Return-Path: Date: Thu, 18 Aug 2016 15:40:19 +0100 From: One Thousand Gnomes To: Greg Kroah-Hartman Cc: "H. Nikolaus Schaller" , Pavel Machek , Rob Herring , Marcel Holtmann , Jiri Slaby , Sebastian Reichel , Peter Hurley , NeilBrown , Arnd Bergmann , Linus Walleij , linux-bluetooth@vger.kernel.org, linux-serial@vger.kernel.org, linux-kernel@vger.kernel.org Subject: Re: [RFC PATCH 0/3] UART slave device bus Message-ID: <20160818154019.40f4df45@lxorguk.ukuu.org.uk> In-Reply-To: <20160818105759.GA642@kroah.com> References: <20160818011445.22726-1-robh@kernel.org> <118926C8-F4D0-41F5-B6A8-690E0312F3FB@goldelico.com> <20160818104749.GB7427@amd> <1ED14975-FC0C-44AB-8F05-782DBC9E98F8@goldelico.com> <20160818105759.GA642@kroah.com> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII List-ID: On Thu, 18 Aug 2016 12:57:59 +0200 Greg Kroah-Hartman wrote: > On Thu, Aug 18, 2016 at 12:54:15PM +0200, H. Nikolaus Schaller wrote: > > Hi Pavel, > > > > > Am 18.08.2016 um 12:47 schrieb Pavel Machek : > > > > > > > > >> > > >> Thereof 4 files, ~260 changes w/o gps demo and documentation/bindings. > > > > > > So what do you use for the serial devices? platform_device was vetoed > > > for that purpose by Greg. > > > > device tree? > > No. > > This patchset from Rob is the way I have been saying it should be done > for years now. Yes, a "bus" takes up more boilerplate code (blame me > for that), but overall, it makes the drivers simpler, and fits into the > rest of the kernel driver/device model much better. The basic problem is that the bus should be tty_ports not uart, fix that and the rest starts to make sense. Alan