pwm: omap-dmtimer: Allow for dynamic pwm period updates
cover pwm: omap-dmtimer: Allow for dynamic pwm period updates
This series fixes minor issues in config callback and allows for on the fly updates for pwm period and duty cycle. This is mainly intended to allow pwm omap dmtimer to be used to generate a 1PPS signal that can be syncronized to PTP clock in CPTS module in AM335x and AM57xx SoCs.
Series tested after applying the following series:
- https://patchwork.kernel.org/patch/11379875/
- https://patchwork.kernel.org/project/linux-omap/list/?series=246183
Full dependencies can be seen here:
https://github.com/lokeshvutla/linux/tree/devel/pwm-1pps-generation
Lokesh Vutla (4):
pwm: omap-dmtimer: Drop unused header file
pwm: omap-dmtimer: Fix pwm enabling sequence
pwm: omap-dmtimer: Do not disable pwm before changing
period/duty_cycle
pwm: omap-dmtimer: Implement .apply callback
drivers/pwm/pwm-omap-dmtimer.c | 163 +++++++++--------- include/clocksource/timer-ti-dm.h | 3 +- .../linux/platform_data/pwm_omap_dmtimer.h | 90 ---------- 3 files changed, 84 insertions(+), 172 deletions(-) delete mode 100644 include/linux/platform_data/pwm_omap_dmtimer.h
Signature · 1 line
2.23.0
PATCH 1/4 pwm: omap-dmtimer: Drop unused header file
pwm_omap_dmtimer.h is used only to typedef struct omap_dm_timer to pwm_omap_dmtimer. Rest of the file is pretty mush unsed. So reuse omap_dm_timer in pwm-omap-dmtimer.c and delete the header file.
--- drivers/pwm/pwm-omap-dmtimer.c | 16 ++-- include/clocksource/timer-ti-dm.h | 3 +- .../linux/platform_data/pwm_omap_dmtimer.h | 90 ------------------- 3 files changed, 8 insertions(+), 101 deletions(-) delete mode 100644 include/linux/platform_data/pwm_omap_dmtimer.h
drivers/pwm/pwm-omap-dmtimer.c +7 -9
@@ -20,8 +20,8 @@#include <linux/mutex.h> #include <linux/of.h> #include <linux/of_platform.h> +#include <clocksource/timer-ti-dm.h> #include <linux/platform_data/dmtimer-omap.h> -#include <linux/platform_data/pwm_omap_dmtimer.h> #include <linux/platform_device.h> #include <linux/pm_runtime.h> #include <linux/pwm.h>
@@ -34,7 +34,7 @@struct pwm_omap_dmtimer_chip {struct pwm_chip chip;struct mutex mutex; -pwm_omap_dmtimer *dm_timer; +struct omap_dm_timer *dm_timer;const struct omap_dm_timer_ops *pdata;struct platform_device *dm_timer_pdev; };
@@ -190,9 +190,8 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip, load_value, load_value, match_value, match_value); omap->pdata->set_pwm (omap->dm_timer, -pwm_get_polarity (pwm) ==PWM_POLARITY_INVERSED , - true, -PWM_OMAP_DMTIMER_TRIGGER_OVERFLOW_AND_COMPARE ); +pwm_get_polarity (pwm) ==PWM_POLARITY_INVERSED , + true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE );/* If config was called while timer was running it must be reenabled. */ if (timer_active)
@@ -220,9 +219,8 @@ static int pwm_omap_dmtimer_set_polarity(struct pwm_chip *chip,*/ mutex_lock (&omap->mutex); omap->pdata->set_pwm (omap->dm_timer, - polarity ==PWM_POLARITY_INVERSED , - true, -PWM_OMAP_DMTIMER_TRIGGER_OVERFLOW_AND_COMPARE ); + polarity ==PWM_POLARITY_INVERSED , + true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE );mutex_unlock (&omap->mutex);return 0 ;
@@ -244,7 +242,7 @@ static int pwm_omap_dmtimer_probe(struct platform_device *pdev)struct pwm_omap_dmtimer_chip *omap;struct dmtimer_platform_data *timer_pdata;const struct omap_dm_timer_ops *pdata; -pwm_omap_dmtimer *dm_timer; +struct omap_dm_timer *dm_timer;u32 v;int ret =0 ;
include/clocksource/timer-ti-dm.h +1 -2
@@ -248,8 +248,7 @@ int omap_dm_timers_active(void);/* * The below are inlined to optimize code size for system timers. Other code -* should not need these at all, see -* include/linux/platform_data/pwm_omap_dmtimer.h +* should not need these at all. */ #if defined(CONFIG_ARCH_OMAP1 ) || defined(CONFIG_ARCH_OMAP2PLUS ) static inline u32 __omap_dm_timer_read(struct omap_dm_timer *timer, u32 reg,
include/linux/platform_data/pwm_omap_dmtimer.h deleted +0 -90
90 lines not on this page · raw message
Signature · 1 line
2.23.0
Re: [PATCH 1/4] pwm: omap-dmtimer: Drop unused header file
Hello,
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_config(struct pwm_chip *chip, · line 194
Quoted 9 lines · Lokesh Vutla
drivers/pwm/pwm-omap-dmtimer.c +2
-3@@ -190,9 +190,8 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip, load_value, load_value, match_value, match_value); omap->pdata->set_pwm (omap->dm_timer, -pwm_get_polarity (pwm) ==PWM_POLARITY_INVERSED , - true, -PWM_OMAP_DMTIMER_TRIGGER_OVERFLOW_AND_COMPARE ); +pwm_get_polarity (pwm) ==PWM_POLARITY_INVERSED , + true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE );
This is unrelated.
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_set_polarity(struct pwm_chip *chip, · line 223
Quoted 12 lines
drivers/pwm/pwm-omap-dmtimer.c +2
-3/* If config was called while timer was running it must be reenabled. */ if (timer_active)@@ -220,9 +219,8 @@ static int pwm_omap_dmtimer_set_polarity(struct pwm_chip *chip,*/ mutex_lock (&omap->mutex); omap->pdata->set_pwm (omap->dm_timer, - polarity ==PWM_POLARITY_INVERSED , - true, -PWM_OMAP_DMTIMER_TRIGGER_OVERFLOW_AND_COMPARE ); + polarity ==PWM_POLARITY_INVERSED , + true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE );
ditto
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_probe(struct platform_device *pdev) · line 248
Quoted 12 lines
drivers/pwm/pwm-omap-dmtimer.c +1
-1mutex_unlock (&omap->mutex);return 0 ;@@ -244,7 +242,7 @@ static int pwm_omap_dmtimer_probe(struct platform_device *pdev)struct pwm_omap_dmtimer_chip *omap;struct dmtimer_platform_data *timer_pdata;const struct omap_dm_timer_ops *pdata; -pwm_omap_dmtimer *dm_timer; +struct omap_dm_timer *dm_timer;u32 v;int ret =0 ;
Other than that looks fine.
Best regards
Uwe
Signature · 2 lines
Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | https://www.pengutronix.de/ |
Re: [PATCH 1/4] pwm: omap-dmtimer: Drop unused header file
Hi Uwe,
Quoted 6 lines · Uwe Kleine-König
Hello,
Quoted 3 lines · Lokesh Vutla
drivers/pwm/pwm-omap-dmtimer.c +0
-0@@ -190,9 +190,8 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip, load_value, load_value, match_value, match_value);
Quoted 8 lines
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_config(struct pwm_chip *chip, · line 194
Quoted 6 lines
drivers/pwm/pwm-omap-dmtimer.c +2
-3omap->pdata->set_pwm (omap->dm_timer, -pwm_get_polarity (pwm) ==PWM_POLARITY_INVERSED , - true, -PWM_OMAP_DMTIMER_TRIGGER_OVERFLOW_AND_COMPARE ); +pwm_get_polarity (pwm) ==PWM_POLARITY_INVERSED , + true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE );This is unrelated.
PWM_OMAP_DMTIMER_TRIGGER_OVERFLOW_AND_COMPARE is deleted along with header file and used OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE. I should have mentioned this in my commit description. Will fix it in v2.
Thanks and regards,
Lokesh
Re: [PATCH 1/4] pwm: omap-dmtimer: Drop unused header file
Quoted 18 lines · Lokesh Vutla
Hi Uwe,
Quoted 6 lines · Uwe Kleine-König
Hello,
Quoted 3 lines · Lokesh Vutla
drivers/pwm/pwm-omap-dmtimer.c +0
-0@@ -190,9 +190,8 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip, load_value, load_value, match_value, match_value);Quoted 8 lines
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_config(struct pwm_chip *chip, · line 194
Quoted 6 lines
drivers/pwm/pwm-omap-dmtimer.c +2
-3omap->pdata->set_pwm (omap->dm_timer, -pwm_get_polarity (pwm) ==PWM_POLARITY_INVERSED , - true, -PWM_OMAP_DMTIMER_TRIGGER_OVERFLOW_AND_COMPARE ); +pwm_get_polarity (pwm) ==PWM_POLARITY_INVERSED , + true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE );This is unrelated.
Quoted 3 lines
PWM_OMAP_DMTIMER_TRIGGER_OVERFLOW_AND_COMPAREis deleted along with header file and usedOMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE. I should have mentioned this in my commit description. Will fix it in v2.
Ah, indeed I missed that the register name was changed in that hunk and considered that an indention only change.
Best regards
Uwe
Signature · 2 lines
Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | https://www.pengutronix.de/ |
PATCH 2/4 pwm: omap-dmtimer: Fix pwm enabling sequence
In reply to Lokesh VutlaTo configure DM timer is pwm mode the following needs to be set in OMAP_TIMER_CTRL_REG using set_pwm callback:
- Set toggle mode on PORTIMERPWM output pin
- Set trigger on overflow and match on PORTIMERPWM output pin.
- Set auto reload
This is a one time configuration and needs to be set before the start of the dm timer. But the current driver tries to set the same configuration for every period/duty cycle update, which is not needed. So move the pwm setup before enabling timer and do not update it in pwm_omap_dmtimer_config.
--- drivers/pwm/pwm-omap-dmtimer.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
drivers/pwm/pwm-omap-dmtimer.c +4 -4
@@ -73,6 +73,10 @@ static int pwm_omap_dmtimer_enable(struct pwm_chip *chip,struct pwm_omap_dmtimer_chip *omap =to_pwm_omap_dmtimer_chip (chip);mutex_lock (&omap->mutex); + omap->pdata->set_pwm (omap->dm_timer, +pwm_get_polarity (pwm) ==PWM_POLARITY_INVERSED , + true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE ); +pwm_omap_dmtimer_start (omap);mutex_unlock (&omap->mutex);
@@ -189,10 +193,6 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip,dev_dbg (chip->dev,"load value: %#08x (%d), match value: %#08x (%d)\n" , load_value, load_value, match_value, match_value); - omap->pdata->set_pwm (omap->dm_timer, -pwm_get_polarity (pwm) ==PWM_POLARITY_INVERSED , - true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE ); -/* If config was called while timer was running it must be reenabled. */ if (timer_active)pwm_omap_dmtimer_start (omap);
Signature · 1 line
2.23.0
Re: [PATCH 2/4] pwm: omap-dmtimer: Fix pwm enabling sequence
Quoted 10 lines · Lokesh Vutla
To configure DM timer is pwm mode the following needs to be set in
OMAP_TIMER_CTRL_REGusingset_pwmcallback:
- Set toggle mode on PORTIMERPWM output pin
- Set trigger on overflow and match on PORTIMERPWM output pin.
- Set auto reloadThis is a one time configuration and needs to be set before the start of the dm timer. But the current driver tries to set the same configuration for every period/duty cycle update, which is not needed. So move the pwm setup before enabling timer and do not update it in
pwm_omap_dmtimer_config.
Is this change kind of moot with the conversion to .apply in the next patch?
Best regards
Uwe
Signature · 2 lines
Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | https://www.pengutronix.de/ |
Re: [PATCH 2/4] pwm: omap-dmtimer: Fix pwm enabling sequence
Hi Uwe,
Quoted 7 lines · Uwe Kleine-König
Quoted 6 lines · Lokesh Vutla
To configure DM timer is pwm mode the following needs to be set in
OMAP_TIMER_CTRL_REGusingset_pwmcallback:
- Set toggle mode on PORTIMERPWM output pin
- Set trigger on overflow and match on PORTIMERPWM output pin.
- Set auto reload
Quoted 7 lines
Quoted 4 lines
This is a one time configuration and needs to be set before the start of the dm timer. But the current driver tries to set the same configuration for every period/duty cycle update, which is not needed. So move the pwm setup before enabling timer and do not update it in
pwm_omap_dmtimer_config.Is this change kind of moot with the conversion to .apply in the next patch?
Yes, but I didn't want to club it with the conversion to .apply as this is functional change wrt to the existing driver.
Thanks and regards,
Lokesh
PATCH 3/4 pwm: omap-dmtimer: Do not disable pwm before changing period/duty_cycle
In reply to Lokesh VutlaOnly the Timer control register(TCLR) can be updated only when the timer is stopped. Registers like Counter register(TCRR), loader register(TLDR), match register(TMAR) can be updated when the counter is running. Since TCLR is not updated in pwm_omap_dmtimer_config(), do not stop the timer for period/duty_cycle update.
--- drivers/pwm/pwm-omap-dmtimer.c | 14 -------------- 1 file changed, 14 deletions(-)
drivers/pwm/pwm-omap-dmtimer.c +0 -14
@@ -102,7 +102,6 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip,u32 load_value, match_value;struct clk *fclk;unsigned long clk_rate; -bool timer_active;dev_dbg (chip->dev,"requested duty cycle: %d ns, period: %d ns\n" , duty_ns, period_ns);
@@ -178,25 +177,12 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip, load_value = (DM_TIMER_MAX - period_cycles) +1 ; match_value = load_value + duty_cycles -1 ; -/* -* We MUST stop the associated dual-mode timer before attempting to -* write its registers, but calls to omap_dm_timer_start/stop must -* be balanced so check if timer is active before calling timer_stop. -*/ - timer_active =pm_runtime_active (&omap->dm_timer_pdev->dev); -if (timer_active) - omap->pdata->stop (omap->dm_timer); - omap->pdata->set_load (omap->dm_timer, true, load_value); omap->pdata->set_match (omap->dm_timer, true, match_value);dev_dbg (chip->dev,"load value: %#08x (%d), match value: %#08x (%d)\n" , load_value, load_value, match_value, match_value); -/* If config was called while timer was running it must be reenabled. */ -if (timer_active) -pwm_omap_dmtimer_start (omap); -mutex_unlock (&omap->mutex);return 0 ;
Signature · 1 line
2.23.0
Re: [PATCH 3/4] pwm: omap-dmtimer: Do not disable pwm before changing period/duty_cycle
Hello,
Quoted 28 lines · Lokesh Vutla
Only the Timer control register(TCLR) can be updated only when the timer is stopped. Registers like Counter register(TCRR), loader register(TLDR), match register(TMAR) can be updated when the counter is running. Since TCLR is not updated in
pwm_omap_dmtimer_config(), do not stop the timer for period/duty_cycle update.Signed-off-by Lokesh Vutla <[email protected]>--- drivers/pwm/pwm-omap-dmtimer.c | 14 -------------- 1 file changed, 14 deletions(-)drivers/pwm/pwm-omap-dmtimer.c +0
-2@@ -102,7 +102,6 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip,u32 load_value, match_value;struct clk *fclk;unsigned long clk_rate; -bool timer_active;dev_dbg (chip->dev,"requested duty cycle: %d ns, period: %d ns\n" , duty_ns, period_ns);@@ -178,25 +177,12 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip, load_value = (DM_TIMER_MAX - period_cycles) +1 ; match_value = load_value + duty_cycles -1 ; -/*
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_config(struct pwm_chip *chip, · line 181
Quoted 10 lines
drivers/pwm/pwm-omap-dmtimer.c +0
-8-* We MUST stop the associated dual-mode timer before attempting to -* write its registers, but calls to omap_dm_timer_start/stop must -* be balanced so check if timer is active before calling timer_stop. -*/ - timer_active =pm_runtime_active (&omap->dm_timer_pdev->dev); -if (timer_active) - omap->pdata->stop (omap->dm_timer); - omap->pdata->set_load (omap->dm_timer, true, load_value); omap->pdata->set_match (omap->dm_timer, true, match_value);
(Without having looked into the depths of the driver I assume .set_load() sets the period of the PWM and .set_match() the duty cycle.)
What happens on a running PWM if you change the period? Consider you change from duty_cycle = 1000, period = 5000 to duty_cycle = 4000, period = 10000. As you set the period first, can it happen the hardware produces a cycle with duty_cycle = 1000, period = 10000?
Best regards
Uwe
Signature · 2 lines
Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | https://www.pengutronix.de/ |
Re: [PATCH 3/4] pwm: omap-dmtimer: Do not disable pwm before changing period/duty_cycle
Hi Uwe,
Quoted 30 lines · Uwe Kleine-König
Hello,
Quoted 27 lines · Lokesh Vutla
Only the Timer control register(TCLR) can be updated only when the timer is stopped. Registers like Counter register(TCRR), loader register(TLDR), match register(TMAR) can be updated when the counter is running. Since TCLR is not updated in
pwm_omap_dmtimer_config(), do not stop the timer for period/duty_cycle update.Signed-off-by Lokesh Vutla <[email protected]>--- drivers/pwm/pwm-omap-dmtimer.c | 14 -------------- 1 file changed, 14 deletions(-)drivers/pwm/pwm-omap-dmtimer.c +0
-1@@ -102,7 +102,6 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip,u32 load_value, match_value;struct clk *fclk;unsigned long clk_rate; -bool timer_active;dev_dbg (chip->dev,"requested duty cycle: %d ns, period: %d ns\n" , duty_ns, period_ns);@@ -178,25 +177,12 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip, load_value = (DM_TIMER_MAX - period_cycles) +1 ; match_value = load_value + duty_cycles -1 ;
Quoted 14 lines
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_config(struct pwm_chip *chip, · line 181
Quoted 11 lines
drivers/pwm/pwm-omap-dmtimer.c +0
-9-/* -* We MUST stop the associated dual-mode timer before attempting to -* write its registers, but calls to omap_dm_timer_start/stop must -* be balanced so check if timer is active before calling timer_stop. -*/ - timer_active =pm_runtime_active (&omap->dm_timer_pdev->dev); -if (timer_active) - omap->pdata->stop (omap->dm_timer); - omap->pdata->set_load (omap->dm_timer, true, load_value); omap->pdata->set_match (omap->dm_timer, true, match_value);(Without having looked into the depths of the driver I assume .set_load() sets the period of the PWM and .set_match() the duty cycle.)
Right.
Quoted 5 lines
What happens on a running PWM if you change the period? Consider you change from
duty_cycle= 1000, period = 5000 toduty_cycle= 4000, period = 10000. As you set the period first, can it happen the hardware produces a cycle withduty_cycle= 1000, period = 10000?
No. So, the current cycle is un affected with duty_cycle = 1000 and period = 5000. Starting from next cycle new settings gets reflected with duty_cycle = 4000 and period = 10000.
Thanks and regards,
Lokesh
Re: [PATCH 3/4] pwm: omap-dmtimer: Do not disable pwm before changing period/duty_cycle
Hello,
Quoted 49 lines · Lokesh Vutla
Quoted 30 lines · Uwe Kleine-König
Hello,
Quoted 27 lines · Lokesh Vutla
Only the Timer control register(TCLR) can be updated only when the timer is stopped. Registers like Counter register(TCRR), loader register(TLDR), match register(TMAR) can be updated when the counter is running. Since TCLR is not updated in
pwm_omap_dmtimer_config(), do not stop the timer for period/duty_cycle update.Signed-off-by Lokesh Vutla <[email protected]>--- drivers/pwm/pwm-omap-dmtimer.c | 14 -------------- 1 file changed, 14 deletions(-)drivers/pwm/pwm-omap-dmtimer.c +0
-1@@ -102,7 +102,6 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip,u32 load_value, match_value;struct clk *fclk;unsigned long clk_rate; -bool timer_active;dev_dbg (chip->dev,"requested duty cycle: %d ns, period: %d ns\n" , duty_ns, period_ns);@@ -178,25 +177,12 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip, load_value = (DM_TIMER_MAX - period_cycles) +1 ; match_value = load_value + duty_cycles -1 ;Quoted 14 lines
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_config(struct pwm_chip *chip, · line 181
Quoted 11 lines
drivers/pwm/pwm-omap-dmtimer.c +0
-9-/* -* We MUST stop the associated dual-mode timer before attempting to -* write its registers, but calls to omap_dm_timer_start/stop must -* be balanced so check if timer is active before calling timer_stop. -*/ - timer_active =pm_runtime_active (&omap->dm_timer_pdev->dev); -if (timer_active) - omap->pdata->stop (omap->dm_timer); - omap->pdata->set_load (omap->dm_timer, true, load_value); omap->pdata->set_match (omap->dm_timer, true, match_value);(Without having looked into the depths of the driver I assume .set_load() sets the period of the PWM and .set_match() the duty cycle.)
Right.
Quoted 1 line
Quoted 8 lines
Quoted 4 lines
What happens on a running PWM if you change the period? Consider you change from
duty_cycle= 1000, period = 5000 toduty_cycle= 4000, period = 10000. As you set the period first, can it happen the hardware produces a cycle withduty_cycle= 1000, period = 10000?No. So, the current cycle is un affected with
duty_cycle= 1000 and period = 5000. Starting from next cycle new settings gets reflected withduty_cycle= 4000 and period = 10000.
Is the reference manual for this hardware publically available?
So the .set_load callback just writes a shadow register and .set_match latches it into hardware atomically with its own register changes? A comment in the source code about this would be good. Also if .set_load doesn't work without .set_match I wonder if it is sane to put their logic in two different functions.
Best regards
Uwe
Signature · 2 lines
Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | https://www.pengutronix.de/ |
Re: [PATCH 3/4] pwm: omap-dmtimer: Do not disable pwm before changing period/duty_cycle
Hi Uwe,
Quoted 57 lines · Uwe Kleine-König
Hello,
Quoted 54 lines · Lokesh Vutla
Quoted 30 lines · Uwe Kleine-König
Hello,
Quoted 27 lines · Lokesh Vutla
Only the Timer control register(TCLR) can be updated only when the timer is stopped. Registers like Counter register(TCRR), loader register(TLDR), match register(TMAR) can be updated when the counter is running. Since TCLR is not updated in
pwm_omap_dmtimer_config(), do not stop the timer for period/duty_cycle update.Signed-off-by Lokesh Vutla <[email protected]>--- drivers/pwm/pwm-omap-dmtimer.c | 14 -------------- 1 file changed, 14 deletions(-)drivers/pwm/pwm-omap-dmtimer.c +0
-1@@ -102,7 +102,6 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip,u32 load_value, match_value;struct clk *fclk;unsigned long clk_rate; -bool timer_active;dev_dbg (chip->dev,"requested duty cycle: %d ns, period: %d ns\n" , duty_ns, period_ns);@@ -178,25 +177,12 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip, load_value = (DM_TIMER_MAX - period_cycles) +1 ; match_value = load_value + duty_cycles -1 ;Quoted 14 lines
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_config(struct pwm_chip *chip, · line 181
Quoted 11 lines
drivers/pwm/pwm-omap-dmtimer.c +0
-9-/* -* We MUST stop the associated dual-mode timer before attempting to -* write its registers, but calls to omap_dm_timer_start/stop must -* be balanced so check if timer is active before calling timer_stop. -*/ - timer_active =pm_runtime_active (&omap->dm_timer_pdev->dev); -if (timer_active) - omap->pdata->stop (omap->dm_timer); - omap->pdata->set_load (omap->dm_timer, true, load_value); omap->pdata->set_match (omap->dm_timer, true, match_value);(Without having looked into the depths of the driver I assume .set_load() sets the period of the PWM and .set_match() the duty cycle.)
Right.
Quoted 5 lines
What happens on a running PWM if you change the period? Consider you change from
duty_cycle= 1000, period = 5000 toduty_cycle= 4000, period = 10000. As you set the period first, can it happen the hardware produces a cycle withduty_cycle= 1000, period = 10000?
Quoted 5 lines
Quoted 3 lines
No. So, the current cycle is un affected with
duty_cycle= 1000 and period = 5000. Starting from next cycle new settings gets reflected withduty_cycle= 4000 and period = 10000.Is the reference manual for this hardware publically available?
AM335x TRM [0] Section 20.1.3.5 Pulse-Width Modulation (Page 4445).
[0] http://www.ti.com/lit/ug/spruh73q/spruh73q.pdf
Quoted 6 lines
So the .set_load callback just writes a shadow register and .set_match latches it into hardware atomically with its own register changes? A comment in the source code about this would be good. Also if .set_load doesn't work without .set_match I wonder if it is sane to put their logic in two different functions.
Just to give a little bit of background:
- The omap timer is an upward counter that can be started and stopped at any time.
- Once the timer counter overflows, it gets loaded with a predefined load value.(Or can be configured to not re load at all).
- Timer has a configurable output pin which can be toggled in the following two cases:
- When the counter overflows
- When the counter matches with a predefined register(match register).
Using this o/p pin the driver tries to generate a PWM with period = (OVERFLOW - LOAD_VALUE) and duty_cycle = (MATCH_VALUE - LOAD_VALUE).
.set_load will configure the load value .set_match will configure the match value. The configured values gets effected only in the next cycle of PWM.
I hope this is clear. Let me know if you need more details.
Thanks and regards,
Lokesh
Re: [PATCH 3/4] pwm: omap-dmtimer: Do not disable pwm before changing period/duty_cycle
Hello Lokesh,
Quoted 22 lines · Lokesh Vutla
Quoted 21 lines · Uwe Kleine-König
Quoted 19 lines · Lokesh Vutla
Quoted 6 lines · Uwe Kleine-König
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_config(struct pwm_chip *chip, · line 181
Quoted 2 lines · Lokesh Vutla
drivers/pwm/pwm-omap-dmtimer.c +0
-0omap->pdata->set_load (omap->dm_timer, true, load_value); omap->pdata->set_match (omap->dm_timer, true, match_value);(Without having looked into the depths of the driver I assume .set_load() sets the period of the PWM and .set_match() the duty cycle.)
Right.
Quoted 5 lines
What happens on a running PWM if you change the period? Consider you change from
duty_cycle= 1000, period = 5000 toduty_cycle= 4000, period = 10000. As you set the period first, can it happen the hardware produces a cycle withduty_cycle= 1000, period = 10000?No. So, the current cycle is un affected with
duty_cycle= 1000 and period = 5000. Starting from next cycle new settings gets reflected withduty_cycle= 4000 and period = 10000.
Quoted 5 lines
Quoted 1 line
Is the reference manual for this hardware publically available?
AM335x TRM [0] Section 20.1.3.5 Pulse-Width Modulation (Page 4445).
Great. This is BTW an opportunity to increase your patch count: Create a patch that adds a reference to this document at the top of the driver.
Quoted 7 lines
Quoted 5 lines
So the .set_load callback just writes a shadow register and .set_match latches it into hardware atomically with its own register changes? A comment in the source code about this would be good. Also if .set_load doesn't work without .set_match I wonder if it is sane to put their logic in two different functions.
Just to give a little bit of background:
Thanks, very appreciated.
Quoted 8 lines
- The omap timer is an upward counter that can be started and stopped at any time.
- Once the timer counter overflows, it gets loaded with a predefined load value.(Or can be configured to not re load at all).
- Timer has a configurable output pin which can be toggled in the following two cases:
- When the counter overflows
- When the counter matches with a predefined register(match register).
Quoted 5 lines
Using this o/p pin the driver tries to generate a PWM with period = (OVERFLOW -
LOAD_VALUE) andduty_cycle= (MATCH_VALUE-LOAD_VALUE)..set_load will configure the load value .set_match will configure the match value. The configured values gets effected only in the next cycle of PWM.
Ah, so back to my original question: If you change from duty_cycle/period = 1000/5000 to duty_cycle/period = 4000/10000 and after you set the period but before you set the duty_cycle a period happens to end, you get indeed a cycle with mixed settings, right?
Best regards
Uwe
Signature · 2 lines
Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | https://www.pengutronix.de/ |
Re: [PATCH 3/4] pwm: omap-dmtimer: Do not disable pwm before changing period/duty_cycle
Hi Uwe,
Quoted 55 lines · Uwe Kleine-König
Hello Lokesh,
Quoted 22 lines · Lokesh Vutla
Quoted 21 lines · Uwe Kleine-König
Quoted 19 lines · Lokesh Vutla
Quoted 6 lines · Uwe Kleine-König
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_config(struct pwm_chip *chip, · line 181
Quoted 2 lines · Lokesh Vutla
drivers/pwm/pwm-omap-dmtimer.c +0
-0omap->pdata->set_load (omap->dm_timer, true, load_value); omap->pdata->set_match (omap->dm_timer, true, match_value);(Without having looked into the depths of the driver I assume .set_load() sets the period of the PWM and .set_match() the duty cycle.)
Right.
Quoted 5 lines
What happens on a running PWM if you change the period? Consider you change from
duty_cycle= 1000, period = 5000 toduty_cycle= 4000, period = 10000. As you set the period first, can it happen the hardware produces a cycle withduty_cycle= 1000, period = 10000?No. So, the current cycle is un affected with
duty_cycle= 1000 and period = 5000. Starting from next cycle new settings gets reflected withduty_cycle= 4000 and period = 10000.Quoted 5 lines
Quoted 1 line
Is the reference manual for this hardware publically available?
AM335x TRM [0] Section 20.1.3.5 Pulse-Width Modulation (Page 4445).
Great. This is BTW an opportunity to increase your patch count: Create a patch that adds a reference to this document at the top of the driver.
Quoted 7 lines
Quoted 5 lines
So the .set_load callback just writes a shadow register and .set_match latches it into hardware atomically with its own register changes? A comment in the source code about this would be good. Also if .set_load doesn't work without .set_match I wonder if it is sane to put their logic in two different functions.
Just to give a little bit of background:
Thanks, very appreciated.
Quoted 11 lines
- The omap timer is an upward counter that can be started and stopped at any time.
- Once the timer counter overflows, it gets loaded with a predefined load value.(Or can be configured to not re load at all).
- Timer has a configurable output pin which can be toggled in the following two cases:
- When the counter overflows
- When the counter matches with a predefined register(match register).Using this o/p pin the driver tries to generate a PWM with period = (OVERFLOW -
LOAD_VALUE) andduty_cycle= (MATCH_VALUE-LOAD_VALUE).
Quoted 7 lines
Quoted 2 lines
.set_load will configure the load value .set_match will configure the match value. The configured values gets effected only in the next cycle of PWM.
Ah, so back to my original question: If you change from duty_cycle/period = 1000/5000 to duty_cycle/period = 4000/10000 and after you set the period but before you set the
duty_cyclea period happens to end, you get indeed a cycle with mixed settings, right?
hmm..you are right but the mixed period happens in a bit different case. Let me explain in bit more detail.
For omap dm timer, the load_value that gets set in the current period, will be reflected only in next cycle, as timer counter has to overflow to load this value. But in case of match register(which determines the duty cycle), the timer counter is continuously matched to it. So below are the cases where a mixed period can happen:
1) When signal is high and new match value is > current timer counter. Then the duty cycle gets reflected in the current cycle.(Duty_cycle for current period= new match value - previous load value).
2) When signal is high and new match value is < current timer counter. Then the period and duty cycle for the current cycle gets effected as well. Because the signal is pulled down only when counter matches with match register, and this happens only in the next cycle(after timer counter overflows). Then:
- new Period for current cycle = (current period + new period)
- new duty cycle for current cycle = (current period + new duty_cycle).
I am able to observe the above mentioned 2 behaviors on the scope using beagle bone black. So the problem is with updating duty cycle when the signal is high. but when signal is low, nothing gets effected to the current cycle.
How do you want to go about this? Should we describe this as limitation in the driver as you asked?
Thanks and regards,
Lokesh
Re: [PATCH 3/4] pwm: omap-dmtimer: Do not disable pwm before changing period/duty_cycle
Quoted 84 lines · Lokesh Vutla
Hi Uwe,
Quoted 55 lines · Uwe Kleine-König
Hello Lokesh,
Quoted 22 lines · Lokesh Vutla
Quoted 21 lines · Uwe Kleine-König
Quoted 19 lines · Lokesh Vutla
Quoted 6 lines · Uwe Kleine-König
Review
Quoted 2 lines · Lokesh Vutla
(quoted diff) +0
-0omap->pdata->set_load(omap->dm_timer, true, load_value); omap->pdata->set_match(omap->dm_timer, true, match_value);(Without having looked into the depths of the driver I assume .set_load() sets the period of the PWM and .set_match() the duty cycle.)
Right.
Quoted 5 lines
What happens on a running PWM if you change the period? Consider you change from
duty_cycle= 1000, period = 5000 toduty_cycle= 4000, period = 10000. As you set the period first, can it happen the hardware produces a cycle withduty_cycle= 1000, period = 10000?No. So, the current cycle is un affected with
duty_cycle= 1000 and period = 5000. Starting from next cycle new settings gets reflected withduty_cycle= 4000 and period = 10000.Quoted 5 lines
Quoted 1 line
Is the reference manual for this hardware publically available?
AM335x TRM [0] Section 20.1.3.5 Pulse-Width Modulation (Page 4445).
Great. This is BTW an opportunity to increase your patch count: Create a patch that adds a reference to this document at the top of the driver.
Quoted 7 lines
Quoted 5 lines
So the .set_load callback just writes a shadow register and .set_match latches it into hardware atomically with its own register changes? A comment in the source code about this would be good. Also if .set_load doesn't work without .set_match I wonder if it is sane to put their logic in two different functions.
Just to give a little bit of background:
Thanks, very appreciated.
Quoted 11 lines
- The omap timer is an upward counter that can be started and stopped at any time.
- Once the timer counter overflows, it gets loaded with a predefined load value.(Or can be configured to not re load at all).
- Timer has a configurable output pin which can be toggled in the following two cases:
- When the counter overflows
- When the counter matches with a predefined register(match register).Using this o/p pin the driver tries to generate a PWM with period = (OVERFLOW -
LOAD_VALUE) andduty_cycle= (MATCH_VALUE-LOAD_VALUE).Quoted 7 lines
Quoted 2 lines
.set_load will configure the load value .set_match will configure the match value. The configured values gets effected only in the next cycle of PWM.
Ah, so back to my original question: If you change from duty_cycle/period = 1000/5000 to duty_cycle/period = 4000/10000 and after you set the period but before you set the
duty_cyclea period happens to end, you get indeed a cycle with mixed settings, right?hmm..you are right but the mixed period happens in a bit different case. Let me explain in bit more detail.
For omap dm timer, the
load_valuethat gets set in the current period, will be reflected only in next cycle, as timer counter has to overflow to load this value. But in case of match register(which determines the duty cycle), the timer counter is continuously matched to it. So below are the cases where a mixed period can happen:
1) When signal is high and new match value is > current timer counter. Then the duty cycle gets reflected in the current cycle.(Duty_cycle for current period= new match value - previous load value).
2) When signal is high and new match value is < current timer counter. Then the period and duty cycle for the current cycle gets effected as well. Because the signal is pulled down only when counter matches with match register, and this happens only in the next cycle(after timer counter overflows). Then:
- new Period for current cycle = (current period + new period)
- new duty cycle for current cycle = (current period + newduty_cycle).
Quoted 6 lines
I am able to observe the above mentioned 2 behaviors on the scope using beagle bone black. So the problem is with updating duty cycle when the signal is high. but when signal is low, nothing gets effected to the current cycle.
How do you want to go about this? Should we describe this as limitation in the driver as you asked?
OK, to sumarize: You have a counter that goes up. When it overflows it gets reloaded with the load value and the output goes up. When $counter = $match, the output goes down.
Having this is a comment at the top of the driver would be very welcome.
(I just noticed I duplicated this question about a racy update in another end of this thread. This pops in my head too automatically :-)
Best regards
Uwe
Signature · 2 lines
Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | https://www.pengutronix.de/ |
PATCH 4/4 pwm: omap-dmtimer: Implement .apply callback
In reply to Lokesh VutlaImplement .apply callback and drop the legacy callbacks(enable, disable, config, set_polarity).
--- drivers/pwm/pwm-omap-dmtimer.c | 141 +++++++++++++++++++-------------- 1 file changed, 80 insertions(+), 61 deletions(-)
drivers/pwm/pwm-omap-dmtimer.c +80 -61
@@ -31,8 +31,18 @@#define DM_TIMER_LOAD_MIN 0xfffffffe#define DM_TIMER_MAX 0xffffffff +/** +* struct pwm_omap_dmtimer_chip - Structure representing a pwm chip +* corresponding to omap dmtimer. +* @chip: PWM chip structure representing PWM controller +* @mutex: Mutex to protect pwm apply state +* @dm_timer: Pointer to omap dm timer. +* @pdata: Pointer to omap dm timer ops. +* dm_timer_pdev: Pointer to omap dm timer platform device +*/ struct pwm_omap_dmtimer_chip {struct pwm_chip chip; +/* Mutex to protect pwm apply state */ struct mutex mutex;struct omap_dm_timer *dm_timer;const struct omap_dm_timer_ops *pdata;
@@ -45,11 +55,22 @@ to_pwm_omap_dmtimer_chip(struct pwm_chip *chip) return container_of(chip, struct pwm_omap_dmtimer_chip, chip); } +/** +* pwm_omap_dmtimer_get_clock_cycles() - Get clock cycles in a time frame +* @clk_rate: pwm timer clock rate +* @ns: time frame in nano seconds. +* +* Return number of clock cycles in a given period(ins ns). +*/ static u32 pwm_omap_dmtimer_get_clock_cycles (unsigned long clk_rate,int ns) {return DIV_ROUND_CLOSEST_ULL ((u64 )clk_rate * ns,NSEC_PER_SEC ); } +/** +* pwm_omap_dmtimer_start() - Start the pwm omap dm timer in pwm mode +* @omap: Pointer to pwm omap dm timer chip +*/ static void pwm_omap_dmtimer_start (struct pwm_omap_dmtimer_chip *omap) {/*
@@ -67,32 +88,16 @@ static void pwm_omap_dmtimer_start(struct pwm_omap_dmtimer_chip *omap) omap->pdata->start (omap->dm_timer); } -static int pwm_omap_dmtimer_enable (struct pwm_chip *chip, -struct pwm_device *pwm) -{ -struct pwm_omap_dmtimer_chip *omap =to_pwm_omap_dmtimer_chip (chip); - -mutex_lock (&omap->mutex); - omap->pdata->set_pwm (omap->dm_timer, -pwm_get_polarity (pwm) ==PWM_POLARITY_INVERSED , - true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE ); - -pwm_omap_dmtimer_start (omap); -mutex_unlock (&omap->mutex); - -return 0 ; -} - -static void pwm_omap_dmtimer_disable (struct pwm_chip *chip, -struct pwm_device *pwm) -{ -struct pwm_omap_dmtimer_chip *omap =to_pwm_omap_dmtimer_chip (chip); - -mutex_lock (&omap->mutex); - omap->pdata->stop (omap->dm_timer); -mutex_unlock (&omap->mutex); -} - +/** +* pwm_omap_dmtimer_config() - Update the configuration of pwm omap dm timer +* @chip: Pointer to PWM controller +* @pwm: Pointer to PWM channel +* @duty_ns: New duty cycle in nano seconds +* @period_ns: New period in nano seconds +* +* Return 0 if successfully changed the period/duty_cycle else appropriate +* error. +*/ static int pwm_omap_dmtimer_config (struct pwm_chip *chip,struct pwm_device *pwm,int duty_ns,int period_ns)
@@ -100,30 +105,26 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip,struct pwm_omap_dmtimer_chip *omap =to_pwm_omap_dmtimer_chip (chip);u32 period_cycles, duty_cycles;u32 load_value, match_value; -struct clk *fclk;unsigned long clk_rate; +struct clk *fclk;dev_dbg (chip->dev,"requested duty cycle: %d ns, period: %d ns\n" , duty_ns, period_ns); -mutex_lock (&omap->mutex);if (duty_ns ==pwm_get_duty_cycle (pwm) && - period_ns ==pwm_get_period (pwm)) { -/* No change - don't cause any transients. */ -mutex_unlock (&omap->mutex); + period_ns ==pwm_get_period (pwm))return 0 ; - } fclk = omap->pdata->get_fclk (omap->dm_timer);if (!fclk) {dev_err (chip->dev,"invalid pmtimer fclk\n" ); - goto err_einval; +return -EINVAL ; } clk_rate =clk_get_rate (fclk);if (!clk_rate) {dev_err (chip->dev,"invalid pmtimer fclk rate\n" ); - goto err_einval; +return -EINVAL ; }dev_dbg (chip->dev,"clk rate: %luHz\n" , clk_rate);
@@ -151,7 +152,7 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip,dev_info (chip->dev,"period %d ns too short for clock rate %lu Hz\n" , period_ns, clk_rate); - goto err_einval; +return -EINVAL ; }if (duty_cycles <1 ) {
@@ -183,54 +184,72 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip,dev_dbg (chip->dev,"load value: %#08x (%d), match value: %#08x (%d)\n" , load_value, load_value, match_value, match_value); -mutex_unlock (&omap->mutex); -return 0 ; - -err_einval: -mutex_unlock (&omap->mutex); - -return -EINVAL ; } -static int pwm_omap_dmtimer_set_polarity (struct pwm_chip *chip, -struct pwm_device *pwm, -enum pwm_polarity polarity) +/** +* pwm_omap_dmtimer_apply() - Changes the state of the pwm omap dm timer. +* @chip: Pointer to PWM controller +* @pwm: Pointer to PWM channel +* @state: New sate to apply +* +* Return 0 if successfully changed the state else appropriate error. +*/ +static int pwm_omap_dmtimer_apply (struct pwm_chip *chip, +struct pwm_device *pwm, +const struct pwm_state *state) {struct pwm_omap_dmtimer_chip *omap =to_pwm_omap_dmtimer_chip (chip); +int ret =0 ; -/* -* PWM core will not call set_polarity while PWM is enabled so it's -* safe to reconfigure the timer here without stopping it first. -*/ mutex_lock (&omap->mutex); - omap->pdata->set_pwm (omap->dm_timer, - polarity ==PWM_POLARITY_INVERSED , - true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE ); + +if (pwm_is_enabled (pwm) && !state->enabled) { + omap->pdata->stop (omap->dm_timer); + goto unlock_mutex; + } + +if (pwm_get_polarity (pwm) != state->polarity) + omap->pdata->set_pwm (omap->dm_timer, + state->polarity ==PWM_POLARITY_INVERSED , + true, +OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE ); + + ret =pwm_omap_dmtimer_config (chip, pwm, state->duty_cycle, + state->period); +if (ret) + goto unlock_mutex; + +if (!pwm_is_enabled (pwm) && state->enabled) { + omap->pdata->set_pwm (omap->dm_timer, + state->polarity ==PWM_POLARITY_INVERSED , + true, +OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE ); +pwm_omap_dmtimer_start (omap); + } + +unlock_mutex:mutex_unlock (&omap->mutex); -return 0 ; +return ret; }static const struct pwm_ops pwm_omap_dmtimer_ops = { - .enable = pwm_omap_dmtimer_enable, - .disable = pwm_omap_dmtimer_disable, - .config = pwm_omap_dmtimer_config, - .set_polarity = pwm_omap_dmtimer_set_polarity, + .apply = pwm_omap_dmtimer_apply, .owner =THIS_MODULE , };static int pwm_omap_dmtimer_probe (struct platform_device *pdev) {struct device_node *np = pdev->dev.of_node; -struct device_node *timer; -struct platform_device *timer_pdev; -struct pwm_omap_dmtimer_chip *omap;struct dmtimer_platform_data *timer_pdata;const struct omap_dm_timer_ops *pdata; +struct platform_device *timer_pdev; +struct pwm_omap_dmtimer_chip *omap;struct omap_dm_timer *dm_timer; -u32 v; +struct device_node *timer;int ret =0 ; +u32 v; timer =of_parse_phandle (np,"ti,timers" ,0 );if (!timer)
Signature · 1 line
2.23.0
Re: [PATCH 4/4] pwm: omap-dmtimer: Implement .apply callback
Quoted 173 lines · Lokesh Vutla
Implement .apply callback and drop the legacy
callbacks(enable, disable, config, set_polarity).Signed-off-by Lokesh Vutla <[email protected]>--- drivers/pwm/pwm-omap-dmtimer.c | 141 +++++++++++++++++++-------------- 1 file changed, 80 insertions(+), 61 deletions(-)drivers/pwm/pwm-omap-dmtimer.c +48
-45@@ -31,8 +31,18 @@#define DM_TIMER_LOAD_MIN 0xfffffffe#define DM_TIMER_MAX 0xffffffff +/** +* struct pwm_omap_dmtimer_chip - Structure representing a pwm chip +* corresponding to omap dmtimer. +* @chip: PWM chip structure representing PWM controller +* @mutex: Mutex to protect pwm apply state +* @dm_timer: Pointer to omap dm timer. +* @pdata: Pointer to omap dm timer ops. +* dm_timer_pdev: Pointer to omap dm timer platform device +*/ struct pwm_omap_dmtimer_chip {struct pwm_chip chip; +/* Mutex to protect pwm apply state */ struct mutex mutex;struct omap_dm_timer *dm_timer;const struct omap_dm_timer_ops *pdata;@@ -45,11 +55,22 @@ to_pwm_omap_dmtimer_chip(struct pwm_chip *chip) return container_of(chip, struct pwm_omap_dmtimer_chip, chip); } +/** +* pwm_omap_dmtimer_get_clock_cycles() - Get clock cycles in a time frame +* @clk_rate: pwm timer clock rate +* @ns: time frame in nano seconds. +* +* Return number of clock cycles in a given period(ins ns). +*/ static u32 pwm_omap_dmtimer_get_clock_cycles (unsigned long clk_rate,int ns) {return DIV_ROUND_CLOSEST_ULL ((u64 )clk_rate * ns,NSEC_PER_SEC ); } +/** +* pwm_omap_dmtimer_start() - Start the pwm omap dm timer in pwm mode +* @omap: Pointer to pwm omap dm timer chip +*/ static void pwm_omap_dmtimer_start (struct pwm_omap_dmtimer_chip *omap) {/* @@ -67,32 +88,16 @@ static void pwm_omap_dmtimer_start(struct pwm_omap_dmtimer_chip *omap) omap->pdata->start (omap->dm_timer); } -static int pwm_omap_dmtimer_enable (struct pwm_chip *chip, -struct pwm_device *pwm) -{ -struct pwm_omap_dmtimer_chip *omap =to_pwm_omap_dmtimer_chip (chip); - -mutex_lock (&omap->mutex); - omap->pdata->set_pwm (omap->dm_timer, -pwm_get_polarity (pwm) ==PWM_POLARITY_INVERSED , - true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE ); - -pwm_omap_dmtimer_start (omap); -mutex_unlock (&omap->mutex); - -return 0 ; -} - -static void pwm_omap_dmtimer_disable (struct pwm_chip *chip, -struct pwm_device *pwm) -{ -struct pwm_omap_dmtimer_chip *omap =to_pwm_omap_dmtimer_chip (chip); - -mutex_lock (&omap->mutex); - omap->pdata->stop (omap->dm_timer); -mutex_unlock (&omap->mutex); -} - +/** +* pwm_omap_dmtimer_config() - Update the configuration of pwm omap dm timer +* @chip: Pointer to PWM controller +* @pwm: Pointer to PWM channel +* @duty_ns: New duty cycle in nano seconds +* @period_ns: New period in nano seconds +* +* Return 0 if successfully changed the period/duty_cycle else appropriate +* error. +*/ static int pwm_omap_dmtimer_config (struct pwm_chip *chip,struct pwm_device *pwm,int duty_ns,int period_ns)@@ -100,30 +105,26 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip,struct pwm_omap_dmtimer_chip *omap =to_pwm_omap_dmtimer_chip (chip);u32 period_cycles, duty_cycles;u32 load_value, match_value; -struct clk *fclk;unsigned long clk_rate; +struct clk *fclk;dev_dbg (chip->dev,"requested duty cycle: %d ns, period: %d ns\n" , duty_ns, period_ns); -mutex_lock (&omap->mutex);if (duty_ns ==pwm_get_duty_cycle (pwm) && - period_ns ==pwm_get_period (pwm)) { -/* No change - don't cause any transients. */ -mutex_unlock (&omap->mutex); + period_ns ==pwm_get_period (pwm))return 0 ; - } fclk = omap->pdata->get_fclk (omap->dm_timer);if (!fclk) {dev_err (chip->dev,"invalid pmtimer fclk\n" ); - goto err_einval; +return -EINVAL ; } clk_rate =clk_get_rate (fclk);if (!clk_rate) {dev_err (chip->dev,"invalid pmtimer fclk rate\n" ); - goto err_einval; +return -EINVAL ; }dev_dbg (chip->dev,"clk rate: %luHz\n" , clk_rate);@@ -151,7 +152,7 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip,dev_info (chip->dev,"period %d ns too short for clock rate %lu Hz\n" , period_ns, clk_rate); - goto err_einval; +return -EINVAL ; }if (duty_cycles <1 ) {@@ -183,54 +184,72 @@ static int pwm_omap_dmtimer_config(struct pwm_chip *chip,dev_dbg (chip->dev,"load value: %#08x (%d), match value: %#08x (%d)\n" , load_value, load_value, match_value, match_value); -mutex_unlock (&omap->mutex); -return 0 ; - -err_einval: -mutex_unlock (&omap->mutex); - -return -EINVAL ; } -static int pwm_omap_dmtimer_set_polarity (struct pwm_chip *chip, -struct pwm_device *pwm, -enum pwm_polarity polarity) +/** +* pwm_omap_dmtimer_apply() - Changes the state of the pwm omap dm timer. +* @chip: Pointer to PWM controller +* @pwm: Pointer to PWM channel +* @state: New sate to apply +* +* Return 0 if successfully changed the state else appropriate error. +*/ +static int pwm_omap_dmtimer_apply (struct pwm_chip *chip, +struct pwm_device *pwm, +const struct pwm_state *state) {struct pwm_omap_dmtimer_chip *omap =to_pwm_omap_dmtimer_chip (chip); +int ret =0 ;
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_config(struct pwm_chip *chip, · line 207
Quoted 10 lines
drivers/pwm/pwm-omap-dmtimer.c +2
-7-/* -* PWM core will not call set_polarity while PWM is enabled so it's -* safe to reconfigure the timer here without stopping it first. -*/ mutex_lock (&omap->mutex); - omap->pdata->set_pwm (omap->dm_timer, - polarity ==PWM_POLARITY_INVERSED , - true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE ); + +if (pwm_is_enabled (pwm) && !state->enabled) {
In my book calling PWM API functions (designed for PWM consumers) is not so nice. I would prefer you checking the hardware registers or cache the state locally instead of relying on the core here.
It would be great to have a general description at the top of the driver (like for example drivers/pwm/pwm-sifive.c) that answers things like:
- Does calling .stop completes the currently running period (it should)?
- Does changing polarity, duty_cycle and period complete the running period?
- How does the hardware behave on disable? (i.e. does it output the state the pin is at in that moment? Does it go High-Z?)
Best regards
Uwe
Signature · 2 lines
Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | https://www.pengutronix.de/ |
Re: [PATCH 4/4] pwm: omap-dmtimer: Implement .apply callback
Hi Uwe,
Quoted 9 lines · Uwe Kleine-König
Quoted 8 lines · Lokesh Vutla
Implement .apply callback and drop the legacy
callbacks(enable, disable, config, set_polarity).Signed-off-by Lokesh Vutla <[email protected]>--- drivers/pwm/pwm-omap-dmtimer.c | 141 +++++++++++++++++++-------------- 1 file changed, 80 insertions(+), 61 deletions(-)
[..snip..]
Quoted 18 lines
Quoted 18 lines
drivers/pwm/pwm-omap-dmtimer.c +12
-3-static int pwm_omap_dmtimer_set_polarity (struct pwm_chip *chip, -struct pwm_device *pwm, -enum pwm_polarity polarity) +/** +* pwm_omap_dmtimer_apply() - Changes the state of the pwm omap dm timer. +* @chip: Pointer to PWM controller +* @pwm: Pointer to PWM channel +* @state: New sate to apply +* +* Return 0 if successfully changed the state else appropriate error. +*/ +static int pwm_omap_dmtimer_apply (struct pwm_chip *chip, +struct pwm_device *pwm, +const struct pwm_state *state) {struct pwm_omap_dmtimer_chip *omap =to_pwm_omap_dmtimer_chip (chip); +int ret =0 ;
Quoted 14 lines
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_config(struct pwm_chip *chip, · line 207
Quoted 10 lines
drivers/pwm/pwm-omap-dmtimer.c +2
-7-/* -* PWM core will not call set_polarity while PWM is enabled so it's -* safe to reconfigure the timer here without stopping it first. -*/ mutex_lock (&omap->mutex); - omap->pdata->set_pwm (omap->dm_timer, - polarity ==PWM_POLARITY_INVERSED , - true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE ); + +if (pwm_is_enabled (pwm) && !state->enabled) {In my book calling PWM API functions (designed for PWM consumers) is not so nice. I would prefer you checking the hardware registers or cache the state locally instead of relying on the core here.
.start and .stop apis does read the hardware registers and check the state before making any changes. Do you want to drop off the pwm_is_enabled(pwm) check here?
Quoted 6 lines
It would be great to have a general description at the top of the driver (like for example
drivers/pwm/pwm-sifive.c) that answers things like:- Does calling .stop completes the currently running period (it should)?
Existing driver implementation abruptly stops the cycle. I can make changes such that it completes the currently running period.
Quoted 2 lines
- Does changing polarity,
duty_cycleand period complete the running period?
- Polarity can be changed only when the pwm is not running. Ill add extra guards to reflect this behavior.
- Changing duty_cycle and period does complete the running period and new values gets reflected in next cycle.
Quoted 2 lines
- How does the hardware behave on disable? (i.e. does it output the state the pin is at in that moment? Does it go High-Z?)
Now that I am making changes to complete the current period on disable, the pin goes to Low after disabling(completing the cycle).
Ill add all these points as you mentioned in v2.
Thanks and regards,
Lokesh
Re: [PATCH 4/4] pwm: omap-dmtimer: Implement .apply callback
Hello Lokesh,
Quoted 48 lines · Lokesh Vutla
Hi Uwe,
Quoted 9 lines · Uwe Kleine-König
Quoted 8 lines · Lokesh Vutla
Implement .apply callback and drop the legacy
callbacks(enable, disable, config, set_polarity).Signed-off-by Lokesh Vutla <[email protected]>--- drivers/pwm/pwm-omap-dmtimer.c | 141 +++++++++++++++++++-------------- 1 file changed, 80 insertions(+), 61 deletions(-)[..snip..]
Quoted 18 lines
Quoted 18 lines
drivers/pwm/pwm-omap-dmtimer.c +12
-3-static int pwm_omap_dmtimer_set_polarity (struct pwm_chip *chip, -struct pwm_device *pwm, -enum pwm_polarity polarity) +/** +* pwm_omap_dmtimer_apply() - Changes the state of the pwm omap dm timer. +* @chip: Pointer to PWM controller +* @pwm: Pointer to PWM channel +* @state: New sate to apply +* +* Return 0 if successfully changed the state else appropriate error. +*/ +static int pwm_omap_dmtimer_apply (struct pwm_chip *chip, +struct pwm_device *pwm, +const struct pwm_state *state) {struct pwm_omap_dmtimer_chip *omap =to_pwm_omap_dmtimer_chip (chip); +int ret =0 ;Quoted 14 lines
drivers/pwm/pwm-omap-dmtimer.c static int pwm_omap_dmtimer_config(struct pwm_chip *chip, · line 207
Quoted 10 lines
drivers/pwm/pwm-omap-dmtimer.c +2
-7-/* -* PWM core will not call set_polarity while PWM is enabled so it's -* safe to reconfigure the timer here without stopping it first. -*/ mutex_lock (&omap->mutex); - omap->pdata->set_pwm (omap->dm_timer, - polarity ==PWM_POLARITY_INVERSED , - true,OMAP_TIMER_TRIGGER_OVERFLOW_AND_COMPARE ); + +if (pwm_is_enabled (pwm) && !state->enabled) {In my book calling PWM API functions (designed for PWM consumers) is not so nice. I would prefer you checking the hardware registers or cache the state locally instead of relying on the core here.
Quoted 3 lines
.start and .stop apis does read the hardware registers and check the state before making any changes. Do you want to drop off the
pwm_is_enabled(pwm)check here?
The IMHO more natural approach would be to look into the hardware registers instead of asking the framework.
Quoted 8 lines
Quoted 5 lines
It would be great to have a general description at the top of the driver (like for example
drivers/pwm/pwm-sifive.c) that answers things like:- Does calling .stop completes the currently running period (it should)?
Existing driver implementation abruptly stops the cycle. I can make changes such that it completes the currently running period.
That would be good for correctness.
Quoted 7 lines
Quoted 2 lines
- Does changing polarity,
duty_cycleand period complete the running period?- Polarity can be changed only when the pwm is not running. Ill add extra guards to reflect this behavior.
- Changingduty_cycleand period does complete the running period and new values gets reflected in next cycle.
Is there are race with the hardware? I.e. can it happen that when a new cycle starts just when you configured the new period but not the duty_cycle yet a mixed cycle is output?
Quoted 7 lines
Quoted 2 lines
- How does the hardware behave on disable? (i.e. does it output the state the pin is at in that moment? Does it go High-Z?)
Now that I am making changes to complete the current period on disable, the pin goes to Low after disabling(completing the cycle).
Ill add all these points as you mentioned in v2.
Great
Best regards
Uwe
Signature · 2 lines
Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | https://www.pengutronix.de/ |
Quoted 12 lines · Lokesh Vutla
This series fixes minor issues in config callback and allows for on the fly updates for pwm period and duty cycle. This is mainly intended to allow pwm omap dmtimer to be used to generate a 1PPS signal that can be syncronized to PTP clock in CPTS module in AM335x and AM57xx SoCs.
Series tested after applying the following series:
- https://patchwork.kernel.org/patch/11379875/
- https://patchwork.kernel.org/project/linux-omap/list/?series=246183Full dependencies can be seen here:
https://github.com/lokeshvutla/linux/tree/devel/pwm-1pps-generation
Quoted 6 lines
Lokesh Vutla (4):
pwm: omap-dmtimer: Drop unused header file
pwm: omap-dmtimer: Fix pwm enabling sequence
pwm: omap-dmtimer: Do not disable pwm before changing
period/duty_cycle
pwm: omap-dmtimer: Implement .apply callback
FYI, things seem to work for me after this set, but as it looks like there will be v2 set of patches I want to test again with v2 out before any Tested-by.
Thanks,
Tony