The current driver assumes all RAPL domains within a CPU package
have the same energy unit. This is no longer true for HSW server
CPUs since DRAM domain has is own fixed energy unit which can be
different than the package energy unit enumerated by package
power MSR. In fact, the default HSW EP package power unit is 61uJ
whereas DRAM domain unit is 15.3uJ. The result is that DRAM power
consumption is counted 4x more than real power reported by energy
counters, similarly for max_energy_range_uj of DRAM domain.
This patch adds domain specific energy unit per cpu type, it allows
domain energy unit to override package energy unit if non zero.
Please see this document for details.
"Intel Xeon Processor E5-1600 and E5-2600 v3 Product Families, Volume 2 of 2.
Datasheet, September 2014, Reference Number: 330784-001 "
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Cc: 3.10+ <stable@vger.kernel.org> # 3.10+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Enable RAPL driver on Xeon cpu id 0x56.
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
RAPL on Atom depends on IOSF_MBI driver for setting
floor frequency. This patch adds explicit dependency on
CONFIG_IOSF_MB.
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
RAPL implementation on Atom has made some changes that are not compatible
with Core CPUs. Specifically, it is different in the way units are computed
as well as floor frequency is enforced.
This patch uses the per CPU model functions to handle the differences. Intel
Software Developers' Manual has also been updated to reflect the changes in
Table 35-7 V3C.
Signed-off-by: Ajay Thomas <ajay.thomas.david.rajamanickam@intel.com>
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
RAPL implementations may vary slightly between Core and Atom CPUs. There are
also potential differences among generations of CPUs within the family.
This patch adds a per model structure to abstract the differences such that
variations can be handled cleanly.
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Many CPUs do not support complete set of RAPL domains, as a
result this detection failed message is very misleading and
can be annoying.
[ 5.082632] intel_rapl: RAPL domain core detection failed
[ 5.088370] intel_rapl: RAPL domain uncore detection failed
So lower the warning message to info and only print out the RAPL
domains that are supported.
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
I've confirmed that monitoring the package power usage as well as setting power
limits appear to be working as expected. Supports the package and dram domains.
Tested aginst cpu:
Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz
Signed-off-by: Jason Baron <jbaron@akamai.com>
Acked-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Add support for Broadwell model 0x3d and Haswell model (0x3c).
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Energy counters may roll slowly for some RAPL domains, checking all
of them can be time consuming and takes unpredictable amount of time.
Therefore, we relax the sanity check by only checking availability of the
MSRs and non-zero value of the energy status counters. It has been shown
sufficient for all the platforms tested to filter out inactive domains.
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Acked-by: Durgadoss R <durgadoss.r@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The purpose of this single series of commits from Srivatsa S Bhat (with
a small piece from Gautham R Shenoy) touching multiple subsystems that use
CPU hotplug notifiers is to provide a way to register them that will not
lead to deadlocks with CPU online/offline operations as described in the
changelog of commit 93ae4f978c (CPU hotplug: Provide lockless versions
of callback registration functions).
The first three commits in the series introduce the API and document it
and the rest simply goes through the users of CPU hotplug notifiers and
converts them to using the new method.
/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)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=SW7Q
-----END PGP SIGNATURE-----
Merge tag 'cpu-hotplug-3.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull CPU hotplug notifiers registration fixes from Rafael Wysocki:
"The purpose of this single series of commits from Srivatsa S Bhat
(with a small piece from Gautham R Shenoy) touching multiple
subsystems that use CPU hotplug notifiers is to provide a way to
register them that will not lead to deadlocks with CPU online/offline
operations as described in the changelog of commit 93ae4f978c ("CPU
hotplug: Provide lockless versions of callback registration
functions").
The first three commits in the series introduce the API and document
it and the rest simply goes through the users of CPU hotplug notifiers
and converts them to using the new method"
* tag 'cpu-hotplug-3.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (52 commits)
net/iucv/iucv.c: Fix CPU hotplug callback registration
net/core/flow.c: Fix CPU hotplug callback registration
mm, zswap: Fix CPU hotplug callback registration
mm, vmstat: Fix CPU hotplug callback registration
profile: Fix CPU hotplug callback registration
trace, ring-buffer: Fix CPU hotplug callback registration
xen, balloon: Fix CPU hotplug callback registration
hwmon, via-cputemp: Fix CPU hotplug callback registration
hwmon, coretemp: Fix CPU hotplug callback registration
thermal, x86-pkg-temp: Fix CPU hotplug callback registration
octeon, watchdog: Fix CPU hotplug callback registration
oprofile, nmi-timer: Fix CPU hotplug callback registration
intel-idle: Fix CPU hotplug callback registration
clocksource, dummy-timer: Fix CPU hotplug callback registration
drivers/base/topology.c: Fix CPU hotplug callback registration
acpi-cpufreq: Fix CPU hotplug callback registration
zsmalloc: Fix CPU hotplug callback registration
scsi, fcoe: Fix CPU hotplug callback registration
scsi, bnx2fc: Fix CPU hotplug callback registration
scsi, bnx2i: Fix CPU hotplug callback registration
...
Subsystems that want to register CPU hotplug callbacks, as well as perform
initialization for the CPUs that are already online, often do it as shown
below:
get_online_cpus();
for_each_online_cpu(cpu)
init_cpu(cpu);
register_cpu_notifier(&foobar_cpu_notifier);
put_online_cpus();
This is wrong, since it is prone to ABBA deadlocks involving the
cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently
with CPU hotplug operations).
Instead, the correct and race-free way of performing the callback
registration is:
cpu_notifier_register_begin();
for_each_online_cpu(cpu)
init_cpu(cpu);
/* Note the use of the double underscored version of the API */
__register_cpu_notifier(&foobar_cpu_notifier);
cpu_notifier_register_done();
Fix the intel-rapl code in the powercap driver by using this latter form
of callback registration. But retain the calls to get/put_online_cpus(),
since they also protect the function rapl_cleanup_data(). By nesting
get/put_online_cpus() *inside* cpu_notifier_register_begin/done(), we avoid
the ABBA deadlock possibility mentioned above.
Cc: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Cc: Ingo Molnar <mingo@kernel.org>
Tested-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Spell out names for supported SoCs.
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Some RAPL domains may not be active at the time driver is being
loaded. Checking energy counter increment may be too strict and
time consuming. So relax the sanity check on energy counters of
these domains.
Otherwise, they may be ignored and become unavailable to the
powercap framework.
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch adds support for RAPL on Intel ValleyView based SoC
platforms, such as Baytrail.
Besides adding CPU ID, special energy unit encoding is handled
for ValleyView.
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
As per the documentation of powercap sysfs, energy_uj field is read only,
if it can't be reset. Currently it always allows write but will fail,
if there is no reset callback.
Changing mode field, to read only if there is no reset callback.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Reported-by: Dirk Brandewie <dirk.j.brandewie@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The newly added power capping framework uses the obsolete .dev_attrs
field of struct class. However this field will be removed in 3.13, so
convert the code to use the .dev_groups field instead.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The Intel Running Average Power Limit (RAPL) technology provides platform
software with the ability to monitor, control, and get notifications on
power usage.
This feature is present in all Sandy Bridge and later Intel processors.
Newer models allow more fine grained controls to be applied. In RAPL,
power control is divided into domains, which include package, DRAM
controller, CPU core (Power Plane 0), graphics uncore (power plane 1), etc.
The purpose of this driver is to expose the RAPL settings to userspace.
Overall, RAPL fits in the new powercap class driver in that platform
level power capping controls are exposed via this generic interface.
This driver is based on an earlier patch from Zhang Rui.
However, while the previous work was mainly focused on thermal monitoring
the focus here is on the usability from user space perspective.
References: https://lkml.org/lkml/2011/5/26/93
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The power capping framework providing a consistent interface between the
kernel and user space that allows power capping drivers to expose their
settings to user space in a uniform way.
The overall design of the framework is described in the documentation
added by the previous patch in this series.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
