On 22.02.2017 16:19, Michel Dänzer wrote:
On 22/02/17 07:58 PM, Nicolai Hähnle wrote:
On 22.02.2017 07:23, Michel Dänzer wrote:
On 22/02/17 12:45 PM, Timothy Arceri wrote:
+get_disk_shader_cache
+^^^^^^^^^^^^^^^^^^^^^
+
+Returns a pointer to driver-specific on-disk shader cache. If the
driver
+failed to create the cache or does not support an on-disk shader
cache NULL is
+returned.
[...]
+ /**
+ * Returns a pointer to driver-specific on-disk shader cache. If
the driver
+ * failed to create the cache or does not support an on-disk
shader cache
+ * NULL is returned.
+ */
+ struct disk_cache *(*get_disk_shader_cache)(struct pipe_screen
*screen);
};
Drivers which don't support an on-disk shader cache don't set this
callback in the first place, right? :) (Just a suggestion for
improvement before landing this patch, not a blocker, no need to resend)
Yeah, but creating the shader cache might have failed, or it might be
disabled. Perhaps a driver that doesn't do its own caching might want to
initialize the cache lazily. I think the interface is more flexible this
way.
Sure. Maybe I was being too cryptic, sorry.
What I mean is that the "or does not support an on-disk shader cache"
language in the callback documentation is slightly misleading, since
such drivers should simply not set the callback in the first place.
That's all.
Ah, yes, that makes sense.
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
