Documentation/ABI/testing/sysfs-devices-system-cpu

   1 What:           /sys/devices/system/cpu/
   2 Date:           pre-git history
   3 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
   4 Description:
   5                 A collection of both global and individual CPU attributes
   6
   7                 Individual CPU attributes are contained in subdirectories
   8                 named by the kernel's logical CPU number, e.g.:
   9
  10                 /sys/devices/system/cpu/cpu#/
  11
  12 What:           /sys/devices/system/cpu/sched_mc_power_savings
  13                 /sys/devices/system/cpu/sched_smt_power_savings
  14 Date:           June 2006
  15 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
  16 Description:    Discover and adjust the kernel's multi-core scheduler support.
  17
  18                 Possible values are:
  19
  20                 0 - No power saving load balance (default value)
  21                 1 - Fill one thread/core/package first for long running threads
  22                 2 - Also bias task wakeups to semi-idle cpu package for power
  23                     savings
  24
  25                 sched_mc_power_savings is dependent upon SCHED_MC, which is
  26                 itself architecture dependent.
  27
  28                 sched_smt_power_savings is dependent upon SCHED_SMT, which
  29                 is itself architecture dependent.
  30
  31                 The two files are independent of each other. It is possible
  32                 that one file may be present without the other.
  33
  34                 Introduced by git commit 5c45bf27.
  35
  36
  37 What:           /sys/devices/system/cpu/kernel_max
  38                 /sys/devices/system/cpu/offline
  39                 /sys/devices/system/cpu/online
  40                 /sys/devices/system/cpu/possible
  41                 /sys/devices/system/cpu/present
  42 Date:           December 2008
  43 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
  44 Description:    CPU topology files that describe kernel limits related to
  45                 hotplug. Briefly:
  46
  47                 kernel_max: the maximum cpu index allowed by the kernel
  48                 configuration.
  49
  50                 offline: cpus that are not online because they have been
  51                 HOTPLUGGED off or exceed the limit of cpus allowed by the
  52                 kernel configuration (kernel_max above).
  53
  54                 online: cpus that are online and being scheduled.
  55
  56                 possible: cpus that have been allocated resources and can be
  57                 brought online if they are present.
  58
  59                 present: cpus that have been identified as being present in
  60                 the system.
  61
  62                 See Documentation/cputopology.txt for more information.
  63
  64
  65 What:           /sys/devices/system/cpu/probe
  66                 /sys/devices/system/cpu/release
  67 Date:           November 2009
  68 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
  69 Description:    Dynamic addition and removal of CPU's.  This is not hotplug
  70                 removal, this is meant complete removal/addition of the CPU
  71                 from the system.
  72
  73                 probe: writes to this file will dynamically add a CPU to the
  74                 system.  Information written to the file to add CPU's is
  75                 architecture specific.
  76
  77                 release: writes to this file dynamically remove a CPU from
  78                 the system.  Information writtento the file to remove CPU's
  79                 is architecture specific.
  80
  81 What:           /sys/devices/system/cpu/cpu#/node
  82 Date:           October 2009
  83 Contact:        Linux memory management mailing list <linux-mm@kvack.org>
  84 Description:    Discover NUMA node a CPU belongs to
  85
  86                 When CONFIG_NUMA is enabled, a symbolic link that points
  87                 to the corresponding NUMA node directory.
  88
  89                 For example, the following symlink is created for cpu42
  90                 in NUMA node 2:
  91
  92                 /sys/devices/system/cpu/cpu42/node2 -> ../../node/node2
  93
  94
  95 What:           /sys/devices/system/cpu/cpu#/node
  96 Date:           October 2009
  97 Contact:        Linux memory management mailing list <linux-mm@kvack.org>
  98 Description:    Discover NUMA node a CPU belongs to
  99
 100                 When CONFIG_NUMA is enabled, a symbolic link that points
 101                 to the corresponding NUMA node directory.
 102
 103                 For example, the following symlink is created for cpu42
 104                 in NUMA node 2:
 105
 106                 /sys/devices/system/cpu/cpu42/node2 -> ../../node/node2
 107
 108
 109 What:           /sys/devices/system/cpu/cpu#/topology/core_id
 110                 /sys/devices/system/cpu/cpu#/topology/core_siblings
 111                 /sys/devices/system/cpu/cpu#/topology/core_siblings_list
 112                 /sys/devices/system/cpu/cpu#/topology/physical_package_id
 113                 /sys/devices/system/cpu/cpu#/topology/thread_siblings
 114                 /sys/devices/system/cpu/cpu#/topology/thread_siblings_list
 115 Date:           December 2008
 116 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
 117 Description:    CPU topology files that describe a logical CPU's relationship
 118                 to other cores and threads in the same physical package.
 119
 120                 One cpu# directory is created per logical CPU in the system,
 121                 e.g. /sys/devices/system/cpu/cpu42/.
 122
 123                 Briefly, the files above are:
 124
 125                 core_id: the CPU core ID of cpu#. Typically it is the
 126                 hardware platform's identifier (rather than the kernel's).
 127                 The actual value is architecture and platform dependent.
 128
 129                 core_siblings: internal kernel map of cpu#'s hardware threads
 130                 within the same physical_package_id.
 131
 132                 core_siblings_list: human-readable list of the logical CPU
 133                 numbers within the same physical_package_id as cpu#.
 134
 135                 physical_package_id: physical package id of cpu#. Typically
 136                 corresponds to a physical socket number, but the actual value
 137                 is architecture and platform dependent.
 138
 139                 thread_siblings: internel kernel map of cpu#'s hardware
 140                 threads within the same core as cpu#
 141
 142                 thread_siblings_list: human-readable list of cpu#'s hardware
 143                 threads within the same core as cpu#
 144
 145                 See Documentation/cputopology.txt for more information.
 146
 147
 148 What:           /sys/devices/system/cpu/cpuidle/current_driver
 149                 /sys/devices/system/cpu/cpuidle/current_governer_ro
 150 Date:           September 2007
 151 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
 152 Description:    Discover cpuidle policy and mechanism
 153
 154                 Various CPUs today support multiple idle levels that are
 155                 differentiated by varying exit latencies and power
 156                 consumption during idle.
 157
 158                 Idle policy (governor) is differentiated from idle mechanism
 159                 (driver)
 160
 161                 current_driver: displays current idle mechanism
 162
 163                 current_governor_ro: displays current idle policy
 164
 165                 See files in Documentation/cpuidle/ for more information.
 166
 167
 168 What:           /sys/devices/system/cpu/cpu#/cpufreq/*
 169 Date:           pre-git history
 170 Contact:        cpufreq@vger.kernel.org
 171 Description:    Discover and change clock speed of CPUs
 172
 173                 Clock scaling allows you to change the clock speed of the
 174                 CPUs on the fly. This is a nice method to save battery
 175                 power, because the lower the clock speed, the less power
 176                 the CPU consumes.
 177
 178                 There are many knobs to tweak in this directory.
 179
 180                 See files in Documentation/cpu-freq/ for more information.
 181
 182                 In particular, read Documentation/cpu-freq/user-guide.txt
 183                 to learn how to control the knobs.
 184
 185
 186 What:           /sys/devices/system/cpu/cpu*/cache/index3/cache_disable_{0,1}
 187 Date:           August 2008
 188 KernelVersion:  2.6.27
 189 Contact:        discuss@x86-64.org
 190 Description:    Disable L3 cache indices
 191
 192                 These files exist in every CPU's cache/index3 directory. Each
 193                 cache_disable_{0,1} file corresponds to one disable slot which
 194                 can be used to disable a cache index. Reading from these files
 195                 on a processor with this functionality will return the currently
 196                 disabled index for that node. There is one L3 structure per
 197                 node, or per internal node on MCM machines. Writing a valid
 198                 index to one of these files will cause the specificed cache
 199                 index to be disabled.
 200
 201                 All AMD processors with L3 caches provide this functionality.
 202                 For details, see BKDGs at
 203                 http://developer.amd.com/documentation/guides/Pages/default.aspx