Re: [PATCH 2/4] drm/i915: Make *_crtc_mode_set work on new_config

[Daniel Vetter <daniel@ffwll.ch>]

On Sun, Oct 19, 2014 at 04:28:57PM +0200, Daniel Vetter wrote:
> On Thu, Oct 09, 2014 at 03:18:03PM +0300, Ander Conselvan de Oliveira wrote:
> > On 10/09/2014 12:11 PM, Daniel Vetter wrote:
> > >On Wed, Oct 08, 2014 at 06:32:21PM +0300, Ander Conselvan de Oliveira wrote:
> > >>From: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
> > >>
> > >>This shouldn't change the behavior of those functions, since they are
> > >>called after the new_config is made effective and that points to the
> > >>current config. In a follow up patch, the mode set sequence will be
> > >>changed so this is called before disabling crtcs, and in that case
> > >>those functions should work on the staged config.
> > >>
> > >>Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
> > >>---
> > 
> > [snip]
> > 
> > >>diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> > >>index 3a06c3d..9d8fe8d 100644
> > >>--- a/drivers/gpu/drm/i915/intel_display.c
> > >>+++ b/drivers/gpu/drm/i915/intel_display.c
> > >>@@ -420,13 +420,32 @@ static bool intel_pipe_has_type(struct drm_crtc *crtc, int type)
> > >>   return false;
> > >>  }
> > >>
> > >>+/**
> > >>+ * Returns whether any output on the specified pipe will have the specified
> > >>+ * type after a staged modeset is complete, i.e., the same as
> > >>+ * intel_pipe_has_type() but looking at encoder->new_crtc instead of
> > >>+ * encoder->crtc.
> > >>+ */
> > >>+static bool intel_pipe_will_have_type(struct drm_crtc *crtc, int type)
> > >
> > >s/drm_crtc/intel_crtc/. In general we use the most specific type
> > >for internal functions to avoid tons of upcasting. So you might want to
> > >throw a patch on top to convert the existing has_type for consistency.
> > 
> > What about the functions in drm_i915_display_funcs (find_dpll,
> > crtc_mode_set)? Are these not considered internal functions or is this
> > leftover from before this rule was in place?
> > 
> > If I write that consistency patch it might be easier to just convert
> > everything.
> 
> leftovers, and we've been slowly converting existing code over to using
> intel structures over the past 2 years. I certainly won't stop you to
> clean up all that stuff, but it's a bit of work. So I'm ok if you just
> have the vfunc signature use the intel_ types and then immediately convert
> to a drm_ type again to avoid code churn.
> 
> Or just convert it all if you feel like it.

But if you do that please split out that rote conversion into a separate
patch for easier reviewing. Either at the start of the series (so I can
merge it right away) or at the end (only done once everything is merged)
to avoid needless rebase pain.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch