http://cgit.freedesktop.org/mesa/drm/commit/?id=5b38e134163cc375e91424c4688cc9328c6e9082
.
This prevents the entire DRM-Upgrade porting approach from being a
dumb merge of Sun's existing port into the current git.
Also: Isolating the actually important changes has not been fun, as
80% of the changes were bloating like sometimes changing // comments
into /* ... */ ones, and other times the other way around (makes
sense!!). Also things like eliminating tabs in header file #define
chains, while systematically putting new \t ' s into other places of
the code.
I had to go through thousands and thousands of lines of the diffs to
delete all the cosmetic stuff, allowing me to see what's really
required to make things work for the DDi/DKI interface. My goal is,
that the code can go upstream into
http://cgit.freedesktop.org/mesa/drm .
And it is halfway trivial to achieve this. Only an abundance of work
is required. That's my current project.
Then it will be easier and cheaper (!) for Sun to keep the DRM code in
sync with developments.
Thoughts?
