Re: [GIT PULL] MT8173 DRM support

From: Daniel Vetter <daniel@ffwll.ch>
To: Eric Anholt <eric@anholt.net>
Cc: Jie Qiu <jie.qiu@mediatek.com>,
	Cawa Cheng <cawa.cheng@mediatek.com>,
	dri-devel@lists.freedesktop.org, kernel@pengutronix.de,
	Matthias Brugger <matthias.bgg@gmail.com>
Subject: Re: [GIT PULL] MT8173 DRM support
Date: Wed, 20 Apr 2016 19:32:51 +0200	[thread overview]
Message-ID: <20160420173251.GO2510@phenom.ffwll.local> (raw)
In-Reply-To: <87potks136.fsf@eliezer.anholt.net>

On Wed, Apr 20, 2016 at 09:21:17AM -0700, Eric Anholt wrote:
> Daniel Vetter <daniel@ffwll.ch> writes:
> 
> > On Wed, Apr 20, 2016 at 03:36:16PM +0200, Matthias Brugger wrote:
> >> 
> >> 
> >> On 19/04/16 15:42, Philipp Zabel wrote:
> >> >Hi Dave,
> >> >
> >> >please consider pulling this tag with initial MediaTek MT8173 DRM
> >> >support, corresponding to v14 of the patch series. These patches have
> >> >been mostly stable for the last few rounds. I'll follow up with the HDMI
> >> >encoder support pending review of the latest version.
> >> >
> >> 
> >> Please don't pull
> >> e34ba70de8c4 ("arm64: dts: mt8173: Add display subsystem related nodes")
> >> If you pull the rest, this patch will go through my branch.
> >
> > So not on top of this at all, but do we have to split up arm drm drivers
> > so much? Generally this stuff goes in through one tree with the driver,
> > with acks from other subsystem as needed. That ack seems to be missing,
> > but imo better to supply it and just get this pull req through. Or double
> > merge a patch, we do that fairly often.
> >
> > Anyway just a comment, but sitting outside watching I think arm has a
> > pretty serious problem with tree proliferation. And it's not helping to
> > get fairly simple drivers like this one merged ...
> 
> DT changes do get merged through the arm DT tree separately.  ARM has
> absurd tree proliferation, but I think in this case it actually makes
> sense.  DT is the most common place I have merge conflicts when working
> on platform enabling across the steaming piles of subsystem trees out
> there, and keeping DT separate means that a merged -next tree can be
> built sanely.

Ok, makes sense. Still I guess just double-merging in this case would be
the simplest option. Git should be clever enough in general to realize
what's going on.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel