The comments and the core_busy variable name in
get_target_pstate_use_performance() are totally confusing,
so modify them to reflect what's going on.
The results of the computations should be the same as before.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Notice that get_avg_pstate() can use sample.core_avg_perf instead of
carrying the same division again, so make it do that.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The core_pct_busy field of struct sample actually contains the
average performace during the last sampling period (in percent)
and not the utilization of the core as suggested by its name
which is confusing.
For this reason, change the name of that field to core_avg_perf
and rename the function that computes its value accordingly.
Also notice that storing this value as percentage requires a costly
integer multiplication to be carried out in a hot path, so instead
store it as an "extended fixed point" value with more fraction bits
and update the code using it accordingly (it is better to change the
name of the field along with its meaning in one go than to make those
two changes separately, as that would likely lead to more
confusion).
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Currently, in intel_pstate_clear_update_util_hook(), after
clearing the utilization update hook, we leverage
synchronize_sched() to deal with synchronization, which
is a little bit time-costly because synchronize_sched()
has to wait for all the CPUs to go through a grace period.
Actually, the synchronize_sched() is not necessary if the utilization
update hook has not been set for the given CPU yet, so make the driver
check if that's the case and avoid the synchronize_sched() call then.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=116371
Tested-by: Tian Ye <yex.tian@intel.com>
Signed-off-by: Chen Yu <yu.c.chen@intel.com>
[ rjw : Rebase ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
intel_pstate_get() contains a local variable that's initialized but
never used and it can be written in fewer lines of code, so clean
it up.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
* pm-cpufreq-fixes:
intel_pstate: Fix intel_pstate_get()
cpufreq: intel_pstate: Fix HWP on boot CPU after system resume
cpufreq: st: enable selective initialization based on the platform
cpufreq: intel_pstate: Fix processing for turbo activation ratio
When HWP (hardware P states) feature is active, the ACPI _PSS and _PPC
is not used. So ignore processing for _PPC limits.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
After commit 8fa520af50 "intel_pstate: Remove freq calculation from
intel_pstate_calc_busy()" intel_pstate_get() calls get_avg_frequency()
to compute the average frequency, which is problematic for two reasons.
First, intel_pstate_get() may be invoked before the driver reads the
CPU feedback registers for the first time and if that happens,
get_avg_frequency() will attempt to divide by zero.
Second, the get_avg_frequency() call in intel_pstate_get() is racy
with respect to intel_pstate_sample() and it may end up returning
completely meaningless values for this reason.
Moreover, after commit 7349ec0470 "intel_pstate: Move
intel_pstate_calc_busy() into get_target_pstate_use_performance()"
sample.core_pct_busy is never computed on Atom, but it is used in
intel_pstate_adjust_busy_pstate() in that case too.
To address those problems notice that if sample.core_pct_busy
was used in the average frequency computation carried out by
get_avg_frequency(), both the divide by zero problem and the
race with respect to intel_pstate_sample() would be avoided.
Accordingly, move the invocation of intel_pstate_calc_busy() from
get_target_pstate_use_performance() to intel_pstate_update_util(),
which also will take care of the uninitialized sample.core_pct_busy
on Atom, and modify get_avg_frequency() to use sample.core_pct_busy
as per the above.
Reported-by: kernel test robot <ying.huang@linux.intel.com>
Link: http://marc.info/?l=linux-kernel&m=146226437623173&w=4
Fixes: 8fa520af50 "intel_pstate: Remove freq calculation from intel_pstate_calc_busy()"
Fixes: 7349ec0470 "intel_pstate: Move intel_pstate_calc_busy() into get_target_pstate_use_performance()"
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Commit 41cfd64cf4 "Update frequencies of policy->cpus only from
->set_policy()" changed the way the intel_pstate driver's ->set_policy
callback updates the HWP (hardware-managed P-states) settings.
A side effect of it is that if those settings are modified on the
boot CPU during system suspend and wakeup, they will never be
restored during subsequent system resume.
To address this problem, allow cpufreq drivers that don't provide
->target or ->target_index callbacks to use ->suspend and ->resume
callbacks and add a ->resume callback to intel_pstate to restore
the HWP settings on the CPUs that belong to the given policy.
Fixes: 41cfd64cf4 "Update frequencies of policy->cpus only from ->set_policy()"
Tested-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
For platforms which are controlled via remove node manager, enable _PPC by
default. These platforms are mostly categorized as enterprise server or
performance servers. These platforms needs to go through some
certifications tests, which tests control via _PPC.
The relative risk of enabling by default is low as this is is less likely
that these systems have broken _PSS table.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
When policy->max is changed via _PPC or sysfs and is more than the max non
turbo frequency, it does not really change resulting performance in some
processors. When policy->max results in a P-State ratio more than the
turbo activation ratio, then processor can choose any P-State up to max
turbo. So the user or _PPC setting has no value, but this can cause
undesirable side effects like:
- Showing reduced max percentage in Intel P-State sysfs
- It can cause reduced max performance under certain boundary conditions:
The requested max scaling frequency either via _PPC or via cpufreq-sysfs,
will be converted into a fixed floating point max percent scale. In
majority of the cases this will result in correct max. But not 100% of the
time. If the _PPC is requested at a point where the calculation lead to a
lower max, this can result in a lower P-State then expected and it will
impact performance.
Example of this condition using a Broadwell laptop with config TDP.
ACPI _PSS table from a Broadwell laptop
2301000 2300000 2200000 2000000 1900000 1800000 1700000 1500000 1400000
1300000 1100000 1000000 900000 800000 600000 500000
The actual results by disabling config TDP so that we can get what is
requested on or below 2300000Khz.
scaling_max_freq Max Requested P-State Resultant scaling
max
---------------------------------------- ----------------------
2400000 18 2900000 (max
turbo)
2300000 17 2300000 (max
physical non turbo)
2200000 15 2100000
2100000 15 2100000
2000000 13 1900000
1900000 13 1900000
1800000 12 1800000
1700000 11 1700000
1600000 10 1600000
1500000 f 1500000
1400000 e 1400000
1300000 d 1300000
1200000 c 1200000
1100000 a 1000000
1000000 a 1000000
900000 9 900000
800000 8 800000
700000 7 700000
600000 6 600000
500000 5 500000
------------------------------------------------------------------
Now set the config TDP level 1 ratio as 0x0b (equivalent to 1100000KHz)
in BIOS (not every system will let you adjust this).
The turbo activation ratio will be set to one less than that, which will
be 0x0a (So any request above 1000000KHz should result in turbo region
assuming no thermal limits).
Here _PPC will request max to 1100000KHz (which basically should still
result in turbo as this is more than the turbo activation ratio up to
max allowable turbo frequency), but actual calculation resulted in a max
ceiling P-State which is 0x0a. So under any load condition, this driver
will not request turbo P-States. This will be a huge performance hit.
When config TDP feature is ON, if the _PPC points to a frequency above
turbo activation ratio, the performance can still reach max turbo. In this
case we don't need to treat this as the reduced frequency in set_policy
callback.
In this change when config TDP is active (by checking if the physical max
non turbo ratio is more than the current max non turbo ratio), any request
above current max non turbo is treated as full performance.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
[ rjw : Minor cleanups ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Use ACPI _PPC notification to limit max P state driver will request.
ACPI _PPC change notification is sent by BIOS to limit max P state
in several cases:
- Reduce impact of platform thermal condition
- When Config TDP feature is used, a changed _PPC is sent to
follow TDP change
- Remote node managers in server want to control platform power
via baseboard management controller (BMC)
This change registers with ACPI processor performance lib so that
_PPC changes are notified to cpufreq core, which in turns will
result in call to .setpolicy() callback. Also the way _PSS
table identifies a turbo frequency is not compatible to max turbo
frequency in intel_pstate, so the very first entry in _PSS needs
to be adjusted.
This feature can be turned on by using kernel parameters:
intel_pstate=support_acpi_ppc
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
[ rjw: Minor cleanups ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
When the config TDP level is not nominal (level = 0), the MSR values for
reading level 1 and level 2 ratios contain power in low 14 bits and actual
ratio bits are at bits [23:16]. The current processing for level 1 and
level 2 is wrong as there is no shift done to get actual ratio.
Fixes: 6a35fc2d6c (cpufreq: intel_pstate: get P1 from TAR when available)
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Cc: 4.4+ <stable@vger.kernel.org> # 4.4+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The result returned by pid_calc() is subtracted from current_pstate
(which is the P-State requested during the last period) in order to
obtain the target P-State for the current iteration.
However, current_pstate may not reflect the real current P-State of
the CPU. In particular, that P-State may be higher because of the
frequency sharing per module.
The theory is:
- The load is the percentage of time spent in C0 and is related to
the average P-State during the same period.
- The last requested P-State can be completely different than the
average P-State (because of frequency sharing or throttling).
- The P-State shift computed by the pid_calc is based on the load
computed at average P-State, so the shift must be relative to
this average P-State.
Using the average P-State instead of current P-State improves power
without significant performance penalty in cases when a task migrates
from one core to other core sharing frequency and voltage.
Performance and power comparison with this patch on Cherry Trail
platform using Android:
Benchmark ?Perf ?Power
FishTank 10.45% 3.1%
SmartBench-Gaming -0.1% -10.4%
SmartBench-Productivity -0.8% -10.4%
CandyCrush n/a -17.4%
AngryBirds n/a -5.9%
videoPlayback n/a -13.9%
audioPlayback n/a -4.9%
IcyRocks-20-50 0.0% -38.4%
iozone RR -0.16% -1.3%
iozone RW 0.74% -1.3%
Signed-off-by: Philippe Longepe <philippe.longepe@linux.intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Jörg Otte reports that commit a4675fbc4a (cpufreq: intel_pstate:
Replace timers with utilization update callbacks) caused the CPUs in
his Haswell-based system to stay in the very high frequency region
even if the system is completely idle.
That turns out to be an existing problem in the intel_pstate driver's
P-state selection algorithm for Core processors. Namely, all
decisions made by that algorithm are based on the average frequency
of the CPU between sampling events and on the P-state requested on
the last invocation, so it may get stuck at a very hight frequency
even if the utilization of the CPU is very low (in fact, it may get
stuck in a inadequate P-state regardless of the CPU utilization).
The only way to kick it out of that limbo is a sufficiently long idle
period (3 times longer than the prescribed sampling interval), but if
that doesn't happen often enough (eg. due to a timing change like
after the above commit), the P-state of the CPU may be inadequate
pretty much all the time.
To address the most egregious manifestations of that issue, reset the
core_busy value used to determine the next P-state to request if the
utilization of the CPU, determined with the help of the MPERF
feedback register and the TSC, is below 1%.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=115771
Reported-and-tested-by: Jörg Otte <jrg.otte@gmail.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Prefix the output using the more common kernel style.
Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
[ rjw: Rebase ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
There are multiple places in intel_pstate where int_tofp() is applied
to both arguments of div_fp(), but this is pointless, because int_tofp()
simply shifts its argument to the left by FRAC_BITS which mathematically
is equivalent to multuplication by 2^FRAC_BITS, so if this is done
to both arguments of a division, the extra factors will cancel each
other during that operation anyway.
Drop the pointless int_tofp() applied to div_fp() arguments throughout
the driver.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
No code change. Only added kernel doc style comments for structures.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
When user sets performance policy using cpufreq interface, it is possible
that because of policy->max limits, the actual performance is still
limited. But the current implementation will silently switch the
policy to powersave and start using powersave limits. If user modifies
any limits using intel_pstate sysfs, this is actually changing powersave
limits.
The current implementation tracks limits under powersave and performance
policy using two different variables. When policy->max is less than
policy->cpuinfo.max_freq, only powersave limit variable is used.
This fix causes the performance limits variable to be used always when
the policy is performance.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Replace the single helper for adding and removing cpufreq utilization
update hooks, cpufreq_set_update_util_data(), with a pair of helpers,
cpufreq_add_update_util_hook() and cpufreq_remove_update_util_hook(),
and modify the users of cpufreq_set_update_util_data() accordingly.
With the new helpers, the code using them doesn't need to worry
about the internals of struct update_util_data and in particular
it doesn't need to worry about populating the func field in it
properly upfront.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
The initialization of intel_pstate for a given CPU involves populating
the fields of its struct cpudata that represent the previous sample,
but currently that is done in a problematic way.
Namely, intel_pstate_init_cpu() makes an extra call to
intel_pstate_sample() so it reads the current register values that
will be used to populate the "previous sample" record during the
next invocation of intel_pstate_sample(). However, after commit
a4675fbc4a (cpufreq: intel_pstate: Replace timers with utilization
update callbacks) that doesn't work for last_sample_time, because
the time value is passed to intel_pstate_sample() as an argument now.
Passing 0 to it from intel_pstate_init_cpu() is problematic, because
that causes cpu->last_sample_time == 0 to be visible in
get_target_pstate_use_performance() (and hence the extra
cpu->last_sample_time > 0 check in there) and effectively allows
the first invocation of intel_pstate_sample() from
intel_pstate_update_util() to happen immediately after the
initialization which may lead to a significant "turn on"
effect in the governor algorithm.
To mitigate that issue, rework the initialization to avoid the
extra intel_pstate_sample() call from intel_pstate_init_cpu().
Instead, make intel_pstate_sample() return false if it has been
called with cpu->sample.time equal to zero, which will make
intel_pstate_update_util() skip the sample in that case, and
reset cpu->sample.time from intel_pstate_set_update_util_hook()
to make the algorithm start properly every time the hook is set.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The utilization update hook in the intel_pstate driver is set too
early, as it only should be set after the policy has been fully
initialized by the core. That may cause intel_pstate_update_util()
to use incorrect data and put the CPUs into incorrect P-states as
a result.
To prevent that from happening, make intel_pstate_set_policy() set
the utilization update hook instead of intel_pstate_init_cpu() so
intel_pstate_update_util() only runs when all things have been
initialized as appropriate.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
After commit a4675fbc4a (cpufreq: intel_pstate: Replace timers with
utilization update callbacks) wrmsrl_on_cpu() cannot be called in the
intel_pstate_adjust_busy_pstate() path as that is executed with
disabled interrupts. However, atom_set_pstate() called from there
via intel_pstate_set_pstate() uses wrmsrl_on_cpu() to update the
IA32_PERF_CTL MSR which triggers the WARN_ON_ONCE() in
smp_call_function_single().
The reason why wrmsrl_on_cpu() is used by atom_set_pstate() is
because intel_pstate_set_pstate() calling it is also invoked during
the initialization and cleanup of the driver and in those cases it is
not guaranteed to be run on the CPU that is being updated. However,
in the case when intel_pstate_set_pstate() is called by
intel_pstate_adjust_busy_pstate(), wrmsrl() can be used to update
the register safely. Moreover, intel_pstate_set_pstate() already
contains code that only is executed if the function is called by
intel_pstate_adjust_busy_pstate() and there is a special argument
passed to it because of that.
To fix the problem at hand, rearrange the code taking the above
observations into account.
First, replace the ->set() callback in struct pstate_funcs with a
->get_val() one that will return the value to be written to the
IA32_PERF_CTL MSR without updating the register.
Second, split intel_pstate_set_pstate() into two functions,
intel_pstate_update_pstate() to be called by
intel_pstate_adjust_busy_pstate() that will contain all of the
intel_pstate_set_pstate() code which only needs to be executed in
that case and will use wrmsrl() to update the MSR (after obtaining
the value to write to it from the ->get_val() callback), and
intel_pstate_set_min_pstate() to be invoked during the
initialization and cleanup that will set the P-state to the
minimum one and will update the MSR using wrmsrl_on_cpu().
Finally, move the code shared between intel_pstate_update_pstate()
and intel_pstate_set_min_pstate() to a new static inline function
intel_pstate_record_pstate() and make them both call it.
Of course, that unifies the handling of the IA32_PERF_CTL MSR writes
between Atom and Core.
Fixes: a4675fbc4a (cpufreq: intel_pstate: Replace timers with utilization update callbacks)
Reported-and-tested-by: Josh Boyer <jwboyer@fedoraproject.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
If the current value of MPERF or the current value of TSC is the
same as the previous one, respectively, intel_pstate_sample() bails
out early and skips the sample.
However, intel_pstate_adjust_busy_pstate() is still called in that
case which is not correct, so modify intel_pstate_sample() to
return a bool value indicating whether or not the sample has been
taken and use it to decide whether or not to call
intel_pstate_adjust_busy_pstate().
While at it, remove redundant parentheses from the MPERF/TSC
check in intel_pstate_sample().
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Use a helper function to compute the average pstate and call it only
where it is needed (only when tracing or in intel_pstate_get).
Signed-off-by: Philippe Longepe <philippe.longepe@linux.intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The cpu_load algorithm doesn't need to invoke intel_pstate_calc_busy(),
so move that call from intel_pstate_sample() to
get_target_pstate_use_performance().
Signed-off-by: Philippe Longepe <philippe.longepe@linux.intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
mul_fp(int_tofp(A), B) expands to:
((A << FRAC_BITS) * B) >> FRAC_BITS, so the same result can be obtained
via simple multiplication A * B. Apply this observation to
max_perf * limits->max_perf and max_perf * limits->min_perf in
intel_pstate_get_min_max()."
Signed-off-by: Philippe Longepe <philippe.longepe@linux.intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
pid->setpoint and pid->deadband can be initialized in fixed point, so we
can avoid the int_tofp in pid_calc.
Signed-off-by: Philippe Longepe <philippe.longepe@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Use the observation that cpufreq_update_util() is only called
by the scheduler with rq->lock held, so the callers of
cpufreq_set_update_util_data() can use synchronize_sched()
instead of synchronize_rcu() to wait for cpufreq_update_util()
to complete. Moreover, if they are updated to do that,
rcu_read_(un)lock() calls in cpufreq_update_util() might be
replaced with rcu_read_(un)lock_sched(), respectively, but
those aren't really necessary, because the scheduler calls
that function from RCU-sched read-side critical sections
already.
In addition to that, if cpufreq_set_update_util_data() checks
the func field in the struct update_util_data before setting
the per-CPU pointer to it, the data->func check may be dropped
from cpufreq_update_util() as well.
Make the above changes to reduce the overhead from
cpufreq_update_util() in the scheduler paths invoking it
and to make the cleanup after removing its callbacks less
heavy-weight somewhat.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Instead of using a per-CPU deferrable timer for utilization sampling
and P-states adjustments, register a utilization update callback that
will be invoked from the scheduler on utilization changes.
The sampling rate is still the same as what was used for the deferrable
timers, so the functional impact of this patch should not be significant.
Based on an earlier patch from Srinivas Pandruvada.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Disable HWP Interrupt notification before enabling HWP. Since we don't
have HWP interrupt handling for possible performance interrupts, there
is not much use of enabling HWP interrupts.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
If the processor supports HWP, enable it by default without checking
for the cpu model. This will allow to enable HWP in all supported
processors without driver change.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The intel-pstate driver is using intel_pstate_hwp_set() from two
separate paths, i.e. ->set_policy() callback and sysfs update path for
the files present in /sys/devices/system/cpu/intel_pstate/ directory.
While an update to the sysfs path applies to all the CPUs being managed
by the driver (which essentially means all the online CPUs), the update
via the ->set_policy() callback applies to a smaller group of CPUs
managed by the policy for which ->set_policy() is called.
And so, intel_pstate_hwp_set() should update frequencies of only the
CPUs that are part of policy->cpus mask, while it is called from
->set_policy() callback.
In order to do that, add a parameter (cpumask) to intel_pstate_hwp_set()
and apply the frequency changes only to the concerned CPUs.
For ->set_policy() path, we are only concerned about policy->cpus, and
so policy->rwsem lock taken by the core prior to calling ->set_policy()
is enough to take care of any races. The larger lock acquired by
get_online_cpus() is required only for the updates to sysfs files.
Add another routine, intel_pstate_hwp_set_online_cpus(), and call it
from the sysfs update paths.
This also fixes a lockdep reported recently, where policy->rwsem and
get_online_cpus() could have been acquired in any order causing an ABBA
deadlock. The sequence of events leading to that was:
intel_pstate_init(...)
...cpufreq_online(...)
down_write(&policy->rwsem); // Locks policy->rwsem
...
cpufreq_init_policy(policy);
...intel_pstate_hwp_set();
get_online_cpus(); // Temporarily locks cpu_hotplug.lock
...
up_write(&policy->rwsem);
pm_suspend(...)
...disable_nonboot_cpus()
_cpu_down()
cpu_hotplug_begin(); // Locks cpu_hotplug.lock
__cpu_notify(CPU_DOWN_PREPARE, ...);
...cpufreq_offline_prepare();
down_write(&policy->rwsem); // Locks policy->rwsem
Reported-and-tested-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
785ee27 ("cpufreq: intel_pstate: Fix limits->max_perf rounding error")
hardcodes the value of FRAC_BITS. This patch fixes that minor issue.
Fixes: 785ee27881 (cpufreq: intel_pstate: Fix limits->max_perf rounding error)
Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
In cases where we have many IOs, the global load becomes low and the
load algorithm will decrease the requested P-State. Because of that,
the IOs overheads will increase and impact the IO performances.
To improve IO bound work, we can count the io-wait time as busy time
in calculating CPU busy.
This change uses get_cpu_iowait_time_us() to obtain the IO wait time value
and converts time into number of cycles spent waiting on IO at the TSC
rate. At the moment, this trick is only used for Atom.
Signed-off-by: Philippe Longepe <philippe.longepe@intel.com>
Signed-off-by: Stephane Gasparini <stephane.gasparini@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The current function to calculate cpu utilization uses the average P-state
ratio (APerf/Mperf) scaled by the ratio of the current P-state to the
max available non-turbo one. This leads to an overestimation of
utilization which causes higher-performance P-states to be selected more
often and that leads to increased energy consumption.
This is a problem for low-power systems, so it is better to use a
different utilization calculation algorithm for them.
Namely, the Percent Busy value (or load) can be estimated as the ratio of the
MPERF counter that runs at a constant rate only during active periods (C0) to
the time stamp counter (TSC) that also runs (at the same rate) during idle.
That is:
Percent Busy = 100 * (delta_mperf / delta_tsc)
Use this algorithm for platforms with SoCs based on the Airmont and Silvermont
Atom cores.
Signed-off-by: Philippe Longepe <philippe.longepe@intel.com>
Signed-off-by: Stephane Gasparini <stephane.gasparini@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Target systems using different cpus have different power and performance
requirements. They may use different algorithms to get the next P-state
based on their power or performance preference.
For example, power-constrained systems may not want to use
high-performance P-states as aggressively as a full-size desktop or a
server platform. A server platform may want to run close to the max to
achieve better performance, while laptop-like systems may prefer
sacrificing performance for longer battery lifes.
For the above reasons, modify intel_pstate to allow the target P-state
selection algorithm to be depend on the CPU ID.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Philippe Longepe <philippe.longepe@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
If hardware-driven P-state selection (HWP) is enabled, the
"performance" mode of intel_pstate should only allow the processor
to use the highest-performance P-state available. That is not
the case currently, so make it actually happen.
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Alexandra Yates <alexandra.yates@linux.intel.com>
[ rjw: Subject and changelog ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
A rounding error was found in the calculation of limits->max_perf
in intel_pstate_set_policy(), which is used to calculate the max and min
pstate values in intel_pstate_get_min_max(). In that code,
limits->max_perf is truncated to 2 hex digits such that, for example,
0x169 was incorrectly calculated to 0x16 instead of 0x17. This resulted in
the pstate being set one level too low. This patch rounds the value of
limits->max_perf up instead of down so that the correct max pstate can
be reached.
Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
I have a Intel (6,63) processor with a "marketing" frequency (from
/proc/cpuinfo) of 2100MHz, and a max turbo frequency of 2600MHz. I
can execute
cpupower frequency-set -g powersave --min 1200MHz --max 2100MHz
and the max_freq_pct is set to 80. When adding load to the system I noticed
that the cpu frequency only reached 2000MHZ and not 2100MHz as expected.
This is because limits->max_policy_pct is calculated as 2100 * 100 /2600 = 80.7
and is rounded down to 80 when it should be rounded up to 81. This patch
adds a DIV_ROUND_UP() which will return the correct value.
Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
There are two flavors of Atom cores to be supported by intel_pstate,
Silvermont and Airmont, so make the driver distinguish between them by
adding separate frequency tables.
Separate the CPU defaults params for each of them and match the CPU IDs
against them as appropriate.
Signed-off-by: Philippe Longepe <philippe.longepe@linux.intel.com>
Signed-off-by: Stephane Gasparini <stephane.gasparini@linux.intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
[ rjw: Subject and changelog ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Rename symbol and function names starting with "BYT" or "byt" to
start with "ATOM" or "atom", respectively, so as to make it clear
that they may apply to Atom in general and not just to Baytrail
(the goal is to support several Atoms architectures eventually).
This should not lead to any functional changes.
Signed-off-by: Philippe Longepe <philippe.longepe@linux.intel.com>
Signed-off-by: Stephane Gasparini <stephane.gasparini@linux.intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
[ rjw : Changelog ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Revert commit 37afb00032 (cpufreq: intel_pstate: Use ACPI perf
configuration) that is reported to cause a regression to happen
on a system where invalid data are returned by the ACPI _PSS object.
Since that commit makes assumptions regarding the _PSS output
correctness that may turn out to be overly optimistic in general,
there is a concern that it may introduce regression on more
systems, so it's better to revert it now and we'll revisit the
underlying issue in the next cycle with a more robust solution.
Conflicts:
drivers/cpufreq/intel_pstate.c
Fixes: 37afb00032 (cpufreq: intel_pstate: Use ACPI perf configuration)
Reported-by: Borislav Petkov <bp@alien8.de>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Revert commit 4ef4514870 (cpufreq: intel_pstate: Avoid calculation for
max/min) as it depends on commit 37afb00032 (cpufreq: intel_pstate: Use
ACPI perf configuration) that causes problems to happen and needs to be
reverted.
Conflicts:
drivers/cpufreq/intel_pstate.c
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
When booting an HWP enabled system the kernel displays one "HWP enabled"
message for each cpu. The messages are superfluous since HWP is globally
enabled across all CPUs. This patch also adds an informational message
when HWP is disabled via intel_pstate=no_hwp.
Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
On systems that initialize the intel_pstate driver with the performance
governor, and then switch to the powersave governor will not transition to
lower cpu frequencies until /sys/devices/system/cpu/intel_pstate/min_perf_pct
is set to a low value.
The behavior of governor switching changed after commit a04759924e
("[cpufreq] intel_pstate: honor user space min_perf_pct override on
resume"). The commit introduced tracking of performance percentage
changes via sysfs in order to restore userspace changes during
suspend/resume. The problem occurs because the global values of the newly
introduced max_sysfs_pct and min_sysfs_pct are not lowered on the governor
change and this causes the powersave governor to inherit the performance
governor's settings.
A simple change would have been to reset max_sysfs_pct to 100 and
min_sysfs_pct to 0 on a governor change, which fixes the problem with
governor switching. However, since we cannot break userspace[1] the fix
is now to give each governor its own limits storage area so that governor
specific changes are tracked.
I successfully tested this by booting with both the performance governor
and the powersave governor by default, and switching between the two
governors (while monitoring /sys/devices/system/cpu/intel_pstate/ values,
and looking at the output of cpupower frequency-info). Suspend/Resume
testing was performed by Doug Smythies.
[1] Systems which suspend/resume using the unmaintained pm-utils package
will always transition to the performance governor before the suspend and
after the resume. This means a system using the powersave governor will
go from powersave to performance, then suspend/resume, performance to
powersave. The simple change during governor changes would have been
overwritten when the governor changed before and after the suspend/resume.
I have submitted https://bugzilla.redhat.com/show_bug.cgi?id=1271225
against Fedora to remove the 94cpufreq file that causes the problem. It
should be noted that pm-utils is obsoleted with newer versions of systemd.
Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Acked-by: Kristen Carlson Accardi <kristen@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
