Dave,

I'm leaving for a two-week workshop tomorrow, so here's a quick summary of 
what r300-bufmgr looks like to me.

1) Texture storage still requires 50% more memory than before. I have a first 
cut of a patch for this in my tree at http://cgit.freedesktop.org/~nh/mesa/. 
The obvious problems are gone, but I'm still seeing some bugs. Unfortunately 
I didn't have the time to test whether they're actually caused by my patch or 
by previous r300-bufmgr work because

2) Some of the Piglit testcases (texCombine, lodbias) seem to take *a lot* 
longer than they used to.

3) There are a number of correctness problems with texture handling even 
without the patch above. They seem mostly related to cubemaps, 
texture_rectangle. Also, maskedClear no longer works correctly which is 
probably due to the renderbuffer patch, but I haven't investigated that one.

Also, unlike git master, the branch leaks very visibly; see my earlier email 
about this problem. It turns out that there are some interesting ordering 
problems. In Glean cases (where multiple GL contexts are created and 
subsequently destroyed), driDestroyDrawable *is* destroyed, but at a 
completely unexpected time *after* the bufmgr corresponding to the context is 
destroyed.

In fact, there's another potential problem, which is texture sharing between 
contexts. I didn't think this through, but it seems like having the bufmgr in 
the GL context is incorrect?

cu,
Nicolai

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Mesa3d-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mesa3d-dev

Reply via email to