From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sam Ravnborg Subject: Re: [PATCH 24/25] drm: kirin: Pass driver data to crtc init and plane init Date: Wed, 24 Apr 2019 23:15:36 +0200 Message-ID: <20190424211536.GB27952@ravnborg.org> References: <1556061656-1733-1-git-send-email-john.stultz@linaro.org> <1556061656-1733-25-git-send-email-john.stultz@linaro.org> <20190424170910.GC21067@ravnborg.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org To: John Stultz Cc: Xu YiPing , David Airlie , Chen Feng , lkml , dri-devel , Xinliang Liu , Xinwei Kong , Rongrong Zou List-Id: dri-devel@lists.freedesktop.org Hi John. > > > I missed where ade_driver_data came from. > > This looks an extra patch to intoduce driver_data, > > that maybe should be merged with an earlier version? > > I'm not sure I'm following you here. Can you clarify a bit more? So I looked at this a bit more - and got the bigger picture in place again. driver_data is assigned using the lookup done at probe() time. For now this is just assigned to ade_driver_data as this is the only option. So an indirection via driver_date or calling ade_driver_data direct is the same. And you have several patches where you migrate to use driver_data rather than calling ade_driver_data direct. It confused me that the patch introducing the lookup at probe() came before all call sites were migrated to use driver_data. But I get it now so it is fine. Maybe a few words in the commit log like: This patch refactor to call functions via driver_data, rather than hardcoding them via ade_driver_data. This is doen so we later can assing another stucture to driver_data to support other chips. PS. I did not complain about your spelling mistakes in the changelog. I have a similar (or worse) keyboard from a spelling point of view. Sam