I second the comment on splitting this patch up into its constituent chunks. It'll be practical to review each of them properly at that point. Like ... it seems wrong to teach plat-omap/devices.c about EVM-specific stuff; it should stay generic. And those EVM-specific bits are done wrong too; those regulators are I2C devices, not platform devices! Plus, their constraints should be board-specific, not generic. Move all that into the omap3evm board code. You still didn't remove the drivers/i2c/chips stuff. Move the *ENTIRE* driver to drivers/regulator, and get rid of all the platform_device stuff. Good to see this driver start moving into its proper home, though. :) On Thursday 04 December 2008, y@xxxxxxxxxxxx wrote: > Not fixed comments: > Not clear on how to implement runtime check for PR785 and would require some > inputs on implementing the same. The TPS driver should not know or care about the PR785 card; it should just respond to the various requests. If there's any knowledge of PR785 in drivers/anything/* you should treat that as a bug (and fix it before merge). Since the chips are pretty much the same other than some bit interpretations, just use the driver_data in the i2c_device_id to describe whatever differences the driver needs to know about. With respect to runtime checking, first make sure the OMAP3 EVM init code is able to cleanly choose between the two power card options. Once that's done you can try turning that into a runtime check. - Dave -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
