On Wed, 2011-05-04 at 11:36 +0900, Misha Shnurapet wrote: > <driconf> > <device screen="0" driver="dri2"> > <application name="Default"> > <option name="vblank_mode" value="0" /> > </application> > </device> We're absolutely never doing this by default. 0 means "never sync with vertical retrace even if the app asks to". The other values for this option are: 1: Default swap interval is 0, but respect app settings 2: Default swap interval is 1 (ie, the refresh rate), but respect app settings 3: Swap interval is never < 1, but otherwise respect app settings 2 is the default, and in the absence of a particularly strong argument to change it, I'd prefer not. > <device screen="0" driver="i915"> > <application name="Default"> > <option name="force_s3tc_enable" value="false" /> > <option name="no_rast" value="false" /> > <option name="always_flush_cache" value="false" /> > <option name="early_z" value="false" /> > <option name="stub_occlusion_query" value="false" /> > <option name="always_flush_batch" value="false" /> > <option name="bo_reuse" value="1" /> > <option name="texture_tiling" value="true" /> Already the default. > <option name="vblank_mode" value="0" /> A repeat of the above. > <option name="allow_large_textures" value="2" /> Already the default. > <option name="fragment_shader" value="false" /> Sigh, such a judgement call with this. The issue here is that people with lame hardware (like, for example, any Intel "gen3" chip, including 915, 945, and Atom chips) really wish they had working ARB_fragment_shader support. But instead, they bought a gen3. You can get _some_ fragment shaders to compile for gen3, but not all, and when they fail you fall back to software in a way that's more expensive than simply having done everything in software in the first place. I think Mesa's gone back and forth on this over time, and I tend to lean towards correctness (ie, that what you've set here would be the default), but then it becomes a battle with the people who say "but my app works _fine_ with shaders enabled". I'm pretty sure we can install a systemwide driconf file to enable it for only some apps, and if someone wanted to maintain that I'd be happy to let them, but in the absence of that I'd rather stick with whatever upstream has picked. - ajax
Attachment:
signature.asc
Description: This is a digitally signed message part
-- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/devel
