tjh.dev
Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
Documentation: thermal: Document thermal throttling on Intel platforms
Add documentation for Intel thermal throttling reporting events.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
[ rjw: Subject adjustment, file name change, minor edits ]
Link: https://patch.msgid.link/20251113212104.221632-1-srinivas.pandruvada@linux.intel.com
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
authored by
Srinivas Pandruvada
and committed by
Rafael J. Wysocki
3402bc01
172880f7
+92
+1
Documentation/admin-guide/thermal/index.rst
+1
Documentation/admin-guide/thermal/index.rst
+91
Documentation/admin-guide/thermal/intel_thermal_throttle.rst
+91
Documentation/admin-guide/thermal/intel_thermal_throttle.rst
···
1
+
.. SPDX-License-Identifier: GPL-2.0
2
+
.. include:: <isonum.txt>
3
+
4
+
=======================================
5
+
Intel thermal throttle events reporting
6
+
=======================================
7
+
8
+
:Author: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
9
+
10
+
Introduction
11
+
------------
12
+
13
+
Intel processors have built in automatic and adaptive thermal monitoring
14
+
mechanisms that force the processor to reduce its power consumption in order
15
+
to operate within predetermined temperature limits.
16
+
17
+
Refer to section "THERMAL MONITORING AND PROTECTION" in the "Intel® 64 and
18
+
IA-32 Architectures Software Developer’s Manual Volume 3 (3A, 3B, 3C, & 3D):
19
+
System Programming Guide" for more details.
20
+
21
+
In general, there are two mechanisms to control the core temperature of the
22
+
processor. They are called "Thermal Monitor 1 (TM1) and Thermal Monitor 2 (TM2)".
23
+
24
+
The status of the temperature sensor that triggers the thermal monitor (TM1/TM2)
25
+
is indicated through the "thermal status flag" and "thermal status log flag" in
26
+
MSR_IA32_THERM_STATUS for core level and MSR_IA32_PACKAGE_THERM_STATUS for
27
+
package level.
28
+
29
+
Thermal Status flag, bit 0 — When set, indicates that the processor core
30
+
temperature is currently at the trip temperature of the thermal monitor and that
31
+
the processor power consumption is being reduced via either TM1 or TM2, depending
32
+
on which is enabled. When clear, the flag indicates that the core temperature is
33
+
below the thermal monitor trip temperature. This flag is read only.
34
+
35
+
Thermal Status Log flag, bit 1 — When set, indicates that the thermal sensor has
36
+
tripped since the last power-up or reset or since the last time that software
37
+
cleared this flag. This flag is a sticky bit; once set it remains set until
38
+
cleared by software or until a power-up or reset of the processor. The default
39
+
state is clear.
40
+
41
+
It is possible that when user reads MSR_IA32_THERM_STATUS or
42
+
MSR_IA32_PACKAGE_THERM_STATUS, TM1/TM2 is not active. In this case,
43
+
"Thermal Status flag" will read "0" and the "Thermal Status Log flag" will be set
44
+
to show any previous "TM1/TM2" activation. But since it needs to be cleared by
45
+
the software, it can't show the number of occurrences of "TM1/TM2" activations.
46
+
47
+
Hence, Linux provides counters of how many times the "Thermal Status flag" was
48
+
set. Also presents how long the "Thermal Status flag" was active in milliseconds.
49
+
Using these counters, users can check if the performance was limited because of
50
+
thermal events. It is recommended to read from sysfs instead of directly reading
51
+
MSRs as the "Thermal Status Log flag" is reset by the driver to implement rate
52
+
control.
53
+
54
+
Sysfs Interface
55
+
---------------
56
+
57
+
Thermal throttling events are presented for each CPU under
58
+
"/sys/devices/system/cpu/cpuX/thermal_throttle/", where "X" is the CPU number.
59
+
60
+
All these counters are read-only. They can't be reset to 0. So, they can potentially
61
+
overflow after reaching the maximum 64 bit unsigned integer.
62
+
63
+
``core_throttle_count``
64
+
Shows the number of times "Thermal Status flag" changed from 0 to 1 for this
65
+
CPU since OS boot and thermal vector is initialized. This is a 64 bit counter.
66
+
67
+
``package_throttle_count``
68
+
Shows the number of times "Thermal Status flag" changed from 0 to 1 for the
69
+
package containing this CPU since OS boot and thermal vector is initialized.
70
+
Package status is broadcast to all CPUs; all CPUs in the package increment
71
+
this count. This is a 64-bit counter.
72
+
73
+
``core_throttle_max_time_ms``
74
+
Shows the maximum amount of time for which "Thermal Status flag" has been
75
+
set to 1 for this CPU at the core level since OS boot and thermal vector
76
+
is initialized.
77
+
78
+
``package_throttle_max_time_ms``
79
+
Shows the maximum amount of time for which "Thermal Status flag" has been
80
+
set to 1 for the package containing this CPU since OS boot and thermal
81
+
vector is initialized.
82
+
83
+
``core_throttle_total_time_ms``
84
+
Shows the cumulative time for which "Thermal Status flag" has been
85
+
set to 1 for this CPU for core level since OS boot and thermal vector
86
+
is initialized.
87
+
88
+
``package_throttle_total_time_ms``
89
+
Shows the cumulative time for which "Thermal Status flag" has been set
90
+
to 1 for the package containing this CPU since OS boot and thermal vector
91
+
is initialized.