Skip to contentSkip to conversation

cpufreq: Expose cpufreq coordination requirements regardless of coordination mechanism

PATCH In review Darrick J. Wong 2 participants 3 messages Mar 21, 2008

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 cpufreq: Expose cpufreq coordination requirements regardless of coordination mechanism

Currently, affected_cpus shows which CPUs need to have their frequency coordinated in software. When hardware coordination is in use, the contents of this file appear the same as when no coordination is required. This can lead to some confusion among user-space programs, for example, that do not know that extra coordination is required to force a CPU core to a particular speed to control power consumption.

To fix this, create a "related_cpus" attribute that always displays the coordination map regardless of whatever coordination strategy the cpufreq driver uses (sw or hw). If the cpufreq driver does not provide a value, fall back to policy->cpus.

---

 arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c |    1 +
 drivers/cpufreq/cpufreq.c                  |   29 +++++++++++++++++++++++-----
 include/linux/cpufreq.h                    |    3 ++-
 3 files changed, 27 insertions(+), 6 deletions(-)
arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c +1 -0
index a962dcb..2e21d35 100644
@@ -591,6 +591,7 @@ static int acpi_cpufreq_cpu_init(struct cpufreq_policy *policy)
 	    policy->shared_type == CPUFREQ_SHARED_TYPE_ANY) {
 		policy->cpus = perf->shared_cpu_map;
 	}
+	policy->related_cpus = perf->shared_cpu_map;
 
 #ifdef CONFIG_SMP
 	dmi_check_system(sw_any_bug_dmi_table);
drivers/cpufreq/cpufreq.c +24 -5
index 35a26a3..38bf03f 100644
@@ -582,15 +582,13 @@ out:
 	i += sprintf(&buf[i], "\n");
 	return i;
 }
-/**
- * show_affected_cpus - show the CPUs affected by each transition
- */
-static ssize_t show_affected_cpus (struct cpufreq_policy * policy, char *buf)
+
+static ssize_t show_cpus(cpumask_t mask, char *buf)
 {
 	ssize_t i = 0;
 	unsigned int cpu;
 
-	for_each_cpu_mask(cpu, policy->cpus) {
+	for_each_cpu_mask(cpu, mask) {
 		if (i)
 			i += scnprintf(&buf[i], (PAGE_SIZE - i - 2), " ");
 		i += scnprintf(&buf[i], (PAGE_SIZE - i - 2), "%u", cpu);
@@ -601,6 +599,25 @@ static ssize_t show_affected_cpus (struct cpufreq_policy * policy, char *buf)
 	return i;
 }
 
+/**
+ * show_related_cpus - show the CPUs affected by each transition even if
+ * hw coordination is in use
+ */
+static ssize_t show_related_cpus(struct cpufreq_policy *policy, char *buf)
+{
+	if (cpus_empty(policy->related_cpus))
+		return show_cpus(policy->cpus, buf);
+	return show_cpus(policy->related_cpus, buf);
+}
+
+/**
+ * show_affected_cpus - show the CPUs affected by each transition
+ */
+static ssize_t show_affected_cpus(struct cpufreq_policy *policy, char *buf)
+{
+	return show_cpus(policy->cpus, buf);
+}
+
 static ssize_t store_scaling_setspeed(struct cpufreq_policy *policy,
 		const char *buf, size_t count)
 {
@@ -645,6 +662,7 @@ define_one_ro(cpuinfo_max_freq);
 define_one_ro(scaling_available_governors);
 define_one_ro(scaling_driver);
 define_one_ro(scaling_cur_freq);
+define_one_ro(related_cpus);
 define_one_ro(affected_cpus);
 define_one_rw(scaling_min_freq);
 define_one_rw(scaling_max_freq);
@@ -657,6 +675,7 @@ static struct attribute * default_attrs[] = {
 	&scaling_min_freq.attr,
 	&scaling_max_freq.attr,
 	&affected_cpus.attr,
+	&related_cpus.attr,
 	&scaling_governor.attr,
 	&scaling_driver.attr,
 	&scaling_available_governors.attr,
include/linux/cpufreq.h +2 -1
index ddd8652..a881fd6 100644
@@ -83,7 +83,8 @@ struct cpufreq_real_policy {
 };
 
 struct cpufreq_policy {
-	cpumask_t		cpus;	/* affected CPUs */
+	cpumask_t		cpus;	/* CPUs requiring sw coordination */
+	cpumask_t		related_cpus; /* CPUs with any coordination */
 	unsigned int		shared_type; /* ANY or ALL affected CPUs
 						should set cpufreq */
 	unsigned int		cpu;    /* cpu nr of registered CPU */
#1 lore raw
Quoted 11 lines · On Thu, 20 Mar 2008 17:43:54 -0700 "Darrick J. Wong" <[email protected]> wrote:

Currently, affected_cpus shows which CPUs need to have their frequency coordinated in software. When hardware coordination is in use, the contents of this file appear the same as when no coordination is required. This can lead to some confusion among user-space programs, for example, that do not know that extra coordination is required to force a CPU core to a particular speed to control power consumption.

To fix this, create a "related_cpus" attribute that always displays the coordination map regardless of whatever coordination strategy the cpufreq driver uses (sw or hw). If the cpufreq driver does not provide a value, fall back to policy->cpus.

I was going to ding you for not updating the kernel->userspace API documentation. But it seems that none of this interface is documented anyway :(

#2 lore raw

PATCH 1/1 cpufreq: Document the currently undocumented parts of the sysfs interface

Quoted 3 lines · Andrew Morton

I was going to ding you for not updating the kernel->userspace API documentation. But it seems that none of this interface is documented anyway :(

Actually, I deserve the ding, for there _is_ a description of some of the sysfs files. However, there are some that are not mentioned in the documentation, so add them to the user's guide.

---

 Documentation/cpu-freq/user-guide.txt |   14 ++++++++++++++
 1 files changed, 14 insertions(+), 0 deletions(-)
Documentation/cpu-freq/user-guide.txt +14 -0
index af3b925..6c442d8 100644
@@ -154,6 +154,11 @@ scaling_governor,		and by "echoing" the name of another
 				that some governors won't load - they only
 				work on some specific architectures or
 				processors.
+
+cpuinfo_cur_freq :		Current speed of the CPU, in KHz.
+
+scaling_available_frequencies : List of available frequencies, in KHz.
+
 scaling_min_freq and
 scaling_max_freq		show the current "policy limits" (in
 				kHz). By echoing new values into these
@@ -162,6 +167,15 @@ scaling_max_freq		show the current "policy limits" (in
 				first set scaling_max_freq, then
 				scaling_min_freq.
 
+affected_cpus :			List of CPUs that require software coordination
+				of frequency.
+
+related_cpus :			List of CPUs that need some sort of frequency
+				coordination, whether software or hardware.
+
+scaling_driver :		Hardware driver for cpufreq.
+
+scaling_cur_freq :		Current frequency of the CPU, in KHz.
 
 If you have selected the "userspace" governor which allows you to
 set the CPU operating frequency to a specific value, you can read out