On Thu, Jan 2, 2014 at 2:33 PM, Baruch Siach [off-list ref] wrote:

On Thu, Jan 02, 2014 at 02:21:12PM +0100, Linus Walleij wrote:

quoted

On Thu, Jan 2, 2014 at 1:18 PM, Baruch Siach [off-list ref] wrote:

quoted

On Thu, Jan 02, 2014 at 01:07:24PM +0100, Linus Walleij wrote:

quoted

On Mon, Dec 16, 2013 at 8:05 AM, Baruch Siach [off-list ref] wrote:

quoted

Prevent build failure when the selected variant does not support GPIO32.

Signed-off-by: Baruch Siach <baruch@tkos.co.il>

(...)

quoted

+config HAVE_XTENSA_GPIO32
+       bool
+
 config GPIO_XTENSA
        bool "Xtensa GPIO32 support"
        depends on XTENSA
+       depends on HAVE_XTENSA_GPIO32

This is just a tautology, you don't need a helper bool
for this.

Instead just select GPIO_XTENSA from your affected
platforms.

But then you force this driver on all users. Why not leave the option to
disable this driver when the user needs direct access to this hardware, or
just wants to remove unused code?

Hmm you're right. But if the build fails on !XTENSA, can you not
put a

depends on XTENSA_VARIANT_FSF || XTENSA_VARIANT_DC232B || XTENSA_VARIANT_DC233C

Into drivers/gpio/Kconfig?

It's a simple oneliner...

IMO, having a 'select HAVE_XTENSA_GPIO32' in arch/xtensa/Kconfig would makes
it clear that this is something to consider when adding another Xtensa
variant. Also, this dependency line might (hopefully will) become quite long.
But if you object I'll change that.

I have no strong opinion on this, I would leave that decision to the
Xtensa maintainer.

Yours,
Linus Walleij