[Intel-gfx] [PATCH] drm/i915: Avoid fluctuating to UNKNOWN connector status during forced probes

[Chris Wilson]

On Thu, Jun 04, 2015 at 07:28:31PM +0300, Ville Syrjälä wrote:
> So besides avoiding the disconnected->unknown transition, the other thing
> we now seem to do is set intel_tv->type to -1 when we find that things
> are disconnected. However that doesn't seem to be user visible, and I
> can't see any place in the code where it should cause problems.

Mostly because after the first refactor, it was not immediately obvious
that doing intel_tv->type = type would be valid, so I only wanted to set
intel_tv->type when doing load_detect. Fortunately, the value is only
used when connected - but it could conceivable be used when the
connector status is UNKNOWN but then it is either still set as
DRM_MODE_CONNECTOR_Unknown from init, or the last known active value.

> It is a bit inconsistent however due to the fact that we initialize type
> to DRM_MODE_CONNECTOR_Unknown, and later we use the -1 to indicate more
> or less the same thing. Seems like we should pick or the other for
> consistency.

Yup. Might as well do that now, or else it will another few years before
the next pass through tumbleweed town.
 
> Apart from the little things like the weird pre-exising extra
> ddc detect and the intel_tv->type inconsistency this patch looks OK to
> me, so
> Reviewed-by: Ville Syrjälä <ville.syrjala at linux.intel.com>

That DDC is definitely weird, but one I'm not inclined to touch without
good reason.  Preparing a v2 with the minor tv_type change, I'll
presumptively stick your r-b on it ;)
-chris

-- 
Chris Wilson, Intel Open Source Technology Centre