Merge branches 'pm-cpuidle' and 'pm-em'

Merge cpuidle and Energy Model changes for 6.13-rc1:

 - Add a built-in idle states table for Granite Rapids Xeon D to the
   intel_idle driver (Artem Bityutskiy).

 - Fix some typos in comments in the cpuidle core and drivers (Shen
   Lichuan).

 - Remove iowait influence from the menu cpuidle governor (Christian
   Loehle).

 - Add min/max available performance state limits to the Energy Model
   management code (Lukasz Luba).

* pm-cpuidle:
  intel_idle: add Granite Rapids Xeon D support
  cpuidle: Correct some typos in comments
  cpuidle: menu: Remove iowait influence

* pm-em:
  PM: EM: Add min/max available performance state limits

Author: rafaeljw

drivers/cpuidle/cpuidle-arm.c

@@ -139,7 +139,7 @@ static int __init arm_idle_init_cpu(int cpu)
  *
  * Initializes arm cpuidle driver for all CPUs, if any CPU fails
  * to register cpuidle driver then rollback to cancel all CPUs
- * registeration.
+ * registration.
  */
 static int __init arm_idle_init(void)
 {

drivers/cpuidle/cpuidle-qcom-spm.c

@@ -48,7 +48,7 @@ static int qcom_cpu_spc(struct spm_driver_data *drv)
 	ret = cpu_suspend(0, qcom_pm_collapse);
 	/*
 	 * ARM common code executes WFI without calling into our driver and
-	 * if the SPM mode is not reset, then we may accidently power down the
+	 * if the SPM mode is not reset, then we may accidentally power down the
 	 * cpu when we intended only to gate the cpu clock.
 	 * Ensure the state is set to standby before returning.
 	 */

drivers/cpuidle/cpuidle.c

@@ -406,7 +406,7 @@ void cpuidle_reflect(struct cpuidle_device *dev, int index)
  * Min polling interval of 10usec is a guess. It is assuming that
  * for most users, the time for a single ping-pong workload like
  * perf bench pipe would generally complete within 10usec but
- * this is hardware dependant. Actual time can be estimated with
+ * this is hardware dependent. Actual time can be estimated with
  *
  * perf bench sched pipe -l 10000
  *

drivers/cpuidle/driver.c

@@ -261,7 +261,7 @@ static void __cpuidle_unregister_driver(struct cpuidle_driver *drv)
  * @drv: a pointer to a valid struct cpuidle_driver
  *
  * Register the driver under a lock to prevent concurrent attempts to
- * [un]register the driver from occuring at the same time.
+ * [un]register the driver from occurring at the same time.
  *
  * Returns 0 on success, a negative error code (returned by
  * __cpuidle_register_driver()) otherwise.
@@ -296,7 +296,7 @@ EXPORT_SYMBOL_GPL(cpuidle_register_driver);
  * @drv: a pointer to a valid struct cpuidle_driver
  *
  * Unregisters the cpuidle driver under a lock to prevent concurrent attempts
- * to [un]register the driver from occuring at the same time.  @drv has to
+ * to [un]register the driver from occurring at the same time.  @drv has to
  * match the currently registered driver.
  */
 void cpuidle_unregister_driver(struct cpuidle_driver *drv)

drivers/cpuidle/governors/menu.c

@@ -19,7 +19,7 @@
 
 #include "gov.h"
 
-#define BUCKETS 12
+#define BUCKETS 6
 #define INTERVAL_SHIFT 3
 #define INTERVALS (1UL << INTERVAL_SHIFT)
 #define RESOLUTION 1024
@@ -29,12 +29,11 @@
 /*
  * Concepts and ideas behind the menu governor
  *
- * For the menu governor, there are 3 decision factors for picking a C
+ * For the menu governor, there are 2 decision factors for picking a C
  * state:
  * 1) Energy break even point
- * 2) Performance impact
- * 3) Latency tolerance (from pmqos infrastructure)
- * These three factors are treated independently.
+ * 2) Latency tolerance (from pmqos infrastructure)
+ * These two factors are treated independently.
  *
  * Energy break even point
  * -----------------------
@@ -75,30 +74,6 @@
  * intervals and if the stand deviation of these 8 intervals is below a
  * threshold value, we use the average of these intervals as prediction.
  *
- * Limiting Performance Impact
- * ---------------------------
- * C states, especially those with large exit latencies, can have a real
- * noticeable impact on workloads, which is not acceptable for most sysadmins,
- * and in addition, less performance has a power price of its own.
- *
- * As a general rule of thumb, menu assumes that the following heuristic
- * holds:
- *     The busier the system, the less impact of C states is acceptable
- *
- * This rule-of-thumb is implemented using a performance-multiplier:
- * If the exit latency times the performance multiplier is longer than
- * the predicted duration, the C state is not considered a candidate
- * for selection due to a too high performance impact. So the higher
- * this multiplier is, the longer we need to be idle to pick a deep C
- * state, and thus the less likely a busy CPU will hit such a deep
- * C state.
- *
- * Currently there is only one value determining the factor:
- * 10 points are added for each process that is waiting for IO on this CPU.
- * (This value was experimentally determined.)
- * Utilization is no longer a factor as it was shown that it never contributed
- * significantly to the performance multiplier in the first place.
- *
  */
 
 struct menu_device {
@@ -112,19 +87,10 @@ struct menu_device {
 	int		interval_ptr;
 };
 
-static inline int which_bucket(u64 duration_ns, unsigned int nr_iowaiters)
+static inline int which_bucket(u64 duration_ns)
 {
 	int bucket = 0;
 
-	/*
-	 * We keep two groups of stats; one with no
-	 * IO pending, one without.
-	 * This allows us to calculate
-	 * E(duration)|iowait
-	 */
-	if (nr_iowaiters)
-		bucket = BUCKETS/2;
-
 	if (duration_ns < 10ULL * NSEC_PER_USEC)
 		return bucket;
 	if (duration_ns < 100ULL * NSEC_PER_USEC)
@@ -138,19 +104,6 @@ static inline int which_bucket(u64 duration_ns, unsigned int nr_iowaiters)
 	return bucket + 5;
 }
 
-/*
- * Return a multiplier for the exit latency that is intended
- * to take performance requirements into account.
- * The more performance critical we estimate the system
- * to be, the higher this multiplier, and thus the higher
- * the barrier to go to an expensive C state.
- */
-static inline int performance_multiplier(unsigned int nr_iowaiters)
-{
-	/* for IO wait tasks (per cpu!) we add 10x each */
-	return 1 + 10 * nr_iowaiters;
-}
-
 static DEFINE_PER_CPU(struct menu_device, menu_devices);
 
 static void menu_update(struct cpuidle_driver *drv, struct cpuidle_device *dev);
@@ -258,8 +211,6 @@ static int menu_select(struct cpuidle_driver *drv, struct cpuidle_device *dev,
 	struct menu_device *data = this_cpu_ptr(&menu_devices);
 	s64 latency_req = cpuidle_governor_latency_req(dev->cpu);
 	u64 predicted_ns;
-	u64 interactivity_req;
-	unsigned int nr_iowaiters;
 	ktime_t delta, delta_tick;
 	int i, idx;
 
@@ -268,8 +219,6 @@ static int menu_select(struct cpuidle_driver *drv, struct cpuidle_device *dev,
 		data->needs_update = 0;
 	}
 
-	nr_iowaiters = nr_iowait_cpu(dev->cpu);
-
 	/* Find the shortest expected idle interval. */
 	predicted_ns = get_typical_interval(data) * NSEC_PER_USEC;
 	if (predicted_ns > RESIDENCY_THRESHOLD_NS) {
@@ -283,7 +232,7 @@ static int menu_select(struct cpuidle_driver *drv, struct cpuidle_device *dev,
 		}
 
 		data->next_timer_ns = delta;
-		data->bucket = which_bucket(data->next_timer_ns, nr_iowaiters);
+		data->bucket = which_bucket(data->next_timer_ns);
 
 		/* Round up the result for half microseconds. */
 		timer_us = div_u64((RESOLUTION * DECAY * NSEC_PER_USEC) / 2 +
@@ -301,7 +250,7 @@ static int menu_select(struct cpuidle_driver *drv, struct cpuidle_device *dev,
 		 */
 		data->next_timer_ns = KTIME_MAX;
 		delta_tick = TICK_NSEC / 2;
-		data->bucket = which_bucket(KTIME_MAX, nr_iowaiters);
+		data->bucket = which_bucket(KTIME_MAX);
 	}
 
 	if (unlikely(drv->state_count <= 1 || latency_req == 0) ||
@@ -328,15 +277,8 @@ static int menu_select(struct cpuidle_driver *drv, struct cpuidle_device *dev,
 		 */
 		if (predicted_ns < TICK_NSEC)
 			predicted_ns = data->next_timer_ns;
-	} else {
-		/*
-		 * Use the performance multiplier and the user-configurable
-		 * latency_req to determine the maximum exit latency.
-		 */
-		interactivity_req = div64_u64(predicted_ns,
-					      performance_multiplier(nr_iowaiters));
-		if (latency_req > interactivity_req)
-			latency_req = interactivity_req;
+	} else if (latency_req > predicted_ns) {
+		latency_req = predicted_ns;
 	}
 
 	/*

drivers/idle/intel_idle.c

@@ -1069,6 +1069,47 @@ static struct cpuidle_state gnr_cstates[] __initdata = {
 		.enter = NULL }
 };
 
+static struct cpuidle_state gnrd_cstates[] __initdata = {
+	{
+		.name = "C1",
+		.desc = "MWAIT 0x00",
+		.flags = MWAIT2flg(0x00),
+		.exit_latency = 1,
+		.target_residency = 1,
+		.enter = &intel_idle,
+		.enter_s2idle = intel_idle_s2idle, },
+	{
+		.name = "C1E",
+		.desc = "MWAIT 0x01",
+		.flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_ALWAYS_ENABLE,
+		.exit_latency = 4,
+		.target_residency = 4,
+		.enter = &intel_idle,
+		.enter_s2idle = intel_idle_s2idle, },
+	{
+		.name = "C6",
+		.desc = "MWAIT 0x20",
+		.flags = MWAIT2flg(0x20) | CPUIDLE_FLAG_TLB_FLUSHED |
+					   CPUIDLE_FLAG_INIT_XSTATE |
+					   CPUIDLE_FLAG_PARTIAL_HINT_MATCH,
+		.exit_latency = 220,
+		.target_residency = 650,
+		.enter = &intel_idle,
+		.enter_s2idle = intel_idle_s2idle, },
+	{
+		.name = "C6P",
+		.desc = "MWAIT 0x21",
+		.flags = MWAIT2flg(0x21) | CPUIDLE_FLAG_TLB_FLUSHED |
+					   CPUIDLE_FLAG_INIT_XSTATE |
+					   CPUIDLE_FLAG_PARTIAL_HINT_MATCH,
+		.exit_latency = 240,
+		.target_residency = 750,
+		.enter = &intel_idle,
+		.enter_s2idle = intel_idle_s2idle, },
+	{
+		.enter = NULL }
+};
+
 static struct cpuidle_state atom_cstates[] __initdata = {
 	{
 		.name = "C1E",
@@ -1508,6 +1549,12 @@ static const struct idle_cpu idle_cpu_gnr __initconst = {
 	.use_acpi = true,
 };
 
+static const struct idle_cpu idle_cpu_gnrd __initconst = {
+	.state_table = gnrd_cstates,
+	.disable_promotion_to_c1e = true,
+	.use_acpi = true,
+};
+
 static const struct idle_cpu idle_cpu_avn __initconst = {
 	.state_table = avn_cstates,
 	.disable_promotion_to_c1e = true,
@@ -1593,6 +1640,7 @@ static const struct x86_cpu_id intel_idle_ids[] __initconst = {
 	X86_MATCH_VFM(INTEL_SAPPHIRERAPIDS_X,	&idle_cpu_spr),
 	X86_MATCH_VFM(INTEL_EMERALDRAPIDS_X,	&idle_cpu_spr),
 	X86_MATCH_VFM(INTEL_GRANITERAPIDS_X,	&idle_cpu_gnr),
+	X86_MATCH_VFM(INTEL_GRANITERAPIDS_D,	&idle_cpu_gnrd),
 	X86_MATCH_VFM(INTEL_XEON_PHI_KNL,	&idle_cpu_knl),
 	X86_MATCH_VFM(INTEL_XEON_PHI_KNM,	&idle_cpu_knl),
 	X86_MATCH_VFM(INTEL_ATOM_GOLDMONT,	&idle_cpu_bxt),

include/linux/energy_model.h

@@ -55,6 +55,8 @@ struct em_perf_table {
  * struct em_perf_domain - Performance domain
  * @em_table:		Pointer to the runtime modifiable em_perf_table
  * @nr_perf_states:	Number of performance states
+ * @min_perf_state:	Minimum allowed Performance State index
+ * @max_perf_state:	Maximum allowed Performance State index
  * @flags:		See "em_perf_domain flags"
  * @cpus:		Cpumask covering the CPUs of the domain. It's here
  *			for performance reasons to avoid potential cache
@@ -70,6 +72,8 @@ struct em_perf_table {
 struct em_perf_domain {
 	struct em_perf_table __rcu *em_table;
 	int nr_perf_states;
+	int min_perf_state;
+	int max_perf_state;
 	unsigned long flags;
 	unsigned long cpus[];
 };
@@ -173,13 +177,14 @@ void em_table_free(struct em_perf_table __rcu *table);
 int em_dev_compute_costs(struct device *dev, struct em_perf_state *table,
 			 int nr_states);
 int em_dev_update_chip_binning(struct device *dev);
+int em_update_performance_limits(struct em_perf_domain *pd,
+		unsigned long freq_min_khz, unsigned long freq_max_khz);
 
 /**
  * em_pd_get_efficient_state() - Get an efficient performance state from the EM
  * @table:		List of performance states, in ascending order
- * @nr_perf_states:	Number of performance states
+ * @pd:			performance domain for which this must be done
  * @max_util:		Max utilization to map with the EM
- * @pd_flags:		Performance Domain flags
  *
  * It is called from the scheduler code quite frequently and as a consequence
  * doesn't implement any check.
@@ -188,13 +193,16 @@ int em_dev_update_chip_binning(struct device *dev);
  * requirement.
  */
 static inline int
-em_pd_get_efficient_state(struct em_perf_state *table, int nr_perf_states,
-			  unsigned long max_util, unsigned long pd_flags)
+em_pd_get_efficient_state(struct em_perf_state *table,
+			  struct em_perf_domain *pd, unsigned long max_util)
 {
+	unsigned long pd_flags = pd->flags;
+	int min_ps = pd->min_perf_state;
+	int max_ps = pd->max_perf_state;
 	struct em_perf_state *ps;
 	int i;
 
-	for (i = 0; i < nr_perf_states; i++) {
+	for (i = min_ps; i <= max_ps; i++) {
 		ps = &table[i];
 		if (ps->performance >= max_util) {
 			if (pd_flags & EM_PERF_DOMAIN_SKIP_INEFFICIENCIES &&
@@ -204,7 +212,7 @@ em_pd_get_efficient_state(struct em_perf_state *table, int nr_perf_states,
 		}
 	}
 
-	return nr_perf_states - 1;
+	return max_ps;
 }
 
 /**
@@ -253,8 +261,7 @@ static inline unsigned long em_cpu_energy(struct em_perf_domain *pd,
 	 * requested performance.
 	 */
 	em_table = rcu_dereference(pd->em_table);
-	i = em_pd_get_efficient_state(em_table->state, pd->nr_perf_states,
-				      max_util, pd->flags);
+	i = em_pd_get_efficient_state(em_table->state, pd, max_util);
 	ps = &em_table->state[i];
 
 	/*
@@ -391,6 +398,12 @@ static inline int em_dev_update_chip_binning(struct device *dev)
 {
 	return -EINVAL;
 }
+static inline
+int em_update_performance_limits(struct em_perf_domain *pd,
+		unsigned long freq_min_khz, unsigned long freq_max_khz)
+{
+	return -EINVAL;
+}
 #endif
 
 #endif