Suman Anna <s-anna@xxxxxx> writes: > Hi Felipe, > > On 10/05/2015 02:48 PM, Balbi, Felipe wrote: >> We actually want these devices to be created because >> we will be moving timers to drivers/clocksource and >> this will prevent them from probing. > > This logic is also used to remove the secure timer from being > registered to the kernel on HS devices, while allowing the timer to be > available on GP devices. Your patch actually would break that > functionality. I suggest that you look at the history of the code that heh, that's a silly way of doing so. Could just detect if we're running on HS or not. > originally added this logic - this function seems to be designed to > actually remove the node. The OMAP DMTimer provides an API to request > timers, and I think this logic was used to eliminate giving out the > timers used for clocksource and clockevent when the driver got adapted > to DT. again not the best way of achieving what you want. In any case, other than ir-rx51.c who's using that API ? Seems like we could just pass a phandle to ir-rx51 and be done with it. >> Signed-off-by: Felipe Balbi <balbi@xxxxxx> >> --- >> >> Tony, I wonder if you can get merged as a fix, or do you >> prefer receiving it together with my timer series ? > > NAK for rc, as it breaks other stuff. a single stuff which is likely easy enough to fix. But fair enough -- balbi
Attachment:
signature.asc
Description: PGP signature
