1 .. SPDX-License-Identifier: GPL-2.0
2
3 ====================
4 The /proc Filesystem
5 ====================
6
7 ===================== =======================
8 /proc/sys Terrehon Bowden <terreho
9 Bodo Bauer <bb@ricochet.
10 2.4.x update Jorge Nerin <comandante@
11 move /proc/sys Shen Feng <shen@cn.fujit
12 fixes/update part 1.1 Stefani Seibold <stefani
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 /pr
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_
38
39 3.2 /proc/<pid>/oom_score - Display curren
40 3.3 /proc/<pid>/io - Display the IO accoun
41 3.4 /proc/<pid>/coredump_filter - Core dum
42 3.5 /proc/<pid>/mountinfo - Information ab
43 3.6 /proc/<pid>/comm & /proc/<pid>/task/<
44 3.7 /proc/<pid>/task/<tid>/children - Info
45 3.8 /proc/<pid>/fdinfo/<fd> - Information
46 3.9 /proc/<pid>/map_files - Information ab
47 3.10 /proc/<pid>/timerslack_ns - Task timer
48 3.11 /proc/<pid>/patch_state - Livepatch pa
49 3.12 /proc/<pid>/arch_status - Task archite
50 3.13 /proc/<pid>/fd - List of symlinks to o
51
52 4 Configuring procfs
53 4.1 Mount options
54
55 5 Filesystem behavior
56
57 Preface
58 =======
59
60 0.1 Introduction/Credits
61 ------------------------
62
63 This documentation is part of a soon (or so
64 the SuSE Linux distribution. As there is no
65 /proc file system and we've used many freely
66 chapters, it seems only fair to give the wor
67 This work is based on the 2.2.* kernel versi
68 afraid it's still far from complete, but we h
69 we know, it is the first 'all-in-one' document
70 is focused on the Intel x86 hardware, so if
71 SPARC, AXP, etc., features, you probably won'
72 It also only covers IPv4 networking, not IPv6
73 additions and patches are welcome and will b
74 mail them to Bodo.
75
76 We'd like to thank Alan Cox, Rik van Riel, a
77 other people for help compiling this documenta
78 special thank you to Andi Kleen for documenta
79 to create this document, as well as the add
80 Thanks to everybody else who contributed so
81 and helped create a great piece of software...
82
83 If you have any comments, corrections or add
84 contact Bodo Bauer at bb@ricochet.net. We'
85 document.
86
87 The latest version of this document
88 https://www.kernel.org/doc/html/latest/filesys
89
90 If the above direction does not works for
91 mailing list at linux-kernel@vger.kernel.or
92 comandante@zaralinux.com.
93
94 0.2 Legal Stuff
95 ---------------
96
97 We don't guarantee the correctness of this
98 complaining about how you screwed up your
99 documentation, we won't feel responsible...
100
101 Chapter 1: Collecting System Information
102 ========================================
103
104 In This Chapter
105 ---------------
106 * Investigating the properties of the pse
107 ability to provide information on the runnin
108 * Examining /proc's structure
109 * Uncovering various information about the
110 on the system
111
112 ----------------------------------------------
113
114 The proc file system acts as an interface to
115 kernel. It can be used to obtain informatio
116 certain kernel parameters at runtime (sysctl).
117
118 First, we'll take a look at the read-only
119 show you how you can use /proc/sys to change s
120
121 1.1 Process-Specific Subdirectories
122 -----------------------------------
123
124 The directory /proc contains (among other t
125 process running on the system, which is named
126
127 The link 'self' points to the process readi
128 subdirectory has the entries listed in Table 1
129
130 Note that an open file descriptor to /proc/<pi
131 contained files or subdirectories does not pre
132 for some other process in the event that <pid>
133 open /proc/<pid> file descriptors correspondin
134 never act on any new process that the kernel m
135 also assigned the process ID <pid>. Instead, o
136 usually fail with ESRCH.
137
138 .. table:: Table 1-1: Process specific entries
139
140 ============= ==============================
141 File Content
142 ============= ==============================
143 clear_refs Clears page referenced bits sh
144 cmdline Command line arguments
145 cpu Current and last cpu in which
146 cwd Link to the current working di
147 environ Values of environment variable
148 exe Link to the executable of this
149 fd Directory, which contains all
150 maps Memory maps to executables and
151 mem Memory held by this process
152 root Link to the root directory of
153 stat Process status
154 statm Process memory status informat
155 status Process status in human readab
156 wchan Present with CONFIG_KALLSYMS=y
157 symbol the task is blocked in
158 pagemap Page table
159 stack Report full stack trace, enabl
160 smaps An extension based on maps, sh
161 each mapping and flags associa
162 smaps_rollup Accumulated smaps stats for al
163 can be derived from smaps, but
164 numa_maps An extension based on maps, sh
165 binding policy as well as mem
166 ============= ==============================
167
168 For example, to get the status information of
169 read the file /proc/PID/status::
170
171 >cat /proc/self/status
172 Name: cat
173 State: R (running)
174 Tgid: 5452
175 Pid: 5452
176 PPid: 743
177 TracerPid: 0
178 Uid: 501 501 501 501
179 Gid: 100 100 100 100
180 FDSize: 256
181 Groups: 100 14 16
182 Kthread: 0
183 VmPeak: 5004 kB
184 VmSize: 5004 kB
185 VmLck: 0 kB
186 VmHWM: 476 kB
187 VmRSS: 476 kB
188 RssAnon: 352 kB
189 RssFile: 120 kB
190 RssShmem: 4 kB
191 VmData: 156 kB
192 VmStk: 88 kB
193 VmExe: 68 kB
194 VmLib: 1412 kB
195 VmPTE: 20 kb
196 VmSwap: 0 kB
197 HugetlbPages: 0 kB
198 CoreDumping: 0
199 THP_enabled: 1
200 Threads: 1
201 SigQ: 0/28578
202 SigPnd: 0000000000000000
203 ShdPnd: 0000000000000000
204 SigBlk: 0000000000000000
205 SigIgn: 0000000000000000
206 SigCgt: 0000000000000000
207 CapInh: 00000000fffffeff
208 CapPrm: 0000000000000000
209 CapEff: 0000000000000000
210 CapBnd: ffffffffffffffff
211 CapAmb: 0000000000000000
212 NoNewPrivs: 0
213 Seccomp: 0
214 Speculation_Store_Bypass: thread vulne
215 SpeculationIndirectBranch: conditional
216 voluntary_ctxt_switches: 0
217 nonvoluntary_ctxt_switches: 1
218
219 This shows you nearly the same information you
220 the ps command. In fact, ps uses the pr
221 information. But you get a more detailed vie
222 file /proc/PID/status. It fields are described
223
224 The statm file contains more detailed in
225 memory usage. Its seven fields are explained i
226 contains detailed information about the proces
227 explained in Table 1-4.
228
229 (for SMP CONFIG users)
230
231 For making accounting scalable, RSS related in
232 asynchronous manner and the value may not be v
233 snapshot of a moment, you can see /proc/<pid>/
234 It's slow but very precise.
235
236 .. table:: Table 1-2: Contents of the status f
237
238 ========================== =================
239 Field Content
240 ========================== =================
241 Name filename of the e
242 Umask file mode creatio
243 State state (R is runni
244 in an uninterrupt
245 T is traced or st
246 Tgid thread group ID
247 Ngid NUMA group ID (0
248 Pid process id
249 PPid process id of the
250 TracerPid PID of process tr
251 the tracer is out
252 Uid Real, effective,
253 Gid Real, effective,
254 FDSize number of file de
255 Groups supplementary gro
256 NStgid descendant namesp
257 NSpid descendant namesp
258 NSpgid descendant namesp
259 NSsid descendant namesp
260 Kthread kernel thread fla
261 VmPeak peak virtual memo
262 VmSize total program siz
263 VmLck locked memory siz
264 VmPin pinned memory siz
265 VmHWM peak resident set
266 VmRSS size of memory po
267 following parts
268 (VmRSS = RssAnon
269 RssAnon size of resident
270 RssFile size of resident
271 RssShmem size of resident
272 mapping of tmpfs
273 VmData size of private d
274 VmStk size of stack seg
275 VmExe size of text segm
276 VmLib size of shared li
277 VmPTE size of page tabl
278 VmSwap amount of swap us
279 (shmem swap usage
280 HugetlbPages size of hugetlb m
281 CoreDumping process's memory
282 (killing the proc
283 THP_enabled process is allowe
284 PR_SET_THP_DISABL
285 Threads number of threads
286 SigQ number of signals
287 SigPnd bitmap of pending
288 ShdPnd bitmap of shared
289 SigBlk bitmap of blocked
290 SigIgn bitmap of ignored
291 SigCgt bitmap of caught
292 CapInh bitmap of inherit
293 CapPrm bitmap of permitt
294 CapEff bitmap of effecti
295 CapBnd bitmap of capabil
296 CapAmb bitmap of ambient
297 NoNewPrivs no_new_privs, lik
298 Seccomp seccomp mode, lik
299 Speculation_Store_Bypass speculative store
300 SpeculationIndirectBranch indirect branch s
301 Cpus_allowed mask of CPUs on w
302 Cpus_allowed_list Same as previous,
303 Mems_allowed mask of memory no
304 Mems_allowed_list Same as previous,
305 voluntary_ctxt_switches number of volunta
306 nonvoluntary_ctxt_switches number of non vol
307 ========================== =================
308
309
310 .. table:: Table 1-3: Contents of the statm fi
311
312 ======== ===============================
313 Field Content
314 ======== ===============================
315 size total program size (pages)
316 resident size of memory portions (pages)
317 shared number of pages that are shared
318
319 trs number of pages that are 'code'
320
321 lrs number of pages of library
322 drs number of pages of data/stack
323
324 dt number of dirty pages
325 ======== ===============================
326
327
328 .. table:: Table 1-4: Contents of the stat fie
329
330 ============= ==============================
331 Field Content
332 ============= ==============================
333 pid process id
334 tcomm filename of the executable
335 state state (R is running, S is slee
336 uninterruptible wait, Z is zom
337 ppid process id of the parent proce
338 pgrp pgrp of the process
339 sid session id
340 tty_nr tty the process uses
341 tty_pgrp pgrp of the tty
342 flags task flags
343 min_flt number of minor faults
344 cmin_flt number of minor faults with ch
345 maj_flt number of major faults
346 cmaj_flt number of major faults with ch
347 utime user mode jiffies
348 stime kernel mode jiffies
349 cutime user mode jiffies with child's
350 cstime kernel mode jiffies with child
351 priority priority level
352 nice nice level
353 num_threads number of threads
354 it_real_value (obsolete, always 0)
355 start_time time the process started after
356 vsize virtual memory size
357 rss resident set memory size
358 rsslim current limit in bytes on the
359 start_code address above which program te
360 end_code address below which program te
361 start_stack address of the start of the ma
362 esp current value of ESP
363 eip current value of EIP
364 pending bitmap of pending signals
365 blocked bitmap of blocked signals
366 sigign bitmap of ignored signals
367 sigcatch bitmap of caught signals
368 0 (place holder, used to be the
369 use /proc/PID/wchan instead)
370 0 (place holder)
371 0 (place holder)
372 exit_signal signal to send to parent threa
373 task_cpu which CPU the task is schedule
374 rt_priority realtime priority
375 policy scheduling policy (man sched_s
376 blkio_ticks time spent waiting for block I
377 gtime guest time of the task in jiff
378 cgtime guest time of the task childre
379 start_data address above which program da
380 end_data address below which program da
381 start_brk address above which program he
382 arg_start address above which program co
383 arg_end address below which program co
384 env_start address above which program en
385 env_end address below which program en
386 exit_code the thread's exit_code in the
387 system call
388 ============= ==============================
389
390 The /proc/PID/maps file contains the currently
391 their access permissions.
392
393 The format is::
394
395 address perms offset dev inod
396
397 08048000-08049000 r-xp 00000000 03:00 8312
398 08049000-0804a000 rw-p 00001000 03:00 8312
399 0804a000-0806b000 rw-p 00000000 00:00 0
400 a7cb1000-a7cb2000 ---p 00000000 00:00 0
401 a7cb2000-a7eb2000 rw-p 00000000 00:00 0
402 a7eb2000-a7eb3000 ---p 00000000 00:00 0
403 a7eb3000-a7ed5000 rw-p 00000000 00:00 0
404 a7ed5000-a8008000 r-xp 00000000 03:00 4222
405 a8008000-a800a000 r--p 00133000 03:00 4222
406 a800a000-a800b000 rw-p 00135000 03:00 4222
407 a800b000-a800e000 rw-p 00000000 00:00 0
408 a800e000-a8022000 r-xp 00000000 03:00 1446
409 a8022000-a8023000 r--p 00013000 03:00 1446
410 a8023000-a8024000 rw-p 00014000 03:00 1446
411 a8024000-a8027000 rw-p 00000000 00:00 0
412 a8027000-a8043000 r-xp 00000000 03:00 8317
413 a8043000-a8044000 r--p 0001b000 03:00 8317
414 a8044000-a8045000 rw-p 0001c000 03:00 8317
415 aff35000-aff4a000 rw-p 00000000 00:00 0
416 ffffe000-fffff000 r-xp 00000000 00:00 0
417
418 where "address" is the address space in the pr
419 is a set of permissions::
420
421 r = read
422 w = write
423 x = execute
424 s = shared
425 p = private (copy on write)
426
427 "offset" is the offset into the mapping, "dev"
428 "inode" is the inode on that device. 0 indic
429 with the memory region, as the case would be w
430 The "pathname" shows the name associated file
431 is not associated with a file:
432
433 =================== ==================
434 [heap] the heap of the pr
435 [stack] the stack of the m
436 [vdso] the "virtual dynam
437 the kernel system
438 [anon:<name>] a private anonymou
439 named by userspace
440 [anon_shmem:<name>] an anonymous share
441 been named by user
442 =================== ==================
443
444 or if empty, the mapping is anonymous.
445
446 Starting with 6.11 kernel, /proc/PID/maps prov
447 ioctl()-based API that gives ability to flexib
448 filter individual VMAs. This interface is bina
449 efficient and easy programmatic use. `struct p
450 linux/fs.h UAPI header, serves as an input/out
451 `PROCMAP_QUERY` ioctl() command. See comments
452 details on query semantics, supported flags, d
453 usage information.
454
455 The /proc/PID/smaps is an extension based on m
456 consumption for each of the process's mappings
457 Memory Area, or VMA) there is a series of line
458
459 08048000-080bc000 r-xp 00000000 03:02 1313
460
461 Size: 1084 kB
462 KernelPageSize: 4 kB
463 MMUPageSize: 4 kB
464 Rss: 892 kB
465 Pss: 374 kB
466 Pss_Dirty: 0 kB
467 Shared_Clean: 892 kB
468 Shared_Dirty: 0 kB
469 Private_Clean: 0 kB
470 Private_Dirty: 0 kB
471 Referenced: 892 kB
472 Anonymous: 0 kB
473 KSM: 0 kB
474 LazyFree: 0 kB
475 AnonHugePages: 0 kB
476 ShmemPmdMapped: 0 kB
477 Shared_Hugetlb: 0 kB
478 Private_Hugetlb: 0 kB
479 Swap: 0 kB
480 SwapPss: 0 kB
481 KernelPageSize: 4 kB
482 MMUPageSize: 4 kB
483 Locked: 0 kB
484 THPeligible: 0
485 VmFlags: rd ex mr mw me dw
486
487 The first of these lines shows the same inform
488 mapping in /proc/PID/maps. Following lines sh
489 (size); the size of each page allocated when b
490 which is usually the same as the size in the p
491 used by the MMU when backing a VMA (in most ca
492 the amount of the mapping that is currently re
493 process' proportional share of this mapping (P
494 dirty shared and private pages in the mapping.
495
496 The "proportional set size" (PSS) of a process
497 in memory, where each page is divided by the n
498 So if a process has 1000 pages all to itself,
499 process, its PSS will be 1500. "Pss_Dirty" is
500 consists of dirty pages. ("Pss_Clean" is not
501 calculated by subtracting "Pss_Dirty" from "Ps
502
503 Note that even a page which is part of a MAP_S
504 a single pte mapped, i.e. is currently used b
505 as private and not as shared.
506
507 "Referenced" indicates the amount of memory cu
508 accessed.
509
510 "Anonymous" shows the amount of memory that do
511 a mapping associated with a file may contain a
512 and a page is modified, the file page is repla
513
514 "KSM" reports how many of the pages are KSM pa
515 are not included, only actual KSM pages.
516
517 "LazyFree" shows the amount of memory which is
518 The memory isn't freed immediately with madvis
519 pressure if the memory is clean. Please note t
520 be lower than the real value due to optimizati
521 implementation. If this is not desirable pleas
522
523 "AnonHugePages" shows the amount of memory bac
524
525 "ShmemPmdMapped" shows the amount of shared (s
526 huge pages.
527
528 "Shared_Hugetlb" and "Private_Hugetlb" show th
529 hugetlbfs page which is *not* counted in "RSS"
530 reasons. And these are not included in {Shared
531
532 "Swap" shows how much would-be-anonymous memor
533
534 For shmem mappings, "Swap" includes also the s
535 replaced by copy-on-write) part of the underly
536 "SwapPss" shows proportional swap share of thi
537 does not take into account swapped out page of
538 "Locked" indicates whether the mapping is lock
539
540 "THPeligible" indicates whether the mapping is
541 naturally aligned THP pages of any currently e
542 otherwise.
543
544 "VmFlags" field deserves a separate descriptio
545 kernel flags associated with the particular vi
546 encoded manner. The codes are the following:
547
548 == ====================================
549 rd readable
550 wr writeable
551 ex executable
552 sh shared
553 mr may read
554 mw may write
555 me may execute
556 ms may share
557 gd stack segment growns down
558 pf pure PFN range
559 dw disabled write to the mapped file
560 lo pages are locked in memory
561 io memory mapped I/O area
562 sr sequential read advise provided
563 rr random read advise provided
564 dc do not copy area on fork
565 de do not expand area on remapping
566 ac area is accountable
567 nr swap space is not reserved for the a
568 ht area uses huge tlb pages
569 sf synchronous page fault
570 ar architecture specific flag
571 wf wipe on fork
572 dd do not include area into core dump
573 sd soft dirty flag
574 mm mixed map area
575 hg huge page advise flag
576 nh no huge page advise flag
577 mg mergeable advise flag
578 bt arm64 BTI guarded page
579 mt arm64 MTE allocation tags are enable
580 um userfaultfd missing tracking
581 uw userfaultfd wr-protect tracking
582 ss shadow stack page
583 sl sealed
584 == ====================================
585
586 Note that there is no guarantee that every fla
587 be present in all further kernel releases. Thi
588 be vanished or the reverse -- new added. Inter
589 might change in future as well. So each consum
590 follow each specific kernel version for the ex
591
592 This file is only present if the CONFIG_MMU ke
593 enabled.
594
595 Note: reading /proc/PID/maps or /proc/PID/smap
596 output can be achieved only in the single read
597
598 This typically manifests when doing partial re
599 memory map is being modified. Despite the rac
600 guarantees:
601
602 1) The mapped addresses never go backwards, wh
603 regions will ever overlap.
604 2) If there is something at a given vaddr duri
605 life of the smaps/maps walk, there will be
606
607 The /proc/PID/smaps_rollup file includes the s
608 but their values are the sums of the correspon
609 the process. Additionally, it contains these
610
611 - Pss_Anon
612 - Pss_File
613 - Pss_Shmem
614
615 They represent the proportional shares of anon
616 described for smaps above. These fields are o
617 mapping identifies the type (anon, file, or sh
618 Thus all information in smaps_rollup can be de
619 significantly higher cost.
620
621 The /proc/PID/clear_refs is used to reset the
622 bits on both physical and virtual pages associ
623 soft-dirty bit on pte (see Documentation/admin
624 for details).
625 To clear the bits for all the pages associated
626
627 > echo 1 > /proc/PID/clear_refs
628
629 To clear the bits for the anonymous pages asso
630
631 > echo 2 > /proc/PID/clear_refs
632
633 To clear the bits for the file mapped pages as
634
635 > echo 3 > /proc/PID/clear_refs
636
637 To clear the soft-dirty bit::
638
639 > echo 4 > /proc/PID/clear_refs
640
641 To reset the peak resident set size ("high wat
642 current value::
643
644 > echo 5 > /proc/PID/clear_refs
645
646 Any other value written to /proc/PID/clear_ref
647
648 The /proc/pid/pagemap gives the PFN, which can
649 using /proc/kpageflags and number of times a p
650 /proc/kpagecount. For detailed explanation, se
651 Documentation/admin-guide/mm/pagemap.rst.
652
653 The /proc/pid/numa_maps is an extension based
654 locality and binding policy, as well as the me
655 each mapping. The output follows a general for
656 summarized separated by blank spaces, one mapp
657
658 address policy mapping details
659
660 00400000 default file=/usr/local/bin/app m
661 00600000 default file=/usr/local/bin/app a
662 3206000000 default file=/lib64/ld-2.12.so
663 320621f000 default file=/lib64/ld-2.12.so
664 3206220000 default file=/lib64/ld-2.12.so
665 3206221000 default anon=1 dirty=1 N3=1 ker
666 3206800000 default file=/lib64/libc-2.12.s
667 320698b000 default file=/lib64/libc-2.12.s
668 3206b8a000 default file=/lib64/libc-2.12.s
669 3206b8e000 default file=/lib64/libc-2.12.s
670 3206b8f000 default anon=3 dirty=3 active=1
671 7f4dc10a2000 default anon=3 dirty=3 N3=3 k
672 7f4dc10b4000 default anon=2 dirty=2 active
673 7f4dc1200000 default file=/anon_hugepage\0
674 7fff335f0000 default stack anon=3 dirty=3
675 7fff3369d000 default mapped=1 mapmax=35 ac
676
677 Where:
678
679 "address" is the starting address for the mapp
680
681 "policy" reports the NUMA memory policy set fo
682
683 "mapping details" summarizes mapping data such
684 node locality page counters (N0 == node0, N1 =
685 size, in KB, that is backing the mapping up.
686
687 1.2 Kernel data
688 ---------------
689
690 Similar to the process entries, the kernel d
691 the running kernel. The files used to obtain t
692 /proc and are listed in Table 1-5. Not all
693 system. It depends on the kernel configurati
694 files are there, and which are missing.
695
696 .. table:: Table 1-5: Kernel info in /proc
697
698 ============ ================================
699 File Content
700 ============ ================================
701 allocinfo Memory allocations profiling inf
702 apm Advanced power management info
703 bootconfig Kernel command line obtained fro
704 and, if there were kernel parame
705 boot loader, a "# Parameters fro
706 line followed by a line containi
707 parameters prefixed by "# ".
708 buddyinfo Kernel memory allocator informat
709 bus Directory containing bus specifi
710 cmdline Kernel command line, both from b
711 in the kernel image
712 cpuinfo Info about the CPU
713 devices Available devices (block and cha
714 dma Used DMS channels
715 filesystems Supported filesystems
716 driver Various drivers grouped here, cu
717 execdomains Execdomains, related to security
718 fb Frame Buffer devices
719 fs File system parameters, currentl
720 ide Directory containing info about
721 interrupts Interrupt usage
722 iomem Memory map
723 ioports I/O port usage
724 irq Masks for irq to cpu affinity
725 isapnp ISA PnP (Plug&Play) Info
726 kcore Kernel core image (can be ELF or
727 kmsg Kernel messages
728 ksyms Kernel symbol table
729 loadavg Load average of last 1, 5 & 15 m
730 number of processes currently
731 total number of processes in s
732 last pid created.
733 All fields are separated by on
734 processes currently runnable"
735 in system", which are separate
736 0.61 0.61 0.55 3/828 22084
737 locks Kernel locks
738 meminfo Memory info
739 misc Miscellaneous
740 modules List of loaded modules
741 mounts Mounted filesystems
742 net Networking info (see text)
743 pagetypeinfo Additional page allocator inform
744 partitions Table of partitions known to the
745 pci Deprecated info of PCI bus (new
746 decoupled by lspci
747 rtc Real time clock
748 scsi SCSI info (see text)
749 slabinfo Slab pool info
750 softirqs softirq usage
751 stat Overall statistics
752 swaps Swap space utilization
753 sys See chapter 2
754 sysvipc Info of SysVIPC Resources (msg,
755 tty Info of tty drivers
756 uptime Wall clock since boot, combined
757 version Kernel version
758 video bttv info of video resources
759 vmallocinfo Show vmalloced areas
760 ============ ================================
761
762 You can, for example, check which interrup
763 they are used for by looking in the file /proc
764
765 > cat /proc/interrupts
766 CPU0
767 0: 8728810 XT-PIC timer
768 1: 895 XT-PIC keyboard
769 2: 0 XT-PIC cascade
770 3: 531695 XT-PIC aha152x
771 4: 2014133 XT-PIC serial
772 5: 44401 XT-PIC pcnet_cs
773 8: 2 XT-PIC rtc
774 11: 8 XT-PIC i82365
775 12: 182918 XT-PIC PS/2 Mouse
776 13: 1 XT-PIC fpu
777 14: 1232265 XT-PIC ide0
778 15: 7 XT-PIC ide1
779 NMI: 0
780
781 In 2.4.* a couple of lines where added to this
782 output of a SMP machine)::
783
784 > cat /proc/interrupts
785
786 CPU0 CPU1
787 0: 1243498 1214548 IO-APIC-edge
788 1: 8949 8958 IO-APIC-edge
789 2: 0 0 XT-PIC
790 5: 11286 10161 IO-APIC-edge
791 8: 1 0 IO-APIC-edge
792 9: 27422 27407 IO-APIC-edge
793 12: 113645 113873 IO-APIC-edge
794 13: 0 0 XT-PIC
795 14: 22491 24012 IO-APIC-edge
796 15: 2183 2415 IO-APIC-edge
797 17: 30564 30414 IO-APIC-level
798 18: 177 164 IO-APIC-level
799 NMI: 2457961 2457959
800 LOC: 2457882 2457881
801 ERR: 2155
802
803 NMI is incremented in this case because every
804 (Non Maskable Interrupt) which is used by the
805
806 LOC is the local interrupt counter of the inte
807
808 ERR is incremented in the case of errors in th
809 connects the CPUs in a SMP system. This means
810 the IO-APIC automatically retry the transmissi
811 problem, but you should read the SMP-FAQ.
812
813 In 2.6.2* /proc/interrupts was expanded again.
814 /proc/interrupts to display every IRQ vector i
815 just those considered 'most important'. The n
816
817 THR
818 interrupt raised when a machine check thresh
819 (typically counting ECC corrected errors of
820 a configurable threshold. Only available on
821
822 TRM
823 a thermal event interrupt occurs when a temp
824 has been exceeded for the CPU. This interru
825 when the temperature drops back to normal.
826
827 SPU
828 a spurious interrupt is some interrupt that
829 by some IO device before it could be fully p
830 the APIC sees the interrupt but does not kno
831 For this case the APIC will generate the int
832 of 0xff. This might also be generated by chi
833
834 RES, CAL, TLB
835 rescheduling, call and TLB flush interrupts
836 sent from one CPU to another per the needs o
837 their statistics are used by kernel develope
838 determine the occurrence of interrupts of th
839
840 The above IRQ vectors are displayed only when
841 the threshold vector does not exist on x86_64
842 suppressed when the system is a uniprocessor.
843 i386 and x86_64 platforms support the new IRQ
844
845 Of some interest is the introduction of the /p
846 It could be used to set IRQ to CPU affinity. T
847 IRQ to only one CPU, or to exclude a CPU of ha
848 irq subdir is one subdir for each IRQ, and two
849 prof_cpu_mask.
850
851 For example::
852
853 > ls /proc/irq/
854 0 10 12 14 16 18 2 4 6 8 prof_cpu_
855 1 11 13 15 17 19 3 5 7 9 default_s
856 > ls /proc/irq/0/
857 smp_affinity
858
859 smp_affinity is a bitmask, in which you can sp
860 IRQ. You can set it by doing::
861
862 > echo 1 > /proc/irq/10/smp_affinity
863
864 This means that only the first CPU will handle
865 5 which means that only the first and third CP
866
867 The contents of each smp_affinity file is the
868
869 > cat /proc/irq/0/smp_affinity
870 ffffffff
871
872 There is an alternate interface, smp_affinity_
873 a CPU range instead of a bitmask::
874
875 > cat /proc/irq/0/smp_affinity_list
876 1024-1031
877
878 The default_smp_affinity mask applies to all n
879 IRQs which have not yet been allocated/activat
880 /proc/irq/[0-9]* directory.
881
882 The node file on an SMP system shows the node
883 reports itself as being attached. This hardwar
884 include information about any possible driver
885
886 prof_cpu_mask specifies which CPUs are to be p
887 profiler. Default value is ffffffff (all CPUs
888
889 The way IRQs are routed is handled by the IO-A
890 between all the CPUs which are allowed to hand
891 more info than you and does a better job than
892 best choice for almost everyone. [Note this a
893 that support "Round Robin" interrupt distribut
894
895 There are three more important subdirectori
896 The general rule is that the contents, o
897 directories, depend on your kernel configurat
898 directory scsi may not exist. The same is t
899 only when networking support is present in the
900
901 The slabinfo file gives information about
902 Linux uses slab pools for memory management
903 Commonly used objects have their own slab
904 directory cache, and so on).
905
906 ::
907
908 > cat /proc/buddyinfo
909
910 Node 0, zone DMA 0 4 5
911 Node 0, zone Normal 1 0 0
912 Node 0, zone HighMem 2 0 0
913
914 External fragmentation is a problem under some
915 useful tool for helping diagnose these problem
916 clue as to how big an area you can safely allo
917 allocation failed.
918
919 Each column represents the number of pages of
920 available. In this case, there are 0 chunks o
921 ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DM
922 available in ZONE_NORMAL, etc...
923
924 More information relevant to external fragment
925 pagetypeinfo::
926
927 > cat /proc/pagetypeinfo
928 Page block order: 9
929 Pages per block: 512
930
931 Free pages count per migrate type at order
932 Node 0, zone DMA, type Unmovabl
933 Node 0, zone DMA, type Reclaimabl
934 Node 0, zone DMA, type Movabl
935 Node 0, zone DMA, type Reserv
936 Node 0, zone DMA, type Isolat
937 Node 0, zone DMA32, type Unmovabl
938 Node 0, zone DMA32, type Reclaimabl
939 Node 0, zone DMA32, type Movabl
940 Node 0, zone DMA32, type Reserv
941 Node 0, zone DMA32, type Isolat
942
943 Number of blocks type Unmovable Recla
944 Node 0, zone DMA 2
945 Node 0, zone DMA32 41
946
947 Fragmentation avoidance in the kernel works by
948 migrate types into the same contiguous regions
949 A page block is typically the size of the defa
950 X86-64. By keeping pages grouped based on thei
951 can reclaim pages within a page block to satis
952
953 The pagetypinfo begins with information on the
954 then gives the same type of information as bud
955 by migrate-type and finishes with details on h
956 type exist.
957
958 If min_free_kbytes has been tuned correctly (r
959 from libhugetlbfs https://github.com/libhugetl
960 make an estimate of the likely number of huge
961 at a given point in time. All the "Movable" bl
962 unless memory has been mlock()'d. Some of the
963 also be allocatable although a lot of filesyst
964 reclaimed to achieve this.
965
966
967 allocinfo
968 ~~~~~~~~~
969
970 Provides information about memory allocations
971 base. Each allocation in the code is identifie
972 number, module (if originates from a loadable
973 the allocation. The number of bytes allocated
974 location are reported. The first line indicate
975 second line is the header listing fields in th
976
977 Example output.
978
979 ::
980
981 > tail -n +3 /proc/allocinfo | sort -rn
982 127664128 31168 mm/page_ext.c:270 func:a
983 56373248 4737 mm/slub.c:2259 func:allo
984 14880768 3633 mm/readahead.c:247 func:
985 14417920 3520 mm/mm_init.c:2530 func:a
986 13377536 234 block/blk-mq.c:3421 func
987 11718656 2861 mm/filemap.c:1919 func:_
988 9192960 2800 kernel/fork.c:307 func:a
989 4206592 4 net/netfilter/nf_conntra
990 4136960 1010 drivers/staging/ctagmod/
991 3940352 962 mm/memory.c:4214 func:al
992 2894464 22613 fs/kernfs/dir.c:615 func
993 ...
994
995
996 meminfo
997 ~~~~~~~
998
999 Provides information about distribution and ut
1000 varies by architecture and compile options.
1001 here overlap. The memory reported by the non
1002 add up to the overall memory usage and the di
1003 can be substantial. In many cases there are
1004 additional memory using subsystem specific in
1005 /proc/net/sockstat for TCP memory allocations
1006
1007 Example output. You may not have all of these
1008
1009 ::
1010
1011 > cat /proc/meminfo
1012
1013 MemTotal: 32858820 kB
1014 MemFree: 21001236 kB
1015 MemAvailable: 27214312 kB
1016 Buffers: 581092 kB
1017 Cached: 5587612 kB
1018 SwapCached: 0 kB
1019 Active: 3237152 kB
1020 Inactive: 7586256 kB
1021 Active(anon): 94064 kB
1022 Inactive(anon): 4570616 kB
1023 Active(file): 3143088 kB
1024 Inactive(file): 3015640 kB
1025 Unevictable: 0 kB
1026 Mlocked: 0 kB
1027 SwapTotal: 0 kB
1028 SwapFree: 0 kB
1029 Zswap: 1904 kB
1030 Zswapped: 7792 kB
1031 Dirty: 12 kB
1032 Writeback: 0 kB
1033 AnonPages: 4654780 kB
1034 Mapped: 266244 kB
1035 Shmem: 9976 kB
1036 KReclaimable: 517708 kB
1037 Slab: 660044 kB
1038 SReclaimable: 517708 kB
1039 SUnreclaim: 142336 kB
1040 KernelStack: 11168 kB
1041 PageTables: 20540 kB
1042 SecPageTables: 0 kB
1043 NFS_Unstable: 0 kB
1044 Bounce: 0 kB
1045 WritebackTmp: 0 kB
1046 CommitLimit: 16429408 kB
1047 Committed_AS: 7715148 kB
1048 VmallocTotal: 34359738367 kB
1049 VmallocUsed: 40444 kB
1050 VmallocChunk: 0 kB
1051 Percpu: 29312 kB
1052 EarlyMemtestBad: 0 kB
1053 HardwareCorrupted: 0 kB
1054 AnonHugePages: 4149248 kB
1055 ShmemHugePages: 0 kB
1056 ShmemPmdMapped: 0 kB
1057 FileHugePages: 0 kB
1058 FilePmdMapped: 0 kB
1059 CmaTotal: 0 kB
1060 CmaFree: 0 kB
1061 HugePages_Total: 0
1062 HugePages_Free: 0
1063 HugePages_Rsvd: 0
1064 HugePages_Surp: 0
1065 Hugepagesize: 2048 kB
1066 Hugetlb: 0 kB
1067 DirectMap4k: 401152 kB
1068 DirectMap2M: 10008576 kB
1069 DirectMap1G: 24117248 kB
1070
1071 MemTotal
1072 Total usable RAM (i.e. physical
1073 bits and the kernel binary code
1074 MemFree
1075 Total free RAM. On highmem syst
1076 MemAvailable
1077 An estimate of how much memory
1078 applications, without swapping.
1079 SReclaimable, the size of the f
1080 watermarks in each zone.
1081 The estimate takes into account
1082 page cache to function well, an
1083 slab will be reclaimable, due t
1084 impact of those factors will va
1085 Buffers
1086 Relatively temporary storage fo
1087 shouldn't get tremendously larg
1088 Cached
1089 In-memory cache for files read
1090 pagecache) as well as tmpfs & s
1091 Doesn't include SwapCached.
1092 SwapCached
1093 Memory that once was swapped ou
1094 still also is in the swapfile (
1095 doesn't need to be swapped out
1096 in the swapfile. This saves I/O
1097 Active
1098 Memory that has been used more
1099 reclaimed unless absolutely nec
1100 Inactive
1101 Memory which has been less rece
1102 eligible to be reclaimed for ot
1103 Unevictable
1104 Memory allocated for userspace
1105 as mlocked pages, ramfs backing
1106 Mlocked
1107 Memory locked with mlock().
1108 HighTotal, HighFree
1109 Highmem is all memory above ~86
1110 Highmem areas are for use by us
1111 for the pagecache. The kernel
1112 this memory, making it slower t
1113 LowTotal, LowFree
1114 Lowmem is memory which can be u
1115 highmem can be used for, but it
1116 kernel's use for its own data s
1117 other things, it is where every
1118 allocated. Bad things happen w
1119 SwapTotal
1120 total amount of swap space avai
1121 SwapFree
1122 Memory which has been evicted f
1123 on the disk
1124 Zswap
1125 Memory consumed by the zswap ba
1126 Zswapped
1127 Amount of anonymous memory stor
1128 Dirty
1129 Memory which is waiting to get
1130 Writeback
1131 Memory which is actively being
1132 AnonPages
1133 Non-file backed pages mapped in
1134 Mapped
1135 files which have been mmapped,
1136 Shmem
1137 Total memory used by shared mem
1138 KReclaimable
1139 Kernel allocations that the ker
1140 under memory pressure. Includes
1141 direct allocations with a shrin
1142 Slab
1143 in-kernel data structures cache
1144 SReclaimable
1145 Part of Slab, that might be rec
1146 SUnreclaim
1147 Part of Slab, that cannot be re
1148 KernelStack
1149 Memory consumed by the kernel s
1150 PageTables
1151 Memory consumed by userspace pa
1152 SecPageTables
1153 Memory consumed by secondary pa
1154 KVM mmu and IOMMU allocations o
1155 NFS_Unstable
1156 Always zero. Previous counted p
1157 the server, but has not been co
1158 Bounce
1159 Memory used for block device "b
1160 WritebackTmp
1161 Memory used by FUSE for tempora
1162 CommitLimit
1163 Based on the overcommit ratio (
1164 this is the total amount of me
1165 be allocated on the system. Thi
1166 if strict overcommit accounting
1167 'vm.overcommit_memory').
1168
1169 The CommitLimit is calculated w
1170
1171 CommitLimit = ([total RAM pag
1172 overcommit_rat
1173
1174 For example, on a system with 1
1175 of swap with a `vm.overcommit_r
1176 yield a CommitLimit of 7.3G.
1177
1178 For more details, see the memor
1179 in mm/overcommit-accounting.
1180 Committed_AS
1181 The amount of memory presently
1182 The committed memory is a sum o
1183 has been allocated by processes
1184 "used" by them as of yet. A pro
1185 of memory, but only touches 300
1186 using 1G. This 1G is memory whi
1187 by the VM and can be used at an
1188 application. With strict overco
1189 (mode 2 in 'vm.overcommit_memor
1190 exceed the CommitLimit (detaile
1191 This is useful if one needs to
1192 not fail due to lack of memory
1193 successfully allocated.
1194 VmallocTotal
1195 total size of vmalloc virtual a
1196 VmallocUsed
1197 amount of vmalloc area which is
1198 VmallocChunk
1199 largest contiguous block of vma
1200 Percpu
1201 Memory allocated to the percpu
1202 allocations. This stat excludes
1203 EarlyMemtestBad
1204 The amount of RAM/memory in kB,
1205 by early memtest. If memtest wa
1206 be displayed at all. Size is ne
1207 That means if 0 kB is reported,
1208 there was at least one pass of
1209 found a single faulty byte of R
1210 HardwareCorrupted
1211 The amount of RAM/memory in KB,
1212 corrupted.
1213 AnonHugePages
1214 Non-file backed huge pages mapp
1215 ShmemHugePages
1216 Memory used by shared memory (s
1217 with huge pages
1218 ShmemPmdMapped
1219 Shared memory mapped into users
1220 FileHugePages
1221 Memory used for filesystem data
1222 with huge pages
1223 FilePmdMapped
1224 Page cache mapped into userspac
1225 CmaTotal
1226 Memory reserved for the Contigu
1227 CmaFree
1228 Free remaining memory in the CM
1229 HugePages_Total, HugePages_Free, HugePages_Rs
1230 See Documentation/admin-guide/m
1231 DirectMap4k, DirectMap2M, DirectMap1G
1232 Breakdown of page table sizes u
1233 identity mapping of RAM
1234
1235 vmallocinfo
1236 ~~~~~~~~~~~
1237
1238 Provides information about vmalloced/vmaped a
1239 containing the virtual address range of the a
1240 caller information of the creator, and option
1241 on the kind of area:
1242
1243 ========== ================================
1244 pages=nr number of pages
1245 phys=addr if a physical address was specif
1246 ioremap I/O mapping (ioremap() and frien
1247 vmalloc vmalloc() area
1248 vmap vmap()ed pages
1249 user VM_USERMAP area
1250 vpages buffer for pages pointers was vm
1251 N<node>=nr (Only on NUMA kernels)
1252 Number of pages allocated on mem
1253 ========== ================================
1254
1255 ::
1256
1257 > cat /proc/vmallocinfo
1258 0xffffc20000000000-0xffffc20000201000 210
1259 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2
1260 0xffffc20000201000-0xffffc20000302000 105
1261 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=6
1262 0xffffc20000302000-0xffffc20000304000
1263 phys=7fee8000 ioremap
1264 0xffffc20000304000-0xffffc20000307000 1
1265 phys=7fee7000 ioremap
1266 0xffffc2000031d000-0xffffc2000031f000
1267 0xffffc2000031f000-0xffffc2000032b000 4
1268 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=
1269 0xffffc2000033a000-0xffffc2000033d000 1
1270 pages=2 vmalloc N1=2
1271 0xffffc20000347000-0xffffc2000034c000 2
1272 /0x130 [x_tables] pages=4 vmalloc N0=4
1273 0xffffffffa0000000-0xffffffffa000f000 6
1274 pages=14 vmalloc N2=14
1275 0xffffffffa000f000-0xffffffffa0014000 2
1276 pages=4 vmalloc N1=4
1277 0xffffffffa0014000-0xffffffffa0017000 1
1278 pages=2 vmalloc N1=2
1279 0xffffffffa0017000-0xffffffffa0022000 4
1280 pages=10 vmalloc N0=10
1281
1282
1283 softirqs
1284 ~~~~~~~~
1285
1286 Provides counts of softirq handlers serviced
1287
1288 ::
1289
1290 > cat /proc/softirqs
1291 CPU0 CPU1 CPU2
1292 HI: 0 0 0
1293 TIMER: 27166 27120 27097
1294 NET_TX: 0 0 0
1295 NET_RX: 42 0 0
1296 BLOCK: 0 0 107
1297 TASKLET: 0 0 0
1298 SCHED: 27035 26983 26971
1299 HRTIMER: 0 0 0
1300 RCU: 1678 1769 2178
1301
1302 1.3 Networking info in /proc/net
1303 --------------------------------
1304
1305 The subdirectory /proc/net follows the us
1306 additional values you get for IP version
1307 support this. Table 1-9 lists the files and t
1308
1309
1310 .. table:: Table 1-8: IPv6 info in /proc/net
1311
1312 ========== =================================
1313 File Content
1314 ========== =================================
1315 udp6 UDP sockets (IPv6)
1316 tcp6 TCP sockets (IPv6)
1317 raw6 Raw device statistics (IPv6)
1318 igmp6 IP multicast addresses, which thi
1319 if_inet6 List of IPv6 interface addresses
1320 ipv6_route Kernel routing table for IPv6
1321 rt6_stats Global IPv6 routing tables statis
1322 sockstat6 Socket statistics (IPv6)
1323 snmp6 Snmp data (IPv6)
1324 ========== =================================
1325
1326 .. table:: Table 1-9: Network info in /proc/n
1327
1328 ============= ==============================
1329 File Content
1330 ============= ==============================
1331 arp Kernel ARP table
1332 dev network devices with statistic
1333 dev_mcast the Layer2 multicast groups a
1334 (interface index, label, numbe
1335 addresses).
1336 dev_stat network device status
1337 ip_fwchains Firewall chain linkage
1338 ip_fwnames Firewall chain names
1339 ip_masq Directory containing the masqu
1340 ip_masquerade Major masquerading table
1341 netstat Network statistics
1342 raw raw device statistics
1343 route Kernel routing table
1344 rpc Directory containing rpc info
1345 rt_cache Routing cache
1346 snmp SNMP data
1347 sockstat Socket statistics
1348 softnet_stat Per-CPU incoming packets queue
1349 tcp TCP sockets
1350 udp UDP sockets
1351 unix UNIX domain sockets
1352 wireless Wireless interface data (Wavel
1353 igmp IP multicast addresses, which
1354 psched Global packet scheduler parame
1355 netlink List of PF_NETLINK sockets
1356 ip_mr_vifs List of multicast virtual inte
1357 ip_mr_cache List of multicast routing cach
1358 ============= ==============================
1359
1360 You can use this information to see which
1361 your system and how much traffic was routed o
1362
1363 > cat /proc/net/dev
1364 Inter-|Receive
1365 face |bytes packets errs drop fifo fram
1366 lo: 908188 5596 0 0 0
1367 ppp0:15475140 20721 410 0 0 41
1368 eth0: 614530 7085 0 0 0
1369
1370 ...] Transmit
1371 ...] bytes packets errs drop fifo colls
1372 ...] 908188 5596 0 0 0 0
1373 ...] 1375103 17405 0 0 0 0
1374 ...] 1703981 5535 0 0 0 3
1375
1376 In addition, each Channel Bond interface has
1377 example, the bond0 device will have a directo
1378 It will contain information that is specific
1379 current slaves of the bond, the link status o
1380 many times the slaves link has failed.
1381
1382 1.4 SCSI info
1383 -------------
1384
1385 If you have a SCSI or ATA host adapter in you
1386 subdirectory named after the driver for this
1387 You'll also see a list of all recognized SCSI
1388
1389 >cat /proc/scsi/scsi
1390 Attached devices:
1391 Host: scsi0 Channel: 00 Id: 00 Lun: 00
1392 Vendor: IBM Model: DGHS09U
1393 Type: Direct-Access
1394 Host: scsi0 Channel: 00 Id: 06 Lun: 00
1395 Vendor: PIONEER Model: CD-ROM DR-U06S
1396 Type: CD-ROM
1397
1398
1399 The directory named after the driver has o
1400 the system. These files contain informatio
1401 the used IRQ and the IO address range. T
1402 dependent on the adapter you use. The exampl
1403 AHA-2940 SCSI adapter::
1404
1405 > cat /proc/scsi/aic7xxx/0
1406
1407 Adaptec AIC7xxx driver version: 5.1.19/3.2.
1408 Compile Options:
1409 TCQ Enabled By Default : Disabled
1410 AIC7XXX_PROC_STATS : Disabled
1411 AIC7XXX_RESET_DELAY : 5
1412 Adapter Configuration:
1413 SCSI Adapter: Adaptec AHA-294X U
1414 Ultra Wide Contr
1415 PCI MMAPed I/O Base: 0xeb001000
1416 Adapter SEEPROM Config: SEEPROM found and
1417 Adaptec SCSI BIOS: Enabled
1418 IRQ: 10
1419 SCBs: Active 0, Max Acti
1420 Allocated 15, HW 1
1421 Interrupts: 160328
1422 BIOS Control Word: 0x18b6
1423 Adapter Control Word: 0x005b
1424 Extended Translation: Enabled
1425 Disconnect Enable Flags: 0xffff
1426 Ultra Enable Flags: 0x0001
1427 Tag Queue Enable Flags: 0x0000
1428 Ordered Queue Tag Flags: 0x0000
1429 Default Tag Queue Depth: 8
1430 Tagged Queue By Device array for aic7xx
1431 {255,255,255,255,255,255,255,255,255,
1432 Actual queue depth per device for aic7x
1433 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}
1434 Statistics:
1435 (scsi0:0:0:0)
1436 Device using Wide/Sync transfers at 40.0
1437 Transinfo settings: current(12/8/1/0), go
1438 Total transfers 160151 (74577 reads and 8
1439 (scsi0:0:6:0)
1440 Device using Narrow/Sync transfers at 5.0
1441 Transinfo settings: current(50/15/0/0), g
1442 Total transfers 0 (0 reads and 0 writes)
1443
1444
1445 1.5 Parallel port info in /proc/parport
1446 ---------------------------------------
1447
1448 The directory /proc/parport contains inform
1449 your system. It has one subdirectory for
1450 number (0,1,2,...).
1451
1452 These directories contain the four files show
1453
1454
1455 .. table:: Table 1-10: Files in /proc/parport
1456
1457 ========= ==================================
1458 File Content
1459 ========= ==================================
1460 autoprobe Any IEEE-1284 device ID informatio
1461 devices list of the device drivers using t
1462 name of the device currently using
1463 against any).
1464 hardware Parallel port's base address, IRQ
1465 irq IRQ that parport is using for that
1466 file to allow you to alter it by w
1467 number or none).
1468 ========= ==================================
1469
1470 1.6 TTY info in /proc/tty
1471 -------------------------
1472
1473 Information about the available and actual
1474 directory /proc/tty. You'll find entries fo
1475 this directory, as shown in Table 1-11.
1476
1477
1478 .. table:: Table 1-11: Files in /proc/tty
1479
1480 ============= ==============================
1481 File Content
1482 ============= ==============================
1483 drivers list of drivers and their usag
1484 ldiscs registered line disciplines
1485 driver/serial usage statistic and status of
1486 ============= ==============================
1487
1488 To see which tty's are currently in use,
1489 /proc/tty/drivers::
1490
1491 > cat /proc/tty/drivers
1492 pty_slave /dev/pts 136 0-
1493 pty_master /dev/ptm 128 0-
1494 pty_slave /dev/ttyp 3 0-
1495 pty_master /dev/pty 2 0-
1496 serial /dev/cua 5 64
1497 serial /dev/ttyS 4 64
1498 /dev/tty0 /dev/tty0 4
1499 /dev/ptmx /dev/ptmx 5
1500 /dev/console /dev/console 5
1501 /dev/tty /dev/tty 5
1502 unknown /dev/tty 4 1
1503
1504
1505 1.7 Miscellaneous kernel statistics in /proc/
1506 ---------------------------------------------
1507
1508 Various pieces of information about kerne
1509 /proc/stat file. All of the numbers report
1510 since the system first booted. For a quick l
1511
1512 > cat /proc/stat
1513 cpu 237902850 368826709 106375398 18735175
1514 cpu0 60045249 91891769 26331539 468411416 4
1515 cpu1 59746288 91759249 26609887 468860630 3
1516 cpu2 59489247 92985423 26904446 467808813 1
1517 cpu3 58622065 92190267 26529524 468436680 1
1518 intr 8688370575 8 3373 0 0 0 0 0 0 1 40791
1519 ctxt 22848221062
1520 btime 1605316999
1521 processes 746787147
1522 procs_running 2
1523 procs_blocked 0
1524 softirq 12121874454 100099120 3938138295 12
1525
1526 The very first "cpu" line aggregates the nu
1527 lines. These numbers identify the amount of
1528 different kinds of work. Time units are in U
1529 second). The meanings of the columns are as
1530
1531 - user: normal processes executing in user mo
1532 - nice: niced processes executing in user mod
1533 - system: processes executing in kernel mode
1534 - idle: twiddling thumbs
1535 - iowait: In a word, iowait stands for waitin
1536 are several problems:
1537
1538 1. CPU will not wait for I/O to complete, i
1539 waiting for I/O to complete. When CPU go
1540 outstanding task I/O, another task will
1541 2. In a multi-core CPU, the task waiting fo
1542 on any CPU, so the iowait of each CPU is
1543 3. The value of iowait field in /proc/stat
1544 conditions.
1545
1546 So, the iowait is not reliable by reading f
1547 - irq: servicing interrupts
1548 - softirq: servicing softirqs
1549 - steal: involuntary wait
1550 - guest: running a normal guest
1551 - guest_nice: running a niced guest
1552
1553 The "intr" line gives counts of interrupts s
1554 of the possible system interrupts. The fir
1555 interrupts serviced including unnumbered a
1556 each subsequent column is the total for tha
1557 Unnumbered interrupts are not shown, only sum
1558
1559 The "ctxt" line gives the total number of con
1560
1561 The "btime" line gives the time at which the
1562 the Unix epoch.
1563
1564 The "processes" line gives the number of pro
1565 includes (but is not limited to) those cre
1566 clone() system calls.
1567
1568 The "procs_running" line gives the total numb
1569 running or ready to run (i.e., the total numb
1570
1571 The "procs_blocked" line gives the number
1572 waiting for I/O to complete.
1573
1574 The "softirq" line gives counts of softirqs s
1575 of the possible system softirqs. The first co
1576 softirqs serviced; each subsequent column is
1577 softirq.
1578
1579
1580 1.8 Ext4 file system parameters
1581 -------------------------------
1582
1583 Information about mounted ext4 file systems c
1584 /proc/fs/ext4. Each mounted filesystem will
1585 /proc/fs/ext4 based on its device name (i.e.,
1586 /proc/fs/ext4/sda9 or /proc/fs/ext4/dm-0).
1587 directory are shown in Table 1-12, below.
1588
1589 .. table:: Table 1-12: Files in /proc/fs/ext4
1590
1591 ============== ============================
1592 File Content
1593 mb_groups details of multiblock alloca
1594 ============== ============================
1595
1596 1.9 /proc/consoles
1597 -------------------
1598 Shows registered system console lines.
1599
1600 To see which character device lines are curre
1601 /dev/console, you may simply look into the fi
1602
1603 > cat /proc/consoles
1604 tty0 -WU (ECp) 4:7
1605 ttyS0 -W- (Ep) 4:64
1606
1607 The columns are:
1608
1609 +--------------------+-----------------------
1610 | device | name of the device
1611 +====================+=======================
1612 | operations | * R = can do read oper
1613 | | * W = can do write ope
1614 | | * U = can do unblank
1615 +--------------------+-----------------------
1616 | flags | * E = it is enabled
1617 | | * C = it is preferred
1618 | | * B = it is primary bo
1619 | | * p = it is used for p
1620 | | * b = it is not a TTY
1621 | | * a = it is safe to us
1622 +--------------------+-----------------------
1623 | major:minor | major and minor number
1624 | | colon
1625 +--------------------+-----------------------
1626
1627 Summary
1628 -------
1629
1630 The /proc file system serves information abou
1631 allows access to process data but also allows
1632 by reading files in the hierarchy.
1633
1634 The directory structure of /proc reflects t
1635 it easy, if not obvious, where to look for sp
1636
1637 Chapter 2: Modifying System Parameters
1638 ======================================
1639
1640 In This Chapter
1641 ---------------
1642
1643 * Modifying kernel parameters by writing into
1644 * Exploring the files which modify certain pa
1645 * Review of the /proc/sys file tree
1646
1647 ---------------------------------------------
1648
1649 A very interesting part of /proc is the dire
1650 a source of information, it also allows yo
1651 kernel. Be very careful when attempting th
1652 but you can also cause it to crash. Ne
1653 production system. Set up a development m
1654 everything works the way you want it to. Y
1655 reboot the machine once an error has been mad
1656
1657 To change a value, simply echo the new v
1658 You need to be root to do this. You can cre
1659 to perform this every time your system boots.
1660
1661 The files in /proc/sys can be used to fine t
1662 general things in the operation of the Linu
1663 can inadvertently disrupt your system, it
1664 documentation and source before actually ma
1665 very careful when writing to any of thes
1666 change slightly between the 2.1.* and the 2.2
1667 review the kernel documentation in the direct
1668 This chapter is heavily based on the docu
1669 kernels, and became part of it in version 2.2
1670
1671 Please see: Documentation/admin-guide/sysctl/
1672 these entries.
1673
1674 Summary
1675 -------
1676
1677 Certain aspects of kernel behavior can be
1678 need to recompile the kernel, or even to re
1679 /proc/sys tree can not only be read, but al
1680 command to write value into these files, ther
1681 of the kernel.
1682
1683
1684 Chapter 3: Per-process Parameters
1685 =================================
1686
1687 3.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_sco
1688 ---------------------------------------------
1689
1690 These files can be used to adjust the badness
1691 process gets killed in out of memory (oom) co
1692
1693 The badness heuristic assigns a value to each
1694 (never kill) to 1000 (always kill) to determi
1695 units are roughly a proportion along that ran
1696 may allocate from based on an estimation of i
1697 For example, if a task is using all allowed m
1698 1000. If it is using half of its allowed mem
1699
1700 The amount of "allowed" memory depends on the
1701 was called. If it is due to the memory assig
1702 being exhausted, the allowed memory represent
1703 cpuset. If it is due to a mempolicy's node(s
1704 memory represents the set of mempolicy nodes.
1705 limit (or swap limit) being reached, the allo
1706 limit. Finally, if it is due to the entire s
1707 allowed memory represents all allocatable res
1708
1709 The value of /proc/<pid>/oom_score_adj is add
1710 is used to determine which task to kill. Acc
1711 (OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_M
1712 polarize the preference for oom killing eithe
1713 task or completely disabling it. The lowest
1714 equivalent to disabling oom killing entirely
1715 report a badness score of 0.
1716
1717 Consequently, it is very simple for userspace
1718 consider for each task. Setting a /proc/<pid
1719 example, is roughly equivalent to allowing th
1720 same system, cpuset, mempolicy, or memory con
1721 50% more memory. A value of -500, on the oth
1722 equivalent to discounting 50% of the task's a
1723 as scoring against the task.
1724
1725 For backwards compatibility with previous ker
1726 be used to tune the badness score. Its accep
1727 (OOM_ADJUST_MIN) to +15 (OOM_ADJUST_MAX) and
1728 (OOM_DISABLE) to disable oom killing entirely
1729 scaled linearly with /proc/<pid>/oom_score_ad
1730
1731 The value of /proc/<pid>/oom_score_adj may be
1732 value set by a CAP_SYS_RESOURCE process. To r
1733 requires CAP_SYS_RESOURCE.
1734
1735
1736 3.2 /proc/<pid>/oom_score - Display current o
1737 ---------------------------------------------
1738
1739 This file can be used to check the current sc
1740 any given <pid>. Use it together with /proc/<
1741 process should be killed in an out-of-memory
1742
1743 Please note that the exported value includes
1744 effectively in range [0,2000].
1745
1746
1747 3.3 /proc/<pid>/io - Display the IO accounti
1748 ---------------------------------------------
1749
1750 This file contains IO statistics for each run
1751
1752 Example
1753 ~~~~~~~
1754
1755 ::
1756
1757 test:/tmp # dd if=/dev/zero of=/tmp/test.
1758 [1] 3828
1759
1760 test:/tmp # cat /proc/3828/io
1761 rchar: 323934931
1762 wchar: 323929600
1763 syscr: 632687
1764 syscw: 632675
1765 read_bytes: 0
1766 write_bytes: 323932160
1767 cancelled_write_bytes: 0
1768
1769
1770 Description
1771 ~~~~~~~~~~~
1772
1773 rchar
1774 ^^^^^
1775
1776 I/O counter: chars read
1777 The number of bytes which this task has cause
1778 is simply the sum of bytes which this process
1779 It includes things like tty IO and it is unaf
1780 physical disk IO was required (the read might
1781 pagecache).
1782
1783
1784 wchar
1785 ^^^^^
1786
1787 I/O counter: chars written
1788 The number of bytes which this task has cause
1789 to disk. Similar caveats apply here as with r
1790
1791
1792 syscr
1793 ^^^^^
1794
1795 I/O counter: read syscalls
1796 Attempt to count the number of read I/O opera
1797 and pread().
1798
1799
1800 syscw
1801 ^^^^^
1802
1803 I/O counter: write syscalls
1804 Attempt to count the number of write I/O oper
1805 write() and pwrite().
1806
1807
1808 read_bytes
1809 ^^^^^^^^^^
1810
1811 I/O counter: bytes read
1812 Attempt to count the number of bytes which th
1813 be fetched from the storage layer. Done at th
1814 accurate for block-backed filesystems. <pleas
1815 CIFS at a later time>
1816
1817
1818 write_bytes
1819 ^^^^^^^^^^^
1820
1821 I/O counter: bytes written
1822 Attempt to count the number of bytes which th
1823 the storage layer. This is done at page-dirty
1824
1825
1826 cancelled_write_bytes
1827 ^^^^^^^^^^^^^^^^^^^^^
1828
1829 The big inaccuracy here is truncate. If a pro
1830 then deletes the file, it will in fact perfor
1831 been accounted as having caused 1MB of write.
1832 In other words: The number of bytes which thi
1833 by truncating pagecache. A task can cause "ne
1834 truncates some dirty pagecache, some IO which
1835 for (in its write_bytes) will not be happenin
1836 from the truncating task's write_bytes, but t
1837 that.
1838
1839
1840 .. Note::
1841
1842 At its current implementation state, this
1843 if process A reads process B's /proc/pid/i
1844 of those 64-bit counters, process A could
1845
1846
1847 More information about this can be found with
1848 Documentation/accounting.
1849
1850 3.4 /proc/<pid>/coredump_filter - Core dump f
1851 ---------------------------------------------
1852 When a process is dumped, all anonymous memor
1853 long as the size of the core file isn't limit
1854 to dump some memory segments, for example, hu
1855 Conversely, sometimes we want to save file-ba
1856 file, not only the individual files.
1857
1858 /proc/<pid>/coredump_filter allows you to cus
1859 will be dumped when the <pid> process is dump
1860 of memory types. If a bit of the bitmask is s
1861 corresponding memory type are dumped, otherwi
1862
1863 The following 9 memory types are supported:
1864
1865 - (bit 0) anonymous private memory
1866 - (bit 1) anonymous shared memory
1867 - (bit 2) file-backed private memory
1868 - (bit 3) file-backed shared memory
1869 - (bit 4) ELF header pages in file-backed p
1870 effective only if the bit 2 is cleared)
1871 - (bit 5) hugetlb private memory
1872 - (bit 6) hugetlb shared memory
1873 - (bit 7) DAX private memory
1874 - (bit 8) DAX shared memory
1875
1876 Note that MMIO pages such as frame buffer a
1877 are always dumped regardless of the bitmask
1878
1879 Note that bits 0-4 don't affect hugetlb or
1880 only affected by bit 5-6, and DAX is only a
1881
1882 The default value of coredump_filter is 0x33;
1883 segments, ELF header pages and hugetlb privat
1884
1885 If you don't want to dump all shared memory s
1886 write 0x31 to the process's proc file::
1887
1888 $ echo 0x31 > /proc/1234/coredump_filter
1889
1890 When a new process is created, the process in
1891 parent. It is useful to set up coredump_filte
1892 For example::
1893
1894 $ echo 0x7 > /proc/self/coredump_filter
1895 $ ./some_program
1896
1897 3.5 /proc/<pid>/mountinfo - Information a
1898 ---------------------------------------------
1899
1900 This file contains lines of the form::
1901
1902 36 35 98:0 /mnt1 /mnt2 rw,noatime master:
1903 (1)(2)(3) (4) (5) (6) (n…m
1904
1905 (1) mount ID: unique identifier
1906 (2) parent ID: ID of parent (or o
1907 (3) major:minor: value of st_dev fo
1908 (4) root: root of the mount
1909 (5) mount point: mount point relati
1910 (6) mount options: per mount options
1911 (n…m) optional fields: zero or more fie
1912 (m+1) separator: marks the end of t
1913 (m+2) filesystem type: name of filesystem
1914 (m+3) mount source: filesystem specifi
1915 (m+4) super options: per super block op
1916
1917 Parsers should ignore all unrecognised option
1918 possible optional fields are:
1919
1920 ================ ===========================
1921 shared:X mount is shared in peer gro
1922 master:X mount is slave to peer grou
1923 propagate_from:X mount is slave and receives
1924 unbindable mount is unbindable
1925 ================ ===========================
1926
1927 .. [#] X is the closest dominant peer group u
1928 X is the immediate master of the mount
1929 group under the same root, then only t
1930 and not the "propagate_from:X" field.
1931
1932 For more information on mount propagation see
1933
1934 Documentation/filesystems/sharedsubtree.rst
1935
1936
1937 3.6 /proc/<pid>/comm & /proc/<pid>/task/
1938 ---------------------------------------------
1939 These files provide a method to access a task
1940 a task to set its own or one of its thread si
1941 is limited in size compared to the cmdline va
1942 then the kernel's TASK_COMM_LEN (currently 16
1943 terminator) will result in a truncated comm v
1944
1945
1946 3.7 /proc/<pid>/task/<tid>/children - Inf
1947 ---------------------------------------------
1948 This file provides a fast way to retrieve fir
1949 of a task pointed by <pid>/<tid> pair. The fo
1950 stream of pids.
1951
1952 Note the "first level" here -- if a child has
1953 not be listed here; one needs to read /proc/<
1954 to obtain the descendants.
1955
1956 Since this interface is intended to be fast a
1957 guarantee to provide precise results and some
1958 skipped, especially if they've exited right a
1959 pids, so one needs to either stop or freeze p
1960 if precise results are needed.
1961
1962
1963 3.8 /proc/<pid>/fdinfo/<fd> - Information
1964 ---------------------------------------------
1965 This file provides information associated wit
1966 files have at least four fields -- 'pos', 'fl
1967 The 'pos' represents the current offset of th
1968 form [see lseek(2) for details], 'flags' deno
1969 file has been created with [see open(2) for d
1970 mount ID of the file system containing the op
1971 /proc/<pid>/mountinfo for details]. 'ino' rep
1972 the file.
1973
1974 A typical output is::
1975
1976 pos: 0
1977 flags: 0100002
1978 mnt_id: 19
1979 ino: 63107
1980
1981 All locks associated with a file descriptor a
1982
1983 lock: 1: FLOCK ADVISORY WRITE 359
1984
1985 The files such as eventfd, fsnotify, signalfd
1986 pair provide additional information particula
1987
1988 Eventfd files
1989 ~~~~~~~~~~~~~
1990
1991 ::
1992
1993 pos: 0
1994 flags: 04002
1995 mnt_id: 9
1996 ino: 63107
1997 eventfd-count: 5a
1998
1999 where 'eventfd-count' is hex value of a count
2000
2001 Signalfd files
2002 ~~~~~~~~~~~~~~
2003
2004 ::
2005
2006 pos: 0
2007 flags: 04002
2008 mnt_id: 9
2009 ino: 63107
2010 sigmask: 0000000000000200
2011
2012 where 'sigmask' is hex value of the signal ma
2013 with a file.
2014
2015 Epoll files
2016 ~~~~~~~~~~~
2017
2018 ::
2019
2020 pos: 0
2021 flags: 02
2022 mnt_id: 9
2023 ino: 63107
2024 tfd: 5 events: 1d data:
2025
2026 where 'tfd' is a target file descriptor numbe
2027 'events' is events mask being watched and the
2028 associated with a target [see epoll(7) for mo
2029
2030 The 'pos' is current offset of the target fil
2031 [see lseek(2)], 'ino' and 'sdev' are inode an
2032 where target file resides, all in hex format.
2033
2034 Fsnotify files
2035 ~~~~~~~~~~~~~~
2036 For inotify files the format is the following
2037
2038 pos: 0
2039 flags: 02000000
2040 mnt_id: 9
2041 ino: 63107
2042 inotify wd:3 ino:9e7e sdev:800013 mas
2043
2044 where 'wd' is a watch descriptor in decimal f
2045 descriptor number, 'ino' and 'sdev' are inode
2046 target file resides and the 'mask' is the mas
2047 form [see inotify(7) for more details].
2048
2049 If the kernel was built with exportfs support
2050 file is encoded as a file handle. The file h
2051 fields 'fhandle-bytes', 'fhandle-type' and 'f
2052 format.
2053
2054 If the kernel is built without exportfs suppo
2055 printed out.
2056
2057 If there is no inotify mark attached yet the
2058
2059 For fanotify files the format is::
2060
2061 pos: 0
2062 flags: 02
2063 mnt_id: 9
2064 ino: 63107
2065 fanotify flags:10 event-flags:0
2066 fanotify mnt_id:12 mflags:40 mask:38
2067 fanotify ino:4f969 sdev:800013 mflags
2068
2069 where fanotify 'flags' and 'event-flags' are
2070 call, 'mnt_id' is the mount point identifier,
2071 flags associated with mark which are tracked
2072 mask. 'ino' and 'sdev' are target inode and d
2073 mask and 'ignored_mask' is the mask of events
2074 All are in hex format. Incorporation of 'mfla
2075 provide information about flags and mask used
2076 call [see fsnotify manpage for details].
2077
2078 While the first three lines are mandatory and
2079 optional and may be omitted if no marks creat
2080
2081 Timerfd files
2082 ~~~~~~~~~~~~~
2083
2084 ::
2085
2086 pos: 0
2087 flags: 02
2088 mnt_id: 9
2089 ino: 63107
2090 clockid: 0
2091 ticks: 0
2092 settime flags: 01
2093 it_value: (0, 49406829)
2094 it_interval: (1, 0)
2095
2096 where 'clockid' is the clock type and 'ticks'
2097 that have occurred [see timerfd_create(2) for
2098 flags in octal form been used to setup the ti
2099 details]. 'it_value' is remaining time until
2100 'it_interval' is the interval for the timer.
2101 with TIMER_ABSTIME option which will be shown
2102 still exhibits timer's remaining time.
2103
2104 DMA Buffer files
2105 ~~~~~~~~~~~~~~~~
2106
2107 ::
2108
2109 pos: 0
2110 flags: 04002
2111 mnt_id: 9
2112 ino: 63107
2113 size: 32768
2114 count: 2
2115 exp_name: system-heap
2116
2117 where 'size' is the size of the DMA buffer in
2118 the DMA buffer file. 'exp_name' is the name o
2119
2120 3.9 /proc/<pid>/map_files - Information a
2121 ---------------------------------------------
2122 This directory contains symbolic links which
2123 the process is maintaining. Example output::
2124
2125 | lr-------- 1 root root 64 Jan 27 11:24
2126 | lr-------- 1 root root 64 Jan 27 11:24
2127 | lr-------- 1 root root 64 Jan 27 11:24
2128 | ...
2129 | lr-------- 1 root root 64 Jan 27 11:24
2130 | lr-------- 1 root root 64 Jan 27 11:24
2131
2132 The name of a link represents the virtual mem
2133 vm_area_struct::vm_start-vm_area_struct::vm_e
2134
2135 The main purpose of the map_files is to retri
2136 files in a fast way instead of parsing /proc/
2137 /proc/<pid>/smaps, both of which contain many
2138 time one can open(2) mappings from the listin
2139 comparing their inode numbers to figure out w
2140 are actually shared.
2141
2142 3.10 /proc/<pid>/timerslack_ns - Task time
2143 ---------------------------------------------
2144 This file provides the value of the task's ti
2145 This value specifies an amount of time that n
2146 in order to coalesce timers and avoid unneces
2147
2148 This allows a task's interactivity vs power c
2149 adjusted.
2150
2151 Writing 0 to the file will set the task's tim
2152
2153 Valid values are from 0 - ULLONG_MAX
2154
2155 An application setting the value must have PT
2156 permissions on the task specified to change i
2157
2158 3.11 /proc/<pid>/patch_state - Livepatch p
2159 ---------------------------------------------
2160 When CONFIG_LIVEPATCH is enabled, this file d
2161 patch state for the task.
2162
2163 A value of '-1' indicates that no patch is in
2164
2165 A value of '0' indicates that a patch is in t
2166 unpatched. If the patch is being enabled, th
2167 patched yet. If the patch is being disabled,
2168 been unpatched.
2169
2170 A value of '1' indicates that a patch is in t
2171 patched. If the patch is being enabled, then
2172 patched. If the patch is being disabled, the
2173 unpatched yet.
2174
2175 3.12 /proc/<pid>/arch_status - task architect
2176 ---------------------------------------------
2177 When CONFIG_PROC_PID_ARCH_STATUS is enabled,
2178 architecture specific status of the task.
2179
2180 Example
2181 ~~~~~~~
2182
2183 ::
2184
2185 $ cat /proc/6753/arch_status
2186 AVX512_elapsed_ms: 8
2187
2188 Description
2189 ~~~~~~~~~~~
2190
2191 x86 specific entries
2192 ~~~~~~~~~~~~~~~~~~~~~
2193
2194 AVX512_elapsed_ms
2195 ^^^^^^^^^^^^^^^^^^
2196
2197 If AVX512 is supported on the machine, this
2198 elapsed since the last time AVX512 usage wa
2199 happens on a best effort basis when a task
2200 that the value depends on two factors:
2201
2202 1) The time which the task spent on the C
2203 out. With CPU isolation and a single r
2204 several seconds.
2205
2206 2) The time since the task was scheduled
2207 reason for being scheduled out (time s
2208 this can be arbitrary long time.
2209
2210 As a consequence the value cannot be consid
2211 information. The application which uses thi
2212 of the overall scenario on the system in or
2213 task is a real AVX512 user or not. Precise
2214 with performance counters.
2215
2216 A special value of '-1' indicates that no A
2217 the task is unlikely an AVX512 user, but de
2218 scheduling scenario, it also could be a fal
2219
2220 3.13 /proc/<pid>/fd - List of symlinks to ope
2221 ---------------------------------------------
2222 This directory contains symbolic links which
2223 the process is maintaining. Example output::
2224
2225 lr-x------ 1 root root 64 Sep 20 17:53 0 ->
2226 l-wx------ 1 root root 64 Sep 20 17:53 1 ->
2227 lrwx------ 1 root root 64 Sep 20 17:53 10 -
2228 lrwx------ 1 root root 64 Sep 20 17:53 11 -
2229 lrwx------ 1 root root 64 Sep 20 17:53 12 -
2230
2231 The number of open files for the process is s
2232 of stat() output for /proc/<pid>/fd for fast
2233 ---------------------------------------------
2234
2235
2236 Chapter 4: Configuring procfs
2237 =============================
2238
2239 4.1 Mount options
2240 ---------------------
2241
2242 The following mount options are supported:
2243
2244 ========= =====================
2245 hidepid= Set /proc/<pid>/ acce
2246 gid= Set the group authori
2247 subset= Show only the specifi
2248 ========= =====================
2249
2250 hidepid=off or hidepid=0 means classic mode -
2251 /proc/<pid>/ directories (default).
2252
2253 hidepid=noaccess or hidepid=1 means users may
2254 directories but their own. Sensitive files l
2255 protected against other users. This makes it
2256 user runs specific program (given the program
2257 behaviour). As an additional bonus, as /proc
2258 other users, poorly written programs passing
2259 arguments are now protected against local eav
2260
2261 hidepid=invisible or hidepid=2 means hidepid=
2262 fully invisible to other users. It doesn't m
2263 process with a specific pid value exists (it
2264 by "kill -0 $PID"), but it hides process' uid
2265 stat()'ing /proc/<pid>/ otherwise. It greatl
2266 gathering information about running processes
2267 elevated privileges, whether other user runs
2268 other users run any program at all, etc.
2269
2270 hidepid=ptraceable or hidepid=4 means that pr
2271 /proc/<pid>/ directories that the caller can
2272
2273 gid= defines a group authorized to learn proc
2274 prohibited by hidepid=. If you use some daem
2275 information about processes information, just
2276
2277 subset=pid hides all top level files and dire
2278 are not related to tasks.
2279
2280 Chapter 5: Filesystem behavior
2281 ==============================
2282
2283 Originally, before the advent of pid namespac
2284 system. It means that there was only one proc
2285
2286 When pid namespace was added, a separate proc
2287 each pid namespace. So, procfs mount options
2288 mountpoints within the same namespace::
2289
2290 # grep ^proc /proc/mounts
2291 proc /proc proc rw,relatime,hidepid=2
2292
2293 # strace -e mount mount -o hidepid=1
2294 mount("proc", "/tmp/proc", "proc", 0,
2295 +++ exited with 0 +++
2296
2297 # grep ^proc /proc/mounts
2298 proc /proc proc rw,relatime,hidepid=2
2299 proc /tmp/proc proc rw,relatime,hidep
2300
2301 and only after remounting procfs mount option
2302 mountpoints::
2303
2304 # mount -o remount,hidepid=1 -t proc
2305
2306 # grep ^proc /proc/mounts
2307 proc /proc proc rw,relatime,hidepid=1
2308 proc /tmp/proc proc rw,relatime,hidep
2309
2310 This behavior is different from the behavior
2311
2312 The new procfs behavior is more like other fi
2313 creates a new procfs instance. Mount options
2314 It means that it became possible to have seve
2315 displaying tasks with different filtering opt
2316
2317 # mount -o hidepid=invisible -t proc
2318 # mount -o hidepid=noaccess -t proc p
2319 # grep ^proc /proc/mounts
2320 proc /proc proc rw,relatime,hidepid=i
2321 proc /tmp/proc proc rw,relatime,hidep
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.