[PATCH v5] drm/i915/selftests: Refactor RC6 power measurement and error handling
Andi Shyti
andi.shyti at linux.intel.com
Tue Apr 1 07:24:17 UTC 2025
Hi Anirban,
On Fri, Mar 28, 2025 at 12:49:24AM +0530, Sk Anirban wrote:
> Revise the power measurement logic to save and evaluate energy values.
How is the logic been revised? Can you be more specific please?
> Previously, the test only checked whether the system had entered the RC6
> state, without considering any potential interruptions in that state.
> This update introduces a threshold check to ensure that the GPU remains
> in the RC6 state properly during the specified sleep duration.
How is the threshold been evaluated? As far as I can see, there
is only one pr_err() added in a path that fails anyway (it's
almost more a debug). Should we evaluate the threshold outside
from the failing path?
Andi
More information about the Intel-gfx
mailing list