ARM: 8621/3: parse cpu capacity-dmips-mhz from DT
authorJuri Lelli <juri.lelli@arm.com>
Sun, 6 Nov 2016 00:34:15 +0000 (01:34 +0100)
committerRussell King <rmk+kernel@armlinux.org.uk>
Tue, 22 Nov 2016 11:45:06 +0000 (11:45 +0000)
With the introduction of cpu capacity-dmips-mhz bindings, CPU capacities
can now be calculated from values extracted from DT and information
coming from cpufreq. Add parsing of DT information at boot time, and
complement it with cpufreq information. We keep code that can produce
same information, based on different DT properties and hard-coded
values, as fall-back for backward compatibility.

Caveat: the information provided by this patch will start to be used in
the future. We need to #define arch_scale_cpu_capacity to something
provided in arch, so that scheduler's default implementation (which gets
used if arch_scale_cpu_capacity is not defined) is overwritten.

Signed-off-by: Juri Lelli <juri.lelli@arm.com>
Acked-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
arch/arm/kernel/topology.c

index ec279d161b3287e1df5b51702c499a2ac0187054..aa63287d9a1054e51fa9b9d7294a582758f24a10 100644 (file)
@@ -12,6 +12,7 @@
  */
 
 #include <linux/cpu.h>
+#include <linux/cpufreq.h>
 #include <linux/cpumask.h>
 #include <linux/export.h>
 #include <linux/init.h>
@@ -78,6 +79,144 @@ static unsigned long *__cpu_capacity;
 #define cpu_capacity(cpu)      __cpu_capacity[cpu]
 
 static unsigned long middle_capacity = 1;
+static bool cap_from_dt = true;
+static u32 *raw_capacity;
+static bool cap_parsing_failed;
+static u32 capacity_scale;
+
+static int __init parse_cpu_capacity(struct device_node *cpu_node, int cpu)
+{
+       int ret = 1;
+       u32 cpu_capacity;
+
+       if (cap_parsing_failed)
+               return !ret;
+
+       ret = of_property_read_u32(cpu_node,
+                                  "capacity-dmips-mhz",
+                                  &cpu_capacity);
+       if (!ret) {
+               if (!raw_capacity) {
+                       raw_capacity = kcalloc(num_possible_cpus(),
+                                              sizeof(*raw_capacity),
+                                              GFP_KERNEL);
+                       if (!raw_capacity) {
+                               pr_err("cpu_capacity: failed to allocate memory for raw capacities\n");
+                               cap_parsing_failed = true;
+                               return !ret;
+                       }
+               }
+               capacity_scale = max(cpu_capacity, capacity_scale);
+               raw_capacity[cpu] = cpu_capacity;
+               pr_debug("cpu_capacity: %s cpu_capacity=%u (raw)\n",
+                       cpu_node->full_name, raw_capacity[cpu]);
+       } else {
+               if (raw_capacity) {
+                       pr_err("cpu_capacity: missing %s raw capacity\n",
+                               cpu_node->full_name);
+                       pr_err("cpu_capacity: partial information: fallback to 1024 for all CPUs\n");
+               }
+               cap_parsing_failed = true;
+               kfree(raw_capacity);
+       }
+
+       return !ret;
+}
+
+static void normalize_cpu_capacity(void)
+{
+       u64 capacity;
+       int cpu;
+
+       if (!raw_capacity || cap_parsing_failed)
+               return;
+
+       pr_debug("cpu_capacity: capacity_scale=%u\n", capacity_scale);
+       for_each_possible_cpu(cpu) {
+               capacity = (raw_capacity[cpu] << SCHED_CAPACITY_SHIFT)
+                       / capacity_scale;
+               set_capacity_scale(cpu, capacity);
+               pr_debug("cpu_capacity: CPU%d cpu_capacity=%lu\n",
+                       cpu, arch_scale_cpu_capacity(NULL, cpu));
+       }
+}
+
+#ifdef CONFIG_CPU_FREQ
+static cpumask_var_t cpus_to_visit;
+static bool cap_parsing_done;
+static void parsing_done_workfn(struct work_struct *work);
+static DECLARE_WORK(parsing_done_work, parsing_done_workfn);
+
+static int
+init_cpu_capacity_callback(struct notifier_block *nb,
+                          unsigned long val,
+                          void *data)
+{
+       struct cpufreq_policy *policy = data;
+       int cpu;
+
+       if (cap_parsing_failed || cap_parsing_done)
+               return 0;
+
+       switch (val) {
+       case CPUFREQ_NOTIFY:
+               pr_debug("cpu_capacity: init cpu capacity for CPUs [%*pbl] (to_visit=%*pbl)\n",
+                               cpumask_pr_args(policy->related_cpus),
+                               cpumask_pr_args(cpus_to_visit));
+               cpumask_andnot(cpus_to_visit,
+                              cpus_to_visit,
+                              policy->related_cpus);
+               for_each_cpu(cpu, policy->related_cpus) {
+                       raw_capacity[cpu] = arch_scale_cpu_capacity(NULL, cpu) *
+                                           policy->cpuinfo.max_freq / 1000UL;
+                       capacity_scale = max(raw_capacity[cpu], capacity_scale);
+               }
+               if (cpumask_empty(cpus_to_visit)) {
+                       normalize_cpu_capacity();
+                       kfree(raw_capacity);
+                       pr_debug("cpu_capacity: parsing done\n");
+                       cap_parsing_done = true;
+                       schedule_work(&parsing_done_work);
+               }
+       }
+       return 0;
+}
+
+static struct notifier_block init_cpu_capacity_notifier = {
+       .notifier_call = init_cpu_capacity_callback,
+};
+
+static int __init register_cpufreq_notifier(void)
+{
+       if (cap_parsing_failed)
+               return -EINVAL;
+
+       if (!alloc_cpumask_var(&cpus_to_visit, GFP_KERNEL)) {
+               pr_err("cpu_capacity: failed to allocate memory for cpus_to_visit\n");
+               return -ENOMEM;
+       }
+       cpumask_copy(cpus_to_visit, cpu_possible_mask);
+
+       return cpufreq_register_notifier(&init_cpu_capacity_notifier,
+                                        CPUFREQ_POLICY_NOTIFIER);
+}
+core_initcall(register_cpufreq_notifier);
+
+static void parsing_done_workfn(struct work_struct *work)
+{
+       cpufreq_unregister_notifier(&init_cpu_capacity_notifier,
+                                        CPUFREQ_POLICY_NOTIFIER);
+}
+
+#else
+static int __init free_raw_capacity(void)
+{
+       kfree(raw_capacity);
+
+       return 0;
+}
+core_initcall(free_raw_capacity);
+#endif
 
 /*
  * Iterate all CPUs' descriptor in DT and compute the efficiency
@@ -99,6 +238,12 @@ static void __init parse_dt_topology(void)
        __cpu_capacity = kcalloc(nr_cpu_ids, sizeof(*__cpu_capacity),
                                 GFP_NOWAIT);
 
+       cn = of_find_node_by_path("/cpus");
+       if (!cn) {
+               pr_err("No CPU information found in DT\n");
+               return;
+       }
+
        for_each_possible_cpu(cpu) {
                const u32 *rate;
                int len;
@@ -110,6 +255,13 @@ static void __init parse_dt_topology(void)
                        continue;
                }
 
+               if (parse_cpu_capacity(cn, cpu)) {
+                       of_node_put(cn);
+                       continue;
+               }
+
+               cap_from_dt = false;
+
                for (cpu_eff = table_efficiency; cpu_eff->compatible; cpu_eff++)
                        if (of_device_is_compatible(cn, cpu_eff->compatible))
                                break;
@@ -151,6 +303,8 @@ static void __init parse_dt_topology(void)
                middle_capacity = ((max_capacity / 3)
                                >> (SCHED_CAPACITY_SHIFT-1)) + 1;
 
+       if (cap_from_dt && !cap_parsing_failed)
+               normalize_cpu_capacity();
 }
 
 /*
@@ -160,7 +314,7 @@ static void __init parse_dt_topology(void)
  */
 static void update_cpu_capacity(unsigned int cpu)
 {
-       if (!cpu_capacity(cpu))
+       if (!cpu_capacity(cpu) || cap_from_dt)
                return;
 
        set_capacity_scale(cpu, cpu_capacity(cpu) / middle_capacity);