[Intel-gfx] [PATCH] drm/i915: Enable non-contiguous pipe fusing

[Ville Syrjälä]

On Thu, Mar 19, 2020 at 01:23:18PM +0530, Anshuman Gupta wrote:
> On 2020-03-18 at 18:48:27 +0200, Ville Syrjälä wrote:
> > On Fri, Mar 13, 2020 at 12:39:17AM -0700, Lucas De Marchi wrote:
> > > On Wed, Mar 11, 2020 at 02:06:32PM +0530, Anshuman Gupta wrote:
> > > >Allow 3-display pipes SKU system with any combination
> > > >in INTEL_INFO pipe mask.
> > > >B.Spec:50075
> > > >
> > > >changes since RFC:
> > > >- using intel_pipe_mask_is_valid() function to check integrity of
> > > >  pipe_mask. [Ville]
> > > >v2:
> > > >- simplify condition in intel_pipe_mask_is_valid(). [Ville]
> > > >v3:
> > > >- removed non-contiguous pipe fusing check. [Lucas]
> > > 
> > > I'd also say in the commit message that the support for non-contiguous
> > > pipe fusing is *already* supported in the driver. So this check here
> > > doesn't make sense anymore and since it's an unlike condition we
> > > can just stop checking.
> > 
> > BTW I think we still have those crtc index==pipe asserts in the code
> > somewhere. Now that all the (known) assumptions have been fixed we can
> > remove the WARNs.
> yes we still have drm_WARN_ON(&dev_priv->drm, drm_crtc_index(&crtc->base) != crtc->pipe)
> in intel_crtc_init.
> AFAIU the idea was to have the WARN_ON to know that we are running a broken driver
> (if there any unknown assumption is still left out).

Unknown assumptions are by definition unknown. The WARN will not help
with those.

-- 
Ville Syrjälä
Intel