[igt-dev] [PATCH i-g-t 4/4] lib/kms: warn if we commit without outputs

[Chris Wilson]

Quoting Daniel Vetter (2018-10-05 16:07:38)
> On Thu, Oct 04, 2018 at 03:21:28PM +0200, Daniel Vetter wrote:
> > With the high-level helpers requiring outputs there's not point
> > in silently ignoring issues anymore. Complain about that if it
> > ever happens.
> > 
> > Cc: Antonio Argenziano <antonio.argenziano at intel.com>
> > Cc: Chris Wilson <chris at chris-wilson.co.uk>
> > Signed-off-by: Daniel Vetter <daniel.vetter at intel.com>
> 
> I guess my motivation with fumbling around with the igt_display_* api
> wasn't entirely clear: It's this patch here, respectively Chris' patch
> which added the silent short circuit.
> 
> Imo that's very brittle api, asking for trouble, and me not recognizing
> right away what's happening in the debugfs is kinda proving the point.
> Silently throwing a request away from the testcase is at least very
> surprising. And inconsistent with both more explicit igt_require/assert
> checks in drivers, and (the style I favour personally, but really not the
> issue here) putting igt_require/assert into the helper library.

I'd argue the opposite. The test case requested us to configure 0
outputs across 0 pipes, and we are obeying it. Not second guessing what
the test case intended.
-Chris