From mboxrd@z Thu Jan 1 00:00:00 1970 From: Simon Horman Date: Tue, 05 Aug 2014 06:54:03 +0000 Subject: Re: [PATCH] ARM: shmobile: Separate APMU resource data into CPU dependant part Message-Id: <20140805065359.GD1534@verge.net.au> List-Id: References: <1407169276-25495-1-git-send-email-ulrich.hecht+renesas@gmail.com> In-Reply-To: <1407169276-25495-1-git-send-email-ulrich.hecht+renesas@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-sh@vger.kernel.org On Tue, Aug 05, 2014 at 02:39:15PM +0900, Magnus Damm wrote: > On Tue, Aug 5, 2014 at 10:15 AM, Simon Horman wrote: > > On Mon, Aug 04, 2014 at 06:21:16PM +0200, Ulrich Hecht wrote: > >> From: Hisashi Nakamura > >> > >> Since APMU resource data were not common data of the R-Car series, > >> it was separated. > >> > >> Signed-off-by: Hisashi Nakamura > >> [uli: moved header file out of arch/arm/mach-shmobile/include/mach] > >> Signed-off-by: Ulrich Hecht > >> --- > >> > >> This is needed for r8a7794 which has a set of CPU cores different from the > >> other Gen2s. > > > > So unfortunate but given the facts this approach seems reasonable to me. > > Magnus (CCed), do you have an opinion on this? > > This information shall really be described via DT in my option. > Perhaps this patch could be one step in the right direction though. > Unfortunately the patch seems to contain incorrect information though, > the r8a7791 SoC only contains two CPU cores so now when we separate > the code perhaps we should do it right? > > My suggestion is to add R-Car E2 support as UP-only first, then we can > add SMP bits incrementally over time. It seems to me that this would be a good candidate to skip straight to supporting using DT rather with the option of switching back to adding legacy-C code if a pressing need arises. As Magnus suggests it may be easiest to achieve by focusing on UP-only support to begin with.