[Mesa-dev] [RFC PATCH 0/26] Implement ARB_texture_multisample
stereotype441 at gmail.com
Fri Jan 4 13:57:54 PST 2013
On 29 December 2012 04:35, Chris Forbes <chrisf at ijw.co.nz> wrote:
> This series adds support for ARB_texture_multisample in core mesa
> and the i965 driver, and initially enables it for Gen6.
> There's a few rough edges still:
> - sample counts aren't quantized correctly
> - various small Gen7-specific things aren't done or adequately tested
> (Gen7 surface_state bits; testing for 8x mode; MCS; Gen7 `ld`
> argument layout)
> - HiZ/STENCIL_INDEX* interactions are almost certainly wrong on Gen6,
> since I'm not setting up HiZ surfaces at all.
> - dispatch is slightly bogus (using exec=mesa). not sure what i'm doing
> wrong here, but the proper exec=dynamic doesn't work. i'm sure im
> missing something really obvious.
> - some other Gen6 restrictions on 'large' multisample formats aren't
> taken into account (the PRM mentions that formats larger than 64bits
> per sample don't work)
> - almost certainly some style violations :)
> Piglits for this feature are in the texture_multisample branch at
> git://github.com/chrisforbes/piglit.git, and cover most things now.
> I'll tidy those changes up into a nice series for the piglit list soon.
Any idea when you'll be ready to submit these piglit tests for review? I'd
like to see the piglit patches land (or at the very least, get some review)
before we try to land this series, so that we can make sure the feature
goes out of the gate well-tested.
> Let me know how I'm doing it wrong :)
> -- Chris
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the mesa-dev