From mboxrd@z Thu Jan 1 00:00:00 1970 From: laurent.pinchart@ideasonboard.com (Laurent Pinchart) Date: Thu, 03 Apr 2014 13:31:02 +0200 Subject: [ARM ATTEND] DeviceTree status In-Reply-To: References: Message-ID: <2325559.Xh72H96OaN@avalon> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Hi Olof, On Wednesday 02 April 2014 22:16:33 Olof Johansson wrote: > Hi, > > Doing a generic reply on an old post, there's been many about DT though: > > On Mon, Mar 24, 2014 at 7:55 AM, Rob Herring wrote: > > I will be attending ELC. I think I may be the only DT maintainer > > attending (the rest were too scared). > > > > For the ARM summit, I can give a summary of what's happened in DT land > > since the last ARM summit. There's been some good progress although > > probably not as much as anyone would like. Given the last summit, I'm > > sure there is more to discuss. If not, I'm available for beers to > > celebrate solving all the issues. ;) > > I'm actually not excited about more discussion. There was a _ton_ of > it in Edinburgh, with many decisions done. Unless people have gone off > to actually try to implement some of the things we agreed need to be > implemented, and now need to come back with "it didn't work, we need > to change everything", then we shouldn't have to meet and spend > another mind-numbing day discussing DT. Or do we? > > > A key question to discuss is: are DT binding reviews improving? > > Hm. Do we need to meet in a room to talk about that, or can we discuss > it over email? > > I would say that they have improved, in particular because we've > started seeing more DT changes go in (and more bindings). There are > some areas that are still difficult, and I think the answer for those > is to find the right people and sit down and hash it out. ELC is > probably a good venue for some of that, but doing it in a room full of > ARM kernel maintainers might not be. That's a good point. Would it make sense to move DT discussions to a BoF at the ELC ? > > I also think the process for handling stable vs. unstable bindings > > needs more discussion. We also need to discuss how to deprecate > > existing "stable" bindings in order to have a way to stop new usage of > > poorly designed bindings we want to phase out. > > Do you have a proposal and a process in mind? Having something > specific to start a discussion off of is more useful than opening it > up for round table talks. -- Regards, Laurent Pinchart