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/kernel_max
  13                 /sys/devices/system/cpu/offline
  14                 /sys/devices/system/cpu/online
  15                 /sys/devices/system/cpu/possible
  16                 /sys/devices/system/cpu/present
  17 Date:           December 2008
  18 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
  19 Description:    CPU topology files that describe kernel limits related to
  20                 hotplug. Briefly:
  21
  22                 kernel_max: the maximum cpu index allowed by the kernel
  23                 configuration.
  24
  25                 offline: cpus that are not online because they have been
  26                 HOTPLUGGED off or exceed the limit of cpus allowed by the
  27                 kernel configuration (kernel_max above).
  28
  29                 online: cpus that are online and being scheduled.
  30
  31                 possible: cpus that have been allocated resources and can be
  32                 brought online if they are present.
  33
  34                 present: cpus that have been identified as being present in
  35                 the system.
  36
  37                 See Documentation/admin-guide/cputopology.rst for more information.
  38
  39
  40 What:           /sys/devices/system/cpu/probe
  41                 /sys/devices/system/cpu/release
  42 Date:           November 2009
  43 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
  44 Description:    Dynamic addition and removal of CPU's.  This is not hotplug
  45                 removal, this is meant complete removal/addition of the CPU
  46                 from the system.
  47
  48                 probe: writes to this file will dynamically add a CPU to the
  49                 system.  Information written to the file to add CPU's is
  50                 architecture specific.
  51
  52                 release: writes to this file dynamically remove a CPU from
  53                 the system.  Information written to the file to remove CPU's
  54                 is architecture specific.
  55
  56 What:           /sys/devices/system/cpu/cpu#/node
  57 Date:           October 2009
  58 Contact:        Linux memory management mailing list <linux-mm@kvack.org>
  59 Description:    Discover NUMA node a CPU belongs to
  60
  61                 When CONFIG_NUMA is enabled, a symbolic link that points
  62                 to the corresponding NUMA node directory.
  63
  64                 For example, the following symlink is created for cpu42
  65                 in NUMA node 2:
  66
  67                 /sys/devices/system/cpu/cpu42/node2 -> ../../node/node2
  68
  69
  70 What:           /sys/devices/system/cpu/cpu#/topology/core_id
  71                 /sys/devices/system/cpu/cpu#/topology/core_siblings
  72                 /sys/devices/system/cpu/cpu#/topology/core_siblings_list
  73                 /sys/devices/system/cpu/cpu#/topology/physical_package_id
  74                 /sys/devices/system/cpu/cpu#/topology/thread_siblings
  75                 /sys/devices/system/cpu/cpu#/topology/thread_siblings_list
  76 Date:           December 2008
  77 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
  78 Description:    CPU topology files that describe a logical CPU's relationship
  79                 to other cores and threads in the same physical package.
  80
  81                 One cpu# directory is created per logical CPU in the system,
  82                 e.g. /sys/devices/system/cpu/cpu42/.
  83
  84                 Briefly, the files above are:
  85
  86                 core_id: the CPU core ID of cpu#. Typically it is the
  87                 hardware platform's identifier (rather than the kernel's).
  88                 The actual value is architecture and platform dependent.
  89
  90                 core_siblings: internal kernel map of cpu#'s hardware threads
  91                 within the same physical_package_id.
  92
  93                 core_siblings_list: human-readable list of the logical CPU
  94                 numbers within the same physical_package_id as cpu#.
  95
  96                 physical_package_id: physical package id of cpu#. Typically
  97                 corresponds to a physical socket number, but the actual value
  98                 is architecture and platform dependent.
  99
 100                 thread_siblings: internal kernel map of cpu#'s hardware
 101                 threads within the same core as cpu#
 102
 103                 thread_siblings_list: human-readable list of cpu#'s hardware
 104                 threads within the same core as cpu#
 105
 106                 See Documentation/admin-guide/cputopology.rst for more information.
 107
 108
 109 What:           /sys/devices/system/cpu/cpuidle/available_governors
 110                 /sys/devices/system/cpu/cpuidle/current_driver
 111                 /sys/devices/system/cpu/cpuidle/current_governor
 112                 /sys/devices/system/cpu/cpuidle/current_governer_ro
 113 Date:           September 2007
 114 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
 115 Description:    Discover cpuidle policy and mechanism
 116
 117                 Various CPUs today support multiple idle levels that are
 118                 differentiated by varying exit latencies and power
 119                 consumption during idle.
 120
 121                 Idle policy (governor) is differentiated from idle mechanism
 122                 (driver).
 123
 124                 available_governors: (RO) displays a space separated list of
 125                 available governors.
 126
 127                 current_driver: (RO) displays current idle mechanism.
 128
 129                 current_governor: (RW) displays current idle policy. Users can
 130                 switch the governor at runtime by writing to this file.
 131
 132                 current_governor_ro: (RO) displays current idle policy.
 133
 134                 See Documentation/admin-guide/pm/cpuidle.rst and
 135                 Documentation/driver-api/pm/cpuidle.rst for more information.
 136
 137
 138 What:           /sys/devices/system/cpu/cpuX/cpuidle/stateN/name
 139                 /sys/devices/system/cpu/cpuX/cpuidle/stateN/latency
 140                 /sys/devices/system/cpu/cpuX/cpuidle/stateN/power
 141                 /sys/devices/system/cpu/cpuX/cpuidle/stateN/time
 142                 /sys/devices/system/cpu/cpuX/cpuidle/stateN/usage
 143                 /sys/devices/system/cpu/cpuX/cpuidle/stateN/above
 144                 /sys/devices/system/cpu/cpuX/cpuidle/stateN/below
 145 Date:           September 2007
 146 KernelVersion:  v2.6.24
 147 Contact:        Linux power management list <linux-pm@vger.kernel.org>
 148 Description:
 149                 The directory /sys/devices/system/cpu/cpuX/cpuidle contains per
 150                 logical CPU specific cpuidle information for each online cpu X.
 151                 The processor idle states which are available for use have the
 152                 following attributes:
 153
 154                 ======== ==== =================================================
 155                 name:    (RO) Name of the idle state (string).
 156
 157                 latency: (RO) The latency to exit out of this idle state (in
 158                               microseconds).
 159
 160                 power:   (RO) The power consumed while in this idle state (in
 161                               milliwatts).
 162
 163                 time:    (RO) The total time spent in this idle state
 164                               (in microseconds).
 165
 166                 usage:   (RO) Number of times this state was entered (a count).
 167
 168                 above:   (RO) Number of times this state was entered, but the
 169                               observed CPU idle duration was too short for it
 170                               (a count).
 171
 172                 below:   (RO) Number of times this state was entered, but the
 173                               observed CPU idle duration was too long for it
 174                               (a count).
 175                 ======== ==== =================================================
 176
 177 What:           /sys/devices/system/cpu/cpuX/cpuidle/stateN/desc
 178 Date:           February 2008
 179 KernelVersion:  v2.6.25
 180 Contact:        Linux power management list <linux-pm@vger.kernel.org>
 181 Description:
 182                 (RO) A small description about the idle state (string).
 183
 184
 185 What:           /sys/devices/system/cpu/cpuX/cpuidle/stateN/disable
 186 Date:           March 2012
 187 KernelVersion:  v3.10
 188 Contact:        Linux power management list <linux-pm@vger.kernel.org>
 189 Description:
 190                 (RW) Option to disable this idle state (bool). The behavior and
 191                 the effect of the disable variable depends on the implementation
 192                 of a particular governor. In the ladder governor, for example,
 193                 it is not coherent, i.e. if one is disabling a light state, then
 194                 all deeper states are disabled as well, but the disable variable
 195                 does not reflect it. Likewise, if one enables a deep state but a
 196                 lighter state still is disabled, then this has no effect.
 197
 198 What:           /sys/devices/system/cpu/cpuX/cpuidle/stateN/default_status
 199 Date:           December 2019
 200 KernelVersion:  v5.6
 201 Contact:        Linux power management list <linux-pm@vger.kernel.org>
 202 Description:
 203                 (RO) The default status of this state, "enabled" or "disabled".
 204
 205 What:           /sys/devices/system/cpu/cpuX/cpuidle/stateN/residency
 206 Date:           March 2014
 207 KernelVersion:  v3.15
 208 Contact:        Linux power management list <linux-pm@vger.kernel.org>
 209 Description:
 210                 (RO) Display the target residency i.e. the minimum amount of
 211                 time (in microseconds) this cpu should spend in this idle state
 212                 to make the transition worth the effort.
 213
 214 What:           /sys/devices/system/cpu/cpuX/cpuidle/stateN/s2idle/
 215 Date:           March 2018
 216 KernelVersion:  v4.17
 217 Contact:        Linux power management list <linux-pm@vger.kernel.org>
 218 Description:
 219                 Idle state usage statistics related to suspend-to-idle.
 220
 221                 This attribute group is only present for states that can be
 222                 used in suspend-to-idle with suspended timekeeping.
 223
 224 What:           /sys/devices/system/cpu/cpuX/cpuidle/stateN/s2idle/time
 225 Date:           March 2018
 226 KernelVersion:  v4.17
 227 Contact:        Linux power management list <linux-pm@vger.kernel.org>
 228 Description:
 229                 Total time spent by the CPU in suspend-to-idle (with scheduler
 230                 tick suspended) after requesting this state.
 231
 232 What:           /sys/devices/system/cpu/cpuX/cpuidle/stateN/s2idle/usage
 233 Date:           March 2018
 234 KernelVersion:  v4.17
 235 Contact:        Linux power management list <linux-pm@vger.kernel.org>
 236 Description:
 237                 Total number of times this state has been requested by the CPU
 238                 while entering suspend-to-idle.
 239
 240 What:           /sys/devices/system/cpu/cpu#/cpufreq/*
 241 Date:           pre-git history
 242 Contact:        linux-pm@vger.kernel.org
 243 Description:    Discover and change clock speed of CPUs
 244
 245                 Clock scaling allows you to change the clock speed of the
 246                 CPUs on the fly. This is a nice method to save battery
 247                 power, because the lower the clock speed, the less power
 248                 the CPU consumes.
 249
 250                 There are many knobs to tweak in this directory.
 251
 252                 See files in Documentation/cpu-freq/ for more information.
 253
 254
 255 What:           /sys/devices/system/cpu/cpu#/cpufreq/freqdomain_cpus
 256 Date:           June 2013
 257 Contact:        linux-pm@vger.kernel.org
 258 Description:    Discover CPUs in the same CPU frequency coordination domain
 259
 260                 freqdomain_cpus is the list of CPUs (online+offline) that share
 261                 the same clock/freq domain (possibly at the hardware level).
 262                 That information may be hidden from the cpufreq core and the
 263                 value of related_cpus may be different from freqdomain_cpus. This
 264                 attribute is useful for user space DVFS controllers to get better
 265                 power/performance results for platforms using acpi-cpufreq.
 266
 267                 This file is only present if the acpi-cpufreq or the cppc-cpufreq
 268                 drivers are in use.
 269
 270
 271 What:           /sys/devices/system/cpu/cpu*/cache/index3/cache_disable_{0,1}
 272 Date:           August 2008
 273 KernelVersion:  2.6.27
 274 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
 275 Description:    Disable L3 cache indices
 276
 277                 These files exist in every CPU's cache/index3 directory. Each
 278                 cache_disable_{0,1} file corresponds to one disable slot which
 279                 can be used to disable a cache index. Reading from these files
 280                 on a processor with this functionality will return the currently
 281                 disabled index for that node. There is one L3 structure per
 282                 node, or per internal node on MCM machines. Writing a valid
 283                 index to one of these files will cause the specified cache
 284                 index to be disabled.
 285
 286                 All AMD processors with L3 caches provide this functionality.
 287                 For details, see BKDGs at
 288                 https://www.amd.com/en/support/tech-docs?keyword=bios+kernel
 289
 290
 291 What:           /sys/devices/system/cpu/cpufreq/boost
 292 Date:           August 2012
 293 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
 294 Description:    Processor frequency boosting control
 295
 296                 This switch controls the boost setting for the whole system.
 297                 Boosting allows the CPU and the firmware to run at a frequency
 298                 beyond it's nominal limit.
 299
 300                 More details can be found in
 301                 Documentation/admin-guide/pm/cpufreq.rst
 302
 303
 304 What:           /sys/devices/system/cpu/cpu#/crash_notes
 305                 /sys/devices/system/cpu/cpu#/crash_notes_size
 306 Date:           April 2013
 307 Contact:        kexec@lists.infradead.org
 308 Description:    address and size of the percpu note.
 309
 310                 crash_notes: the physical address of the memory that holds the
 311                 note of cpu#.
 312
 313                 crash_notes_size: size of the note of cpu#.
 314
 315
 316 What:           /sys/devices/system/cpu/intel_pstate/max_perf_pct
 317                 /sys/devices/system/cpu/intel_pstate/min_perf_pct
 318                 /sys/devices/system/cpu/intel_pstate/no_turbo
 319 Date:           February 2013
 320 Contact:        linux-pm@vger.kernel.org
 321 Description:    Parameters for the Intel P-state driver
 322
 323                 Logic for selecting the current P-state in Intel
 324                 Sandybridge+ processors. The three knobs control
 325                 limits for the P-state that will be requested by the
 326                 driver.
 327
 328                 max_perf_pct: limits the maximum P state that will be requested by
 329                 the driver stated as a percentage of the available performance.
 330
 331                 min_perf_pct: limits the minimum P state that will be requested by
 332                 the driver stated as a percentage of the available performance.
 333
 334                 no_turbo: limits the driver to selecting P states below the turbo
 335                 frequency range.
 336
 337                 More details can be found in
 338                 Documentation/admin-guide/pm/intel_pstate.rst
 339
 340 What:           /sys/devices/system/cpu/cpu*/cache/index*/<set_of_attributes_mentioned_below>
 341 Date:           July 2014(documented, existed before August 2008)
 342 Contact:        Sudeep Holla <sudeep.holla@arm.com>
 343                 Linux kernel mailing list <linux-kernel@vger.kernel.org>
 344 Description:    Parameters for the CPU cache attributes
 345
 346                 allocation_policy:
 347                         - WriteAllocate:
 348                                         allocate a memory location to a cache line
 349                                         on a cache miss because of a write
 350                         - ReadAllocate:
 351                                         allocate a memory location to a cache line
 352                                         on a cache miss because of a read
 353                         - ReadWriteAllocate:
 354                                         both writeallocate and readallocate
 355
 356                 attributes:
 357                             LEGACY used only on IA64 and is same as write_policy
 358
 359                 coherency_line_size:
 360                                      the minimum amount of data in bytes that gets
 361                                      transferred from memory to cache
 362
 363                 level:
 364                         the cache hierarchy in the multi-level cache configuration
 365
 366                 number_of_sets:
 367                                 total number of sets in the cache, a set is a
 368                                 collection of cache lines with the same cache index
 369
 370                 physical_line_partition:
 371                                 number of physical cache line per cache tag
 372
 373                 shared_cpu_list:
 374                                 the list of logical cpus sharing the cache
 375
 376                 shared_cpu_map:
 377                                 logical cpu mask containing the list of cpus sharing
 378                                 the cache
 379
 380                 size:
 381                         the total cache size in kB
 382
 383                 type:
 384                         - Instruction: cache that only holds instructions
 385                         - Data: cache that only caches data
 386                         - Unified: cache that holds both data and instructions
 387
 388                 ways_of_associativity:
 389                         degree of freedom in placing a particular block
 390                         of memory in the cache
 391
 392                 write_policy:
 393                         - WriteThrough:
 394                                         data is written to both the cache line
 395                                         and to the block in the lower-level memory
 396                         - WriteBack:
 397                                      data is written only to the cache line and
 398                                      the modified cache line is written to main
 399                                      memory only when it is replaced
 400
 401
 402 What:           /sys/devices/system/cpu/cpu*/cache/index*/id
 403 Date:           September 2016
 404 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
 405 Description:    Cache id
 406
 407                 The id provides a unique number for a specific instance of
 408                 a cache of a particular type. E.g. there may be a level
 409                 3 unified cache on each socket in a server and we may
 410                 assign them ids 0, 1, 2, ...
 411
 412                 Note that id value can be non-contiguous. E.g. level 1
 413                 caches typically exist per core, but there may not be a
 414                 power of two cores on a socket, so these caches may be
 415                 numbered 0, 1, 2, 3, 4, 5, 8, 9, 10, ...
 416
 417 What:           /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats
 418                 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/turbo_stat
 419                 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/sub_turbo_stat
 420                 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/unthrottle
 421                 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/powercap
 422                 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/overtemp
 423                 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/supply_fault
 424                 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/overcurrent
 425                 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/occ_reset
 426 Date:           March 2016
 427 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
 428                 Linux for PowerPC mailing list <linuxppc-dev@ozlabs.org>
 429 Description:    POWERNV CPUFreq driver's frequency throttle stats directory and
 430                 attributes
 431
 432                 'cpuX/cpufreq/throttle_stats' directory contains the CPU frequency
 433                 throttle stat attributes for the chip. The throttle stats of a cpu
 434                 is common across all the cpus belonging to a chip. Below are the
 435                 throttle attributes exported in the 'throttle_stats' directory:
 436
 437                 - turbo_stat : This file gives the total number of times the max
 438                   frequency is throttled to lower frequency in turbo (at and above
 439                   nominal frequency) range of frequencies.
 440
 441                 - sub_turbo_stat : This file gives the total number of times the
 442                   max frequency is throttled to lower frequency in sub-turbo(below
 443                   nominal frequency) range of frequencies.
 444
 445                 - unthrottle : This file gives the total number of times the max
 446                   frequency is unthrottled after being throttled.
 447
 448                 - powercap : This file gives the total number of times the max
 449                   frequency is throttled due to 'Power Capping'.
 450
 451                 - overtemp : This file gives the total number of times the max
 452                   frequency is throttled due to 'CPU Over Temperature'.
 453
 454                 - supply_fault : This file gives the total number of times the
 455                   max frequency is throttled due to 'Power Supply Failure'.
 456
 457                 - overcurrent : This file gives the total number of times the
 458                   max frequency is throttled due to 'Overcurrent'.
 459
 460                 - occ_reset : This file gives the total number of times the max
 461                   frequency is throttled due to 'OCC Reset'.
 462
 463                 The sysfs attributes representing different throttle reasons like
 464                 powercap, overtemp, supply_fault, overcurrent and occ_reset map to
 465                 the reasons provided by OCC firmware for throttling the frequency.
 466
 467 What:           /sys/devices/system/cpu/cpufreq/policyX/throttle_stats
 468                 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/turbo_stat
 469                 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/sub_turbo_stat
 470                 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/unthrottle
 471                 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/powercap
 472                 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/overtemp
 473                 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/supply_fault
 474                 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/overcurrent
 475                 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/occ_reset
 476 Date:           March 2016
 477 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
 478                 Linux for PowerPC mailing list <linuxppc-dev@ozlabs.org>
 479 Description:    POWERNV CPUFreq driver's frequency throttle stats directory and
 480                 attributes
 481
 482                 'policyX/throttle_stats' directory and all the attributes are same as
 483                 the /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats directory and
 484                 attributes which give the frequency throttle information of the chip.
 485
 486 What:           /sys/devices/system/cpu/cpuX/regs/
 487                 /sys/devices/system/cpu/cpuX/regs/identification/
 488                 /sys/devices/system/cpu/cpuX/regs/identification/midr_el1
 489                 /sys/devices/system/cpu/cpuX/regs/identification/revidr_el1
 490 Date:           June 2016
 491 Contact:        Linux ARM Kernel Mailing list <linux-arm-kernel@lists.infradead.org>
 492 Description:    AArch64 CPU registers
 493
 494                 'identification' directory exposes the CPU ID registers for
 495                 identifying model and revision of the CPU.
 496
 497 What:           /sys/devices/system/cpu/aarch32_el0
 498 Date:           May 2021
 499 Contact:        Linux ARM Kernel Mailing list <linux-arm-kernel@lists.infradead.org>
 500 Description:    Identifies the subset of CPUs in the system that can execute
 501                 AArch32 (32-bit ARM) applications. If present, the same format as
 502                 /sys/devices/system/cpu/{offline,online,possible,present} is used.
 503                 If absent, then all or none of the CPUs can execute AArch32
 504                 applications and execve() will behave accordingly.
 505
 506 What:           /sys/devices/system/cpu/cpu#/cpu_capacity
 507 Date:           December 2016
 508 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
 509 Description:    information about CPUs heterogeneity.
 510
 511                 cpu_capacity: capacity of cpu#.
 512
 513 What:           /sys/devices/system/cpu/vulnerabilities
 514                 /sys/devices/system/cpu/vulnerabilities/meltdown
 515                 /sys/devices/system/cpu/vulnerabilities/spectre_v1
 516                 /sys/devices/system/cpu/vulnerabilities/spectre_v2
 517                 /sys/devices/system/cpu/vulnerabilities/spec_store_bypass
 518                 /sys/devices/system/cpu/vulnerabilities/l1tf
 519                 /sys/devices/system/cpu/vulnerabilities/mds
 520                 /sys/devices/system/cpu/vulnerabilities/srbds
 521                 /sys/devices/system/cpu/vulnerabilities/tsx_async_abort
 522                 /sys/devices/system/cpu/vulnerabilities/itlb_multihit
 523 Date:           January 2018
 524 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
 525 Description:    Information about CPU vulnerabilities
 526
 527                 The files are named after the code names of CPU
 528                 vulnerabilities. The output of those files reflects the
 529                 state of the CPUs in the system. Possible output values:
 530
 531                 ================  ==============================================
 532                 "Not affected"    CPU is not affected by the vulnerability
 533                 "Vulnerable"      CPU is affected and no mitigation in effect
 534                 "Mitigation: $M"  CPU is affected and mitigation $M is in effect
 535                 ================  ==============================================
 536
 537                 See also: Documentation/admin-guide/hw-vuln/index.rst
 538
 539 What:           /sys/devices/system/cpu/smt
 540                 /sys/devices/system/cpu/smt/active
 541                 /sys/devices/system/cpu/smt/control
 542 Date:           June 2018
 543 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
 544 Description:    Control Symmetric Multi Threading (SMT)
 545
 546                 active:  Tells whether SMT is active (enabled and siblings online)
 547
 548                 control: Read/write interface to control SMT. Possible
 549                          values:
 550
 551                          ================ =========================================
 552                          "on"             SMT is enabled
 553                          "off"            SMT is disabled
 554                          "forceoff"       SMT is force disabled. Cannot be changed.
 555                          "notsupported"   SMT is not supported by the CPU
 556                          "notimplemented" SMT runtime toggling is not
 557                                           implemented for the architecture
 558                          ================ =========================================
 559
 560                          If control status is "forceoff" or "notsupported" writes
 561                          are rejected.
 562
 563 What:           /sys/devices/system/cpu/cpu#/power/energy_perf_bias
 564 Date:           March 2019
 565 Contact:        linux-pm@vger.kernel.org
 566 Description:    Intel Energy and Performance Bias Hint (EPB)
 567
 568                 EPB for the given CPU in a sliding scale 0 - 15, where a value
 569                 of 0 corresponds to a hint preference for highest performance
 570                 and a value of 15 corresponds to the maximum energy savings.
 571
 572                 In order to change the EPB value for the CPU, write either
 573                 a number in the 0 - 15 sliding scale above, or one of the
 574                 strings: "performance", "balance-performance", "normal",
 575                 "balance-power", "power" (that represent values reflected by
 576                 their meaning), to this attribute.
 577
 578                 This attribute is present for all online CPUs supporting the
 579                 Intel EPB feature.
 580
 581 What:           /sys/devices/system/cpu/umwait_control
 582                 /sys/devices/system/cpu/umwait_control/enable_c02
 583                 /sys/devices/system/cpu/umwait_control/max_time
 584 Date:           May 2019
 585 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
 586 Description:    Umwait control
 587
 588                 enable_c02: Read/write interface to control umwait C0.2 state
 589                         Read returns C0.2 state status:
 590                                 0: C0.2 is disabled
 591                                 1: C0.2 is enabled
 592
 593                         Write 'y' or '1'  or 'on' to enable C0.2 state.
 594                         Write 'n' or '0'  or 'off' to disable C0.2 state.
 595
 596                         The interface is case insensitive.
 597
 598                 max_time: Read/write interface to control umwait maximum time
 599                           in TSC-quanta that the CPU can reside in either C0.1
 600                           or C0.2 state. The time is an unsigned 32-bit number.
 601                           Note that a value of zero means there is no limit.
 602                           Low order two bits must be zero.
 603
 604 What:           /sys/devices/system/cpu/svm
 605 Date:           August 2019
 606 Contact:        Linux kernel mailing list <linux-kernel@vger.kernel.org>
 607                 Linux for PowerPC mailing list <linuxppc-dev@ozlabs.org>
 608 Description:    Secure Virtual Machine
 609
 610                 If 1, it means the system is using the Protected Execution
 611                 Facility in POWER9 and newer processors. i.e., it is a Secure
 612                 Virtual Machine.
 613
 614 What:           /sys/devices/system/cpu/cpuX/purr
 615 Date:           Apr 2005
 616 Contact:        Linux for PowerPC mailing list <linuxppc-dev@ozlabs.org>
 617 Description:    PURR ticks for this CPU since the system boot.
 618
 619                 The Processor Utilization Resources Register (PURR) is
 620                 a 64-bit counter which provides an estimate of the
 621                 resources used by the CPU thread. The contents of this
 622                 register increases monotonically. This sysfs interface
 623                 exposes the number of PURR ticks for cpuX.
 624
 625 What:           /sys/devices/system/cpu/cpuX/spurr
 626 Date:           Dec 2006
 627 Contact:        Linux for PowerPC mailing list <linuxppc-dev@ozlabs.org>
 628 Description:    SPURR ticks for this CPU since the system boot.
 629
 630                 The Scaled Processor Utilization Resources Register
 631                 (SPURR) is a 64-bit counter that provides a frequency
 632                 invariant estimate of the resources used by the CPU
 633                 thread. The contents of this register increases
 634                 monotonically. This sysfs interface exposes the number
 635                 of SPURR ticks for cpuX.
 636
 637 What:           /sys/devices/system/cpu/cpuX/idle_purr
 638 Date:           Apr 2020
 639 Contact:        Linux for PowerPC mailing list <linuxppc-dev@ozlabs.org>
 640 Description:    PURR ticks for cpuX when it was idle.
 641
 642                 This sysfs interface exposes the number of PURR ticks
 643                 for cpuX when it was idle.
 644
 645 What:           /sys/devices/system/cpu/cpuX/idle_spurr
 646 Date:           Apr 2020
 647 Contact:        Linux for PowerPC mailing list <linuxppc-dev@ozlabs.org>
 648 Description:    SPURR ticks for cpuX when it was idle.
 649
 650                 This sysfs interface exposes the number of SPURR ticks
 651                 for cpuX when it was idle.
 652
 653 What:           /sys/devices/system/cpu/cpuX/mte_tcf_preferred
 654 Date:           July 2021
 655 Contact:        Linux ARM Kernel Mailing list <linux-arm-kernel@lists.infradead.org>
 656 Description:    Preferred MTE tag checking mode
 657
 658                 When a user program specifies more than one MTE tag checking
 659                 mode, this sysfs node is used to specify which mode should
 660                 be preferred when scheduling a task on that CPU. Possible
 661                 values:
 662
 663                 ================  ==============================================
 664                 "sync"            Prefer synchronous mode
 665                 "async"           Prefer asynchronous mode
 666                 ================  ==============================================
 667
 668                 See also: Documentation/arm64/memory-tagging-extension.rst