Peter De Schrijver wrote at Monday, November 21, 2011 7:29 AM:

On Fri, Nov 18, 2011 at 10:41:16PM +0100, Olof Johansson wrote:

quoted

On Thu, Nov 17, 2011 at 06:19:20PM +0200, Peter De Schrijver wrote:

quoted

This patch modifies the pinmux code to be useable for multiple tegra variants.
Some tegra20 specific constants will be replaced by variables which will be
initialized to the appropriate value at runtime.

...

quoted

quoted

diff --git a/arch/arm/mach-tegra/board-harmony-pcie.c b/arch/arm/mach-tegra/board-harmony-pcie.c

...

quoted

quoted

+#include <mach/pinmux-tegra20.h>

Boards shouldn't have to include this. The idea is that you should only
have to do board code against the pinmux.h interface, which internally
abstracts it for tegra 20 vs tegra 30.

The pinmux naming is still SoC specific. Unless we move this to devicetree
(which should be part of a different patchset), I don't see how we can solve
this, except by renaming all the pingroups. That would cause a lot more
changes though.

I believe that Peter's explanation makes sense.

BTW, re: the new pinmux API: I have a driver for both Tegra20 and Tegra30
now which should work for muxing. All the data is present for all the pin
configs too (tri-state, pull, ...) but we need to define the core pinctrl
API for this before I can expose that. Similarly, we need DT bindings for
all of this before we can make use of it; something I hope to start work
on in the near future.

-- 
nvpublic