* [PATCH 1/2] hwmon: pwm-fan: Ensure that calculation doesn't discard big period values @ 2020-12-15 9:20 Uwe Kleine-König 2020-12-15 9:20 ` [PATCH 2/2] hwmon: pwm-fan: stop using legacy PWM functions and some cleanups Uwe Kleine-König ` (2 more replies) 0 siblings, 3 replies; 10+ messages in thread From: Uwe Kleine-König @ 2020-12-15 9:20 UTC (permalink / raw) To: Bartlomiej Zolnierkiewicz, Jean Delvare, Guenter Roeck Cc: Thierry Reding, Lee Jones, linux-hwmon, linux-pwm, kernel With MAX_PWM being defined to 255 the code unsigned long period; ... period = ctx->pwm->args.period; state.duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); calculates a too small value for duty_cycle if the configured period is big (either by discarding the 64 bit value ctx->pwm->args.period or by overflowing the multiplication). As this results in a too slow fan and so maybe an overheating machine better be safe than sorry and error out in .probe. Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> --- drivers/hwmon/pwm-fan.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) diff --git a/drivers/hwmon/pwm-fan.c b/drivers/hwmon/pwm-fan.c index 1f63807c0399..ec171f2b684a 100644 --- a/drivers/hwmon/pwm-fan.c +++ b/drivers/hwmon/pwm-fan.c @@ -324,8 +324,18 @@ static int pwm_fan_probe(struct platform_device *pdev) ctx->pwm_value = MAX_PWM; - /* Set duty cycle to maximum allowed and enable PWM output */ pwm_init_state(ctx->pwm, &state); + /* + * __set_pwm assumes that MAX_PWM * (period - 1) fits into an unsigned + * long. Check this here to prevent the fan running at a too low + * frequency. + */ + if (state.period > ULONG_MAX / MAX_PWM + 1) { + dev_err(dev, "Configured period too big\n"); + return -EINVAL; + } + + /* Set duty cycle to maximum allowed and enable PWM output */ state.duty_cycle = ctx->pwm->args.period - 1; state.enabled = true; base-commit: 2c85ebc57b3e1817b6ce1a6b703928e113a90442 -- 2.29.2 ^ permalink raw reply related [flat|nested] 10+ messages in thread
* [PATCH 2/2] hwmon: pwm-fan: stop using legacy PWM functions and some cleanups 2020-12-15 9:20 [PATCH 1/2] hwmon: pwm-fan: Ensure that calculation doesn't discard big period values Uwe Kleine-König @ 2020-12-15 9:20 ` Uwe Kleine-König 2020-12-15 11:45 ` Paul Barker 2020-12-30 16:35 ` Guenter Roeck 2020-12-15 11:29 ` [PATCH 1/2] hwmon: pwm-fan: Ensure that calculation doesn't discard big period values Paul Barker 2020-12-30 16:20 ` Guenter Roeck 2 siblings, 2 replies; 10+ messages in thread From: Uwe Kleine-König @ 2020-12-15 9:20 UTC (permalink / raw) To: Bartlomiej Zolnierkiewicz, Jean Delvare, Guenter Roeck Cc: Thierry Reding, Lee Jones, linux-hwmon, linux-pwm, kernel pwm_apply_state() does what the legacy functions pwm_config() and pwm_{en,dis}able() do in a single function call. This simplifies error handling and is more efficient for new-style PWM hardware drivers. Instead of repeatedly querying the PWM framework about the initial PWM configuration, cache the settings in driver data. Also use __set_pwm() in .probe() to have the algorithm calculating the PWM state in a single place. Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> --- drivers/hwmon/pwm-fan.c | 47 +++++++++++++++++------------------------ 1 file changed, 19 insertions(+), 28 deletions(-) diff --git a/drivers/hwmon/pwm-fan.c b/drivers/hwmon/pwm-fan.c index ec171f2b684a..4ccad5a87019 100644 --- a/drivers/hwmon/pwm-fan.c +++ b/drivers/hwmon/pwm-fan.c @@ -25,6 +25,7 @@ struct pwm_fan_ctx { struct mutex lock; struct pwm_device *pwm; + struct pwm_state pwm_state; struct regulator *reg_en; int irq; @@ -73,18 +74,17 @@ static int __set_pwm(struct pwm_fan_ctx *ctx, unsigned long pwm) { unsigned long period; int ret = 0; - struct pwm_state state = { }; + struct pwm_state *state = &ctx->pwm_state; mutex_lock(&ctx->lock); if (ctx->pwm_value == pwm) goto exit_set_pwm_err; - pwm_init_state(ctx->pwm, &state); - period = ctx->pwm->args.period; - state.duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); - state.enabled = pwm ? true : false; + period = state->period; + state->duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); + state->enabled = pwm ? true : false; - ret = pwm_apply_state(ctx->pwm, &state); + ret = pwm_apply_state(ctx->pwm, state); if (!ret) ctx->pwm_value = pwm; exit_set_pwm_err: @@ -274,7 +274,9 @@ static void pwm_fan_regulator_disable(void *data) static void pwm_fan_pwm_disable(void *__ctx) { struct pwm_fan_ctx *ctx = __ctx; - pwm_disable(ctx->pwm); + + ctx->pwm_state.enabled = false; + pwm_apply_state(ctx->pwm, &ctx->pwm_state); del_timer_sync(&ctx->rpm_timer); } @@ -285,7 +287,6 @@ static int pwm_fan_probe(struct platform_device *pdev) struct pwm_fan_ctx *ctx; struct device *hwmon; int ret; - struct pwm_state state = { }; u32 ppr = 2; ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL); @@ -324,22 +325,20 @@ static int pwm_fan_probe(struct platform_device *pdev) ctx->pwm_value = MAX_PWM; - pwm_init_state(ctx->pwm, &state); + pwm_init_state(ctx->pwm, &ctx->pwm_state); + /* * __set_pwm assumes that MAX_PWM * (period - 1) fits into an unsigned * long. Check this here to prevent the fan running at a too low * frequency. */ - if (state.period > ULONG_MAX / MAX_PWM + 1) { + if (ctx->pwm_state.period > ULONG_MAX / MAX_PWM + 1) { dev_err(dev, "Configured period too big\n"); return -EINVAL; } /* Set duty cycle to maximum allowed and enable PWM output */ - state.duty_cycle = ctx->pwm->args.period - 1; - state.enabled = true; - - ret = pwm_apply_state(ctx->pwm, &state); + ret = __set_pwm(ctx, MAX_PWM); if (ret) { dev_err(dev, "Failed to configure PWM: %d\n", ret); return ret; @@ -399,17 +398,16 @@ static int pwm_fan_probe(struct platform_device *pdev) static int pwm_fan_disable(struct device *dev) { struct pwm_fan_ctx *ctx = dev_get_drvdata(dev); - struct pwm_args args; int ret; - pwm_get_args(ctx->pwm, &args); - if (ctx->pwm_value) { - ret = pwm_config(ctx->pwm, 0, args.period); + /* keep ctx->pwm_state unmodified for pwm_fan_resume() */ + struct pwm_state state = ctx->pwm_state; + state.duty_cycle = 0; + state.enabled = false; + ret = pwm_apply_state(ctx->pwm, &state); if (ret < 0) return ret; - - pwm_disable(ctx->pwm); } if (ctx->reg_en) { @@ -437,8 +435,6 @@ static int pwm_fan_suspend(struct device *dev) static int pwm_fan_resume(struct device *dev) { struct pwm_fan_ctx *ctx = dev_get_drvdata(dev); - struct pwm_args pargs; - unsigned long duty; int ret; if (ctx->reg_en) { @@ -452,12 +448,7 @@ static int pwm_fan_resume(struct device *dev) if (ctx->pwm_value == 0) return 0; - pwm_get_args(ctx->pwm, &pargs); - duty = DIV_ROUND_UP_ULL(ctx->pwm_value * (pargs.period - 1), MAX_PWM); - ret = pwm_config(ctx->pwm, duty, pargs.period); - if (ret) - return ret; - return pwm_enable(ctx->pwm); + return pwm_apply_state(ctx->pwm, &ctx->pwm_state); } #endif -- 2.29.2 ^ permalink raw reply related [flat|nested] 10+ messages in thread
* Re: [PATCH 2/2] hwmon: pwm-fan: stop using legacy PWM functions and some cleanups 2020-12-15 9:20 ` [PATCH 2/2] hwmon: pwm-fan: stop using legacy PWM functions and some cleanups Uwe Kleine-König @ 2020-12-15 11:45 ` Paul Barker 2020-12-30 16:35 ` Guenter Roeck 1 sibling, 0 replies; 10+ messages in thread From: Paul Barker @ 2020-12-15 11:45 UTC (permalink / raw) To: Uwe Kleine-König Cc: Bartlomiej Zolnierkiewicz, Jean Delvare, Guenter Roeck, Thierry Reding, Lee Jones, linux-hwmon, linux-pwm, kernel On Tue, 15 Dec 2020 at 09:23, Uwe Kleine-König <u.kleine-koenig@pengutronix.de> wrote: > > pwm_apply_state() does what the legacy functions pwm_config() and > pwm_{en,dis}able() do in a single function call. This simplifies error > handling and is more efficient for new-style PWM hardware drivers. > > Instead of repeatedly querying the PWM framework about the initial PWM > configuration, cache the settings in driver data. > > Also use __set_pwm() in .probe() to have the algorithm calculating the PWM > state in a single place. > > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> > --- > drivers/hwmon/pwm-fan.c | 47 +++++++++++++++++------------------------ > 1 file changed, 19 insertions(+), 28 deletions(-) > > diff --git a/drivers/hwmon/pwm-fan.c b/drivers/hwmon/pwm-fan.c > index ec171f2b684a..4ccad5a87019 100644 > --- a/drivers/hwmon/pwm-fan.c > +++ b/drivers/hwmon/pwm-fan.c > @@ -25,6 +25,7 @@ > struct pwm_fan_ctx { > struct mutex lock; > struct pwm_device *pwm; > + struct pwm_state pwm_state; > struct regulator *reg_en; > > int irq; > @@ -73,18 +74,17 @@ static int __set_pwm(struct pwm_fan_ctx *ctx, unsigned long pwm) > { > unsigned long period; > int ret = 0; > - struct pwm_state state = { }; > + struct pwm_state *state = &ctx->pwm_state; > > mutex_lock(&ctx->lock); > if (ctx->pwm_value == pwm) > goto exit_set_pwm_err; > > - pwm_init_state(ctx->pwm, &state); > - period = ctx->pwm->args.period; > - state.duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); > - state.enabled = pwm ? true : false; > + period = state->period; > + state->duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); > + state->enabled = pwm ? true : false; > > - ret = pwm_apply_state(ctx->pwm, &state); > + ret = pwm_apply_state(ctx->pwm, state); > if (!ret) > ctx->pwm_value = pwm; > exit_set_pwm_err: > @@ -274,7 +274,9 @@ static void pwm_fan_regulator_disable(void *data) > static void pwm_fan_pwm_disable(void *__ctx) > { > struct pwm_fan_ctx *ctx = __ctx; > - pwm_disable(ctx->pwm); > + > + ctx->pwm_state.enabled = false; > + pwm_apply_state(ctx->pwm, &ctx->pwm_state); > del_timer_sync(&ctx->rpm_timer); > } > > @@ -285,7 +287,6 @@ static int pwm_fan_probe(struct platform_device *pdev) > struct pwm_fan_ctx *ctx; > struct device *hwmon; > int ret; > - struct pwm_state state = { }; > u32 ppr = 2; > > ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL); > @@ -324,22 +325,20 @@ static int pwm_fan_probe(struct platform_device *pdev) > > ctx->pwm_value = MAX_PWM; > > - pwm_init_state(ctx->pwm, &state); > + pwm_init_state(ctx->pwm, &ctx->pwm_state); > + > /* > * __set_pwm assumes that MAX_PWM * (period - 1) fits into an unsigned > * long. Check this here to prevent the fan running at a too low > * frequency. > */ > - if (state.period > ULONG_MAX / MAX_PWM + 1) { > + if (ctx->pwm_state.period > ULONG_MAX / MAX_PWM + 1) { > dev_err(dev, "Configured period too big\n"); > return -EINVAL; > } > > /* Set duty cycle to maximum allowed and enable PWM output */ > - state.duty_cycle = ctx->pwm->args.period - 1; > - state.enabled = true; > - > - ret = pwm_apply_state(ctx->pwm, &state); > + ret = __set_pwm(ctx, MAX_PWM); > if (ret) { > dev_err(dev, "Failed to configure PWM: %d\n", ret); > return ret; > @@ -399,17 +398,16 @@ static int pwm_fan_probe(struct platform_device *pdev) > static int pwm_fan_disable(struct device *dev) > { > struct pwm_fan_ctx *ctx = dev_get_drvdata(dev); > - struct pwm_args args; > int ret; > > - pwm_get_args(ctx->pwm, &args); > - > if (ctx->pwm_value) { > - ret = pwm_config(ctx->pwm, 0, args.period); > + /* keep ctx->pwm_state unmodified for pwm_fan_resume() */ > + struct pwm_state state = ctx->pwm_state; > + state.duty_cycle = 0; > + state.enabled = false; > + ret = pwm_apply_state(ctx->pwm, &state); > if (ret < 0) > return ret; > - > - pwm_disable(ctx->pwm); > } > > if (ctx->reg_en) { > @@ -437,8 +435,6 @@ static int pwm_fan_suspend(struct device *dev) > static int pwm_fan_resume(struct device *dev) > { > struct pwm_fan_ctx *ctx = dev_get_drvdata(dev); > - struct pwm_args pargs; > - unsigned long duty; > int ret; > > if (ctx->reg_en) { > @@ -452,12 +448,7 @@ static int pwm_fan_resume(struct device *dev) > if (ctx->pwm_value == 0) > return 0; > > - pwm_get_args(ctx->pwm, &pargs); > - duty = DIV_ROUND_UP_ULL(ctx->pwm_value * (pargs.period - 1), MAX_PWM); > - ret = pwm_config(ctx->pwm, duty, pargs.period); > - if (ret) > - return ret; > - return pwm_enable(ctx->pwm); > + return pwm_apply_state(ctx->pwm, &ctx->pwm_state); > } > #endif > > -- > 2.29.2 > All looks good to me at first glance. This does conflict with the changes I proposed (https://lore.kernel.org/linux-hwmon/20201212195008.6036-1-pbarker@konsulko.com/T/#t) so we'll need to figure out that out depending which order they get applied in (assuming both changes get accepted). Thanks, -- Paul Barker Konsulko Group ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH 2/2] hwmon: pwm-fan: stop using legacy PWM functions and some cleanups 2020-12-15 9:20 ` [PATCH 2/2] hwmon: pwm-fan: stop using legacy PWM functions and some cleanups Uwe Kleine-König 2020-12-15 11:45 ` Paul Barker @ 2020-12-30 16:35 ` Guenter Roeck 2021-01-12 19:13 ` [PATCH] " Uwe Kleine-König 1 sibling, 1 reply; 10+ messages in thread From: Guenter Roeck @ 2020-12-30 16:35 UTC (permalink / raw) To: Uwe Kleine-König Cc: Bartlomiej Zolnierkiewicz, Jean Delvare, Thierry Reding, Lee Jones, linux-hwmon, linux-pwm, kernel On Tue, Dec 15, 2020 at 10:20:31AM +0100, Uwe Kleine-König wrote: > pwm_apply_state() does what the legacy functions pwm_config() and > pwm_{en,dis}able() do in a single function call. This simplifies error > handling and is more efficient for new-style PWM hardware drivers. > > Instead of repeatedly querying the PWM framework about the initial PWM > configuration, cache the settings in driver data. > > Also use __set_pwm() in .probe() to have the algorithm calculating the PWM > state in a single place. > > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Can you rebase this patch on top of hwmon-next ? Thanks, Guenter > --- > drivers/hwmon/pwm-fan.c | 47 +++++++++++++++++------------------------ > 1 file changed, 19 insertions(+), 28 deletions(-) > > diff --git a/drivers/hwmon/pwm-fan.c b/drivers/hwmon/pwm-fan.c > index ec171f2b684a..4ccad5a87019 100644 > --- a/drivers/hwmon/pwm-fan.c > +++ b/drivers/hwmon/pwm-fan.c > @@ -25,6 +25,7 @@ > struct pwm_fan_ctx { > struct mutex lock; > struct pwm_device *pwm; > + struct pwm_state pwm_state; > struct regulator *reg_en; > > int irq; > @@ -73,18 +74,17 @@ static int __set_pwm(struct pwm_fan_ctx *ctx, unsigned long pwm) > { > unsigned long period; > int ret = 0; > - struct pwm_state state = { }; > + struct pwm_state *state = &ctx->pwm_state; > > mutex_lock(&ctx->lock); > if (ctx->pwm_value == pwm) > goto exit_set_pwm_err; > > - pwm_init_state(ctx->pwm, &state); > - period = ctx->pwm->args.period; > - state.duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); > - state.enabled = pwm ? true : false; > + period = state->period; > + state->duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); > + state->enabled = pwm ? true : false; > > - ret = pwm_apply_state(ctx->pwm, &state); > + ret = pwm_apply_state(ctx->pwm, state); > if (!ret) > ctx->pwm_value = pwm; > exit_set_pwm_err: > @@ -274,7 +274,9 @@ static void pwm_fan_regulator_disable(void *data) > static void pwm_fan_pwm_disable(void *__ctx) > { > struct pwm_fan_ctx *ctx = __ctx; > - pwm_disable(ctx->pwm); > + > + ctx->pwm_state.enabled = false; > + pwm_apply_state(ctx->pwm, &ctx->pwm_state); > del_timer_sync(&ctx->rpm_timer); > } > > @@ -285,7 +287,6 @@ static int pwm_fan_probe(struct platform_device *pdev) > struct pwm_fan_ctx *ctx; > struct device *hwmon; > int ret; > - struct pwm_state state = { }; > u32 ppr = 2; > > ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL); > @@ -324,22 +325,20 @@ static int pwm_fan_probe(struct platform_device *pdev) > > ctx->pwm_value = MAX_PWM; > > - pwm_init_state(ctx->pwm, &state); > + pwm_init_state(ctx->pwm, &ctx->pwm_state); > + > /* > * __set_pwm assumes that MAX_PWM * (period - 1) fits into an unsigned > * long. Check this here to prevent the fan running at a too low > * frequency. > */ > - if (state.period > ULONG_MAX / MAX_PWM + 1) { > + if (ctx->pwm_state.period > ULONG_MAX / MAX_PWM + 1) { > dev_err(dev, "Configured period too big\n"); > return -EINVAL; > } > > /* Set duty cycle to maximum allowed and enable PWM output */ > - state.duty_cycle = ctx->pwm->args.period - 1; > - state.enabled = true; > - > - ret = pwm_apply_state(ctx->pwm, &state); > + ret = __set_pwm(ctx, MAX_PWM); > if (ret) { > dev_err(dev, "Failed to configure PWM: %d\n", ret); > return ret; > @@ -399,17 +398,16 @@ static int pwm_fan_probe(struct platform_device *pdev) > static int pwm_fan_disable(struct device *dev) > { > struct pwm_fan_ctx *ctx = dev_get_drvdata(dev); > - struct pwm_args args; > int ret; > > - pwm_get_args(ctx->pwm, &args); > - > if (ctx->pwm_value) { > - ret = pwm_config(ctx->pwm, 0, args.period); > + /* keep ctx->pwm_state unmodified for pwm_fan_resume() */ > + struct pwm_state state = ctx->pwm_state; > + state.duty_cycle = 0; > + state.enabled = false; > + ret = pwm_apply_state(ctx->pwm, &state); > if (ret < 0) > return ret; > - > - pwm_disable(ctx->pwm); > } > > if (ctx->reg_en) { > @@ -437,8 +435,6 @@ static int pwm_fan_suspend(struct device *dev) > static int pwm_fan_resume(struct device *dev) > { > struct pwm_fan_ctx *ctx = dev_get_drvdata(dev); > - struct pwm_args pargs; > - unsigned long duty; > int ret; > > if (ctx->reg_en) { > @@ -452,12 +448,7 @@ static int pwm_fan_resume(struct device *dev) > if (ctx->pwm_value == 0) > return 0; > > - pwm_get_args(ctx->pwm, &pargs); > - duty = DIV_ROUND_UP_ULL(ctx->pwm_value * (pargs.period - 1), MAX_PWM); > - ret = pwm_config(ctx->pwm, duty, pargs.period); > - if (ret) > - return ret; > - return pwm_enable(ctx->pwm); > + return pwm_apply_state(ctx->pwm, &ctx->pwm_state); > } > #endif > ^ permalink raw reply [flat|nested] 10+ messages in thread
* [PATCH] hwmon: pwm-fan: stop using legacy PWM functions and some cleanups 2020-12-30 16:35 ` Guenter Roeck @ 2021-01-12 19:13 ` Uwe Kleine-König 2021-01-23 16:37 ` Guenter Roeck 0 siblings, 1 reply; 10+ messages in thread From: Uwe Kleine-König @ 2021-01-12 19:13 UTC (permalink / raw) To: Bartlomiej Zolnierkiewicz, Jean Delvare, Guenter Roeck Cc: linux-hwmon, linux-pwm, Thierry Reding, Lee Jones, kernel, Paul Barker pwm_apply_state() does what the legacy functions pwm_config() and pwm_{en,dis}able() do in a single function call. This simplifies error handling and is more efficient for new-style PWM hardware drivers. Instead of repeatedly querying the PWM framework about the initial PWM configuration, cache the settings in driver data. Also use __set_pwm() in .probe() to have the algorithm calculating the PWM state in a single place. Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> --- On Wed, Dec 30, 2020 at 08:35:20AM -0800, Guenter Roeck wrote: > Can you rebase this patch on top of hwmon-next ? Sure, here it is. Best regards Uwe drivers/hwmon/pwm-fan.c | 47 +++++++++++++++++------------------------ 1 file changed, 19 insertions(+), 28 deletions(-) diff --git a/drivers/hwmon/pwm-fan.c b/drivers/hwmon/pwm-fan.c index b5204aeb8085..41cb829784a2 100644 --- a/drivers/hwmon/pwm-fan.c +++ b/drivers/hwmon/pwm-fan.c @@ -31,6 +31,7 @@ struct pwm_fan_tach { struct pwm_fan_ctx { struct mutex lock; struct pwm_device *pwm; + struct pwm_state pwm_state; struct regulator *reg_en; int tach_count; @@ -95,18 +96,17 @@ static int __set_pwm(struct pwm_fan_ctx *ctx, unsigned long pwm) { unsigned long period; int ret = 0; - struct pwm_state state = { }; + struct pwm_state *state = &ctx->pwm_state; mutex_lock(&ctx->lock); if (ctx->pwm_value == pwm) goto exit_set_pwm_err; - pwm_init_state(ctx->pwm, &state); - period = ctx->pwm->args.period; - state.duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); - state.enabled = pwm ? true : false; + period = state->period; + state->duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); + state->enabled = pwm ? true : false; - ret = pwm_apply_state(ctx->pwm, &state); + ret = pwm_apply_state(ctx->pwm, state); if (!ret) ctx->pwm_value = pwm; exit_set_pwm_err: @@ -288,7 +288,9 @@ static void pwm_fan_regulator_disable(void *data) static void pwm_fan_pwm_disable(void *__ctx) { struct pwm_fan_ctx *ctx = __ctx; - pwm_disable(ctx->pwm); + + ctx->pwm_state.enabled = false; + pwm_apply_state(ctx->pwm, &ctx->pwm_state); del_timer_sync(&ctx->rpm_timer); } @@ -299,7 +301,6 @@ static int pwm_fan_probe(struct platform_device *pdev) struct pwm_fan_ctx *ctx; struct device *hwmon; int ret; - struct pwm_state state = { }; const struct hwmon_channel_info **channels; u32 *fan_channel_config; int channel_count = 1; /* We always have a PWM channel. */ @@ -337,22 +338,20 @@ static int pwm_fan_probe(struct platform_device *pdev) ctx->pwm_value = MAX_PWM; - pwm_init_state(ctx->pwm, &state); + pwm_init_state(ctx->pwm, &ctx->pwm_state); + /* * __set_pwm assumes that MAX_PWM * (period - 1) fits into an unsigned * long. Check this here to prevent the fan running at a too low * frequency. */ - if (state.period > ULONG_MAX / MAX_PWM + 1) { + if (ctx->pwm_state.period > ULONG_MAX / MAX_PWM + 1) { dev_err(dev, "Configured period too big\n"); return -EINVAL; } /* Set duty cycle to maximum allowed and enable PWM output */ - state.duty_cycle = ctx->pwm->args.period - 1; - state.enabled = true; - - ret = pwm_apply_state(ctx->pwm, &state); + ret = __set_pwm(ctx, MAX_PWM); if (ret) { dev_err(dev, "Failed to configure PWM: %d\n", ret); return ret; @@ -468,17 +467,16 @@ static int pwm_fan_probe(struct platform_device *pdev) static int pwm_fan_disable(struct device *dev) { struct pwm_fan_ctx *ctx = dev_get_drvdata(dev); - struct pwm_args args; int ret; - pwm_get_args(ctx->pwm, &args); - if (ctx->pwm_value) { - ret = pwm_config(ctx->pwm, 0, args.period); + /* keep ctx->pwm_state unmodified for pwm_fan_resume() */ + struct pwm_state state = ctx->pwm_state; + state.duty_cycle = 0; + state.enabled = false; + ret = pwm_apply_state(ctx->pwm, &state); if (ret < 0) return ret; - - pwm_disable(ctx->pwm); } if (ctx->reg_en) { @@ -506,8 +504,6 @@ static int pwm_fan_suspend(struct device *dev) static int pwm_fan_resume(struct device *dev) { struct pwm_fan_ctx *ctx = dev_get_drvdata(dev); - struct pwm_args pargs; - unsigned long duty; int ret; if (ctx->reg_en) { @@ -521,12 +517,7 @@ static int pwm_fan_resume(struct device *dev) if (ctx->pwm_value == 0) return 0; - pwm_get_args(ctx->pwm, &pargs); - duty = DIV_ROUND_UP_ULL(ctx->pwm_value * (pargs.period - 1), MAX_PWM); - ret = pwm_config(ctx->pwm, duty, pargs.period); - if (ret) - return ret; - return pwm_enable(ctx->pwm); + return pwm_apply_state(ctx->pwm, &ctx->pwm_state); } #endif base-commit: d1f7b079ce5b69c88c813439eea6a9c133f0846b -- 2.29.2 ^ permalink raw reply related [flat|nested] 10+ messages in thread
* Re: [PATCH] hwmon: pwm-fan: stop using legacy PWM functions and some cleanups 2021-01-12 19:13 ` [PATCH] " Uwe Kleine-König @ 2021-01-23 16:37 ` Guenter Roeck 0 siblings, 0 replies; 10+ messages in thread From: Guenter Roeck @ 2021-01-23 16:37 UTC (permalink / raw) To: Uwe Kleine-König Cc: Bartlomiej Zolnierkiewicz, Jean Delvare, linux-hwmon, linux-pwm, Thierry Reding, Lee Jones, kernel, Paul Barker On Tue, Jan 12, 2021 at 08:13:14PM +0100, Uwe Kleine-König wrote: > pwm_apply_state() does what the legacy functions pwm_config() and > pwm_{en,dis}able() do in a single function call. This simplifies error > handling and is more efficient for new-style PWM hardware drivers. > > Instead of repeatedly querying the PWM framework about the initial PWM > configuration, cache the settings in driver data. > > Also use __set_pwm() in .probe() to have the algorithm calculating the PWM > state in a single place. > > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Applied. Thanks, Guenter > --- > > On Wed, Dec 30, 2020 at 08:35:20AM -0800, Guenter Roeck wrote: > > Can you rebase this patch on top of hwmon-next ? > > Sure, here it is. > > Best regards > Uwe > > drivers/hwmon/pwm-fan.c | 47 +++++++++++++++++------------------------ > 1 file changed, 19 insertions(+), 28 deletions(-) > > > base-commit: d1f7b079ce5b69c88c813439eea6a9c133f0846b > > diff --git a/drivers/hwmon/pwm-fan.c b/drivers/hwmon/pwm-fan.c > index b5204aeb8085..41cb829784a2 100644 > --- a/drivers/hwmon/pwm-fan.c > +++ b/drivers/hwmon/pwm-fan.c > @@ -31,6 +31,7 @@ struct pwm_fan_tach { > struct pwm_fan_ctx { > struct mutex lock; > struct pwm_device *pwm; > + struct pwm_state pwm_state; > struct regulator *reg_en; > > int tach_count; > @@ -95,18 +96,17 @@ static int __set_pwm(struct pwm_fan_ctx *ctx, unsigned long pwm) > { > unsigned long period; > int ret = 0; > - struct pwm_state state = { }; > + struct pwm_state *state = &ctx->pwm_state; > > mutex_lock(&ctx->lock); > if (ctx->pwm_value == pwm) > goto exit_set_pwm_err; > > - pwm_init_state(ctx->pwm, &state); > - period = ctx->pwm->args.period; > - state.duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); > - state.enabled = pwm ? true : false; > + period = state->period; > + state->duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); > + state->enabled = pwm ? true : false; > > - ret = pwm_apply_state(ctx->pwm, &state); > + ret = pwm_apply_state(ctx->pwm, state); > if (!ret) > ctx->pwm_value = pwm; > exit_set_pwm_err: > @@ -288,7 +288,9 @@ static void pwm_fan_regulator_disable(void *data) > static void pwm_fan_pwm_disable(void *__ctx) > { > struct pwm_fan_ctx *ctx = __ctx; > - pwm_disable(ctx->pwm); > + > + ctx->pwm_state.enabled = false; > + pwm_apply_state(ctx->pwm, &ctx->pwm_state); > del_timer_sync(&ctx->rpm_timer); > } > > @@ -299,7 +301,6 @@ static int pwm_fan_probe(struct platform_device *pdev) > struct pwm_fan_ctx *ctx; > struct device *hwmon; > int ret; > - struct pwm_state state = { }; > const struct hwmon_channel_info **channels; > u32 *fan_channel_config; > int channel_count = 1; /* We always have a PWM channel. */ > @@ -337,22 +338,20 @@ static int pwm_fan_probe(struct platform_device *pdev) > > ctx->pwm_value = MAX_PWM; > > - pwm_init_state(ctx->pwm, &state); > + pwm_init_state(ctx->pwm, &ctx->pwm_state); > + > /* > * __set_pwm assumes that MAX_PWM * (period - 1) fits into an unsigned > * long. Check this here to prevent the fan running at a too low > * frequency. > */ > - if (state.period > ULONG_MAX / MAX_PWM + 1) { > + if (ctx->pwm_state.period > ULONG_MAX / MAX_PWM + 1) { > dev_err(dev, "Configured period too big\n"); > return -EINVAL; > } > > /* Set duty cycle to maximum allowed and enable PWM output */ > - state.duty_cycle = ctx->pwm->args.period - 1; > - state.enabled = true; > - > - ret = pwm_apply_state(ctx->pwm, &state); > + ret = __set_pwm(ctx, MAX_PWM); > if (ret) { > dev_err(dev, "Failed to configure PWM: %d\n", ret); > return ret; > @@ -468,17 +467,16 @@ static int pwm_fan_probe(struct platform_device *pdev) > static int pwm_fan_disable(struct device *dev) > { > struct pwm_fan_ctx *ctx = dev_get_drvdata(dev); > - struct pwm_args args; > int ret; > > - pwm_get_args(ctx->pwm, &args); > - > if (ctx->pwm_value) { > - ret = pwm_config(ctx->pwm, 0, args.period); > + /* keep ctx->pwm_state unmodified for pwm_fan_resume() */ > + struct pwm_state state = ctx->pwm_state; > + state.duty_cycle = 0; > + state.enabled = false; > + ret = pwm_apply_state(ctx->pwm, &state); > if (ret < 0) > return ret; > - > - pwm_disable(ctx->pwm); > } > > if (ctx->reg_en) { > @@ -506,8 +504,6 @@ static int pwm_fan_suspend(struct device *dev) > static int pwm_fan_resume(struct device *dev) > { > struct pwm_fan_ctx *ctx = dev_get_drvdata(dev); > - struct pwm_args pargs; > - unsigned long duty; > int ret; > > if (ctx->reg_en) { > @@ -521,12 +517,7 @@ static int pwm_fan_resume(struct device *dev) > if (ctx->pwm_value == 0) > return 0; > > - pwm_get_args(ctx->pwm, &pargs); > - duty = DIV_ROUND_UP_ULL(ctx->pwm_value * (pargs.period - 1), MAX_PWM); > - ret = pwm_config(ctx->pwm, duty, pargs.period); > - if (ret) > - return ret; > - return pwm_enable(ctx->pwm); > + return pwm_apply_state(ctx->pwm, &ctx->pwm_state); > } > #endif > ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH 1/2] hwmon: pwm-fan: Ensure that calculation doesn't discard big period values 2020-12-15 9:20 [PATCH 1/2] hwmon: pwm-fan: Ensure that calculation doesn't discard big period values Uwe Kleine-König 2020-12-15 9:20 ` [PATCH 2/2] hwmon: pwm-fan: stop using legacy PWM functions and some cleanups Uwe Kleine-König @ 2020-12-15 11:29 ` Paul Barker 2020-12-15 12:56 ` Uwe Kleine-König 2020-12-30 16:20 ` Guenter Roeck 2 siblings, 1 reply; 10+ messages in thread From: Paul Barker @ 2020-12-15 11:29 UTC (permalink / raw) To: Uwe Kleine-König Cc: Bartlomiej Zolnierkiewicz, Jean Delvare, Guenter Roeck, Thierry Reding, Lee Jones, linux-hwmon, linux-pwm, kernel On Tue, 15 Dec 2020 at 09:23, Uwe Kleine-König <u.kleine-koenig@pengutronix.de> wrote: > > With MAX_PWM being defined to 255 the code > > unsigned long period; > ... > period = ctx->pwm->args.period; > state.duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); Reviewing this I noticed that in pwm_fan_resume() we use DIV_ROUND_UP_ULL for what looks like essentially the same calculation. Could we just switch this line to DIV_ROUND_UP_ULL instead? > > calculates a too small value for duty_cycle if the configured period is > big (either by discarding the 64 bit value ctx->pwm->args.period or by > overflowing the multiplication). As this results in a too slow fan and > so maybe an overheating machine better be safe than sorry and error out > in .probe. > > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> > --- > drivers/hwmon/pwm-fan.c | 12 +++++++++++- > 1 file changed, 11 insertions(+), 1 deletion(-) > > diff --git a/drivers/hwmon/pwm-fan.c b/drivers/hwmon/pwm-fan.c > index 1f63807c0399..ec171f2b684a 100644 > --- a/drivers/hwmon/pwm-fan.c > +++ b/drivers/hwmon/pwm-fan.c > @@ -324,8 +324,18 @@ static int pwm_fan_probe(struct platform_device *pdev) > > ctx->pwm_value = MAX_PWM; > > - /* Set duty cycle to maximum allowed and enable PWM output */ > pwm_init_state(ctx->pwm, &state); > + /* > + * __set_pwm assumes that MAX_PWM * (period - 1) fits into an unsigned > + * long. Check this here to prevent the fan running at a too low > + * frequency. > + */ > + if (state.period > ULONG_MAX / MAX_PWM + 1) { > + dev_err(dev, "Configured period too big\n"); > + return -EINVAL; > + } > + > + /* Set duty cycle to maximum allowed and enable PWM output */ > state.duty_cycle = ctx->pwm->args.period - 1; > state.enabled = true; > > > base-commit: 2c85ebc57b3e1817b6ce1a6b703928e113a90442 > -- > 2.29.2 > -- Paul Barker Konsulko Group ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH 1/2] hwmon: pwm-fan: Ensure that calculation doesn't discard big period values 2020-12-15 11:29 ` [PATCH 1/2] hwmon: pwm-fan: Ensure that calculation doesn't discard big period values Paul Barker @ 2020-12-15 12:56 ` Uwe Kleine-König 2020-12-15 13:29 ` Paul Barker 0 siblings, 1 reply; 10+ messages in thread From: Uwe Kleine-König @ 2020-12-15 12:56 UTC (permalink / raw) To: Paul Barker Cc: linux-hwmon, linux-pwm, Jean Delvare, Bartlomiej Zolnierkiewicz, Thierry Reding, kernel, Lee Jones, Guenter Roeck [-- Attachment #1: Type: text/plain, Size: 1057 bytes --] On Tue, Dec 15, 2020 at 11:29:39AM +0000, Paul Barker wrote: > On Tue, 15 Dec 2020 at 09:23, Uwe Kleine-König > <u.kleine-koenig@pengutronix.de> wrote: > > > > With MAX_PWM being defined to 255 the code > > > > unsigned long period; > > ... > > period = ctx->pwm->args.period; > > state.duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); > > Reviewing this I noticed that in pwm_fan_resume() we use > DIV_ROUND_UP_ULL for what looks like essentially the same calculation. After my second patch this isn't true any more. With it applied __set_pwm is the only place in the driver that calculates this stuff. > Could we just switch this line to DIV_ROUND_UP_ULL instead? Yes that would work, but actually I don't expect someone specifiying a period big enough to justify the additional overhead of a 64 bit division. Best regards Uwe -- Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | https://www.pengutronix.de/ | [-- Attachment #2: signature.asc --] [-- Type: application/pgp-signature, Size: 488 bytes --] ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH 1/2] hwmon: pwm-fan: Ensure that calculation doesn't discard big period values 2020-12-15 12:56 ` Uwe Kleine-König @ 2020-12-15 13:29 ` Paul Barker 0 siblings, 0 replies; 10+ messages in thread From: Paul Barker @ 2020-12-15 13:29 UTC (permalink / raw) To: Uwe Kleine-König Cc: linux-hwmon, linux-pwm, Jean Delvare, Bartlomiej Zolnierkiewicz, Thierry Reding, kernel, Lee Jones, Guenter Roeck On Tue, 15 Dec 2020 at 12:56, Uwe Kleine-König <u.kleine-koenig@pengutronix.de> wrote: > > On Tue, Dec 15, 2020 at 11:29:39AM +0000, Paul Barker wrote: > > On Tue, 15 Dec 2020 at 09:23, Uwe Kleine-König > > <u.kleine-koenig@pengutronix.de> wrote: > > > > > > With MAX_PWM being defined to 255 the code > > > > > > unsigned long period; > > > ... > > > period = ctx->pwm->args.period; > > > state.duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); > > > > Reviewing this I noticed that in pwm_fan_resume() we use > > DIV_ROUND_UP_ULL for what looks like essentially the same calculation. > > After my second patch this isn't true any more. With it applied > __set_pwm is the only place in the driver that calculates this stuff. > > > Could we just switch this line to DIV_ROUND_UP_ULL instead? > > Yes that would work, but actually I don't expect someone specifiying a > period big enough to justify the additional overhead of a 64 bit > division. So ULONG_MAX / (MAX_PWM + 1) is 16.7 million on 32-bit platforms. As the period is in nanoseconds (if I understand correctly), this would allow a period of up to 16.7ms and so a minimum frequency of around 60Hz. That does seem fairly reasonable to me but it's probably worth making a note of these limits in the commit message for future reference. Thanks, -- Paul Barker Konsulko Group ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH 1/2] hwmon: pwm-fan: Ensure that calculation doesn't discard big period values 2020-12-15 9:20 [PATCH 1/2] hwmon: pwm-fan: Ensure that calculation doesn't discard big period values Uwe Kleine-König 2020-12-15 9:20 ` [PATCH 2/2] hwmon: pwm-fan: stop using legacy PWM functions and some cleanups Uwe Kleine-König 2020-12-15 11:29 ` [PATCH 1/2] hwmon: pwm-fan: Ensure that calculation doesn't discard big period values Paul Barker @ 2020-12-30 16:20 ` Guenter Roeck 2 siblings, 0 replies; 10+ messages in thread From: Guenter Roeck @ 2020-12-30 16:20 UTC (permalink / raw) To: Uwe Kleine-König Cc: Bartlomiej Zolnierkiewicz, Jean Delvare, Thierry Reding, Lee Jones, linux-hwmon, linux-pwm, kernel On Tue, Dec 15, 2020 at 10:20:30AM +0100, Uwe Kleine-König wrote: > With MAX_PWM being defined to 255 the code > > unsigned long period; > ... > period = ctx->pwm->args.period; > state.duty_cycle = DIV_ROUND_UP(pwm * (period - 1), MAX_PWM); > > calculates a too small value for duty_cycle if the configured period is > big (either by discarding the 64 bit value ctx->pwm->args.period or by > overflowing the multiplication). As this results in a too slow fan and > so maybe an overheating machine better be safe than sorry and error out > in .probe. > > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Applied. Thanks, Guenter > --- > drivers/hwmon/pwm-fan.c | 12 +++++++++++- > 1 file changed, 11 insertions(+), 1 deletion(-) > > > base-commit: 2c85ebc57b3e1817b6ce1a6b703928e113a90442 > > diff --git a/drivers/hwmon/pwm-fan.c b/drivers/hwmon/pwm-fan.c > index 1f63807c0399..ec171f2b684a 100644 > --- a/drivers/hwmon/pwm-fan.c > +++ b/drivers/hwmon/pwm-fan.c > @@ -324,8 +324,18 @@ static int pwm_fan_probe(struct platform_device *pdev) > > ctx->pwm_value = MAX_PWM; > > - /* Set duty cycle to maximum allowed and enable PWM output */ > pwm_init_state(ctx->pwm, &state); > + /* > + * __set_pwm assumes that MAX_PWM * (period - 1) fits into an unsigned > + * long. Check this here to prevent the fan running at a too low > + * frequency. > + */ > + if (state.period > ULONG_MAX / MAX_PWM + 1) { > + dev_err(dev, "Configured period too big\n"); > + return -EINVAL; > + } > + > + /* Set duty cycle to maximum allowed and enable PWM output */ > state.duty_cycle = ctx->pwm->args.period - 1; > state.enabled = true; > ^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2021-01-23 16:38 UTC | newest] Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2020-12-15 9:20 [PATCH 1/2] hwmon: pwm-fan: Ensure that calculation doesn't discard big period values Uwe Kleine-König 2020-12-15 9:20 ` [PATCH 2/2] hwmon: pwm-fan: stop using legacy PWM functions and some cleanups Uwe Kleine-König 2020-12-15 11:45 ` Paul Barker 2020-12-30 16:35 ` Guenter Roeck 2021-01-12 19:13 ` [PATCH] " Uwe Kleine-König 2021-01-23 16:37 ` Guenter Roeck 2020-12-15 11:29 ` [PATCH 1/2] hwmon: pwm-fan: Ensure that calculation doesn't discard big period values Paul Barker 2020-12-15 12:56 ` Uwe Kleine-König 2020-12-15 13:29 ` Paul Barker 2020-12-30 16:20 ` Guenter Roeck
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).