From mboxrd@z Thu Jan  1 00:00:00 1970
From: tony@atomide.com (Tony Lindgren)
Date: Tue, 14 Feb 2012 12:43:39 -0800
Subject: [PATCH 00/15] mach/io.h cleanup and removal
In-Reply-To: <4F3AB8FE.2060704@gmail.com>
References: <1329169408-17253-1-git-send-email-robherring2@gmail.com>
 <4F39D2EE.80806@gmail.com> <20120214172423.GO1426@atomide.com>
 <201202141757.03748.arnd@arndb.de> <4F3AB8FE.2060704@gmail.com>
Message-ID: <20120214204339.GV1426@atomide.com>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

* Rob Herring <robherring2@gmail.com> [120214 11:10]:
> On 02/14/2012 11:57 AM, Arnd Bergmann wrote:
> > On Tuesday 14 February 2012, Tony Lindgren wrote:
> >> Yes should be just the legacy drivers needing this for most part,
> >> so that's currently most of omap1 drivers for us.
> >>
> >> Anyways I'm using plat/omap-iomap.h for the name, so if somebody
> >> has better ideas for naming to avoid further search and replace
> >> later on, let me know.
> >>
> >> I guess in the long run we could have
> >>
> >> #include <mach-omap1/iomap.h>
> >>
> >> instead of
> >>
> >> #include <plat/omap-iomap.h>
> >>
> >> as the plat can't be used for multi-subarch builds.
> > 
> > I think it /could/ be used, we just need to make a definite
> > decision which way we want to go for header files that
> > are defined by a platform and used by code outside of that
> > platform such as device drivers.
> > 
> > The two main approaches that I can see are
> > 
> > a) make every header file name unique for platforms that you want to
> > build together, and just add every path at the compiler command line.
> > Not too much work, but somewhat error prone when you start having
> > file name conflicts.
> > 
> > b) move all platform specific header files into a directory named after
> > the platform and change all device drivers using this. Lots of work, but
> > probably better in the long run.
> > 
> 
> c) Don't allow mach includes in drivers and sound dirs for
> multi-platform kernels. This is already the case for any multi-arch
> driver. A lot of the headers are platform_data structs or things that
> should be cleaned up or need common infrastructure. Some cases I've
> found seem like the include is unnecessary. Also, just fixing up the
> name or path is no guarantee of avoiding namespace collisions.

Out of the three options c) makes most sense for multi-subarch kernels.
And that avoids having to sort out the name collisions with defines.

For dealing with legacy platforms, I too would prefer b).

Regards,

Tony