From mboxrd@z Thu Jan 1 00:00:00 1970 From: broonie@opensource.wolfsonmicro.com (Mark Brown) Date: Wed, 9 May 2012 15:32:00 +0100 Subject: [PATCH 10/10] spi: s3c64xx: add device tree support In-Reply-To: References: <1336514694-22393-1-git-send-email-thomas.abraham@linaro.org> <1336514694-22393-11-git-send-email-thomas.abraham@linaro.org> <20120509090716.GC28702@sirena.org.uk> Message-ID: <20120509143159.GU3955@opensource.wolfsonmicro.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, May 09, 2012 at 10:13:28PM +0800, Thomas Abraham wrote: > On 9 May 2012 17:07, Mark Brown wrote: > > On Wed, May 09, 2012 at 03:34:54AM +0530, Thomas Abraham wrote: > >> +- gpios: The gpio specifier for clock, mosi and miso interface lines (in no > >> + ?particular order). The format of the gpio specifier depends on the gpio > >> + ?controller. > > This seems odd... ?This isn't a bitbanging controller, and surely the > > driver will need to know which signal is which? ?I suspect this is > > actually for pinmux rather than to identify the signals but that should > > at least be made clear and really should be being done using the pinmux > > API. > The driver retrieves the list of gpio's that it is allowed to use. The > gpio numbers for miso, mosi and clk are mandatory but the order in > which they are specified is not important since the driver never needs > to which gpio is which interface line. I agree the pinmux api should > be used here, but the call to pinmux api would be a incremental change > here, not changing the code this patch is adding. I'd suggest just specifying the order - someone might want to use it later for some reason and it's not really a hardship for someone to use it. Avoids any "how does that work?" questions like I had. > >> + ?- samsung,spi-cs-gpio: A gpio specifier that specifies the gpio line used as > >> + ? ?the slave select line by the spi controller. The format of the gpio > >> + ? ?specifier depends on the gpio controller. > > We should really have a binding for this at the SPI level (and ideally > > some code to manage setting the GPIO too) - it's pretty common to use a > > GPIO as /CS. > The existing implementations vary in the way the nCS gpio lines are > specified. For some controllers, the nCS gpio's are included in the > spi device node whereas in this implementation, the nCS gpio is listed > in the spi slave device node. Yeah, I know. I'm saying we should try to come up with a binding for this that can be used by new SPI contollers going forward so things are consistent. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 836 bytes Desc: Digital signature URL: