On 08/19/2013 04:47 PM, Tomasz Figa wrote: > Hi Markus, > > Please see my comments inline. > > On Monday 19 of August 2013 11:59:19 Markus Mayer wrote: >> This patch adds the GPIO driver for the bcm281xx family of chips. >> diff --git a/Documentation/devicetree/bindings/gpio/gpio-bcm-kona.txt >> +- interrupts: The interrupt outputs from the controller. There is one >> GPIO + interrupt per GPIO bank. The number of GPIO banks is HW >> configurable. > > What do you mean by HW configurable? Does it mean that it depends from > SoC or you can have different bank configurations on the same SoC? the wording is a little odd; I was going to comment on it, but thought it was probably clear enough. > How does the driver know the layout of GPIO banks? The code suggests that > the number of interrupts is used as bank count. While this saves you from > adding any new property or any other way of getting this information, I'm > not sure if it's completely correct. I'd like to hear others' opinion on > this, though. I expect a lookup table with key being the exact compatible value, and output being the # banks would be a reasonable driver implementation. Keying off the number of entries in interrupts seems fine too, since it should work out OK (Tegra does this IIRC). -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
