[igt-dev] [PATCH i-g-t] kms_atomic_transition: Make the available pipe discovery dynamic

[Petri Latvala]

On Mon, Apr 20, 2020 at 03:03:12PM +0100, Chris Wilson wrote:
> All of the per-plane tests are run on all available matching pipes. A
> natural case for igt_subtest_with_dynamic + igt_dynamic_f
> 
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> Cc: Petri Latvala <petri.latvala at intel.com>
> ---


A side note, now that this change has been in effect for a bit and
properly skips when no matching pipe/output combos are found:

We're using for_each_pipe_with_valid_output, omitting the loop
elements with a (non-)internal panel. However, that loop only loops
through pipes, trying one output valid for that pipe. In effect, we
only test internal panels even if we have non-internal outputs
available, if the pipe can also drive the internal panel's output.

Maarten, do we really need the non/internal panel separation for this
test?


-- 
Petri Latvala