1What: /sys/devices/system/cpu/ 2Date: pre-git history 3Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 4Description: 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 12What: /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 17Date: December 2008 18Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 19Description: 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 40What: /sys/devices/system/cpu/probe 41 /sys/devices/system/cpu/release 42Date: November 2009 43Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 44Description: 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 writtento the file to remove CPU's 54 is architecture specific. 55 56What: /sys/devices/system/cpu/cpu#/node 57Date: October 2009 58Contact: Linux memory management mailing list <linux-mm@kvack.org> 59Description: 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 70What: /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 76Date: December 2008 77Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 78Description: 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: internel 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 109What: /sys/devices/system/cpu/cpuidle/current_driver 110 /sys/devices/system/cpu/cpuidle/current_governer_ro 111 /sys/devices/system/cpu/cpuidle/available_governors 112 /sys/devices/system/cpu/cpuidle/current_governor 113Date: September 2007 114Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 115Description: 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 current_driver: (RO) displays current idle mechanism 125 126 current_governor_ro: (RO) displays current idle policy 127 128 With the cpuidle_sysfs_switch boot option enabled (meant for 129 developer testing), the following three attributes are visible 130 instead: 131 132 current_driver: same as described above 133 134 available_governors: (RO) displays a space separated list of 135 available governors 136 137 current_governor: (RW) displays current idle policy. Users can 138 switch the governor at runtime by writing to this file. 139 140 See Documentation/admin-guide/pm/cpuidle.rst and 141 Documentation/driver-api/pm/cpuidle.rst for more information. 142 143 144What: /sys/devices/system/cpu/cpuX/cpuidle/stateN/name 145 /sys/devices/system/cpu/cpuX/cpuidle/stateN/latency 146 /sys/devices/system/cpu/cpuX/cpuidle/stateN/power 147 /sys/devices/system/cpu/cpuX/cpuidle/stateN/time 148 /sys/devices/system/cpu/cpuX/cpuidle/stateN/usage 149 /sys/devices/system/cpu/cpuX/cpuidle/stateN/above 150 /sys/devices/system/cpu/cpuX/cpuidle/stateN/below 151Date: September 2007 152KernelVersion: v2.6.24 153Contact: Linux power management list <linux-pm@vger.kernel.org> 154Description: 155 The directory /sys/devices/system/cpu/cpuX/cpuidle contains per 156 logical CPU specific cpuidle information for each online cpu X. 157 The processor idle states which are available for use have the 158 following attributes: 159 160 name: (RO) Name of the idle state (string). 161 162 latency: (RO) The latency to exit out of this idle state (in 163 microseconds). 164 165 power: (RO) The power consumed while in this idle state (in 166 milliwatts). 167 168 time: (RO) The total time spent in this idle state (in microseconds). 169 170 usage: (RO) Number of times this state was entered (a count). 171 172 above: (RO) Number of times this state was entered, but the 173 observed CPU idle duration was too short for it (a count). 174 175 below: (RO) Number of times this state was entered, but the 176 observed CPU idle duration was too long for it (a count). 177 178What: /sys/devices/system/cpu/cpuX/cpuidle/stateN/desc 179Date: February 2008 180KernelVersion: v2.6.25 181Contact: Linux power management list <linux-pm@vger.kernel.org> 182Description: 183 (RO) A small description about the idle state (string). 184 185 186What: /sys/devices/system/cpu/cpuX/cpuidle/stateN/disable 187Date: March 2012 188KernelVersion: v3.10 189Contact: Linux power management list <linux-pm@vger.kernel.org> 190Description: 191 (RW) Option to disable this idle state (bool). The behavior and 192 the effect of the disable variable depends on the implementation 193 of a particular governor. In the ladder governor, for example, 194 it is not coherent, i.e. if one is disabling a light state, then 195 all deeper states are disabled as well, but the disable variable 196 does not reflect it. Likewise, if one enables a deep state but a 197 lighter state still is disabled, then this has no effect. 198 199 200What: /sys/devices/system/cpu/cpuX/cpuidle/stateN/residency 201Date: March 2014 202KernelVersion: v3.15 203Contact: Linux power management list <linux-pm@vger.kernel.org> 204Description: 205 (RO) Display the target residency i.e. the minimum amount of 206 time (in microseconds) this cpu should spend in this idle state 207 to make the transition worth the effort. 208 209What: /sys/devices/system/cpu/cpuX/cpuidle/stateN/s2idle/ 210Date: March 2018 211KernelVersion: v4.17 212Contact: Linux power management list <linux-pm@vger.kernel.org> 213Description: 214 Idle state usage statistics related to suspend-to-idle. 215 216 This attribute group is only present for states that can be 217 used in suspend-to-idle with suspended timekeeping. 218 219What: /sys/devices/system/cpu/cpuX/cpuidle/stateN/s2idle/time 220Date: March 2018 221KernelVersion: v4.17 222Contact: Linux power management list <linux-pm@vger.kernel.org> 223Description: 224 Total time spent by the CPU in suspend-to-idle (with scheduler 225 tick suspended) after requesting this state. 226 227What: /sys/devices/system/cpu/cpuX/cpuidle/stateN/s2idle/usage 228Date: March 2018 229KernelVersion: v4.17 230Contact: Linux power management list <linux-pm@vger.kernel.org> 231Description: 232 Total number of times this state has been requested by the CPU 233 while entering suspend-to-idle. 234 235What: /sys/devices/system/cpu/cpu#/cpufreq/* 236Date: pre-git history 237Contact: linux-pm@vger.kernel.org 238Description: Discover and change clock speed of CPUs 239 240 Clock scaling allows you to change the clock speed of the 241 CPUs on the fly. This is a nice method to save battery 242 power, because the lower the clock speed, the less power 243 the CPU consumes. 244 245 There are many knobs to tweak in this directory. 246 247 See files in Documentation/cpu-freq/ for more information. 248 249 250What: /sys/devices/system/cpu/cpu#/cpufreq/freqdomain_cpus 251Date: June 2013 252Contact: linux-pm@vger.kernel.org 253Description: Discover CPUs in the same CPU frequency coordination domain 254 255 freqdomain_cpus is the list of CPUs (online+offline) that share 256 the same clock/freq domain (possibly at the hardware level). 257 That information may be hidden from the cpufreq core and the 258 value of related_cpus may be different from freqdomain_cpus. This 259 attribute is useful for user space DVFS controllers to get better 260 power/performance results for platforms using acpi-cpufreq. 261 262 This file is only present if the acpi-cpufreq driver is in use. 263 264 265What: /sys/devices/system/cpu/cpu*/cache/index3/cache_disable_{0,1} 266Date: August 2008 267KernelVersion: 2.6.27 268Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 269Description: Disable L3 cache indices 270 271 These files exist in every CPU's cache/index3 directory. Each 272 cache_disable_{0,1} file corresponds to one disable slot which 273 can be used to disable a cache index. Reading from these files 274 on a processor with this functionality will return the currently 275 disabled index for that node. There is one L3 structure per 276 node, or per internal node on MCM machines. Writing a valid 277 index to one of these files will cause the specificed cache 278 index to be disabled. 279 280 All AMD processors with L3 caches provide this functionality. 281 For details, see BKDGs at 282 http://developer.amd.com/documentation/guides/Pages/default.aspx 283 284 285What: /sys/devices/system/cpu/cpufreq/boost 286Date: August 2012 287Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 288Description: Processor frequency boosting control 289 290 This switch controls the boost setting for the whole system. 291 Boosting allows the CPU and the firmware to run at a frequency 292 beyound it's nominal limit. 293 More details can be found in 294 Documentation/admin-guide/pm/cpufreq.rst 295 296 297What: /sys/devices/system/cpu/cpu#/crash_notes 298 /sys/devices/system/cpu/cpu#/crash_notes_size 299Date: April 2013 300Contact: kexec@lists.infradead.org 301Description: address and size of the percpu note. 302 303 crash_notes: the physical address of the memory that holds the 304 note of cpu#. 305 306 crash_notes_size: size of the note of cpu#. 307 308 309What: /sys/devices/system/cpu/intel_pstate/max_perf_pct 310 /sys/devices/system/cpu/intel_pstate/min_perf_pct 311 /sys/devices/system/cpu/intel_pstate/no_turbo 312Date: February 2013 313Contact: linux-pm@vger.kernel.org 314Description: Parameters for the Intel P-state driver 315 316 Logic for selecting the current P-state in Intel 317 Sandybridge+ processors. The three knobs control 318 limits for the P-state that will be requested by the 319 driver. 320 321 max_perf_pct: limits the maximum P state that will be requested by 322 the driver stated as a percentage of the available performance. 323 324 min_perf_pct: limits the minimum P state that will be requested by 325 the driver stated as a percentage of the available performance. 326 327 no_turbo: limits the driver to selecting P states below the turbo 328 frequency range. 329 330 More details can be found in 331 Documentation/admin-guide/pm/intel_pstate.rst 332 333What: /sys/devices/system/cpu/cpu*/cache/index*/<set_of_attributes_mentioned_below> 334Date: July 2014(documented, existed before August 2008) 335Contact: Sudeep Holla <sudeep.holla@arm.com> 336 Linux kernel mailing list <linux-kernel@vger.kernel.org> 337Description: Parameters for the CPU cache attributes 338 339 allocation_policy: 340 - WriteAllocate: allocate a memory location to a cache line 341 on a cache miss because of a write 342 - ReadAllocate: allocate a memory location to a cache line 343 on a cache miss because of a read 344 - ReadWriteAllocate: both writeallocate and readallocate 345 346 attributes: LEGACY used only on IA64 and is same as write_policy 347 348 coherency_line_size: the minimum amount of data in bytes that gets 349 transferred from memory to cache 350 351 level: the cache hierarchy in the multi-level cache configuration 352 353 number_of_sets: total number of sets in the cache, a set is a 354 collection of cache lines with the same cache index 355 356 physical_line_partition: number of physical cache line per cache tag 357 358 shared_cpu_list: the list of logical cpus sharing the cache 359 360 shared_cpu_map: logical cpu mask containing the list of cpus sharing 361 the cache 362 363 size: the total cache size in kB 364 365 type: 366 - Instruction: cache that only holds instructions 367 - Data: cache that only caches data 368 - Unified: cache that holds both data and instructions 369 370 ways_of_associativity: degree of freedom in placing a particular block 371 of memory in the cache 372 373 write_policy: 374 - WriteThrough: data is written to both the cache line 375 and to the block in the lower-level memory 376 - WriteBack: data is written only to the cache line and 377 the modified cache line is written to main 378 memory only when it is replaced 379 380 381What: /sys/devices/system/cpu/cpu*/cache/index*/id 382Date: September 2016 383Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 384Description: Cache id 385 386 The id provides a unique number for a specific instance of 387 a cache of a particular type. E.g. there may be a level 388 3 unified cache on each socket in a server and we may 389 assign them ids 0, 1, 2, ... 390 391 Note that id value can be non-contiguous. E.g. level 1 392 caches typically exist per core, but there may not be a 393 power of two cores on a socket, so these caches may be 394 numbered 0, 1, 2, 3, 4, 5, 8, 9, 10, ... 395 396What: /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats 397 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/turbo_stat 398 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/sub_turbo_stat 399 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/unthrottle 400 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/powercap 401 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/overtemp 402 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/supply_fault 403 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/overcurrent 404 /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats/occ_reset 405Date: March 2016 406Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 407 Linux for PowerPC mailing list <linuxppc-dev@ozlabs.org> 408Description: POWERNV CPUFreq driver's frequency throttle stats directory and 409 attributes 410 411 'cpuX/cpufreq/throttle_stats' directory contains the CPU frequency 412 throttle stat attributes for the chip. The throttle stats of a cpu 413 is common across all the cpus belonging to a chip. Below are the 414 throttle attributes exported in the 'throttle_stats' directory: 415 416 - turbo_stat : This file gives the total number of times the max 417 frequency is throttled to lower frequency in turbo (at and above 418 nominal frequency) range of frequencies. 419 420 - sub_turbo_stat : This file gives the total number of times the 421 max frequency is throttled to lower frequency in sub-turbo(below 422 nominal frequency) range of frequencies. 423 424 - unthrottle : This file gives the total number of times the max 425 frequency is unthrottled after being throttled. 426 427 - powercap : This file gives the total number of times the max 428 frequency is throttled due to 'Power Capping'. 429 430 - overtemp : This file gives the total number of times the max 431 frequency is throttled due to 'CPU Over Temperature'. 432 433 - supply_fault : This file gives the total number of times the 434 max frequency is throttled due to 'Power Supply Failure'. 435 436 - overcurrent : This file gives the total number of times the 437 max frequency is throttled due to 'Overcurrent'. 438 439 - occ_reset : This file gives the total number of times the max 440 frequency is throttled due to 'OCC Reset'. 441 442 The sysfs attributes representing different throttle reasons like 443 powercap, overtemp, supply_fault, overcurrent and occ_reset map to 444 the reasons provided by OCC firmware for throttling the frequency. 445 446What: /sys/devices/system/cpu/cpufreq/policyX/throttle_stats 447 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/turbo_stat 448 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/sub_turbo_stat 449 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/unthrottle 450 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/powercap 451 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/overtemp 452 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/supply_fault 453 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/overcurrent 454 /sys/devices/system/cpu/cpufreq/policyX/throttle_stats/occ_reset 455Date: March 2016 456Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 457 Linux for PowerPC mailing list <linuxppc-dev@ozlabs.org> 458Description: POWERNV CPUFreq driver's frequency throttle stats directory and 459 attributes 460 461 'policyX/throttle_stats' directory and all the attributes are same as 462 the /sys/devices/system/cpu/cpuX/cpufreq/throttle_stats directory and 463 attributes which give the frequency throttle information of the chip. 464 465What: /sys/devices/system/cpu/cpuX/regs/ 466 /sys/devices/system/cpu/cpuX/regs/identification/ 467 /sys/devices/system/cpu/cpuX/regs/identification/midr_el1 468 /sys/devices/system/cpu/cpuX/regs/identification/revidr_el1 469Date: June 2016 470Contact: Linux ARM Kernel Mailing list <linux-arm-kernel@lists.infradead.org> 471Description: AArch64 CPU registers 472 'identification' directory exposes the CPU ID registers for 473 identifying model and revision of the CPU. 474 475What: /sys/devices/system/cpu/cpu#/cpu_capacity 476Date: December 2016 477Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 478Description: information about CPUs heterogeneity. 479 480 cpu_capacity: capacity of cpu#. 481 482What: /sys/devices/system/cpu/vulnerabilities 483 /sys/devices/system/cpu/vulnerabilities/meltdown 484 /sys/devices/system/cpu/vulnerabilities/spectre_v1 485 /sys/devices/system/cpu/vulnerabilities/spectre_v2 486 /sys/devices/system/cpu/vulnerabilities/spec_store_bypass 487 /sys/devices/system/cpu/vulnerabilities/l1tf 488 /sys/devices/system/cpu/vulnerabilities/mds 489Date: January 2018 490Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 491Description: Information about CPU vulnerabilities 492 493 The files are named after the code names of CPU 494 vulnerabilities. The output of those files reflects the 495 state of the CPUs in the system. Possible output values: 496 497 "Not affected" CPU is not affected by the vulnerability 498 "Vulnerable" CPU is affected and no mitigation in effect 499 "Mitigation: $M" CPU is affected and mitigation $M is in effect 500 501 See also: Documentation/admin-guide/hw-vuln/index.rst 502 503What: /sys/devices/system/cpu/smt 504 /sys/devices/system/cpu/smt/active 505 /sys/devices/system/cpu/smt/control 506Date: June 2018 507Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 508Description: Control Symetric Multi Threading (SMT) 509 510 active: Tells whether SMT is active (enabled and siblings online) 511 512 control: Read/write interface to control SMT. Possible 513 values: 514 515 "on" SMT is enabled 516 "off" SMT is disabled 517 "forceoff" SMT is force disabled. Cannot be changed. 518 "notsupported" SMT is not supported by the CPU 519 "notimplemented" SMT runtime toggling is not 520 implemented for the architecture 521 522 If control status is "forceoff" or "notsupported" writes 523 are rejected. 524 525What: /sys/devices/system/cpu/cpu#/power/energy_perf_bias 526Date: March 2019 527Contact: linux-pm@vger.kernel.org 528Description: Intel Energy and Performance Bias Hint (EPB) 529 530 EPB for the given CPU in a sliding scale 0 - 15, where a value 531 of 0 corresponds to a hint preference for highest performance 532 and a value of 15 corresponds to the maximum energy savings. 533 534 In order to change the EPB value for the CPU, write either 535 a number in the 0 - 15 sliding scale above, or one of the 536 strings: "performance", "balance-performance", "normal", 537 "balance-power", "power" (that represent values reflected by 538 their meaning), to this attribute. 539 540 This attribute is present for all online CPUs supporting the 541 Intel EPB feature. 542 543What: /sys/devices/system/cpu/umwait_control 544 /sys/devices/system/cpu/umwait_control/enable_c02 545 /sys/devices/system/cpu/umwait_control/max_time 546Date: May 2019 547Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org> 548Description: Umwait control 549 550 enable_c02: Read/write interface to control umwait C0.2 state 551 Read returns C0.2 state status: 552 0: C0.2 is disabled 553 1: C0.2 is enabled 554 555 Write 'y' or '1' or 'on' to enable C0.2 state. 556 Write 'n' or '0' or 'off' to disable C0.2 state. 557 558 The interface is case insensitive. 559 560 max_time: Read/write interface to control umwait maximum time 561 in TSC-quanta that the CPU can reside in either C0.1 562 or C0.2 state. The time is an unsigned 32-bit number. 563 Note that a value of zero means there is no limit. 564 Low order two bits must be zero. 565