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 275 Threads number of threads 276 SigQ number of signals queued/max. number for queue 277 SigPnd bitmap of pending signals for the thread 278 ShdPnd bitmap of shared pending signals for the process 279 SigBlk bitmap of blocked signals 280 SigIgn bitmap of ignored signals 281 SigCgt bitmap of caught signals 282 CapInh bitmap of inheritable capabilities 283 CapPrm bitmap of permitted capabilities 284 CapEff bitmap of effective capabilities 285 CapBnd bitmap of capabilities bounding set 286 CapAmb bitmap of ambient capabilities 287 NoNewPrivs no_new_privs, like prctl(PR_GET_NO_NEW_PRIV, ...) 288 Seccomp seccomp mode, like prctl(PR_GET_SECCOMP, ...) 289 Speculation_Store_Bypass speculative store bypass mitigation status 290 SpeculationIndirectBranch indirect branch speculation mode 291 Cpus_allowed mask of CPUs on which this process may run 292 Cpus_allowed_list Same as previous, but in "list format" 293 Mems_allowed mask of memory nodes allowed to this process 294 Mems_allowed_list Same as previous, but in "list format" 295 voluntary_ctxt_switches number of voluntary context switches 296 nonvoluntary_ctxt_switches number of non voluntary context switches 297 ========================== =================================================== 298 299 300.. table:: Table 1-3: Contents of the statm fields (as of 2.6.8-rc3) 301 302 ======== =============================== ============================== 303 Field Content 304 ======== =============================== ============================== 305 size total program size (pages) (same as VmSize in status) 306 resident size of memory portions (pages) (same as VmRSS in status) 307 shared number of pages that are shared (i.e. backed by a file, same 308 as RssFile+RssShmem in status) 309 trs number of pages that are 'code' (not including libs; broken, 310 includes data segment) 311 lrs number of pages of library (always 0 on 2.6) 312 drs number of pages of data/stack (including libs; broken, 313 includes library text) 314 dt number of dirty pages (always 0 on 2.6) 315 ======== =============================== ============================== 316 317 318.. table:: Table 1-4: Contents of the stat fields (as of 2.6.30-rc7) 319 320 ============= =============================================================== 321 Field Content 322 ============= =============================================================== 323 pid process id 324 tcomm filename of the executable 325 state state (R is running, S is sleeping, D is sleeping in an 326 uninterruptible wait, Z is zombie, T is traced or stopped) 327 ppid process id of the parent process 328 pgrp pgrp of the process 329 sid session id 330 tty_nr tty the process uses 331 tty_pgrp pgrp of the tty 332 flags task flags 333 min_flt number of minor faults 334 cmin_flt number of minor faults with child's 335 maj_flt number of major faults 336 cmaj_flt number of major faults with child's 337 utime user mode jiffies 338 stime kernel mode jiffies 339 cutime user mode jiffies with child's 340 cstime kernel mode jiffies with child's 341 priority priority level 342 nice nice level 343 num_threads number of threads 344 it_real_value (obsolete, always 0) 345 start_time time the process started after system boot 346 vsize virtual memory size 347 rss resident set memory size 348 rsslim current limit in bytes on the rss 349 start_code address above which program text can run 350 end_code address below which program text can run 351 start_stack address of the start of the main process stack 352 esp current value of ESP 353 eip current value of EIP 354 pending bitmap of pending signals 355 blocked bitmap of blocked signals 356 sigign bitmap of ignored signals 357 sigcatch bitmap of caught signals 358 0 (place holder, used to be the wchan address, 359 use /proc/PID/wchan instead) 360 0 (place holder) 361 0 (place holder) 362 exit_signal signal to send to parent thread on exit 363 task_cpu which CPU the task is scheduled on 364 rt_priority realtime priority 365 policy scheduling policy (man sched_setscheduler) 366 blkio_ticks time spent waiting for block IO 367 gtime guest time of the task in jiffies 368 cgtime guest time of the task children in jiffies 369 start_data address above which program data+bss is placed 370 end_data address below which program data+bss is placed 371 start_brk address above which program heap can be expanded with brk() 372 arg_start address above which program command line is placed 373 arg_end address below which program command line is placed 374 env_start address above which program environment is placed 375 env_end address below which program environment is placed 376 exit_code the thread's exit_code in the form reported by the waitpid 377 system call 378 ============= =============================================================== 379 380The /proc/PID/maps file contains the currently mapped memory regions and 381their access permissions. 382 383The format is:: 384 385 address perms offset dev inode pathname 386 387 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test 388 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test 389 0804a000-0806b000 rw-p 00000000 00:00 0 [heap] 390 a7cb1000-a7cb2000 ---p 00000000 00:00 0 391 a7cb2000-a7eb2000 rw-p 00000000 00:00 0 392 a7eb2000-a7eb3000 ---p 00000000 00:00 0 393 a7eb3000-a7ed5000 rw-p 00000000 00:00 0 394 a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6 395 a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6 396 a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6 397 a800b000-a800e000 rw-p 00000000 00:00 0 398 a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0 399 a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0 400 a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0 401 a8024000-a8027000 rw-p 00000000 00:00 0 402 a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2 403 a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2 404 a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2 405 aff35000-aff4a000 rw-p 00000000 00:00 0 [stack] 406 ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso] 407 408where "address" is the address space in the process that it occupies, "perms" 409is a set of permissions:: 410 411 r = read 412 w = write 413 x = execute 414 s = shared 415 p = private (copy on write) 416 417"offset" is the offset into the mapping, "dev" is the device (major:minor), and 418"inode" is the inode on that device. 0 indicates that no inode is associated 419with the memory region, as the case would be with BSS (uninitialized data). 420The "pathname" shows the name associated file for this mapping. If the mapping 421is not associated with a file: 422 423 =================== =========================================== 424 [heap] the heap of the program 425 [stack] the stack of the main process 426 [vdso] the "virtual dynamic shared object", 427 the kernel system call handler 428 [anon:<name>] a private anonymous mapping that has been 429 named by userspace 430 [anon_shmem:<name>] an anonymous shared memory mapping that has 431 been named by userspace 432 =================== =========================================== 433 434 or if empty, the mapping is anonymous. 435 436Starting with 6.11 kernel, /proc/PID/maps provides an alternative 437ioctl()-based API that gives ability to flexibly and efficiently query and 438filter individual VMAs. This interface is binary and is meant for more 439efficient and easy programmatic use. `struct procmap_query`, defined in 440linux/fs.h UAPI header, serves as an input/output argument to the 441`PROCMAP_QUERY` ioctl() command. See comments in linus/fs.h UAPI header for 442details on query semantics, supported flags, data returned, and general API 443usage information. 444 445The /proc/PID/smaps is an extension based on maps, showing the memory 446consumption for each of the process's mappings. For each mapping (aka Virtual 447Memory Area, or VMA) there is a series of lines such as the following:: 448 449 08048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash 450 451 Size: 1084 kB 452 KernelPageSize: 4 kB 453 MMUPageSize: 4 kB 454 Rss: 892 kB 455 Pss: 374 kB 456 Pss_Dirty: 0 kB 457 Shared_Clean: 892 kB 458 Shared_Dirty: 0 kB 459 Private_Clean: 0 kB 460 Private_Dirty: 0 kB 461 Referenced: 892 kB 462 Anonymous: 0 kB 463 KSM: 0 kB 464 LazyFree: 0 kB 465 AnonHugePages: 0 kB 466 ShmemPmdMapped: 0 kB 467 Shared_Hugetlb: 0 kB 468 Private_Hugetlb: 0 kB 469 Swap: 0 kB 470 SwapPss: 0 kB 471 KernelPageSize: 4 kB 472 MMUPageSize: 4 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 size of each page allocated when backing a VMA 480(KernelPageSize), which is usually the same as the size in the page table 481entries; the page size used by the MMU when backing a VMA (in most cases, 482the same as KernelPageSize); the amount of the mapping that is currently 483resident in RAM (RSS); the process's proportional share of this mapping 484(PSS); and the number of clean and dirty shared and private pages in the 485mapping. 486 487The "proportional set size" (PSS) of a process is the count of pages it has 488in memory, where each page is divided by the number of processes sharing it. 489So if a process has 1000 pages all to itself, and 1000 shared with one other 490process, its PSS will be 1500. "Pss_Dirty" is the portion of PSS which 491consists of dirty pages. ("Pss_Clean" is not included, but it can be 492calculated by subtracting "Pss_Dirty" from "Pss".) 493 494Traditionally, a page is accounted as "private" if it is mapped exactly once, 495and a page is accounted as "shared" when mapped multiple times, even when 496mapped in the same process multiple times. Note that this accounting is 497independent of MAP_SHARED. 498 499In some kernel configurations, the semantics of pages part of a larger 500allocation (e.g., THP) can differ: a page is accounted as "private" if all 501pages part of the corresponding large allocation are *certainly* mapped in the 502same process, even if the page is mapped multiple times in that process. A 503page is accounted as "shared" if any page page of the larger allocation 504is *maybe* mapped in a different process. In some cases, a large allocation 505might be treated as "maybe mapped by multiple processes" even though this 506is no longer the case. 507 508Some kernel configurations do not track the precise number of times a page part 509of a larger allocation is mapped. In this case, when calculating the PSS, the 510average number of mappings per page in this larger allocation might be used 511as an approximation for the number of mappings of a page. The PSS calculation 512will be imprecise in this case. 513 514"Referenced" indicates the amount of memory currently marked as referenced or 515accessed. 516 517"Anonymous" shows the amount of memory that does not belong to any file. Even 518a mapping associated with a file may contain anonymous pages: when MAP_PRIVATE 519and a page is modified, the file page is replaced by a private anonymous copy. 520 521"KSM" reports how many of the pages are KSM pages. Note that KSM-placed zeropages 522are not included, only actual KSM pages. 523 524"LazyFree" shows the amount of memory which is marked by madvise(MADV_FREE). 525The memory isn't freed immediately with madvise(). It's freed in memory 526pressure if the memory is clean. Please note that the printed value might 527be lower than the real value due to optimizations used in the current 528implementation. If this is not desirable please file a bug report. 529 530"AnonHugePages" shows the amount of memory backed by transparent hugepage. 531 532"ShmemPmdMapped" shows the amount of shared (shmem/tmpfs) memory backed by 533huge pages. 534 535"Shared_Hugetlb" and "Private_Hugetlb" show the amounts of memory backed by 536hugetlbfs page which is *not* counted in "RSS" or "PSS" field for historical 537reasons. And these are not included in {Shared,Private}_{Clean,Dirty} field. 538 539"Swap" shows how much would-be-anonymous memory is also used, but out on swap. 540 541For shmem mappings, "Swap" includes also the size of the mapped (and not 542replaced by copy-on-write) part of the underlying shmem object out on swap. 543"SwapPss" shows proportional swap share of this mapping. Unlike "Swap", this 544does not take into account swapped out page of underlying shmem objects. 545"Locked" indicates whether the mapping is locked in memory or not. 546 547"THPeligible" indicates whether the mapping is eligible for allocating 548naturally aligned THP pages of any currently enabled size. 1 if true, 0 549otherwise. 550 551"VmFlags" field deserves a separate description. This member represents the 552kernel flags associated with the particular virtual memory area in two letter 553encoded manner. The codes are the following: 554 555 == ======================================= 556 rd readable 557 wr writeable 558 ex executable 559 sh shared 560 mr may read 561 mw may write 562 me may execute 563 ms may share 564 gd stack segment growns down 565 pf pure PFN range 566 lo pages are locked in memory 567 io memory mapped I/O area 568 sr sequential read advise provided 569 rr random read advise provided 570 dc do not copy area on fork 571 de do not expand area on remapping 572 ac area is accountable 573 nr swap space is not reserved for the area 574 ht area uses huge tlb pages 575 sf synchronous page fault 576 ar architecture specific flag 577 wf wipe on fork 578 dd do not include area into core dump 579 sd soft dirty flag 580 mm mixed map area 581 hg huge page advise flag 582 nh no huge page advise flag 583 mg mergeable advise flag 584 bt arm64 BTI guarded page 585 mt arm64 MTE allocation tags are enabled 586 um userfaultfd missing tracking 587 uw userfaultfd wr-protect tracking 588 ui userfaultfd minor fault 589 ss shadow/guarded control stack page 590 sl sealed 591 lf lock on fault pages 592 dp always lazily freeable mapping 593 == ======================================= 594 595Note that there is no guarantee that every flag and associated mnemonic will 596be present in all further kernel releases. Things get changed, the flags may 597be vanished or the reverse -- new added. Interpretation of their meaning 598might change in future as well. So each consumer of these flags has to 599follow each specific kernel version for the exact semantic. 600 601This file is only present if the CONFIG_MMU kernel configuration option is 602enabled. 603 604Note: reading /proc/PID/maps or /proc/PID/smaps is inherently racy (consistent 605output can be achieved only in the single read call). 606 607This typically manifests when doing partial reads of these files while the 608memory map is being modified. Despite the races, we do provide the following 609guarantees: 610 6111) The mapped addresses never go backwards, which implies no two 612 regions will ever overlap. 6132) If there is something at a given vaddr during the entirety of the 614 life of the smaps/maps walk, there will be some output for it. 615 616The /proc/PID/smaps_rollup file includes the same fields as /proc/PID/smaps, 617but their values are the sums of the corresponding values for all mappings of 618the process. Additionally, it contains these fields: 619 620- Pss_Anon 621- Pss_File 622- Pss_Shmem 623 624They represent the proportional shares of anonymous, file, and shmem pages, as 625described for smaps above. These fields are omitted in smaps since each 626mapping identifies the type (anon, file, or shmem) of all pages it contains. 627Thus all information in smaps_rollup can be derived from smaps, but at a 628significantly higher cost. 629 630The /proc/PID/clear_refs is used to reset the PG_Referenced and ACCESSED/YOUNG 631bits on both physical and virtual pages associated with a process, and the 632soft-dirty bit on pte (see Documentation/admin-guide/mm/soft-dirty.rst 633for details). 634To clear the bits for all the pages associated with the process:: 635 636 > echo 1 > /proc/PID/clear_refs 637 638To clear the bits for the anonymous pages associated with the process:: 639 640 > echo 2 > /proc/PID/clear_refs 641 642To clear the bits for the file mapped pages associated with the process:: 643 644 > echo 3 > /proc/PID/clear_refs 645 646To clear the soft-dirty bit:: 647 648 > echo 4 > /proc/PID/clear_refs 649 650To reset the peak resident set size ("high water mark") to the process's 651current value:: 652 653 > echo 5 > /proc/PID/clear_refs 654 655Any other value written to /proc/PID/clear_refs will have no effect. 656 657The /proc/pid/pagemap gives the PFN, which can be used to find the pageflags 658using /proc/kpageflags and number of times a page is mapped using 659/proc/kpagecount. For detailed explanation, see 660Documentation/admin-guide/mm/pagemap.rst. 661 662The /proc/pid/numa_maps is an extension based on maps, showing the memory 663locality and binding policy, as well as the memory usage (in pages) of 664each mapping. The output follows a general format where mapping details get 665summarized separated by blank spaces, one mapping per each file line:: 666 667 address policy mapping details 668 669 00400000 default file=/usr/local/bin/app mapped=1 active=0 N3=1 kernelpagesize_kB=4 670 00600000 default file=/usr/local/bin/app anon=1 dirty=1 N3=1 kernelpagesize_kB=4 671 3206000000 default file=/lib64/ld-2.12.so mapped=26 mapmax=6 N0=24 N3=2 kernelpagesize_kB=4 672 320621f000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 673 3206220000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 674 3206221000 default anon=1 dirty=1 N3=1 kernelpagesize_kB=4 675 3206800000 default file=/lib64/libc-2.12.so mapped=59 mapmax=21 active=55 N0=41 N3=18 kernelpagesize_kB=4 676 320698b000 default file=/lib64/libc-2.12.so 677 3206b8a000 default file=/lib64/libc-2.12.so anon=2 dirty=2 N3=2 kernelpagesize_kB=4 678 3206b8e000 default file=/lib64/libc-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 679 3206b8f000 default anon=3 dirty=3 active=1 N3=3 kernelpagesize_kB=4 680 7f4dc10a2000 default anon=3 dirty=3 N3=3 kernelpagesize_kB=4 681 7f4dc10b4000 default anon=2 dirty=2 active=1 N3=2 kernelpagesize_kB=4 682 7f4dc1200000 default file=/anon_hugepage\040(deleted) huge anon=1 dirty=1 N3=1 kernelpagesize_kB=2048 683 7fff335f0000 default stack anon=3 dirty=3 N3=3 kernelpagesize_kB=4 684 7fff3369d000 default mapped=1 mapmax=35 active=0 N3=1 kernelpagesize_kB=4 685 686Where: 687 688"address" is the starting address for the mapping; 689 690"policy" reports the NUMA memory policy set for the mapping (see Documentation/admin-guide/mm/numa_memory_policy.rst); 691 692"mapping details" summarizes mapping data such as mapping type, page usage counters, 693node locality page counters (N0 == node0, N1 == node1, ...) and the kernel page 694size, in KB, that is backing the mapping up. 695 696Note that some kernel configurations do not track the precise number of times 697a page part of a larger allocation (e.g., THP) is mapped. In these 698configurations, "mapmax" might corresponds to the average number of mappings 699per page in such a larger allocation instead. 700 7011.2 Kernel data 702--------------- 703 704Similar to the process entries, the kernel data files give information about 705the running kernel. The files used to obtain this information are contained in 706/proc and are listed in Table 1-5. Not all of these will be present in your 707system. It depends on the kernel configuration and the loaded modules, which 708files are there, and which are missing. 709 710.. table:: Table 1-5: Kernel info in /proc 711 712 ============ =============================================================== 713 File Content 714 ============ =============================================================== 715 allocinfo Memory allocations profiling information 716 apm Advanced power management info 717 bootconfig Kernel command line obtained from boot config, 718 and, if there were kernel parameters from the 719 boot loader, a "# Parameters from bootloader:" 720 line followed by a line containing those 721 parameters prefixed by "# ". (5.5) 722 buddyinfo Kernel memory allocator information (see text) (2.5) 723 bus Directory containing bus specific information 724 cmdline Kernel command line, both from bootloader and embedded 725 in the kernel image 726 cpuinfo Info about the CPU 727 devices Available devices (block and character) 728 dma Used DMS channels 729 filesystems Supported filesystems 730 driver Various drivers grouped here, currently rtc (2.4) 731 execdomains Execdomains, related to security (2.4) 732 fb Frame Buffer devices (2.4) 733 fs File system parameters, currently nfs/exports (2.4) 734 ide Directory containing info about the IDE subsystem 735 interrupts Interrupt usage 736 iomem Memory map (2.4) 737 ioports I/O port usage 738 irq Masks for irq to cpu affinity (2.4)(smp?) 739 isapnp ISA PnP (Plug&Play) Info (2.4) 740 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4)) 741 kmsg Kernel messages 742 ksyms Kernel symbol table 743 loadavg Load average of last 1, 5 & 15 minutes; 744 number of processes currently runnable (running or on ready queue); 745 total number of processes in system; 746 last pid created. 747 All fields are separated by one space except "number of 748 processes currently runnable" and "total number of processes 749 in system", which are separated by a slash ('/'). Example: 750 0.61 0.61 0.55 3/828 22084 751 locks Kernel locks 752 meminfo Memory info 753 misc Miscellaneous 754 modules List of loaded modules 755 mounts Mounted filesystems 756 net Networking info (see text) 757 pagetypeinfo Additional page allocator information (see text) (2.5) 758 partitions Table of partitions known to the system 759 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/, 760 decoupled by lspci (2.4) 761 rtc Real time clock 762 scsi SCSI info (see text) 763 slabinfo Slab pool info 764 softirqs softirq usage 765 stat Overall statistics 766 swaps Swap space utilization 767 sys See chapter 2 768 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4) 769 tty Info of tty drivers 770 uptime Wall clock since boot, combined idle time of all cpus 771 version Kernel version 772 video bttv info of video resources (2.4) 773 vmallocinfo Show vmalloced areas 774 ============ =============================================================== 775 776You can, for example, check which interrupts are currently in use and what 777they are used for by looking in the file /proc/interrupts:: 778 779 > cat /proc/interrupts 780 CPU0 781 0: 8728810 XT-PIC timer 782 1: 895 XT-PIC keyboard 783 2: 0 XT-PIC cascade 784 3: 531695 XT-PIC aha152x 785 4: 2014133 XT-PIC serial 786 5: 44401 XT-PIC pcnet_cs 787 8: 2 XT-PIC rtc 788 11: 8 XT-PIC i82365 789 12: 182918 XT-PIC PS/2 Mouse 790 13: 1 XT-PIC fpu 791 14: 1232265 XT-PIC ide0 792 15: 7 XT-PIC ide1 793 NMI: 0 794 795In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the 796output of a SMP machine):: 797 798 > cat /proc/interrupts 799 800 CPU0 CPU1 801 0: 1243498 1214548 IO-APIC-edge timer 802 1: 8949 8958 IO-APIC-edge keyboard 803 2: 0 0 XT-PIC cascade 804 5: 11286 10161 IO-APIC-edge soundblaster 805 8: 1 0 IO-APIC-edge rtc 806 9: 27422 27407 IO-APIC-edge 3c503 807 12: 113645 113873 IO-APIC-edge PS/2 Mouse 808 13: 0 0 XT-PIC fpu 809 14: 22491 24012 IO-APIC-edge ide0 810 15: 2183 2415 IO-APIC-edge ide1 811 17: 30564 30414 IO-APIC-level eth0 812 18: 177 164 IO-APIC-level bttv 813 NMI: 2457961 2457959 814 LOC: 2457882 2457881 815 ERR: 2155 816 817NMI is incremented in this case because every timer interrupt generates a NMI 818(Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups. 819 820LOC is the local interrupt counter of the internal APIC of every CPU. 821 822ERR is incremented in the case of errors in the IO-APIC bus (the bus that 823connects the CPUs in a SMP system. This means that an error has been detected, 824the IO-APIC automatically retry the transmission, so it should not be a big 825problem, but you should read the SMP-FAQ. 826 827In 2.6.2* /proc/interrupts was expanded again. This time the goal was for 828/proc/interrupts to display every IRQ vector in use by the system, not 829just those considered 'most important'. The new vectors are: 830 831THR 832 interrupt raised when a machine check threshold counter 833 (typically counting ECC corrected errors of memory or cache) exceeds 834 a configurable threshold. Only available on some systems. 835 836TRM 837 a thermal event interrupt occurs when a temperature threshold 838 has been exceeded for the CPU. This interrupt may also be generated 839 when the temperature drops back to normal. 840 841SPU 842 a spurious interrupt is some interrupt that was raised then lowered 843 by some IO device before it could be fully processed by the APIC. Hence 844 the APIC sees the interrupt but does not know what device it came from. 845 For this case the APIC will generate the interrupt with a IRQ vector 846 of 0xff. This might also be generated by chipset bugs. 847 848RES, CAL, TLB 849 rescheduling, call and TLB flush interrupts are 850 sent from one CPU to another per the needs of the OS. Typically, 851 their statistics are used by kernel developers and interested users to 852 determine the occurrence of interrupts of the given type. 853 854The above IRQ vectors are displayed only when relevant. For example, 855the threshold vector does not exist on x86_64 platforms. Others are 856suppressed when the system is a uniprocessor. As of this writing, only 857i386 and x86_64 platforms support the new IRQ vector displays. 858 859Of some interest is the introduction of the /proc/irq directory to 2.4. 860It could be used to set IRQ to CPU affinity. This means that you can "hook" an 861IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the 862irq subdir is one subdir for each IRQ, and two files; default_smp_affinity and 863prof_cpu_mask. 864 865For example:: 866 867 > ls /proc/irq/ 868 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask 869 1 11 13 15 17 19 3 5 7 9 default_smp_affinity 870 > ls /proc/irq/0/ 871 smp_affinity 872 873smp_affinity is a bitmask, in which you can specify which CPUs can handle the 874IRQ. You can set it by doing:: 875 876 > echo 1 > /proc/irq/10/smp_affinity 877 878This means that only the first CPU will handle the IRQ, but you can also echo 8795 which means that only the first and third CPU can handle the IRQ. 880 881The contents of each smp_affinity file is the same by default:: 882 883 > cat /proc/irq/0/smp_affinity 884 ffffffff 885 886There is an alternate interface, smp_affinity_list which allows specifying 887a CPU range instead of a bitmask:: 888 889 > cat /proc/irq/0/smp_affinity_list 890 1024-1031 891 892The default_smp_affinity mask applies to all non-active IRQs, which are the 893IRQs which have not yet been allocated/activated, and hence which lack a 894/proc/irq/[0-9]* directory. 895 896The node file on an SMP system shows the node to which the device using the IRQ 897reports itself as being attached. This hardware locality information does not 898include information about any possible driver locality preference. 899 900prof_cpu_mask specifies which CPUs are to be profiled by the system wide 901profiler. Default value is ffffffff (all CPUs if there are only 32 of them). 902 903The way IRQs are routed is handled by the IO-APIC, and it's Round Robin 904between all the CPUs which are allowed to handle it. As usual the kernel has 905more info than you and does a better job than you, so the defaults are the 906best choice for almost everyone. [Note this applies only to those IO-APIC's 907that support "Round Robin" interrupt distribution.] 908 909There are three more important subdirectories in /proc: net, scsi, and sys. 910The general rule is that the contents, or even the existence of these 911directories, depend on your kernel configuration. If SCSI is not enabled, the 912directory scsi may not exist. The same is true with the net, which is there 913only when networking support is present in the running kernel. 914 915The slabinfo file gives information about memory usage at the slab level. 916Linux uses slab pools for memory management above page level in version 2.2. 917Commonly used objects have their own slab pool (such as network buffers, 918directory cache, and so on). 919 920:: 921 922 > cat /proc/buddyinfo 923 924 Node 0, zone DMA 0 4 5 4 4 3 ... 925 Node 0, zone Normal 1 0 0 1 101 8 ... 926 Node 0, zone HighMem 2 0 0 1 1 0 ... 927 928External fragmentation is a problem under some workloads, and buddyinfo is a 929useful tool for helping diagnose these problems. Buddyinfo will give you a 930clue as to how big an area you can safely allocate, or why a previous 931allocation failed. 932 933Each column represents the number of pages of a certain order which are 934available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in 935ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE 936available in ZONE_NORMAL, etc... 937 938More information relevant to external fragmentation can be found in 939pagetypeinfo:: 940 941 > cat /proc/pagetypeinfo 942 Page block order: 9 943 Pages per block: 512 944 945 Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 946 Node 0, zone DMA, type Unmovable 0 0 0 1 1 1 1 1 1 1 0 947 Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 948 Node 0, zone DMA, type Movable 1 1 2 1 2 1 1 0 1 0 2 949 Node 0, zone DMA, type Reserve 0 0 0 0 0 0 0 0 0 1 0 950 Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 951 Node 0, zone DMA32, type Unmovable 103 54 77 1 1 1 11 8 7 1 9 952 Node 0, zone DMA32, type Reclaimable 0 0 2 1 0 0 0 0 1 0 0 953 Node 0, zone DMA32, type Movable 169 152 113 91 77 54 39 13 6 1 452 954 Node 0, zone DMA32, type Reserve 1 2 2 2 2 0 1 1 1 1 0 955 Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 956 957 Number of blocks type Unmovable Reclaimable Movable Reserve Isolate 958 Node 0, zone DMA 2 0 5 1 0 959 Node 0, zone DMA32 41 6 967 2 0 960 961Fragmentation avoidance in the kernel works by grouping pages of different 962migrate types into the same contiguous regions of memory called page blocks. 963A page block is typically the size of the default hugepage size, e.g. 2MB on 964X86-64. By keeping pages grouped based on their ability to move, the kernel 965can reclaim pages within a page block to satisfy a high-order allocation. 966 967The pagetypinfo begins with information on the size of a page block. It 968then gives the same type of information as buddyinfo except broken down 969by migrate-type and finishes with details on how many page blocks of each 970type exist. 971 972If min_free_kbytes has been tuned correctly (recommendations made by hugeadm 973from libhugetlbfs https://github.com/libhugetlbfs/libhugetlbfs/), one can 974make an estimate of the likely number of huge pages that can be allocated 975at a given point in time. All the "Movable" blocks should be allocatable 976unless memory has been mlock()'d. Some of the Reclaimable blocks should 977also be allocatable although a lot of filesystem metadata may have to be 978reclaimed to achieve this. 979 980 981allocinfo 982~~~~~~~~~ 983 984Provides information about memory allocations at all locations in the code 985base. Each allocation in the code is identified by its source file, line 986number, module (if originates from a loadable module) and the function calling 987the allocation. The number of bytes allocated and number of calls at each 988location are reported. The first line indicates the version of the file, the 989second line is the header listing fields in the file. 990 991Example output. 992 993:: 994 995 > tail -n +3 /proc/allocinfo | sort -rn 996 127664128 31168 mm/page_ext.c:270 func:alloc_page_ext 997 56373248 4737 mm/slub.c:2259 func:alloc_slab_page 998 14880768 3633 mm/readahead.c:247 func:page_cache_ra_unbounded 999 14417920 3520 mm/mm_init.c:2530 func:alloc_large_system_hash 1000 13377536 234 block/blk-mq.c:3421 func:blk_mq_alloc_rqs 1001 11718656 2861 mm/filemap.c:1919 func:__filemap_get_folio 1002 9192960 2800 kernel/fork.c:307 func:alloc_thread_stack_node 1003 4206592 4 net/netfilter/nf_conntrack_core.c:2567 func:nf_ct_alloc_hashtable 1004 4136960 1010 drivers/staging/ctagmod/ctagmod.c:20 [ctagmod] func:ctagmod_start 1005 3940352 962 mm/memory.c:4214 func:alloc_anon_folio 1006 2894464 22613 fs/kernfs/dir.c:615 func:__kernfs_new_node 1007 ... 1008 1009 1010meminfo 1011~~~~~~~ 1012 1013Provides information about distribution and utilization of memory. This 1014varies by architecture and compile options. Some of the counters reported 1015here overlap. The memory reported by the non overlapping counters may not 1016add up to the overall memory usage and the difference for some workloads 1017can be substantial. In many cases there are other means to find out 1018additional memory using subsystem specific interfaces, for instance 1019/proc/net/sockstat for TCP memory allocations. 1020 1021Example output. You may not have all of these fields. 1022 1023:: 1024 1025 > cat /proc/meminfo 1026 1027 MemTotal: 32858820 kB 1028 MemFree: 21001236 kB 1029 MemAvailable: 27214312 kB 1030 Buffers: 581092 kB 1031 Cached: 5587612 kB 1032 SwapCached: 0 kB 1033 Active: 3237152 kB 1034 Inactive: 7586256 kB 1035 Active(anon): 94064 kB 1036 Inactive(anon): 4570616 kB 1037 Active(file): 3143088 kB 1038 Inactive(file): 3015640 kB 1039 Unevictable: 0 kB 1040 Mlocked: 0 kB 1041 SwapTotal: 0 kB 1042 SwapFree: 0 kB 1043 Zswap: 1904 kB 1044 Zswapped: 7792 kB 1045 Dirty: 12 kB 1046 Writeback: 0 kB 1047 AnonPages: 4654780 kB 1048 Mapped: 266244 kB 1049 Shmem: 9976 kB 1050 KReclaimable: 517708 kB 1051 Slab: 660044 kB 1052 SReclaimable: 517708 kB 1053 SUnreclaim: 142336 kB 1054 KernelStack: 11168 kB 1055 PageTables: 20540 kB 1056 SecPageTables: 0 kB 1057 NFS_Unstable: 0 kB 1058 Bounce: 0 kB 1059 WritebackTmp: 0 kB 1060 CommitLimit: 16429408 kB 1061 Committed_AS: 7715148 kB 1062 VmallocTotal: 34359738367 kB 1063 VmallocUsed: 40444 kB 1064 VmallocChunk: 0 kB 1065 Percpu: 29312 kB 1066 EarlyMemtestBad: 0 kB 1067 HardwareCorrupted: 0 kB 1068 AnonHugePages: 4149248 kB 1069 ShmemHugePages: 0 kB 1070 ShmemPmdMapped: 0 kB 1071 FileHugePages: 0 kB 1072 FilePmdMapped: 0 kB 1073 CmaTotal: 0 kB 1074 CmaFree: 0 kB 1075 Unaccepted: 0 kB 1076 Balloon: 0 kB 1077 HugePages_Total: 0 1078 HugePages_Free: 0 1079 HugePages_Rsvd: 0 1080 HugePages_Surp: 0 1081 Hugepagesize: 2048 kB 1082 Hugetlb: 0 kB 1083 DirectMap4k: 401152 kB 1084 DirectMap2M: 10008576 kB 1085 DirectMap1G: 24117248 kB 1086 1087MemTotal 1088 Total usable RAM (i.e. physical RAM minus a few reserved 1089 bits and the kernel binary code) 1090MemFree 1091 Total free RAM. On highmem systems, the sum of LowFree+HighFree 1092MemAvailable 1093 An estimate of how much memory is available for starting new 1094 applications, without swapping. Calculated from MemFree, 1095 SReclaimable, the size of the file LRU lists, and the low 1096 watermarks in each zone. 1097 The estimate takes into account that the system needs some 1098 page cache to function well, and that not all reclaimable 1099 slab will be reclaimable, due to items being in use. The 1100 impact of those factors will vary from system to system. 1101Buffers 1102 Relatively temporary storage for raw disk blocks 1103 shouldn't get tremendously large (20MB or so) 1104Cached 1105 In-memory cache for files read from the disk (the 1106 pagecache) as well as tmpfs & shmem. 1107 Doesn't include SwapCached. 1108SwapCached 1109 Memory that once was swapped out, is swapped back in but 1110 still also is in the swapfile (if memory is needed it 1111 doesn't need to be swapped out AGAIN because it is already 1112 in the swapfile. This saves I/O) 1113Active 1114 Memory that has been used more recently and usually not 1115 reclaimed unless absolutely necessary. 1116Inactive 1117 Memory which has been less recently used. It is more 1118 eligible to be reclaimed for other purposes 1119Unevictable 1120 Memory allocated for userspace which cannot be reclaimed, such 1121 as mlocked pages, ramfs backing pages, secret memfd pages etc. 1122Mlocked 1123 Memory locked with mlock(). 1124HighTotal, HighFree 1125 Highmem is all memory above ~860MB of physical memory. 1126 Highmem areas are for use by userspace programs, or 1127 for the pagecache. The kernel must use tricks to access 1128 this memory, making it slower to access than lowmem. 1129LowTotal, LowFree 1130 Lowmem is memory which can be used for everything that 1131 highmem can be used for, but it is also available for the 1132 kernel's use for its own data structures. Among many 1133 other things, it is where everything from the Slab is 1134 allocated. Bad things happen when you're out of lowmem. 1135SwapTotal 1136 total amount of swap space available 1137SwapFree 1138 Memory which has been evicted from RAM, and is temporarily 1139 on the disk 1140Zswap 1141 Memory consumed by the zswap backend (compressed size) 1142Zswapped 1143 Amount of anonymous memory stored in zswap (original size) 1144Dirty 1145 Memory which is waiting to get written back to the disk 1146Writeback 1147 Memory which is actively being written back to the disk 1148AnonPages 1149 Non-file backed pages mapped into userspace page tables. Note that 1150 some kernel configurations might consider all pages part of a 1151 larger allocation (e.g., THP) as "mapped", as soon as a single 1152 page is mapped. 1153Mapped 1154 files which have been mmapped, such as libraries. Note that some 1155 kernel configurations might consider all pages part of a larger 1156 allocation (e.g., THP) as "mapped", as soon as a single page is 1157 mapped. 1158Shmem 1159 Total memory used by shared memory (shmem) and tmpfs 1160KReclaimable 1161 Kernel allocations that the kernel will attempt to reclaim 1162 under memory pressure. Includes SReclaimable (below), and other 1163 direct allocations with a shrinker. 1164Slab 1165 in-kernel data structures cache 1166SReclaimable 1167 Part of Slab, that might be reclaimed, such as caches 1168SUnreclaim 1169 Part of Slab, that cannot be reclaimed on memory pressure 1170KernelStack 1171 Memory consumed by the kernel stacks of all tasks 1172PageTables 1173 Memory consumed by userspace page tables 1174SecPageTables 1175 Memory consumed by secondary page tables, this currently includes 1176 KVM mmu and IOMMU allocations on x86 and arm64. 1177NFS_Unstable 1178 Always zero. Previously counted pages which had been written to 1179 the server, but has not been committed to stable storage. 1180Bounce 1181 Always zero. Previously memory used for block device 1182 "bounce buffers". 1183WritebackTmp 1184 Always zero. Previously memory used by FUSE for temporary 1185 writeback buffers. 1186CommitLimit 1187 Based on the overcommit ratio ('vm.overcommit_ratio'), 1188 this is the total amount of memory currently available to 1189 be allocated on the system. This limit is only adhered to 1190 if strict overcommit accounting is enabled (mode 2 in 1191 'vm.overcommit_memory'). 1192 1193 The CommitLimit is calculated with the following formula:: 1194 1195 CommitLimit = ([total RAM pages] - [total huge TLB pages]) * 1196 overcommit_ratio / 100 + [total swap pages] 1197 1198 For example, on a system with 1G of physical RAM and 7G 1199 of swap with a `vm.overcommit_ratio` of 30 it would 1200 yield a CommitLimit of 7.3G. 1201 1202 For more details, see the memory overcommit documentation 1203 in mm/overcommit-accounting. 1204Committed_AS 1205 The amount of memory presently allocated on the system. 1206 The committed memory is a sum of all of the memory which 1207 has been allocated by processes, even if it has not been 1208 "used" by them as of yet. A process which malloc()'s 1G 1209 of memory, but only touches 300M of it will show up as 1210 using 1G. This 1G is memory which has been "committed" to 1211 by the VM and can be used at any time by the allocating 1212 application. With strict overcommit enabled on the system 1213 (mode 2 in 'vm.overcommit_memory'), allocations which would 1214 exceed the CommitLimit (detailed above) will not be permitted. 1215 This is useful if one needs to guarantee that processes will 1216 not fail due to lack of memory once that memory has been 1217 successfully allocated. 1218VmallocTotal 1219 total size of vmalloc virtual address space 1220VmallocUsed 1221 amount of vmalloc area which is used 1222VmallocChunk 1223 largest contiguous block of vmalloc area which is free 1224Percpu 1225 Memory allocated to the percpu allocator used to back percpu 1226 allocations. This stat excludes the cost of metadata. 1227EarlyMemtestBad 1228 The amount of RAM/memory in kB, that was identified as corrupted 1229 by early memtest. If memtest was not run, this field will not 1230 be displayed at all. Size is never rounded down to 0 kB. 1231 That means if 0 kB is reported, you can safely assume 1232 there was at least one pass of memtest and none of the passes 1233 found a single faulty byte of RAM. 1234HardwareCorrupted 1235 The amount of RAM/memory in KB, the kernel identifies as 1236 corrupted. 1237AnonHugePages 1238 Non-file backed huge pages mapped into userspace page tables 1239ShmemHugePages 1240 Memory used by shared memory (shmem) and tmpfs allocated 1241 with huge pages 1242ShmemPmdMapped 1243 Shared memory mapped into userspace with huge pages 1244FileHugePages 1245 Memory used for filesystem data (page cache) allocated 1246 with huge pages 1247FilePmdMapped 1248 Page cache mapped into userspace with huge pages 1249CmaTotal 1250 Memory reserved for the Contiguous Memory Allocator (CMA) 1251CmaFree 1252 Free remaining memory in the CMA reserves 1253Unaccepted 1254 Memory that has not been accepted by the guest 1255Balloon 1256 Memory returned to Host by VM Balloon Drivers 1257HugePages_Total, HugePages_Free, HugePages_Rsvd, HugePages_Surp, Hugepagesize, Hugetlb 1258 See Documentation/admin-guide/mm/hugetlbpage.rst. 1259DirectMap4k, DirectMap2M, DirectMap1G 1260 Breakdown of page table sizes used in the kernel's 1261 identity mapping of RAM 1262 1263vmallocinfo 1264~~~~~~~~~~~ 1265 1266Provides information about vmalloced/vmaped areas. One line per area, 1267containing the virtual address range of the area, size in bytes, 1268caller information of the creator, and optional information depending 1269on the kind of area: 1270 1271 ========== =================================================== 1272 pages=nr number of pages 1273 phys=addr if a physical address was specified 1274 ioremap I/O mapping (ioremap() and friends) 1275 vmalloc vmalloc() area 1276 vmap vmap()ed pages 1277 user VM_USERMAP area 1278 vpages buffer for pages pointers was vmalloced (huge area) 1279 N<node>=nr (Only on NUMA kernels) 1280 Number of pages allocated on memory node <node> 1281 ========== =================================================== 1282 1283:: 1284 1285 > cat /proc/vmallocinfo 1286 0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ... 1287 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128 1288 0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ... 1289 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64 1290 0xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f... 1291 phys=7fee8000 ioremap 1292 0xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f... 1293 phys=7fee7000 ioremap 1294 0xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210 1295 0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ... 1296 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3 1297 0xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ... 1298 pages=2 vmalloc N1=2 1299 0xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ... 1300 /0x130 [x_tables] pages=4 vmalloc N0=4 1301 0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ... 1302 pages=14 vmalloc N2=14 1303 0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ... 1304 pages=4 vmalloc N1=4 1305 0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ... 1306 pages=2 vmalloc N1=2 1307 0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ... 1308 pages=10 vmalloc N0=10 1309 1310 1311softirqs 1312~~~~~~~~ 1313 1314Provides counts of softirq handlers serviced since boot time, for each CPU. 1315 1316:: 1317 1318 > cat /proc/softirqs 1319 CPU0 CPU1 CPU2 CPU3 1320 HI: 0 0 0 0 1321 TIMER: 27166 27120 27097 27034 1322 NET_TX: 0 0 0 17 1323 NET_RX: 42 0 0 39 1324 BLOCK: 0 0 107 1121 1325 TASKLET: 0 0 0 290 1326 SCHED: 27035 26983 26971 26746 1327 HRTIMER: 0 0 0 0 1328 RCU: 1678 1769 2178 2250 1329 13301.3 Networking info in /proc/net 1331-------------------------------- 1332 1333The subdirectory /proc/net follows the usual pattern. Table 1-8 shows the 1334additional values you get for IP version 6 if you configure the kernel to 1335support this. Table 1-9 lists the files and their meaning. 1336 1337 1338.. table:: Table 1-8: IPv6 info in /proc/net 1339 1340 ========== ===================================================== 1341 File Content 1342 ========== ===================================================== 1343 udp6 UDP sockets (IPv6) 1344 tcp6 TCP sockets (IPv6) 1345 raw6 Raw device statistics (IPv6) 1346 igmp6 IP multicast addresses, which this host joined (IPv6) 1347 if_inet6 List of IPv6 interface addresses 1348 ipv6_route Kernel routing table for IPv6 1349 rt6_stats Global IPv6 routing tables statistics 1350 sockstat6 Socket statistics (IPv6) 1351 snmp6 Snmp data (IPv6) 1352 ========== ===================================================== 1353 1354.. table:: Table 1-9: Network info in /proc/net 1355 1356 ============= ================================================================ 1357 File Content 1358 ============= ================================================================ 1359 arp Kernel ARP table 1360 dev network devices with statistics 1361 dev_mcast the Layer2 multicast groups a device is listening too 1362 (interface index, label, number of references, number of bound 1363 addresses). 1364 dev_stat network device status 1365 ip_fwchains Firewall chain linkage 1366 ip_fwnames Firewall chain names 1367 ip_masq Directory containing the masquerading tables 1368 ip_masquerade Major masquerading table 1369 netstat Network statistics 1370 raw raw device statistics 1371 route Kernel routing table 1372 rpc Directory containing rpc info 1373 rt_cache Routing cache 1374 snmp SNMP data 1375 sockstat Socket statistics 1376 softnet_stat Per-CPU incoming packets queues statistics of online CPUs 1377 tcp TCP sockets 1378 udp UDP sockets 1379 unix UNIX domain sockets 1380 wireless Wireless interface data (Wavelan etc) 1381 igmp IP multicast addresses, which this host joined 1382 psched Global packet scheduler parameters. 1383 netlink List of PF_NETLINK sockets 1384 ip_mr_vifs List of multicast virtual interfaces 1385 ip_mr_cache List of multicast routing cache 1386 ============= ================================================================ 1387 1388You can use this information to see which network devices are available in 1389your system and how much traffic was routed over those devices:: 1390 1391 > cat /proc/net/dev 1392 Inter-|Receive |[... 1393 face |bytes packets errs drop fifo frame compressed multicast|[... 1394 lo: 908188 5596 0 0 0 0 0 0 [... 1395 ppp0:15475140 20721 410 0 0 410 0 0 [... 1396 eth0: 614530 7085 0 0 0 0 0 1 [... 1397 1398 ...] Transmit 1399 ...] bytes packets errs drop fifo colls carrier compressed 1400 ...] 908188 5596 0 0 0 0 0 0 1401 ...] 1375103 17405 0 0 0 0 0 0 1402 ...] 1703981 5535 0 0 0 3 0 0 1403 1404In addition, each Channel Bond interface has its own directory. For 1405example, the bond0 device will have a directory called /proc/net/bond0/. 1406It will contain information that is specific to that bond, such as the 1407current slaves of the bond, the link status of the slaves, and how 1408many times the slaves link has failed. 1409 14101.4 SCSI info 1411------------- 1412 1413If you have a SCSI or ATA host adapter in your system, you'll find a 1414subdirectory named after the driver for this adapter in /proc/scsi. 1415You'll also see a list of all recognized SCSI devices in /proc/scsi:: 1416 1417 >cat /proc/scsi/scsi 1418 Attached devices: 1419 Host: scsi0 Channel: 00 Id: 00 Lun: 00 1420 Vendor: IBM Model: DGHS09U Rev: 03E0 1421 Type: Direct-Access ANSI SCSI revision: 03 1422 Host: scsi0 Channel: 00 Id: 06 Lun: 00 1423 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04 1424 Type: CD-ROM ANSI SCSI revision: 02 1425 1426 1427The directory named after the driver has one file for each adapter found in 1428the system. These files contain information about the controller, including 1429the used IRQ and the IO address range. The amount of information shown is 1430dependent on the adapter you use. The example shows the output for an Adaptec 1431AHA-2940 SCSI adapter:: 1432 1433 > cat /proc/scsi/aic7xxx/0 1434 1435 Adaptec AIC7xxx driver version: 5.1.19/3.2.4 1436 Compile Options: 1437 TCQ Enabled By Default : Disabled 1438 AIC7XXX_PROC_STATS : Disabled 1439 AIC7XXX_RESET_DELAY : 5 1440 Adapter Configuration: 1441 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter 1442 Ultra Wide Controller 1443 PCI MMAPed I/O Base: 0xeb001000 1444 Adapter SEEPROM Config: SEEPROM found and used. 1445 Adaptec SCSI BIOS: Enabled 1446 IRQ: 10 1447 SCBs: Active 0, Max Active 2, 1448 Allocated 15, HW 16, Page 255 1449 Interrupts: 160328 1450 BIOS Control Word: 0x18b6 1451 Adapter Control Word: 0x005b 1452 Extended Translation: Enabled 1453 Disconnect Enable Flags: 0xffff 1454 Ultra Enable Flags: 0x0001 1455 Tag Queue Enable Flags: 0x0000 1456 Ordered Queue Tag Flags: 0x0000 1457 Default Tag Queue Depth: 8 1458 Tagged Queue By Device array for aic7xxx host instance 0: 1459 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255} 1460 Actual queue depth per device for aic7xxx host instance 0: 1461 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1} 1462 Statistics: 1463 (scsi0:0:0:0) 1464 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8 1465 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0) 1466 Total transfers 160151 (74577 reads and 85574 writes) 1467 (scsi0:0:6:0) 1468 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15 1469 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0) 1470 Total transfers 0 (0 reads and 0 writes) 1471 1472 14731.5 Parallel port info in /proc/parport 1474--------------------------------------- 1475 1476The directory /proc/parport contains information about the parallel ports of 1477your system. It has one subdirectory for each port, named after the port 1478number (0,1,2,...). 1479 1480These directories contain the four files shown in Table 1-10. 1481 1482 1483.. table:: Table 1-10: Files in /proc/parport 1484 1485 ========= ==================================================================== 1486 File Content 1487 ========= ==================================================================== 1488 autoprobe Any IEEE-1284 device ID information that has been acquired. 1489 devices list of the device drivers using that port. A + will appear by the 1490 name of the device currently using the port (it might not appear 1491 against any). 1492 hardware Parallel port's base address, IRQ line and DMA channel. 1493 irq IRQ that parport is using for that port. This is in a separate 1494 file to allow you to alter it by writing a new value in (IRQ 1495 number or none). 1496 ========= ==================================================================== 1497 14981.6 TTY info in /proc/tty 1499------------------------- 1500 1501Information about the available and actually used tty's can be found in the 1502directory /proc/tty. You'll find entries for drivers and line disciplines in 1503this directory, as shown in Table 1-11. 1504 1505 1506.. table:: Table 1-11: Files in /proc/tty 1507 1508 ============= ============================================== 1509 File Content 1510 ============= ============================================== 1511 drivers list of drivers and their usage 1512 ldiscs registered line disciplines 1513 driver/serial usage statistic and status of single tty lines 1514 ============= ============================================== 1515 1516To see which tty's are currently in use, you can simply look into the file 1517/proc/tty/drivers:: 1518 1519 > cat /proc/tty/drivers 1520 pty_slave /dev/pts 136 0-255 pty:slave 1521 pty_master /dev/ptm 128 0-255 pty:master 1522 pty_slave /dev/ttyp 3 0-255 pty:slave 1523 pty_master /dev/pty 2 0-255 pty:master 1524 serial /dev/cua 5 64-67 serial:callout 1525 serial /dev/ttyS 4 64-67 serial 1526 /dev/tty0 /dev/tty0 4 0 system:vtmaster 1527 /dev/ptmx /dev/ptmx 5 2 system 1528 /dev/console /dev/console 5 1 system:console 1529 /dev/tty /dev/tty 5 0 system:/dev/tty 1530 unknown /dev/tty 4 1-63 console 1531 1532 15331.7 Miscellaneous kernel statistics in /proc/stat 1534------------------------------------------------- 1535 1536Various pieces of information about kernel activity are available in the 1537/proc/stat file. All of the numbers reported in this file are aggregates 1538since the system first booted. For a quick look, simply cat the file:: 1539 1540 > cat /proc/stat 1541 cpu 237902850 368826709 106375398 1873517540 1135548 0 14507935 0 0 0 1542 cpu0 60045249 91891769 26331539 468411416 495718 0 5739640 0 0 0 1543 cpu1 59746288 91759249 26609887 468860630 312281 0 4384817 0 0 0 1544 cpu2 59489247 92985423 26904446 467808813 171668 0 2268998 0 0 0 1545 cpu3 58622065 92190267 26529524 468436680 155879 0 2114478 0 0 0 1546 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> 1547 ctxt 22848221062 1548 btime 1605316999 1549 processes 746787147 1550 procs_running 2 1551 procs_blocked 0 1552 softirq 12121874454 100099120 3938138295 127375644 2795979 187870761 0 173808342 3072582055 52608 224184354 1553 1554The very first "cpu" line aggregates the numbers in all of the other "cpuN" 1555lines. These numbers identify the amount of time the CPU has spent performing 1556different kinds of work. Time units are in USER_HZ (typically hundredths of a 1557second). The meanings of the columns are as follows, from left to right: 1558 1559- user: normal processes executing in user mode 1560- nice: niced processes executing in user mode 1561- system: processes executing in kernel mode 1562- idle: twiddling thumbs 1563- iowait: In a word, iowait stands for waiting for I/O to complete. But there 1564 are several problems: 1565 1566 1. CPU will not wait for I/O to complete, iowait is the time that a task is 1567 waiting for I/O to complete. When CPU goes into idle state for 1568 outstanding task I/O, another task will be scheduled on this CPU. 1569 2. In a multi-core CPU, the task waiting for I/O to complete is not running 1570 on any CPU, so the iowait of each CPU is difficult to calculate. 1571 3. The value of iowait field in /proc/stat will decrease in certain 1572 conditions. 1573 1574 So, the iowait is not reliable by reading from /proc/stat. 1575- irq: servicing interrupts 1576- softirq: servicing softirqs 1577- steal: involuntary wait 1578- guest: running a normal guest 1579- guest_nice: running a niced guest 1580 1581The "intr" line gives counts of interrupts serviced since boot time, for each 1582of the possible system interrupts. The first column is the total of all 1583interrupts serviced including unnumbered architecture specific interrupts; 1584each subsequent column is the total for that particular numbered interrupt. 1585Unnumbered interrupts are not shown, only summed into the total. 1586 1587The "ctxt" line gives the total number of context switches across all CPUs. 1588 1589The "btime" line gives the time at which the system booted, in seconds since 1590the Unix epoch. 1591 1592The "processes" line gives the number of processes and threads created, which 1593includes (but is not limited to) those created by calls to the fork() and 1594clone() system calls. 1595 1596The "procs_running" line gives the total number of threads that are 1597running or ready to run (i.e., the total number of runnable threads). 1598 1599The "procs_blocked" line gives the number of processes currently blocked, 1600waiting for I/O to complete. 1601 1602The "softirq" line gives counts of softirqs serviced since boot time, for each 1603of the possible system softirqs. The first column is the total of all 1604softirqs serviced; each subsequent column is the total for that particular 1605softirq. 1606 1607 16081.8 Ext4 file system parameters 1609------------------------------- 1610 1611Information about mounted ext4 file systems can be found in 1612/proc/fs/ext4. Each mounted filesystem will have a directory in 1613/proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or 1614/proc/fs/ext4/sda9 or /proc/fs/ext4/dm-0). The files in each per-device 1615directory are shown in Table 1-12, below. 1616 1617.. table:: Table 1-12: Files in /proc/fs/ext4/<devname> 1618 1619 ============== ========================================================== 1620 File Content 1621 mb_groups details of multiblock allocator buddy cache of free blocks 1622 ============== ========================================================== 1623 16241.9 /proc/consoles 1625------------------- 1626Shows registered system console lines. 1627 1628To see which character device lines are currently used for the system console 1629/dev/console, you may simply look into the file /proc/consoles:: 1630 1631 > cat /proc/consoles 1632 tty0 -WU (ECp) 4:7 1633 ttyS0 -W- (Ep) 4:64 1634 1635The columns are: 1636 1637+--------------------+-------------------------------------------------------+ 1638| device | name of the device | 1639+====================+=======================================================+ 1640| operations | * R = can do read operations | 1641| | * W = can do write operations | 1642| | * U = can do unblank | 1643+--------------------+-------------------------------------------------------+ 1644| flags | * E = it is enabled | 1645| | * C = it is preferred console | 1646| | * B = it is primary boot console | 1647| | * p = it is used for printk buffer | 1648| | * b = it is not a TTY but a Braille device | 1649| | * a = it is safe to use when cpu is offline | 1650+--------------------+-------------------------------------------------------+ 1651| major:minor | major and minor number of the device separated by a | 1652| | colon | 1653+--------------------+-------------------------------------------------------+ 1654 1655Summary 1656------- 1657 1658The /proc file system serves information about the running system. It not only 1659allows access to process data but also allows you to request the kernel status 1660by reading files in the hierarchy. 1661 1662The directory structure of /proc reflects the types of information and makes 1663it easy, if not obvious, where to look for specific data. 1664 1665Chapter 2: Modifying System Parameters 1666====================================== 1667 1668In This Chapter 1669--------------- 1670 1671* Modifying kernel parameters by writing into files found in /proc/sys 1672* Exploring the files which modify certain parameters 1673* Review of the /proc/sys file tree 1674 1675------------------------------------------------------------------------------ 1676 1677A very interesting part of /proc is the directory /proc/sys. This is not only 1678a source of information, it also allows you to change parameters within the 1679kernel. Be very careful when attempting this. You can optimize your system, 1680but you can also cause it to crash. Never alter kernel parameters on a 1681production system. Set up a development machine and test to make sure that 1682everything works the way you want it to. You may have no alternative but to 1683reboot the machine once an error has been made. 1684 1685To change a value, simply echo the new value into the file. 1686You need to be root to do this. You can create your own boot script 1687to perform this every time your system boots. 1688 1689The files in /proc/sys can be used to fine tune and monitor miscellaneous and 1690general things in the operation of the Linux kernel. Since some of the files 1691can inadvertently disrupt your system, it is advisable to read both 1692documentation and source before actually making adjustments. In any case, be 1693very careful when writing to any of these files. The entries in /proc may 1694change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt 1695review the kernel documentation in the directory linux/Documentation. 1696This chapter is heavily based on the documentation included in the pre 2.2 1697kernels, and became part of it in version 2.2.1 of the Linux kernel. 1698 1699Please see: Documentation/admin-guide/sysctl/ directory for descriptions of 1700these entries. 1701 1702Summary 1703------- 1704 1705Certain aspects of kernel behavior can be modified at runtime, without the 1706need to recompile the kernel, or even to reboot the system. The files in the 1707/proc/sys tree can not only be read, but also modified. You can use the echo 1708command to write value into these files, thereby changing the default settings 1709of the kernel. 1710 1711 1712Chapter 3: Per-process Parameters 1713================================= 1714 17153.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj- Adjust the oom-killer score 1716-------------------------------------------------------------------------------- 1717 1718These files can be used to adjust the badness heuristic used to select which 1719process gets killed in out of memory (oom) conditions. 1720 1721The badness heuristic assigns a value to each candidate task ranging from 0 1722(never kill) to 1000 (always kill) to determine which process is targeted. The 1723units are roughly a proportion along that range of allowed memory the process 1724may allocate from based on an estimation of its current memory and swap use. 1725For example, if a task is using all allowed memory, its badness score will be 17261000. If it is using half of its allowed memory, its score will be 500. 1727 1728The amount of "allowed" memory depends on the context in which the oom killer 1729was called. If it is due to the memory assigned to the allocating task's cpuset 1730being exhausted, the allowed memory represents the set of mems assigned to that 1731cpuset. If it is due to a mempolicy's node(s) being exhausted, the allowed 1732memory represents the set of mempolicy nodes. If it is due to a memory 1733limit (or swap limit) being reached, the allowed memory is that configured 1734limit. Finally, if it is due to the entire system being out of memory, the 1735allowed memory represents all allocatable resources. 1736 1737The value of /proc/<pid>/oom_score_adj is added to the badness score before it 1738is used to determine which task to kill. Acceptable values range from -1000 1739(OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_MAX). This allows userspace to 1740polarize the preference for oom killing either by always preferring a certain 1741task or completely disabling it. The lowest possible value, -1000, is 1742equivalent to disabling oom killing entirely for that task since it will always 1743report a badness score of 0. 1744 1745Consequently, it is very simple for userspace to define the amount of memory to 1746consider for each task. Setting a /proc/<pid>/oom_score_adj value of +500, for 1747example, is roughly equivalent to allowing the remainder of tasks sharing the 1748same system, cpuset, mempolicy, or memory controller resources to use at least 174950% more memory. A value of -500, on the other hand, would be roughly 1750equivalent to discounting 50% of the task's allowed memory from being considered 1751as scoring against the task. 1752 1753For backwards compatibility with previous kernels, /proc/<pid>/oom_adj may also 1754be used to tune the badness score. Its acceptable values range from -16 1755(OOM_ADJUST_MIN) to +15 (OOM_ADJUST_MAX) and a special value of -17 1756(OOM_DISABLE) to disable oom killing entirely for that task. Its value is 1757scaled linearly with /proc/<pid>/oom_score_adj. 1758 1759The value of /proc/<pid>/oom_score_adj may be reduced no lower than the last 1760value set by a CAP_SYS_RESOURCE process. To reduce the value any lower 1761requires CAP_SYS_RESOURCE. 1762 1763 17643.2 /proc/<pid>/oom_score - Display current oom-killer score 1765------------------------------------------------------------- 1766 1767This file can be used to check the current score used by the oom-killer for 1768any given <pid>. Use it together with /proc/<pid>/oom_score_adj to tune which 1769process should be killed in an out-of-memory situation. 1770 1771Please note that the exported value includes oom_score_adj so it is 1772effectively in range [0,2000]. 1773 1774 17753.3 /proc/<pid>/io - Display the IO accounting fields 1776------------------------------------------------------- 1777 1778This file contains IO statistics for each running process. 1779 1780Example 1781~~~~~~~ 1782 1783:: 1784 1785 test:/tmp # dd if=/dev/zero of=/tmp/test.dat & 1786 [1] 3828 1787 1788 test:/tmp # cat /proc/3828/io 1789 rchar: 323934931 1790 wchar: 323929600 1791 syscr: 632687 1792 syscw: 632675 1793 read_bytes: 0 1794 write_bytes: 323932160 1795 cancelled_write_bytes: 0 1796 1797 1798Description 1799~~~~~~~~~~~ 1800 1801rchar 1802^^^^^ 1803 1804I/O counter: chars read 1805The number of bytes which this task has caused to be read from storage. This 1806is simply the sum of bytes which this process passed to read() and pread(). 1807It includes things like tty IO and it is unaffected by whether or not actual 1808physical disk IO was required (the read might have been satisfied from 1809pagecache). 1810 1811 1812wchar 1813^^^^^ 1814 1815I/O counter: chars written 1816The number of bytes which this task has caused, or shall cause to be written 1817to disk. Similar caveats apply here as with rchar. 1818 1819 1820syscr 1821^^^^^ 1822 1823I/O counter: read syscalls 1824Attempt to count the number of read I/O operations, i.e. syscalls like read() 1825and pread(). 1826 1827 1828syscw 1829^^^^^ 1830 1831I/O counter: write syscalls 1832Attempt to count the number of write I/O operations, i.e. syscalls like 1833write() and pwrite(). 1834 1835 1836read_bytes 1837^^^^^^^^^^ 1838 1839I/O counter: bytes read 1840Attempt to count the number of bytes which this process really did cause to 1841be fetched from the storage layer. Done at the submit_bio() level, so it is 1842accurate for block-backed filesystems. <please add status regarding NFS and 1843CIFS at a later time> 1844 1845 1846write_bytes 1847^^^^^^^^^^^ 1848 1849I/O counter: bytes written 1850Attempt to count the number of bytes which this process caused to be sent to 1851the storage layer. This is done at page-dirtying time. 1852 1853 1854cancelled_write_bytes 1855^^^^^^^^^^^^^^^^^^^^^ 1856 1857The big inaccuracy here is truncate. If a process writes 1MB to a file and 1858then deletes the file, it will in fact perform no writeout. But it will have 1859been accounted as having caused 1MB of write. 1860In other words: The number of bytes which this process caused to not happen, 1861by truncating pagecache. A task can cause "negative" IO too. If this task 1862truncates some dirty pagecache, some IO which another task has been accounted 1863for (in its write_bytes) will not be happening. We _could_ just subtract that 1864from the truncating task's write_bytes, but there is information loss in doing 1865that. 1866 1867 1868.. Note:: 1869 1870 At its current implementation state, this is a bit racy on 32-bit machines: 1871 if process A reads process B's /proc/pid/io while process B is updating one 1872 of those 64-bit counters, process A could see an intermediate result. 1873 1874 1875More information about this can be found within the taskstats documentation in 1876Documentation/accounting. 1877 18783.4 /proc/<pid>/coredump_filter - Core dump filtering settings 1879--------------------------------------------------------------- 1880When a process is dumped, all anonymous memory is written to a core file as 1881long as the size of the core file isn't limited. But sometimes we don't want 1882to dump some memory segments, for example, huge shared memory or DAX. 1883Conversely, sometimes we want to save file-backed memory segments into a core 1884file, not only the individual files. 1885 1886/proc/<pid>/coredump_filter allows you to customize which memory segments 1887will be dumped when the <pid> process is dumped. coredump_filter is a bitmask 1888of memory types. If a bit of the bitmask is set, memory segments of the 1889corresponding memory type are dumped, otherwise they are not dumped. 1890 1891The following 9 memory types are supported: 1892 1893 - (bit 0) anonymous private memory 1894 - (bit 1) anonymous shared memory 1895 - (bit 2) file-backed private memory 1896 - (bit 3) file-backed shared memory 1897 - (bit 4) ELF header pages in file-backed private memory areas (it is 1898 effective only if the bit 2 is cleared) 1899 - (bit 5) hugetlb private memory 1900 - (bit 6) hugetlb shared memory 1901 - (bit 7) DAX private memory 1902 - (bit 8) DAX shared memory 1903 1904 Note that MMIO pages such as frame buffer are never dumped and vDSO pages 1905 are always dumped regardless of the bitmask status. 1906 1907 Note that bits 0-4 don't affect hugetlb or DAX memory. hugetlb memory is 1908 only affected by bit 5-6, and DAX is only affected by bits 7-8. 1909 1910The default value of coredump_filter is 0x33; this means all anonymous memory 1911segments, ELF header pages and hugetlb private memory are dumped. 1912 1913If you don't want to dump all shared memory segments attached to pid 1234, 1914write 0x31 to the process's proc file:: 1915 1916 $ echo 0x31 > /proc/1234/coredump_filter 1917 1918When a new process is created, the process inherits the bitmask status from its 1919parent. It is useful to set up coredump_filter before the program runs. 1920For example:: 1921 1922 $ echo 0x7 > /proc/self/coredump_filter 1923 $ ./some_program 1924 19253.5 /proc/<pid>/mountinfo - Information about mounts 1926-------------------------------------------------------- 1927 1928This file contains lines of the form:: 1929 1930 36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue 1931 (1)(2)(3) (4) (5) (6) (n…m) (m+1)(m+2) (m+3) (m+4) 1932 1933 (1) mount ID: unique identifier of the mount (may be reused after umount) 1934 (2) parent ID: ID of parent (or of self for the top of the mount tree) 1935 (3) major:minor: value of st_dev for files on filesystem 1936 (4) root: root of the mount within the filesystem 1937 (5) mount point: mount point relative to the process's root 1938 (6) mount options: per mount options 1939 (n…m) optional fields: zero or more fields of the form "tag[:value]" 1940 (m+1) separator: marks the end of the optional fields 1941 (m+2) filesystem type: name of filesystem of the form "type[.subtype]" 1942 (m+3) mount source: filesystem specific information or "none" 1943 (m+4) super options: per super block options 1944 1945Parsers should ignore all unrecognised optional fields. Currently the 1946possible optional fields are: 1947 1948================ ============================================================== 1949shared:X mount is shared in peer group X 1950master:X mount is slave to peer group X 1951propagate_from:X mount is slave and receives propagation from peer group X [#]_ 1952unbindable mount is unbindable 1953================ ============================================================== 1954 1955.. [#] X is the closest dominant peer group under the process's root. If 1956 X is the immediate master of the mount, or if there's no dominant peer 1957 group under the same root, then only the "master:X" field is present 1958 and not the "propagate_from:X" field. 1959 1960For more information on mount propagation see: 1961 1962 Documentation/filesystems/sharedsubtree.rst 1963 1964 19653.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm 1966-------------------------------------------------------- 1967These files provide a method to access a task's comm value. It also allows for 1968a task to set its own or one of its thread siblings comm value. The comm value 1969is limited in size compared to the cmdline value, so writing anything longer 1970then the kernel's TASK_COMM_LEN (currently 16 chars, including the NUL 1971terminator) will result in a truncated comm value. 1972 1973 19743.7 /proc/<pid>/task/<tid>/children - Information about task children 1975------------------------------------------------------------------------- 1976This file provides a fast way to retrieve first level children pids 1977of a task pointed by <pid>/<tid> pair. The format is a space separated 1978stream of pids. 1979 1980Note the "first level" here -- if a child has its own children they will 1981not be listed here; one needs to read /proc/<children-pid>/task/<tid>/children 1982to obtain the descendants. 1983 1984Since this interface is intended to be fast and cheap it doesn't 1985guarantee to provide precise results and some children might be 1986skipped, especially if they've exited right after we printed their 1987pids, so one needs to either stop or freeze processes being inspected 1988if precise results are needed. 1989 1990 19913.8 /proc/<pid>/fdinfo/<fd> - Information about opened file 1992--------------------------------------------------------------- 1993This file provides information associated with an opened file. The regular 1994files have at least four fields -- 'pos', 'flags', 'mnt_id' and 'ino'. 1995The 'pos' represents the current offset of the opened file in decimal 1996form [see lseek(2) for details], 'flags' denotes the octal O_xxx mask the 1997file has been created with [see open(2) for details] and 'mnt_id' represents 1998mount ID of the file system containing the opened file [see 3.5 1999/proc/<pid>/mountinfo for details]. 'ino' represents the inode number of 2000the file. 2001 2002A typical output is:: 2003 2004 pos: 0 2005 flags: 0100002 2006 mnt_id: 19 2007 ino: 63107 2008 2009All locks associated with a file descriptor are shown in its fdinfo too:: 2010 2011 lock: 1: FLOCK ADVISORY WRITE 359 00:13:11691 0 EOF 2012 2013The files such as eventfd, fsnotify, signalfd, epoll among the regular pos/flags 2014pair provide additional information particular to the objects they represent. 2015 2016Eventfd files 2017~~~~~~~~~~~~~ 2018 2019:: 2020 2021 pos: 0 2022 flags: 04002 2023 mnt_id: 9 2024 ino: 63107 2025 eventfd-count: 5a 2026 2027where 'eventfd-count' is hex value of a counter. 2028 2029Signalfd files 2030~~~~~~~~~~~~~~ 2031 2032:: 2033 2034 pos: 0 2035 flags: 04002 2036 mnt_id: 9 2037 ino: 63107 2038 sigmask: 0000000000000200 2039 2040where 'sigmask' is hex value of the signal mask associated 2041with a file. 2042 2043Epoll files 2044~~~~~~~~~~~ 2045 2046:: 2047 2048 pos: 0 2049 flags: 02 2050 mnt_id: 9 2051 ino: 63107 2052 tfd: 5 events: 1d data: ffffffffffffffff pos:0 ino:61af sdev:7 2053 2054where 'tfd' is a target file descriptor number in decimal form, 2055'events' is events mask being watched and the 'data' is data 2056associated with a target [see epoll(7) for more details]. 2057 2058The 'pos' is current offset of the target file in decimal form 2059[see lseek(2)], 'ino' and 'sdev' are inode and device numbers 2060where target file resides, all in hex format. 2061 2062Fsnotify files 2063~~~~~~~~~~~~~~ 2064For inotify files the format is the following:: 2065 2066 pos: 0 2067 flags: 02000000 2068 mnt_id: 9 2069 ino: 63107 2070 inotify wd:3 ino:9e7e sdev:800013 mask:800afce ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:7e9e0000640d1b6d 2071 2072where 'wd' is a watch descriptor in decimal form, i.e. a target file 2073descriptor number, 'ino' and 'sdev' are inode and device where the 2074target file resides and the 'mask' is the mask of events, all in hex 2075form [see inotify(7) for more details]. 2076 2077If the kernel was built with exportfs support, the path to the target 2078file is encoded as a file handle. The file handle is provided by three 2079fields 'fhandle-bytes', 'fhandle-type' and 'f_handle', all in hex 2080format. 2081 2082If the kernel is built without exportfs support the file handle won't be 2083printed out. 2084 2085If there is no inotify mark attached yet the 'inotify' line will be omitted. 2086 2087For fanotify files the format is:: 2088 2089 pos: 0 2090 flags: 02 2091 mnt_id: 9 2092 ino: 63107 2093 fanotify flags:10 event-flags:0 2094 fanotify mnt_id:12 mflags:40 mask:38 ignored_mask:40000003 2095 fanotify ino:4f969 sdev:800013 mflags:0 mask:3b ignored_mask:40000000 fhandle-bytes:8 fhandle-type:1 f_handle:69f90400c275b5b4 2096 2097where fanotify 'flags' and 'event-flags' are values used in fanotify_init 2098call, 'mnt_id' is the mount point identifier, 'mflags' is the value of 2099flags associated with mark which are tracked separately from events 2100mask. 'ino' and 'sdev' are target inode and device, 'mask' is the events 2101mask and 'ignored_mask' is the mask of events which are to be ignored. 2102All are in hex format. Incorporation of 'mflags', 'mask' and 'ignored_mask' 2103provide information about flags and mask used in fanotify_mark 2104call [see fsnotify manpage for details]. 2105 2106While the first three lines are mandatory and always printed, the rest is 2107optional and may be omitted if no marks created yet. 2108 2109Timerfd files 2110~~~~~~~~~~~~~ 2111 2112:: 2113 2114 pos: 0 2115 flags: 02 2116 mnt_id: 9 2117 ino: 63107 2118 clockid: 0 2119 ticks: 0 2120 settime flags: 01 2121 it_value: (0, 49406829) 2122 it_interval: (1, 0) 2123 2124where 'clockid' is the clock type and 'ticks' is the number of the timer expirations 2125that have occurred [see timerfd_create(2) for details]. 'settime flags' are 2126flags in octal form been used to setup the timer [see timerfd_settime(2) for 2127details]. 'it_value' is remaining time until the timer expiration. 2128'it_interval' is the interval for the timer. Note the timer might be set up 2129with TIMER_ABSTIME option which will be shown in 'settime flags', but 'it_value' 2130still exhibits timer's remaining time. 2131 2132DMA Buffer files 2133~~~~~~~~~~~~~~~~ 2134 2135:: 2136 2137 pos: 0 2138 flags: 04002 2139 mnt_id: 9 2140 ino: 63107 2141 size: 32768 2142 count: 2 2143 exp_name: system-heap 2144 2145where 'size' is the size of the DMA buffer in bytes. 'count' is the file count of 2146the DMA buffer file. 'exp_name' is the name of the DMA buffer exporter. 2147 21483.9 /proc/<pid>/map_files - Information about memory mapped files 2149--------------------------------------------------------------------- 2150This directory contains symbolic links which represent memory mapped files 2151the process is maintaining. Example output:: 2152 2153 | lr-------- 1 root root 64 Jan 27 11:24 333c600000-333c620000 -> /usr/lib64/ld-2.18.so 2154 | lr-------- 1 root root 64 Jan 27 11:24 333c81f000-333c820000 -> /usr/lib64/ld-2.18.so 2155 | lr-------- 1 root root 64 Jan 27 11:24 333c820000-333c821000 -> /usr/lib64/ld-2.18.so 2156 | ... 2157 | lr-------- 1 root root 64 Jan 27 11:24 35d0421000-35d0422000 -> /usr/lib64/libselinux.so.1 2158 | lr-------- 1 root root 64 Jan 27 11:24 400000-41a000 -> /usr/bin/ls 2159 2160The name of a link represents the virtual memory bounds of a mapping, i.e. 2161vm_area_struct::vm_start-vm_area_struct::vm_end. 2162 2163The main purpose of the map_files is to retrieve a set of memory mapped 2164files in a fast way instead of parsing /proc/<pid>/maps or 2165/proc/<pid>/smaps, both of which contain many more records. At the same 2166time one can open(2) mappings from the listings of two processes and 2167comparing their inode numbers to figure out which anonymous memory areas 2168are actually shared. 2169 21703.10 /proc/<pid>/timerslack_ns - Task timerslack value 2171--------------------------------------------------------- 2172This file provides the value of the task's timerslack value in nanoseconds. 2173This value specifies an amount of time that normal timers may be deferred 2174in order to coalesce timers and avoid unnecessary wakeups. 2175 2176This allows a task's interactivity vs power consumption tradeoff to be 2177adjusted. 2178 2179Writing 0 to the file will set the task's timerslack to the default value. 2180 2181Valid values are from 0 - ULLONG_MAX 2182 2183An application setting the value must have PTRACE_MODE_ATTACH_FSCREDS level 2184permissions on the task specified to change its timerslack_ns value. 2185 21863.11 /proc/<pid>/patch_state - Livepatch patch operation state 2187----------------------------------------------------------------- 2188When CONFIG_LIVEPATCH is enabled, this file displays the value of the 2189patch state for the task. 2190 2191A value of '-1' indicates that no patch is in transition. 2192 2193A value of '0' indicates that a patch is in transition and the task is 2194unpatched. If the patch is being enabled, then the task hasn't been 2195patched yet. If the patch is being disabled, then the task has already 2196been unpatched. 2197 2198A value of '1' indicates that a patch is in transition and the task is 2199patched. If the patch is being enabled, then the task has already been 2200patched. If the patch is being disabled, then the task hasn't been 2201unpatched yet. 2202 22033.12 /proc/<pid>/arch_status - task architecture specific status 2204------------------------------------------------------------------- 2205When CONFIG_PROC_PID_ARCH_STATUS is enabled, this file displays the 2206architecture specific status of the task. 2207 2208Example 2209~~~~~~~ 2210 2211:: 2212 2213 $ cat /proc/6753/arch_status 2214 AVX512_elapsed_ms: 8 2215 2216Description 2217~~~~~~~~~~~ 2218 2219x86 specific entries 2220~~~~~~~~~~~~~~~~~~~~~ 2221 2222AVX512_elapsed_ms 2223^^^^^^^^^^^^^^^^^^ 2224 2225 If AVX512 is supported on the machine, this entry shows the milliseconds 2226 elapsed since the last time AVX512 usage was recorded. The recording 2227 happens on a best effort basis when a task is scheduled out. This means 2228 that the value depends on two factors: 2229 2230 1) The time which the task spent on the CPU without being scheduled 2231 out. With CPU isolation and a single runnable task this can take 2232 several seconds. 2233 2234 2) The time since the task was scheduled out last. Depending on the 2235 reason for being scheduled out (time slice exhausted, syscall ...) 2236 this can be arbitrary long time. 2237 2238 As a consequence the value cannot be considered precise and authoritative 2239 information. The application which uses this information has to be aware 2240 of the overall scenario on the system in order to determine whether a 2241 task is a real AVX512 user or not. Precise information can be obtained 2242 with performance counters. 2243 2244 A special value of '-1' indicates that no AVX512 usage was recorded, thus 2245 the task is unlikely an AVX512 user, but depends on the workload and the 2246 scheduling scenario, it also could be a false negative mentioned above. 2247 22483.13 /proc/<pid>/fd - List of symlinks to open files 2249------------------------------------------------------- 2250This directory contains symbolic links which represent open files 2251the process is maintaining. Example output:: 2252 2253 lr-x------ 1 root root 64 Sep 20 17:53 0 -> /dev/null 2254 l-wx------ 1 root root 64 Sep 20 17:53 1 -> /dev/null 2255 lrwx------ 1 root root 64 Sep 20 17:53 10 -> 'socket:[12539]' 2256 lrwx------ 1 root root 64 Sep 20 17:53 11 -> 'socket:[12540]' 2257 lrwx------ 1 root root 64 Sep 20 17:53 12 -> 'socket:[12542]' 2258 2259The number of open files for the process is stored in 'size' member 2260of stat() output for /proc/<pid>/fd for fast access. 2261------------------------------------------------------- 2262 22633.14 /proc/<pid/ksm_stat - Information about the process's ksm status 2264--------------------------------------------------------------------- 2265When CONFIG_KSM is enabled, each process has this file which displays 2266the information of ksm merging status. 2267 2268Example 2269~~~~~~~ 2270 2271:: 2272 2273 / # cat /proc/self/ksm_stat 2274 ksm_rmap_items 0 2275 ksm_zero_pages 0 2276 ksm_merging_pages 0 2277 ksm_process_profit 0 2278 ksm_merge_any: no 2279 ksm_mergeable: no 2280 2281Description 2282~~~~~~~~~~~ 2283 2284ksm_rmap_items 2285^^^^^^^^^^^^^^ 2286 2287The number of ksm_rmap_item structures in use. The structure 2288ksm_rmap_item stores the reverse mapping information for virtual 2289addresses. KSM will generate a ksm_rmap_item for each ksm-scanned page of 2290the process. 2291 2292ksm_zero_pages 2293^^^^^^^^^^^^^^ 2294 2295When /sys/kernel/mm/ksm/use_zero_pages is enabled, it represent how many 2296empty pages are merged with kernel zero pages by KSM. 2297 2298ksm_merging_pages 2299^^^^^^^^^^^^^^^^^ 2300 2301It represents how many pages of this process are involved in KSM merging 2302(not including ksm_zero_pages). It is the same with what 2303/proc/<pid>/ksm_merging_pages shows. 2304 2305ksm_process_profit 2306^^^^^^^^^^^^^^^^^^ 2307 2308The profit that KSM brings (Saved bytes). KSM can save memory by merging 2309identical pages, but also can consume additional memory, because it needs 2310to generate a number of rmap_items to save each scanned page's brief rmap 2311information. Some of these pages may be merged, but some may not be abled 2312to be merged after being checked several times, which are unprofitable 2313memory consumed. 2314 2315ksm_merge_any 2316^^^^^^^^^^^^^ 2317 2318It specifies whether the process's 'mm is added by prctl() into the 2319candidate list of KSM or not, and if KSM scanning is fully enabled at 2320process level. 2321 2322ksm_mergeable 2323^^^^^^^^^^^^^ 2324 2325It specifies whether any VMAs of the process''s mms are currently 2326applicable to KSM. 2327 2328More information about KSM can be found in 2329Documentation/admin-guide/mm/ksm.rst. 2330 2331 2332Chapter 4: Configuring procfs 2333============================= 2334 23354.1 Mount options 2336--------------------- 2337 2338The following mount options are supported: 2339 2340 ========= ======================================================== 2341 hidepid= Set /proc/<pid>/ access mode. 2342 gid= Set the group authorized to learn processes information. 2343 subset= Show only the specified subset of procfs. 2344 ========= ======================================================== 2345 2346hidepid=off or hidepid=0 means classic mode - everybody may access all 2347/proc/<pid>/ directories (default). 2348 2349hidepid=noaccess or hidepid=1 means users may not access any /proc/<pid>/ 2350directories but their own. Sensitive files like cmdline, sched*, status are now 2351protected against other users. This makes it impossible to learn whether any 2352user runs specific program (given the program doesn't reveal itself by its 2353behaviour). As an additional bonus, as /proc/<pid>/cmdline is unaccessible for 2354other users, poorly written programs passing sensitive information via program 2355arguments are now protected against local eavesdroppers. 2356 2357hidepid=invisible or hidepid=2 means hidepid=1 plus all /proc/<pid>/ will be 2358fully invisible to other users. It doesn't mean that it hides a fact whether a 2359process with a specific pid value exists (it can be learned by other means, e.g. 2360by "kill -0 $PID"), but it hides process's uid and gid, which may be learned by 2361stat()'ing /proc/<pid>/ otherwise. It greatly complicates an intruder's task of 2362gathering information about running processes, whether some daemon runs with 2363elevated privileges, whether other user runs some sensitive program, whether 2364other users run any program at all, etc. 2365 2366hidepid=ptraceable or hidepid=4 means that procfs should only contain 2367/proc/<pid>/ directories that the caller can ptrace. 2368 2369gid= defines a group authorized to learn processes information otherwise 2370prohibited by hidepid=. If you use some daemon like identd which needs to learn 2371information about processes information, just add identd to this group. 2372 2373subset=pid hides all top level files and directories in the procfs that 2374are not related to tasks. 2375 2376Chapter 5: Filesystem behavior 2377============================== 2378 2379Originally, before the advent of pid namespace, procfs was a global file 2380system. It means that there was only one procfs instance in the system. 2381 2382When pid namespace was added, a separate procfs instance was mounted in 2383each pid namespace. So, procfs mount options are global among all 2384mountpoints within the same namespace:: 2385 2386 # grep ^proc /proc/mounts 2387 proc /proc proc rw,relatime,hidepid=2 0 0 2388 2389 # strace -e mount mount -o hidepid=1 -t proc proc /tmp/proc 2390 mount("proc", "/tmp/proc", "proc", 0, "hidepid=1") = 0 2391 +++ exited with 0 +++ 2392 2393 # grep ^proc /proc/mounts 2394 proc /proc proc rw,relatime,hidepid=2 0 0 2395 proc /tmp/proc proc rw,relatime,hidepid=2 0 0 2396 2397and only after remounting procfs mount options will change at all 2398mountpoints:: 2399 2400 # mount -o remount,hidepid=1 -t proc proc /tmp/proc 2401 2402 # grep ^proc /proc/mounts 2403 proc /proc proc rw,relatime,hidepid=1 0 0 2404 proc /tmp/proc proc rw,relatime,hidepid=1 0 0 2405 2406This behavior is different from the behavior of other filesystems. 2407 2408The new procfs behavior is more like other filesystems. Each procfs mount 2409creates a new procfs instance. Mount options affect own procfs instance. 2410It means that it became possible to have several procfs instances 2411displaying tasks with different filtering options in one pid namespace:: 2412 2413 # mount -o hidepid=invisible -t proc proc /proc 2414 # mount -o hidepid=noaccess -t proc proc /tmp/proc 2415 # grep ^proc /proc/mounts 2416 proc /proc proc rw,relatime,hidepid=invisible 0 0 2417 proc /tmp/proc proc rw,relatime,hidepid=noaccess 0 0 2418