| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <5725513.D8MRamvf82@vostro.rjw.lan>
Date: Thu, 25 Feb 2016 22:20:17 +0100
From: "Rafael J. Wysocki" <rjw@...ysocki.net>
To: Linux PM list <linux-pm@...r.kernel.org>
Cc: Juri Lelli <juri.lelli@....com>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Viresh Kumar <viresh.kumar@...aro.org>,
Srinivas Pandruvada <srinivas.pandruvada@...ux.intel.com>,
Peter Zijlstra <peterz@...radead.org>,
Steve Muckle <steve.muckle@...aro.org>,
Ingo Molnar <mingo@...nel.org>
Subject: [RFC/RFT][PATCH v4 2/2] cpufreq: schedutil: Switching frequencies from interrupt context
From: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
Subject: [PATCH] cpufreq: schedutil: Switching frequencies from interrupt context
Modify the ACPI cpufreq driver to provide a method for switching
CPU frequencies from interrupt context and update the cpufreq core
and the schedutil governor to use that method if available.
Introduce a new cpufreq driver callback, ->fast_switch, to be
invoked for frequency switching from interrupt context via
new helper function cpufreq_driver_fast_switch().
Modify the schedutil governor to call cpufreq_driver_fast_switch()
from its sugov_update_commit() function and avoid queuing up the
irq_work if that is successful.
Implement the ->fast_switch callback in the ACPI cpufreq driver
(with a limited coverage for the time being).
In addition to the above, cpufreq_governor_limits() is modified so
it doesn't call __cpufreq_driver_target() to enforce the new limits
immediately if the fast_switch_enabled flag is set for the policy,
because in that case the frequency will be updated immediately
using the new limits anyway and the additional invocation of
__cpufreq_driver_target() might be racing with that violating
the cpufreq_driver_fast_switch() requirements.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
---
Changes from v3:
- Rebase on the v4 of patch [1/2].
- Change the table lookup algo in acpi_cpufreq_fast_switch() to take the table
ordering into account and to prefer higher frequencies (if the target falls
between two of them).
Changes from v2:
- Rework due to the revised handling of policy->min and policy->max changes.
v2 was the first version of this patch.
---
drivers/cpufreq/acpi-cpufreq.c | 63 ++++++++++++++++++++++++++++++++++++
drivers/cpufreq/cpufreq.c | 31 +++++++++++++++++
drivers/cpufreq/cpufreq_governor.c | 13 ++++---
drivers/cpufreq/cpufreq_governor.h | 1
drivers/cpufreq/cpufreq_schedutil.c | 16 ++++++++-
include/linux/cpufreq.h | 5 ++
6 files changed, 122 insertions(+), 7 deletions(-)
Index: linux-pm/drivers/cpufreq/acpi-cpufreq.c
===================================================================
--- linux-pm.orig/drivers/cpufreq/acpi-cpufreq.c
+++ linux-pm/drivers/cpufreq/acpi-cpufreq.c
@@ -70,6 +70,7 @@ struct acpi_cpufreq_data {
unsigned int cpu_feature;
unsigned int acpi_perf_cpu;
cpumask_var_t freqdomain_cpus;
+ void (*cpu_freq_fast_write)(u32 val);
};
/* acpi_perf_data is a pointer to percpu data. */
@@ -243,6 +244,15 @@ static unsigned extract_freq(u32 val, st
}
}
+void cpu_freq_fast_write_intel(u32 val)
+{
+ u32 lo, hi;
+
+ rdmsr(MSR_IA32_PERF_CTL, lo, hi);
+ lo = (lo & ~INTEL_MSR_RANGE) | (val & INTEL_MSR_RANGE);
+ wrmsr(MSR_IA32_PERF_CTL, lo, hi);
+}
+
struct msr_addr {
u32 reg;
};
@@ -484,6 +494,53 @@ out:
return result;
}
+unsigned int acpi_cpufreq_fast_switch(struct cpufreq_policy *policy,
+ unsigned int target_freq)
+{
+ struct acpi_cpufreq_data *data = policy->driver_data;
+ struct acpi_processor_performance *perf;
+ struct cpufreq_frequency_table *entry, *next;
+ unsigned int next_perf_state, next_freq, freq;
+
+ /*
+ * Find the closest frequency above target_freq or equal to it.
+ *
+ * The table is sorted in the reverse order with respect to the
+ * frequency and all of the entires are valid (see the initialization).
+ */
+ entry = data->freq_table;
+ do {
+ next = entry;
+ entry++;
+ freq = entry->frequency;
+ } while (freq >= target_freq && freq != CPUFREQ_TABLE_END);
+
+ next_freq = next->frequency;
+ /*
+ * The frequency below the target may be more suitable, so check and
+ * choose it if that's the case.
+ */
+ if (next_freq != target_freq && freq != CPUFREQ_TABLE_END &&
+ (freq + ((next_freq - freq) >> 2) > target_freq)) {
+ next_freq = freq;
+ next_perf_state = entry->driver_data;
+ } else {
+ next_perf_state = next->driver_data;
+ }
+
+ perf = to_perf_data(data);
+ if (perf->state == next_perf_state) {
+ if (unlikely(data->resume))
+ data->resume = 0;
+ else
+ return next_freq;
+ }
+
+ data->cpu_freq_fast_write(perf->states[next_perf_state].control);
+ perf->state = next_perf_state;
+ return next_freq;
+}
+
static unsigned long
acpi_cpufreq_guess_freq(struct acpi_cpufreq_data *data, unsigned int cpu)
{
@@ -745,6 +802,7 @@ static int acpi_cpufreq_cpu_init(struct
pr_debug("HARDWARE addr space\n");
if (check_est_cpu(cpu)) {
data->cpu_feature = SYSTEM_INTEL_MSR_CAPABLE;
+ data->cpu_freq_fast_write = cpu_freq_fast_write_intel;
break;
}
if (check_amd_hwpstate_cpu(cpu)) {
@@ -760,6 +818,10 @@ static int acpi_cpufreq_cpu_init(struct
goto err_unreg;
}
+ policy->fast_switch_possible = data->cpu_freq_fast_write != NULL &&
+ !acpi_pstate_strict && !(policy_is_shared(policy) &&
+ policy->shared_type != CPUFREQ_SHARED_TYPE_ANY);
+
data->freq_table = kzalloc(sizeof(*data->freq_table) *
(perf->state_count+1), GFP_KERNEL);
if (!data->freq_table) {
@@ -894,6 +956,7 @@ static struct freq_attr *acpi_cpufreq_at
static struct cpufreq_driver acpi_cpufreq_driver = {
.verify = cpufreq_generic_frequency_table_verify,
.target_index = acpi_cpufreq_target,
+ .fast_switch = acpi_cpufreq_fast_switch,
.bios_limit = acpi_processor_get_bios_limit,
.init = acpi_cpufreq_cpu_init,
.exit = acpi_cpufreq_cpu_exit,
Index: linux-pm/include/linux/cpufreq.h
===================================================================
--- linux-pm.orig/include/linux/cpufreq.h
+++ linux-pm/include/linux/cpufreq.h
@@ -82,6 +82,7 @@ struct cpufreq_policy {
void *governor_data;
bool governor_enabled; /* governor start/stop flag */
char last_governor[CPUFREQ_NAME_LEN]; /* last governor used */
+ bool fast_switch_possible;
struct work_struct update; /* if update_policy() needs to be
* called, but you're in IRQ context */
@@ -271,6 +272,8 @@ struct cpufreq_driver {
unsigned int relation); /* Deprecated */
int (*target_index)(struct cpufreq_policy *policy,
unsigned int index);
+ unsigned int (*fast_switch)(struct cpufreq_policy *policy,
+ unsigned int target_freq);
/*
* Only for drivers with target_index() and CPUFREQ_ASYNC_NOTIFICATION
* unset.
@@ -485,6 +488,8 @@ struct cpufreq_governor {
};
/* Pass a target to the cpufreq driver */
+void cpufreq_driver_fast_switch(struct cpufreq_policy *policy,
+ unsigned int target_freq);
int cpufreq_driver_target(struct cpufreq_policy *policy,
unsigned int target_freq,
unsigned int relation);
Index: linux-pm/drivers/cpufreq/cpufreq.c
===================================================================
--- linux-pm.orig/drivers/cpufreq/cpufreq.c
+++ linux-pm/drivers/cpufreq/cpufreq.c
@@ -1814,6 +1814,37 @@ EXPORT_SYMBOL(cpufreq_unregister_notifie
* GOVERNORS *
*********************************************************************/
+/**
+ * cpufreq_driver_fast_switch - Carry out a fast CPU frequency switch.
+ * @policy: cpufreq policy to switch the frequency for.
+ * @target_freq: New frequency to set (may be approximate).
+ *
+ * Carry out a fast frequency switch from interrupt context.
+ *
+ * This function must not be called if policy->fast_switch_possible is unset.
+ *
+ * Governors calling this function must guarantee that it will never be invoked
+ * twice in parallel for the same policy and that it will never be called in
+ * parallel with either ->target() or ->target_index() for the same policy.
+ *
+ * If CPUFREQ_ENTRY_INVALID is returned by the driver's ->fast_switch()
+ * callback, the hardware configuration must be preserved.
+ */
+void cpufreq_driver_fast_switch(struct cpufreq_policy *policy,
+ unsigned int target_freq)
+{
+ unsigned int freq;
+
+ if (target_freq == policy->cur)
+ return;
+
+ freq = cpufreq_driver->fast_switch(policy, target_freq);
+ if (freq != CPUFREQ_ENTRY_INVALID) {
+ policy->cur = freq;
+ trace_cpu_frequency(freq, smp_processor_id());
+ }
+}
+
/* Must set freqs->new to intermediate frequency */
static int __target_intermediate(struct cpufreq_policy *policy,
struct cpufreq_freqs *freqs, int index)
Index: linux-pm/drivers/cpufreq/cpufreq_governor.c
===================================================================
--- linux-pm.orig/drivers/cpufreq/cpufreq_governor.c
+++ linux-pm/drivers/cpufreq/cpufreq_governor.c
@@ -613,15 +613,18 @@ static int cpufreq_governor_limits(struc
mutex_lock(&policy_dbs->timer_mutex);
- if (policy->max < policy->cur)
- __cpufreq_driver_target(policy, policy->max, CPUFREQ_RELATION_H);
- else if (policy->min > policy->cur)
- __cpufreq_driver_target(policy, policy->min, CPUFREQ_RELATION_L);
+ if (!policy_dbs->fast_switch_enabled) {
+ if (policy->max < policy->cur)
+ __cpufreq_driver_target(policy, policy->max,
+ CPUFREQ_RELATION_H);
+ else if (policy->min > policy->cur)
+ __cpufreq_driver_target(policy, policy->min,
+ CPUFREQ_RELATION_L);
+ }
gov_update_sample_delay(policy_dbs, 0);
mutex_unlock(&policy_dbs->timer_mutex);
-
return 0;
}
Index: linux-pm/drivers/cpufreq/cpufreq_schedutil.c
===================================================================
--- linux-pm.orig/drivers/cpufreq/cpufreq_schedutil.c
+++ linux-pm/drivers/cpufreq/cpufreq_schedutil.c
@@ -86,8 +86,19 @@ static void sugov_update_commit(struct p
struct sugov_policy *sg_policy = to_sg_policy(policy_dbs);
policy_dbs->last_sample_time = time;
- if (sg_policy->next_freq != next_freq) {
- sg_policy->next_freq = next_freq;
+ if (sg_policy->next_freq == next_freq)
+ return;
+
+ sg_policy->next_freq = next_freq;
+ if (policy_dbs->fast_switch_enabled) {
+ cpufreq_driver_fast_switch(policy_dbs->policy, next_freq);
+ /*
+ * Restore the sample delay in case it has been set to 0 from
+ * sysfs in the meantime.
+ */
+ gov_update_sample_delay(policy_dbs,
+ policy_dbs->dbs_data->sampling_rate);
+ } else {
policy_dbs->work_in_progress = true;
irq_work_queue(&policy_dbs->irq_work);
}
@@ -191,6 +202,7 @@ static bool sugov_start(struct cpufreq_p
gov_update_sample_delay(policy_dbs, policy_dbs->dbs_data->sampling_rate);
policy_dbs->last_sample_time = 0;
+ policy_dbs->fast_switch_enabled = policy->fast_switch_possible;
sg_policy->next_freq = UINT_MAX;
for_each_cpu(cpu, policy->cpus) {
Index: linux-pm/drivers/cpufreq/cpufreq_governor.h
===================================================================
--- linux-pm.orig/drivers/cpufreq/cpufreq_governor.h
+++ linux-pm/drivers/cpufreq/cpufreq_governor.h
@@ -124,6 +124,7 @@ struct policy_dbs_info {
/* Status indicators */
bool is_shared; /* This object is used by multiple CPUs */
bool work_in_progress; /* Work is being queued up or in progress */
+ bool fast_switch_enabled; /* Switch frequencies from interrup context */
};
static inline void gov_update_sample_delay(struct policy_dbs_info *policy_dbs,
Powered by blists - more mailing lists