From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tony Lindgren Subject: Re: [PATCH 2/4] ARM: msm: Remove 7x00 support Date: Tue, 29 Oct 2013 11:40:45 -0700 Message-ID: <20131029184044.GF15154@atomide.com> References: <1382993006-27359-1-git-send-email-davidb@codeaurora.org> <1382993006-27359-3-git-send-email-davidb@codeaurora.org> <20131029132043.GA28165@fifo99.com> <20131029170817.GA13047@fifo99.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mho-02-ewr.mailhop.org ([204.13.248.72]:14123 "EHLO mho-02-ewr.mailhop.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751511Ab3J2Skz (ORCPT ); Tue, 29 Oct 2013 14:40:55 -0400 Content-Disposition: inline In-Reply-To: Sender: linux-arm-msm-owner@vger.kernel.org List-Id: linux-arm-msm@vger.kernel.org To: Olof Johansson Cc: Daniel Walker , David Brown , Bryan Huntsman , Russell King , Arnd Bergmann , Kevin Hilman , "linux-kernel@vger.kernel.org" , linux-arm-msm@vger.kernel.org, "linux-arm-kernel@lists.infradead.org" * Olof Johansson [131029 10:40]: > On Tue, Oct 29, 2013 at 10:08 AM, Daniel Walker wrote: > > > Personally I think splitting mach- stuff isn't very useful or > > interesting.. There's just no technical reason for it, for example x86 > > and x86_64 was a win from my perspective , there's a lot more reason to > > keep similar things together than to split things up. > > There are definitely valid technical reasons for it; the old and new > platforms share no code, and the legacy platforms are unlikely to be > updated to modern infrastructure anytime soon. Other platforms are > managed in similar manners, such as OMAP, imx/mxs, etc. Yeah there are still few valid reasons to have separate mach directories. The main reason why mach-omap2 was originally set up separately from mach-omap1 was because the IO space was different. And we could not properly deal with that until CONFIG_ARM_PATCH_PHYS_VIRT few years ago. So we placed the shared code into plat-omap, which worked OK but is not really needed any longer with device tree. We have only dmtimer and legacy DMA code left in plat-omap pretty much. And those will be moved to live under drivers/. Even with most issues fixed, it still does not not make sense to merge mach-omap1 and mach-omap2. For example, even if somebody wanted to do it as a hobby project, we'd have to compile things with v4 or v5 flags, which won't work properly for SMP cores at least :) Regards, Tony