pwm: Improvements for PWM_DEBUG checks
cover pwm: Improvements for PWM_DEBUG checks
Hello,
while working on the pwm-mediatek driver I found a few corners in the debug checks that result in false positive diagnostics. Find the two resulting patches in reply to this mail.
Best regards
Uwe
Uwe Kleine-König (2):
pwm: Disable PWM_DEBUG check for disabled states
pwm: Check actual period and duty_cycle for ignored polarity test
drivers/pwm/core.c | 34 +++++++++++++++++++++------------- 1 file changed, 21 insertions(+), 13 deletions(-)
base-commit: edd3bcb1801e1bb98f4f81485140e18c86406ced
Signature · 1 line
2.49.0
PATCH 1/2 pwm: Disable PWM_DEBUG check for disabled states
When a PWM is requested to be disabled, the result is unspecified, the only intention is to save some power. So skip all checks in this case.
All but two checks already only triggered for states with .enabled = true. The first resulted in some false positive diagnostics, the other checked for a condition that depending on hardware might not be implementable.
Similar if the lowlevel driver disabled the hardware this might be a valid reaction and with .enabled = false all other state parameters are unreliable, so skip further tests in this case, too.
All later usages of .enabled can be assumed to yield true, and so several if conditions can be simplified.
--- drivers/pwm/core.c | 32 ++++++++++++++++++++------------ 1 file changed, 20 insertions(+), 12 deletions(-)
drivers/pwm/core.c +20 -12
@@ -496,6 +496,13 @@ static void pwm_apply_debug(struct pwm_device *pwm,if (!chip->ops->get_state)return ; +/* +* If a disabled PWM was requested the result is unspecified, so nothing +* to check. +*/ +if (!state->enabled) +return ; +/* * *state was just applied. Read out the hardware state and do some * checks.
@@ -507,16 +514,23 @@ static void pwm_apply_debug(struct pwm_device *pwm,/* If that failed there isn't much to debug */ return ; +/* +* If the PWM was disabled that's maybe strange but there is nothing +* that can be sensibly checked then. So return early. +*/ +if (!s1.enabled) +return ; +/* * The lowlevel driver either ignored .polarity (which is a bug) or as * best effort inverted .polarity and fixed .duty_cycle respectively. * Undo this inversion and fixup for further tests. */ -if (s1.enabled && s1.polarity != state->polarity) { +if (s1.polarity != state->polarity) { s2.polarity = state->polarity; s2.duty_cycle = s1.period - s1.duty_cycle; s2.period = s1.period; - s2.enabled = s1.enabled; + s2.enabled = true; }else { s2 = s1; }
@@ -525,8 +539,7 @@ static void pwm_apply_debug(struct pwm_device *pwm, state->duty_cycle < state->period)dev_warn (pwmchip_parent (chip),".apply ignored .polarity\n" ); -if (state->enabled && s2.enabled && - last->polarity == state->polarity && +if (last->polarity == state->polarity && last->period > s2.period && last->period <= state->period) dev_warn(pwmchip_parent(chip),
@@ -537,13 +550,12 @@ static void pwm_apply_debug(struct pwm_device *pwm,* Rounding period up is fine only if duty_cycle is 0 then, because a * flat line doesn't have a characteristic period. */ -if (state->enabled && s2.enabled && state->period < s2.period && s2.duty_cycle) +if (state->period < s2.period && s2.duty_cycle)dev_warn (pwmchip_parent (chip),".apply is supposed to round down period (requested: %llu, applied: %llu)\n" , state->period, s2.period); -if (state->enabled && - last->polarity == state->polarity && +if (last->polarity == state->polarity && last->period == s2.period && last->duty_cycle > s2.duty_cycle && last->duty_cycle <= state->duty_cycle)
@@ -553,16 +565,12 @@ static void pwm_apply_debug(struct pwm_device *pwm, s2.duty_cycle, s2.period, last->duty_cycle, last->period); -if (state->enabled && s2.enabled && state->duty_cycle < s2.duty_cycle) +if (state->duty_cycle < s2.duty_cycle)dev_warn (pwmchip_parent (chip),".apply is supposed to round down duty_cycle (requested: %llu/%llu, applied: %llu/%llu)\n" , state->duty_cycle, state->period, s2.duty_cycle, s2.period); -if (!state->enabled && s2.enabled && s2.duty_cycle >0 ) -dev_warn (pwmchip_parent (chip), -"requested disabled, but yielded enabled with duty > 0\n" ); -/* reapply the state that the driver reported being configured. */ err = chip->ops->apply (chip, pwm, &s1);trace_pwm_apply (pwm, &s1, err);
Signature · 1 line
2.49.0
PATCH 2/2 pwm: Check actual period and duty_cycle for ignored polarity test
In reply to Uwe Kleine-KönigWhen a lowlevel driver configures the wrong period that might (historically) be ok if the emitted signal has a 100% relative duty_cycle as that just corresponds to rounding down the duty_cycle to 0 which is an allowed thing to do for a lowlevel driver.
--- drivers/pwm/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/pwm/core.c +1 -1
@@ -536,7 +536,7 @@ static void pwm_apply_debug(struct pwm_device *pwm, }if (s2.polarity != state->polarity && - state->duty_cycle < state->period) + s2.duty_cycle < s2.period)dev_warn (pwmchip_parent (chip),".apply ignored .polarity\n" ); if (last->polarity == state->polarity &&
Signature · 1 line
2.49.0
Hello,
Quoted 3 lines · Uwe Kleine-König
while working on the pwm-mediatek driver I found a few corners in the debug checks that result in false positive diagnostics. Find the two resulting patches in reply to this mail.
I applied this series to
https://git.kernel.org/pub/scm/linux/kernel/git/ukleinek/linux.git pwm/for-nexxt
as 6.18-rc1 material. Will push it to pwm/for-next after 6.17-rc1.
Best regards
Uwe