* Felipe Balbi <balbi@xxxxxx> [151005 18:02]: > > Hi, > > Sebastian Reichel <sre@xxxxxxxxxx> writes: > > On Mon, Oct 05, 2015 at 06:41:37PM -0500, Suman Anna wrote: > >> We will gain a user from OMAP remoteproc driver as well (out of tree at > >> the moment, but it does follow the DT phandle and > >> omap_dm_timer_request_by_node semantics). I do not think ir-rx51.c is > >> converted to DT, and also some of the API are alive just because of the > >> continued OMAP3 legacy boot support. Guess, it is a just a question of > >> not breaking existing API until we kill some of them. > > > > Right, ir-rx51 has not yet been converted to DT. Additionally it > > depends !ARCH_MULTIPLATFORM, since it requires <plat/dmtimer.h>. > > > > I remember a patchset adding a omap-pwm driver using dmtimers. If > > that was added to the kernel, ir-rx51 could probably be switched to > > the PWM API + hrtimer API instead of depending on the dmtimer API. > > At the end not much is gained, though, since then the PWM driver > > would probably require the dmtimer API. > > Tony has been talking about an HRTimer API for some time, but I guess it > didn't happen yet. Until then, cleanups will have to continue to cope > with legacy users. Yeh.. Well meanwhile, here's how we can easily support hardware timers for pwm in a non-intrusive way: 1. We add irqchip support to gptimer code 2. We create omap-gptimer-pwm.c and pass hwtimer_get/set type function pointers in pdata to avoid exposing the gptimer custom functions to drivers until we have a hwtimer framework 3. Then omap-gptimer-pwm.c does request_irq on the hardware timer(s) configured in the board specific dts file to get the right timer 4. Then ir-rx51 can just use he Linux PWM API and omap-gptimer-pwm.c 5. We get rid of the pdata when we have a generic API available Regards, Tony -- 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
