Merge tag 'pm-6.1-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm

Pull power management fixes from Rafael Wysocki:
 "These revert a recent change in the schedutil cpufreq governor that
  had not been expected to make any functional difference, but turned
  out to introduce a performance regression, fix an initialization issue
  in the amd-pstate driver and make it actually replace the venerable
  ACPI cpufreq driver on the supported systems by default.

  Specifics:

   - Revert a recent schedutil cpufreq governor change that introduced a
     performace regression on Pixel 6 (Sam Wu)

   - Fix amd-pstate driver initialization after running the kernel via
     kexec (Wyes Karny)

   - Turn amd-pstate into a built-in driver which allows it to take
     precedence over acpi-cpufreq by default on supported systems and
     amend it with a mechanism to disable this behavior (Perry Yuan)

   - Update amd-pstate documentation in accordance with the other
     changes made to it (Perry Yuan)"

* tag 'pm-6.1-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  Documentation: add amd-pstate kernel command line options
  Documentation: amd-pstate: add driver working mode introduction
  cpufreq: amd-pstate: add amd-pstate driver parameter for mode selection
  cpufreq: amd-pstate: change amd-pstate driver to be built-in type
  cpufreq: amd-pstate: cpufreq: amd-pstate: reset MSR_AMD_PERF_CTL register at init
  Revert "cpufreq: schedutil: Move max CPU capacity to sugov_policy"

1 files changed, 15 insertions, 15 deletions

diff --git a/kernel/sched/cpufreq_schedutil.c b/kernel/sched/cpufreq_schedutil.c
index 9161d1136d01..1207c78f85c1 100644
--- a/kernel/sched/cpufreq_schedutil.c
+++ b/kernel/sched/cpufreq_schedutil.c
@@ -25,9 +25,6 @@ struct sugov_policy {
 	unsigned int		next_freq;
 	unsigned int		cached_raw_freq;
 
-	/* max CPU capacity, which is equal for all CPUs in freq. domain */
-	unsigned long		max;
-
 	/* The next fields are only needed if fast switch cannot be used: */
 	struct			irq_work irq_work;
 	struct			kthread_work work;
@@ -51,6 +48,7 @@ struct sugov_cpu {
 
 	unsigned long		util;
 	unsigned long		bw_dl;
+	unsigned long		max;
 
 	/* The field below is for single-CPU policies only: */
 #ifdef CONFIG_NO_HZ_COMMON
@@ -160,6 +158,7 @@ static void sugov_get_util(struct sugov_cpu *sg_cpu)
 {
 	struct rq *rq = cpu_rq(sg_cpu->cpu);
 
+	sg_cpu->max = arch_scale_cpu_capacity(sg_cpu->cpu);
 	sg_cpu->bw_dl = cpu_bw_dl(rq);
 	sg_cpu->util = effective_cpu_util(sg_cpu->cpu, cpu_util_cfs(sg_cpu->cpu),
 					  FREQUENCY_UTIL, NULL);
@@ -254,7 +253,6 @@ static void sugov_iowait_boost(struct sugov_cpu *sg_cpu, u64 time,
  */
 static void sugov_iowait_apply(struct sugov_cpu *sg_cpu, u64 time)
 {
-	struct sugov_policy *sg_policy = sg_cpu->sg_policy;
 	unsigned long boost;
 
 	/* No boost currently required */
@@ -282,8 +280,7 @@ static void sugov_iowait_apply(struct sugov_cpu *sg_cpu, u64 time)
 	 * sg_cpu->util is already in capacity scale; convert iowait_boost
 	 * into the same scale so we can compare.
 	 */
-	boost = sg_cpu->iowait_boost * sg_policy->max;
-	boost >>= SCHED_CAPACITY_SHIFT;
+	boost = (sg_cpu->iowait_boost * sg_cpu->max) >> SCHED_CAPACITY_SHIFT;
 	boost = uclamp_rq_util_with(cpu_rq(sg_cpu->cpu), boost, NULL);
 	if (sg_cpu->util < boost)
 		sg_cpu->util = boost;
@@ -340,7 +337,7 @@ static void sugov_update_single_freq(struct update_util_data *hook, u64 time,
 	if (!sugov_update_single_common(sg_cpu, time, flags))
 		return;
 
-	next_f = get_next_freq(sg_policy, sg_cpu->util, sg_policy->max);
+	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
 	 * recently, as the reduction is likely to be premature then.
@@ -376,7 +373,6 @@ 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);
-	struct sugov_policy *sg_policy = sg_cpu->sg_policy;
 	unsigned long prev_util = sg_cpu->util;
 
 	/*
@@ -403,8 +399,7 @@ static void sugov_update_single_perf(struct update_util_data *hook, u64 time,
 		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_policy->max);
+				   map_util_perf(sg_cpu->util), sg_cpu->max);
 
 	sg_cpu->sg_policy->last_freq_update_time = time;
 }
@@ -413,19 +408,25 @@ static unsigned int sugov_next_freq_shared(struct sugov_cpu *sg_cpu, u64 time)
 {
 	struct sugov_policy *sg_policy = sg_cpu->sg_policy;
 	struct cpufreq_policy *policy = sg_policy->policy;
-	unsigned long util = 0;
+	unsigned long util = 0, max = 1;
 	unsigned int j;
 
 	for_each_cpu(j, policy->cpus) {
 		struct sugov_cpu *j_sg_cpu = &per_cpu(sugov_cpu, j);
+		unsigned long j_util, j_max;
 
 		sugov_get_util(j_sg_cpu);
 		sugov_iowait_apply(j_sg_cpu, time);
+		j_util = j_sg_cpu->util;
+		j_max = j_sg_cpu->max;
 
-		util = max(j_sg_cpu->util, util);
+		if (j_util * max > j_max * util) {
+			util = j_util;
+			max = j_max;
+		}
 	}
 
-	return get_next_freq(sg_policy, util, sg_policy->max);
+	return get_next_freq(sg_policy, util, max);
 }
 
 static void
@@ -751,7 +752,7 @@ 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 = cpumask_first(policy->cpus);
+	unsigned int cpu;
 
 	sg_policy->freq_update_delay_ns	= sg_policy->tunables->rate_limit_us * NSEC_PER_USEC;
 	sg_policy->last_freq_update_time	= 0;
@@ -759,7 +760,6 @@ static int sugov_start(struct cpufreq_policy *policy)
 	sg_policy->work_in_progress		= false;
 	sg_policy->limits_changed		= false;
 	sg_policy->cached_raw_freq		= 0;
-	sg_policy->max				= arch_scale_cpu_capacity(cpu);
 
 	sg_policy->need_freq_update = cpufreq_driver_test_flags(CPUFREQ_NEED_UPDATE_LIMITS);