| Message ID | c9f030f57a10c1be00f8732ae2190fcc55248db5.1434622147.git.viresh.kumar@linaro.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 2015-06-18 12:54, Viresh Kumar wrote: > Migrate vf_pit driver to the new 'set-state' interface provided by > clockevents core, the earlier 'set-mode' interface is marked obsolete > now. > > This also enables us to implement callbacks for new states of clockevent > devices, for example: ONESHOT_STOPPED. > > Cc: Jingchang Lu <b35083@freescale.com> > Cc: Stefan Agner <stefan@agner.ch> > Cc: Shawn Guo <shawn.guo@linaro.org> > Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> > --- > drivers/clocksource/vf_pit_timer.c | 27 ++++++++++++--------------- > 1 file changed, 12 insertions(+), 15 deletions(-) > > diff --git a/drivers/clocksource/vf_pit_timer.c > b/drivers/clocksource/vf_pit_timer.c > index b45ac6229b57..f07ba9932171 100644 > --- a/drivers/clocksource/vf_pit_timer.c > +++ b/drivers/clocksource/vf_pit_timer.c > @@ -86,20 +86,16 @@ static int pit_set_next_event(unsigned long delta, > return 0; > } > > -static void pit_set_mode(enum clock_event_mode mode, > - struct clock_event_device *evt) > +static int pit_shutdown(struct clock_event_device *evt) > { > - switch (mode) { > - case CLOCK_EVT_MODE_PERIODIC: > - pit_set_next_event(cycle_per_jiffy, evt); > - break; > - case CLOCK_EVT_MODE_SHUTDOWN: > - case CLOCK_EVT_MODE_UNUSED: > - pit_timer_disable(); > - break; > - default: > - break; > - } > + pit_timer_disable(); > + return 0; > +} > + > +static int pit_set_periodic(struct clock_event_device *evt) > +{ > + pit_set_next_event(cycle_per_jiffy, evt); > + return 0; > } > > static irqreturn_t pit_timer_interrupt(int irq, void *dev_id) > @@ -114,7 +110,7 @@ static irqreturn_t pit_timer_interrupt(int irq, > void *dev_id) > * and start the counter again. So software need to disable the timer > * to stop the counter loop in ONESHOT mode. > */ > - if (likely(evt->mode == CLOCK_EVT_MODE_ONESHOT)) > + if (likely(clockevent_state_oneshot(evt))) > pit_timer_disable(); > > evt->event_handler(evt); > @@ -125,7 +121,8 @@ static irqreturn_t pit_timer_interrupt(int irq, > void *dev_id) > static struct clock_event_device clockevent_pit = { > .name = "VF pit timer", > .features = CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT, > - .set_mode = pit_set_mode, > + .set_state_shutdown = pit_shutdown, > + .set_state_periodic = pit_set_periodic, I'm not really familiar with the interface, but given that we announce the feature CLOCK_EVT_FEAT_ONESHOT shouldn't we add a set_state_oneshot callback here? -- Stefan > .set_next_event = pit_set_next_event, > .rating = 300, > };
On 03-07-15, 10:10, Stefan Agner wrote: > > .features = CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT, > > - .set_mode = pit_set_mode, > > + .set_state_shutdown = pit_shutdown, > > + .set_state_periodic = pit_set_periodic, > > I'm not really familiar with the interface, but given that we announce > the feature CLOCK_EVT_FEAT_ONESHOT shouldn't we add a set_state_oneshot > callback here? We weren't doing anything in pit_set_mode(ONESHOT) and so that callback is not implemented. In case you need to do something in set_state_oneshot(), we can add it back.
On 2015-07-03 10:57, Viresh Kumar wrote: > On 03-07-15, 10:10, Stefan Agner wrote: >> > .features = CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT, >> > - .set_mode = pit_set_mode, >> > + .set_state_shutdown = pit_shutdown, >> > + .set_state_periodic = pit_set_periodic, >> >> I'm not really familiar with the interface, but given that we announce >> the feature CLOCK_EVT_FEAT_ONESHOT shouldn't we add a set_state_oneshot >> callback here? > > We weren't doing anything in pit_set_mode(ONESHOT) and so that > callback is not implemented. In case you need to do something in > set_state_oneshot(), we can add it back. True, weren't doing anything. I wonder if that is right. Afaik, we should set the same timer for oneshot too, hence call pit_set_next_event. With your change we can just reuse the same function (pit_set_periodic) for set_state_oneshot. To maintain the atomicity of the changes, this would need to be fixed in a separate patch anyway. So this change looks good to me: Acked-by: Stefan Agner <stefan@agner.ch> I guess "clockevents: Allow set-state callbacks to be optional" makes it before this patch? Otherwise we would call a null pointer... -- Stefan
On 03-07-15, 13:11, Stefan Agner wrote: > On 2015-07-03 10:57, Viresh Kumar wrote: > > On 03-07-15, 10:10, Stefan Agner wrote: > >> > .features = CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT, > >> > - .set_mode = pit_set_mode, > >> > + .set_state_shutdown = pit_shutdown, > >> > + .set_state_periodic = pit_set_periodic, > >> > >> I'm not really familiar with the interface, but given that we announce > >> the feature CLOCK_EVT_FEAT_ONESHOT shouldn't we add a set_state_oneshot > >> callback here? > > > > We weren't doing anything in pit_set_mode(ONESHOT) and so that > > callback is not implemented. In case you need to do something in > > set_state_oneshot(), we can add it back. > > True, weren't doing anything. I wonder if that is right. Afaik, we > should set the same timer for oneshot too, hence call > pit_set_next_event. With your change we can just reuse the same function > (pit_set_periodic) for set_state_oneshot. pit_set_next_event() will be called by clockevents core directly after tying to set the device in oneshot mode. And so no changes are required. > To maintain the atomicity of the changes, this would need to be fixed in > a separate patch anyway. So this change looks good to me: > > Acked-by: Stefan Agner <stefan@agner.ch> Thanks. > I guess "clockevents: Allow set-state callbacks to be optional" makes it > before this patch? Otherwise we would call a null pointer... Yeah, I have mentioned this in the cover-letter that there are dependencies over clockevent core's next branch.
diff --git a/drivers/clocksource/vf_pit_timer.c b/drivers/clocksource/vf_pit_timer.c index b45ac6229b57..f07ba9932171 100644 --- a/drivers/clocksource/vf_pit_timer.c +++ b/drivers/clocksource/vf_pit_timer.c @@ -86,20 +86,16 @@ static int pit_set_next_event(unsigned long delta, return 0; } -static void pit_set_mode(enum clock_event_mode mode, - struct clock_event_device *evt) +static int pit_shutdown(struct clock_event_device *evt) { - switch (mode) { - case CLOCK_EVT_MODE_PERIODIC: - pit_set_next_event(cycle_per_jiffy, evt); - break; - case CLOCK_EVT_MODE_SHUTDOWN: - case CLOCK_EVT_MODE_UNUSED: - pit_timer_disable(); - break; - default: - break; - } + pit_timer_disable(); + return 0; +} + +static int pit_set_periodic(struct clock_event_device *evt) +{ + pit_set_next_event(cycle_per_jiffy, evt); + return 0; } static irqreturn_t pit_timer_interrupt(int irq, void *dev_id) @@ -114,7 +110,7 @@ static irqreturn_t pit_timer_interrupt(int irq, void *dev_id) * and start the counter again. So software need to disable the timer * to stop the counter loop in ONESHOT mode. */ - if (likely(evt->mode == CLOCK_EVT_MODE_ONESHOT)) + if (likely(clockevent_state_oneshot(evt))) pit_timer_disable(); evt->event_handler(evt); @@ -125,7 +121,8 @@ static irqreturn_t pit_timer_interrupt(int irq, void *dev_id) static struct clock_event_device clockevent_pit = { .name = "VF pit timer", .features = CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT, - .set_mode = pit_set_mode, + .set_state_shutdown = pit_shutdown, + .set_state_periodic = pit_set_periodic, .set_next_event = pit_set_next_event, .rating = 300, };
Migrate vf_pit driver to the new 'set-state' interface provided by clockevents core, the earlier 'set-mode' interface is marked obsolete now. This also enables us to implement callbacks for new states of clockevent devices, for example: ONESHOT_STOPPED. Cc: Jingchang Lu <b35083@freescale.com> Cc: Stefan Agner <stefan@agner.ch> Cc: Shawn Guo <shawn.guo@linaro.org> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> --- drivers/clocksource/vf_pit_timer.c | 27 ++++++++++++--------------- 1 file changed, 12 insertions(+), 15 deletions(-)