Il 05/09/23 09:58, Michael Walle ha scritto:
At this point, I think that it would be sane to change this function to
return a signed type, so that we can return -ENOENT if we couldn't find
any DDP path (so, if we couldn't find any possible crtc).
Fair enough, but should it be part of the fixes commit or a different one?
-michael
I would say that this should be part of the Fixes commit: after all, you're
fixing a case in which the possible_crtcs calculation *may fail*, so the
error handling for this possible failure is, indeed, one part of the fix :-)
Angelo