Hi Linus,

On Thu, Jun 15, 2017 at 02:57:53PM +0200, Linus Walleij wrote:

On Thu, Jun 15, 2017 at 10:55 AM, Geert Uytterhoeven
[off-list ref] wrote:

quoted

If clocks and resets are provided by the same hardware module, you can
have a single (platform) driver registering both the clock and reset
controllers.
Cfr. drivers/clk/renesas/renesas-cpg-mssr.c.

That is indeed an option.

So I would say, clk & reset maintainers: would you prefer that I merge the
reset control into the clock driver as well, ask Philipp to drop the pending
reset control patches from his subsystem tree and have you manage the
combined driver and bindings?

The reset/next pull requests are not merged into the arm-soc tree yet.
I suppose I could retract the pull requests and drop the Gemini reset
patches, if the patches in arm-soc/gemeni/dts are also dropped from
arm-soc/for-next.

It seems to me as very ugly from a divide & conquer subsystem and file
split point of view.

I seems elegant from the "make clocks a platform device" point of view.

I am happy with either approach as long as it works.

I guess it is up to the taste of the subsystem maintainers, especially
clk.

If I get some time I might just hack this up and send the patches so
it is on the table as an alternative to the current v5 patch. Certainly it is
better than going back and augmenting the DT bindings.

I have a slight preference for keeping the DT bindings simple, even if
that means merging the reset controller into the clock driver.

regards
Philipp