1.. SPDX-License-Identifier: GPL-2.0 2 3==================== 4The /proc Filesystem 5==================== 6 7===================== ======================================= ================ 8/proc/sys Terrehon Bowden <terrehon@pacbell.net>, October 7 1999 9 Bodo Bauer <bb@ricochet.net> 102.4.x update Jorge Nerin <comandante@zaralinux.com> November 14 2000 11move /proc/sys Shen Feng <shen@cn.fujitsu.com> April 1 2009 12fixes/update part 1.1 Stefani Seibold <stefani@seibold.net> June 9 2009 13===================== ======================================= ================ 14 15 16 17.. Table of Contents 18 19 0 Preface 20 0.1 Introduction/Credits 21 0.2 Legal Stuff 22 23 1 Collecting System Information 24 1.1 Process-Specific Subdirectories 25 1.2 Kernel data 26 1.3 IDE devices in /proc/ide 27 1.4 Networking info in /proc/net 28 1.5 SCSI info 29 1.6 Parallel port info in /proc/parport 30 1.7 TTY info in /proc/tty 31 1.8 Miscellaneous kernel statistics in /proc/stat 32 1.9 Ext4 file system parameters 33 34 2 Modifying System Parameters 35 36 3 Per-Process Parameters 37 3.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj - Adjust the oom-killer 38 score 39 3.2 /proc/<pid>/oom_score - Display current oom-killer score 40 3.3 /proc/<pid>/io - Display the IO accounting fields 41 3.4 /proc/<pid>/coredump_filter - Core dump filtering settings 42 3.5 /proc/<pid>/mountinfo - Information about mounts 43 3.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm 44 3.7 /proc/<pid>/task/<tid>/children - Information about task children 45 3.8 /proc/<pid>/fdinfo/<fd> - Information about opened file 46 3.9 /proc/<pid>/map_files - Information about memory mapped files 47 3.10 /proc/<pid>/timerslack_ns - Task timerslack value 48 3.11 /proc/<pid>/patch_state - Livepatch patch operation state 49 3.12 /proc/<pid>/arch_status - Task architecture specific information 50 3.13 /proc/<pid>/fd - List of symlinks to open files 51 3.14 /proc/<pid>/ksm_stat - Information about the process's ksm status. 52 53 4 Configuring procfs 54 4.1 Mount options 55 56 5 Filesystem behavior 57 58Preface 59======= 60 610.1 Introduction/Credits 62------------------------ 63 64We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of 65other people for help compiling this documentation. We'd also like to extend a 66special thank you to Andi Kleen for documentation, which we relied on heavily 67to create this document, as well as the additional information he provided. 68Thanks to everybody else who contributed source or docs to the Linux kernel 69and helped create a great piece of software... :) 70 71The latest version of this document is available online at 72https://www.kernel.org/doc/html/latest/filesystems/proc.html 73 740.2 Legal Stuff 75--------------- 76 77We don't guarantee the correctness of this document, and if you come to us 78complaining about how you screwed up your system because of incorrect 79documentation, we won't feel responsible... 80 81Chapter 1: Collecting System Information 82======================================== 83 84In This Chapter 85--------------- 86* Investigating the properties of the pseudo file system /proc and its 87 ability to provide information on the running Linux system 88* Examining /proc's structure 89* Uncovering various information about the kernel and the processes running 90 on the system 91 92------------------------------------------------------------------------------ 93 94The proc file system acts as an interface to internal data structures in the 95kernel. It can be used to obtain information about the system and to change 96certain kernel parameters at runtime (sysctl). 97 98First, we'll take a look at the read-only parts of /proc. In Chapter 2, we 99show you how you can use /proc/sys to change settings. 100 1011.1 Process-Specific Subdirectories 102----------------------------------- 103 104The directory /proc contains (among other things) one subdirectory for each 105process running on the system, which is named after the process ID (PID). 106 107The link 'self' points to the process reading the file system. Each process 108subdirectory has the entries listed in Table 1-1. 109 110A process can read its own information from /proc/PID/* with no extra 111permissions. When reading /proc/PID/* information for other processes, reading 112process is required to have either CAP_SYS_PTRACE capability with 113PTRACE_MODE_READ access permissions, or, alternatively, CAP_PERFMON 114capability. This applies to all read-only information like `maps`, `environ`, 115`pagemap`, etc. The only exception is `mem` file due to its read-write nature, 116which requires CAP_SYS_PTRACE capabilities with more elevated 117PTRACE_MODE_ATTACH permissions; CAP_PERFMON capability does not grant access 118to /proc/PID/mem for other processes. 119 120Note that an open file descriptor to /proc/<pid> or to any of its 121contained files or subdirectories does not prevent <pid> being reused 122for some other process in the event that <pid> exits. Operations on 123open /proc/<pid> file descriptors corresponding to dead processes 124never act on any new process that the kernel may, through chance, have 125also assigned the process ID <pid>. Instead, operations on these FDs 126usually fail with ESRCH. 127 128.. table:: Table 1-1: Process specific entries in /proc 129 130 ============= =============================================================== 131 File Content 132 ============= =============================================================== 133 clear_refs Clears page referenced bits shown in smaps output 134 cmdline Command line arguments 135 cpu Current and last cpu in which it was executed (2.4)(smp) 136 cwd Link to the current working directory 137 environ Values of environment variables 138 exe Link to the executable of this process 139 fd Directory, which contains all file descriptors 140 maps Memory maps to executables and library files (2.4) 141 mem Memory held by this process 142 root Link to the root directory of this process 143 stat Process status 144 statm Process memory status information 145 status Process status in human readable form 146 wchan Present with CONFIG_KALLSYMS=y: it shows the kernel function 147 symbol the task is blocked in - or "0" if not blocked. 148 pagemap Page table 149 stack Report full stack trace, enable via CONFIG_STACKTRACE 150 smaps An extension based on maps, showing the memory consumption of 151 each mapping and flags associated with it 152 smaps_rollup Accumulated smaps stats for all mappings of the process. This 153 can be derived from smaps, but is faster and more convenient 154 numa_maps An extension based on maps, showing the memory locality and 155 binding policy as well as mem usage (in pages) of each mapping. 156 ============= =============================================================== 157 158For example, to get the status information of a process, all you have to do is 159read the file /proc/PID/status:: 160 161 >cat /proc/self/status 162 Name: cat 163 State: R (running) 164 Tgid: 5452 165 Pid: 5452 166 PPid: 743 167 TracerPid: 0 (2.4) 168 Uid: 501 501 501 501 169 Gid: 100 100 100 100 170 FDSize: 256 171 Groups: 100 14 16 172 Kthread: 0 173 VmPeak: 5004 kB 174 VmSize: 5004 kB 175 VmLck: 0 kB 176 VmHWM: 476 kB 177 VmRSS: 476 kB 178 RssAnon: 352 kB 179 RssFile: 120 kB 180 RssShmem: 4 kB 181 VmData: 156 kB 182 VmStk: 88 kB 183 VmExe: 68 kB 184 VmLib: 1412 kB 185 VmPTE: 20 kb 186 VmSwap: 0 kB 187 HugetlbPages: 0 kB 188 CoreDumping: 0 189 THP_enabled: 1 190 Threads: 1 191 SigQ: 0/28578 192 SigPnd: 0000000000000000 193 ShdPnd: 0000000000000000 194 SigBlk: 0000000000000000 195 SigIgn: 0000000000000000 196 SigCgt: 0000000000000000 197 CapInh: 00000000fffffeff 198 CapPrm: 0000000000000000 199 CapEff: 0000000000000000 200 CapBnd: ffffffffffffffff 201 CapAmb: 0000000000000000 202 NoNewPrivs: 0 203 Seccomp: 0 204 Speculation_Store_Bypass: thread vulnerable 205 SpeculationIndirectBranch: conditional enabled 206 voluntary_ctxt_switches: 0 207 nonvoluntary_ctxt_switches: 1 208 209This shows you nearly the same information you would get if you viewed it with 210the ps command. In fact, ps uses the proc file system to obtain its 211information. But you get a more detailed view of the process by reading the 212file /proc/PID/status. It fields are described in table 1-2. 213 214The statm file contains more detailed information about the process 215memory usage. Its seven fields are explained in Table 1-3. The stat file 216contains detailed information about the process itself. Its fields are 217explained in Table 1-4. 218 219(for SMP CONFIG users) 220 221For making accounting scalable, RSS related information are handled in an 222asynchronous manner and the value may not be very precise. To see a precise 223snapshot of a moment, you can see /proc/<pid>/smaps file and scan page table. 224It's slow but very precise. 225 226.. table:: Table 1-2: Contents of the status fields (as of 4.19) 227 228 ========================== =================================================== 229 Field Content 230 ========================== =================================================== 231 Name filename of the executable 232 Umask file mode creation mask 233 State state (R is running, S is sleeping, D is sleeping 234 in an uninterruptible wait, Z is zombie, 235 T is traced or stopped) 236 Tgid thread group ID 237 Ngid NUMA group ID (0 if none) 238 Pid process id 239 PPid process id of the parent process 240 TracerPid PID of process tracing this process (0 if not, or 241 the tracer is outside of the current pid namespace) 242 Uid Real, effective, saved set, and file system UIDs 243 Gid Real, effective, saved set, and file system GIDs 244 FDSize number of file descriptor slots currently allocated 245 Groups supplementary group list 246 NStgid descendant namespace thread group ID hierarchy 247 NSpid descendant namespace process ID hierarchy 248 NSpgid descendant namespace process group ID hierarchy 249 NSsid descendant namespace session ID hierarchy 250 Kthread kernel thread flag, 1 is yes, 0 is no 251 VmPeak peak virtual memory size 252 VmSize total program size 253 VmLck locked memory size 254 VmPin pinned memory size 255 VmHWM peak resident set size ("high water mark") 256 VmRSS size of memory portions. It contains the three 257 following parts 258 (VmRSS = RssAnon + RssFile + RssShmem) 259 RssAnon size of resident anonymous memory 260 RssFile size of resident file mappings 261 RssShmem size of resident shmem memory (includes SysV shm, 262 mapping of tmpfs and shared anonymous mappings) 263 VmData size of private data segments 264 VmStk size of stack segments 265 VmExe size of text segment 266 VmLib size of shared library code 267 VmPTE size of page table entries 268 VmSwap amount of swap used by anonymous private data 269 (shmem swap usage is not included) 270 HugetlbPages size of hugetlb memory portions 271 CoreDumping process's memory is currently being dumped 272 (killing the process may lead to a corrupted core) 273 THP_enabled process is allowed to use THP (returns 0 when 274 PR_SET_THP_DISABLE is set on the process to disable 275 THP completely, not just partially) 276 Threads number of threads 277 SigQ number of signals queued/max. number for queue 278 SigPnd bitmap of pending signals for the thread 279 ShdPnd bitmap of shared pending signals for the process 280 SigBlk bitmap of blocked signals 281 SigIgn bitmap of ignored signals 282 SigCgt bitmap of caught signals 283 CapInh bitmap of inheritable capabilities 284 CapPrm bitmap of permitted capabilities 285 CapEff bitmap of effective capabilities 286 CapBnd bitmap of capabilities bounding set 287 CapAmb bitmap of ambient capabilities 288 NoNewPrivs no_new_privs, like prctl(PR_GET_NO_NEW_PRIV, ...) 289 Seccomp seccomp mode, like prctl(PR_GET_SECCOMP, ...) 290 Speculation_Store_Bypass speculative store bypass mitigation status 291 SpeculationIndirectBranch indirect branch speculation mode 292 Cpus_allowed mask of CPUs on which this process may run 293 Cpus_allowed_list Same as previous, but in "list format" 294 Mems_allowed mask of memory nodes allowed to this process 295 Mems_allowed_list Same as previous, but in "list format" 296 voluntary_ctxt_switches number of voluntary context switches 297 nonvoluntary_ctxt_switches number of non voluntary context switches 298 ========================== =================================================== 299 300 301.. table:: Table 1-3: Contents of the statm fields (as of 2.6.8-rc3) 302 303 ======== =============================== ============================== 304 Field Content 305 ======== =============================== ============================== 306 size total program size (pages) (same as VmSize in status) 307 resident size of memory portions (pages) (same as VmRSS in status) 308 shared number of pages that are shared (i.e. backed by a file, same 309 as RssFile+RssShmem in status) 310 trs number of pages that are 'code' (not including libs; broken, 311 includes data segment) 312 lrs number of pages of library (always 0 on 2.6) 313 drs number of pages of data/stack (including libs; broken, 314 includes library text) 315 dt number of dirty pages (always 0 on 2.6) 316 ======== =============================== ============================== 317 318 319.. table:: Table 1-4: Contents of the stat fields (as of 2.6.30-rc7) 320 321 ============= =============================================================== 322 Field Content 323 ============= =============================================================== 324 pid process id 325 tcomm filename of the executable 326 state state (R is running, S is sleeping, D is sleeping in an 327 uninterruptible wait, Z is zombie, T is traced or stopped) 328 ppid process id of the parent process 329 pgrp pgrp of the process 330 sid session id 331 tty_nr tty the process uses 332 tty_pgrp pgrp of the tty 333 flags task flags 334 min_flt number of minor faults 335 cmin_flt number of minor faults with child's 336 maj_flt number of major faults 337 cmaj_flt number of major faults with child's 338 utime user mode jiffies 339 stime kernel mode jiffies 340 cutime user mode jiffies with child's 341 cstime kernel mode jiffies with child's 342 priority priority level 343 nice nice level 344 num_threads number of threads 345 it_real_value (obsolete, always 0) 346 start_time time the process started after system boot 347 vsize virtual memory size 348 rss resident set memory size 349 rsslim current limit in bytes on the rss 350 start_code address above which program text can run 351 end_code address below which program text can run 352 start_stack address of the start of the main process stack 353 esp current value of ESP 354 eip current value of EIP 355 pending bitmap of pending signals 356 blocked bitmap of blocked signals 357 sigign bitmap of ignored signals 358 sigcatch bitmap of caught signals 359 0 (place holder, used to be the wchan address, 360 use /proc/PID/wchan instead) 361 0 (place holder) 362 0 (place holder) 363 exit_signal signal to send to parent thread on exit 364 task_cpu which CPU the task is scheduled on 365 rt_priority realtime priority 366 policy scheduling policy (man sched_setscheduler) 367 blkio_ticks time spent waiting for block IO 368 gtime guest time of the task in jiffies 369 cgtime guest time of the task children in jiffies 370 start_data address above which program data+bss is placed 371 end_data address below which program data+bss is placed 372 start_brk address above which program heap can be expanded with brk() 373 arg_start address above which program command line is placed 374 arg_end address below which program command line is placed 375 env_start address above which program environment is placed 376 env_end address below which program environment is placed 377 exit_code the thread's exit_code in the form reported by the waitpid 378 system call 379 ============= =============================================================== 380 381The /proc/PID/maps file contains the currently mapped memory regions and 382their access permissions. 383 384The format is:: 385 386 address perms offset dev inode pathname 387 388 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test 389 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test 390 0804a000-0806b000 rw-p 00000000 00:00 0 [heap] 391 a7cb1000-a7cb2000 ---p 00000000 00:00 0 392 a7cb2000-a7eb2000 rw-p 00000000 00:00 0 393 a7eb2000-a7eb3000 ---p 00000000 00:00 0 394 a7eb3000-a7ed5000 rw-p 00000000 00:00 0 395 a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6 396 a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6 397 a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6 398 a800b000-a800e000 rw-p 00000000 00:00 0 399 a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0 400 a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0 401 a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0 402 a8024000-a8027000 rw-p 00000000 00:00 0 403 a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2 404 a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2 405 a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2 406 aff35000-aff4a000 rw-p 00000000 00:00 0 [stack] 407 ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso] 408 409where "address" is the address space in the process that it occupies, "perms" 410is a set of permissions:: 411 412 r = read 413 w = write 414 x = execute 415 s = shared 416 p = private (copy on write) 417 418"offset" is the offset into the mapping, "dev" is the device (major:minor), and 419"inode" is the inode on that device. 0 indicates that no inode is associated 420with the memory region, as the case would be with BSS (uninitialized data). 421The "pathname" shows the name associated file for this mapping. If the mapping 422is not associated with a file: 423 424 =================== =========================================== 425 [heap] the heap of the program 426 [stack] the stack of the main process 427 [vdso] the "virtual dynamic shared object", 428 the kernel system call handler 429 [anon:<name>] a private anonymous mapping that has been 430 named by userspace 431 [anon_shmem:<name>] an anonymous shared memory mapping that has 432 been named by userspace 433 =================== =========================================== 434 435 or if empty, the mapping is anonymous. 436 437Starting with 6.11 kernel, /proc/PID/maps provides an alternative 438ioctl()-based API that gives ability to flexibly and efficiently query and 439filter individual VMAs. This interface is binary and is meant for more 440efficient and easy programmatic use. `struct procmap_query`, defined in 441linux/fs.h UAPI header, serves as an input/output argument to the 442`PROCMAP_QUERY` ioctl() command. See comments in linus/fs.h UAPI header for 443details on query semantics, supported flags, data returned, and general API 444usage information. 445 446The /proc/PID/smaps is an extension based on maps, showing the memory 447consumption for each of the process's mappings. For each mapping (aka Virtual 448Memory Area, or VMA) there is a series of lines such as the following:: 449 450 08048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash 451 452 Size: 1084 kB 453 KernelPageSize: 4 kB 454 MMUPageSize: 4 kB 455 Rss: 892 kB 456 Pss: 374 kB 457 Pss_Dirty: 0 kB 458 Shared_Clean: 892 kB 459 Shared_Dirty: 0 kB 460 Private_Clean: 0 kB 461 Private_Dirty: 0 kB 462 Referenced: 892 kB 463 Anonymous: 0 kB 464 KSM: 0 kB 465 LazyFree: 0 kB 466 AnonHugePages: 0 kB 467 FilePmdMapped: 0 kB 468 ShmemPmdMapped: 0 kB 469 Shared_Hugetlb: 0 kB 470 Private_Hugetlb: 0 kB 471 Swap: 0 kB 472 SwapPss: 0 kB 473 Locked: 0 kB 474 THPeligible: 0 475 VmFlags: rd ex mr mw me dw 476 477The first of these lines shows the same information as is displayed for 478the mapping in /proc/PID/maps. Following lines show the size of the 479mapping (size); the smallest possible page size allocated when backing a 480VMA (KernelPageSize), which is the granularity in which VMA modifications 481can be performed; the smallest possible page size that could be used by the 482MMU (MMUPageSize) when backing a VMA; the amount of the mapping that is 483currently resident in RAM (RSS); the process's proportional share of this 484mapping (PSS); and the number of clean and dirty shared and private pages 485in the mapping. 486 487"KernelPageSize" always corresponds to "MMUPageSize", except when a larger 488kernel page size is emulated on a system with a smaller page size used by the 489MMU, which is the case for some PPC64 setups with hugetlb. Furthermore, 490"KernelPageSize" and "MMUPageSize" always correspond to the smallest 491possible granularity (fallback) that can be encountered in a VMA throughout 492its lifetime. These values are not affected by Transparent Huge Pages 493being in effect, or any usage of larger MMU page sizes (either through 494architectural huge-page mappings or other explicit/implicit coalescing of 495virtual ranges performed by the MMU). "AnonHugePages", "ShmemPmdMapped" and 496"FilePmdMapped" provide insight into the usage of PMD-level architectural 497huge-page mappings. 498 499The "proportional set size" (PSS) of a process is the count of pages it has 500in memory, where each page is divided by the number of processes sharing it. 501So if a process has 1000 pages all to itself, and 1000 shared with one other 502process, its PSS will be 1500. "Pss_Dirty" is the portion of PSS which 503consists of dirty pages. ("Pss_Clean" is not included, but it can be 504calculated by subtracting "Pss_Dirty" from "Pss".) 505 506Traditionally, a page is accounted as "private" if it is mapped exactly once, 507and a page is accounted as "shared" when mapped multiple times, even when 508mapped in the same process multiple times. Note that this accounting is 509independent of MAP_SHARED. 510 511In some kernel configurations, the semantics of pages part of a larger 512allocation (e.g., THP) can differ: a page is accounted as "private" if all 513pages part of the corresponding large allocation are *certainly* mapped in the 514same process, even if the page is mapped multiple times in that process. A 515page is accounted as "shared" if any page page of the larger allocation 516is *maybe* mapped in a different process. In some cases, a large allocation 517might be treated as "maybe mapped by multiple processes" even though this 518is no longer the case. 519 520Some kernel configurations do not track the precise number of times a page part 521of a larger allocation is mapped. In this case, when calculating the PSS, the 522average number of mappings per page in this larger allocation might be used 523as an approximation for the number of mappings of a page. The PSS calculation 524will be imprecise in this case. 525 526"Referenced" indicates the amount of memory currently marked as referenced or 527accessed. 528 529"Anonymous" shows the amount of memory that does not belong to any file. Even 530a mapping associated with a file may contain anonymous pages: when MAP_PRIVATE 531and a page is modified, the file page is replaced by a private anonymous copy. 532 533"KSM" reports how many of the pages are KSM pages. Note that KSM-placed zeropages 534are not included, only actual KSM pages. 535 536"LazyFree" shows the amount of memory which is marked by madvise(MADV_FREE). 537The memory isn't freed immediately with madvise(). It's freed in memory 538pressure if the memory is clean. Please note that the printed value might 539be lower than the real value due to optimizations used in the current 540implementation. If this is not desirable please file a bug report. 541 542"AnonHugePages", "ShmemPmdMapped" and "FilePmdMapped" show the amount of 543memory backed by Transparent Huge Pages that are currently mapped by 544architectural huge-page mappings at the PMD level. "AnonHugePages" 545corresponds to memory that does not belong to a file, "ShmemPmdMapped" to 546shared memory (shmem/tmpfs) and "FilePmdMapped" to file-backed memory 547(excluding shmem/tmpfs). 548 549There are no dedicated entries for Transparent Huge Pages (or similar concepts) 550that are not mapped by architectural huge-page mappings at the PMD level. 551 552"Shared_Hugetlb" and "Private_Hugetlb" show the amounts of memory backed by 553hugetlbfs page which is *not* counted in "RSS" or "PSS" field for historical 554reasons. And these are not included in {Shared,Private}_{Clean,Dirty} field. 555 556"Swap" shows how much would-be-anonymous memory is also used, but out on swap. 557 558For shmem mappings, "Swap" includes also the size of the mapped (and not 559replaced by copy-on-write) part of the underlying shmem object out on swap. 560"SwapPss" shows proportional swap share of this mapping. Unlike "Swap", this 561does not take into account swapped out page of underlying shmem objects. 562"Locked" indicates whether the mapping is locked in memory or not. 563 564"THPeligible" indicates whether the mapping is eligible for allocating 565naturally aligned THP pages of any currently enabled size. 1 if true, 0 566otherwise. 567 568"VmFlags" field deserves a separate description. This member represents the 569kernel flags associated with the particular virtual memory area in two letter 570encoded manner. The codes are the following: 571 572 == ============================================================= 573 rd readable 574 wr writeable 575 ex executable 576 sh shared 577 mr may read 578 mw may write 579 me may execute 580 ms may share 581 gd stack segment growns down 582 pf pure PFN range 583 lo pages are locked in memory 584 io memory mapped I/O area 585 sr sequential read advise provided 586 rr random read advise provided 587 dc do not copy area on fork 588 de do not expand area on remapping 589 ac area is accountable 590 nr swap space is not reserved for the area 591 ht area uses huge tlb pages 592 sf synchronous page fault 593 ar architecture specific flag 594 wf wipe on fork 595 dd do not include area into core dump 596 sd soft dirty flag 597 mm mixed map area 598 hg huge page advise flag 599 nh no huge page advise flag 600 mg mergeable advise flag 601 bt arm64 BTI guarded page 602 mt arm64 MTE allocation tags are enabled 603 um userfaultfd missing tracking 604 uw userfaultfd wr-protect tracking 605 ui userfaultfd minor fault 606 ss shadow/guarded control stack page 607 sl sealed 608 lf lock on fault pages 609 dp always lazily freeable mapping 610 gu maybe contains guard regions (if not set, definitely doesn't) 611 == ============================================================= 612 613Note that there is no guarantee that every flag and associated mnemonic will 614be present in all further kernel releases. Things get changed, the flags may 615be vanished or the reverse -- new added. Interpretation of their meaning 616might change in future as well. So each consumer of these flags has to 617follow each specific kernel version for the exact semantic. 618 619This file is only present if the CONFIG_MMU kernel configuration option is 620enabled. 621 622Note: reading /proc/PID/maps or /proc/PID/smaps is inherently racy (consistent 623output can be achieved only in the single read call). 624 625This typically manifests when doing partial reads of these files while the 626memory map is being modified. Despite the races, we do provide the following 627guarantees: 628 6291) The mapped addresses never go backwards, which implies no two 630 regions will ever overlap. 6312) If there is something at a given vaddr during the entirety of the 632 life of the smaps/maps walk, there will be some output for it. 633 634The /proc/PID/smaps_rollup file includes the same fields as /proc/PID/smaps, 635but their values are the sums of the corresponding values for all mappings of 636the process. Additionally, it contains these fields: 637 638- Pss_Anon 639- Pss_File 640- Pss_Shmem 641 642They represent the proportional shares of anonymous, file, and shmem pages, as 643described for smaps above. These fields are omitted in smaps since each 644mapping identifies the type (anon, file, or shmem) of all pages it contains. 645Thus all information in smaps_rollup can be derived from smaps, but at a 646significantly higher cost. 647 648The /proc/PID/clear_refs is used to reset the PG_Referenced and ACCESSED/YOUNG 649bits on both physical and virtual pages associated with a process, and the 650soft-dirty bit on pte (see Documentation/admin-guide/mm/soft-dirty.rst 651for details). 652To clear the bits for all the pages associated with the process:: 653 654 > echo 1 > /proc/PID/clear_refs 655 656To clear the bits for the anonymous pages associated with the process:: 657 658 > echo 2 > /proc/PID/clear_refs 659 660To clear the bits for the file mapped pages associated with the process:: 661 662 > echo 3 > /proc/PID/clear_refs 663 664To clear the soft-dirty bit:: 665 666 > echo 4 > /proc/PID/clear_refs 667 668To reset the peak resident set size ("high water mark") to the process's 669current value:: 670 671 > echo 5 > /proc/PID/clear_refs 672 673Any other value written to /proc/PID/clear_refs will have no effect. 674 675The /proc/pid/pagemap gives the PFN, which can be used to find the pageflags 676using /proc/kpageflags and number of times a page is mapped using 677/proc/kpagecount. For detailed explanation, see 678Documentation/admin-guide/mm/pagemap.rst. 679 680The /proc/pid/numa_maps is an extension based on maps, showing the memory 681locality and binding policy, as well as the memory usage (in pages) of 682each mapping. The output follows a general format where mapping details get 683summarized separated by blank spaces, one mapping per each file line:: 684 685 address policy mapping details 686 687 00400000 default file=/usr/local/bin/app mapped=1 active=0 N3=1 kernelpagesize_kB=4 688 00600000 default file=/usr/local/bin/app anon=1 dirty=1 N3=1 kernelpagesize_kB=4 689 3206000000 default file=/lib64/ld-2.12.so mapped=26 mapmax=6 N0=24 N3=2 kernelpagesize_kB=4 690 320621f000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 691 3206220000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 692 3206221000 default anon=1 dirty=1 N3=1 kernelpagesize_kB=4 693 3206800000 default file=/lib64/libc-2.12.so mapped=59 mapmax=21 active=55 N0=41 N3=18 kernelpagesize_kB=4 694 320698b000 default file=/lib64/libc-2.12.so 695 3206b8a000 default file=/lib64/libc-2.12.so anon=2 dirty=2 N3=2 kernelpagesize_kB=4 696 3206b8e000 default file=/lib64/libc-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 697 3206b8f000 default anon=3 dirty=3 active=1 N3=3 kernelpagesize_kB=4 698 7f4dc10a2000 default anon=3 dirty=3 N3=3 kernelpagesize_kB=4 699 7f4dc10b4000 default anon=2 dirty=2 active=1 N3=2 kernelpagesize_kB=4 700 7f4dc1200000 default file=/anon_hugepage\040(deleted) huge anon=1 dirty=1 N3=1 kernelpagesize_kB=2048 701 7fff335f0000 default stack anon=3 dirty=3 N3=3 kernelpagesize_kB=4 702 7fff3369d000 default mapped=1 mapmax=35 active=0 N3=1 kernelpagesize_kB=4 703 704Where: 705 706"address" is the starting address for the mapping; 707 708"policy" reports the NUMA memory policy set for the mapping (see Documentation/admin-guide/mm/numa_memory_policy.rst); 709 710"mapping details" summarizes mapping data such as mapping type, page usage counters, 711node locality page counters (N0 == node0, N1 == node1, ...) and the kernel page 712size, in KB, that is backing the mapping up. 713 714Note that some kernel configurations do not track the precise number of times 715a page part of a larger allocation (e.g., THP) is mapped. In these 716configurations, "mapmax" might corresponds to the average number of mappings 717per page in such a larger allocation instead. 718 7191.2 Kernel data 720--------------- 721 722Similar to the process entries, the kernel data files give information about 723the running kernel. The files used to obtain this information are contained in 724/proc and are listed in Table 1-5. Not all of these will be present in your 725system. It depends on the kernel configuration and the loaded modules, which 726files are there, and which are missing. 727 728.. table:: Table 1-5: Kernel info in /proc 729 730 ============ =============================================================== 731 File Content 732 ============ =============================================================== 733 allocinfo Memory allocations profiling information 734 apm Advanced power management info 735 bootconfig Kernel command line obtained from boot config, 736 and, if there were kernel parameters from the 737 boot loader, a "# Parameters from bootloader:" 738 line followed by a line containing those 739 parameters prefixed by "# ". (5.5) 740 buddyinfo Kernel memory allocator information (see text) (2.5) 741 bus Directory containing bus specific information 742 cmdline Kernel command line, both from bootloader and embedded 743 in the kernel image 744 cpuinfo Info about the CPU 745 devices Available devices (block and character) 746 dma Used DMA channels 747 filesystems Supported filesystems 748 driver Various drivers grouped here, currently rtc (2.4) 749 execdomains Execdomains, related to security (2.4) 750 fb Frame Buffer devices (2.4) 751 fs File system parameters, currently nfs/exports (2.4) 752 ide Directory containing info about the IDE subsystem 753 interrupts Interrupt usage 754 iomem Memory map (2.4) 755 ioports I/O port usage 756 irq Masks for irq to cpu affinity (2.4)(smp?) 757 isapnp ISA PnP (Plug&Play) Info (2.4) 758 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4)) 759 kmsg Kernel messages 760 ksyms Kernel symbol table 761 loadavg Load average of last 1, 5 & 15 minutes; 762 number of processes currently runnable (running or on ready queue); 763 total number of processes in system; 764 last pid created. 765 All fields are separated by one space except "number of 766 processes currently runnable" and "total number of processes 767 in system", which are separated by a slash ('/'). Example: 768 0.61 0.61 0.55 3/828 22084 769 locks Kernel locks 770 meminfo Memory info 771 misc Miscellaneous 772 modules List of loaded modules 773 mounts Mounted filesystems 774 net Networking info (see text) 775 pagetypeinfo Additional page allocator information (see text) (2.5) 776 partitions Table of partitions known to the system 777 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/, 778 decoupled by lspci (2.4) 779 rtc Real time clock 780 scsi SCSI info (see text) 781 slabinfo Slab pool info 782 softirqs softirq usage 783 stat Overall statistics 784 swaps Swap space utilization 785 sys See chapter 2 786 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4) 787 tty Info of tty drivers 788 uptime Wall clock since boot, combined idle time of all cpus 789 version Kernel version 790 video bttv info of video resources (2.4) 791 vmallocinfo Show vmalloced areas 792 ============ =============================================================== 793 794You can, for example, check which interrupts are currently in use and what 795they are used for by looking in the file /proc/interrupts:: 796 797 > cat /proc/interrupts 798 CPU0 799 0: 8728810 XT-PIC timer 800 1: 895 XT-PIC keyboard 801 2: 0 XT-PIC cascade 802 3: 531695 XT-PIC aha152x 803 4: 2014133 XT-PIC serial 804 5: 44401 XT-PIC pcnet_cs 805 8: 2 XT-PIC rtc 806 11: 8 XT-PIC i82365 807 12: 182918 XT-PIC PS/2 Mouse 808 13: 1 XT-PIC fpu 809 14: 1232265 XT-PIC ide0 810 15: 7 XT-PIC ide1 811 NMI: 0 812 813In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the 814output of a SMP machine):: 815 816 > cat /proc/interrupts 817 818 CPU0 CPU1 819 0: 1243498 1214548 IO-APIC-edge timer 820 1: 8949 8958 IO-APIC-edge keyboard 821 2: 0 0 XT-PIC cascade 822 5: 11286 10161 IO-APIC-edge soundblaster 823 8: 1 0 IO-APIC-edge rtc 824 9: 27422 27407 IO-APIC-edge 3c503 825 12: 113645 113873 IO-APIC-edge PS/2 Mouse 826 13: 0 0 XT-PIC fpu 827 14: 22491 24012 IO-APIC-edge ide0 828 15: 2183 2415 IO-APIC-edge ide1 829 17: 30564 30414 IO-APIC-level eth0 830 18: 177 164 IO-APIC-level bttv 831 NMI: 2457961 2457959 832 LOC: 2457882 2457881 833 ERR: 2155 834 835NMI is incremented in this case because every timer interrupt generates a NMI 836(Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups. 837 838LOC is the local interrupt counter of the internal APIC of every CPU. 839 840ERR is incremented in the case of errors in the IO-APIC bus (the bus that 841connects the CPUs in a SMP system. This means that an error has been detected, 842the IO-APIC automatically retry the transmission, so it should not be a big 843problem, but you should read the SMP-FAQ. 844 845In 2.6.2* /proc/interrupts was expanded again. This time the goal was for 846/proc/interrupts to display every IRQ vector in use by the system, not 847just those considered 'most important'. The new vectors are: 848 849THR 850 interrupt raised when a machine check threshold counter 851 (typically counting ECC corrected errors of memory or cache) exceeds 852 a configurable threshold. Only available on some systems. 853 854TRM 855 a thermal event interrupt occurs when a temperature threshold 856 has been exceeded for the CPU. This interrupt may also be generated 857 when the temperature drops back to normal. 858 859SPU 860 a spurious interrupt is some interrupt that was raised then lowered 861 by some IO device before it could be fully processed by the APIC. Hence 862 the APIC sees the interrupt but does not know what device it came from. 863 For this case the APIC will generate the interrupt with a IRQ vector 864 of 0xff. This might also be generated by chipset bugs. 865 866RES, CAL, TLB 867 rescheduling, call and TLB flush interrupts are 868 sent from one CPU to another per the needs of the OS. Typically, 869 their statistics are used by kernel developers and interested users to 870 determine the occurrence of interrupts of the given type. 871 872The above IRQ vectors are displayed only when relevant. For example, 873the threshold vector does not exist on x86_64 platforms. Others are 874suppressed when the system is a uniprocessor. As of this writing, only 875i386 and x86_64 platforms support the new IRQ vector displays. 876 877Of some interest is the introduction of the /proc/irq directory to 2.4. 878It could be used to set IRQ to CPU affinity. This means that you can "hook" an 879IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the 880irq subdir is one subdir for each IRQ, and default_smp_affinity. 881 882For example:: 883 884 > ls /proc/irq/ 885 0 10 12 14 16 18 2 4 6 8 default_smp_affinity 886 1 11 13 15 17 19 3 5 7 9 887 > ls /proc/irq/0/ 888 smp_affinity 889 890smp_affinity is a bitmask, in which you can specify which CPUs can handle the 891IRQ. You can set it by doing:: 892 893 > echo 1 > /proc/irq/10/smp_affinity 894 895This means that only the first CPU will handle the IRQ, but you can also echo 8965 which means that only the first and third CPU can handle the IRQ. 897 898The contents of each smp_affinity file is the same by default:: 899 900 > cat /proc/irq/0/smp_affinity 901 ffffffff 902 903There is an alternate interface, smp_affinity_list which allows specifying 904a CPU range instead of a bitmask:: 905 906 > cat /proc/irq/0/smp_affinity_list 907 1024-1031 908 909The default_smp_affinity mask applies to all non-active IRQs, which are the 910IRQs which have not yet been allocated/activated, and hence which lack a 911/proc/irq/[0-9]* directory. 912 913The node file on an SMP system shows the node to which the device using the IRQ 914reports itself as being attached. This hardware locality information does not 915include information about any possible driver locality preference. 916 917The way IRQs are routed is handled by the IO-APIC, and it's Round Robin 918between all the CPUs which are allowed to handle it. As usual the kernel has 919more info than you and does a better job than you, so the defaults are the 920best choice for almost everyone. [Note this applies only to those IO-APIC's 921that support "Round Robin" interrupt distribution.] 922 923There are three more important subdirectories in /proc: net, scsi, and sys. 924The general rule is that the contents, or even the existence of these 925directories, depend on your kernel configuration. If SCSI is not enabled, the 926directory scsi may not exist. The same is true with the net, which is there 927only when networking support is present in the running kernel. 928 929The slabinfo file gives information about memory usage at the slab level. 930Linux uses slab pools for memory management above page level in version 2.2. 931Commonly used objects have their own slab pool (such as network buffers, 932directory cache, and so on). 933 934:: 935 936 > cat /proc/buddyinfo 937 938 Node 0, zone DMA 0 4 5 4 4 3 ... 939 Node 0, zone Normal 1 0 0 1 101 8 ... 940 Node 0, zone HighMem 2 0 0 1 1 0 ... 941 942External fragmentation is a problem under some workloads, and buddyinfo is a 943useful tool for helping diagnose these problems. Buddyinfo will give you a 944clue as to how big an area you can safely allocate, or why a previous 945allocation failed. 946 947Each column represents the number of pages of a certain order which are 948available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in 949ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE 950available in ZONE_NORMAL, etc... 951 952More information relevant to external fragmentation can be found in 953pagetypeinfo:: 954 955 > cat /proc/pagetypeinfo 956 Page block order: 9 957 Pages per block: 512 958 959 Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 960 Node 0, zone DMA, type Unmovable 0 0 0 1 1 1 1 1 1 1 0 961 Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 962 Node 0, zone DMA, type Movable 1 1 2 1 2 1 1 0 1 0 2 963 Node 0, zone DMA, type Reserve 0 0 0 0 0 0 0 0 0 1 0 964 Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 965 Node 0, zone DMA32, type Unmovable 103 54 77 1 1 1 11 8 7 1 9 966 Node 0, zone DMA32, type Reclaimable 0 0 2 1 0 0 0 0 1 0 0 967 Node 0, zone DMA32, type Movable 169 152 113 91 77 54 39 13 6 1 452 968 Node 0, zone DMA32, type Reserve 1 2 2 2 2 0 1 1 1 1 0 969 Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 970 971 Number of blocks type Unmovable Reclaimable Movable Reserve Isolate 972 Node 0, zone DMA 2 0 5 1 0 973 Node 0, zone DMA32 41 6 967 2 0 974 975Fragmentation avoidance in the kernel works by grouping pages of different 976migrate types into the same contiguous regions of memory called page blocks. 977A page block is typically the size of the default hugepage size, e.g. 2MB on 978X86-64. By keeping pages grouped based on their ability to move, the kernel 979can reclaim pages within a page block to satisfy a high-order allocation. 980 981The pagetypinfo begins with information on the size of a page block. It 982then gives the same type of information as buddyinfo except broken down 983by migrate-type and finishes with details on how many page blocks of each 984type exist. 985 986If min_free_kbytes has been tuned correctly (recommendations made by hugeadm 987from libhugetlbfs https://github.com/libhugetlbfs/libhugetlbfs/), one can 988make an estimate of the likely number of huge pages that can be allocated 989at a given point in time. All the "Movable" blocks should be allocatable 990unless memory has been mlock()'d. Some of the Reclaimable blocks should 991also be allocatable although a lot of filesystem metadata may have to be 992reclaimed to achieve this. 993 994 995allocinfo 996~~~~~~~~~ 997 998Provides information about memory allocations at all locations in the code 999base. Each allocation in the code is identified by its source file, line 1000number, module (if originates from a loadable module) and the function calling 1001the allocation. The number of bytes allocated and number of calls at each 1002location are reported. The first line indicates the version of the file, the 1003second line is the header listing fields in the file. 1004If file version is 2.0 or higher then each line may contain additional 1005<key>:<value> pairs representing extra information about the call site. 1006For example if the counters are not accurate, the line will be appended with 1007"accurate:no" pair. 1008 1009Supported markers in v2: 1010accurate:no 1011 1012 Absolute values of the counters in this line are not accurate 1013 because of the failure to allocate memory to track some of the 1014 allocations made at this location. Deltas in these counters are 1015 accurate, therefore counters can be used to track allocation size 1016 and count changes. 1017 1018Example output. 1019 1020:: 1021 1022 > tail -n +3 /proc/allocinfo | sort -rn 1023 127664128 31168 mm/page_ext.c:270 func:alloc_page_ext 1024 56373248 4737 mm/slub.c:2259 func:alloc_slab_page 1025 14880768 3633 mm/readahead.c:247 func:page_cache_ra_unbounded 1026 14417920 3520 mm/mm_init.c:2530 func:alloc_large_system_hash 1027 13377536 234 block/blk-mq.c:3421 func:blk_mq_alloc_rqs 1028 11718656 2861 mm/filemap.c:1919 func:__filemap_get_folio 1029 9192960 2800 kernel/fork.c:307 func:alloc_thread_stack_node 1030 4206592 4 net/netfilter/nf_conntrack_core.c:2567 func:nf_ct_alloc_hashtable 1031 4136960 1010 drivers/staging/ctagmod/ctagmod.c:20 [ctagmod] func:ctagmod_start 1032 3940352 962 mm/memory.c:4214 func:alloc_anon_folio 1033 2894464 22613 fs/kernfs/dir.c:615 func:__kernfs_new_node 1034 ... 1035 1036 1037meminfo 1038~~~~~~~ 1039 1040Provides information about distribution and utilization of memory. This 1041varies by architecture and compile options. Some of the counters reported 1042here overlap. The memory reported by the non overlapping counters may not 1043add up to the overall memory usage and the difference for some workloads 1044can be substantial. In many cases there are other means to find out 1045additional memory using subsystem specific interfaces, for instance 1046/proc/net/sockstat for TCP memory allocations. 1047 1048Example output. You may not have all of these fields. 1049 1050:: 1051 1052 > cat /proc/meminfo 1053 1054 MemTotal: 32858820 kB 1055 MemFree: 21001236 kB 1056 MemAvailable: 27214312 kB 1057 Buffers: 581092 kB 1058 Cached: 5587612 kB 1059 SwapCached: 0 kB 1060 Active: 3237152 kB 1061 Inactive: 7586256 kB 1062 Active(anon): 94064 kB 1063 Inactive(anon): 4570616 kB 1064 Active(file): 3143088 kB 1065 Inactive(file): 3015640 kB 1066 Unevictable: 0 kB 1067 Mlocked: 0 kB 1068 SwapTotal: 0 kB 1069 SwapFree: 0 kB 1070 Zswap: 1904 kB 1071 Zswapped: 7792 kB 1072 Dirty: 12 kB 1073 Writeback: 0 kB 1074 AnonPages: 4654780 kB 1075 Mapped: 266244 kB 1076 Shmem: 9976 kB 1077 KReclaimable: 517708 kB 1078 Slab: 660044 kB 1079 SReclaimable: 517708 kB 1080 SUnreclaim: 142336 kB 1081 KernelStack: 11168 kB 1082 PageTables: 20540 kB 1083 SecPageTables: 0 kB 1084 NFS_Unstable: 0 kB 1085 Bounce: 0 kB 1086 WritebackTmp: 0 kB 1087 CommitLimit: 16429408 kB 1088 Committed_AS: 7715148 kB 1089 VmallocTotal: 34359738367 kB 1090 VmallocUsed: 40444 kB 1091 VmallocChunk: 0 kB 1092 Percpu: 29312 kB 1093 EarlyMemtestBad: 0 kB 1094 HardwareCorrupted: 0 kB 1095 AnonHugePages: 4149248 kB 1096 ShmemHugePages: 0 kB 1097 ShmemPmdMapped: 0 kB 1098 FileHugePages: 0 kB 1099 FilePmdMapped: 0 kB 1100 CmaTotal: 0 kB 1101 CmaFree: 0 kB 1102 Unaccepted: 0 kB 1103 Balloon: 0 kB 1104 HugePages_Total: 0 1105 HugePages_Free: 0 1106 HugePages_Rsvd: 0 1107 HugePages_Surp: 0 1108 Hugepagesize: 2048 kB 1109 Hugetlb: 0 kB 1110 DirectMap4k: 401152 kB 1111 DirectMap2M: 10008576 kB 1112 DirectMap1G: 24117248 kB 1113 1114MemTotal 1115 Total usable RAM (i.e. physical RAM minus a few reserved 1116 bits and the kernel binary code) 1117MemFree 1118 Total free RAM. On highmem systems, the sum of LowFree+HighFree 1119MemAvailable 1120 An estimate of how much memory is available for starting new 1121 applications, without swapping. Calculated from MemFree, 1122 SReclaimable, the size of the file LRU lists, and the low 1123 watermarks in each zone. 1124 The estimate takes into account that the system needs some 1125 page cache to function well, and that not all reclaimable 1126 slab will be reclaimable, due to items being in use. The 1127 impact of those factors will vary from system to system. 1128Buffers 1129 Relatively temporary storage for raw disk blocks 1130 shouldn't get tremendously large (20MB or so) 1131Cached 1132 In-memory cache for files read from the disk (the 1133 pagecache) as well as tmpfs & shmem. 1134 Doesn't include SwapCached. 1135SwapCached 1136 Memory that once was swapped out, is swapped back in but 1137 still also is in the swapfile (if memory is needed it 1138 doesn't need to be swapped out AGAIN because it is already 1139 in the swapfile. This saves I/O) 1140Active 1141 Memory that has been used more recently and usually not 1142 reclaimed unless absolutely necessary. 1143Inactive 1144 Memory which has been less recently used. It is more 1145 eligible to be reclaimed for other purposes 1146Unevictable 1147 Memory allocated for userspace which cannot be reclaimed, such 1148 as mlocked pages, ramfs backing pages, secret memfd pages etc. 1149Mlocked 1150 Memory locked with mlock(). 1151HighTotal, HighFree 1152 Highmem is all memory above ~860MB of physical memory. 1153 Highmem areas are for use by userspace programs, or 1154 for the pagecache. The kernel must use tricks to access 1155 this memory, making it slower to access than lowmem. 1156LowTotal, LowFree 1157 Lowmem is memory which can be used for everything that 1158 highmem can be used for, but it is also available for the 1159 kernel's use for its own data structures. Among many 1160 other things, it is where everything from the Slab is 1161 allocated. Bad things happen when you're out of lowmem. 1162SwapTotal 1163 total amount of swap space available 1164SwapFree 1165 Memory which has been evicted from RAM, and is temporarily 1166 on the disk 1167Zswap 1168 Memory consumed by the zswap backend (compressed size) 1169Zswapped 1170 Amount of anonymous memory stored in zswap (original size) 1171Dirty 1172 Memory which is waiting to get written back to the disk 1173Writeback 1174 Memory which is actively being written back to the disk 1175AnonPages 1176 Non-file backed pages mapped into userspace page tables. Note that 1177 some kernel configurations might consider all pages part of a 1178 larger allocation (e.g., THP) as "mapped", as soon as a single 1179 page is mapped. 1180Mapped 1181 files which have been mmapped, such as libraries. Note that some 1182 kernel configurations might consider all pages part of a larger 1183 allocation (e.g., THP) as "mapped", as soon as a single page is 1184 mapped. 1185Shmem 1186 Total memory used by shared memory (shmem) and tmpfs 1187KReclaimable 1188 Kernel allocations that the kernel will attempt to reclaim 1189 under memory pressure. Includes SReclaimable (below), and other 1190 direct allocations with a shrinker. 1191Slab 1192 in-kernel data structures cache 1193SReclaimable 1194 Part of Slab, that might be reclaimed, such as caches 1195SUnreclaim 1196 Part of Slab, that cannot be reclaimed on memory pressure 1197KernelStack 1198 Memory consumed by the kernel stacks of all tasks 1199PageTables 1200 Memory consumed by userspace page tables 1201SecPageTables 1202 Memory consumed by secondary page tables, this currently includes 1203 KVM mmu and IOMMU allocations on x86 and arm64. 1204NFS_Unstable 1205 Always zero. Previously counted pages which had been written to 1206 the server, but has not been committed to stable storage. 1207Bounce 1208 Always zero. Previously memory used for block device 1209 "bounce buffers". 1210WritebackTmp 1211 Always zero. Previously memory used by FUSE for temporary 1212 writeback buffers. 1213CommitLimit 1214 Based on the overcommit ratio ('vm.overcommit_ratio'), 1215 this is the total amount of memory currently available to 1216 be allocated on the system. This limit is only adhered to 1217 if strict overcommit accounting is enabled (mode 2 in 1218 'vm.overcommit_memory'). 1219 1220 The CommitLimit is calculated with the following formula:: 1221 1222 CommitLimit = ([total RAM pages] - [total huge TLB pages]) * 1223 overcommit_ratio / 100 + [total swap pages] 1224 1225 For example, on a system with 1G of physical RAM and 7G 1226 of swap with a `vm.overcommit_ratio` of 30 it would 1227 yield a CommitLimit of 7.3G. 1228 1229 For more details, see the memory overcommit documentation 1230 in mm/overcommit-accounting. 1231Committed_AS 1232 The amount of memory presently allocated on the system. 1233 The committed memory is a sum of all of the memory which 1234 has been allocated by processes, even if it has not been 1235 "used" by them as of yet. A process which malloc()'s 1G 1236 of memory, but only touches 300M of it will show up as 1237 using 1G. This 1G is memory which has been "committed" to 1238 by the VM and can be used at any time by the allocating 1239 application. With strict overcommit enabled on the system 1240 (mode 2 in 'vm.overcommit_memory'), allocations which would 1241 exceed the CommitLimit (detailed above) will not be permitted. 1242 This is useful if one needs to guarantee that processes will 1243 not fail due to lack of memory once that memory has been 1244 successfully allocated. 1245VmallocTotal 1246 total size of vmalloc virtual address space 1247VmallocUsed 1248 amount of vmalloc area which is used 1249VmallocChunk 1250 largest contiguous block of vmalloc area which is free 1251Percpu 1252 Memory allocated to the percpu allocator used to back percpu 1253 allocations. This stat excludes the cost of metadata. 1254EarlyMemtestBad 1255 The amount of RAM/memory in kB, that was identified as corrupted 1256 by early memtest. If memtest was not run, this field will not 1257 be displayed at all. Size is never rounded down to 0 kB. 1258 That means if 0 kB is reported, you can safely assume 1259 there was at least one pass of memtest and none of the passes 1260 found a single faulty byte of RAM. 1261HardwareCorrupted 1262 The amount of RAM/memory in KB, the kernel identifies as 1263 corrupted. 1264AnonHugePages 1265 Non-file backed huge pages mapped into userspace page tables 1266ShmemHugePages 1267 Memory used by shared memory (shmem) and tmpfs allocated 1268 with huge pages 1269ShmemPmdMapped 1270 Shared memory mapped into userspace with huge pages 1271FileHugePages 1272 Memory used for filesystem data (page cache) allocated 1273 with huge pages 1274FilePmdMapped 1275 Page cache mapped into userspace with huge pages 1276CmaTotal 1277 Memory reserved for the Contiguous Memory Allocator (CMA) 1278CmaFree 1279 Free remaining memory in the CMA reserves 1280Unaccepted 1281 Memory that has not been accepted by the guest 1282Balloon 1283 Memory returned to Host by VM Balloon Drivers 1284HugePages_Total, HugePages_Free, HugePages_Rsvd, HugePages_Surp, Hugepagesize, Hugetlb 1285 See Documentation/admin-guide/mm/hugetlbpage.rst. 1286DirectMap4k, DirectMap2M, DirectMap1G 1287 Breakdown of page table sizes used in the kernel's 1288 identity mapping of RAM 1289 1290vmallocinfo 1291~~~~~~~~~~~ 1292 1293Provides information about vmalloced/vmaped areas. One line per area, 1294containing the virtual address range of the area, size in bytes, 1295caller information of the creator, and optional information depending 1296on the kind of area: 1297 1298 ========== =================================================== 1299 pages=nr number of pages 1300 phys=addr if a physical address was specified 1301 ioremap I/O mapping (ioremap() and friends) 1302 vmalloc vmalloc() area 1303 vmap vmap()ed pages 1304 user VM_USERMAP area 1305 vpages buffer for pages pointers was vmalloced (huge area) 1306 N<node>=nr (Only on NUMA kernels) 1307 Number of pages allocated on memory node <node> 1308 ========== =================================================== 1309 1310:: 1311 1312 > cat /proc/vmallocinfo 1313 0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ... 1314 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128 1315 0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ... 1316 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64 1317 0xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f... 1318 phys=7fee8000 ioremap 1319 0xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f... 1320 phys=7fee7000 ioremap 1321 0xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210 1322 0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ... 1323 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3 1324 0xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ... 1325 pages=2 vmalloc N1=2 1326 0xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ... 1327 /0x130 [x_tables] pages=4 vmalloc N0=4 1328 0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ... 1329 pages=14 vmalloc N2=14 1330 0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ... 1331 pages=4 vmalloc N1=4 1332 0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ... 1333 pages=2 vmalloc N1=2 1334 0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ... 1335 pages=10 vmalloc N0=10 1336 1337 1338softirqs 1339~~~~~~~~ 1340 1341Provides counts of softirq handlers serviced since boot time, for each CPU. 1342 1343:: 1344 1345 > cat /proc/softirqs 1346 CPU0 CPU1 CPU2 CPU3 1347 HI: 0 0 0 0 1348 TIMER: 27166 27120 27097 27034 1349 NET_TX: 0 0 0 17 1350 NET_RX: 42 0 0 39 1351 BLOCK: 0 0 107 1121 1352 TASKLET: 0 0 0 290 1353 SCHED: 27035 26983 26971 26746 1354 HRTIMER: 0 0 0 0 1355 RCU: 1678 1769 2178 2250 1356 13571.3 Networking info in /proc/net 1358-------------------------------- 1359 1360The subdirectory /proc/net follows the usual pattern. Table 1-8 shows the 1361additional values you get for IP version 6 if you configure the kernel to 1362support this. Table 1-9 lists the files and their meaning. 1363 1364 1365.. table:: Table 1-8: IPv6 info in /proc/net 1366 1367 ========== ===================================================== 1368 File Content 1369 ========== ===================================================== 1370 udp6 UDP sockets (IPv6) 1371 tcp6 TCP sockets (IPv6) 1372 raw6 Raw device statistics (IPv6) 1373 igmp6 IP multicast addresses, which this host joined (IPv6) 1374 if_inet6 List of IPv6 interface addresses 1375 ipv6_route Kernel routing table for IPv6 1376 rt6_stats Global IPv6 routing tables statistics 1377 sockstat6 Socket statistics (IPv6) 1378 snmp6 Snmp data (IPv6) 1379 ========== ===================================================== 1380 1381.. table:: Table 1-9: Network info in /proc/net 1382 1383 ============= ================================================================ 1384 File Content 1385 ============= ================================================================ 1386 arp Kernel ARP table 1387 dev network devices with statistics 1388 dev_mcast the Layer2 multicast groups a device is listening too 1389 (interface index, label, number of references, number of bound 1390 addresses). 1391 dev_stat network device status 1392 ip_fwchains Firewall chain linkage 1393 ip_fwnames Firewall chain names 1394 ip_masq Directory containing the masquerading tables 1395 ip_masquerade Major masquerading table 1396 netstat Network statistics 1397 raw raw device statistics 1398 route Kernel routing table 1399 rpc Directory containing rpc info 1400 rt_cache Routing cache 1401 snmp SNMP data 1402 sockstat Socket statistics 1403 softnet_stat Per-CPU incoming packets queues statistics of online CPUs 1404 tcp TCP sockets 1405 udp UDP sockets 1406 unix UNIX domain sockets 1407 wireless Wireless interface data (Wavelan etc) 1408 igmp IP multicast addresses, which this host joined 1409 psched Global packet scheduler parameters. 1410 netlink List of PF_NETLINK sockets 1411 ip_mr_vifs List of multicast virtual interfaces 1412 ip_mr_cache List of multicast routing cache 1413 ============= ================================================================ 1414 1415You can use this information to see which network devices are available in 1416your system and how much traffic was routed over those devices:: 1417 1418 > cat /proc/net/dev 1419 Inter-|Receive |[... 1420 face |bytes packets errs drop fifo frame compressed multicast|[... 1421 lo: 908188 5596 0 0 0 0 0 0 [... 1422 ppp0:15475140 20721 410 0 0 410 0 0 [... 1423 eth0: 614530 7085 0 0 0 0 0 1 [... 1424 1425 ...] Transmit 1426 ...] bytes packets errs drop fifo colls carrier compressed 1427 ...] 908188 5596 0 0 0 0 0 0 1428 ...] 1375103 17405 0 0 0 0 0 0 1429 ...] 1703981 5535 0 0 0 3 0 0 1430 1431In addition, each Channel Bond interface has its own directory. For 1432example, the bond0 device will have a directory called /proc/net/bond0/. 1433It will contain information that is specific to that bond, such as the 1434current slaves of the bond, the link status of the slaves, and how 1435many times the slaves link has failed. 1436 14371.4 SCSI info 1438------------- 1439 1440If you have a SCSI or ATA host adapter in your system, you'll find a 1441subdirectory named after the driver for this adapter in /proc/scsi. 1442You'll also see a list of all recognized SCSI devices in /proc/scsi:: 1443 1444 >cat /proc/scsi/scsi 1445 Attached devices: 1446 Host: scsi0 Channel: 00 Id: 00 Lun: 00 1447 Vendor: IBM Model: DGHS09U Rev: 03E0 1448 Type: Direct-Access ANSI SCSI revision: 03 1449 Host: scsi0 Channel: 00 Id: 06 Lun: 00 1450 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04 1451 Type: CD-ROM ANSI SCSI revision: 02 1452 1453 1454The directory named after the driver has one file for each adapter found in 1455the system. These files contain information about the controller, including 1456the used IRQ and the IO address range. The amount of information shown is 1457dependent on the adapter you use. The example shows the output for an Adaptec 1458AHA-2940 SCSI adapter:: 1459 1460 > cat /proc/scsi/aic7xxx/0 1461 1462 Adaptec AIC7xxx driver version: 5.1.19/3.2.4 1463 Compile Options: 1464 TCQ Enabled By Default : Disabled 1465 AIC7XXX_PROC_STATS : Disabled 1466 AIC7XXX_RESET_DELAY : 5 1467 Adapter Configuration: 1468 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter 1469 Ultra Wide Controller 1470 PCI MMAPed I/O Base: 0xeb001000 1471 Adapter SEEPROM Config: SEEPROM found and used. 1472 Adaptec SCSI BIOS: Enabled 1473 IRQ: 10 1474 SCBs: Active 0, Max Active 2, 1475 Allocated 15, HW 16, Page 255 1476 Interrupts: 160328 1477 BIOS Control Word: 0x18b6 1478 Adapter Control Word: 0x005b 1479 Extended Translation: Enabled 1480 Disconnect Enable Flags: 0xffff 1481 Ultra Enable Flags: 0x0001 1482 Tag Queue Enable Flags: 0x0000 1483 Ordered Queue Tag Flags: 0x0000 1484 Default Tag Queue Depth: 8 1485 Tagged Queue By Device array for aic7xxx host instance 0: 1486 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255} 1487 Actual queue depth per device for aic7xxx host instance 0: 1488 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1} 1489 Statistics: 1490 (scsi0:0:0:0) 1491 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8 1492 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0) 1493 Total transfers 160151 (74577 reads and 85574 writes) 1494 (scsi0:0:6:0) 1495 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15 1496 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0) 1497 Total transfers 0 (0 reads and 0 writes) 1498 1499 15001.5 Parallel port info in /proc/parport 1501--------------------------------------- 1502 1503The directory /proc/parport contains information about the parallel ports of 1504your system. It has one subdirectory for each port, named after the port 1505number (0,1,2,...). 1506 1507These directories contain the four files shown in Table 1-10. 1508 1509 1510.. table:: Table 1-10: Files in /proc/parport 1511 1512 ========= ==================================================================== 1513 File Content 1514 ========= ==================================================================== 1515 autoprobe Any IEEE-1284 device ID information that has been acquired. 1516 devices list of the device drivers using that port. A + will appear by the 1517 name of the device currently using the port (it might not appear 1518 against any). 1519 hardware Parallel port's base address, IRQ line and DMA channel. 1520 irq IRQ that parport is using for that port. This is in a separate 1521 file to allow you to alter it by writing a new value in (IRQ 1522 number or none). 1523 ========= ==================================================================== 1524 15251.6 TTY info in /proc/tty 1526------------------------- 1527 1528Information about the available and actually used tty's can be found in the 1529directory /proc/tty. You'll find entries for drivers and line disciplines in 1530this directory, as shown in Table 1-11. 1531 1532 1533.. table:: Table 1-11: Files in /proc/tty 1534 1535 ============= ============================================== 1536 File Content 1537 ============= ============================================== 1538 drivers list of drivers and their usage 1539 ldiscs registered line disciplines 1540 driver/serial usage statistic and status of single tty lines 1541 ============= ============================================== 1542 1543To see which tty's are currently in use, you can simply look into the file 1544/proc/tty/drivers:: 1545 1546 > cat /proc/tty/drivers 1547 pty_slave /dev/pts 136 0-255 pty:slave 1548 pty_master /dev/ptm 128 0-255 pty:master 1549 pty_slave /dev/ttyp 3 0-255 pty:slave 1550 pty_master /dev/pty 2 0-255 pty:master 1551 serial /dev/cua 5 64-67 serial:callout 1552 serial /dev/ttyS 4 64-67 serial 1553 /dev/tty0 /dev/tty0 4 0 system:vtmaster 1554 /dev/ptmx /dev/ptmx 5 2 system 1555 /dev/console /dev/console 5 1 system:console 1556 /dev/tty /dev/tty 5 0 system:/dev/tty 1557 unknown /dev/tty 4 1-63 console 1558 1559 15601.7 Miscellaneous kernel statistics in /proc/stat 1561------------------------------------------------- 1562 1563Various pieces of information about kernel activity are available in the 1564/proc/stat file. All of the numbers reported in this file are aggregates 1565since the system first booted. For a quick look, simply cat the file:: 1566 1567 > cat /proc/stat 1568 cpu 237902850 368826709 106375398 1873517540 1135548 0 14507935 0 0 0 1569 cpu0 60045249 91891769 26331539 468411416 495718 0 5739640 0 0 0 1570 cpu1 59746288 91759249 26609887 468860630 312281 0 4384817 0 0 0 1571 cpu2 59489247 92985423 26904446 467808813 171668 0 2268998 0 0 0 1572 cpu3 58622065 92190267 26529524 468436680 155879 0 2114478 0 0 0 1573 intr 8688370575 8 3373 0 0 0 0 0 0 1 40791 0 0 353317 0 0 0 0 224789828 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 190974333 41958554 123983334 43 0 224593 0 0 0 <more 0's deleted> 1574 ctxt 22848221062 1575 btime 1605316999 1576 processes 746787147 1577 procs_running 2 1578 procs_blocked 0 1579 softirq 12121874454 100099120 3938138295 127375644 2795979 187870761 0 173808342 3072582055 52608 224184354 1580 1581The very first "cpu" line aggregates the numbers in all of the other "cpuN" 1582lines. These numbers identify the amount of time the CPU has spent performing 1583different kinds of work. Time units are in USER_HZ (typically hundredths of a 1584second). The meanings of the columns are as follows, from left to right: 1585 1586- user: normal processes executing in user mode 1587- nice: niced processes executing in user mode 1588- system: processes executing in kernel mode 1589- idle: twiddling thumbs 1590- iowait: In a word, iowait stands for waiting for I/O to complete. But there 1591 are several problems: 1592 1593 1. CPU will not wait for I/O to complete, iowait is the time that a task is 1594 waiting for I/O to complete. When CPU goes into idle state for 1595 outstanding task I/O, another task will be scheduled on this CPU. 1596 2. In a multi-core CPU, the task waiting for I/O to complete is not running 1597 on any CPU, so the iowait of each CPU is difficult to calculate. 1598 3. The value of iowait field in /proc/stat will decrease in certain 1599 conditions. 1600 1601 So, the iowait is not reliable by reading from /proc/stat. 1602- irq: servicing interrupts 1603- softirq: servicing softirqs 1604- steal: involuntary wait 1605- guest: running a normal guest 1606- guest_nice: running a niced guest 1607 1608The "intr" line gives counts of interrupts serviced since boot time, for each 1609of the possible system interrupts. The first column is the total of all 1610interrupts serviced including unnumbered architecture specific interrupts; 1611each subsequent column is the total for that particular numbered interrupt. 1612Unnumbered interrupts are not shown, only summed into the total. 1613 1614The "ctxt" line gives the total number of context switches across all CPUs. 1615 1616The "btime" line gives the time at which the system booted, in seconds since 1617the Unix epoch. 1618 1619The "processes" line gives the number of processes and threads created, which 1620includes (but is not limited to) those created by calls to the fork() and 1621clone() system calls. 1622 1623The "procs_running" line gives the total number of threads that are 1624running or ready to run (i.e., the total number of runnable threads). 1625 1626The "procs_blocked" line gives the number of processes currently blocked, 1627waiting for I/O to complete. 1628 1629The "softirq" line gives counts of softirqs serviced since boot time, for each 1630of the possible system softirqs. The first column is the total of all 1631softirqs serviced; each subsequent column is the total for that particular 1632softirq. 1633 1634 16351.8 Ext4 file system parameters 1636------------------------------- 1637 1638Information about mounted ext4 file systems can be found in 1639/proc/fs/ext4. Each mounted filesystem will have a directory in 1640/proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or 1641/proc/fs/ext4/sda9 or /proc/fs/ext4/dm-0). The files in each per-device 1642directory are shown in Table 1-12, below. 1643 1644.. table:: Table 1-12: Files in /proc/fs/ext4/<devname> 1645 1646 ============== ========================================================== 1647 File Content 1648 mb_groups details of multiblock allocator buddy cache of free blocks 1649 ============== ========================================================== 1650 16511.9 /proc/consoles 1652------------------- 1653Shows registered system console lines. 1654 1655To see which character device lines are currently used for the system console 1656/dev/console, you may simply look into the file /proc/consoles:: 1657 1658 > cat /proc/consoles 1659 tty0 -WU (ECp) 4:7 1660 ttyS0 -W- (Ep) 4:64 1661 1662The columns are: 1663 1664+--------------------+-------------------------------------------------------+ 1665| device | name of the device | 1666+====================+=======================================================+ 1667| operations | * R = can do read operations | 1668| | * W = can do write operations | 1669| | * U = can do unblank | 1670+--------------------+-------------------------------------------------------+ 1671| flags | * E = it is enabled | 1672| | * C = it is preferred console | 1673| | * B = it is primary boot console | 1674| | * p = it is used for printk buffer | 1675| | * b = it is not a TTY but a Braille device | 1676| | * a = it is safe to use when cpu is offline | 1677+--------------------+-------------------------------------------------------+ 1678| major:minor | major and minor number of the device separated by a | 1679| | colon | 1680+--------------------+-------------------------------------------------------+ 1681 1682Summary 1683------- 1684 1685The /proc file system serves information about the running system. It not only 1686allows access to process data but also allows you to request the kernel status 1687by reading files in the hierarchy. 1688 1689The directory structure of /proc reflects the types of information and makes 1690it easy, if not obvious, where to look for specific data. 1691 1692Chapter 2: Modifying System Parameters 1693====================================== 1694 1695In This Chapter 1696--------------- 1697 1698* Modifying kernel parameters by writing into files found in /proc/sys 1699* Exploring the files which modify certain parameters 1700* Review of the /proc/sys file tree 1701 1702------------------------------------------------------------------------------ 1703 1704A very interesting part of /proc is the directory /proc/sys. This is not only 1705a source of information, it also allows you to change parameters within the 1706kernel. Be very careful when attempting this. You can optimize your system, 1707but you can also cause it to crash. Never alter kernel parameters on a 1708production system. Set up a development machine and test to make sure that 1709everything works the way you want it to. You may have no alternative but to 1710reboot the machine once an error has been made. 1711 1712To change a value, simply echo the new value into the file. 1713You need to be root to do this. You can create your own boot script 1714to perform this every time your system boots. 1715 1716The files in /proc/sys can be used to fine tune and monitor miscellaneous and 1717general things in the operation of the Linux kernel. Since some of the files 1718can inadvertently disrupt your system, it is advisable to read both 1719documentation and source before actually making adjustments. In any case, be 1720very careful when writing to any of these files. The entries in /proc may 1721change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt 1722review the kernel documentation in the directory linux/Documentation. 1723This chapter is heavily based on the documentation included in the pre 2.2 1724kernels, and became part of it in version 2.2.1 of the Linux kernel. 1725 1726Please see: Documentation/admin-guide/sysctl/ directory for descriptions of 1727these entries. 1728 1729Summary 1730------- 1731 1732Certain aspects of kernel behavior can be modified at runtime, without the 1733need to recompile the kernel, or even to reboot the system. The files in the 1734/proc/sys tree can not only be read, but also modified. You can use the echo 1735command to write value into these files, thereby changing the default settings 1736of the kernel. 1737 1738 1739Chapter 3: Per-process Parameters 1740================================= 1741 17423.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj- Adjust the oom-killer score 1743-------------------------------------------------------------------------------- 1744 1745These files can be used to adjust the badness heuristic used to select which 1746process gets killed in out of memory (oom) conditions. 1747 1748The badness heuristic assigns a value to each candidate task ranging from 0 1749(never kill) to 1000 (always kill) to determine which process is targeted. The 1750units are roughly a proportion along that range of allowed memory the process 1751may allocate from based on an estimation of its current memory and swap use. 1752For example, if a task is using all allowed memory, its badness score will be 17531000. If it is using half of its allowed memory, its score will be 500. 1754 1755The amount of "allowed" memory depends on the context in which the oom killer 1756was called. If it is due to the memory assigned to the allocating task's cpuset 1757being exhausted, the allowed memory represents the set of mems assigned to that 1758cpuset. If it is due to a mempolicy's node(s) being exhausted, the allowed 1759memory represents the set of mempolicy nodes. If it is due to a memory 1760limit (or swap limit) being reached, the allowed memory is that configured 1761limit. Finally, if it is due to the entire system being out of memory, the 1762allowed memory represents all allocatable resources. 1763 1764The value of /proc/<pid>/oom_score_adj is added to the badness score before it 1765is used to determine which task to kill. Acceptable values range from -1000 1766(OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_MAX). This allows userspace to 1767polarize the preference for oom killing either by always preferring a certain 1768task or completely disabling it. The lowest possible value, -1000, is 1769equivalent to disabling oom killing entirely for that task since it will always 1770report a badness score of 0. 1771 1772Consequently, it is very simple for userspace to define the amount of memory to 1773consider for each task. Setting a /proc/<pid>/oom_score_adj value of +500, for 1774example, is roughly equivalent to allowing the remainder of tasks sharing the 1775same system, cpuset, mempolicy, or memory controller resources to use at least 177650% more memory. A value of -500, on the other hand, would be roughly 1777equivalent to discounting 50% of the task's allowed memory from being considered 1778as scoring against the task. 1779 1780For backwards compatibility with previous kernels, /proc/<pid>/oom_adj may also 1781be used to tune the badness score. Its acceptable values range from -16 1782(OOM_ADJUST_MIN) to +15 (OOM_ADJUST_MAX) and a special value of -17 1783(OOM_DISABLE) to disable oom killing entirely for that task. Its value is 1784scaled linearly with /proc/<pid>/oom_score_adj. 1785 1786The value of /proc/<pid>/oom_score_adj may be reduced no lower than the last 1787value set by a CAP_SYS_RESOURCE process. To reduce the value any lower 1788requires CAP_SYS_RESOURCE. 1789 1790 17913.2 /proc/<pid>/oom_score - Display current oom-killer score 1792------------------------------------------------------------- 1793 1794This file can be used to check the current score used by the oom-killer for 1795any given <pid>. Use it together with /proc/<pid>/oom_score_adj to tune which 1796process should be killed in an out-of-memory situation. 1797 1798Please note that the exported value includes oom_score_adj so it is 1799effectively in range [0,2000]. 1800 1801 18023.3 /proc/<pid>/io - Display the IO accounting fields 1803------------------------------------------------------- 1804 1805This file contains IO statistics for each running process. 1806 1807Example 1808~~~~~~~ 1809 1810:: 1811 1812 test:/tmp # dd if=/dev/zero of=/tmp/test.dat & 1813 [1] 3828 1814 1815 test:/tmp # cat /proc/3828/io 1816 rchar: 323934931 1817 wchar: 323929600 1818 syscr: 632687 1819 syscw: 632675 1820 read_bytes: 0 1821 write_bytes: 323932160 1822 cancelled_write_bytes: 0 1823 1824 1825Description 1826~~~~~~~~~~~ 1827 1828rchar 1829^^^^^ 1830 1831I/O counter: chars read 1832The number of bytes which this task has caused to be read from storage. This 1833is simply the sum of bytes which this process passed to read() and pread(). 1834It includes things like tty IO and it is unaffected by whether or not actual 1835physical disk IO was required (the read might have been satisfied from 1836pagecache). 1837 1838 1839wchar 1840^^^^^ 1841 1842I/O counter: chars written 1843The number of bytes which this task has caused, or shall cause to be written 1844to disk. Similar caveats apply here as with rchar. 1845 1846 1847syscr 1848^^^^^ 1849 1850I/O counter: read syscalls 1851Attempt to count the number of read I/O operations, i.e. syscalls like read() 1852and pread(). 1853 1854 1855syscw 1856^^^^^ 1857 1858I/O counter: write syscalls 1859Attempt to count the number of write I/O operations, i.e. syscalls like 1860write() and pwrite(). 1861 1862 1863read_bytes 1864^^^^^^^^^^ 1865 1866I/O counter: bytes read 1867Attempt to count the number of bytes which this process really did cause to 1868be fetched from the storage layer. Done at the submit_bio() level, so it is 1869accurate for block-backed filesystems. <please add status regarding NFS and 1870CIFS at a later time> 1871 1872 1873write_bytes 1874^^^^^^^^^^^ 1875 1876I/O counter: bytes written 1877Attempt to count the number of bytes which this process caused to be sent to 1878the storage layer. This is done at page-dirtying time. 1879 1880 1881cancelled_write_bytes 1882^^^^^^^^^^^^^^^^^^^^^ 1883 1884The big inaccuracy here is truncate. If a process writes 1MB to a file and 1885then deletes the file, it will in fact perform no writeout. But it will have 1886been accounted as having caused 1MB of write. 1887In other words: The number of bytes which this process caused to not happen, 1888by truncating pagecache. A task can cause "negative" IO too. If this task 1889truncates some dirty pagecache, some IO which another task has been accounted 1890for (in its write_bytes) will not be happening. We _could_ just subtract that 1891from the truncating task's write_bytes, but there is information loss in doing 1892that. 1893 1894 1895.. Note:: 1896 1897 At its current implementation state, this is a bit racy on 32-bit machines: 1898 if process A reads process B's /proc/pid/io while process B is updating one 1899 of those 64-bit counters, process A could see an intermediate result. 1900 1901 1902More information about this can be found within the taskstats documentation in 1903Documentation/accounting. 1904 19053.4 /proc/<pid>/coredump_filter - Core dump filtering settings 1906--------------------------------------------------------------- 1907When a process is dumped, all anonymous memory is written to a core file as 1908long as the size of the core file isn't limited. But sometimes we don't want 1909to dump some memory segments, for example, huge shared memory or DAX. 1910Conversely, sometimes we want to save file-backed memory segments into a core 1911file, not only the individual files. 1912 1913/proc/<pid>/coredump_filter allows you to customize which memory segments 1914will be dumped when the <pid> process is dumped. coredump_filter is a bitmask 1915of memory types. If a bit of the bitmask is set, memory segments of the 1916corresponding memory type are dumped, otherwise they are not dumped. 1917 1918The following 9 memory types are supported: 1919 1920 - (bit 0) anonymous private memory 1921 - (bit 1) anonymous shared memory 1922 - (bit 2) file-backed private memory 1923 - (bit 3) file-backed shared memory 1924 - (bit 4) ELF header pages in file-backed private memory areas (it is 1925 effective only if the bit 2 is cleared) 1926 - (bit 5) hugetlb private memory 1927 - (bit 6) hugetlb shared memory 1928 - (bit 7) DAX private memory 1929 - (bit 8) DAX shared memory 1930 1931 Note that MMIO pages such as frame buffer are never dumped and vDSO pages 1932 are always dumped regardless of the bitmask status. 1933 1934 Note that bits 0-4 don't affect hugetlb or DAX memory. hugetlb memory is 1935 only affected by bit 5-6, and DAX is only affected by bits 7-8. 1936 1937The default value of coredump_filter is 0x33; this means all anonymous memory 1938segments, ELF header pages and hugetlb private memory are dumped. 1939 1940If you don't want to dump all shared memory segments attached to pid 1234, 1941write 0x31 to the process's proc file:: 1942 1943 $ echo 0x31 > /proc/1234/coredump_filter 1944 1945When a new process is created, the process inherits the bitmask status from its 1946parent. It is useful to set up coredump_filter before the program runs. 1947For example:: 1948 1949 $ echo 0x7 > /proc/self/coredump_filter 1950 $ ./some_program 1951 19523.5 /proc/<pid>/mountinfo - Information about mounts 1953-------------------------------------------------------- 1954 1955This file contains lines of the form:: 1956 1957 36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue 1958 (1)(2)(3) (4) (5) (6) (n…m) (m+1)(m+2) (m+3) (m+4) 1959 1960 (1) mount ID: unique identifier of the mount (may be reused after umount) 1961 (2) parent ID: ID of parent (or of self for the top of the mount tree) 1962 (3) major:minor: value of st_dev for files on filesystem 1963 (4) root: root of the mount within the filesystem 1964 (5) mount point: mount point relative to the process's root 1965 (6) mount options: per mount options 1966 (n…m) optional fields: zero or more fields of the form "tag[:value]" 1967 (m+1) separator: marks the end of the optional fields 1968 (m+2) filesystem type: name of filesystem of the form "type[.subtype]" 1969 (m+3) mount source: filesystem specific information or "none" 1970 (m+4) super options: per super block options 1971 1972Parsers should ignore all unrecognised optional fields. Currently the 1973possible optional fields are: 1974 1975================ ============================================================== 1976shared:X mount is shared in peer group X 1977master:X mount is slave to peer group X 1978propagate_from:X mount is slave and receives propagation from peer group X [#]_ 1979unbindable mount is unbindable 1980================ ============================================================== 1981 1982.. [#] X is the closest dominant peer group under the process's root. If 1983 X is the immediate master of the mount, or if there's no dominant peer 1984 group under the same root, then only the "master:X" field is present 1985 and not the "propagate_from:X" field. 1986 1987For more information on mount propagation see: 1988 1989 Documentation/filesystems/sharedsubtree.rst 1990 1991 19923.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm 1993-------------------------------------------------------- 1994These files provide a method to access a task's comm value. It also allows for 1995a task to set its own or one of its thread siblings comm value. The comm value 1996is limited in size compared to the cmdline value, so writing anything longer 1997then the kernel's TASK_COMM_LEN (currently 16 chars, including the NUL 1998terminator) will result in a truncated comm value. 1999 2000 20013.7 /proc/<pid>/task/<tid>/children - Information about task children 2002------------------------------------------------------------------------- 2003This file provides a fast way to retrieve first level children pids 2004of a task pointed by <pid>/<tid> pair. The format is a space separated 2005stream of pids. 2006 2007Note the "first level" here -- if a child has its own children they will 2008not be listed here; one needs to read /proc/<children-pid>/task/<tid>/children 2009to obtain the descendants. 2010 2011Since this interface is intended to be fast and cheap it doesn't 2012guarantee to provide precise results and some children might be 2013skipped, especially if they've exited right after we printed their 2014pids, so one needs to either stop or freeze processes being inspected 2015if precise results are needed. 2016 2017 20183.8 /proc/<pid>/fdinfo/<fd> - Information about opened file 2019--------------------------------------------------------------- 2020This file provides information associated with an opened file. The regular 2021files have at least four fields -- 'pos', 'flags', 'mnt_id' and 'ino'. 2022The 'pos' represents the current offset of the opened file in decimal 2023form [see lseek(2) for details], 'flags' denotes the octal O_xxx mask the 2024file has been created with [see open(2) for details] and 'mnt_id' represents 2025mount ID of the file system containing the opened file [see 3.5 2026/proc/<pid>/mountinfo for details]. 'ino' represents the inode number of 2027the file. 2028 2029A typical output is:: 2030 2031 pos: 0 2032 flags: 0100002 2033 mnt_id: 19 2034 ino: 63107 2035 2036All locks associated with a file descriptor are shown in its fdinfo too:: 2037 2038 lock: 1: FLOCK ADVISORY WRITE 359 00:13:11691 0 EOF 2039 2040The files such as eventfd, fsnotify, signalfd, epoll among the regular pos/flags 2041pair provide additional information particular to the objects they represent. 2042 2043Eventfd files 2044~~~~~~~~~~~~~ 2045 2046:: 2047 2048 pos: 0 2049 flags: 04002 2050 mnt_id: 9 2051 ino: 63107 2052 eventfd-count: 5a 2053 2054where 'eventfd-count' is hex value of a counter. 2055 2056Signalfd files 2057~~~~~~~~~~~~~~ 2058 2059:: 2060 2061 pos: 0 2062 flags: 04002 2063 mnt_id: 9 2064 ino: 63107 2065 sigmask: 0000000000000200 2066 2067where 'sigmask' is hex value of the signal mask associated 2068with a file. 2069 2070Epoll files 2071~~~~~~~~~~~ 2072 2073:: 2074 2075 pos: 0 2076 flags: 02 2077 mnt_id: 9 2078 ino: 63107 2079 tfd: 5 events: 1d data: ffffffffffffffff pos:0 ino:61af sdev:7 2080 2081where 'tfd' is a target file descriptor number in decimal form, 2082'events' is events mask being watched and the 'data' is data 2083associated with a target [see epoll(7) for more details]. 2084 2085The 'pos' is current offset of the target file in decimal form 2086[see lseek(2)], 'ino' and 'sdev' are inode and device numbers 2087where target file resides, all in hex format. 2088 2089Fsnotify files 2090~~~~~~~~~~~~~~ 2091For inotify files the format is the following:: 2092 2093 pos: 0 2094 flags: 02000000 2095 mnt_id: 9 2096 ino: 63107 2097 inotify wd:3 ino:9e7e sdev:800013 mask:800afce ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:7e9e0000640d1b6d 2098 2099where 'wd' is a watch descriptor in decimal form, i.e. a target file 2100descriptor number, 'ino' and 'sdev' are inode and device where the 2101target file resides and the 'mask' is the mask of events, all in hex 2102form [see inotify(7) for more details]. 2103 2104If the kernel was built with exportfs support, the path to the target 2105file is encoded as a file handle. The file handle is provided by three 2106fields 'fhandle-bytes', 'fhandle-type' and 'f_handle', all in hex 2107format. 2108 2109If the kernel is built without exportfs support the file handle won't be 2110printed out. 2111 2112If there is no inotify mark attached yet the 'inotify' line will be omitted. 2113 2114For fanotify files the format is:: 2115 2116 pos: 0 2117 flags: 02 2118 mnt_id: 9 2119 ino: 63107 2120 fanotify flags:10 event-flags:0 2121 fanotify mnt_id:12 mflags:40 mask:38 ignored_mask:40000003 2122 fanotify ino:4f969 sdev:800013 mflags:0 mask:3b ignored_mask:40000000 fhandle-bytes:8 fhandle-type:1 f_handle:69f90400c275b5b4 2123 2124where fanotify 'flags' and 'event-flags' are values used in fanotify_init 2125call, 'mnt_id' is the mount point identifier, 'mflags' is the value of 2126flags associated with mark which are tracked separately from events 2127mask. 'ino' and 'sdev' are target inode and device, 'mask' is the events 2128mask and 'ignored_mask' is the mask of events which are to be ignored. 2129All are in hex format. Incorporation of 'mflags', 'mask' and 'ignored_mask' 2130provide information about flags and mask used in fanotify_mark 2131call [see fsnotify manpage for details]. 2132 2133While the first three lines are mandatory and always printed, the rest is 2134optional and may be omitted if no marks created yet. 2135 2136Timerfd files 2137~~~~~~~~~~~~~ 2138 2139:: 2140 2141 pos: 0 2142 flags: 02 2143 mnt_id: 9 2144 ino: 63107 2145 clockid: 0 2146 ticks: 0 2147 settime flags: 01 2148 it_value: (0, 49406829) 2149 it_interval: (1, 0) 2150 2151where 'clockid' is the clock type and 'ticks' is the number of the timer expirations 2152that have occurred [see timerfd_create(2) for details]. 'settime flags' are 2153flags in octal form been used to setup the timer [see timerfd_settime(2) for 2154details]. 'it_value' is remaining time until the timer expiration. 2155'it_interval' is the interval for the timer. Note the timer might be set up 2156with TIMER_ABSTIME option which will be shown in 'settime flags', but 'it_value' 2157still exhibits timer's remaining time. 2158 2159DMA Buffer files 2160~~~~~~~~~~~~~~~~ 2161 2162:: 2163 2164 pos: 0 2165 flags: 04002 2166 mnt_id: 9 2167 ino: 63107 2168 size: 32768 2169 count: 2 2170 exp_name: system-heap 2171 2172where 'size' is the size of the DMA buffer in bytes. 'count' is the file count of 2173the DMA buffer file. 'exp_name' is the name of the DMA buffer exporter. 2174 2175VFIO Device files 2176~~~~~~~~~~~~~~~~~ 2177 2178:: 2179 2180 pos: 0 2181 flags: 02000002 2182 mnt_id: 17 2183 ino: 5122 2184 vfio-device-syspath: /sys/devices/pci0000:e0/0000:e0:01.1/0000:e1:00.0/0000:e2:05.0/0000:e8:00.0 2185 2186where 'vfio-device-syspath' is the sysfs path corresponding to the VFIO device 2187file. 2188 21893.9 /proc/<pid>/map_files - Information about memory mapped files 2190--------------------------------------------------------------------- 2191This directory contains symbolic links which represent memory mapped files 2192the process is maintaining. Example output:: 2193 2194 | lr-------- 1 root root 64 Jan 27 11:24 333c600000-333c620000 -> /usr/lib64/ld-2.18.so 2195 | lr-------- 1 root root 64 Jan 27 11:24 333c81f000-333c820000 -> /usr/lib64/ld-2.18.so 2196 | lr-------- 1 root root 64 Jan 27 11:24 333c820000-333c821000 -> /usr/lib64/ld-2.18.so 2197 | ... 2198 | lr-------- 1 root root 64 Jan 27 11:24 35d0421000-35d0422000 -> /usr/lib64/libselinux.so.1 2199 | lr-------- 1 root root 64 Jan 27 11:24 400000-41a000 -> /usr/bin/ls 2200 2201The name of a link represents the virtual memory bounds of a mapping, i.e. 2202vm_area_struct::vm_start-vm_area_struct::vm_end. 2203 2204The main purpose of the map_files is to retrieve a set of memory mapped 2205files in a fast way instead of parsing /proc/<pid>/maps or 2206/proc/<pid>/smaps, both of which contain many more records. At the same 2207time one can open(2) mappings from the listings of two processes and 2208comparing their inode numbers to figure out which anonymous memory areas 2209are actually shared. 2210 22113.10 /proc/<pid>/timerslack_ns - Task timerslack value 2212--------------------------------------------------------- 2213This file provides the value of the task's timerslack value in nanoseconds. 2214This value specifies an amount of time that normal timers may be deferred 2215in order to coalesce timers and avoid unnecessary wakeups. 2216 2217This allows a task's interactivity vs power consumption tradeoff to be 2218adjusted. 2219 2220Writing 0 to the file will set the task's timerslack to the default value. 2221 2222Valid values are from 0 - ULLONG_MAX 2223 2224An application setting the value must have PTRACE_MODE_ATTACH_FSCREDS level 2225permissions on the task specified to change its timerslack_ns value. 2226 22273.11 /proc/<pid>/patch_state - Livepatch patch operation state 2228----------------------------------------------------------------- 2229When CONFIG_LIVEPATCH is enabled, this file displays the value of the 2230patch state for the task. 2231 2232A value of '-1' indicates that no patch is in transition. 2233 2234A value of '0' indicates that a patch is in transition and the task is 2235unpatched. If the patch is being enabled, then the task hasn't been 2236patched yet. If the patch is being disabled, then the task has already 2237been unpatched. 2238 2239A value of '1' indicates that a patch is in transition and the task is 2240patched. If the patch is being enabled, then the task has already been 2241patched. If the patch is being disabled, then the task hasn't been 2242unpatched yet. 2243 22443.12 /proc/<pid>/arch_status - task architecture specific status 2245------------------------------------------------------------------- 2246When CONFIG_PROC_PID_ARCH_STATUS is enabled, this file displays the 2247architecture specific status of the task. 2248 2249Example 2250~~~~~~~ 2251 2252:: 2253 2254 $ cat /proc/6753/arch_status 2255 AVX512_elapsed_ms: 8 2256 2257Description 2258~~~~~~~~~~~ 2259 2260x86 specific entries 2261~~~~~~~~~~~~~~~~~~~~~ 2262 2263AVX512_elapsed_ms 2264^^^^^^^^^^^^^^^^^^ 2265 2266 If AVX512 is supported on the machine, this entry shows the milliseconds 2267 elapsed since the last time AVX512 usage was recorded. The recording 2268 happens on a best effort basis when a task is scheduled out. This means 2269 that the value depends on two factors: 2270 2271 1) The time which the task spent on the CPU without being scheduled 2272 out. With CPU isolation and a single runnable task this can take 2273 several seconds. 2274 2275 2) The time since the task was scheduled out last. Depending on the 2276 reason for being scheduled out (time slice exhausted, syscall ...) 2277 this can be arbitrary long time. 2278 2279 As a consequence the value cannot be considered precise and authoritative 2280 information. The application which uses this information has to be aware 2281 of the overall scenario on the system in order to determine whether a 2282 task is a real AVX512 user or not. Precise information can be obtained 2283 with performance counters. 2284 2285 A special value of '-1' indicates that no AVX512 usage was recorded, thus 2286 the task is unlikely an AVX512 user, but depends on the workload and the 2287 scheduling scenario, it also could be a false negative mentioned above. 2288 22893.13 /proc/<pid>/fd - List of symlinks to open files 2290------------------------------------------------------- 2291This directory contains symbolic links which represent open files 2292the process is maintaining. Example output:: 2293 2294 lr-x------ 1 root root 64 Sep 20 17:53 0 -> /dev/null 2295 l-wx------ 1 root root 64 Sep 20 17:53 1 -> /dev/null 2296 lrwx------ 1 root root 64 Sep 20 17:53 10 -> 'socket:[12539]' 2297 lrwx------ 1 root root 64 Sep 20 17:53 11 -> 'socket:[12540]' 2298 lrwx------ 1 root root 64 Sep 20 17:53 12 -> 'socket:[12542]' 2299 2300The number of open files for the process is stored in 'size' member 2301of stat() output for /proc/<pid>/fd for fast access. 2302------------------------------------------------------- 2303 23043.14 /proc/<pid>/ksm_stat - Information about the process's ksm status 2305---------------------------------------------------------------------- 2306When CONFIG_KSM is enabled, each process has this file which displays 2307the information of ksm merging status. 2308 2309Example 2310~~~~~~~ 2311 2312:: 2313 2314 / # cat /proc/self/ksm_stat 2315 ksm_rmap_items 0 2316 ksm_zero_pages 0 2317 ksm_merging_pages 0 2318 ksm_process_profit 0 2319 ksm_merge_any: no 2320 ksm_mergeable: no 2321 2322Description 2323~~~~~~~~~~~ 2324 2325ksm_rmap_items 2326^^^^^^^^^^^^^^ 2327 2328The number of ksm_rmap_item structures in use. The structure 2329ksm_rmap_item stores the reverse mapping information for virtual 2330addresses. KSM will generate a ksm_rmap_item for each ksm-scanned page of 2331the process. 2332 2333ksm_zero_pages 2334^^^^^^^^^^^^^^ 2335 2336When /sys/kernel/mm/ksm/use_zero_pages is enabled, it represent how many 2337empty pages are merged with kernel zero pages by KSM. 2338 2339ksm_merging_pages 2340^^^^^^^^^^^^^^^^^ 2341 2342It represents how many pages of this process are involved in KSM merging 2343(not including ksm_zero_pages). It is the same with what 2344/proc/<pid>/ksm_merging_pages shows. 2345 2346ksm_process_profit 2347^^^^^^^^^^^^^^^^^^ 2348 2349The profit that KSM brings (Saved bytes). KSM can save memory by merging 2350identical pages, but also can consume additional memory, because it needs 2351to generate a number of rmap_items to save each scanned page's brief rmap 2352information. Some of these pages may be merged, but some may not be abled 2353to be merged after being checked several times, which are unprofitable 2354memory consumed. 2355 2356ksm_merge_any 2357^^^^^^^^^^^^^ 2358 2359It specifies whether the process's 'mm is added by prctl() into the 2360candidate list of KSM or not, and if KSM scanning is fully enabled at 2361process level. 2362 2363ksm_mergeable 2364^^^^^^^^^^^^^ 2365 2366It specifies whether any VMAs of the process''s mms are currently 2367applicable to KSM. 2368 2369More information about KSM can be found in 2370Documentation/admin-guide/mm/ksm.rst. 2371 2372 2373Chapter 4: Configuring procfs 2374============================= 2375 23764.1 Mount options 2377--------------------- 2378 2379The following mount options are supported: 2380 2381 ========= ======================================================== 2382 hidepid= Set /proc/<pid>/ access mode. 2383 gid= Set the group authorized to learn processes information. 2384 subset= Show only the specified subset of procfs. 2385 pidns= Specify a the namespace used by this procfs. 2386 ========= ======================================================== 2387 2388hidepid=off or hidepid=0 means classic mode - everybody may access all 2389/proc/<pid>/ directories (default). 2390 2391hidepid=noaccess or hidepid=1 means users may not access any /proc/<pid>/ 2392directories but their own. Sensitive files like cmdline, sched*, status are now 2393protected against other users. This makes it impossible to learn whether any 2394user runs specific program (given the program doesn't reveal itself by its 2395behaviour). As an additional bonus, as /proc/<pid>/cmdline is unaccessible for 2396other users, poorly written programs passing sensitive information via program 2397arguments are now protected against local eavesdroppers. 2398 2399hidepid=invisible or hidepid=2 means hidepid=1 plus all /proc/<pid>/ will be 2400fully invisible to other users. It doesn't mean that it hides a fact whether a 2401process with a specific pid value exists (it can be learned by other means, e.g. 2402by "kill -0 $PID"), but it hides process's uid and gid, which may be learned by 2403stat()'ing /proc/<pid>/ otherwise. It greatly complicates an intruder's task of 2404gathering information about running processes, whether some daemon runs with 2405elevated privileges, whether other user runs some sensitive program, whether 2406other users run any program at all, etc. 2407 2408hidepid=ptraceable or hidepid=4 means that procfs should only contain 2409/proc/<pid>/ directories that the caller can ptrace. 2410 2411gid= defines a group authorized to learn processes information otherwise 2412prohibited by hidepid=. If you use some daemon like identd which needs to learn 2413information about processes information, just add identd to this group. 2414 2415subset=pid hides all top level files and directories in the procfs that 2416are not related to tasks. 2417 2418pidns= specifies a pid namespace (either as a string path to something like 2419`/proc/$pid/ns/pid`, or a file descriptor when using `FSCONFIG_SET_FD`) that 2420will be used by the procfs instance when translating pids. By default, procfs 2421will use the calling process's active pid namespace. Note that the pid 2422namespace of an existing procfs instance cannot be modified (attempting to do 2423so will give an `-EBUSY` error). 2424 2425Chapter 5: Filesystem behavior 2426============================== 2427 2428Originally, before the advent of pid namespace, procfs was a global file 2429system. It means that there was only one procfs instance in the system. 2430 2431When pid namespace was added, a separate procfs instance was mounted in 2432each pid namespace. So, procfs mount options are global among all 2433mountpoints within the same namespace:: 2434 2435 # grep ^proc /proc/mounts 2436 proc /proc proc rw,relatime,hidepid=2 0 0 2437 2438 # strace -e mount mount -o hidepid=1 -t proc proc /tmp/proc 2439 mount("proc", "/tmp/proc", "proc", 0, "hidepid=1") = 0 2440 +++ exited with 0 +++ 2441 2442 # grep ^proc /proc/mounts 2443 proc /proc proc rw,relatime,hidepid=2 0 0 2444 proc /tmp/proc proc rw,relatime,hidepid=2 0 0 2445 2446and only after remounting procfs mount options will change at all 2447mountpoints:: 2448 2449 # mount -o remount,hidepid=1 -t proc proc /tmp/proc 2450 2451 # grep ^proc /proc/mounts 2452 proc /proc proc rw,relatime,hidepid=1 0 0 2453 proc /tmp/proc proc rw,relatime,hidepid=1 0 0 2454 2455This behavior is different from the behavior of other filesystems. 2456 2457The new procfs behavior is more like other filesystems. Each procfs mount 2458creates a new procfs instance. Mount options affect own procfs instance. 2459It means that it became possible to have several procfs instances 2460displaying tasks with different filtering options in one pid namespace:: 2461 2462 # mount -o hidepid=invisible -t proc proc /proc 2463 # mount -o hidepid=noaccess -t proc proc /tmp/proc 2464 # grep ^proc /proc/mounts 2465 proc /proc proc rw,relatime,hidepid=invisible 0 0 2466 proc /tmp/proc proc rw,relatime,hidepid=noaccess 0 0 2467