On 08/04/2019 16:57, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2019-04-08 15:54:01)
On 08/04/2019 10:17, Chris Wilson wrote:
As we push for better compartmentalisation, it is more convenient to
copy the default sseu configuration from the engine into the derived
logical context, than it is to dig it out from i915->runtime_info.
Why it is more convenient?
Read on in the series.
+static inline struct intel_sseu
+intel_device_default_sseu(const struct sseu_dev_info *sseu)
+{
I think function parameter change is dangerous. If caller can pass in
any pointer then it is slightly less "intel_*device_default*_sseu. It is
more a convert and copy from device sseu to intel_sseu. Like
intel_convert_device_sseu.
Although if the function just kept taking i915 then the dilemma would
not even be there.
What i915? How about if it took the parameter that described the device
and converted it into the internal format?
Don't know what i915 - function name says "default". If the caller
passes in the pointer it is not a default but could be anything. So you
can go with my first alternative, which is to give function a better
name if you want to stick with the parameter change.
Regards,
Tvrtko
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
