* @cpu: CPU number for this instance data
* @update_util: CPUFreq utility callback information
* @update_util_set: CPUFreq utility callback is set
+ * @iowait_boost: iowait-related boost fraction
+ * @last_update: Time of the last update.
* @pstate: Stores P state limits for this CPU
* @vid: Stores VID limits for this CPU
* @pid: Stores PID parameters for this CPU
struct vid_data vid;
struct _pid pid;
+ u64 last_update;
u64 last_sample_time;
u64 prev_aperf;
u64 prev_mperf;
struct acpi_processor_performance acpi_perf_data;
bool valid_pss_table;
#endif
+ unsigned int iowait_boost;
};
static struct cpudata **all_cpu_data;
* @p_gain_pct: PID proportional gain
* @i_gain_pct: PID integral gain
* @d_gain_pct: PID derivative gain
+ * @boost_iowait: Whether or not to use iowait boosting.
*
* Stores per CPU model static PID configuration data.
*/
int p_gain_pct;
int d_gain_pct;
int i_gain_pct;
+ bool boost_iowait;
};
/**
int min, hw_min, max, hw_max, cpu, range, adj_range;
u64 value, cap;
- rdmsrl(MSR_HWP_CAPABILITIES, cap);
- hw_min = HWP_LOWEST_PERF(cap);
- hw_max = HWP_HIGHEST_PERF(cap);
- range = hw_max - hw_min;
-
for_each_cpu(cpu, cpumask) {
+ rdmsrl_on_cpu(cpu, MSR_HWP_CAPABILITIES, &cap);
+ hw_min = HWP_LOWEST_PERF(cap);
+ hw_max = HWP_HIGHEST_PERF(cap);
+ range = hw_max - hw_min;
+
rdmsrl_on_cpu(cpu, MSR_HWP_REQUEST, &value);
adj_range = limits->min_perf_pct * range / 100;
min = hw_min + adj_range;
},
};
-static struct cpu_defaults silvermont_params = {
+static const struct cpu_defaults silvermont_params = {
.pid_policy = {
.sample_rate_ms = 10,
.deadband = 0,
.p_gain_pct = 14,
.d_gain_pct = 0,
.i_gain_pct = 4,
+ .boost_iowait = true,
},
.funcs = {
.get_max = atom_get_max_pstate,
},
};
-static struct cpu_defaults airmont_params = {
+static const struct cpu_defaults airmont_params = {
.pid_policy = {
.sample_rate_ms = 10,
.deadband = 0,
.p_gain_pct = 14,
.d_gain_pct = 0,
.i_gain_pct = 4,
+ .boost_iowait = true,
},
.funcs = {
.get_max = atom_get_max_pstate,
},
};
-static struct cpu_defaults knl_params = {
+static const struct cpu_defaults knl_params = {
.pid_policy = {
.sample_rate_ms = 10,
.deadband = 0,
},
};
-static struct cpu_defaults bxt_params = {
+static const struct cpu_defaults bxt_params = {
.pid_policy = {
.sample_rate_ms = 10,
.deadband = 0,
.p_gain_pct = 14,
.d_gain_pct = 0,
.i_gain_pct = 4,
+ .boost_iowait = true,
},
.funcs = {
.get_max = core_get_max_pstate,
static inline int32_t get_target_pstate_use_cpu_load(struct cpudata *cpu)
{
struct sample *sample = &cpu->sample;
- u64 cummulative_iowait, delta_iowait_us;
- u64 delta_iowait_mperf;
- u64 mperf, now;
- int32_t cpu_load;
+ int32_t busy_frac, boost;
+ int target, avg_pstate;
- cummulative_iowait = get_cpu_iowait_time_us(cpu->cpu, &now);
+ busy_frac = div_fp(sample->mperf, sample->tsc);
- /*
- * Convert iowait time into number of IO cycles spent at max_freq.
- * IO is considered as busy only for the cpu_load algorithm. For
- * performance this is not needed since we always try to reach the
- * maximum P-State, so we are already boosting the IOs.
- */
- delta_iowait_us = cummulative_iowait - cpu->prev_cummulative_iowait;
- delta_iowait_mperf = div64_u64(delta_iowait_us * cpu->pstate.scaling *
- cpu->pstate.max_pstate, MSEC_PER_SEC);
+ boost = cpu->iowait_boost;
+ cpu->iowait_boost >>= 1;
+
+ if (busy_frac < boost)
+ busy_frac = boost;
- mperf = cpu->sample.mperf + delta_iowait_mperf;
- cpu->prev_cummulative_iowait = cummulative_iowait;
+ sample->busy_scaled = busy_frac * 100;
+
+ target = limits->no_turbo || limits->turbo_disabled ?
+ cpu->pstate.max_pstate : cpu->pstate.turbo_pstate;
+ target += target >> 2;
+ target = mul_fp(target, busy_frac);
+ if (target < cpu->pstate.min_pstate)
+ target = cpu->pstate.min_pstate;
/*
- * The load can be estimated as the ratio of the mperf counter
- * running at a constant frequency during active periods
- * (C0) and the time stamp counter running at the same frequency
- * also during C-states.
+ * If the average P-state during the previous cycle was higher than the
+ * current target, add 50% of the difference to the target to reduce
+ * possible performance oscillations and offset possible performance
+ * loss related to moving the workload from one CPU to another within
+ * a package/module.
*/
- cpu_load = div64_u64(int_tofp(100) * mperf, sample->tsc);
- cpu->sample.busy_scaled = cpu_load;
+ avg_pstate = get_avg_pstate(cpu);
+ if (avg_pstate > target)
+ target += (avg_pstate - target) >> 1;
- return get_avg_pstate(cpu) - pid_calc(&cpu->pid, cpu_load);
+ return target;
}
static inline int32_t get_target_pstate_use_performance(struct cpudata *cpu)
u64 duration_ns;
/*
- * perf_scaled is the average performance during the last sampling
- * period scaled by the ratio of the maximum P-state to the P-state
- * requested last time (in percent). That measures the system's
- * response to the previous P-state selection.
+ * perf_scaled is the ratio of the average P-state during the last
+ * sampling period to the P-state requested last time (in percent).
+ *
+ * That measures the system's response to the previous P-state
+ * selection.
*/
max_pstate = cpu->pstate.max_pstate_physical;
current_pstate = cpu->pstate.current_pstate;
sample->mperf,
sample->aperf,
sample->tsc,
- get_avg_frequency(cpu));
+ get_avg_frequency(cpu),
+ fp_toint(cpu->iowait_boost * 100));
}
static void intel_pstate_update_util(struct update_util_data *data, u64 time,
- unsigned long util, unsigned long max)
+ unsigned int flags)
{
struct cpudata *cpu = container_of(data, struct cpudata, update_util);
- u64 delta_ns = time - cpu->sample.time;
+ u64 delta_ns;
+
+ if (pid_params.boost_iowait) {
+ if (flags & SCHED_CPUFREQ_IOWAIT) {
+ cpu->iowait_boost = int_tofp(1);
+ } else if (cpu->iowait_boost) {
+ /* Clear iowait_boost if the CPU may have been idle. */
+ delta_ns = time - cpu->last_update;
+ if (delta_ns > TICK_NSEC)
+ cpu->iowait_boost = 0;
+ }
+ cpu->last_update = time;
+ }
+ delta_ns = time - cpu->sample.time;
if ((s64)delta_ns >= pid_params.sample_rate_ns) {
bool sample_taken = intel_pstate_sample(cpu, time);
static const struct x86_cpu_id intel_pstate_cpu_oob_ids[] __initconst = {
ICPU(INTEL_FAM6_BROADWELL_XEON_D, core_params),
+ ICPU(INTEL_FAM6_BROADWELL_X, core_params),
+ ICPU(INTEL_FAM6_SKYLAKE_X, core_params),
{}
};
projects / cascardo / linux.git / blobdiff