Glamor - next step

[Jamey Sharp]

On Fri, Oct 21, 2011 at 11:00:20AM +0800, Zhigang Gong wrote:
> Hi folks,
> 
> Thanks for your valuable advices. To incrementally merge glamor with current
> video driver is indeed a more reasonable and less risk solution than merge it
> directly into xorg.
> 
> The idea to bind BOs to a textures works fine. Actually, I already did so when I 
> try to get DRI2 work with glamor.
> 
> Now, it become clear for next step.
> 
> 1. Extract the device independent part of glamor into an independent library which
>   could be linked by other ddx driver.
> 2. Incrementally change current intel-video driver's rendering path to GL-based
>   rendering path.
> 
> If there is still any misunderstanding, please help to point it out.
> Thanks.

That looks perfect. I'm hoping you will eventually port the Xephyr bits
into xf86-video-nested, too, but demonstrating on one driver would be a
great start, and helps illustrate what to do if somebody else wants to
try it with another driver.

This plan also has the advantage that you only need your co-workers'
approval to merge your work, not the rest of us on xorg-devel. :-)

Jamey

> -- zhigang
> 
> > -----Original Message-----
> > From: xorg-devel-bounces+zhigang.gong=linux.intel.com at lists.x.org
> > [mailto:xorg-devel-bounces+zhigang.gong=linux.intel.com at lists.x.org] On
> > Behalf Of Eric Anholt
> > Sent: Thursday, October 20, 2011 11:32 PM
> > To: Jamey Sharp; Zhigang Gong; Jeremy Huddleston
> > Cc: xorg-devel at lists.x.org
> > Subject: Re: [Pull v2] Glamor - fixed build failure, fixed coding style
> > problem.
> > 
> > On Thu, 20 Oct 2011 09:54:29 +0200, Jamey Sharp <jamey at minilop.net>
> > wrote:
> > Non-text part: multipart/mixed
> > Non-text part: multipart/signed
> > > This is also like Xgl, as I recall, in that Xgl was a large
> > > development effort happening inside a company (Novell, in that case)
> > > and then being dropped on the X.Org community. I'd say that had a lot
> > > to do with it eventually failing.
> > >
> > > A bunch of this work has Eric Anholt's and Kristian Høgsberg's names
> > > on it, and I'd have expected them to remember that debacle.
> > 
> > I didn't get very far in the work before deciding that getting DRI2 working
> > was going to be hard and that further work wasn't useful until I had some
> > sort of plan for it.
> > 
> > My intention back when I was working on it was to do incremental merging
> > once DRI2 was figured out -- there's an obvious merge point where
> > correct all-software-fallbacks are possible, then another when the first
> > couple of blit paths are in, then Render accel.
> > 
> > But now, in order to get DRI2 working with GL-based 2D, we're thinking of
> > flipping the whole thing on its head and using our current driver, binding
> > its BOs into GL, and doing incremental migration of our current rendering
> > to GL rendering that way.  We've burned ourselves switching to new AAs
> > both times we've done it, so any plan that involves incremental changes
> > has a lot recommending it.
> 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.x.org/archives/xorg-devel/attachments/20111021/c3519c89/attachment.pgp>