amd-pstate preferred core

amd-pstate driver also uses SCHED_MC_PRIO, so decouple the requirement of CPU_SUP_INTEL from the dependencies to allow compilation in kernels without Intel CPU support.

Reviewed-by Mario Limonciello <[email protected]>

Reviewed-by Huang Rui <[email protected]>

Signed-off-by Meng Li <[email protected]>

---
 arch/x86/Kconfig | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

arch/x86/Kconfig +3 -2

index 66bfabae8814..a2e163acf623 100644
@@ -1054,8 +1054,9 @@ config SCHED_MC
 
 config SCHED_MC_PRIO
 	bool "CPU core priorities scheduler support"
-	depends on SCHED_MC && CPU_SUP_INTEL
-	select X86_INTEL_PSTATE
+	depends on SCHED_MC
+	select X86_INTEL_PSTATE if CPU_SUP_INTEL
+	select X86_AMD_PSTATE if CPU_SUP_AMD && ACPI
 	select CPU_FREQ
 	default y
 	help

Signature · 1 line

2.34.1

amd-pstate driver also uses SCHED_MC_PRIO, so decouple the requirement of CPU_SUP_INTEL from the dependencies to allow compilation in kernels without Intel CPU support.

Reviewed-by Mario Limonciello <[email protected]>

Reviewed-by Huang Rui <[email protected]>

Signed-off-by Meng Li <[email protected]>

---
 arch/x86/Kconfig | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

arch/x86/Kconfig +0 -0

index 66bfabae8814..a2e163acf623 100644
@@ -1054,8 +1054,9 @@ config SCHED_MC
 

amd-pstate driver utilizes the functions and data structures provided by the ITMT architecture to enable the scheduler to favor scheduling on cores which can be get a higher frequency with lower voltage. We call it amd-pstate preferrred core.

Here sched_set_itmt_core_prio() is called to set priorities and sched_set_itmt_support() is called to enable ITMT feature. amd-pstate driver uses the highest performance value to indicate the priority of CPU. The higher value has a higher priority.

The initial core rankings are set up by amd-pstate when the system boots.

Add a variable hw_prefcore in cpudata structure. It will check if the processor and power firmware support preferred core feature.

Add one new early parameter disable to allow user to disable the preferred core.

Only when hardware supports preferred core and user set enabled in early parameter, amd pstate driver supports preferred core featue.

Reviewed-by Mario Limonciello <[email protected]>

Co-developed-by Perry Yuan <[email protected]>

Signed-off-by Perry Yuan <[email protected]>

Signed-off-by Meng Li <[email protected]>

---
 drivers/cpufreq/amd-pstate.c | 152 +++++++++++++++++++++++++++++++----
 include/linux/amd-pstate.h   |   4 +
 2 files changed, 140 insertions(+), 16 deletions(-)

drivers/cpufreq/amd-pstate.c +136 -16

index 9a1e194d5cf8..6ac8939fce5a 100644
@@ -37,6 +37,7 @@
 #include <linux/uaccess.h>
 #include <linux/static_call.h>
 #include <linux/amd-pstate.h>
+#include <linux/topology.h>
 
 #include <acpi/processor.h>
 #include <acpi/cppc_acpi.h>

@@ -49,6 +50,8 @@
 
 #define AMD_PSTATE_TRANSITION_LATENCY	20000
 #define AMD_PSTATE_TRANSITION_DELAY	1000
+#define AMD_PSTATE_PREFCORE_THRESHOLD	166
+#define AMD_PSTATE_MAX_CPPC_PERF	255
 
 /*
  * TODO: We need more time to fine tune processors with shared memory solution

@@ -64,6 +67,7 @@ static struct cpufreq_driver amd_pstate_driver;
 static struct cpufreq_driver amd_pstate_epp_driver;
 static int cppc_state = AMD_PSTATE_UNDEFINED;
 static bool cppc_enabled;
+static bool amd_pstate_prefcore = true;
 
 /*
  * AMD Energy Preference Performance (EPP)

@@ -290,23 +294,21 @@ static inline int amd_pstate_enable(bool enable)
 static int pstate_init_perf(struct amd_cpudata *cpudata)
 {
 	u64 cap1;
-	u32 highest_perf;
 
 	int ret = rdmsrl_safe_on_cpu(cpudata->cpu, MSR_AMD_CPPC_CAP1,
 				     &cap1);
 	if (ret)
 		return ret;
 
-	/*
-	 * TODO: Introduce AMD specific power feature.
-	 *
-	 * CPPC entry doesn't indicate the highest performance in some ASICs.
+	/* For platforms that do not support the preferred core feature, the
+	 * highest_pef may be configured with 166 or 255, to avoid max frequency
+	 * calculated wrongly. we take the AMD_CPPC_HIGHEST_PERF(cap1) value as
+	 * the default max perf.
 	 */
-	highest_perf = amd_get_highest_perf();
-	if (highest_perf > AMD_CPPC_HIGHEST_PERF(cap1))
-		highest_perf = AMD_CPPC_HIGHEST_PERF(cap1);
-
-	WRITE_ONCE(cpudata->highest_perf, highest_perf);
+	if (cpudata->hw_prefcore)
+		WRITE_ONCE(cpudata->highest_perf, AMD_PSTATE_PREFCORE_THRESHOLD);
+	else
+		WRITE_ONCE(cpudata->highest_perf, AMD_CPPC_HIGHEST_PERF(cap1));
 
 	WRITE_ONCE(cpudata->nominal_perf, AMD_CPPC_NOMINAL_PERF(cap1));
 	WRITE_ONCE(cpudata->lowest_nonlinear_perf, AMD_CPPC_LOWNONLIN_PERF(cap1));

@@ -318,17 +320,15 @@ static int pstate_init_perf(struct amd_cpudata *cpudata)
 static int cppc_init_perf(struct amd_cpudata *cpudata)
 {
 	struct cppc_perf_caps cppc_perf;
-	u32 highest_perf;
 
 	int ret = cppc_get_perf_caps(cpudata->cpu, &cppc_perf);
 	if (ret)
 		return ret;
 
-	highest_perf = amd_get_highest_perf();
-	if (highest_perf > cppc_perf.highest_perf)
-		highest_perf = cppc_perf.highest_perf;
-
-	WRITE_ONCE(cpudata->highest_perf, highest_perf);
+	if (cpudata->hw_prefcore)
+		WRITE_ONCE(cpudata->highest_perf, AMD_PSTATE_PREFCORE_THRESHOLD);
+	else
+		WRITE_ONCE(cpudata->highest_perf, cppc_perf.highest_perf);
 
 	WRITE_ONCE(cpudata->nominal_perf, cppc_perf.nominal_perf);
 	WRITE_ONCE(cpudata->lowest_nonlinear_perf,

@@ -676,6 +676,90 @@ static void amd_perf_ctl_reset(unsigned int cpu)
 	wrmsrl_on_cpu(cpu, MSR_AMD_PERF_CTL, 0);
 }
 
+/*
+ * Set amd-pstate preferred core enable can't be done directly from cpufreq callbacks
+ * due to locking, so queue the work for later.
+ */
+static void amd_pstste_sched_prefcore_workfn(struct work_struct *work)
+{
+	sched_set_itmt_support();
+}
+static DECLARE_WORK(sched_prefcore_work, amd_pstste_sched_prefcore_workfn);
+
+/*
+ * Get the highest performance register value.
+ * @cpu: CPU from which to get highest performance.
+ * @highest_perf: Return address.
+ *
+ * Return: 0 for success, -EIO otherwise.
+ */
+static int amd_pstate_get_highest_perf(int cpu, u32 *highest_perf)
+{
+	int ret;
+	u64 cppc_highest_perf;
+
+	if (boot_cpu_has(X86_FEATURE_CPPC)) {
+		u64 cap1;
+
+		ret = rdmsrl_safe_on_cpu(cpu, MSR_AMD_CPPC_CAP1, &cap1);
+		if (ret)
+			return ret;
+		WRITE_ONCE(*highest_perf, AMD_CPPC_HIGHEST_PERF(cap1));
+	} else {
+		ret = cppc_get_highest_perf(cpu, &cppc_highest_perf);
+		*highest_perf = (u32)(cppc_highest_perf & 0xFFFF);
+	}
+
+	return (ret);
+}
+
+static void amd_pstate_init_prefcore(struct amd_cpudata *cpudata)
+{
+	int ret;
+	u32 highest_perf;
+	static u32 max_highest_perf = 0, min_highest_perf = U32_MAX;
+
+	ret = amd_pstate_get_highest_perf(cpudata->cpu, &highest_perf);
+	if (ret)
+		return;
+
+	cpudata->hw_prefcore = true;
+	/* check if CPPC preferred core feature is enabled*/
+	if (highest_perf == AMD_PSTATE_MAX_CPPC_PERF) {
+		pr_debug("AMD CPPC preferred core is unsupported!\n");
+		cpudata->hw_prefcore = false;
+		return;
+	}
+
+	if (!amd_pstate_prefcore)
+		return;
+
+	/*
+	 * The priorities can be set regardless of whether or not
+	 * sched_set_itmt_support(true) has been called and it is valid to
+	 * update them at any time after it has been called.
+	 */
+	sched_set_itmt_core_prio(highest_perf, cpudata->cpu);
+
+	if (max_highest_perf <= min_highest_perf) {
+		if (highest_perf > max_highest_perf)
+			max_highest_perf = highest_perf;
+
+		if (highest_perf < min_highest_perf)
+			min_highest_perf = highest_perf;
+
+		if (max_highest_perf > min_highest_perf) {
+			/*
+			 * This code can be run during CPU online under the
+			 * CPU hotplug locks, so sched_set_itmt_support()
+			 * cannot be called from here.  Queue up a work item
+			 * to invoke it.
+			 */
+			schedule_work(&sched_prefcore_work);
+		}
+	}
+}
+
 static int amd_pstate_cpu_init(struct cpufreq_policy *policy)
 {
 	int min_freq, max_freq, nominal_freq, lowest_nonlinear_freq, ret;

@@ -697,6 +781,8 @@ static int amd_pstate_cpu_init(struct cpufreq_policy *policy)
 
 	cpudata->cpu = policy->cpu;
 
+	amd_pstate_init_prefcore(cpudata);
+
 	ret = amd_pstate_init_perf(cpudata);
 	if (ret)
 		goto free_cpudata1;

@@ -845,6 +931,17 @@ static ssize_t show_amd_pstate_highest_perf(struct cpufreq_policy *policy,
 	return sysfs_emit(buf, "%u\n", perf);
 }
 
+static ssize_t show_amd_pstate_hw_prefcore(struct cpufreq_policy *policy,
+					   char *buf)
+{
+	bool hw_prefcore;
+	struct amd_cpudata *cpudata = policy->driver_data;
+
+	hw_prefcore = READ_ONCE(cpudata->hw_prefcore);
+
+	return sysfs_emit(buf, "%s\n", hw_prefcore ? "supported" : "unsupported");
+}
+
 static ssize_t show_energy_performance_available_preferences(
 				struct cpufreq_policy *policy, char *buf)
 {

@@ -1037,18 +1134,27 @@ static ssize_t status_store(struct device *a, struct device_attribute *b,
 	return ret < 0 ? ret : count;
 }
 
+static ssize_t prefcore_show(struct device *dev,
+			     struct device_attribute *attr, char *buf)
+{
+	return sysfs_emit(buf, "%s\n", amd_pstate_prefcore ? "enabled" : "disabled");
+}
+
 cpufreq_freq_attr_ro(amd_pstate_max_freq);
 cpufreq_freq_attr_ro(amd_pstate_lowest_nonlinear_freq);
 
 cpufreq_freq_attr_ro(amd_pstate_highest_perf);
+cpufreq_freq_attr_ro(amd_pstate_hw_prefcore);
 cpufreq_freq_attr_rw(energy_performance_preference);
 cpufreq_freq_attr_ro(energy_performance_available_preferences);
 static DEVICE_ATTR_RW(status);
+static DEVICE_ATTR_RO(prefcore);
 
 static struct freq_attr *amd_pstate_attr[] = {
 	&amd_pstate_max_freq,
 	&amd_pstate_lowest_nonlinear_freq,
 	&amd_pstate_highest_perf,
+	&amd_pstate_hw_prefcore,
 	NULL,
 };
 

@@ -1056,6 +1162,7 @@ static struct freq_attr *amd_pstate_epp_attr[] = {
 	&amd_pstate_max_freq,
 	&amd_pstate_lowest_nonlinear_freq,
 	&amd_pstate_highest_perf,
+	&amd_pstate_hw_prefcore,
 	&energy_performance_preference,
 	&energy_performance_available_preferences,
 	NULL,

@@ -1063,6 +1170,7 @@ static struct freq_attr *amd_pstate_epp_attr[] = {
 
 static struct attribute *pstate_global_attributes[] = {
 	&dev_attr_status.attr,
+	&dev_attr_prefcore.attr,
 	NULL
 };
 

@@ -1114,6 +1222,8 @@ static int amd_pstate_epp_cpu_init(struct cpufreq_policy *policy)
 	cpudata->cpu = policy->cpu;
 	cpudata->epp_policy = 0;
 
+	amd_pstate_init_prefcore(cpudata);
+
 	ret = amd_pstate_init_perf(cpudata);
 	if (ret)
 		goto free_cpudata1;

@@ -1527,7 +1637,17 @@ static int __init amd_pstate_param(char *str)
 
 	return amd_pstate_set_driver(mode_idx);
 }
+
+static int __init amd_prefcore_param(char *str)
+{
+	if (!strcmp(str, "disable"))
+		amd_pstate_prefcore = false;
+
+	return 0;
+}
+
 early_param("amd_pstate", amd_pstate_param);
+early_param("amd_prefcore", amd_prefcore_param);
 
 MODULE_AUTHOR("Huang Rui <[email protected]>");
 MODULE_DESCRIPTION("AMD Processor P-state Frequency Driver");

include/linux/amd-pstate.h +4 -0

index 446394f84606..87e140e9e6db 100644
@@ -52,6 +52,9 @@ struct amd_aperf_mperf {
  * @prev: Last Aperf/Mperf/tsc count value read from register
  * @freq: current cpu frequency value
  * @boost_supported: check whether the Processor or SBIOS supports boost mode
+ * @hw_prefcore: check whether HW supports preferred core featue.
+ * 		  Only when hw_prefcore and early prefcore param are true,
+ * 		  AMD P-State driver supports preferred core featue.
  * @epp_policy: Last saved policy used to set energy-performance preference
  * @epp_cached: Cached CPPC energy-performance preference value
  * @policy: Cpufreq policy value

@@ -81,6 +84,7 @@ struct amd_cpudata {
 
 	u64	freq;
 	bool	boost_supported;
+	bool	hw_prefcore;
 
 	/* EPP feature related attributes*/
 	s16	epp_policy;

Signature · 1 line

2.34.1

drivers/cpufreq/amd-pstate.c +17 -0

+static void amd_pstate_init_prefcore(struct amd_cpudata *cpudata)
+{
+	int ret;
+	u32 highest_perf;
+	static u32 max_highest_perf = 0, min_highest_perf = U32_MAX;
+
+	ret = amd_pstate_get_highest_perf(cpudata->cpu, &highest_perf);
+	if (ret)
+		return;
+
+	cpudata->hw_prefcore = true;
+	/* check if CPPC preferred core feature is enabled*/
+	if (highest_perf == AMD_PSTATE_MAX_CPPC_PERF) {
+		pr_debug("AMD CPPC preferred core is unsupported!\n");
+		cpudata->hw_prefcore = false;
+		return;
+	}

drivers/cpufreq/amd-pstate.c +19 -0

+
+	if (max_highest_perf <= min_highest_perf) {
+		if (highest_perf > max_highest_perf)
+			max_highest_perf = highest_perf;
+
+		if (highest_perf < min_highest_perf)
+			min_highest_perf = highest_perf;
+
+		if (max_highest_perf > min_highest_perf) {
+			/*
+			 * This code can be run during CPU online under the
+			 * CPU hotplug locks, so sched_set_itmt_support()
+			 * cannot be called from here.  Queue up a work item
+			 * to invoke it.
+			 */
+			schedule_work(&sched_prefcore_work);
+		}
+	}
+}

-----Original Message-----
From: Peter Zijlstra <[email protected]>
Sent: Tuesday, October 10, 2023 6:36 PM
To: Meng, Li (Jassmine) <[email protected]>
Cc: Rafael J . Wysocki <[email protected]>; Huang, Ray <[email protected]>; [email protected]; linux- [email protected]; [email protected]; [email protected]; Shuah Khan <[email protected]>; [email protected]; Fontenot, Nathan <[email protected]>; Sharma, Deepak <[email protected]>; Deucher, Alexander <[email protected]>; Limonciello, Mario <[email protected]>; Huang, Shimmer <[email protected]>; Yuan, Perry <[email protected]>; Du, Xiaojian <[email protected]>; Viresh Kumar <[email protected]>; Borislav Petkov <[email protected]>
Subject: Re: [PATCH V8 3/7] cpufreq: amd-pstate: Enable amd-pstate preferred core supporting.

Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.

Quoted 16 lines · Meng Li

drivers/cpufreq/amd-pstate.c +16 -0

+static void amd_pstate_init_prefcore(struct amd_cpudata *cpudata) {
+     int ret;
+     u32 highest_perf;
+     static u32 max_highest_perf = 0, min_highest_perf = U32_MAX;
+
+     ret = amd_pstate_get_highest_perf(cpudata->cpu, &highest_perf);
+     if (ret)
+             return;
+
+     cpudata->hw_prefcore = true;
+     /* check if CPPC preferred core feature is enabled*/
+     if (highest_perf == AMD_PSTATE_MAX_CPPC_PERF) {
+             pr_debug("AMD CPPC preferred core is unsupported!\n");
+             cpudata->hw_prefcore = false;
+             return;
+     }

drivers/cpufreq/amd-pstate.c static void amd_perf_ctl_reset(unsigned int cpu) · line 742

Quoted 10 lines

drivers/cpufreq/amd-pstate.c +10 -0

+
+     if (!amd_pstate_prefcore)
+             return;
+
+     /*
+      * The priorities can be set regardless of whether or not
+      * sched_set_itmt_support(true) has been called and it is valid to
+      * update them at any time after it has been called.
+      */
+     sched_set_itmt_core_prio(highest_perf, cpudata->cpu);

You still got the whole u32 vs int thing confused, I've only pointed that out *TWICE* before.

Boris, can you pull out the clue hammer please?

Quoted 19 lines

drivers/cpufreq/amd-pstate.c +19 -0

+
+     if (max_highest_perf <= min_highest_perf) {
+             if (highest_perf > max_highest_perf)
+                     max_highest_perf = highest_perf;
+
+             if (highest_perf < min_highest_perf)
+                     min_highest_perf = highest_perf;
+
+             if (max_highest_perf > min_highest_perf) {
+                     /*
+                      * This code can be run during CPU online under the
+                      * CPU hotplug locks, so sched_set_itmt_support()
+                      * cannot be called from here.  Queue up a work item
+                      * to invoke it.
+                      */
+                     schedule_work(&sched_prefcore_work);
+             }
+     }
+}

ACPI 6.5 section 8.4.6.1.1.1 specifies that Notify event 0x85 can be emmitted to cause the the OSPM to re-evaluate the highest performance register. Add support for this event.

Reviewed-by Mario Limonciello <[email protected]>

Reviewed-by Huang Rui <[email protected]>

Signed-off-by Meng Li <[email protected]>

Link https://uefi.org/specs/ACPI/6.5/05_ACPI_Software_Programming_Model.html#processor-device-notification-values

---
 drivers/acpi/processor_driver.c |  6 ++++++
 drivers/cpufreq/cpufreq.c       | 13 +++++++++++++
 include/linux/cpufreq.h         |  5 +++++
 3 files changed, 24 insertions(+)

drivers/acpi/processor_driver.c +6 -0

index 4bd16b3f0781..29b2fb68a35d 100644
@@ -27,6 +27,7 @@
 #define ACPI_PROCESSOR_NOTIFY_PERFORMANCE 0x80
 #define ACPI_PROCESSOR_NOTIFY_POWER	0x81
 #define ACPI_PROCESSOR_NOTIFY_THROTTLING	0x82
+#define ACPI_PROCESSOR_NOTIFY_HIGEST_PERF_CHANGED	0x85
 
 MODULE_AUTHOR("Paul Diefenbaugh");
 MODULE_DESCRIPTION("ACPI Processor Driver");

@@ -83,6 +84,11 @@ static void acpi_processor_notify(acpi_handle handle, u32 event, void *data)
 		acpi_bus_generate_netlink_event(device->pnp.device_class,
 						  dev_name(&device->dev), event, 0);
 		break;
+	case ACPI_PROCESSOR_NOTIFY_HIGEST_PERF_CHANGED:
+		cpufreq_update_highest_perf(pr->id);
+		acpi_bus_generate_netlink_event(device->pnp.device_class,
+						  dev_name(&device->dev), event, 0);
+		break;
 	default:
 		acpi_handle_debug(handle, "Unsupported event [0x%x]\n", event);
 		break;

drivers/cpufreq/cpufreq.c +13 -0

index 60ed89000e82..4ada787ff105 100644
@@ -2718,6 +2718,19 @@ void cpufreq_update_limits(unsigned int cpu)
 }
 EXPORT_SYMBOL_GPL(cpufreq_update_limits);
 
+/**
+ * cpufreq_update_highest_perf - Update highest performance for a given CPU.
+ * @cpu: CPU to update the highest performance for.
+ *
+ * Invoke the driver's ->update_highest_perf callback if present
+ */
+void cpufreq_update_highest_perf(unsigned int cpu)
+{
+	if (cpufreq_driver->update_highest_perf)
+		cpufreq_driver->update_highest_perf(cpu);
+}
+EXPORT_SYMBOL_GPL(cpufreq_update_highest_perf);
+
 /*********************************************************************
  *               BOOST						     *
  *********************************************************************/

include/linux/cpufreq.h +5 -0

index 71d186d6933a..1cc1241fb698 100644
@@ -235,6 +235,7 @@ int cpufreq_get_policy(struct cpufreq_policy *policy, unsigned int cpu);
 void refresh_frequency_limits(struct cpufreq_policy *policy);
 void cpufreq_update_policy(unsigned int cpu);
 void cpufreq_update_limits(unsigned int cpu);
+void cpufreq_update_highest_perf(unsigned int cpu);
 bool have_governor_per_policy(void);
 bool cpufreq_supports_freq_invariance(void);
 struct kobject *get_governor_parent_kobj(struct cpufreq_policy *policy);

@@ -263,6 +264,7 @@ static inline bool cpufreq_supports_freq_invariance(void)
 	return false;
 }
 static inline void disable_cpufreq(void) { }
+static inline void cpufreq_update_highest_perf(unsigned int cpu) { }
 #endif
 
 #ifdef CONFIG_CPU_FREQ_STAT

@@ -380,6 +382,9 @@ struct cpufreq_driver {
 	/* Called to update policy limits on firmware notifications. */
 	void		(*update_limits)(unsigned int cpu);
 
+	/* Called to update highest performance on firmware notifications. */
+	void		(*update_highest_perf)(unsigned int cpu);
+
 	/* optional */
 	int		(*bios_limit)(int cpu, unsigned int *limit);
 

Signature · 1 line

2.34.1

Preferred core rankings can be changed dynamically by the platform based on the workload and platform conditions and accounting for thermals and aging.
When this occurs, cpu priority need to be set.

Reviewed-by Mario Limonciello <[email protected]>

Reviewed-by Wyes Karny <[email protected]>

Reviewed-by Huang Rui <[email protected]>

Signed-off-by Meng Li <[email protected]>

---
 drivers/cpufreq/amd-pstate.c | 34 ++++++++++++++++++++++++++++++++--
 include/linux/amd-pstate.h   |  6 ++++++
 2 files changed, 38 insertions(+), 2 deletions(-)

drivers/cpufreq/amd-pstate.c +2 -0

index 6ac8939fce5a..d3369247c6c9 100644
@@ -313,6 +313,7 @@ static int pstate_init_perf(struct amd_cpudata *cpudata)
 	WRITE_ONCE(cpudata->nominal_perf, AMD_CPPC_NOMINAL_PERF(cap1));
 	WRITE_ONCE(cpudata->lowest_nonlinear_perf, AMD_CPPC_LOWNONLIN_PERF(cap1));
 	WRITE_ONCE(cpudata->lowest_perf, AMD_CPPC_LOWEST_PERF(cap1));
+	WRITE_ONCE(cpudata->prefcore_ranking, AMD_CPPC_HIGHEST_PERF(cap1));
 
 	return 0;
 }

@@ -334,6 +335,7 @@ static int cppc_init_perf(struct amd_cpudata *cpudata)
 	WRITE_ONCE(cpudata->lowest_nonlinear_perf,
 		   cppc_perf.lowest_nonlinear_perf);
 	WRITE_ONCE(cpudata->lowest_perf, cppc_perf.lowest_perf);
+	WRITE_ONCE(cpudata->prefcore_ranking, cppc_perf.highest_perf);
 
 	if (cppc_state == AMD_PSTATE_ACTIVE)
 		return 0;

drivers/cpufreq/amd-pstate.c +26 -0

 	if (_min_perf < capacity)
 		min_perf = DIV_ROUND_UP(cap_perf * _min_perf, capacity);
 

@@ -760,6 +762,32 @@ static void amd_pstate_init_prefcore(struct amd_cpudata *cpudata)
 	}
 }
 
+static void amd_pstate_update_highest_perf(unsigned int cpu)
+{
+	struct cpufreq_policy *policy;
+	struct amd_cpudata *cpudata;
+	u32 prev_high = 0, cur_high = 0;
+	int ret;
+
+	if ((!amd_pstate_prefcore) || (!cpudata->hw_prefcore))
+		return;
+
+	ret = amd_pstate_get_highest_perf(cpu, &cur_high);
+	if (ret)
+		return;
+
+	policy = cpufreq_cpu_get(cpu);
+	cpudata = policy->driver_data;
+	prev_high = READ_ONCE(cpudata->prefcore_ranking);
+
+	if (prev_high != cur_high) {
+		WRITE_ONCE(cpudata->prefcore_ranking, cur_high);
+		sched_set_itmt_core_prio(cur_high, cpu);
+	}
+
+	cpufreq_cpu_put(policy);
+}
+
 static int amd_pstate_cpu_init(struct cpufreq_policy *policy)
 {
 	int min_freq, max_freq, nominal_freq, lowest_nonlinear_freq, ret;

drivers/cpufreq/amd-pstate.c +2 -0

 	return sysfs_emit(buf, "%u\n", perf);
 }

@@ -1502,6 +1530,7 @@ static struct cpufreq_driver amd_pstate_driver = {
 	.suspend	= amd_pstate_cpu_suspend,
 	.resume		= amd_pstate_cpu_resume,
 	.set_boost	= amd_pstate_set_boost,
+	.update_highest_perf	= amd_pstate_update_highest_perf,
 	.name		= "amd-pstate",
 	.attr		= amd_pstate_attr,
 };

@@ -1516,6 +1545,7 @@ static struct cpufreq_driver amd_pstate_epp_driver = {
 	.online		= amd_pstate_epp_cpu_online,
 	.suspend	= amd_pstate_epp_suspend,
 	.resume		= amd_pstate_epp_resume,
+	.update_highest_perf	= amd_pstate_update_highest_perf,
 	.name		= "amd-pstate-epp",
 	.attr		= amd_pstate_epp_attr,
 };

include/linux/amd-pstate.h +6 -0

index 87e140e9e6db..426822612373 100644
@@ -39,11 +39,16 @@ struct amd_aperf_mperf {
  * @cppc_req_cached: cached performance request hints
  * @highest_perf: the maximum performance an individual processor may reach,
  *		  assuming ideal conditions
+ *		  For platforms that do not support the preferred core feature, the
+ *		  highest_pef may be configured with 166 or 255, to avoid max frequency
+ *		  calculated wrongly. we take the fixed value as the highest_perf.
  * @nominal_perf: the maximum sustained performance level of the processor,
  *		  assuming ideal operating conditions
  * @lowest_nonlinear_perf: the lowest performance level at which nonlinear power
  *			   savings are achieved
  * @lowest_perf: the absolute lowest performance level of the processor
+ * @prefcore_ranking: the preferred core ranking, the higher value indicates a higher
+ * 		  priority.
  * @max_freq: the frequency that mapped to highest_perf
  * @min_freq: the frequency that mapped to lowest_perf
  * @nominal_freq: the frequency that mapped to nominal_perf

@@ -73,6 +78,7 @@ struct amd_cpudata {
 	u32	nominal_perf;
 	u32	lowest_nonlinear_perf;
 	u32	lowest_perf;
+	u32     prefcore_ranking;
 
 	u32	max_freq;
 	u32	min_freq;

Signature · 1 line

2.34.1

-----Original Message-----
From: Karny, Wyes <[email protected]>
Sent: Monday, October 9, 2023 2:19 PM
To: Meng, Li (Jassmine) <[email protected]>
Cc: Karny, Wyes <[email protected]>; Rafael J . Wysocki <[email protected]>; Huang, Ray <[email protected]>; linux- [email protected]; [email protected]; [email protected]; linux- [email protected]; Shuah Khan <[email protected]>; linux- [email protected]; Fontenot, Nathan <[email protected]>; Sharma, Deepak <[email protected]>; Deucher, Alexander <[email protected]>; Limonciello, Mario <[email protected]>; Huang, Shimmer <[email protected]>; Yuan, Perry <[email protected]>; Du, Xiaojian <[email protected]>; Viresh Kumar <[email protected]>; Borislav Petkov <[email protected]>
Subject: Re: [PATCH V8 5/7] cpufreq: amd-pstate: Update amd-pstate preferred core ranking dynamically

Hi Meng Li,

Quoted 11 lines · Meng Li

Preferred core rankings can be changed dynamically by the platform based on the workload and platform conditions and accounting for thermals and aging.
When this occurs, cpu priority need to be set.

Reviewed-by Mario Limonciello <[email protected]>

Reviewed-by Wyes Karny <[email protected]>

Reviewed-by Huang Rui <[email protected]>

Signed-off-by Meng Li <[email protected]>

---
 drivers/cpufreq/amd-pstate.c | 34

++++++++++++++++++++++++++++++++--

drivers/cpufreq/amd-pstate.c static int pstate_init_perf(struct amd_cpudata

Quoted 8 lines

 include/linux/amd-pstate.h   |  6 ++++++
 2 files changed, 38 insertions(+), 2 deletions(-)

(quoted diff) +0 -0

b/drivers/cpufreq/amd-pstate.c index 6ac8939fce5a..d3369247c6c9 100644

drivers/cpufreq/amd-pstate.c +0 -0

@@ -313,6 +313,7 @@ static int pstate_init_perf(struct amd_cpudata

*cpudata)

drivers/cpufreq/amd-pstate.c static int pstate_init_perf(struct amd_cpudata · line 313

Quoted 1 line

drivers/cpufreq/amd-pstate.c +0 -0

    WRITE_ONCE(cpudata->nominal_perf,

AMD_CPPC_NOMINAL_PERF(cap1));

drivers/cpufreq/amd-pstate.c static int cppc_init_perf(struct amd_cpudata *cpudata) · line 335

Quoted 1 line

drivers/cpufreq/amd-pstate.c +0 -0

    WRITE_ONCE(cpudata->lowest_nonlinear_perf,

AMD_CPPC_LOWNONLIN_PERF(cap1));

drivers/cpufreq/amd-pstate.c static int pstate_init_perf(struct amd_cpudata · line 315

Quoted 1 line

drivers/cpufreq/amd-pstate.c +0 -0

    WRITE_ONCE(cpudata->lowest_perf,

AMD_CPPC_LOWEST_PERF(cap1));

drivers/cpufreq/amd-pstate.c static int pstate_init_perf(struct amd_cpudata · line 316

Quoted 1 line

drivers/cpufreq/amd-pstate.c +1 -0

+   WRITE_ONCE(cpudata->prefcore_ranking,

AMD_CPPC_HIGHEST_PERF(cap1));

drivers/cpufreq/amd-pstate.c static int cppc_init_perf(struct amd_cpudata

Quoted 4 lines

drivers/cpufreq/amd-pstate.c +0 -0

    return 0;
 }

@@ -334,6 +335,7 @@ static int cppc_init_perf(struct amd_cpudata

*cpudata)

drivers/cpufreq/amd-pstate.c static void amd_pstate_adjust_perf(unsigned int

Quoted 8 lines

drivers/cpufreq/amd-pstate.c +1 -0

    WRITE_ONCE(cpudata->lowest_nonlinear_perf,
               cppc_perf.lowest_nonlinear_perf);
    WRITE_ONCE(cpudata->lowest_perf, cppc_perf.lowest_perf);
+   WRITE_ONCE(cpudata->prefcore_ranking, cppc_perf.highest_perf);
 
    if (cppc_state == AMD_PSTATE_ACTIVE)
            return 0;

@@ -540,7 +542,7 @@ static void amd_pstate_adjust_perf(unsigned int

cpu,

drivers/cpufreq/amd-pstate.c static void amd_pstate_adjust_perf(unsigned int · line 543

Quoted 2 lines

drivers/cpufreq/amd-pstate.c +0 -0

    if (target_perf < capacity)
            des_perf = DIV_ROUND_UP(cap_perf * target_perf,

capacity);

drivers/cpufreq/amd-pstate.c static void amd_pstate_adjust_perf(unsigned int cpu, · line 545

Quoted 3 lines

drivers/cpufreq/amd-pstate.c +1 -1

-   min_perf = READ_ONCE(cpudata->highest_perf);
+   min_perf = READ_ONCE(cpudata->lowest_perf);

This seems to be a fix. So, this could be a separate patch.

drivers/cpufreq/amd-pstate.c static void amd_pstate_init_prefcore(struct

Quoted 5 lines

drivers/cpufreq/amd-pstate.c +0 -0

    if (_min_perf < capacity)
            min_perf = DIV_ROUND_UP(cap_perf * _min_perf, capacity);
 

@@ -760,6 +762,32 @@ static void amd_pstate_init_prefcore(struct

amd_cpudata *cpudata)

Quoted 20 lines

drivers/cpufreq/amd-pstate.c +17 -0

    }
 }
 
+static void amd_pstate_update_highest_perf(unsigned int cpu) {
+   struct cpufreq_policy *policy;
+   struct amd_cpudata *cpudata;
+   u32 prev_high = 0, cur_high = 0;
+   int ret;
+
+   if ((!amd_pstate_prefcore) || (!cpudata->hw_prefcore))
+           return;
+
+   ret = amd_pstate_get_highest_perf(cpu, &cur_high);
+   if (ret)
+           return;
+
+   policy = cpufreq_cpu_get(cpu);
+   cpudata = policy->driver_data;
+   prev_high = READ_ONCE(cpudata->prefcore_ranking);
+

drivers/cpufreq/amd-pstate.c static void amd_pstate_init_prefcore(struct amd_cpudata *cpudata) · line 793

Quoted 10 lines

drivers/cpufreq/amd-pstate.c +8 -0

+   if (prev_high != cur_high) {
+           WRITE_ONCE(cpudata->prefcore_ranking, cur_high);
+           sched_set_itmt_core_prio(cur_high, cpu);
+   }
+
+   cpufreq_cpu_put(policy);
+}
+
 static int amd_pstate_cpu_init(struct cpufreq_policy *policy)  {
    int min_freq, max_freq, nominal_freq, lowest_nonlinear_freq, ret;

@@

drivers/cpufreq/amd-pstate.c static void amd_pstate_init_prefcore(struct · line 765

Quoted 1 line

drivers/cpufreq/amd-pstate.c +0 -1

-926,7 +954,7 @@ static ssize_t show_amd_pstate_highest_perf(struct

cpufreq_policy *policy,

drivers/cpufreq/amd-pstate.c static ssize_t show_amd_pstate_highest_perf(struct cpufreq_policy *policy, · line 957

Quoted 5 lines

drivers/cpufreq/amd-pstate.c +1 -1

    u32 perf;
    struct amd_cpudata *cpudata = policy->driver_data;
 
-   perf = READ_ONCE(cpudata->highest_perf);
+   perf = READ_ONCE(cpudata->prefcore_ranking);

I think this should show cpudata->highest_perf.

Thanks,
Wyes

drivers/cpufreq/amd-pstate.c static struct cpufreq_driver

Quoted 11 lines

    return sysfs_emit(buf, "%u\n", perf);  } @@ -1502,6 +1530,7 @@
static struct cpufreq_driver amd_pstate_driver = {
    .suspend        = amd_pstate_cpu_suspend,
    .resume         = amd_pstate_cpu_resume,
    .set_boost      = amd_pstate_set_boost,

drivers/cpufreq/amd-pstate.c +1 -0

+   .update_highest_perf    = amd_pstate_update_highest_perf,
    .name           = "amd-pstate",
    .attr           = amd_pstate_attr,
 };

@@ -1516,6 +1545,7 @@ static struct cpufreq_driver

amd_pstate_epp_driver = {

Quoted 12 lines

drivers/cpufreq/amd-pstate.c +1 -0

    .online         = amd_pstate_epp_cpu_online,
    .suspend        = amd_pstate_epp_suspend,
    .resume         = amd_pstate_epp_resume,
+   .update_highest_perf    = amd_pstate_update_highest_perf,
    .name           = "amd-pstate-epp",
    .attr           = amd_pstate_epp_attr,
 };

include/linux/amd-pstate.h +0 -0

index 87e140e9e6db..426822612373 100644

include/linux/amd-pstate.h struct amd_aperf_mperf { · line 40

Quoted 2 lines

include/linux/amd-pstate.h +0 -0

@@ -39,11 +39,16 @@ struct amd_aperf_mperf {
  * @cppc_req_cached: cached performance request hints
  * @highest_perf: the maximum performance an individual processor may

reach,

include/linux/amd-pstate.h struct amd_aperf_mperf { · line 42

Quoted 2 lines

include/linux/amd-pstate.h +1 -0

  *           assuming ideal conditions
+ *           For platforms that do not support the preferred core

feature, the

include/linux/amd-pstate.h struct amd_aperf_mperf { · line 43

Quoted 1 line

include/linux/amd-pstate.h +1 -0

+ *           highest_pef may be configured with 166 or 255, to avoid

max frequency

include/linux/amd-pstate.h struct amd_aperf_mperf { · line 44

Quoted 1 line

include/linux/amd-pstate.h +1 -0

+ *           calculated wrongly. we take the fixed value as the

highest_perf.

include/linux/amd-pstate.h struct amd_aperf_mperf { · line 45

Quoted 1 line

include/linux/amd-pstate.h +0 -0

  * @nominal_perf: the maximum sustained performance level of the

processor,

include/linux/amd-pstate.h struct amd_aperf_mperf { · line 47

Quoted 2 lines

include/linux/amd-pstate.h +0 -0

  *           assuming ideal operating conditions
  * @lowest_nonlinear_perf: the lowest performance level at which

nonlinear power

Quoted 4 lines

include/linux/amd-pstate.h +0 -0

  *                    savings are achieved
  * @lowest_perf: the absolute lowest performance level of the

processor
+ * @prefcore_ranking: the preferred core ranking, the higher value

indicates a higher

Quoted 15 lines

+ * priority.
* @max_freq: the frequency that mapped to highest_perf
* @min_freq: the frequency that mapped to lowest_perf
* @nominal_freq: the frequency that mapped to nominal_perf @@ -73,6

+78,7 @@ struct amd_cpudata {
    u32     nominal_perf;
    u32     lowest_nonlinear_perf;
    u32     lowest_perf;

+ u32 prefcore_ranking;

    u32     max_freq;
    u32     min_freq;

--
2.34.1

Introduce amd-pstate preferred core.

check preferred core state set by the kernel parameter:

$ cat /sys/devices/system/cpu/amd-pstate/prefcore

Reviewed-by Mario Limonciello <[email protected]>

Reviewed-by Huang Rui <[email protected]>

Signed-off-by Meng Li <[email protected]>

---
 Documentation/admin-guide/pm/amd-pstate.rst | 59 ++++++++++++++++++++-
 1 file changed, 57 insertions(+), 2 deletions(-)

Documentation/admin-guide/pm/amd-pstate.rst +57 -2

index 1cf40f69278c..0b832ff529db 100644
@@ -300,8 +300,8 @@ platforms. The AMD P-States mechanism is the more performance and energy
 efficiency frequency management method on AMD processors.
 
 
-AMD Pstate Driver Operation Modes
-=================================
+``amd-pstate`` Driver Operation Modes
+======================================
 
 ``amd_pstate`` CPPC has 3 operation modes: autonomous (active) mode,
 non-autonomous (passive) mode and guided autonomous (guided) mode.

@@ -353,6 +353,48 @@ is activated.  In this mode, driver requests minimum and maximum performance
 level and the platform autonomously selects a performance level in this range
 and appropriate to the current workload.
 
+``amd-pstate`` Preferred Core
+=================================
+
+The core frequency is subjected to the process variation in semiconductors.
+Not all cores are able to reach the maximum frequency respecting the
+infrastructure limits. Consequently, AMD has redefined the concept of
+maximum frequency of a part. This means that a fraction of cores can reach
+maximum frequency. To find the best process scheduling policy for a given
+scenario, OS needs to know the core ordering informed by the platform through
+highest performance capability register of the CPPC interface.
+
+``amd-pstate`` preferred core enables the scheduler to prefer scheduling on
+cores that can achieve a higher frequency with lower voltage. The preferred
+core rankings can dynamically change based on the workload, platform conditions,
+thermals and ageing.
+
+The priority metric will be initialized by the ``amd-pstate`` driver. The ``amd-pstate``
+driver will also determine whether or not ``amd-pstate`` preferred core is
+supported by the platform.
+
+``amd-pstate`` driver will provide an initial core ordering when the system boots.
+The platform uses the CPPC interfaces to communicate the core ranking to the
+operating system and scheduler to make sure that OS is choosing the cores
+with highest performance firstly for scheduling the process. When ``amd-pstate``
+driver receives a message with the highest performance change, it will
+update the core ranking and set the cpu's priority.
+
+``amd-pstate`` Preferred Core Switch
+=================================
+Kernel Parameters
+-----------------
+
+``amd-pstate`` peferred core`` has two states: enable and disable.
+Enable/disable states can be chosen by different kernel parameters.
+Default enable ``amd-pstate`` preferred core.
+
+``amd_prefcore=disable``
+
+For systems that support ``amd-pstate`` preferred core, the core rankings will
+always be advertised by the platform. But OS can choose to ignore that via the
+kernel parameter ``amd_prefcore=disable``.
+
 User Space Interface in ``sysfs`` - General
 ===========================================
 

@@ -385,6 +427,19 @@ control its functionality at the system level.  They are located in the
         to the operation mode represented by that string - or to be
         unregistered in the "disable" case.
 
+``prefcore``
+	Preferred core state of the driver: "enabled" or "disabled".
+
+	"enabled"
+		Enable the ``amd-pstate`` preferred core.
+
+	"disabled"
+		Disable the ``amd-pstate`` preferred core
+
+
+        This attribute is read-only to check the state of preferred core set
+        by the kernel parameter.
+
 ``cpupower`` tool support for ``amd-pstate``
 ===============================================
 

Signature · 1 line

2.34.1

Hi all:

The core frequency is subjected to the process variation in semiconductors. Not all cores are able to reach the maximum frequency respecting the infrastructure limits. Consequently, AMD has redefined the concept of maximum frequency of a part. This means that a fraction of cores can reach maximum frequency. To find the best process scheduling policy for a given scenario, OS needs to know the core ordering informed by the platform through highest performance capability register of the CPPC interface.

Earlier implementations of amd-pstate preferred core only support a static core ranking and targeted performance. Now it has the ability to dynamically change the preferred core based on the workload and platform conditions and accounting for thermals and aging.

Amd-pstate driver utilizes the functions and data structures provided by the ITMT architecture to enable the scheduler to favor scheduling on cores which can be get a higher frequency with lower voltage.
We call it amd-pstate preferred core.

Here sched_set_itmt_core_prio() is called to set priorities and sched_set_itmt_support() is called to enable ITMT feature. Amd-pstate driver uses the highest performance value to indicate the priority of CPU. The higher value has a higher priority.

Amd-pstate driver will provide an initial core ordering at boot time. It relies on the CPPC interface to communicate the core ranking to the operating system and scheduler to make sure that OS is choosing the cores with highest performance firstly for scheduling the process. When amd-pstate driver receives a message with the highest performance change, it will update the core ranking.

Changes form V7->V8:
- all:
- - pick up Review-By flag added by Mario and Ray.
- cpufreq: amd-pstate:
- - use hw_prefcore embeds into cpudata structure.
- - delete preferred core init from cpu online/off.

Changes form V6->V7:
- x86:
- - Modify kconfig about X86_AMD_PSTATE.
- cpufreq: amd-pstate:
- - modify incorrect comments about scheduler_work().
- - convert highest_perf data type.
- - modify preferred core init when cpu init and online.
- acpi: cppc:
- - modify link of CPPC highest performance.
- cpufreq:
- - modify link of CPPC highest performance changed.

Changes form V5->V6:
- cpufreq: amd-pstate:
- - modify the wrong tag order.
- - modify warning about hw_prefcore sysfs attribute.
- - delete duplicate comments.
- - modify the variable name cppc_highest_perf to prefcore_ranking.
- - modify judgment conditions for setting highest_perf.
- - modify sysfs attribute for CPPC highest perf to pr_debug message.
- Documentation: amd-pstate:
- - modify warning: title underline too short.

Changes form V4->V5:
- cpufreq: amd-pstate:
- - modify sysfs attribute for CPPC highest perf.
- - modify warning about comments
- - rebase linux-next
- cpufreq:
- - Moidfy warning about function declarations.
- Documentation: amd-pstate:
- - align with ``amd-pstat``

Changes form V3->V4:
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.

Changes form V2->V3:
- x86:
- - Modify kconfig and description.
- cpufreq: amd-pstate:
- - Add Co-developed-by tag in commit message.
- cpufreq:
- - Modify commit message.
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.

Changes form V1->V2:
- acpi: cppc:
- - Add reference link.
- cpufreq:
- - Moidfy link error.
- cpufreq: amd-pstate:
- - Init the priorities of all online CPUs
- - Use a single variable to represent the status of preferred core.
- Documentation:
- - Default enabled preferred core.
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.
- - Default enabled preferred core.
- - Use a single variable to represent the status of preferred core.

Meng Li (7):
x86: Drop CPU_SUP_INTEL from SCHED_MC_PRIO for the expansion.
acpi: cppc: Add get the highest performance cppc control
cpufreq: amd-pstate: Enable amd-pstate preferred core supporting.
cpufreq: Add a notification message that the highest perf has changed
cpufreq: amd-pstate: Update amd-pstate preferred core ranking
dynamically
Documentation: amd-pstate: introduce amd-pstate preferred core
Documentation: introduce amd-pstate preferrd core mode kernel command
line options

 .../admin-guide/kernel-parameters.txt         |   5 +
 Documentation/admin-guide/pm/amd-pstate.rst   |  59 +++++-
 arch/x86/Kconfig                              |   5 +-
 drivers/acpi/cppc_acpi.c                      |  13 ++
 drivers/acpi/processor_driver.c               |   6 +
 drivers/cpufreq/amd-pstate.c                  | 186 ++++++++++++++++--
 drivers/cpufreq/cpufreq.c                     |  13 ++
 include/acpi/cppc_acpi.h                      |   5 +
 include/linux/amd-pstate.h                    |  10 +
 include/linux/cpufreq.h                       |   5 +
 10 files changed, 285 insertions(+), 22 deletions(-)

-----Original Message-----
From: Oleksandr Natalenko <[email protected]>
Sent: Monday, October 9, 2023 2:55 PM
To: Rafael J . Wysocki <[email protected]>; Huang, Ray <[email protected]>; Meng, Li (Jassmine) <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Shuah Khan <[email protected]>; [email protected]; Fontenot, Nathan <[email protected]>; Sharma, Deepak <[email protected]>; Deucher, Alexander <[email protected]>; Limonciello, Mario <[email protected]>; Huang, Shimmer <[email protected]>; Yuan, Perry <[email protected]>; Du, Xiaojian <[email protected]>; Viresh Kumar <[email protected]>; Borislav Petkov <[email protected]>; Meng, Li (Jassmine) <[email protected]>
Subject: Re: [PATCH V8 0/7] amd-pstate preferred core

Hello.

Quoted 9 lines · On pondělí 9. října 2023 4:49:25 CEST Meng Li wrote:

Hi all:

The core frequency is subjected to the process variation in semiconductors. Not all cores are able to reach the maximum frequency respecting the infrastructure limits. Consequently, AMD has redefined the concept of maximum frequency of a part. This means that a fraction of cores can reach maximum frequency. To find the best process scheduling policy for a given scenario, OS needs to know the core ordering informed by the platform through highest performance capability register of the CPPC

interface.

Quoted 23 lines

Earlier implementations of amd-pstate preferred core only support a static core ranking and targeted performance. Now it has the ability to dynamically change the preferred core based on the workload and platform conditions and accounting for thermals and aging.

Amd-pstate driver utilizes the functions and data structures provided by the ITMT architecture to enable the scheduler to favor scheduling on cores which can be get a higher frequency with lower voltage. We call it amd-pstate preferred core.

Here sched_set_itmt_core_prio() is called to set priorities and sched_set_itmt_support() is called to enable ITMT feature. Amd-pstate driver uses the highest performance value to indicate the priority of CPU. The higher value has a higher priority.

Amd-pstate driver will provide an initial core ordering at boot time. It relies on the CPPC interface to communicate the core ranking to the operating system and scheduler to make sure that OS is choosing the cores with highest performance firstly for scheduling the process. When amd-pstate driver receives a message with the highest performance change, it will update the core ranking.

Quoted 6 lines

Changes form V7->V8:
- all:
- - pick up Review-By flag added by Mario and Ray.
- cpufreq: amd-pstate:
- - use hw_prefcore embeds into cpudata structure.
- - delete preferred core init from cpu online/off.

Could you please let me know if this change means a fix for the report I've sent previously? [1]

Would you also be able to Cc me on the next iteration of this patchset?

Thank you!

[1] https://lore.kernel.org/lkml/[email protected]/

Quoted 63 lines

Changes form V6->V7:
- x86:
- - Modify kconfig about X86_AMD_PSTATE.
- cpufreq: amd-pstate:
- - modify incorrect comments about scheduler_work().
- - convert highest_perf data type.
- - modify preferred core init when cpu init and online.
- acpi: cppc:
- - modify link of CPPC highest performance.
- cpufreq:
- - modify link of CPPC highest performance changed.

Changes form V5->V6:
- cpufreq: amd-pstate:
- - modify the wrong tag order.
- - modify warning about hw_prefcore sysfs attribute.
- - delete duplicate comments.
- - modify the variable name cppc_highest_perf to prefcore_ranking.
- - modify judgment conditions for setting highest_perf.
- - modify sysfs attribute for CPPC highest perf to pr_debug message.
- Documentation: amd-pstate:
- - modify warning: title underline too short.

Changes form V4->V5:
- cpufreq: amd-pstate:
- - modify sysfs attribute for CPPC highest perf.
- - modify warning about comments
- - rebase linux-next
- cpufreq:
- - Moidfy warning about function declarations.
- Documentation: amd-pstate:
- - align with ``amd-pstat``

Changes form V3->V4:
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.

Changes form V2->V3:
- x86:
- - Modify kconfig and description.
- cpufreq: amd-pstate:
- - Add Co-developed-by tag in commit message.
- cpufreq:
- - Modify commit message.
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.

Changes form V1->V2:
- acpi: cppc:
- - Add reference link.
- cpufreq:
- - Moidfy link error.
- cpufreq: amd-pstate:
- - Init the priorities of all online CPUs
- - Use a single variable to represent the status of preferred core.
- Documentation:
- - Default enabled preferred core.
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.
- - Default enabled preferred core.
- - Use a single variable to represent the status of preferred core.

Quoted 9 lines

Meng Li (7):
x86: Drop CPU_SUP_INTEL from SCHED_MC_PRIO for the expansion.
acpi: cppc: Add get the highest performance cppc control
cpufreq: amd-pstate: Enable amd-pstate preferred core supporting.
cpufreq: Add a notification message that the highest perf has changed
cpufreq: amd-pstate: Update amd-pstate preferred core ranking
dynamically
Documentation: amd-pstate: introduce amd-pstate preferred core
Documentation: introduce amd-pstate preferrd core mode kernel

command

Quoted 15 lines

line options

 .../admin-guide/kernel-parameters.txt         |   5 +
 Documentation/admin-guide/pm/amd-pstate.rst   |  59 +++++-
 arch/x86/Kconfig                              |   5 +-
 drivers/acpi/cppc_acpi.c                      |  13 ++
 drivers/acpi/processor_driver.c               |   6 +
 drivers/cpufreq/amd-pstate.c                  | 186 ++++++++++++++++--
 drivers/cpufreq/cpufreq.c                     |  13 ++
 include/acpi/cppc_acpi.h                      |   5 +
 include/linux/amd-pstate.h                    |  10 +
 include/linux/cpufreq.h                       |   5 +
 10 files changed, 285 insertions(+), 22 deletions(-)

--
Oleksandr Natalenko (post-factum)

[AMD Official Use Only - General]

Hi Oleksandr:

Quoted 60 lines

-----Original Message-----
From: Oleksandr Natalenko <[email protected]>
Sent: Monday, October 9, 2023 2:55 PM
To: Rafael J . Wysocki <[email protected]>; Huang, Ray <[email protected]>; Meng, Li (Jassmine) <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Shuah Khan <[email protected]>; [email protected]; Fontenot, Nathan <[email protected]>; Sharma, Deepak <[email protected]>; Deucher, Alexander <[email protected]>; Limonciello, Mario <[email protected]>; Huang, Shimmer <[email protected]>; Yuan, Perry <[email protected]>; Du, Xiaojian <[email protected]>; Viresh Kumar <[email protected]>; Borislav Petkov <[email protected]>; Meng, Li (Jassmine) <[email protected]>
Subject: Re: [PATCH V8 0/7] amd-pstate preferred core

Hello.

Quoted 9 lines · On pondělí 9. října 2023 4:49:25 CEST Meng Li wrote:

Hi all:

The core frequency is subjected to the process variation in semiconductors. Not all cores are able to reach the maximum frequency respecting the infrastructure limits. Consequently, AMD has redefined the concept of maximum frequency of a part. This means that a fraction of cores can reach maximum frequency. To find the best process scheduling policy for a given scenario, OS needs to know the core ordering informed by the platform through highest performance capability register of the CPPC

interface.

Quoted 29 lines

Earlier implementations of amd-pstate preferred core only support a static core ranking and targeted performance. Now it has the ability to dynamically change the preferred core based on the workload and platform conditions and accounting for thermals and aging.

Amd-pstate driver utilizes the functions and data structures provided by the ITMT architecture to enable the scheduler to favor scheduling on cores which can be get a higher frequency with lower voltage. We call it amd-pstate preferred core.

Here sched_set_itmt_core_prio() is called to set priorities and sched_set_itmt_support() is called to enable ITMT feature. Amd-pstate driver uses the highest performance value to indicate the priority of CPU. The higher value has a higher priority.

Amd-pstate driver will provide an initial core ordering at boot time. It relies on the CPPC interface to communicate the core ranking to the operating system and scheduler to make sure that OS is choosing the cores with highest performance firstly for scheduling the process. When amd-pstate driver receives a message with the highest performance change, it will update the core ranking.

Changes form V7->V8:
- all:
- - pick up Review-By flag added by Mario and Ray.
- cpufreq: amd-pstate:
- - use hw_prefcore embeds into cpudata structure.
- - delete preferred core init from cpu online/off.

Quoted 3 lines

Could you please let me know if this change means a fix for the report I've sent previously? [1]

[Meng, Li (Jassmine)] Yes.
I have deleted online handle function of amd pstate driver. It doesn't re-initialize preferred core.
This online function will set incorrect des perf value.

Quoted 1 line

Would you also be able to Cc me on the next iteration of this patchset?

[Meng, Li (Jassmine)] OK.

Quoted 97 lines

Thank you!

[1] https://lore.kernel.org/lkml/[email protected]/

Quoted 63 lines

Changes form V6->V7:
- x86:
- - Modify kconfig about X86_AMD_PSTATE.
- cpufreq: amd-pstate:
- - modify incorrect comments about scheduler_work().
- - convert highest_perf data type.
- - modify preferred core init when cpu init and online.
- acpi: cppc:
- - modify link of CPPC highest performance.
- cpufreq:
- - modify link of CPPC highest performance changed.

Changes form V5->V6:
- cpufreq: amd-pstate:
- - modify the wrong tag order.
- - modify warning about hw_prefcore sysfs attribute.
- - delete duplicate comments.
- - modify the variable name cppc_highest_perf to prefcore_ranking.
- - modify judgment conditions for setting highest_perf.
- - modify sysfs attribute for CPPC highest perf to pr_debug message.
- Documentation: amd-pstate:
- - modify warning: title underline too short.

Changes form V4->V5:
- cpufreq: amd-pstate:
- - modify sysfs attribute for CPPC highest perf.
- - modify warning about comments
- - rebase linux-next
- cpufreq:
- - Moidfy warning about function declarations.
- Documentation: amd-pstate:
- - align with ``amd-pstat``

Changes form V3->V4:
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.

Changes form V2->V3:
- x86:
- - Modify kconfig and description.
- cpufreq: amd-pstate:
- - Add Co-developed-by tag in commit message.
- cpufreq:
- - Modify commit message.
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.

Changes form V1->V2:
- acpi: cppc:
- - Add reference link.
- cpufreq:
- - Moidfy link error.
- cpufreq: amd-pstate:
- - Init the priorities of all online CPUs
- - Use a single variable to represent the status of preferred core.
- Documentation:
- - Default enabled preferred core.
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.
- - Default enabled preferred core.
- - Use a single variable to represent the status of preferred core.

Quoted 9 lines

Meng Li (7):
x86: Drop CPU_SUP_INTEL from SCHED_MC_PRIO for the expansion.
acpi: cppc: Add get the highest performance cppc control
cpufreq: amd-pstate: Enable amd-pstate preferred core supporting.
cpufreq: Add a notification message that the highest perf has changed
cpufreq: amd-pstate: Update amd-pstate preferred core ranking
dynamically
Documentation: amd-pstate: introduce amd-pstate preferred core
Documentation: introduce amd-pstate preferrd core mode kernel

command

Quoted 15 lines

line options

 .../admin-guide/kernel-parameters.txt         |   5 +
 Documentation/admin-guide/pm/amd-pstate.rst   |  59 +++++-
 arch/x86/Kconfig                              |   5 +-
 drivers/acpi/cppc_acpi.c                      |  13 ++
 drivers/acpi/processor_driver.c               |   6 +
 drivers/cpufreq/amd-pstate.c                  | 186 ++++++++++++++++--
 drivers/cpufreq/cpufreq.c                     |  13 ++
 include/acpi/cppc_acpi.h                      |   5 +
 include/linux/amd-pstate.h                    |  10 +
 include/linux/cpufreq.h                       |   5 +
 10 files changed, 285 insertions(+), 22 deletions(-)

--
Oleksandr Natalenko (post-factum)

-----Original Message-----
From: Oleksandr Natalenko <[email protected]>
Sent: Monday, October 9, 2023 9:00 PM
To: Rafael J . Wysocki <[email protected]>; Huang, Ray <[email protected]>; Meng, Li (Jassmine) <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Shuah Khan <[email protected]>; [email protected]; Fontenot, Nathan <[email protected]>; Sharma, Deepak <[email protected]>; Deucher, Alexander <[email protected]>; Limonciello, Mario <[email protected]>; Huang, Shimmer <[email protected]>; Yuan, Perry <[email protected]>; Du, Xiaojian <[email protected]>; Viresh Kumar <[email protected]>; Borislav Petkov <[email protected]>
Subject: Re: [PATCH V8 0/7] amd-pstate preferred core

Hello.

Quoted 16 lines · Li (Jassmine)

[AMD Official Use Only - General]

Hi Oleksandr:

Quoted 12 lines

-----Original Message-----
From: Oleksandr Natalenko <[email protected]>
Sent: Monday, October 9, 2023 2:55 PM
To: Rafael J . Wysocki <[email protected]>; Huang, Ray <[email protected]>; Meng, Li (Jassmine) <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Shuah Khan <[email protected]>; [email protected]; Fontenot, Nathan <[email protected]>; Sharma, Deepak <[email protected]>; Deucher, Alexander <[email protected]>; Limonciello, Mario <[email protected]>; Huang, Shimmer

<[email protected]>;

Quoted 6 lines

Quoted 6 lines

Yuan, Perry <[email protected]>; Du, Xiaojian <[email protected]>; Viresh Kumar <[email protected]>; Borislav Petkov <[email protected]>; Meng, Li (Jassmine) <[email protected]>
Subject: Re: [PATCH V8 0/7] amd-pstate preferred core

Quoted 6 lines

Quoted 6 lines

Hello.

Quoted 3 lines · On pondělí 9. října 2023 4:49:25 CEST Meng Li wrote:

Hi all:

The core frequency is subjected to the process variation in

semiconductors.

Quoted 7 lines

Quoted 7 lines

Quoted 7 lines

Not all cores are able to reach the maximum frequency respecting the infrastructure limits. Consequently, AMD has redefined the concept of maximum frequency of a part. This means that a fraction of cores can reach maximum frequency. To find the best process scheduling policy for a given scenario, OS needs to know the core ordering informed by the platform through highest performance capability register of the CPPC

Quoted 6 lines

Quoted 6 lines

interface.

Quoted 5 lines

Earlier implementations of amd-pstate preferred core only support a static core ranking and targeted performance. Now it has the ability to dynamically change the preferred core based on the workload and platform conditions and accounting for thermals and

aging.

Quoted 4 lines

Quoted 4 lines

Quoted 4 lines

Amd-pstate driver utilizes the functions and data structures provided by the ITMT architecture to enable the scheduler to favor scheduling on cores which can be get a higher frequency with lower

voltage.

Quoted 11 lines

Quoted 11 lines

Quoted 11 lines

We call it amd-pstate preferred core.

Here sched_set_itmt_core_prio() is called to set priorities and sched_set_itmt_support() is called to enable ITMT feature. Amd-pstate driver uses the highest performance value to indicate the priority of CPU. The higher value has a higher priority.

Amd-pstate driver will provide an initial core ordering at boot time. It relies on the CPPC interface to communicate the core ranking to the operating system and scheduler to make sure that OS is choosing the cores with highest performance firstly for scheduling the

process.

Quoted 10 lines

Quoted 10 lines

Quoted 9 lines

When amd-pstate driver receives a message with the highest performance change, it will update the core ranking.

Changes form V7->V8:
- all:
- - pick up Review-By flag added by Mario and Ray.
- cpufreq: amd-pstate:
- - use hw_prefcore embeds into cpudata structure.
- - delete preferred core init from cpu online/off.

Quoted 7 lines

Quoted 3 lines

Could you please let me know if this change means a fix for the report I've sent previously? [1]

[Meng, Li (Jassmine)] Yes.
I have deleted online handle function of amd pstate driver. It doesn't re-initialize preferred core.
This online function will set incorrect des perf value.

Thank you for the confirmation. I've built v6.5.5 with this patchset applied, and now the frequency is as expected after the suspend-resume cycle.

I've also added the following modification to accommodate recent feedback:

```
commit 1450ac395434c532f995521e1a2497d09ddf106c

Author: Oleksandr Natalenko <[email protected]>
Date:   Mon Oct 9 11:19:50 2023 +0200

cpufreq/amd-pstate: show prefcore_ranking separately

Signed-off-by Oleksandr Natalenko <[email protected]>

drivers/cpufreq/amd-pstate.c +11 -0

index d3369247c6c9c..86999d861e87b 100644
@@ -954,6 +954,17 @@ static ssize_t show_amd_pstate_highest_perf(struct
 cpufreq_policy *policy,
      u32 perf;
      struct amd_cpudata *cpudata = policy->driver_data;
 
+     perf = READ_ONCE(cpudata->highest_perf);
+
+     return sysfs_emit(buf, "%u\n", perf);
+}
+
+static ssize_t show_amd_pstate_prefcore_ranking(struct cpufreq_policy
 *policy,
+                                             char *buf)
+{
+     u32 perf;
+     struct amd_cpudata *cpudata = policy->driver_data;
+
      perf = READ_ONCE(cpudata->prefcore_ranking);

      return sysfs_emit(buf, "%u\n", perf);

drivers/cpufreq/amd-pstate.c +3 -0

@@ -1172,6 +1183,7 @@ cpufreq_freq_attr_ro(amd_pstate_max_freq);
 cpufreq_freq_attr_ro(amd_pstate_lowest_nonlinear_freq);
 
 cpufreq_freq_attr_ro(amd_pstate_highest_perf);
+cpufreq_freq_attr_ro(amd_pstate_prefcore_ranking);
 cpufreq_freq_attr_ro(amd_pstate_hw_prefcore);
 cpufreq_freq_attr_rw(energy_performance_preference);
 cpufreq_freq_attr_ro(energy_performance_available_preferences);

@@ -1182,6 +1194,7 @@ static struct freq_attr *amd_pstate_attr[] = {
      &amd_pstate_max_freq,
      &amd_pstate_lowest_nonlinear_freq,
      &amd_pstate_highest_perf,
+     &amd_pstate_prefcore_ranking,
      &amd_pstate_hw_prefcore,
      NULL,
 };

@@ -1190,6 +1203,7 @@ static struct freq_attr *amd_pstate_epp_attr[] = {
      &amd_pstate_max_freq,
      &amd_pstate_lowest_nonlinear_freq,
      &amd_pstate_highest_perf,
+     &amd_pstate_prefcore_ranking,
      &amd_pstate_hw_prefcore,
      &energy_performance_preference,
      &energy_performance_available_preferences,

```

with the following output as a result:

```

[~]> grep . /sys/devices/system/cpu*/cpufreq/policy*/amd_pstate_highest_perf /sys/devices/system/cpu/cpufreq/policy0/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy1/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy2/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy3/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy4/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy5/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy6/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy7/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy8/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy9/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy10/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy11/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy12/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy13/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy14/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy15/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy16/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy17/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy18/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy19/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy20/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy21/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy22/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy23/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy24/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy25/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy26/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy27/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy28/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy29/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy30/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy31/amd_pstate_highest_perf:166

[~]> grep . /sys/devices/system/cpu*/cpufreq/policy*/amd_pstate_hw_prefcore /sys/devices/system/cpu/cpufreq/policy0/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy1/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy2/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy3/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy4/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy5/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy6/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy7/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy8/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy9/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy10/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy11/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy12/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy13/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy14/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy15/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy16/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy17/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy18/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy19/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy20/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy21/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy22/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy23/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy24/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy25/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy26/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy27/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy28/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy29/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy30/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy31/amd_pstate_hw_prefcore:supp orted

[~]> grep . /sys/devices/system/cpu*/cpufreq/policy*/amd_pstate_prefcore_ranking /sys/devices/system/cpu/cpufreq/policy0/amd_pstate_prefcore_ranking:22 6 /sys/devices/system/cpu/cpufreq/policy1/amd_pstate_prefcore_ranking:23 1 /sys/devices/system/cpu/cpufreq/policy2/amd_pstate_prefcore_ranking:21 1 /sys/devices/system/cpu/cpufreq/policy3/amd_pstate_prefcore_ranking:23 6 /sys/devices/system/cpu/cpufreq/policy4/amd_pstate_prefcore_ranking:21 6 /sys/devices/system/cpu/cpufreq/policy5/amd_pstate_prefcore_ranking:23 6 /sys/devices/system/cpu/cpufreq/policy6/amd_pstate_prefcore_ranking:20 6 /sys/devices/system/cpu/cpufreq/policy7/amd_pstate_prefcore_ranking:22 1 /sys/devices/system/cpu/cpufreq/policy8/amd_pstate_prefcore_ranking:19 1 /sys/devices/system/cpu/cpufreq/policy9/amd_pstate_prefcore_ranking:20 1 /sys/devices/system/cpu/cpufreq/policy10/amd_pstate_prefcore_ranking:1 86 /sys/devices/system/cpu/cpufreq/policy11/amd_pstate_prefcore_ranking:1 96 /sys/devices/system/cpu/cpufreq/policy12/amd_pstate_prefcore_ranking:1 71 /sys/devices/system/cpu/cpufreq/policy13/amd_pstate_prefcore_ranking:1 66 /sys/devices/system/cpu/cpufreq/policy14/amd_pstate_prefcore_ranking:1 76 /sys/devices/system/cpu/cpufreq/policy15/amd_pstate_prefcore_ranking:1 81 /sys/devices/system/cpu/cpufreq/policy16/amd_pstate_prefcore_ranking:2 26 /sys/devices/system/cpu/cpufreq/policy17/amd_pstate_prefcore_ranking:2 31 /sys/devices/system/cpu/cpufreq/policy18/amd_pstate_prefcore_ranking:2 11 /sys/devices/system/cpu/cpufreq/policy19/amd_pstate_prefcore_ranking:2 36 /sys/devices/system/cpu/cpufreq/policy20/amd_pstate_prefcore_ranking:2 16 /sys/devices/system/cpu/cpufreq/policy21/amd_pstate_prefcore_ranking:2 36 /sys/devices/system/cpu/cpufreq/policy22/amd_pstate_prefcore_ranking:2 06 /sys/devices/system/cpu/cpufreq/policy23/amd_pstate_prefcore_ranking:2 21 /sys/devices/system/cpu/cpufreq/policy24/amd_pstate_prefcore_ranking:1 91 /sys/devices/system/cpu/cpufreq/policy25/amd_pstate_prefcore_ranking:2 01 /sys/devices/system/cpu/cpufreq/policy26/amd_pstate_prefcore_ranking:1 86 /sys/devices/system/cpu/cpufreq/policy27/amd_pstate_prefcore_ranking:1 96 /sys/devices/system/cpu/cpufreq/policy28/amd_pstate_prefcore_ranking:1 71 /sys/devices/system/cpu/cpufreq/policy29/amd_pstate_prefcore_ranking:1 66 /sys/devices/system/cpu/cpufreq/policy30/amd_pstate_prefcore_ranking:1 76 /sys/devices/system/cpu/cpufreq/policy31/amd_pstate_prefcore_ranking:1 81
```

When I run dd if=/dev/zero of=/dev/null, the load lands onto cores 3, 5, 19 or 21, IOW, those that have the highest amd_pstate_prefcore_ranking value given schedutil is in use.

If all of the above is as expected, please add:

Tested-by Oleksandr Natalenko <[email protected]>

Quoted 2 lines

Quoted 1 line

Would you also be able to Cc me on the next iteration of this patchset?

[Meng, Li (Jassmine)] OK.

Thanks.

Quoted 95 lines

Quoted 95 lines

Thank you!

[1] https://lore.kernel.org/lkml/[email protected]/

Quoted 63 lines

Changes form V6->V7:
- x86:
- - Modify kconfig about X86_AMD_PSTATE.
- cpufreq: amd-pstate:
- - modify incorrect comments about scheduler_work().
- - convert highest_perf data type.
- - modify preferred core init when cpu init and online.
- acpi: cppc:
- - modify link of CPPC highest performance.
- cpufreq:
- - modify link of CPPC highest performance changed.

Changes form V5->V6:
- cpufreq: amd-pstate:
- - modify the wrong tag order.
- - modify warning about hw_prefcore sysfs attribute.
- - delete duplicate comments.
- - modify the variable name cppc_highest_perf to prefcore_ranking.
- - modify judgment conditions for setting highest_perf.
- - modify sysfs attribute for CPPC highest perf to pr_debug message.
- Documentation: amd-pstate:
- - modify warning: title underline too short.

Changes form V4->V5:
- cpufreq: amd-pstate:
- - modify sysfs attribute for CPPC highest perf.
- - modify warning about comments
- - rebase linux-next
- cpufreq:
- - Moidfy warning about function declarations.
- Documentation: amd-pstate:
- - align with ``amd-pstat``

Changes form V3->V4:
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.

Changes form V2->V3:
- x86:
- - Modify kconfig and description.
- cpufreq: amd-pstate:
- - Add Co-developed-by tag in commit message.
- cpufreq:
- - Modify commit message.
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.

Changes form V1->V2:
- acpi: cppc:
- - Add reference link.
- cpufreq:
- - Moidfy link error.
- cpufreq: amd-pstate:
- - Init the priorities of all online CPUs
- - Use a single variable to represent the status of preferred core.
- Documentation:
- - Default enabled preferred core.
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.
- - Default enabled preferred core.
- - Use a single variable to represent the status of preferred core.

Quoted 9 lines

Meng Li (7):
x86: Drop CPU_SUP_INTEL from SCHED_MC_PRIO for the expansion.
acpi: cppc: Add get the highest performance cppc control
cpufreq: amd-pstate: Enable amd-pstate preferred core supporting.
cpufreq: Add a notification message that the highest perf has changed
cpufreq: amd-pstate: Update amd-pstate preferred core ranking
dynamically
Documentation: amd-pstate: introduce amd-pstate preferred core
Documentation: introduce amd-pstate preferrd core mode kernel

command

Quoted 15 lines

line options

 .../admin-guide/kernel-parameters.txt         |   5 +
 Documentation/admin-guide/pm/amd-pstate.rst   |  59 +++++-
 arch/x86/Kconfig                              |   5 +-
 drivers/acpi/cppc_acpi.c                      |  13 ++
 drivers/acpi/processor_driver.c               |   6 +
 drivers/cpufreq/amd-pstate.c                  | 186 ++++++++++++++++--
 drivers/cpufreq/cpufreq.c                     |  13 ++
 include/acpi/cppc_acpi.h                      |   5 +
 include/linux/amd-pstate.h                    |  10 +
 include/linux/cpufreq.h                       |   5 +
 10 files changed, 285 insertions(+), 22 deletions(-)

Quoted 3 lines

Quoted 2 lines

--
Oleksandr Natalenko (post-factum)

--
Oleksandr Natalenko (post-factum)

Hi all:

The core frequency is subjected to the process variation in semiconductors. Not all cores are able to reach the maximum frequency respecting the infrastructure limits. Consequently, AMD has redefined the concept of maximum frequency of a part. This means that a fraction of cores can reach maximum frequency. To find the best process scheduling policy for a given scenario, OS needs to know the core ordering informed by the platform through highest performance capability register of the CPPC interface.

Earlier implementations of amd-pstate preferred core only support a static core ranking and targeted performance. Now it has the ability to dynamically change the preferred core based on the workload and platform conditions and accounting for thermals and aging.

Amd-pstate driver utilizes the functions and data structures provided by the ITMT architecture to enable the scheduler to favor scheduling on cores which can be get a higher frequency with lower voltage.
We call it amd-pstate preferred core.

Here sched_set_itmt_core_prio() is called to set priorities and sched_set_itmt_support() is called to enable ITMT feature. Amd-pstate driver uses the highest performance value to indicate the priority of CPU. The higher value has a higher priority.

Amd-pstate driver will provide an initial core ordering at boot time. It relies on the CPPC interface to communicate the core ranking to the operating system and scheduler to make sure that OS is choosing the cores with highest performance firstly for scheduling the process. When amd-pstate driver receives a message with the highest performance change, it will update the core ranking.

Changes form V7->V8:
- all:
- - pick up Review-By flag added by Mario and Ray.
- cpufreq: amd-pstate:
- - use hw_prefcore embeds into cpudata structure.
- - delete preferred core init from cpu online/off.

Changes form V6->V7:
- x86:
- - Modify kconfig about X86_AMD_PSTATE.
- cpufreq: amd-pstate:
- - modify incorrect comments about scheduler_work().
- - convert highest_perf data type.
- - modify preferred core init when cpu init and online.
- acpi: cppc:
- - modify link of CPPC highest performance.
- cpufreq:
- - modify link of CPPC highest performance changed.

Changes form V5->V6:
- cpufreq: amd-pstate:
- - modify the wrong tag order.
- - modify warning about hw_prefcore sysfs attribute.
- - delete duplicate comments.
- - modify the variable name cppc_highest_perf to prefcore_ranking.
- - modify judgment conditions for setting highest_perf.
- - modify sysfs attribute for CPPC highest perf to pr_debug message.
- Documentation: amd-pstate:
- - modify warning: title underline too short.

Changes form V4->V5:
- cpufreq: amd-pstate:
- - modify sysfs attribute for CPPC highest perf.
- - modify warning about comments
- - rebase linux-next
- cpufreq:
- - Moidfy warning about function declarations.
- Documentation: amd-pstate:
- - align with ``amd-pstat``

Changes form V3->V4:
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.

Changes form V2->V3:
- x86:
- - Modify kconfig and description.
- cpufreq: amd-pstate:
- - Add Co-developed-by tag in commit message.
- cpufreq:
- - Modify commit message.
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.

Changes form V1->V2:
- acpi: cppc:
- - Add reference link.
- cpufreq:
- - Moidfy link error.
- cpufreq: amd-pstate:
- - Init the priorities of all online CPUs
- - Use a single variable to represent the status of preferred core.
- Documentation:
- - Default enabled preferred core.
- Documentation: amd-pstate:
- - Modify inappropriate descriptions.
- - Default enabled preferred core.
- - Use a single variable to represent the status of preferred core.

Meng Li (7):
x86: Drop CPU_SUP_INTEL from SCHED_MC_PRIO for the expansion.
acpi: cppc: Add get the highest performance cppc control
cpufreq: amd-pstate: Enable amd-pstate preferred core supporting.
cpufreq: Add a notification message that the highest perf has changed
cpufreq: amd-pstate: Update amd-pstate preferred core ranking
dynamically
Documentation: amd-pstate: introduce amd-pstate preferred core
Documentation: introduce amd-pstate preferrd core mode kernel command
line options

 .../admin-guide/kernel-parameters.txt         |   5 +
 Documentation/admin-guide/pm/amd-pstate.rst   |  59 +++++-
 arch/x86/Kconfig                              |   5 +-
 drivers/acpi/cppc_acpi.c                      |  13 ++
 drivers/acpi/processor_driver.c               |   6 +
 drivers/cpufreq/amd-pstate.c                  | 186 ++++++++++++++++--
 drivers/cpufreq/cpufreq.c                     |  13 ++
 include/acpi/cppc_acpi.h                      |   5 +
 include/linux/amd-pstate.h                    |  10 +
 include/linux/cpufreq.h                       |   5 +
 10 files changed, 285 insertions(+), 22 deletions(-)

Signature · 1 line

2.34.1