[PATCH] [cpufreq] Add APERF/MPERF support for AMD processors
From: Mark Langsdorf
Date: Mon Mar 01 2010 - 13:58:28 EST
Starting with model 10 of Family 0x10, AMD processors may have
support for APERF/MPERF. Add support for identifying it and using
it within cpufreq. Move the APERF/MPERF functions out of the
acpi-cpufreq code and into their own file so they can easily be
shared.
Signed-off-by: Mark Langsdorf <mark.langsdorf@xxxxxxx>
---
arch/x86/kernel/cpu/amd.c | 6 +++
arch/x86/kernel/cpu/cpufreq/Makefile | 4 +-
arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c | 42 +-----------------------
arch/x86/kernel/cpu/cpufreq/mperf.c | 50
++++++++++++++++++++++++++++
arch/x86/kernel/cpu/cpufreq/mperf.h | 9 +++++
arch/x86/kernel/cpu/cpufreq/powernow-k8.c | 4 ++-
arch/x86/kernel/cpu/perf_event_amd.c | 2 +-
7 files changed, 72 insertions(+), 45 deletions(-)
create mode 100644 arch/x86/kernel/cpu/cpufreq/mperf.c
create mode 100644 arch/x86/kernel/cpu/cpufreq/mperf.h
diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index e485825..796f662 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -537,6 +537,12 @@ static void __cpuinit init_amd(struct cpuinfo_x86 *c)
set_cpu_cap(c, X86_FEATURE_MFENCE_RDTSC);
}
+ if (c->cpuid_level >= 6) {
+ unsigned ecx = cpuid_ecx(6);
+ if (ecx & 0x01)
+ set_cpu_cap(c, X86_FEATURE_APERFMPERF);
+ }
+
#ifdef CONFIG_X86_64
if (c->x86 == 0x10) {
/* do this for boot cpu */
diff --git a/arch/x86/kernel/cpu/cpufreq/Makefile
b/arch/x86/kernel/cpu/cpufreq/Makefile
index 509296d..39cc160 100644
--- a/arch/x86/kernel/cpu/cpufreq/Makefile
+++ b/arch/x86/kernel/cpu/cpufreq/Makefile
@@ -2,8 +2,8 @@
# K8 systems. ACPI is preferred to all other hardware-specific drivers.
# speedstep-* is preferred over p4-clockmod.
-obj-$(CONFIG_X86_POWERNOW_K8) += powernow-k8.o
-obj-$(CONFIG_X86_ACPI_CPUFREQ) += acpi-cpufreq.o
+obj-$(CONFIG_X86_POWERNOW_K8) += powernow-k8.o mperf.o
+obj-$(CONFIG_X86_ACPI_CPUFREQ) += acpi-cpufreq.o mperf.o
obj-$(CONFIG_X86_POWERNOW_K6) += powernow-k6.o
obj-$(CONFIG_X86_POWERNOW_K7) += powernow-k7.o
obj-$(CONFIG_X86_LONGHAUL) += longhaul.o
diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
index 1b1920f..b3bd6b3 100644
--- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
+++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
@@ -45,6 +45,7 @@
#include <asm/msr.h>
#include <asm/processor.h>
#include <asm/cpufeature.h>
+#include "mperf.h"
#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, \
"acpi-cpufreq", msg)
@@ -70,8 +71,6 @@ struct acpi_cpufreq_data {
static DEFINE_PER_CPU(struct acpi_cpufreq_data *, acfreq_data);
-static DEFINE_PER_CPU(struct aperfmperf, acfreq_old_perf);
-
/* acpi_perf_data is a pointer to percpu data. */
static struct acpi_processor_performance *acpi_perf_data;
@@ -239,45 +238,6 @@ static u32 get_cur_val(const struct cpumask *mask)
return cmd.val;
}
-/* Called via smp_call_function_single(), on the target CPU */
-static void read_measured_perf_ctrs(void *_cur)
-{
- struct aperfmperf *am = _cur;
-
- get_aperfmperf(am);
-}
-
-/*
- * Return the measured active (C0) frequency on this CPU since last call
- * to this function.
- * Input: cpu number
- * Return: Average CPU frequency in terms of max frequency (zero on error)
- *
- * We use IA32_MPERF and IA32_APERF MSRs to get the measured performance
- * over a period of time, while CPU is in C0 state.
- * IA32_MPERF counts at the rate of max advertised frequency
- * IA32_APERF counts at the rate of actual CPU frequency
- * Only IA32_APERF/IA32_MPERF ratio is architecturally defined and
- * no meaning should be associated with absolute values of these MSRs.
- */
-static unsigned int get_measured_perf(struct cpufreq_policy *policy,
- unsigned int cpu)
-{
- struct aperfmperf perf;
- unsigned long ratio;
- unsigned int retval;
-
- if (smp_call_function_single(cpu, read_measured_perf_ctrs, &perf, 1))
- return 0;
-
- ratio = calc_aperfmperf_ratio(&per_cpu(acfreq_old_perf, cpu), &perf);
- per_cpu(acfreq_old_perf, cpu) = perf;
-
- retval = (policy->cpuinfo.max_freq * ratio) >> APERFMPERF_SHIFT;
-
- return retval;
-}
-
static unsigned int get_cur_freq_on_cpu(unsigned int cpu)
{
struct acpi_cpufreq_data *data = per_cpu(acfreq_data, cpu);
diff --git a/arch/x86/kernel/cpu/cpufreq/mperf.c
b/arch/x86/kernel/cpu/cpufreq/mperf.c
new file mode 100644
index 0000000..7a0a753
--- /dev/null
+++ b/arch/x86/kernel/cpu/cpufreq/mperf.c
@@ -0,0 +1,50 @@
+#include <linux/kernel.h>
+#include <linux/smp.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/cpufreq.h>
+#include <linux/slab.h>
+
+#include "mperf.h"
+
+static DEFINE_PER_CPU(struct aperfmperf, acfreq_old_perf);
+
+/* Called via smp_call_function_single(), on the target CPU */
+static void read_measured_perf_ctrs(void *_cur)
+{
+ struct aperfmperf *am = _cur;
+
+ get_aperfmperf(am);
+}
+
+/*
+ * Return the measured active (C0) frequency on this CPU since last call
+ * to this function.
+ * Input: cpu number
+ * Return: Average CPU frequency in terms of max frequency (zero on error)
+ *
+ * We use IA32_MPERF and IA32_APERF MSRs to get the measured performance
+ * over a period of time, while CPU is in C0 state.
+ * IA32_MPERF counts at the rate of max advertised frequency
+ * IA32_APERF counts at the rate of actual CPU frequency
+ * Only IA32_APERF/IA32_MPERF ratio is architecturally defined and
+ * no meaning should be associated with absolute values of these MSRs.
+ */
+unsigned int get_measured_perf(struct cpufreq_policy *policy,
+ unsigned int cpu)
+{
+ struct aperfmperf perf;
+ unsigned long ratio;
+ unsigned int retval;
+
+ if (smp_call_function_single(cpu, read_measured_perf_ctrs, &perf, 1))
+ return 0;
+
+ ratio = calc_aperfmperf_ratio(&per_cpu(acfreq_old_perf, cpu), &perf);
+ per_cpu(acfreq_old_perf, cpu) = perf;
+
+ retval = (policy->cpuinfo.max_freq * ratio) >> APERFMPERF_SHIFT;
+
+ return retval;
+}
+EXPORT_SYMBOL_GPL(get_measured_perf);
diff --git a/arch/x86/kernel/cpu/cpufreq/mperf.h
b/arch/x86/kernel/cpu/cpufreq/mperf.h
new file mode 100644
index 0000000..67a7b64
--- /dev/null
+++ b/arch/x86/kernel/cpu/cpufreq/mperf.h
@@ -0,0 +1,9 @@
+/*
+ * (c) 2003-2006 Advanced Micro Devices, Inc.
+ * Your use of this code is subject to the terms and conditions of the
+ * GNU general public license version 2. See "COPYING" or
+ * http://www.gnu.org/licenses/gpl.html
+ */
+
+unsigned int get_measured_perf(struct cpufreq_policy *policy,
+ unsigned int cpu);
diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
index 6e44519..ed9c4c2 100644
--- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
+++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
@@ -46,6 +46,7 @@
#define PFX "powernow-k8: "
#define VERSION "version 2.20.00"
#include "powernow-k8.h"
+#include "mperf.h"
/* serialize freq changes */
static DEFINE_MUTEX(fidvid_mutex);
@@ -929,7 +930,8 @@ static int fill_powernow_table_pstate(struct
powernow_k8_data *data,
powernow_table[i].index = index;
/* Frequency may be rounded for these */
- if (boot_cpu_data.x86 == 0x10 || boot_cpu_data.x86 == 0x11) {
+ if ((boot_cpu_data.x86 == 0x10 && boot_cpu_data.x86_model < 10)
+ || boot_cpu_data.x86 == 0x11) {
powernow_table[i].frequency =
freq_from_fid_did(lo & 0x3f, (lo >> 6) & 7);
} else
diff --git a/arch/x86/kernel/cpu/perf_event_amd.c
b/arch/x86/kernel/cpu/perf_event_amd.c
index 6d28e08..8f3dbfd 100644
--- a/arch/x86/kernel/cpu/perf_event_amd.c
+++ b/arch/x86/kernel/cpu/perf_event_amd.c
@@ -1,6 +1,6 @@
#ifdef CONFIG_CPU_SUP_AMD
-static raw_spinlock_t amd_nb_lock;
+static DEFINE_RAW_SPINLOCK(amd_nb_lock);
static __initconst u64 amd_hw_cache_event_ids
[PERF_COUNT_HW_CACHE_MAX]
--
1.6.0.2
3
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/