From mboxrd@z Thu Jan 1 00:00:00 1970 From: pawel.moll@arm.com (Pawel Moll) Date: Wed, 13 Mar 2013 11:46:15 +0000 Subject: [PATCH v3 06/11] ARM: vexpress: use clocksource_of_init for sp804 In-Reply-To: References: <1363151142-32162-1-git-send-email-haojian.zhuang@linaro.org> <1363151142-32162-7-git-send-email-haojian.zhuang@linaro.org> <1363173037.3100.22.camel@hornet> Message-ID: <1363175175.3100.37.camel@hornet> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, 2013-03-13 at 11:42 +0000, Haojian Zhuang wrote: > > I love the possibility of removing the SP804 code from the VE machine > > (thanks for your work!), but I'd like to avoid it being replaced by > > cryptic properties in the tree. > > > > Now, to get this right: I'm not saying that there should be completly no > > OS-specific (but neutral) properties for the clocking system. > > > > I'm saying that in "my" (the VE) case, there are four, effectively > > identical timers (each pair sharing an interrupt though). Every single > > one of them is perfectly capable of serving as clocksource or > > clockevent. Therefore I believe the driver/framework/OS should make its > > own decision. > > > I can move those properties into dts file instead. I'm not sure I understand you here... > If clocksource uses alias to configure source or event. There's still a lot > of work. At least we need to split dual timer into two timer devices in DTS. My point is: at least for VE there is _no need_ for any alias or special property in the tree (DTS). All timers are equal. The system can (and should) pick any one it wants. Pawe?