1.. SPDX-License-Identifier: GPL-2.0 2 3==================== 4The /proc Filesystem 5==================== 6 7===================== ======================================= ================ 8/proc/sys Terrehon Bowden <[email protected]>, October 7 1999 9 Bodo Bauer <[email protected]> 102.4.x update Jorge Nerin <[email protected]> November 14 2000 11move /proc/sys Shen Feng <[email protected]> April 1 2009 12fixes/update part 1.1 Stefani Seibold <[email protected]> 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 64This documentation is part of a soon (or so we hope) to be released book on 65the SuSE Linux distribution. As there is no complete documentation for the 66/proc file system and we've used many freely available sources to write these 67chapters, it seems only fair to give the work back to the Linux community. 68This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm 69afraid it's still far from complete, but we hope it will be useful. As far as 70we know, it is the first 'all-in-one' document about the /proc file system. It 71is focused on the Intel x86 hardware, so if you are looking for PPC, ARM, 72SPARC, AXP, etc., features, you probably won't find what you are looking for. 73It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But 74additions and patches are welcome and will be added to this document if you 75mail them to Bodo. 76 77We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of 78other people for help compiling this documentation. We'd also like to extend a 79special thank you to Andi Kleen for documentation, which we relied on heavily 80to create this document, as well as the additional information he provided. 81Thanks to everybody else who contributed source or docs to the Linux kernel 82and helped create a great piece of software... :) 83 84If you have any comments, corrections or additions, please don't hesitate to 85contact Bodo Bauer at [email protected]. We'll be happy to add them to this 86document. 87 88The latest version of this document is available online at 89https://www.kernel.org/doc/html/latest/filesystems/proc.html 90 91If the above direction does not works for you, you could try the kernel 92mailing list at [email protected] and/or try to reach me at 93[email protected]. 94 950.2 Legal Stuff 96--------------- 97 98We don't guarantee the correctness of this document, and if you come to us 99complaining about how you screwed up your system because of incorrect 100documentation, we won't feel responsible... 101 102Chapter 1: Collecting System Information 103======================================== 104 105In This Chapter 106--------------- 107* Investigating the properties of the pseudo file system /proc and its 108 ability to provide information on the running Linux system 109* Examining /proc's structure 110* Uncovering various information about the kernel and the processes running 111 on the system 112 113------------------------------------------------------------------------------ 114 115The proc file system acts as an interface to internal data structures in the 116kernel. It can be used to obtain information about the system and to change 117certain kernel parameters at runtime (sysctl). 118 119First, we'll take a look at the read-only parts of /proc. In Chapter 2, we 120show you how you can use /proc/sys to change settings. 121 1221.1 Process-Specific Subdirectories 123----------------------------------- 124 125The directory /proc contains (among other things) one subdirectory for each 126process running on the system, which is named after the process ID (PID). 127 128The link 'self' points to the process reading the file system. Each process 129subdirectory has the entries listed in Table 1-1. 130 131Note that an open file descriptor to /proc/<pid> or to any of its 132contained files or subdirectories does not prevent <pid> being reused 133for some other process in the event that <pid> exits. Operations on 134open /proc/<pid> file descriptors corresponding to dead processes 135never act on any new process that the kernel may, through chance, have 136also assigned the process ID <pid>. Instead, operations on these FDs 137usually fail with ESRCH. 138 139.. table:: Table 1-1: Process specific entries in /proc 140 141 ============= =============================================================== 142 File Content 143 ============= =============================================================== 144 clear_refs Clears page referenced bits shown in smaps output 145 cmdline Command line arguments 146 cpu Current and last cpu in which it was executed (2.4)(smp) 147 cwd Link to the current working directory 148 environ Values of environment variables 149 exe Link to the executable of this process 150 fd Directory, which contains all file descriptors 151 maps Memory maps to executables and library files (2.4) 152 mem Memory held by this process 153 root Link to the root directory of this process 154 stat Process status 155 statm Process memory status information 156 status Process status in human readable form 157 wchan Present with CONFIG_KALLSYMS=y: it shows the kernel function 158 symbol the task is blocked in - or "0" if not blocked. 159 pagemap Page table 160 stack Report full stack trace, enable via CONFIG_STACKTRACE 161 smaps An extension based on maps, showing the memory consumption of 162 each mapping and flags associated with it 163 smaps_rollup Accumulated smaps stats for all mappings of the process. This 164 can be derived from smaps, but is faster and more convenient 165 numa_maps An extension based on maps, showing the memory locality and 166 binding policy as well as mem usage (in pages) of each mapping. 167 ============= =============================================================== 168 169For example, to get the status information of a process, all you have to do is 170read the file /proc/PID/status:: 171 172 >cat /proc/self/status 173 Name: cat 174 State: R (running) 175 Tgid: 5452 176 Pid: 5452 177 PPid: 743 178 TracerPid: 0 (2.4) 179 Uid: 501 501 501 501 180 Gid: 100 100 100 100 181 FDSize: 256 182 Groups: 100 14 16 183 Kthread: 0 184 VmPeak: 5004 kB 185 VmSize: 5004 kB 186 VmLck: 0 kB 187 VmHWM: 476 kB 188 VmRSS: 476 kB 189 RssAnon: 352 kB 190 RssFile: 120 kB 191 RssShmem: 4 kB 192 VmData: 156 kB 193 VmStk: 88 kB 194 VmExe: 68 kB 195 VmLib: 1412 kB 196 VmPTE: 20 kb 197 VmSwap: 0 kB 198 HugetlbPages: 0 kB 199 CoreDumping: 0 200 THP_enabled: 1 201 Threads: 1 202 SigQ: 0/28578 203 SigPnd: 0000000000000000 204 ShdPnd: 0000000000000000 205 SigBlk: 0000000000000000 206 SigIgn: 0000000000000000 207 SigCgt: 0000000000000000 208 CapInh: 00000000fffffeff 209 CapPrm: 0000000000000000 210 CapEff: 0000000000000000 211 CapBnd: ffffffffffffffff 212 CapAmb: 0000000000000000 213 NoNewPrivs: 0 214 Seccomp: 0 215 Speculation_Store_Bypass: thread vulnerable 216 SpeculationIndirectBranch: conditional enabled 217 voluntary_ctxt_switches: 0 218 nonvoluntary_ctxt_switches: 1 219 220This shows you nearly the same information you would get if you viewed it with 221the ps command. In fact, ps uses the proc file system to obtain its 222information. But you get a more detailed view of the process by reading the 223file /proc/PID/status. It fields are described in table 1-2. 224 225The statm file contains more detailed information about the process 226memory usage. Its seven fields are explained in Table 1-3. The stat file 227contains detailed information about the process itself. Its fields are 228explained in Table 1-4. 229 230(for SMP CONFIG users) 231 232For making accounting scalable, RSS related information are handled in an 233asynchronous manner and the value may not be very precise. To see a precise 234snapshot of a moment, you can see /proc/<pid>/smaps file and scan page table. 235It's slow but very precise. 236 237.. table:: Table 1-2: Contents of the status fields (as of 4.19) 238 239 ========================== =================================================== 240 Field Content 241 ========================== =================================================== 242 Name filename of the executable 243 Umask file mode creation mask 244 State state (R is running, S is sleeping, D is sleeping 245 in an uninterruptible wait, Z is zombie, 246 T is traced or stopped) 247 Tgid thread group ID 248 Ngid NUMA group ID (0 if none) 249 Pid process id 250 PPid process id of the parent process 251 TracerPid PID of process tracing this process (0 if not, or 252 the tracer is outside of the current pid namespace) 253 Uid Real, effective, saved set, and file system UIDs 254 Gid Real, effective, saved set, and file system GIDs 255 FDSize number of file descriptor slots currently allocated 256 Groups supplementary group list 257 NStgid descendant namespace thread group ID hierarchy 258 NSpid descendant namespace process ID hierarchy 259 NSpgid descendant namespace process group ID hierarchy 260 NSsid descendant namespace session ID hierarchy 261 Kthread kernel thread flag, 1 is yes, 0 is no 262 VmPeak peak virtual memory size 263 VmSize total program size 264 VmLck locked memory size 265 VmPin pinned memory size 266 VmHWM peak resident set size ("high water mark") 267 VmRSS size of memory portions. It contains the three 268 following parts 269 (VmRSS = RssAnon + RssFile + RssShmem) 270 RssAnon size of resident anonymous memory 271 RssFile size of resident file mappings 272 RssShmem size of resident shmem memory (includes SysV shm, 273 mapping of tmpfs and shared anonymous mappings) 274 VmData size of private data segments 275 VmStk size of stack segments 276 VmExe size of text segment 277 VmLib size of shared library code 278 VmPTE size of page table entries 279 VmSwap amount of swap used by anonymous private data 280 (shmem swap usage is not included) 281 HugetlbPages size of hugetlb memory portions 282 CoreDumping process's memory is currently being dumped 283 (killing the process may lead to a corrupted core) 284 THP_enabled process is allowed to use THP (returns 0 when 285 PR_SET_THP_DISABLE is set on the process 286 Threads number of threads 287 SigQ number of signals queued/max. number for queue 288 SigPnd bitmap of pending signals for the thread 289 ShdPnd bitmap of shared pending signals for the process 290 SigBlk bitmap of blocked signals 291 SigIgn bitmap of ignored signals 292 SigCgt bitmap of caught signals 293 CapInh bitmap of inheritable capabilities 294 CapPrm bitmap of permitted capabilities 295 CapEff bitmap of effective capabilities 296 CapBnd bitmap of capabilities bounding set 297 CapAmb bitmap of ambient capabilities 298 NoNewPrivs no_new_privs, like prctl(PR_GET_NO_NEW_PRIV, ...) 299 Seccomp seccomp mode, like prctl(PR_GET_SECCOMP, ...) 300 Speculation_Store_Bypass speculative store bypass mitigation status 301 SpeculationIndirectBranch indirect branch speculation mode 302 Cpus_allowed mask of CPUs on which this process may run 303 Cpus_allowed_list Same as previous, but in "list format" 304 Mems_allowed mask of memory nodes allowed to this process 305 Mems_allowed_list Same as previous, but in "list format" 306 voluntary_ctxt_switches number of voluntary context switches 307 nonvoluntary_ctxt_switches number of non voluntary context switches 308 ========================== =================================================== 309 310 311.. table:: Table 1-3: Contents of the statm fields (as of 2.6.8-rc3) 312 313 ======== =============================== ============================== 314 Field Content 315 ======== =============================== ============================== 316 size total program size (pages) (same as VmSize in status) 317 resident size of memory portions (pages) (same as VmRSS in status) 318 shared number of pages that are shared (i.e. backed by a file, same 319 as RssFile+RssShmem in status) 320 trs number of pages that are 'code' (not including libs; broken, 321 includes data segment) 322 lrs number of pages of library (always 0 on 2.6) 323 drs number of pages of data/stack (including libs; broken, 324 includes library text) 325 dt number of dirty pages (always 0 on 2.6) 326 ======== =============================== ============================== 327 328 329.. table:: Table 1-4: Contents of the stat fields (as of 2.6.30-rc7) 330 331 ============= =============================================================== 332 Field Content 333 ============= =============================================================== 334 pid process id 335 tcomm filename of the executable 336 state state (R is running, S is sleeping, D is sleeping in an 337 uninterruptible wait, Z is zombie, T is traced or stopped) 338 ppid process id of the parent process 339 pgrp pgrp of the process 340 sid session id 341 tty_nr tty the process uses 342 tty_pgrp pgrp of the tty 343 flags task flags 344 min_flt number of minor faults 345 cmin_flt number of minor faults with child's 346 maj_flt number of major faults 347 cmaj_flt number of major faults with child's 348 utime user mode jiffies 349 stime kernel mode jiffies 350 cutime user mode jiffies with child's 351 cstime kernel mode jiffies with child's 352 priority priority level 353 nice nice level 354 num_threads number of threads 355 it_real_value (obsolete, always 0) 356 start_time time the process started after system boot 357 vsize virtual memory size 358 rss resident set memory size 359 rsslim current limit in bytes on the rss 360 start_code address above which program text can run 361 end_code address below which program text can run 362 start_stack address of the start of the main process stack 363 esp current value of ESP 364 eip current value of EIP 365 pending bitmap of pending signals 366 blocked bitmap of blocked signals 367 sigign bitmap of ignored signals 368 sigcatch bitmap of caught signals 369 0 (place holder, used to be the wchan address, 370 use /proc/PID/wchan instead) 371 0 (place holder) 372 0 (place holder) 373 exit_signal signal to send to parent thread on exit 374 task_cpu which CPU the task is scheduled on 375 rt_priority realtime priority 376 policy scheduling policy (man sched_setscheduler) 377 blkio_ticks time spent waiting for block IO 378 gtime guest time of the task in jiffies 379 cgtime guest time of the task children in jiffies 380 start_data address above which program data+bss is placed 381 end_data address below which program data+bss is placed 382 start_brk address above which program heap can be expanded with brk() 383 arg_start address above which program command line is placed 384 arg_end address below which program command line is placed 385 env_start address above which program environment is placed 386 env_end address below which program environment is placed 387 exit_code the thread's exit_code in the form reported by the waitpid 388 system call 389 ============= =============================================================== 390 391The /proc/PID/maps file contains the currently mapped memory regions and 392their access permissions. 393 394The format is:: 395 396 address perms offset dev inode pathname 397 398 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test 399 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test 400 0804a000-0806b000 rw-p 00000000 00:00 0 [heap] 401 a7cb1000-a7cb2000 ---p 00000000 00:00 0 402 a7cb2000-a7eb2000 rw-p 00000000 00:00 0 403 a7eb2000-a7eb3000 ---p 00000000 00:00 0 404 a7eb3000-a7ed5000 rw-p 00000000 00:00 0 405 a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6 406 a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6 407 a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6 408 a800b000-a800e000 rw-p 00000000 00:00 0 409 a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0 410 a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0 411 a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0 412 a8024000-a8027000 rw-p 00000000 00:00 0 413 a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2 414 a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2 415 a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2 416 aff35000-aff4a000 rw-p 00000000 00:00 0 [stack] 417 ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso] 418 419where "address" is the address space in the process that it occupies, "perms" 420is a set of permissions:: 421 422 r = read 423 w = write 424 x = execute 425 s = shared 426 p = private (copy on write) 427 428"offset" is the offset into the mapping, "dev" is the device (major:minor), and 429"inode" is the inode on that device. 0 indicates that no inode is associated 430with the memory region, as the case would be with BSS (uninitialized data). 431The "pathname" shows the name associated file for this mapping. If the mapping 432is not associated with a file: 433 434 =================== =========================================== 435 [heap] the heap of the program 436 [stack] the stack of the main process 437 [vdso] the "virtual dynamic shared object", 438 the kernel system call handler 439 [anon:<name>] a private anonymous mapping that has been 440 named by userspace 441 [anon_shmem:<name>] an anonymous shared memory mapping that has 442 been named by userspace 443 =================== =========================================== 444 445 or if empty, the mapping is anonymous. 446 447Starting with 6.11 kernel, /proc/PID/maps provides an alternative 448ioctl()-based API that gives ability to flexibly and efficiently query and 449filter individual VMAs. This interface is binary and is meant for more 450efficient and easy programmatic use. `struct procmap_query`, defined in 451linux/fs.h UAPI header, serves as an input/output argument to the 452`PROCMAP_QUERY` ioctl() command. See comments in linus/fs.h UAPI header for 453details on query semantics, supported flags, data returned, and general API 454usage information. 455 456The /proc/PID/smaps is an extension based on maps, showing the memory 457consumption for each of the process's mappings. For each mapping (aka Virtual 458Memory Area, or VMA) there is a series of lines such as the following:: 459 460 08048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash 461 462 Size: 1084 kB 463 KernelPageSize: 4 kB 464 MMUPageSize: 4 kB 465 Rss: 892 kB 466 Pss: 374 kB 467 Pss_Dirty: 0 kB 468 Shared_Clean: 892 kB 469 Shared_Dirty: 0 kB 470 Private_Clean: 0 kB 471 Private_Dirty: 0 kB 472 Referenced: 892 kB 473 Anonymous: 0 kB 474 KSM: 0 kB 475 LazyFree: 0 kB 476 AnonHugePages: 0 kB 477 ShmemPmdMapped: 0 kB 478 Shared_Hugetlb: 0 kB 479 Private_Hugetlb: 0 kB 480 Swap: 0 kB 481 SwapPss: 0 kB 482 KernelPageSize: 4 kB 483 MMUPageSize: 4 kB 484 Locked: 0 kB 485 THPeligible: 0 486 VmFlags: rd ex mr mw me dw 487 488The first of these lines shows the same information as is displayed for 489the mapping in /proc/PID/maps. Following lines show the size of the 490mapping (size); the size of each page allocated when backing a VMA 491(KernelPageSize), which is usually the same as the size in the page table 492entries; the page size used by the MMU when backing a VMA (in most cases, 493the same as KernelPageSize); the amount of the mapping that is currently 494resident in RAM (RSS); the process's proportional share of this mapping 495(PSS); and the number of clean and dirty shared and private pages in the 496mapping. 497 498The "proportional set size" (PSS) of a process is the count of pages it has 499in memory, where each page is divided by the number of processes sharing it. 500So if a process has 1000 pages all to itself, and 1000 shared with one other 501process, its PSS will be 1500. "Pss_Dirty" is the portion of PSS which 502consists of dirty pages. ("Pss_Clean" is not included, but it can be 503calculated by subtracting "Pss_Dirty" from "Pss".) 504 505Note that even a page which is part of a MAP_SHARED mapping, but has only 506a single pte mapped, i.e. is currently used by only one process, is accounted 507as private and not as shared. 508 509"Referenced" indicates the amount of memory currently marked as referenced or 510accessed. 511 512"Anonymous" shows the amount of memory that does not belong to any file. Even 513a mapping associated with a file may contain anonymous pages: when MAP_PRIVATE 514and a page is modified, the file page is replaced by a private anonymous copy. 515 516"KSM" reports how many of the pages are KSM pages. Note that KSM-placed zeropages 517are not included, only actual KSM pages. 518 519"LazyFree" shows the amount of memory which is marked by madvise(MADV_FREE). 520The memory isn't freed immediately with madvise(). It's freed in memory 521pressure if the memory is clean. Please note that the printed value might 522be lower than the real value due to optimizations used in the current 523implementation. If this is not desirable please file a bug report. 524 525"AnonHugePages" shows the amount of memory backed by transparent hugepage. 526 527"ShmemPmdMapped" shows the amount of shared (shmem/tmpfs) memory backed by 528huge pages. 529 530"Shared_Hugetlb" and "Private_Hugetlb" show the amounts of memory backed by 531hugetlbfs page which is *not* counted in "RSS" or "PSS" field for historical 532reasons. And these are not included in {Shared,Private}_{Clean,Dirty} field. 533 534"Swap" shows how much would-be-anonymous memory is also used, but out on swap. 535 536For shmem mappings, "Swap" includes also the size of the mapped (and not 537replaced by copy-on-write) part of the underlying shmem object out on swap. 538"SwapPss" shows proportional swap share of this mapping. Unlike "Swap", this 539does not take into account swapped out page of underlying shmem objects. 540"Locked" indicates whether the mapping is locked in memory or not. 541 542"THPeligible" indicates whether the mapping is eligible for allocating 543naturally aligned THP pages of any currently enabled size. 1 if true, 0 544otherwise. 545 546"VmFlags" field deserves a separate description. This member represents the 547kernel flags associated with the particular virtual memory area in two letter 548encoded manner. The codes are the following: 549 550 == ======================================= 551 rd readable 552 wr writeable 553 ex executable 554 sh shared 555 mr may read 556 mw may write 557 me may execute 558 ms may share 559 gd stack segment growns down 560 pf pure PFN range 561 dw disabled write to the mapped file 562 lo pages are locked in memory 563 io memory mapped I/O area 564 sr sequential read advise provided 565 rr random read advise provided 566 dc do not copy area on fork 567 de do not expand area on remapping 568 ac area is accountable 569 nr swap space is not reserved for the area 570 ht area uses huge tlb pages 571 sf synchronous page fault 572 ar architecture specific flag 573 wf wipe on fork 574 dd do not include area into core dump 575 sd soft dirty flag 576 mm mixed map area 577 hg huge page advise flag 578 nh no huge page advise flag 579 mg mergeable advise flag 580 bt arm64 BTI guarded page 581 mt arm64 MTE allocation tags are enabled 582 um userfaultfd missing tracking 583 uw userfaultfd wr-protect tracking 584 ss shadow/guarded control stack page 585 sl sealed 586 == ======================================= 587 588Note that there is no guarantee that every flag and associated mnemonic will 589be present in all further kernel releases. Things get changed, the flags may 590be vanished or the reverse -- new added. Interpretation of their meaning 591might change in future as well. So each consumer of these flags has to 592follow each specific kernel version for the exact semantic. 593 594This file is only present if the CONFIG_MMU kernel configuration option is 595enabled. 596 597Note: reading /proc/PID/maps or /proc/PID/smaps is inherently racy (consistent 598output can be achieved only in the single read call). 599 600This typically manifests when doing partial reads of these files while the 601memory map is being modified. Despite the races, we do provide the following 602guarantees: 603 6041) The mapped addresses never go backwards, which implies no two 605 regions will ever overlap. 6062) If there is something at a given vaddr during the entirety of the 607 life of the smaps/maps walk, there will be some output for it. 608 609The /proc/PID/smaps_rollup file includes the same fields as /proc/PID/smaps, 610but their values are the sums of the corresponding values for all mappings of 611the process. Additionally, it contains these fields: 612 613- Pss_Anon 614- Pss_File 615- Pss_Shmem 616 617They represent the proportional shares of anonymous, file, and shmem pages, as 618described for smaps above. These fields are omitted in smaps since each 619mapping identifies the type (anon, file, or shmem) of all pages it contains. 620Thus all information in smaps_rollup can be derived from smaps, but at a 621significantly higher cost. 622 623The /proc/PID/clear_refs is used to reset the PG_Referenced and ACCESSED/YOUNG 624bits on both physical and virtual pages associated with a process, and the 625soft-dirty bit on pte (see Documentation/admin-guide/mm/soft-dirty.rst 626for details). 627To clear the bits for all the pages associated with the process:: 628 629 > echo 1 > /proc/PID/clear_refs 630 631To clear the bits for the anonymous pages associated with the process:: 632 633 > echo 2 > /proc/PID/clear_refs 634 635To clear the bits for the file mapped pages associated with the process:: 636 637 > echo 3 > /proc/PID/clear_refs 638 639To clear the soft-dirty bit:: 640 641 > echo 4 > /proc/PID/clear_refs 642 643To reset the peak resident set size ("high water mark") to the process's 644current value:: 645 646 > echo 5 > /proc/PID/clear_refs 647 648Any other value written to /proc/PID/clear_refs will have no effect. 649 650The /proc/pid/pagemap gives the PFN, which can be used to find the pageflags 651using /proc/kpageflags and number of times a page is mapped using 652/proc/kpagecount. For detailed explanation, see 653Documentation/admin-guide/mm/pagemap.rst. 654 655The /proc/pid/numa_maps is an extension based on maps, showing the memory 656locality and binding policy, as well as the memory usage (in pages) of 657each mapping. The output follows a general format where mapping details get 658summarized separated by blank spaces, one mapping per each file line:: 659 660 address policy mapping details 661 662 00400000 default file=/usr/local/bin/app mapped=1 active=0 N3=1 kernelpagesize_kB=4 663 00600000 default file=/usr/local/bin/app anon=1 dirty=1 N3=1 kernelpagesize_kB=4 664 3206000000 default file=/lib64/ld-2.12.so mapped=26 mapmax=6 N0=24 N3=2 kernelpagesize_kB=4 665 320621f000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 666 3206220000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 667 3206221000 default anon=1 dirty=1 N3=1 kernelpagesize_kB=4 668 3206800000 default file=/lib64/libc-2.12.so mapped=59 mapmax=21 active=55 N0=41 N3=18 kernelpagesize_kB=4 669 320698b000 default file=/lib64/libc-2.12.so 670 3206b8a000 default file=/lib64/libc-2.12.so anon=2 dirty=2 N3=2 kernelpagesize_kB=4 671 3206b8e000 default file=/lib64/libc-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 672 3206b8f000 default anon=3 dirty=3 active=1 N3=3 kernelpagesize_kB=4 673 7f4dc10a2000 default anon=3 dirty=3 N3=3 kernelpagesize_kB=4 674 7f4dc10b4000 default anon=2 dirty=2 active=1 N3=2 kernelpagesize_kB=4 675 7f4dc1200000 default file=/anon_hugepage\040(deleted) huge anon=1 dirty=1 N3=1 kernelpagesize_kB=2048 676 7fff335f0000 default stack anon=3 dirty=3 N3=3 kernelpagesize_kB=4 677 7fff3369d000 default mapped=1 mapmax=35 active=0 N3=1 kernelpagesize_kB=4 678 679Where: 680 681"address" is the starting address for the mapping; 682 683"policy" reports the NUMA memory policy set for the mapping (see Documentation/admin-guide/mm/numa_memory_policy.rst); 684 685"mapping details" summarizes mapping data such as mapping type, page usage counters, 686node locality page counters (N0 == node0, N1 == node1, ...) and the kernel page 687size, in KB, that is backing the mapping up. 688 6891.2 Kernel data 690--------------- 691 692Similar to the process entries, the kernel data files give information about 693the running kernel. The files used to obtain this information are contained in 694/proc and are listed in Table 1-5. Not all of these will be present in your 695system. It depends on the kernel configuration and the loaded modules, which 696files are there, and which are missing. 697 698.. table:: Table 1-5: Kernel info in /proc 699 700 ============ =============================================================== 701 File Content 702 ============ =============================================================== 703 allocinfo Memory allocations profiling information 704 apm Advanced power management info 705 bootconfig Kernel command line obtained from boot config, 706 and, if there were kernel parameters from the 707 boot loader, a "# Parameters from bootloader:" 708 line followed by a line containing those 709 parameters prefixed by "# ". (5.5) 710 buddyinfo Kernel memory allocator information (see text) (2.5) 711 bus Directory containing bus specific information 712 cmdline Kernel command line, both from bootloader and embedded 713 in the kernel image 714 cpuinfo Info about the CPU 715 devices Available devices (block and character) 716 dma Used DMS channels 717 filesystems Supported filesystems 718 driver Various drivers grouped here, currently rtc (2.4) 719 execdomains Execdomains, related to security (2.4) 720 fb Frame Buffer devices (2.4) 721 fs File system parameters, currently nfs/exports (2.4) 722 ide Directory containing info about the IDE subsystem 723 interrupts Interrupt usage 724 iomem Memory map (2.4) 725 ioports I/O port usage 726 irq Masks for irq to cpu affinity (2.4)(smp?) 727 isapnp ISA PnP (Plug&Play) Info (2.4) 728 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4)) 729 kmsg Kernel messages 730 ksyms Kernel symbol table 731 loadavg Load average of last 1, 5 & 15 minutes; 732 number of processes currently runnable (running or on ready queue); 733 total number of processes in system; 734 last pid created. 735 All fields are separated by one space except "number of 736 processes currently runnable" and "total number of processes 737 in system", which are separated by a slash ('/'). Example: 738 0.61 0.61 0.55 3/828 22084 739 locks Kernel locks 740 meminfo Memory info 741 misc Miscellaneous 742 modules List of loaded modules 743 mounts Mounted filesystems 744 net Networking info (see text) 745 pagetypeinfo Additional page allocator information (see text) (2.5) 746 partitions Table of partitions known to the system 747 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/, 748 decoupled by lspci (2.4) 749 rtc Real time clock 750 scsi SCSI info (see text) 751 slabinfo Slab pool info 752 softirqs softirq usage 753 stat Overall statistics 754 swaps Swap space utilization 755 sys See chapter 2 756 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4) 757 tty Info of tty drivers 758 uptime Wall clock since boot, combined idle time of all cpus 759 version Kernel version 760 video bttv info of video resources (2.4) 761 vmallocinfo Show vmalloced areas 762 ============ =============================================================== 763 764You can, for example, check which interrupts are currently in use and what 765they are used for by looking in the file /proc/interrupts:: 766 767 > cat /proc/interrupts 768 CPU0 769 0: 8728810 XT-PIC timer 770 1: 895 XT-PIC keyboard 771 2: 0 XT-PIC cascade 772 3: 531695 XT-PIC aha152x 773 4: 2014133 XT-PIC serial 774 5: 44401 XT-PIC pcnet_cs 775 8: 2 XT-PIC rtc 776 11: 8 XT-PIC i82365 777 12: 182918 XT-PIC PS/2 Mouse 778 13: 1 XT-PIC fpu 779 14: 1232265 XT-PIC ide0 780 15: 7 XT-PIC ide1 781 NMI: 0 782 783In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the 784output of a SMP machine):: 785 786 > cat /proc/interrupts 787 788 CPU0 CPU1 789 0: 1243498 1214548 IO-APIC-edge timer 790 1: 8949 8958 IO-APIC-edge keyboard 791 2: 0 0 XT-PIC cascade 792 5: 11286 10161 IO-APIC-edge soundblaster 793 8: 1 0 IO-APIC-edge rtc 794 9: 27422 27407 IO-APIC-edge 3c503 795 12: 113645 113873 IO-APIC-edge PS/2 Mouse 796 13: 0 0 XT-PIC fpu 797 14: 22491 24012 IO-APIC-edge ide0 798 15: 2183 2415 IO-APIC-edge ide1 799 17: 30564 30414 IO-APIC-level eth0 800 18: 177 164 IO-APIC-level bttv 801 NMI: 2457961 2457959 802 LOC: 2457882 2457881 803 ERR: 2155 804 805NMI is incremented in this case because every timer interrupt generates a NMI 806(Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups. 807 808LOC is the local interrupt counter of the internal APIC of every CPU. 809 810ERR is incremented in the case of errors in the IO-APIC bus (the bus that 811connects the CPUs in a SMP system. This means that an error has been detected, 812the IO-APIC automatically retry the transmission, so it should not be a big 813problem, but you should read the SMP-FAQ. 814 815In 2.6.2* /proc/interrupts was expanded again. This time the goal was for 816/proc/interrupts to display every IRQ vector in use by the system, not 817just those considered 'most important'. The new vectors are: 818 819THR 820 interrupt raised when a machine check threshold counter 821 (typically counting ECC corrected errors of memory or cache) exceeds 822 a configurable threshold. Only available on some systems. 823 824TRM 825 a thermal event interrupt occurs when a temperature threshold 826 has been exceeded for the CPU. This interrupt may also be generated 827 when the temperature drops back to normal. 828 829SPU 830 a spurious interrupt is some interrupt that was raised then lowered 831 by some IO device before it could be fully processed by the APIC. Hence 832 the APIC sees the interrupt but does not know what device it came from. 833 For this case the APIC will generate the interrupt with a IRQ vector 834 of 0xff. This might also be generated by chipset bugs. 835 836RES, CAL, TLB 837 rescheduling, call and TLB flush interrupts are 838 sent from one CPU to another per the needs of the OS. Typically, 839 their statistics are used by kernel developers and interested users to 840 determine the occurrence of interrupts of the given type. 841 842The above IRQ vectors are displayed only when relevant. For example, 843the threshold vector does not exist on x86_64 platforms. Others are 844suppressed when the system is a uniprocessor. As of this writing, only 845i386 and x86_64 platforms support the new IRQ vector displays. 846 847Of some interest is the introduction of the /proc/irq directory to 2.4. 848It could be used to set IRQ to CPU affinity. This means that you can "hook" an 849IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the 850irq subdir is one subdir for each IRQ, and two files; default_smp_affinity and 851prof_cpu_mask. 852 853For example:: 854 855 > ls /proc/irq/ 856 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask 857 1 11 13 15 17 19 3 5 7 9 default_smp_affinity 858 > ls /proc/irq/0/ 859 smp_affinity 860 861smp_affinity is a bitmask, in which you can specify which CPUs can handle the 862IRQ. You can set it by doing:: 863 864 > echo 1 > /proc/irq/10/smp_affinity 865 866This means that only the first CPU will handle the IRQ, but you can also echo 8675 which means that only the first and third CPU can handle the IRQ. 868 869The contents of each smp_affinity file is the same by default:: 870 871 > cat /proc/irq/0/smp_affinity 872 ffffffff 873 874There is an alternate interface, smp_affinity_list which allows specifying 875a CPU range instead of a bitmask:: 876 877 > cat /proc/irq/0/smp_affinity_list 878 1024-1031 879 880The default_smp_affinity mask applies to all non-active IRQs, which are the 881IRQs which have not yet been allocated/activated, and hence which lack a 882/proc/irq/[0-9]* directory. 883 884The node file on an SMP system shows the node to which the device using the IRQ 885reports itself as being attached. This hardware locality information does not 886include information about any possible driver locality preference. 887 888prof_cpu_mask specifies which CPUs are to be profiled by the system wide 889profiler. Default value is ffffffff (all CPUs if there are only 32 of them). 890 891The way IRQs are routed is handled by the IO-APIC, and it's Round Robin 892between all the CPUs which are allowed to handle it. As usual the kernel has 893more info than you and does a better job than you, so the defaults are the 894best choice for almost everyone. [Note this applies only to those IO-APIC's 895that support "Round Robin" interrupt distribution.] 896 897There are three more important subdirectories in /proc: net, scsi, and sys. 898The general rule is that the contents, or even the existence of these 899directories, depend on your kernel configuration. If SCSI is not enabled, the 900directory scsi may not exist. The same is true with the net, which is there 901only when networking support is present in the running kernel. 902 903The slabinfo file gives information about memory usage at the slab level. 904Linux uses slab pools for memory management above page level in version 2.2. 905Commonly used objects have their own slab pool (such as network buffers, 906directory cache, and so on). 907 908:: 909 910 > cat /proc/buddyinfo 911 912 Node 0, zone DMA 0 4 5 4 4 3 ... 913 Node 0, zone Normal 1 0 0 1 101 8 ... 914 Node 0, zone HighMem 2 0 0 1 1 0 ... 915 916External fragmentation is a problem under some workloads, and buddyinfo is a 917useful tool for helping diagnose these problems. Buddyinfo will give you a 918clue as to how big an area you can safely allocate, or why a previous 919allocation failed. 920 921Each column represents the number of pages of a certain order which are 922available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in 923ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE 924available in ZONE_NORMAL, etc... 925 926More information relevant to external fragmentation can be found in 927pagetypeinfo:: 928 929 > cat /proc/pagetypeinfo 930 Page block order: 9 931 Pages per block: 512 932 933 Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 934 Node 0, zone DMA, type Unmovable 0 0 0 1 1 1 1 1 1 1 0 935 Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 936 Node 0, zone DMA, type Movable 1 1 2 1 2 1 1 0 1 0 2 937 Node 0, zone DMA, type Reserve 0 0 0 0 0 0 0 0 0 1 0 938 Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 939 Node 0, zone DMA32, type Unmovable 103 54 77 1 1 1 11 8 7 1 9 940 Node 0, zone DMA32, type Reclaimable 0 0 2 1 0 0 0 0 1 0 0 941 Node 0, zone DMA32, type Movable 169 152 113 91 77 54 39 13 6 1 452 942 Node 0, zone DMA32, type Reserve 1 2 2 2 2 0 1 1 1 1 0 943 Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 944 945 Number of blocks type Unmovable Reclaimable Movable Reserve Isolate 946 Node 0, zone DMA 2 0 5 1 0 947 Node 0, zone DMA32 41 6 967 2 0 948 949Fragmentation avoidance in the kernel works by grouping pages of different 950migrate types into the same contiguous regions of memory called page blocks. 951A page block is typically the size of the default hugepage size, e.g. 2MB on 952X86-64. By keeping pages grouped based on their ability to move, the kernel 953can reclaim pages within a page block to satisfy a high-order allocation. 954 955The pagetypinfo begins with information on the size of a page block. It 956then gives the same type of information as buddyinfo except broken down 957by migrate-type and finishes with details on how many page blocks of each 958type exist. 959 960If min_free_kbytes has been tuned correctly (recommendations made by hugeadm 961from libhugetlbfs https://github.com/libhugetlbfs/libhugetlbfs/), one can 962make an estimate of the likely number of huge pages that can be allocated 963at a given point in time. All the "Movable" blocks should be allocatable 964unless memory has been mlock()'d. Some of the Reclaimable blocks should 965also be allocatable although a lot of filesystem metadata may have to be 966reclaimed to achieve this. 967 968 969allocinfo 970~~~~~~~~~ 971 972Provides information about memory allocations at all locations in the code 973base. Each allocation in the code is identified by its source file, line 974number, module (if originates from a loadable module) and the function calling 975the allocation. The number of bytes allocated and number of calls at each 976location are reported. The first line indicates the version of the file, the 977second line is the header listing fields in the file. 978 979Example output. 980 981:: 982 983 > tail -n +3 /proc/allocinfo | sort -rn 984 127664128 31168 mm/page_ext.c:270 func:alloc_page_ext 985 56373248 4737 mm/slub.c:2259 func:alloc_slab_page 986 14880768 3633 mm/readahead.c:247 func:page_cache_ra_unbounded 987 14417920 3520 mm/mm_init.c:2530 func:alloc_large_system_hash 988 13377536 234 block/blk-mq.c:3421 func:blk_mq_alloc_rqs 989 11718656 2861 mm/filemap.c:1919 func:__filemap_get_folio 990 9192960 2800 kernel/fork.c:307 func:alloc_thread_stack_node 991 4206592 4 net/netfilter/nf_conntrack_core.c:2567 func:nf_ct_alloc_hashtable 992 4136960 1010 drivers/staging/ctagmod/ctagmod.c:20 [ctagmod] func:ctagmod_start 993 3940352 962 mm/memory.c:4214 func:alloc_anon_folio 994 2894464 22613 fs/kernfs/dir.c:615 func:__kernfs_new_node 995 ... 996 997 998meminfo 999~~~~~~~ 1000 1001Provides information about distribution and utilization of memory. This 1002varies by architecture and compile options. Some of the counters reported 1003here overlap. The memory reported by the non overlapping counters may not 1004add up to the overall memory usage and the difference for some workloads 1005can be substantial. In many cases there are other means to find out 1006additional memory using subsystem specific interfaces, for instance 1007/proc/net/sockstat for TCP memory allocations. 1008 1009Example output. You may not have all of these fields. 1010 1011:: 1012 1013 > cat /proc/meminfo 1014 1015 MemTotal: 32858820 kB 1016 MemFree: 21001236 kB 1017 MemAvailable: 27214312 kB 1018 Buffers: 581092 kB 1019 Cached: 5587612 kB 1020 SwapCached: 0 kB 1021 Active: 3237152 kB 1022 Inactive: 7586256 kB 1023 Active(anon): 94064 kB 1024 Inactive(anon): 4570616 kB 1025 Active(file): 3143088 kB 1026 Inactive(file): 3015640 kB 1027 Unevictable: 0 kB 1028 Mlocked: 0 kB 1029 SwapTotal: 0 kB 1030 SwapFree: 0 kB 1031 Zswap: 1904 kB 1032 Zswapped: 7792 kB 1033 Dirty: 12 kB 1034 Writeback: 0 kB 1035 AnonPages: 4654780 kB 1036 Mapped: 266244 kB 1037 Shmem: 9976 kB 1038 KReclaimable: 517708 kB 1039 Slab: 660044 kB 1040 SReclaimable: 517708 kB 1041 SUnreclaim: 142336 kB 1042 KernelStack: 11168 kB 1043 PageTables: 20540 kB 1044 SecPageTables: 0 kB 1045 NFS_Unstable: 0 kB 1046 Bounce: 0 kB 1047 WritebackTmp: 0 kB 1048 CommitLimit: 16429408 kB 1049 Committed_AS: 7715148 kB 1050 VmallocTotal: 34359738367 kB 1051 VmallocUsed: 40444 kB 1052 VmallocChunk: 0 kB 1053 Percpu: 29312 kB 1054 EarlyMemtestBad: 0 kB 1055 HardwareCorrupted: 0 kB 1056 AnonHugePages: 4149248 kB 1057 ShmemHugePages: 0 kB 1058 ShmemPmdMapped: 0 kB 1059 FileHugePages: 0 kB 1060 FilePmdMapped: 0 kB 1061 CmaTotal: 0 kB 1062 CmaFree: 0 kB 1063 HugePages_Total: 0 1064 HugePages_Free: 0 1065 HugePages_Rsvd: 0 1066 HugePages_Surp: 0 1067 Hugepagesize: 2048 kB 1068 Hugetlb: 0 kB 1069 DirectMap4k: 401152 kB 1070 DirectMap2M: 10008576 kB 1071 DirectMap1G: 24117248 kB 1072 1073MemTotal 1074 Total usable RAM (i.e. physical RAM minus a few reserved 1075 bits and the kernel binary code) 1076MemFree 1077 Total free RAM. On highmem systems, the sum of LowFree+HighFree 1078MemAvailable 1079 An estimate of how much memory is available for starting new 1080 applications, without swapping. Calculated from MemFree, 1081 SReclaimable, the size of the file LRU lists, and the low 1082 watermarks in each zone. 1083 The estimate takes into account that the system needs some 1084 page cache to function well, and that not all reclaimable 1085 slab will be reclaimable, due to items being in use. The 1086 impact of those factors will vary from system to system. 1087Buffers 1088 Relatively temporary storage for raw disk blocks 1089 shouldn't get tremendously large (20MB or so) 1090Cached 1091 In-memory cache for files read from the disk (the 1092 pagecache) as well as tmpfs & shmem. 1093 Doesn't include SwapCached. 1094SwapCached 1095 Memory that once was swapped out, is swapped back in but 1096 still also is in the swapfile (if memory is needed it 1097 doesn't need to be swapped out AGAIN because it is already 1098 in the swapfile. This saves I/O) 1099Active 1100 Memory that has been used more recently and usually not 1101 reclaimed unless absolutely necessary. 1102Inactive 1103 Memory which has been less recently used. It is more 1104 eligible to be reclaimed for other purposes 1105Unevictable 1106 Memory allocated for userspace which cannot be reclaimed, such 1107 as mlocked pages, ramfs backing pages, secret memfd pages etc. 1108Mlocked 1109 Memory locked with mlock(). 1110HighTotal, HighFree 1111 Highmem is all memory above ~860MB of physical memory. 1112 Highmem areas are for use by userspace programs, or 1113 for the pagecache. The kernel must use tricks to access 1114 this memory, making it slower to access than lowmem. 1115LowTotal, LowFree 1116 Lowmem is memory which can be used for everything that 1117 highmem can be used for, but it is also available for the 1118 kernel's use for its own data structures. Among many 1119 other things, it is where everything from the Slab is 1120 allocated. Bad things happen when you're out of lowmem. 1121SwapTotal 1122 total amount of swap space available 1123SwapFree 1124 Memory which has been evicted from RAM, and is temporarily 1125 on the disk 1126Zswap 1127 Memory consumed by the zswap backend (compressed size) 1128Zswapped 1129 Amount of anonymous memory stored in zswap (original size) 1130Dirty 1131 Memory which is waiting to get written back to the disk 1132Writeback 1133 Memory which is actively being written back to the disk 1134AnonPages 1135 Non-file backed pages mapped into userspace page tables 1136Mapped 1137 files which have been mmapped, such as libraries 1138Shmem 1139 Total memory used by shared memory (shmem) and tmpfs 1140KReclaimable 1141 Kernel allocations that the kernel will attempt to reclaim 1142 under memory pressure. Includes SReclaimable (below), and other 1143 direct allocations with a shrinker. 1144Slab 1145 in-kernel data structures cache 1146SReclaimable 1147 Part of Slab, that might be reclaimed, such as caches 1148SUnreclaim 1149 Part of Slab, that cannot be reclaimed on memory pressure 1150KernelStack 1151 Memory consumed by the kernel stacks of all tasks 1152PageTables 1153 Memory consumed by userspace page tables 1154SecPageTables 1155 Memory consumed by secondary page tables, this currently includes 1156 KVM mmu and IOMMU allocations on x86 and arm64. 1157NFS_Unstable 1158 Always zero. Previous counted pages which had been written to 1159 the server, but has not been committed to stable storage. 1160Bounce 1161 Memory used for block device "bounce buffers" 1162WritebackTmp 1163 Memory used by FUSE for temporary writeback buffers 1164CommitLimit 1165 Based on the overcommit ratio ('vm.overcommit_ratio'), 1166 this is the total amount of memory currently available to 1167 be allocated on the system. This limit is only adhered to 1168 if strict overcommit accounting is enabled (mode 2 in 1169 'vm.overcommit_memory'). 1170 1171 The CommitLimit is calculated with the following formula:: 1172 1173 CommitLimit = ([total RAM pages] - [total huge TLB pages]) * 1174 overcommit_ratio / 100 + [total swap pages] 1175 1176 For example, on a system with 1G of physical RAM and 7G 1177 of swap with a `vm.overcommit_ratio` of 30 it would 1178 yield a CommitLimit of 7.3G. 1179 1180 For more details, see the memory overcommit documentation 1181 in mm/overcommit-accounting. 1182Committed_AS 1183 The amount of memory presently allocated on the system. 1184 The committed memory is a sum of all of the memory which 1185 has been allocated by processes, even if it has not been 1186 "used" by them as of yet. A process which malloc()'s 1G 1187 of memory, but only touches 300M of it will show up as 1188 using 1G. This 1G is memory which has been "committed" to 1189 by the VM and can be used at any time by the allocating 1190 application. With strict overcommit enabled on the system 1191 (mode 2 in 'vm.overcommit_memory'), allocations which would 1192 exceed the CommitLimit (detailed above) will not be permitted. 1193 This is useful if one needs to guarantee that processes will 1194 not fail due to lack of memory once that memory has been 1195 successfully allocated. 1196VmallocTotal 1197 total size of vmalloc virtual address space 1198VmallocUsed 1199 amount of vmalloc area which is used 1200VmallocChunk 1201 largest contiguous block of vmalloc area which is free 1202Percpu 1203 Memory allocated to the percpu allocator used to back percpu 1204 allocations. This stat excludes the cost of metadata. 1205EarlyMemtestBad 1206 The amount of RAM/memory in kB, that was identified as corrupted 1207 by early memtest. If memtest was not run, this field will not 1208 be displayed at all. Size is never rounded down to 0 kB. 1209 That means if 0 kB is reported, you can safely assume 1210 there was at least one pass of memtest and none of the passes 1211 found a single faulty byte of RAM. 1212HardwareCorrupted 1213 The amount of RAM/memory in KB, the kernel identifies as 1214 corrupted. 1215AnonHugePages 1216 Non-file backed huge pages mapped into userspace page tables 1217ShmemHugePages 1218 Memory used by shared memory (shmem) and tmpfs allocated 1219 with huge pages 1220ShmemPmdMapped 1221 Shared memory mapped into userspace with huge pages 1222FileHugePages 1223 Memory used for filesystem data (page cache) allocated 1224 with huge pages 1225FilePmdMapped 1226 Page cache mapped into userspace with huge pages 1227CmaTotal 1228 Memory reserved for the Contiguous Memory Allocator (CMA) 1229CmaFree 1230 Free remaining memory in the CMA reserves 1231HugePages_Total, HugePages_Free, HugePages_Rsvd, HugePages_Surp, Hugepagesize, Hugetlb 1232 See Documentation/admin-guide/mm/hugetlbpage.rst. 1233DirectMap4k, DirectMap2M, DirectMap1G 1234 Breakdown of page table sizes used in the kernel's 1235 identity mapping of RAM 1236 1237vmallocinfo 1238~~~~~~~~~~~ 1239 1240Provides information about vmalloced/vmaped areas. One line per area, 1241containing the virtual address range of the area, size in bytes, 1242caller information of the creator, and optional information depending 1243on the kind of area: 1244 1245 ========== =================================================== 1246 pages=nr number of pages 1247 phys=addr if a physical address was specified 1248 ioremap I/O mapping (ioremap() and friends) 1249 vmalloc vmalloc() area 1250 vmap vmap()ed pages 1251 user VM_USERMAP area 1252 vpages buffer for pages pointers was vmalloced (huge area) 1253 N<node>=nr (Only on NUMA kernels) 1254 Number of pages allocated on memory node <node> 1255 ========== =================================================== 1256 1257:: 1258 1259 > cat /proc/vmallocinfo 1260 0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ... 1261 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128 1262 0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ... 1263 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64 1264 0xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f... 1265 phys=7fee8000 ioremap 1266 0xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f... 1267 phys=7fee7000 ioremap 1268 0xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210 1269 0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ... 1270 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3 1271 0xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ... 1272 pages=2 vmalloc N1=2 1273 0xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ... 1274 /0x130 [x_tables] pages=4 vmalloc N0=4 1275 0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ... 1276 pages=14 vmalloc N2=14 1277 0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ... 1278 pages=4 vmalloc N1=4 1279 0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ... 1280 pages=2 vmalloc N1=2 1281 0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ... 1282 pages=10 vmalloc N0=10 1283 1284 1285softirqs 1286~~~~~~~~ 1287 1288Provides counts of softirq handlers serviced since boot time, for each CPU. 1289 1290:: 1291 1292 > cat /proc/softirqs 1293 CPU0 CPU1 CPU2 CPU3 1294 HI: 0 0 0 0 1295 TIMER: 27166 27120 27097 27034 1296 NET_TX: 0 0 0 17 1297 NET_RX: 42 0 0 39 1298 BLOCK: 0 0 107 1121 1299 TASKLET: 0 0 0 290 1300 SCHED: 27035 26983 26971 26746 1301 HRTIMER: 0 0 0 0 1302 RCU: 1678 1769 2178 2250 1303 13041.3 Networking info in /proc/net 1305-------------------------------- 1306 1307The subdirectory /proc/net follows the usual pattern. Table 1-8 shows the 1308additional values you get for IP version 6 if you configure the kernel to 1309support this. Table 1-9 lists the files and their meaning. 1310 1311 1312.. table:: Table 1-8: IPv6 info in /proc/net 1313 1314 ========== ===================================================== 1315 File Content 1316 ========== ===================================================== 1317 udp6 UDP sockets (IPv6) 1318 tcp6 TCP sockets (IPv6) 1319 raw6 Raw device statistics (IPv6) 1320 igmp6 IP multicast addresses, which this host joined (IPv6) 1321 if_inet6 List of IPv6 interface addresses 1322 ipv6_route Kernel routing table for IPv6 1323 rt6_stats Global IPv6 routing tables statistics 1324 sockstat6 Socket statistics (IPv6) 1325 snmp6 Snmp data (IPv6) 1326 ========== ===================================================== 1327 1328.. table:: Table 1-9: Network info in /proc/net 1329 1330 ============= ================================================================ 1331 File Content 1332 ============= ================================================================ 1333 arp Kernel ARP table 1334 dev network devices with statistics 1335 dev_mcast the Layer2 multicast groups a device is listening too 1336 (interface index, label, number of references, number of bound 1337 addresses). 1338 dev_stat network device status 1339 ip_fwchains Firewall chain linkage 1340 ip_fwnames Firewall chain names 1341 ip_masq Directory containing the masquerading tables 1342 ip_masquerade Major masquerading table 1343 netstat Network statistics 1344 raw raw device statistics 1345 route Kernel routing table 1346 rpc Directory containing rpc info 1347 rt_cache Routing cache 1348 snmp SNMP data 1349 sockstat Socket statistics 1350 softnet_stat Per-CPU incoming packets queues statistics of online CPUs 1351 tcp TCP sockets 1352 udp UDP sockets 1353 unix UNIX domain sockets 1354 wireless Wireless interface data (Wavelan etc) 1355 igmp IP multicast addresses, which this host joined 1356 psched Global packet scheduler parameters. 1357 netlink List of PF_NETLINK sockets 1358 ip_mr_vifs List of multicast virtual interfaces 1359 ip_mr_cache List of multicast routing cache 1360 ============= ================================================================ 1361 1362You can use this information to see which network devices are available in 1363your system and how much traffic was routed over those devices:: 1364 1365 > cat /proc/net/dev 1366 Inter-|Receive |[... 1367 face |bytes packets errs drop fifo frame compressed multicast|[... 1368 lo: 908188 5596 0 0 0 0 0 0 [... 1369 ppp0:15475140 20721 410 0 0 410 0 0 [... 1370 eth0: 614530 7085 0 0 0 0 0 1 [... 1371 1372 ...] Transmit 1373 ...] bytes packets errs drop fifo colls carrier compressed 1374 ...] 908188 5596 0 0 0 0 0 0 1375 ...] 1375103 17405 0 0 0 0 0 0 1376 ...] 1703981 5535 0 0 0 3 0 0 1377 1378In addition, each Channel Bond interface has its own directory. For 1379example, the bond0 device will have a directory called /proc/net/bond0/. 1380It will contain information that is specific to that bond, such as the 1381current slaves of the bond, the link status of the slaves, and how 1382many times the slaves link has failed. 1383 13841.4 SCSI info 1385------------- 1386 1387If you have a SCSI or ATA host adapter in your system, you'll find a 1388subdirectory named after the driver for this adapter in /proc/scsi. 1389You'll also see a list of all recognized SCSI devices in /proc/scsi:: 1390 1391 >cat /proc/scsi/scsi 1392 Attached devices: 1393 Host: scsi0 Channel: 00 Id: 00 Lun: 00 1394 Vendor: IBM Model: DGHS09U Rev: 03E0 1395 Type: Direct-Access ANSI SCSI revision: 03 1396 Host: scsi0 Channel: 00 Id: 06 Lun: 00 1397 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04 1398 Type: CD-ROM ANSI SCSI revision: 02 1399 1400 1401The directory named after the driver has one file for each adapter found in 1402the system. These files contain information about the controller, including 1403the used IRQ and the IO address range. The amount of information shown is 1404dependent on the adapter you use. The example shows the output for an Adaptec 1405AHA-2940 SCSI adapter:: 1406 1407 > cat /proc/scsi/aic7xxx/0 1408 1409 Adaptec AIC7xxx driver version: 5.1.19/3.2.4 1410 Compile Options: 1411 TCQ Enabled By Default : Disabled 1412 AIC7XXX_PROC_STATS : Disabled 1413 AIC7XXX_RESET_DELAY : 5 1414 Adapter Configuration: 1415 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter 1416 Ultra Wide Controller 1417 PCI MMAPed I/O Base: 0xeb001000 1418 Adapter SEEPROM Config: SEEPROM found and used. 1419 Adaptec SCSI BIOS: Enabled 1420 IRQ: 10 1421 SCBs: Active 0, Max Active 2, 1422 Allocated 15, HW 16, Page 255 1423 Interrupts: 160328 1424 BIOS Control Word: 0x18b6 1425 Adapter Control Word: 0x005b 1426 Extended Translation: Enabled 1427 Disconnect Enable Flags: 0xffff 1428 Ultra Enable Flags: 0x0001 1429 Tag Queue Enable Flags: 0x0000 1430 Ordered Queue Tag Flags: 0x0000 1431 Default Tag Queue Depth: 8 1432 Tagged Queue By Device array for aic7xxx host instance 0: 1433 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255} 1434 Actual queue depth per device for aic7xxx host instance 0: 1435 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1} 1436 Statistics: 1437 (scsi0:0:0:0) 1438 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8 1439 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0) 1440 Total transfers 160151 (74577 reads and 85574 writes) 1441 (scsi0:0:6:0) 1442 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15 1443 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0) 1444 Total transfers 0 (0 reads and 0 writes) 1445 1446 14471.5 Parallel port info in /proc/parport 1448--------------------------------------- 1449 1450The directory /proc/parport contains information about the parallel ports of 1451your system. It has one subdirectory for each port, named after the port 1452number (0,1,2,...). 1453 1454These directories contain the four files shown in Table 1-10. 1455 1456 1457.. table:: Table 1-10: Files in /proc/parport 1458 1459 ========= ==================================================================== 1460 File Content 1461 ========= ==================================================================== 1462 autoprobe Any IEEE-1284 device ID information that has been acquired. 1463 devices list of the device drivers using that port. A + will appear by the 1464 name of the device currently using the port (it might not appear 1465 against any). 1466 hardware Parallel port's base address, IRQ line and DMA channel. 1467 irq IRQ that parport is using for that port. This is in a separate 1468 file to allow you to alter it by writing a new value in (IRQ 1469 number or none). 1470 ========= ==================================================================== 1471 14721.6 TTY info in /proc/tty 1473------------------------- 1474 1475Information about the available and actually used tty's can be found in the 1476directory /proc/tty. You'll find entries for drivers and line disciplines in 1477this directory, as shown in Table 1-11. 1478 1479 1480.. table:: Table 1-11: Files in /proc/tty 1481 1482 ============= ============================================== 1483 File Content 1484 ============= ============================================== 1485 drivers list of drivers and their usage 1486 ldiscs registered line disciplines 1487 driver/serial usage statistic and status of single tty lines 1488 ============= ============================================== 1489 1490To see which tty's are currently in use, you can simply look into the file 1491/proc/tty/drivers:: 1492 1493 > cat /proc/tty/drivers 1494 pty_slave /dev/pts 136 0-255 pty:slave 1495 pty_master /dev/ptm 128 0-255 pty:master 1496 pty_slave /dev/ttyp 3 0-255 pty:slave 1497 pty_master /dev/pty 2 0-255 pty:master 1498 serial /dev/cua 5 64-67 serial:callout 1499 serial /dev/ttyS 4 64-67 serial 1500 /dev/tty0 /dev/tty0 4 0 system:vtmaster 1501 /dev/ptmx /dev/ptmx 5 2 system 1502 /dev/console /dev/console 5 1 system:console 1503 /dev/tty /dev/tty 5 0 system:/dev/tty 1504 unknown /dev/tty 4 1-63 console 1505 1506 15071.7 Miscellaneous kernel statistics in /proc/stat 1508------------------------------------------------- 1509 1510Various pieces of information about kernel activity are available in the 1511/proc/stat file. All of the numbers reported in this file are aggregates 1512since the system first booted. For a quick look, simply cat the file:: 1513 1514 > cat /proc/stat 1515 cpu 237902850 368826709 106375398 1873517540 1135548 0 14507935 0 0 0 1516 cpu0 60045249 91891769 26331539 468411416 495718 0 5739640 0 0 0 1517 cpu1 59746288 91759249 26609887 468860630 312281 0 4384817 0 0 0 1518 cpu2 59489247 92985423 26904446 467808813 171668 0 2268998 0 0 0 1519 cpu3 58622065 92190267 26529524 468436680 155879 0 2114478 0 0 0 1520 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> 1521 ctxt 22848221062 1522 btime 1605316999 1523 processes 746787147 1524 procs_running 2 1525 procs_blocked 0 1526 softirq 12121874454 100099120 3938138295 127375644 2795979 187870761 0 173808342 3072582055 52608 224184354 1527 1528The very first "cpu" line aggregates the numbers in all of the other "cpuN" 1529lines. These numbers identify the amount of time the CPU has spent performing 1530different kinds of work. Time units are in USER_HZ (typically hundredths of a 1531second). The meanings of the columns are as follows, from left to right: 1532 1533- user: normal processes executing in user mode 1534- nice: niced processes executing in user mode 1535- system: processes executing in kernel mode 1536- idle: twiddling thumbs 1537- iowait: In a word, iowait stands for waiting for I/O to complete. But there 1538 are several problems: 1539 1540 1. CPU will not wait for I/O to complete, iowait is the time that a task is 1541 waiting for I/O to complete. When CPU goes into idle state for 1542 outstanding task I/O, another task will be scheduled on this CPU. 1543 2. In a multi-core CPU, the task waiting for I/O to complete is not running 1544 on any CPU, so the iowait of each CPU is difficult to calculate. 1545 3. The value of iowait field in /proc/stat will decrease in certain 1546 conditions. 1547 1548 So, the iowait is not reliable by reading from /proc/stat. 1549- irq: servicing interrupts 1550- softirq: servicing softirqs 1551- steal: involuntary wait 1552- guest: running a normal guest 1553- guest_nice: running a niced guest 1554 1555The "intr" line gives counts of interrupts serviced since boot time, for each 1556of the possible system interrupts. The first column is the total of all 1557interrupts serviced including unnumbered architecture specific interrupts; 1558each subsequent column is the total for that particular numbered interrupt. 1559Unnumbered interrupts are not shown, only summed into the total. 1560 1561The "ctxt" line gives the total number of context switches across all CPUs. 1562 1563The "btime" line gives the time at which the system booted, in seconds since 1564the Unix epoch. 1565 1566The "processes" line gives the number of processes and threads created, which 1567includes (but is not limited to) those created by calls to the fork() and 1568clone() system calls. 1569 1570The "procs_running" line gives the total number of threads that are 1571running or ready to run (i.e., the total number of runnable threads). 1572 1573The "procs_blocked" line gives the number of processes currently blocked, 1574waiting for I/O to complete. 1575 1576The "softirq" line gives counts of softirqs serviced since boot time, for each 1577of the possible system softirqs. The first column is the total of all 1578softirqs serviced; each subsequent column is the total for that particular 1579softirq. 1580 1581 15821.8 Ext4 file system parameters 1583------------------------------- 1584 1585Information about mounted ext4 file systems can be found in 1586/proc/fs/ext4. Each mounted filesystem will have a directory in 1587/proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or 1588/proc/fs/ext4/sda9 or /proc/fs/ext4/dm-0). The files in each per-device 1589directory are shown in Table 1-12, below. 1590 1591.. table:: Table 1-12: Files in /proc/fs/ext4/<devname> 1592 1593 ============== ========================================================== 1594 File Content 1595 mb_groups details of multiblock allocator buddy cache of free blocks 1596 ============== ========================================================== 1597 15981.9 /proc/consoles 1599------------------- 1600Shows registered system console lines. 1601 1602To see which character device lines are currently used for the system console 1603/dev/console, you may simply look into the file /proc/consoles:: 1604 1605 > cat /proc/consoles 1606 tty0 -WU (ECp) 4:7 1607 ttyS0 -W- (Ep) 4:64 1608 1609The columns are: 1610 1611+--------------------+-------------------------------------------------------+ 1612| device | name of the device | 1613+====================+=======================================================+ 1614| operations | * R = can do read operations | 1615| | * W = can do write operations | 1616| | * U = can do unblank | 1617+--------------------+-------------------------------------------------------+ 1618| flags | * E = it is enabled | 1619| | * C = it is preferred console | 1620| | * B = it is primary boot console | 1621| | * p = it is used for printk buffer | 1622| | * b = it is not a TTY but a Braille device | 1623| | * a = it is safe to use when cpu is offline | 1624+--------------------+-------------------------------------------------------+ 1625| major:minor | major and minor number of the device separated by a | 1626| | colon | 1627+--------------------+-------------------------------------------------------+ 1628 1629Summary 1630------- 1631 1632The /proc file system serves information about the running system. It not only 1633allows access to process data but also allows you to request the kernel status 1634by reading files in the hierarchy. 1635 1636The directory structure of /proc reflects the types of information and makes 1637it easy, if not obvious, where to look for specific data. 1638 1639Chapter 2: Modifying System Parameters 1640====================================== 1641 1642In This Chapter 1643--------------- 1644 1645* Modifying kernel parameters by writing into files found in /proc/sys 1646* Exploring the files which modify certain parameters 1647* Review of the /proc/sys file tree 1648 1649------------------------------------------------------------------------------ 1650 1651A very interesting part of /proc is the directory /proc/sys. This is not only 1652a source of information, it also allows you to change parameters within the 1653kernel. Be very careful when attempting this. You can optimize your system, 1654but you can also cause it to crash. Never alter kernel parameters on a 1655production system. Set up a development machine and test to make sure that 1656everything works the way you want it to. You may have no alternative but to 1657reboot the machine once an error has been made. 1658 1659To change a value, simply echo the new value into the file. 1660You need to be root to do this. You can create your own boot script 1661to perform this every time your system boots. 1662 1663The files in /proc/sys can be used to fine tune and monitor miscellaneous and 1664general things in the operation of the Linux kernel. Since some of the files 1665can inadvertently disrupt your system, it is advisable to read both 1666documentation and source before actually making adjustments. In any case, be 1667very careful when writing to any of these files. The entries in /proc may 1668change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt 1669review the kernel documentation in the directory linux/Documentation. 1670This chapter is heavily based on the documentation included in the pre 2.2 1671kernels, and became part of it in version 2.2.1 of the Linux kernel. 1672 1673Please see: Documentation/admin-guide/sysctl/ directory for descriptions of 1674these entries. 1675 1676Summary 1677------- 1678 1679Certain aspects of kernel behavior can be modified at runtime, without the 1680need to recompile the kernel, or even to reboot the system. The files in the 1681/proc/sys tree can not only be read, but also modified. You can use the echo 1682command to write value into these files, thereby changing the default settings 1683of the kernel. 1684 1685 1686Chapter 3: Per-process Parameters 1687================================= 1688 16893.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj- Adjust the oom-killer score 1690-------------------------------------------------------------------------------- 1691 1692These files can be used to adjust the badness heuristic used to select which 1693process gets killed in out of memory (oom) conditions. 1694 1695The badness heuristic assigns a value to each candidate task ranging from 0 1696(never kill) to 1000 (always kill) to determine which process is targeted. The 1697units are roughly a proportion along that range of allowed memory the process 1698may allocate from based on an estimation of its current memory and swap use. 1699For example, if a task is using all allowed memory, its badness score will be 17001000. If it is using half of its allowed memory, its score will be 500. 1701 1702The amount of "allowed" memory depends on the context in which the oom killer 1703was called. If it is due to the memory assigned to the allocating task's cpuset 1704being exhausted, the allowed memory represents the set of mems assigned to that 1705cpuset. If it is due to a mempolicy's node(s) being exhausted, the allowed 1706memory represents the set of mempolicy nodes. If it is due to a memory 1707limit (or swap limit) being reached, the allowed memory is that configured 1708limit. Finally, if it is due to the entire system being out of memory, the 1709allowed memory represents all allocatable resources. 1710 1711The value of /proc/<pid>/oom_score_adj is added to the badness score before it 1712is used to determine which task to kill. Acceptable values range from -1000 1713(OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_MAX). This allows userspace to 1714polarize the preference for oom killing either by always preferring a certain 1715task or completely disabling it. The lowest possible value, -1000, is 1716equivalent to disabling oom killing entirely for that task since it will always 1717report a badness score of 0. 1718 1719Consequently, it is very simple for userspace to define the amount of memory to 1720consider for each task. Setting a /proc/<pid>/oom_score_adj value of +500, for 1721example, is roughly equivalent to allowing the remainder of tasks sharing the 1722same system, cpuset, mempolicy, or memory controller resources to use at least 172350% more memory. A value of -500, on the other hand, would be roughly 1724equivalent to discounting 50% of the task's allowed memory from being considered 1725as scoring against the task. 1726 1727For backwards compatibility with previous kernels, /proc/<pid>/oom_adj may also 1728be used to tune the badness score. Its acceptable values range from -16 1729(OOM_ADJUST_MIN) to +15 (OOM_ADJUST_MAX) and a special value of -17 1730(OOM_DISABLE) to disable oom killing entirely for that task. Its value is 1731scaled linearly with /proc/<pid>/oom_score_adj. 1732 1733The value of /proc/<pid>/oom_score_adj may be reduced no lower than the last 1734value set by a CAP_SYS_RESOURCE process. To reduce the value any lower 1735requires CAP_SYS_RESOURCE. 1736 1737 17383.2 /proc/<pid>/oom_score - Display current oom-killer score 1739------------------------------------------------------------- 1740 1741This file can be used to check the current score used by the oom-killer for 1742any given <pid>. Use it together with /proc/<pid>/oom_score_adj to tune which 1743process should be killed in an out-of-memory situation. 1744 1745Please note that the exported value includes oom_score_adj so it is 1746effectively in range [0,2000]. 1747 1748 17493.3 /proc/<pid>/io - Display the IO accounting fields 1750------------------------------------------------------- 1751 1752This file contains IO statistics for each running process. 1753 1754Example 1755~~~~~~~ 1756 1757:: 1758 1759 test:/tmp # dd if=/dev/zero of=/tmp/test.dat & 1760 [1] 3828 1761 1762 test:/tmp # cat /proc/3828/io 1763 rchar: 323934931 1764 wchar: 323929600 1765 syscr: 632687 1766 syscw: 632675 1767 read_bytes: 0 1768 write_bytes: 323932160 1769 cancelled_write_bytes: 0 1770 1771 1772Description 1773~~~~~~~~~~~ 1774 1775rchar 1776^^^^^ 1777 1778I/O counter: chars read 1779The number of bytes which this task has caused to be read from storage. This 1780is simply the sum of bytes which this process passed to read() and pread(). 1781It includes things like tty IO and it is unaffected by whether or not actual 1782physical disk IO was required (the read might have been satisfied from 1783pagecache). 1784 1785 1786wchar 1787^^^^^ 1788 1789I/O counter: chars written 1790The number of bytes which this task has caused, or shall cause to be written 1791to disk. Similar caveats apply here as with rchar. 1792 1793 1794syscr 1795^^^^^ 1796 1797I/O counter: read syscalls 1798Attempt to count the number of read I/O operations, i.e. syscalls like read() 1799and pread(). 1800 1801 1802syscw 1803^^^^^ 1804 1805I/O counter: write syscalls 1806Attempt to count the number of write I/O operations, i.e. syscalls like 1807write() and pwrite(). 1808 1809 1810read_bytes 1811^^^^^^^^^^ 1812 1813I/O counter: bytes read 1814Attempt to count the number of bytes which this process really did cause to 1815be fetched from the storage layer. Done at the submit_bio() level, so it is 1816accurate for block-backed filesystems. <please add status regarding NFS and 1817CIFS at a later time> 1818 1819 1820write_bytes 1821^^^^^^^^^^^ 1822 1823I/O counter: bytes written 1824Attempt to count the number of bytes which this process caused to be sent to 1825the storage layer. This is done at page-dirtying time. 1826 1827 1828cancelled_write_bytes 1829^^^^^^^^^^^^^^^^^^^^^ 1830 1831The big inaccuracy here is truncate. If a process writes 1MB to a file and 1832then deletes the file, it will in fact perform no writeout. But it will have 1833been accounted as having caused 1MB of write. 1834In other words: The number of bytes which this process caused to not happen, 1835by truncating pagecache. A task can cause "negative" IO too. If this task 1836truncates some dirty pagecache, some IO which another task has been accounted 1837for (in its write_bytes) will not be happening. We _could_ just subtract that 1838from the truncating task's write_bytes, but there is information loss in doing 1839that. 1840 1841 1842.. Note:: 1843 1844 At its current implementation state, this is a bit racy on 32-bit machines: 1845 if process A reads process B's /proc/pid/io while process B is updating one 1846 of those 64-bit counters, process A could see an intermediate result. 1847 1848 1849More information about this can be found within the taskstats documentation in 1850Documentation/accounting. 1851 18523.4 /proc/<pid>/coredump_filter - Core dump filtering settings 1853--------------------------------------------------------------- 1854When a process is dumped, all anonymous memory is written to a core file as 1855long as the size of the core file isn't limited. But sometimes we don't want 1856to dump some memory segments, for example, huge shared memory or DAX. 1857Conversely, sometimes we want to save file-backed memory segments into a core 1858file, not only the individual files. 1859 1860/proc/<pid>/coredump_filter allows you to customize which memory segments 1861will be dumped when the <pid> process is dumped. coredump_filter is a bitmask 1862of memory types. If a bit of the bitmask is set, memory segments of the 1863corresponding memory type are dumped, otherwise they are not dumped. 1864 1865The following 9 memory types are supported: 1866 1867 - (bit 0) anonymous private memory 1868 - (bit 1) anonymous shared memory 1869 - (bit 2) file-backed private memory 1870 - (bit 3) file-backed shared memory 1871 - (bit 4) ELF header pages in file-backed private memory areas (it is 1872 effective only if the bit 2 is cleared) 1873 - (bit 5) hugetlb private memory 1874 - (bit 6) hugetlb shared memory 1875 - (bit 7) DAX private memory 1876 - (bit 8) DAX shared memory 1877 1878 Note that MMIO pages such as frame buffer are never dumped and vDSO pages 1879 are always dumped regardless of the bitmask status. 1880 1881 Note that bits 0-4 don't affect hugetlb or DAX memory. hugetlb memory is 1882 only affected by bit 5-6, and DAX is only affected by bits 7-8. 1883 1884The default value of coredump_filter is 0x33; this means all anonymous memory 1885segments, ELF header pages and hugetlb private memory are dumped. 1886 1887If you don't want to dump all shared memory segments attached to pid 1234, 1888write 0x31 to the process's proc file:: 1889 1890 $ echo 0x31 > /proc/1234/coredump_filter 1891 1892When a new process is created, the process inherits the bitmask status from its 1893parent. It is useful to set up coredump_filter before the program runs. 1894For example:: 1895 1896 $ echo 0x7 > /proc/self/coredump_filter 1897 $ ./some_program 1898 18993.5 /proc/<pid>/mountinfo - Information about mounts 1900-------------------------------------------------------- 1901 1902This file contains lines of the form:: 1903 1904 36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue 1905 (1)(2)(3) (4) (5) (6) (n…m) (m+1)(m+2) (m+3) (m+4) 1906 1907 (1) mount ID: unique identifier of the mount (may be reused after umount) 1908 (2) parent ID: ID of parent (or of self for the top of the mount tree) 1909 (3) major:minor: value of st_dev for files on filesystem 1910 (4) root: root of the mount within the filesystem 1911 (5) mount point: mount point relative to the process's root 1912 (6) mount options: per mount options 1913 (n…m) optional fields: zero or more fields of the form "tag[:value]" 1914 (m+1) separator: marks the end of the optional fields 1915 (m+2) filesystem type: name of filesystem of the form "type[.subtype]" 1916 (m+3) mount source: filesystem specific information or "none" 1917 (m+4) super options: per super block options 1918 1919Parsers should ignore all unrecognised optional fields. Currently the 1920possible optional fields are: 1921 1922================ ============================================================== 1923shared:X mount is shared in peer group X 1924master:X mount is slave to peer group X 1925propagate_from:X mount is slave and receives propagation from peer group X [#]_ 1926unbindable mount is unbindable 1927================ ============================================================== 1928 1929.. [#] X is the closest dominant peer group under the process's root. If 1930 X is the immediate master of the mount, or if there's no dominant peer 1931 group under the same root, then only the "master:X" field is present 1932 and not the "propagate_from:X" field. 1933 1934For more information on mount propagation see: 1935 1936 Documentation/filesystems/sharedsubtree.rst 1937 1938 19393.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm 1940-------------------------------------------------------- 1941These files provide a method to access a task's comm value. It also allows for 1942a task to set its own or one of its thread siblings comm value. The comm value 1943is limited in size compared to the cmdline value, so writing anything longer 1944then the kernel's TASK_COMM_LEN (currently 16 chars, including the NUL 1945terminator) will result in a truncated comm value. 1946 1947 19483.7 /proc/<pid>/task/<tid>/children - Information about task children 1949------------------------------------------------------------------------- 1950This file provides a fast way to retrieve first level children pids 1951of a task pointed by <pid>/<tid> pair. The format is a space separated 1952stream of pids. 1953 1954Note the "first level" here -- if a child has its own children they will 1955not be listed here; one needs to read /proc/<children-pid>/task/<tid>/children 1956to obtain the descendants. 1957 1958Since this interface is intended to be fast and cheap it doesn't 1959guarantee to provide precise results and some children might be 1960skipped, especially if they've exited right after we printed their 1961pids, so one needs to either stop or freeze processes being inspected 1962if precise results are needed. 1963 1964 19653.8 /proc/<pid>/fdinfo/<fd> - Information about opened file 1966--------------------------------------------------------------- 1967This file provides information associated with an opened file. The regular 1968files have at least four fields -- 'pos', 'flags', 'mnt_id' and 'ino'. 1969The 'pos' represents the current offset of the opened file in decimal 1970form [see lseek(2) for details], 'flags' denotes the octal O_xxx mask the 1971file has been created with [see open(2) for details] and 'mnt_id' represents 1972mount ID of the file system containing the opened file [see 3.5 1973/proc/<pid>/mountinfo for details]. 'ino' represents the inode number of 1974the file. 1975 1976A typical output is:: 1977 1978 pos: 0 1979 flags: 0100002 1980 mnt_id: 19 1981 ino: 63107 1982 1983All locks associated with a file descriptor are shown in its fdinfo too:: 1984 1985 lock: 1: FLOCK ADVISORY WRITE 359 00:13:11691 0 EOF 1986 1987The files such as eventfd, fsnotify, signalfd, epoll among the regular pos/flags 1988pair provide additional information particular to the objects they represent. 1989 1990Eventfd files 1991~~~~~~~~~~~~~ 1992 1993:: 1994 1995 pos: 0 1996 flags: 04002 1997 mnt_id: 9 1998 ino: 63107 1999 eventfd-count: 5a 2000 2001where 'eventfd-count' is hex value of a counter. 2002 2003Signalfd files 2004~~~~~~~~~~~~~~ 2005 2006:: 2007 2008 pos: 0 2009 flags: 04002 2010 mnt_id: 9 2011 ino: 63107 2012 sigmask: 0000000000000200 2013 2014where 'sigmask' is hex value of the signal mask associated 2015with a file. 2016 2017Epoll files 2018~~~~~~~~~~~ 2019 2020:: 2021 2022 pos: 0 2023 flags: 02 2024 mnt_id: 9 2025 ino: 63107 2026 tfd: 5 events: 1d data: ffffffffffffffff pos:0 ino:61af sdev:7 2027 2028where 'tfd' is a target file descriptor number in decimal form, 2029'events' is events mask being watched and the 'data' is data 2030associated with a target [see epoll(7) for more details]. 2031 2032The 'pos' is current offset of the target file in decimal form 2033[see lseek(2)], 'ino' and 'sdev' are inode and device numbers 2034where target file resides, all in hex format. 2035 2036Fsnotify files 2037~~~~~~~~~~~~~~ 2038For inotify files the format is the following:: 2039 2040 pos: 0 2041 flags: 02000000 2042 mnt_id: 9 2043 ino: 63107 2044 inotify wd:3 ino:9e7e sdev:800013 mask:800afce ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:7e9e0000640d1b6d 2045 2046where 'wd' is a watch descriptor in decimal form, i.e. a target file 2047descriptor number, 'ino' and 'sdev' are inode and device where the 2048target file resides and the 'mask' is the mask of events, all in hex 2049form [see inotify(7) for more details]. 2050 2051If the kernel was built with exportfs support, the path to the target 2052file is encoded as a file handle. The file handle is provided by three 2053fields 'fhandle-bytes', 'fhandle-type' and 'f_handle', all in hex 2054format. 2055 2056If the kernel is built without exportfs support the file handle won't be 2057printed out. 2058 2059If there is no inotify mark attached yet the 'inotify' line will be omitted. 2060 2061For fanotify files the format is:: 2062 2063 pos: 0 2064 flags: 02 2065 mnt_id: 9 2066 ino: 63107 2067 fanotify flags:10 event-flags:0 2068 fanotify mnt_id:12 mflags:40 mask:38 ignored_mask:40000003 2069 fanotify ino:4f969 sdev:800013 mflags:0 mask:3b ignored_mask:40000000 fhandle-bytes:8 fhandle-type:1 f_handle:69f90400c275b5b4 2070 2071where fanotify 'flags' and 'event-flags' are values used in fanotify_init 2072call, 'mnt_id' is the mount point identifier, 'mflags' is the value of 2073flags associated with mark which are tracked separately from events 2074mask. 'ino' and 'sdev' are target inode and device, 'mask' is the events 2075mask and 'ignored_mask' is the mask of events which are to be ignored. 2076All are in hex format. Incorporation of 'mflags', 'mask' and 'ignored_mask' 2077provide information about flags and mask used in fanotify_mark 2078call [see fsnotify manpage for details]. 2079 2080While the first three lines are mandatory and always printed, the rest is 2081optional and may be omitted if no marks created yet. 2082 2083Timerfd files 2084~~~~~~~~~~~~~ 2085 2086:: 2087 2088 pos: 0 2089 flags: 02 2090 mnt_id: 9 2091 ino: 63107 2092 clockid: 0 2093 ticks: 0 2094 settime flags: 01 2095 it_value: (0, 49406829) 2096 it_interval: (1, 0) 2097 2098where 'clockid' is the clock type and 'ticks' is the number of the timer expirations 2099that have occurred [see timerfd_create(2) for details]. 'settime flags' are 2100flags in octal form been used to setup the timer [see timerfd_settime(2) for 2101details]. 'it_value' is remaining time until the timer expiration. 2102'it_interval' is the interval for the timer. Note the timer might be set up 2103with TIMER_ABSTIME option which will be shown in 'settime flags', but 'it_value' 2104still exhibits timer's remaining time. 2105 2106DMA Buffer files 2107~~~~~~~~~~~~~~~~ 2108 2109:: 2110 2111 pos: 0 2112 flags: 04002 2113 mnt_id: 9 2114 ino: 63107 2115 size: 32768 2116 count: 2 2117 exp_name: system-heap 2118 2119where 'size' is the size of the DMA buffer in bytes. 'count' is the file count of 2120the DMA buffer file. 'exp_name' is the name of the DMA buffer exporter. 2121 21223.9 /proc/<pid>/map_files - Information about memory mapped files 2123--------------------------------------------------------------------- 2124This directory contains symbolic links which represent memory mapped files 2125the process is maintaining. Example output:: 2126 2127 | lr-------- 1 root root 64 Jan 27 11:24 333c600000-333c620000 -> /usr/lib64/ld-2.18.so 2128 | lr-------- 1 root root 64 Jan 27 11:24 333c81f000-333c820000 -> /usr/lib64/ld-2.18.so 2129 | lr-------- 1 root root 64 Jan 27 11:24 333c820000-333c821000 -> /usr/lib64/ld-2.18.so 2130 | ... 2131 | lr-------- 1 root root 64 Jan 27 11:24 35d0421000-35d0422000 -> /usr/lib64/libselinux.so.1 2132 | lr-------- 1 root root 64 Jan 27 11:24 400000-41a000 -> /usr/bin/ls 2133 2134The name of a link represents the virtual memory bounds of a mapping, i.e. 2135vm_area_struct::vm_start-vm_area_struct::vm_end. 2136 2137The main purpose of the map_files is to retrieve a set of memory mapped 2138files in a fast way instead of parsing /proc/<pid>/maps or 2139/proc/<pid>/smaps, both of which contain many more records. At the same 2140time one can open(2) mappings from the listings of two processes and 2141comparing their inode numbers to figure out which anonymous memory areas 2142are actually shared. 2143 21443.10 /proc/<pid>/timerslack_ns - Task timerslack value 2145--------------------------------------------------------- 2146This file provides the value of the task's timerslack value in nanoseconds. 2147This value specifies an amount of time that normal timers may be deferred 2148in order to coalesce timers and avoid unnecessary wakeups. 2149 2150This allows a task's interactivity vs power consumption tradeoff to be 2151adjusted. 2152 2153Writing 0 to the file will set the task's timerslack to the default value. 2154 2155Valid values are from 0 - ULLONG_MAX 2156 2157An application setting the value must have PTRACE_MODE_ATTACH_FSCREDS level 2158permissions on the task specified to change its timerslack_ns value. 2159 21603.11 /proc/<pid>/patch_state - Livepatch patch operation state 2161----------------------------------------------------------------- 2162When CONFIG_LIVEPATCH is enabled, this file displays the value of the 2163patch state for the task. 2164 2165A value of '-1' indicates that no patch is in transition. 2166 2167A value of '0' indicates that a patch is in transition and the task is 2168unpatched. If the patch is being enabled, then the task hasn't been 2169patched yet. If the patch is being disabled, then the task has already 2170been unpatched. 2171 2172A value of '1' indicates that a patch is in transition and the task is 2173patched. If the patch is being enabled, then the task has already been 2174patched. If the patch is being disabled, then the task hasn't been 2175unpatched yet. 2176 21773.12 /proc/<pid>/arch_status - task architecture specific status 2178------------------------------------------------------------------- 2179When CONFIG_PROC_PID_ARCH_STATUS is enabled, this file displays the 2180architecture specific status of the task. 2181 2182Example 2183~~~~~~~ 2184 2185:: 2186 2187 $ cat /proc/6753/arch_status 2188 AVX512_elapsed_ms: 8 2189 2190Description 2191~~~~~~~~~~~ 2192 2193x86 specific entries 2194~~~~~~~~~~~~~~~~~~~~~ 2195 2196AVX512_elapsed_ms 2197^^^^^^^^^^^^^^^^^^ 2198 2199 If AVX512 is supported on the machine, this entry shows the milliseconds 2200 elapsed since the last time AVX512 usage was recorded. The recording 2201 happens on a best effort basis when a task is scheduled out. This means 2202 that the value depends on two factors: 2203 2204 1) The time which the task spent on the CPU without being scheduled 2205 out. With CPU isolation and a single runnable task this can take 2206 several seconds. 2207 2208 2) The time since the task was scheduled out last. Depending on the 2209 reason for being scheduled out (time slice exhausted, syscall ...) 2210 this can be arbitrary long time. 2211 2212 As a consequence the value cannot be considered precise and authoritative 2213 information. The application which uses this information has to be aware 2214 of the overall scenario on the system in order to determine whether a 2215 task is a real AVX512 user or not. Precise information can be obtained 2216 with performance counters. 2217 2218 A special value of '-1' indicates that no AVX512 usage was recorded, thus 2219 the task is unlikely an AVX512 user, but depends on the workload and the 2220 scheduling scenario, it also could be a false negative mentioned above. 2221 22223.13 /proc/<pid>/fd - List of symlinks to open files 2223------------------------------------------------------- 2224This directory contains symbolic links which represent open files 2225the process is maintaining. Example output:: 2226 2227 lr-x------ 1 root root 64 Sep 20 17:53 0 -> /dev/null 2228 l-wx------ 1 root root 64 Sep 20 17:53 1 -> /dev/null 2229 lrwx------ 1 root root 64 Sep 20 17:53 10 -> 'socket:[12539]' 2230 lrwx------ 1 root root 64 Sep 20 17:53 11 -> 'socket:[12540]' 2231 lrwx------ 1 root root 64 Sep 20 17:53 12 -> 'socket:[12542]' 2232 2233The number of open files for the process is stored in 'size' member 2234of stat() output for /proc/<pid>/fd for fast access. 2235------------------------------------------------------- 2236 22373.14 /proc/<pid/ksm_stat - Information about the process's ksm status 2238--------------------------------------------------------------------- 2239When CONFIG_KSM is enabled, each process has this file which displays 2240the information of ksm merging status. 2241 2242Example 2243~~~~~~~ 2244 2245:: 2246 2247 / # cat /proc/self/ksm_stat 2248 ksm_rmap_items 0 2249 ksm_zero_pages 0 2250 ksm_merging_pages 0 2251 ksm_process_profit 0 2252 ksm_merge_any: no 2253 ksm_mergeable: no 2254 2255Description 2256~~~~~~~~~~~ 2257 2258ksm_rmap_items 2259^^^^^^^^^^^^^^ 2260 2261The number of ksm_rmap_item structures in use. The structure 2262ksm_rmap_item stores the reverse mapping information for virtual 2263addresses. KSM will generate a ksm_rmap_item for each ksm-scanned page of 2264the process. 2265 2266ksm_zero_pages 2267^^^^^^^^^^^^^^ 2268 2269When /sys/kernel/mm/ksm/use_zero_pages is enabled, it represent how many 2270empty pages are merged with kernel zero pages by KSM. 2271 2272ksm_merging_pages 2273^^^^^^^^^^^^^^^^^ 2274 2275It represents how many pages of this process are involved in KSM merging 2276(not including ksm_zero_pages). It is the same with what 2277/proc/<pid>/ksm_merging_pages shows. 2278 2279ksm_process_profit 2280^^^^^^^^^^^^^^^^^^ 2281 2282The profit that KSM brings (Saved bytes). KSM can save memory by merging 2283identical pages, but also can consume additional memory, because it needs 2284to generate a number of rmap_items to save each scanned page's brief rmap 2285information. Some of these pages may be merged, but some may not be abled 2286to be merged after being checked several times, which are unprofitable 2287memory consumed. 2288 2289ksm_merge_any 2290^^^^^^^^^^^^^ 2291 2292It specifies whether the process's 'mm is added by prctl() into the 2293candidate list of KSM or not, and if KSM scanning is fully enabled at 2294process level. 2295 2296ksm_mergeable 2297^^^^^^^^^^^^^ 2298 2299It specifies whether any VMAs of the process''s mms are currently 2300applicable to KSM. 2301 2302More information about KSM can be found in 2303Documentation/admin-guide/mm/ksm.rst. 2304 2305 2306Chapter 4: Configuring procfs 2307============================= 2308 23094.1 Mount options 2310--------------------- 2311 2312The following mount options are supported: 2313 2314 ========= ======================================================== 2315 hidepid= Set /proc/<pid>/ access mode. 2316 gid= Set the group authorized to learn processes information. 2317 subset= Show only the specified subset of procfs. 2318 ========= ======================================================== 2319 2320hidepid=off or hidepid=0 means classic mode - everybody may access all 2321/proc/<pid>/ directories (default). 2322 2323hidepid=noaccess or hidepid=1 means users may not access any /proc/<pid>/ 2324directories but their own. Sensitive files like cmdline, sched*, status are now 2325protected against other users. This makes it impossible to learn whether any 2326user runs specific program (given the program doesn't reveal itself by its 2327behaviour). As an additional bonus, as /proc/<pid>/cmdline is unaccessible for 2328other users, poorly written programs passing sensitive information via program 2329arguments are now protected against local eavesdroppers. 2330 2331hidepid=invisible or hidepid=2 means hidepid=1 plus all /proc/<pid>/ will be 2332fully invisible to other users. It doesn't mean that it hides a fact whether a 2333process with a specific pid value exists (it can be learned by other means, e.g. 2334by "kill -0 $PID"), but it hides process's uid and gid, which may be learned by 2335stat()'ing /proc/<pid>/ otherwise. It greatly complicates an intruder's task of 2336gathering information about running processes, whether some daemon runs with 2337elevated privileges, whether other user runs some sensitive program, whether 2338other users run any program at all, etc. 2339 2340hidepid=ptraceable or hidepid=4 means that procfs should only contain 2341/proc/<pid>/ directories that the caller can ptrace. 2342 2343gid= defines a group authorized to learn processes information otherwise 2344prohibited by hidepid=. If you use some daemon like identd which needs to learn 2345information about processes information, just add identd to this group. 2346 2347subset=pid hides all top level files and directories in the procfs that 2348are not related to tasks. 2349 2350Chapter 5: Filesystem behavior 2351============================== 2352 2353Originally, before the advent of pid namespace, procfs was a global file 2354system. It means that there was only one procfs instance in the system. 2355 2356When pid namespace was added, a separate procfs instance was mounted in 2357each pid namespace. So, procfs mount options are global among all 2358mountpoints within the same namespace:: 2359 2360 # grep ^proc /proc/mounts 2361 proc /proc proc rw,relatime,hidepid=2 0 0 2362 2363 # strace -e mount mount -o hidepid=1 -t proc proc /tmp/proc 2364 mount("proc", "/tmp/proc", "proc", 0, "hidepid=1") = 0 2365 +++ exited with 0 +++ 2366 2367 # grep ^proc /proc/mounts 2368 proc /proc proc rw,relatime,hidepid=2 0 0 2369 proc /tmp/proc proc rw,relatime,hidepid=2 0 0 2370 2371and only after remounting procfs mount options will change at all 2372mountpoints:: 2373 2374 # mount -o remount,hidepid=1 -t proc proc /tmp/proc 2375 2376 # grep ^proc /proc/mounts 2377 proc /proc proc rw,relatime,hidepid=1 0 0 2378 proc /tmp/proc proc rw,relatime,hidepid=1 0 0 2379 2380This behavior is different from the behavior of other filesystems. 2381 2382The new procfs behavior is more like other filesystems. Each procfs mount 2383creates a new procfs instance. Mount options affect own procfs instance. 2384It means that it became possible to have several procfs instances 2385displaying tasks with different filtering options in one pid namespace:: 2386 2387 # mount -o hidepid=invisible -t proc proc /proc 2388 # mount -o hidepid=noaccess -t proc proc /tmp/proc 2389 # grep ^proc /proc/mounts 2390 proc /proc proc rw,relatime,hidepid=invisible 0 0 2391 proc /tmp/proc proc rw,relatime,hidepid=noaccess 0 0 2392