[Intel-gfx] [PATCH 00/13] rip out i830_memory, part 1: everything but xvmc

Daniel Vetter daniel at ffwll.ch
Tue Mar 2 21:08:35 CET 2010


On Tue, Mar 02, 2010 at 11:56:14AM -0800, Carl Worth wrote:
> On Tue, 02 Mar 2010 09:58:37 -0800, Eric Anholt <eric at anholt.net> wrote:
> > > Patch 2 is something I've stumbled upon while doing this cleanup. Totally
> > > untested because I don't have i965 class hw, but it makes much more sense this
> > > way. Please review carefully.
> 
> I didn't test this directly, but I agree that the code being changed
> looks entirely insane before the change, (one of those, "How could this
> even work?!" cases).

Just hope that the buffer is always freshly allocated, then ->offset is
initialized to 0.

[snip]

> I've got the whole series ready to push. I did notice the following
> warning though:
> 
> i830_memory.c:234: warning: ‘i830_allocate_memory’ defined but not used
> 
> for which I've written the patch below. With an ACK on that, I'll push
> out the whole series.

Oops, I've forgotten to compile test without XVMC enabled. jbarnes and krh
told me on irc to send the second patch pile, too. That one kills the old
i830_memory allocator completely, so your patch won't be needed anymore.

-Daniel
-- 
Daniel Vetter
Mail: daniel at ffwll.ch
Mobile: +41 (0)79 365 57 48



More information about the Intel-gfx mailing list