From: Rafael J. Wysocki Date: Mon, 14 Dec 2020 20:08:00 +0000 (+0100) Subject: cpufreq: Add special-purpose fast-switching callback for drivers X-Git-Tag: microblaze-v5.12~36^2^2~2 X-Git-Url: http://git.monstr.eu/?p=linux-2.6-microblaze.git;a=commitdiff_plain;h=ee2cc4276ba4909438f5894a218877660e1536d9 cpufreq: Add special-purpose fast-switching callback for drivers First off, some cpufreq drivers (eg. intel_pstate) can pass hints beyond the current target frequency to the hardware and there are no provisions for doing that in the cpufreq framework. In particular, today the driver has to assume that it should not allow the frequency to fall below the one requested by the governor (or the required capacity may not be provided) which may not be the case and which may lead to excessive energy usage in some scenarios. Second, the hints passed by these drivers to the hardware need not be in terms of the frequency, so representing the utilization numbers coming from the scheduler as frequency before passing them to those drivers is not really useful. Address the two points above by adding a special-purpose replacement for the ->fast_switch callback, called ->adjust_perf, allowing the governor to pass abstract performance level (rather than frequency) values for the minimum (required) and target (desired) performance along with the CPU capacity to compare them to. Also update the schedutil governor to use the new callback instead of ->fast_switch if present and if the utilization mertics are frequency-invariant (that is requisite for the direct mapping between the utilization and the CPU performance levels to be a reasonable approximation). Signed-off-by: Rafael J. Wysocki Acked-by: Viresh Kumar --- diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c index c17aa2973c44..d0a3525ce27f 100644 --- a/drivers/cpufreq/cpufreq.c +++ b/drivers/cpufreq/cpufreq.c @@ -2097,6 +2097,46 @@ unsigned int cpufreq_driver_fast_switch(struct cpufreq_policy *policy, } EXPORT_SYMBOL_GPL(cpufreq_driver_fast_switch); +/** + * cpufreq_driver_adjust_perf - Adjust CPU performance level in one go. + * @cpu: Target CPU. + * @min_perf: Minimum (required) performance level (units of @capacity). + * @target_perf: Terget (desired) performance level (units of @capacity). + * @capacity: Capacity of the target CPU. + * + * Carry out a fast performance level switch of @cpu without sleeping. + * + * The driver's ->adjust_perf() callback invoked by this function must be + * suitable for being called from within RCU-sched read-side critical sections + * and it is expected to select a suitable performance level equal to or above + * @min_perf and preferably equal to or below @target_perf. + * + * This function must not be called if policy->fast_switch_enabled is unset. + * + * Governors calling this function must guarantee that it will never be invoked + * twice in parallel for the same CPU and that it will never be called in + * parallel with either ->target() or ->target_index() or ->fast_switch() for + * the same CPU. + */ +void cpufreq_driver_adjust_perf(unsigned int cpu, + unsigned long min_perf, + unsigned long target_perf, + unsigned long capacity) +{ + cpufreq_driver->adjust_perf(cpu, min_perf, target_perf, capacity); +} + +/** + * cpufreq_driver_has_adjust_perf - Check "direct fast switch" callback. + * + * Return 'true' if the ->adjust_perf callback is present for the + * current driver or 'false' otherwise. + */ +bool cpufreq_driver_has_adjust_perf(void) +{ + return !!cpufreq_driver->adjust_perf; +} + /* Must set freqs->new to intermediate frequency */ static int __target_intermediate(struct cpufreq_policy *policy, struct cpufreq_freqs *freqs, int index) diff --git a/include/linux/cpufreq.h b/include/linux/cpufreq.h index 584fccd4fcab..9c8b7437b6cd 100644 --- a/include/linux/cpufreq.h +++ b/include/linux/cpufreq.h @@ -320,6 +320,15 @@ struct cpufreq_driver { unsigned int index); unsigned int (*fast_switch)(struct cpufreq_policy *policy, unsigned int target_freq); + /* + * ->fast_switch() replacement for drivers that use an internal + * representation of performance levels and can pass hints other than + * the target performance level to the hardware. + */ + void (*adjust_perf)(unsigned int cpu, + unsigned long min_perf, + unsigned long target_perf, + unsigned long capacity); /* * Caches and returns the lowest driver-supported frequency greater than @@ -588,6 +597,11 @@ struct cpufreq_governor { /* Pass a target to the cpufreq driver */ unsigned int cpufreq_driver_fast_switch(struct cpufreq_policy *policy, unsigned int target_freq); +void cpufreq_driver_adjust_perf(unsigned int cpu, + unsigned long min_perf, + unsigned long target_perf, + unsigned long capacity); +bool cpufreq_driver_has_adjust_perf(void); int cpufreq_driver_target(struct cpufreq_policy *policy, unsigned int target_freq, unsigned int relation); diff --git a/include/linux/sched/cpufreq.h b/include/linux/sched/cpufreq.h index 3ed5aa18593f..6205578ab6ee 100644 --- a/include/linux/sched/cpufreq.h +++ b/include/linux/sched/cpufreq.h @@ -28,6 +28,11 @@ static inline unsigned long map_util_freq(unsigned long util, { return (freq + (freq >> 2)) * util / cap; } + +static inline unsigned long map_util_perf(unsigned long util) +{ + return util + (util >> 2); +} #endif /* CONFIG_CPU_FREQ */ #endif /* _LINUX_SCHED_CPUFREQ_H */ diff --git a/kernel/sched/cpufreq_schedutil.c b/kernel/sched/cpufreq_schedutil.c index 319a270d13c1..803bcb30db27 100644 --- a/kernel/sched/cpufreq_schedutil.c +++ b/kernel/sched/cpufreq_schedutil.c @@ -432,13 +432,10 @@ static inline void ignore_dl_rate_limit(struct sugov_cpu *sg_cpu, struct sugov_p sg_policy->limits_changed = true; } -static void sugov_update_single(struct update_util_data *hook, u64 time, - unsigned int flags) +static inline bool sugov_update_single_common(struct sugov_cpu *sg_cpu, + u64 time, unsigned int flags) { - struct sugov_cpu *sg_cpu = container_of(hook, struct sugov_cpu, update_util); struct sugov_policy *sg_policy = sg_cpu->sg_policy; - unsigned int cached_freq = sg_policy->cached_raw_freq; - unsigned int next_f; sugov_iowait_boost(sg_cpu, time, flags); sg_cpu->last_update = time; @@ -446,11 +443,25 @@ static void sugov_update_single(struct update_util_data *hook, u64 time, ignore_dl_rate_limit(sg_cpu, sg_policy); if (!sugov_should_update_freq(sg_policy, time)) - return; + return false; sugov_get_util(sg_cpu); sugov_iowait_apply(sg_cpu, time); + return true; +} + +static void sugov_update_single_freq(struct update_util_data *hook, u64 time, + unsigned int flags) +{ + struct sugov_cpu *sg_cpu = container_of(hook, struct sugov_cpu, update_util); + struct sugov_policy *sg_policy = sg_cpu->sg_policy; + unsigned int cached_freq = sg_policy->cached_raw_freq; + unsigned int next_f; + + if (!sugov_update_single_common(sg_cpu, time, flags)) + return; + next_f = get_next_freq(sg_policy, sg_cpu->util, sg_cpu->max); /* * Do not reduce the frequency if the CPU has not been idle @@ -477,6 +488,38 @@ static void sugov_update_single(struct update_util_data *hook, u64 time, } } +static void sugov_update_single_perf(struct update_util_data *hook, u64 time, + unsigned int flags) +{ + struct sugov_cpu *sg_cpu = container_of(hook, struct sugov_cpu, update_util); + unsigned long prev_util = sg_cpu->util; + + /* + * Fall back to the "frequency" path if frequency invariance is not + * supported, because the direct mapping between the utilization and + * the performance levels depends on the frequency invariance. + */ + if (!arch_scale_freq_invariant()) { + sugov_update_single_freq(hook, time, flags); + return; + } + + if (!sugov_update_single_common(sg_cpu, time, flags)) + return; + + /* + * Do not reduce the target performance level if the CPU has not been + * idle recently, as the reduction is likely to be premature then. + */ + if (sugov_cpu_is_busy(sg_cpu) && sg_cpu->util < prev_util) + sg_cpu->util = prev_util; + + cpufreq_driver_adjust_perf(sg_cpu->cpu, map_util_perf(sg_cpu->bw_dl), + map_util_perf(sg_cpu->util), sg_cpu->max); + + sg_cpu->sg_policy->last_freq_update_time = time; +} + static unsigned int sugov_next_freq_shared(struct sugov_cpu *sg_cpu, u64 time) { struct sugov_policy *sg_policy = sg_cpu->sg_policy; @@ -815,6 +858,7 @@ static void sugov_exit(struct cpufreq_policy *policy) static int sugov_start(struct cpufreq_policy *policy) { struct sugov_policy *sg_policy = policy->governor_data; + void (*uu)(struct update_util_data *data, u64 time, unsigned int flags); unsigned int cpu; sg_policy->freq_update_delay_ns = sg_policy->tunables->rate_limit_us * NSEC_PER_USEC; @@ -834,13 +878,17 @@ static int sugov_start(struct cpufreq_policy *policy) sg_cpu->sg_policy = sg_policy; } + if (policy_is_shared(policy)) + uu = sugov_update_shared; + else if (policy->fast_switch_enabled && cpufreq_driver_has_adjust_perf()) + uu = sugov_update_single_perf; + else + uu = sugov_update_single_freq; + for_each_cpu(cpu, policy->cpus) { struct sugov_cpu *sg_cpu = &per_cpu(sugov_cpu, cpu); - cpufreq_add_update_util_hook(cpu, &sg_cpu->update_util, - policy_is_shared(policy) ? - sugov_update_shared : - sugov_update_single); + cpufreq_add_update_util_hook(cpu, &sg_cpu->update_util, uu); } return 0; }