x86: Add generic aperf/mperf code

Move some of the aperf/mperf code out from the cpufreq driver
thingy so that other people can enjoy it too.

Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Cc: Yanmin <yanmin_zhang@linux.intel.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Len Brown <len.brown@intel.com>
Cc: Yinghai Lu <yhlu.kernel@gmail.com>
Cc: cpufreq@vger.kernel.org
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

2 files changed, 39 insertions, 70 deletions

diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h
index e08ea043e085..4ae2ccfed638 100644
--- a/arch/x86/include/asm/processor.h
+++ b/arch/x86/include/asm/processor.h
@@ -27,6 +27,7 @@ struct mm_struct;
 #include <linux/cpumask.h>
 #include <linux/cache.h>
 #include <linux/threads.h>
+#include <linux/math64.h>
 #include <linux/init.h>
 
 /*
@@ -1020,4 +1021,33 @@ extern void start_thread(struct pt_regs *regs, unsigned long new_ip,
 extern int get_tsc_mode(unsigned long adr);
 extern int set_tsc_mode(unsigned int val);
 
+struct aperfmperf {
+        u64 aperf, mperf;
+};
+
+static inline void get_aperfmperf(struct aperfmperf *am)
+{
+        WARN_ON_ONCE(!boot_cpu_has(X86_FEATURE_APERFMPERF));
+
+        rdmsrl(MSR_IA32_APERF, am->aperf);
+        rdmsrl(MSR_IA32_MPERF, am->mperf);
+}
+
+#define APERFMPERF_SHIFT 10
+
+static inline
+unsigned long calc_aperfmperf_ratio(struct aperfmperf *old,
+                                    struct aperfmperf *new)
+{
+        u64 aperf = new->aperf - old->aperf;
+        u64 mperf = new->mperf - old->mperf;
+        unsigned long ratio = aperf;
+
+        mperf >>= APERFMPERF_SHIFT;
+        if (mperf)
+                ratio = div64_u64(aperf, mperf);
+
+        return ratio;
+}
+
 #endif /* _ASM_X86_PROCESSOR_H */
diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
index 509e6a7db719..4109679863c1 100644
--- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
+++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
@@ -70,11 +70,7 @@ struct acpi_cpufreq_data {
 
 static DEFINE_PER_CPU(struct acpi_cpufreq_data *, drv_data);
 
-struct acpi_msr_data {
-        u64 saved_aperf, saved_mperf;
-};
-
-static DEFINE_PER_CPU(struct acpi_msr_data, msr_data);
+static DEFINE_PER_CPU(struct aperfmperf, old_perf);
 
 DEFINE_TRACE(power_mark);
 
@@ -243,23 +239,12 @@ static u32 get_cur_val(const struct cpumask *mask)
         return cmd.val;
 }
 
-struct perf_pair {
-        union {
-                struct {
-                        u32 lo;
-                        u32 hi;
-                } split;
-                u64 whole;
-        } aperf, mperf;
-};
-
 /* Called via smp_call_function_single(), on the target CPU */
 static void read_measured_perf_ctrs(void *_cur)
 {
-        struct perf_pair *cur = _cur;
+        struct aperfmperf *am = _cur;
 
-        rdmsr(MSR_IA32_APERF, cur->aperf.split.lo, cur->aperf.split.hi);
-        rdmsr(MSR_IA32_MPERF, cur->mperf.split.lo, cur->mperf.split.hi);
+        get_aperfmperf(am);
 }
 
 /*
@@ -278,63 +263,17 @@ static void read_measured_perf_ctrs(void *_cur)
 static unsigned int get_measured_perf(struct cpufreq_policy *policy,
                                       unsigned int cpu)
 {
-        struct perf_pair readin, cur;
-        unsigned int perf_percent;
+        struct aperfmperf perf;
+        unsigned long ratio;
         unsigned int retval;
 
-        if (smp_call_function_single(cpu, read_measured_perf_ctrs, &readin, 1))
+        if (smp_call_function_single(cpu, read_measured_perf_ctrs, &perf, 1))
                 return 0;
 
-        cur.aperf.whole = readin.aperf.whole -
-                                per_cpu(msr_data, cpu).saved_aperf;
-        cur.mperf.whole = readin.mperf.whole -
-                                per_cpu(msr_data, cpu).saved_mperf;
-        per_cpu(msr_data, cpu).saved_aperf = readin.aperf.whole;
-        per_cpu(msr_data, cpu).saved_mperf = readin.mperf.whole;
-
-#ifdef __i386__
-        /*
-         * We dont want to do 64 bit divide with 32 bit kernel
-         * Get an approximate value. Return failure in case we cannot get
-         * an approximate value.
-         */
-        if (unlikely(cur.aperf.split.hi || cur.mperf.split.hi)) {
-                int shift_count;
-                u32 h;
-
-                h = max_t(u32, cur.aperf.split.hi, cur.mperf.split.hi);
-                shift_count = fls(h);
-
-                cur.aperf.whole >>= shift_count;
-                cur.mperf.whole >>= shift_count;
-        }
-
-        if (((unsigned long)(-1) / 100) < cur.aperf.split.lo) {
-                int shift_count = 7;
-                cur.aperf.split.lo >>= shift_count;
-                cur.mperf.split.lo >>= shift_count;
-        }
-
-        if (cur.aperf.split.lo && cur.mperf.split.lo)
-                perf_percent = (cur.aperf.split.lo * 100) / cur.mperf.split.lo;
-        else
-                perf_percent = 0;
-
-#else
-        if (unlikely(((unsigned long)(-1) / 100) < cur.aperf.whole)) {
-                int shift_count = 7;
-                cur.aperf.whole >>= shift_count;
-                cur.mperf.whole >>= shift_count;
-        }
-
-        if (cur.aperf.whole && cur.mperf.whole)
-                perf_percent = (cur.aperf.whole * 100) / cur.mperf.whole;
-        else
-                perf_percent = 0;
-
-#endif
+        ratio = calc_aperfmperf_ratio(&per_cpu(old_perf, cpu), &perf);
+        per_cpu(old_perf, cpu) = perf;
 
-        retval = (policy->cpuinfo.max_freq * perf_percent) / 100;
+        retval = (policy->cpuinfo.max_freq * ratio) >> APERFMPERF_SHIFT;
 
         return retval;
 }