Hi, Currently, many drivers implement their own strategy when trying to find which timer to use as the clocksource or the clockevent. The main issue is that this selection is happen early in the boot process and the kernel doesn't always have all the information to take that decision. So we end up with suboptimal solutions, especially in a multiplatform kernel setting, as the MXC_USE_EPIT or ATMEL_TCB_CLKSRC_BLOCK kernel config option. There is also the clocksource kernel parameter only implemented in mach-omap2. Other drivers are registering the first seen timer as a clockevent, the other one as a clocksource. Also, this will help in the goal of separating clocksource and clockevent drivers (see 376bc27150f180d9f5eddec6a14117780177589d) Patch 1 documents the binding, patch 2 implements the parsing of the chosen node and finally, patch 3 makes use of the parsing in a driver to give an overview of how it is working. Alexandre Belloni (3): dt-bindings: chosen: Add clocksource and clockevent selection clocksource/drivers: timer-of: parse the chosen node clocksource/drivers: integrator-ap: parse the chosen node Documentation/devicetree/bindings/chosen.txt | 20 ++++++++++++++++++++ drivers/clocksource/Kconfig | 1 + drivers/clocksource/timer-integrator-ap.c | 11 +++++++++++ drivers/clocksource/timer-of.c | 22 ++++++++++++++++++++++ drivers/clocksource/timer-of.h | 3 +++ 5 files changed, 57 insertions(+) -- 2.15.1 -- 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
