· 6 years ago · Sep 02, 2019, 02:18 AM
1
2------------------------------------------------------------------------------
3 T H E /proc F I L E S Y S T E M
4------------------------------------------------------------------------------
5/proc/sys Terrehon Bowden <terrehon@pacbell.net> October 7 1999
6 Bodo Bauer <bb@ricochet.net>
7
82.4.x update Jorge Nerin <comandante@zaralinux.com> November 14 2000
9move /proc/sys Shen Feng <shen@cn.fujitsu.com> April 1 2009
10------------------------------------------------------------------------------
11Version 1.3 Kernel version 2.2.12
12 Kernel version 2.4.0-test11-pre4
13------------------------------------------------------------------------------
14fixes/update part 1.1 Stefani Seibold <stefani@seibold.net> June 9 2009
15
16Table of Contents
17-----------------
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
51 4 Configuring procfs
52 4.1 Mount options
53
54------------------------------------------------------------------------------
55Preface
56------------------------------------------------------------------------------
57
580.1 Introduction/Credits
59------------------------
60
61This documentation is part of a soon (or so we hope) to be released book on
62the SuSE Linux distribution. As there is no complete documentation for the
63/proc file system and we've used many freely available sources to write these
64chapters, it seems only fair to give the work back to the Linux community.
65This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm
66afraid it's still far from complete, but we hope it will be useful. As far as
67we know, it is the first 'all-in-one' document about the /proc file system. It
68is focused on the Intel x86 hardware, so if you are looking for PPC, ARM,
69SPARC, AXP, etc., features, you probably won't find what you are looking for.
70It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But
71additions and patches are welcome and will be added to this document if you
72mail them to Bodo.
73
74We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of
75other people for help compiling this documentation. We'd also like to extend a
76special thank you to Andi Kleen for documentation, which we relied on heavily
77to create this document, as well as the additional information he provided.
78Thanks to everybody else who contributed source or docs to the Linux kernel
79and helped create a great piece of software... :)
80
81If you have any comments, corrections or additions, please don't hesitate to
82contact Bodo Bauer at bb@ricochet.net. We'll be happy to add them to this
83document.
84
85The latest version of this document is available online at
86http://tldp.org/LDP/Linux-Filesystem-Hierarchy/html/proc.html
87
88If the above direction does not works for you, you could try the kernel
89mailing list at linux-kernel@vger.kernel.org and/or try to reach me at
90comandante@zaralinux.com.
91
920.2 Legal Stuff
93---------------
94
95We don't guarantee the correctness of this document, and if you come to us
96complaining about how you screwed up your system because of incorrect
97documentation, we won't feel responsible...
98
99------------------------------------------------------------------------------
100CHAPTER 1: COLLECTING SYSTEM INFORMATION
101------------------------------------------------------------------------------
102
103------------------------------------------------------------------------------
104In This Chapter
105------------------------------------------------------------------------------
106* Investigating the properties of the pseudo file system /proc and its
107 ability to provide information on the running Linux system
108* Examining /proc's structure
109* Uncovering various information about the kernel and the processes running
110 on the system
111------------------------------------------------------------------------------
112
113
114The proc file system acts as an interface to internal data structures in the
115kernel. It can be used to obtain information about the system and to change
116certain kernel parameters at runtime (sysctl).
117
118First, we'll take a look at the read-only parts of /proc. In Chapter 2, we
119show you how you can use /proc/sys to change settings.
120
1211.1 Process-Specific Subdirectories
122-----------------------------------
123
124The directory /proc contains (among other things) one subdirectory for each
125process running on the system, which is named after the process ID (PID).
126
127The link self points to the process reading the file system. Each process
128subdirectory has the entries listed in Table 1-1.
129
130Note that an open a file descriptor to /proc/<pid> or to any of its
131contained files or subdirectories does not prevent <pid> being reused
132for some other process in the event that <pid> exits. Operations on
133open /proc/<pid> file descriptors corresponding to dead processes
134never act on any new process that the kernel may, through chance, have
135also assigned the process ID <pid>. Instead, operations on these FDs
136usually fail with ESRCH.
137
138Table 1-1: Process specific entries in /proc
139..............................................................................
140 File Content
141 clear_refs Clears page referenced bits shown in smaps output
142 cmdline Command line arguments
143 cpu Current and last cpu in which it was executed (2.4)(smp)
144 cwd Link to the current working directory
145 environ Values of environment variables
146 exe Link to the executable of this process
147 fd Directory, which contains all file descriptors
148 maps Memory maps to executables and library files (2.4)
149 mem Memory held by this process
150 root Link to the root directory of this process
151 stat Process status
152 statm Process memory status information
153 status Process status in human readable form
154 wchan Present with CONFIG_KALLSYMS=y: it shows the kernel function
155 symbol the task is blocked in - or "0" if not blocked.
156 pagemap Page table
157 stack Report full stack trace, enable via CONFIG_STACKTRACE
158 smaps An extension based on maps, showing the memory consumption of
159 each mapping and flags associated with it
160 smaps_rollup Accumulated smaps stats for all mappings of the process. This
161 can be derived from smaps, but is faster and more convenient
162 numa_maps An extension based on maps, showing the memory locality and
163 binding policy as well as mem usage (in pages) of each mapping.
164..............................................................................
165
166For example, to get the status information of a process, all you have to do is
167read the file /proc/PID/status:
168
169 >cat /proc/self/status
170 Name: cat
171 State: R (running)
172 Tgid: 5452
173 Pid: 5452
174 PPid: 743
175 TracerPid: 0 (2.4)
176 Uid: 501 501 501 501
177 Gid: 100 100 100 100
178 FDSize: 256
179 Groups: 100 14 16
180 VmPeak: 5004 kB
181 VmSize: 5004 kB
182 VmLck: 0 kB
183 VmHWM: 476 kB
184 VmRSS: 476 kB
185 RssAnon: 352 kB
186 RssFile: 120 kB
187 RssShmem: 4 kB
188 VmData: 156 kB
189 VmStk: 88 kB
190 VmExe: 68 kB
191 VmLib: 1412 kB
192 VmPTE: 20 kb
193 VmSwap: 0 kB
194 HugetlbPages: 0 kB
195 CoreDumping: 0
196 THP_enabled: 1
197 Threads: 1
198 SigQ: 0/28578
199 SigPnd: 0000000000000000
200 ShdPnd: 0000000000000000
201 SigBlk: 0000000000000000
202 SigIgn: 0000000000000000
203 SigCgt: 0000000000000000
204 CapInh: 00000000fffffeff
205 CapPrm: 0000000000000000
206 CapEff: 0000000000000000
207 CapBnd: ffffffffffffffff
208 CapAmb: 0000000000000000
209 NoNewPrivs: 0
210 Seccomp: 0
211 Speculation_Store_Bypass: thread vulnerable
212 voluntary_ctxt_switches: 0
213 nonvoluntary_ctxt_switches: 1
214
215This shows you nearly the same information you would get if you viewed it with
216the ps command. In fact, ps uses the proc file system to obtain its
217information. But you get a more detailed view of the process by reading the
218file /proc/PID/status. It fields are described in table 1-2.
219
220The statm file contains more detailed information about the process
221memory usage. Its seven fields are explained in Table 1-3. The stat file
222contains details information about the process itself. Its fields are
223explained in Table 1-4.
224
225(for SMP CONFIG users)
226For making accounting scalable, RSS related information are handled in an
227asynchronous manner and the value may not be very precise. To see a precise
228snapshot of a moment, you can see /proc/<pid>/smaps file and scan page table.
229It's slow but very precise.
230
231Table 1-2: Contents of the status files (as of 4.19)
232..............................................................................
233 Field Content
234 Name filename of the executable
235 Umask file mode creation mask
236 State state (R is running, S is sleeping, D is sleeping
237 in an uninterruptible wait, Z is zombie,
238 T is traced or stopped)
239 Tgid thread group ID
240 Ngid NUMA group ID (0 if none)
241 Pid process id
242 PPid process id of the parent process
243 TracerPid PID of process tracing this process (0 if not)
244 Uid Real, effective, saved set, and file system UIDs
245 Gid Real, effective, saved set, and file system GIDs
246 FDSize number of file descriptor slots currently allocated
247 Groups supplementary group list
248 NStgid descendant namespace thread group ID hierarchy
249 NSpid descendant namespace process ID hierarchy
250 NSpgid descendant namespace process group ID hierarchy
251 NSsid descendant namespace session ID hierarchy
252 VmPeak peak virtual memory size
253 VmSize total program size
254 VmLck locked memory size
255 VmPin pinned memory size
256 VmHWM peak resident set size ("high water mark")
257 VmRSS size of memory portions. It contains the three
258 following parts (VmRSS = RssAnon + RssFile + RssShmem)
259 RssAnon size of resident anonymous memory
260 RssFile size of resident file mappings
261 RssShmem size of resident shmem memory (includes SysV shm,
262 mapping of tmpfs and shared anonymous mappings)
263 VmData size of private data segments
264 VmStk size of stack segments
265 VmExe size of text segment
266 VmLib size of shared library code
267 VmPTE size of page table entries
268 VmSwap amount of swap used by anonymous private data
269 (shmem swap usage is not included)
270 HugetlbPages size of hugetlb memory portions
271 CoreDumping process's memory is currently being dumped
272 (killing the process may lead to a corrupted core)
273 THP_enabled process is allowed to use THP (returns 0 when
274 PR_SET_THP_DISABLE is set on the process
275 Threads number of threads
276 SigQ number of signals queued/max. number for queue
277 SigPnd bitmap of pending signals for the thread
278 ShdPnd bitmap of shared pending signals for the process
279 SigBlk bitmap of blocked signals
280 SigIgn bitmap of ignored signals
281 SigCgt bitmap of caught signals
282 CapInh bitmap of inheritable capabilities
283 CapPrm bitmap of permitted capabilities
284 CapEff bitmap of effective capabilities
285 CapBnd bitmap of capabilities bounding set
286 CapAmb bitmap of ambient capabilities
287 NoNewPrivs no_new_privs, like prctl(PR_GET_NO_NEW_PRIV, ...)
288 Seccomp seccomp mode, like prctl(PR_GET_SECCOMP, ...)
289 Speculation_Store_Bypass speculative store bypass mitigation status
290 Cpus_allowed mask of CPUs on which this process may run
291 Cpus_allowed_list Same as previous, but in "list format"
292 Mems_allowed mask of memory nodes allowed to this process
293 Mems_allowed_list Same as previous, but in "list format"
294 voluntary_ctxt_switches number of voluntary context switches
295 nonvoluntary_ctxt_switches number of non voluntary context switches
296..............................................................................
297
298Table 1-3: Contents of the statm files (as of 2.6.8-rc3)
299..............................................................................
300 Field Content
301 size total program size (pages) (same as VmSize in status)
302 resident size of memory portions (pages) (same as VmRSS in status)
303 shared number of pages that are shared (i.e. backed by a file, same
304 as RssFile+RssShmem in status)
305 trs number of pages that are 'code' (not including libs; broken,
306 includes data segment)
307 lrs number of pages of library (always 0 on 2.6)
308 drs number of pages of data/stack (including libs; broken,
309 includes library text)
310 dt number of dirty pages (always 0 on 2.6)
311..............................................................................
312
313
314Table 1-4: Contents of the stat files (as of 2.6.30-rc7)
315..............................................................................
316 Field Content
317 pid process id
318 tcomm filename of the executable
319 state state (R is running, S is sleeping, D is sleeping in an
320 uninterruptible wait, Z is zombie, T is traced or stopped)
321 ppid process id of the parent process
322 pgrp pgrp of the process
323 sid session id
324 tty_nr tty the process uses
325 tty_pgrp pgrp of the tty
326 flags task flags
327 min_flt number of minor faults
328 cmin_flt number of minor faults with child's
329 maj_flt number of major faults
330 cmaj_flt number of major faults with child's
331 utime user mode jiffies
332 stime kernel mode jiffies
333 cutime user mode jiffies with child's
334 cstime kernel mode jiffies with child's
335 priority priority level
336 nice nice level
337 num_threads number of threads
338 it_real_value (obsolete, always 0)
339 start_time time the process started after system boot
340 vsize virtual memory size
341 rss resident set memory size
342 rsslim current limit in bytes on the rss
343 start_code address above which program text can run
344 end_code address below which program text can run
345 start_stack address of the start of the main process stack
346 esp current value of ESP
347 eip current value of EIP
348 pending bitmap of pending signals
349 blocked bitmap of blocked signals
350 sigign bitmap of ignored signals
351 sigcatch bitmap of caught signals
352 0 (place holder, used to be the wchan address, use /proc/PID/wchan instead)
353 0 (place holder)
354 0 (place holder)
355 exit_signal signal to send to parent thread on exit
356 task_cpu which CPU the task is scheduled on
357 rt_priority realtime priority
358 policy scheduling policy (man sched_setscheduler)
359 blkio_ticks time spent waiting for block IO
360 gtime guest time of the task in jiffies
361 cgtime guest time of the task children in jiffies
362 start_data address above which program data+bss is placed
363 end_data address below which program data+bss is placed
364 start_brk address above which program heap can be expanded with brk()
365 arg_start address above which program command line is placed
366 arg_end address below which program command line is placed
367 env_start address above which program environment is placed
368 env_end address below which program environment is placed
369 exit_code the thread's exit_code in the form reported by the waitpid system call
370..............................................................................
371
372The /proc/PID/maps file contains the currently mapped memory regions and
373their access permissions.
374
375The format is:
376
377address perms offset dev inode pathname
378
37908048000-08049000 r-xp 00000000 03:00 8312 /opt/test
38008049000-0804a000 rw-p 00001000 03:00 8312 /opt/test
3810804a000-0806b000 rw-p 00000000 00:00 0 [heap]
382a7cb1000-a7cb2000 ---p 00000000 00:00 0
383a7cb2000-a7eb2000 rw-p 00000000 00:00 0
384a7eb2000-a7eb3000 ---p 00000000 00:00 0
385a7eb3000-a7ed5000 rw-p 00000000 00:00 0
386a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6
387a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6
388a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6
389a800b000-a800e000 rw-p 00000000 00:00 0
390a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0
391a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0
392a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0
393a8024000-a8027000 rw-p 00000000 00:00 0
394a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2
395a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2
396a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2
397aff35000-aff4a000 rw-p 00000000 00:00 0 [stack]
398ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso]
399
400where "address" is the address space in the process that it occupies, "perms"
401is a set of permissions:
402
403 r = read
404 w = write
405 x = execute
406 s = shared
407 p = private (copy on write)
408
409"offset" is the offset into the mapping, "dev" is the device (major:minor), and
410"inode" is the inode on that device. 0 indicates that no inode is associated
411with the memory region, as the case would be with BSS (uninitialized data).
412The "pathname" shows the name associated file for this mapping. If the mapping
413is not associated with a file:
414
415 [heap] = the heap of the program
416 [stack] = the stack of the main process
417 [vdso] = the "virtual dynamic shared object",
418 the kernel system call handler
419
420 or if empty, the mapping is anonymous.
421
422The /proc/PID/smaps is an extension based on maps, showing the memory
423consumption for each of the process's mappings. For each mapping (aka Virtual
424Memory Area, or VMA) there is a series of lines such as the following:
425
42608048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash
427
428Size: 1084 kB
429KernelPageSize: 4 kB
430MMUPageSize: 4 kB
431Rss: 892 kB
432Pss: 374 kB
433Shared_Clean: 892 kB
434Shared_Dirty: 0 kB
435Private_Clean: 0 kB
436Private_Dirty: 0 kB
437Referenced: 892 kB
438Anonymous: 0 kB
439LazyFree: 0 kB
440AnonHugePages: 0 kB
441ShmemPmdMapped: 0 kB
442Shared_Hugetlb: 0 kB
443Private_Hugetlb: 0 kB
444Swap: 0 kB
445SwapPss: 0 kB
446KernelPageSize: 4 kB
447MMUPageSize: 4 kB
448Locked: 0 kB
449THPeligible: 0
450VmFlags: rd ex mr mw me dw
451
452The first of these lines shows the same information as is displayed for the
453mapping in /proc/PID/maps. Following lines show the size of the mapping
454(size); the size of each page allocated when backing a VMA (KernelPageSize),
455which is usually the same as the size in the page table entries; the page size
456used by the MMU when backing a VMA (in most cases, the same as KernelPageSize);
457the amount of the mapping that is currently resident in RAM (RSS); the
458process' proportional share of this mapping (PSS); and the number of clean and
459dirty shared and private pages in the mapping.
460
461The "proportional set size" (PSS) of a process is the count of pages it has
462in memory, where each page is divided by the number of processes sharing it.
463So if a process has 1000 pages all to itself, and 1000 shared with one other
464process, its PSS will be 1500.
465Note that even a page which is part of a MAP_SHARED mapping, but has only
466a single pte mapped, i.e. is currently used by only one process, is accounted
467as private and not as shared.
468"Referenced" indicates the amount of memory currently marked as referenced or
469accessed.
470"Anonymous" shows the amount of memory that does not belong to any file. Even
471a mapping associated with a file may contain anonymous pages: when MAP_PRIVATE
472and a page is modified, the file page is replaced by a private anonymous copy.
473"LazyFree" shows the amount of memory which is marked by madvise(MADV_FREE).
474The memory isn't freed immediately with madvise(). It's freed in memory
475pressure if the memory is clean. Please note that the printed value might
476be lower than the real value due to optimizations used in the current
477implementation. If this is not desirable please file a bug report.
478"AnonHugePages" shows the ammount of memory backed by transparent hugepage.
479"ShmemPmdMapped" shows the ammount of shared (shmem/tmpfs) memory backed by
480huge pages.
481"Shared_Hugetlb" and "Private_Hugetlb" show the ammounts of memory backed by
482hugetlbfs page which is *not* counted in "RSS" or "PSS" field for historical
483reasons. And these are not included in {Shared,Private}_{Clean,Dirty} field.
484"Swap" shows how much would-be-anonymous memory is also used, but out on swap.
485For shmem mappings, "Swap" includes also the size of the mapped (and not
486replaced by copy-on-write) part of the underlying shmem object out on swap.
487"SwapPss" shows proportional swap share of this mapping. Unlike "Swap", this
488does not take into account swapped out page of underlying shmem objects.
489"Locked" indicates whether the mapping is locked in memory or not.
490"THPeligible" indicates whether the mapping is eligible for allocating THP
491pages - 1 if true, 0 otherwise. It just shows the current status.
492
493"VmFlags" field deserves a separate description. This member represents the kernel
494flags associated with the particular virtual memory area in two letter encoded
495manner. The codes are the following:
496 rd - readable
497 wr - writeable
498 ex - executable
499 sh - shared
500 mr - may read
501 mw - may write
502 me - may execute
503 ms - may share
504 gd - stack segment growns down
505 pf - pure PFN range
506 dw - disabled write to the mapped file
507 lo - pages are locked in memory
508 io - memory mapped I/O area
509 sr - sequential read advise provided
510 rr - random read advise provided
511 dc - do not copy area on fork
512 de - do not expand area on remapping
513 ac - area is accountable
514 nr - swap space is not reserved for the area
515 ht - area uses huge tlb pages
516 ar - architecture specific flag
517 dd - do not include area into core dump
518 sd - soft-dirty flag
519 mm - mixed map area
520 hg - huge page advise flag
521 nh - no-huge page advise flag
522 mg - mergable advise flag
523
524Note that there is no guarantee that every flag and associated mnemonic will
525be present in all further kernel releases. Things get changed, the flags may
526be vanished or the reverse -- new added. Interpretation of their meaning
527might change in future as well. So each consumer of these flags has to
528follow each specific kernel version for the exact semantic.
529
530This file is only present if the CONFIG_MMU kernel configuration option is
531enabled.
532
533Note: reading /proc/PID/maps or /proc/PID/smaps is inherently racy (consistent
534output can be achieved only in the single read call).
535This typically manifests when doing partial reads of these files while the
536memory map is being modified. Despite the races, we do provide the following
537guarantees:
538
5391) The mapped addresses never go backwards, which implies no two
540 regions will ever overlap.
5412) If there is something at a given vaddr during the entirety of the
542 life of the smaps/maps walk, there will be some output for it.
543
544The /proc/PID/smaps_rollup file includes the same fields as /proc/PID/smaps,
545but their values are the sums of the corresponding values for all mappings of
546the process. Additionally, it contains these fields:
547
548Pss_Anon
549Pss_File
550Pss_Shmem
551
552They represent the proportional shares of anonymous, file, and shmem pages, as
553described for smaps above. These fields are omitted in smaps since each
554mapping identifies the type (anon, file, or shmem) of all pages it contains.
555Thus all information in smaps_rollup can be derived from smaps, but at a
556significantly higher cost.
557
558The /proc/PID/clear_refs is used to reset the PG_Referenced and ACCESSED/YOUNG
559bits on both physical and virtual pages associated with a process, and the
560soft-dirty bit on pte (see Documentation/admin-guide/mm/soft-dirty.rst
561for details).
562To clear the bits for all the pages associated with the process
563 > echo 1 > /proc/PID/clear_refs
564
565To clear the bits for the anonymous pages associated with the process
566 > echo 2 > /proc/PID/clear_refs
567
568To clear the bits for the file mapped pages associated with the process
569 > echo 3 > /proc/PID/clear_refs
570
571To clear the soft-dirty bit
572 > echo 4 > /proc/PID/clear_refs
573
574To reset the peak resident set size ("high water mark") to the process's
575current value:
576 > echo 5 > /proc/PID/clear_refs
577
578Any other value written to /proc/PID/clear_refs will have no effect.
579
580The /proc/pid/pagemap gives the PFN, which can be used to find the pageflags
581using /proc/kpageflags and number of times a page is mapped using
582/proc/kpagecount. For detailed explanation, see
583Documentation/admin-guide/mm/pagemap.rst.
584
585The /proc/pid/numa_maps is an extension based on maps, showing the memory
586locality and binding policy, as well as the memory usage (in pages) of
587each mapping. The output follows a general format where mapping details get
588summarized separated by blank spaces, one mapping per each file line:
589
590address policy mapping details
591
59200400000 default file=/usr/local/bin/app mapped=1 active=0 N3=1 kernelpagesize_kB=4
59300600000 default file=/usr/local/bin/app anon=1 dirty=1 N3=1 kernelpagesize_kB=4
5943206000000 default file=/lib64/ld-2.12.so mapped=26 mapmax=6 N0=24 N3=2 kernelpagesize_kB=4
595320621f000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4
5963206220000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4
5973206221000 default anon=1 dirty=1 N3=1 kernelpagesize_kB=4
5983206800000 default file=/lib64/libc-2.12.so mapped=59 mapmax=21 active=55 N0=41 N3=18 kernelpagesize_kB=4
599320698b000 default file=/lib64/libc-2.12.so
6003206b8a000 default file=/lib64/libc-2.12.so anon=2 dirty=2 N3=2 kernelpagesize_kB=4
6013206b8e000 default file=/lib64/libc-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4
6023206b8f000 default anon=3 dirty=3 active=1 N3=3 kernelpagesize_kB=4
6037f4dc10a2000 default anon=3 dirty=3 N3=3 kernelpagesize_kB=4
6047f4dc10b4000 default anon=2 dirty=2 active=1 N3=2 kernelpagesize_kB=4
6057f4dc1200000 default file=/anon_hugepage\040(deleted) huge anon=1 dirty=1 N3=1 kernelpagesize_kB=2048
6067fff335f0000 default stack anon=3 dirty=3 N3=3 kernelpagesize_kB=4
6077fff3369d000 default mapped=1 mapmax=35 active=0 N3=1 kernelpagesize_kB=4
608
609Where:
610"address" is the starting address for the mapping;
611"policy" reports the NUMA memory policy set for the mapping (see Documentation/admin-guide/mm/numa_memory_policy.rst);
612"mapping details" summarizes mapping data such as mapping type, page usage counters,
613node locality page counters (N0 == node0, N1 == node1, ...) and the kernel page
614size, in KB, that is backing the mapping up.
615
6161.2 Kernel data
617---------------
618
619Similar to the process entries, the kernel data files give information about
620the running kernel. The files used to obtain this information are contained in
621/proc and are listed in Table 1-5. Not all of these will be present in your
622system. It depends on the kernel configuration and the loaded modules, which
623files are there, and which are missing.
624
625Table 1-5: Kernel info in /proc
626..............................................................................
627 File Content
628 apm Advanced power management info
629 buddyinfo Kernel memory allocator information (see text) (2.5)
630 bus Directory containing bus specific information
631 cmdline Kernel command line
632 cpuinfo Info about the CPU
633 devices Available devices (block and character)
634 dma Used DMS channels
635 filesystems Supported filesystems
636 driver Various drivers grouped here, currently rtc (2.4)
637 execdomains Execdomains, related to security (2.4)
638 fb Frame Buffer devices (2.4)
639 fs File system parameters, currently nfs/exports (2.4)
640 ide Directory containing info about the IDE subsystem
641 interrupts Interrupt usage
642 iomem Memory map (2.4)
643 ioports I/O port usage
644 irq Masks for irq to cpu affinity (2.4)(smp?)
645 isapnp ISA PnP (Plug&Play) Info (2.4)
646 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4))
647 kmsg Kernel messages
648 ksyms Kernel symbol table
649 loadavg Load average of last 1, 5 & 15 minutes
650 locks Kernel locks
651 meminfo Memory info
652 misc Miscellaneous
653 modules List of loaded modules
654 mounts Mounted filesystems
655 net Networking info (see text)
656 pagetypeinfo Additional page allocator information (see text) (2.5)
657 partitions Table of partitions known to the system
658 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/,
659 decoupled by lspci (2.4)
660 rtc Real time clock
661 scsi SCSI info (see text)
662 slabinfo Slab pool info
663 softirqs softirq usage
664 stat Overall statistics
665 swaps Swap space utilization
666 sys See chapter 2
667 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4)
668 tty Info of tty drivers
669 uptime Wall clock since boot, combined idle time of all cpus
670 version Kernel version
671 video bttv info of video resources (2.4)
672 vmallocinfo Show vmalloced areas
673..............................................................................
674
675You can, for example, check which interrupts are currently in use and what
676they are used for by looking in the file /proc/interrupts:
677
678 > cat /proc/interrupts
679 CPU0
680 0: 8728810 XT-PIC timer
681 1: 895 XT-PIC keyboard
682 2: 0 XT-PIC cascade
683 3: 531695 XT-PIC aha152x
684 4: 2014133 XT-PIC serial
685 5: 44401 XT-PIC pcnet_cs
686 8: 2 XT-PIC rtc
687 11: 8 XT-PIC i82365
688 12: 182918 XT-PIC PS/2 Mouse
689 13: 1 XT-PIC fpu
690 14: 1232265 XT-PIC ide0
691 15: 7 XT-PIC ide1
692 NMI: 0
693
694In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the
695output of a SMP machine):
696
697 > cat /proc/interrupts
698
699 CPU0 CPU1
700 0: 1243498 1214548 IO-APIC-edge timer
701 1: 8949 8958 IO-APIC-edge keyboard
702 2: 0 0 XT-PIC cascade
703 5: 11286 10161 IO-APIC-edge soundblaster
704 8: 1 0 IO-APIC-edge rtc
705 9: 27422 27407 IO-APIC-edge 3c503
706 12: 113645 113873 IO-APIC-edge PS/2 Mouse
707 13: 0 0 XT-PIC fpu
708 14: 22491 24012 IO-APIC-edge ide0
709 15: 2183 2415 IO-APIC-edge ide1
710 17: 30564 30414 IO-APIC-level eth0
711 18: 177 164 IO-APIC-level bttv
712 NMI: 2457961 2457959
713 LOC: 2457882 2457881
714 ERR: 2155
715
716NMI is incremented in this case because every timer interrupt generates a NMI
717(Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups.
718
719LOC is the local interrupt counter of the internal APIC of every CPU.
720
721ERR is incremented in the case of errors in the IO-APIC bus (the bus that
722connects the CPUs in a SMP system. This means that an error has been detected,
723the IO-APIC automatically retry the transmission, so it should not be a big
724problem, but you should read the SMP-FAQ.
725
726In 2.6.2* /proc/interrupts was expanded again. This time the goal was for
727/proc/interrupts to display every IRQ vector in use by the system, not
728just those considered 'most important'. The new vectors are:
729
730 THR -- interrupt raised when a machine check threshold counter
731 (typically counting ECC corrected errors of memory or cache) exceeds
732 a configurable threshold. Only available on some systems.
733
734 TRM -- a thermal event interrupt occurs when a temperature threshold
735 has been exceeded for the CPU. This interrupt may also be generated
736 when the temperature drops back to normal.
737
738 SPU -- a spurious interrupt is some interrupt that was raised then lowered
739 by some IO device before it could be fully processed by the APIC. Hence
740 the APIC sees the interrupt but does not know what device it came from.
741 For this case the APIC will generate the interrupt with a IRQ vector
742 of 0xff. This might also be generated by chipset bugs.
743
744 RES, CAL, TLB -- rescheduling, call and TLB flush interrupts are
745 sent from one CPU to another per the needs of the OS. Typically,
746 their statistics are used by kernel developers and interested users to
747 determine the occurrence of interrupts of the given type.
748
749The above IRQ vectors are displayed only when relevant. For example,
750the threshold vector does not exist on x86_64 platforms. Others are
751suppressed when the system is a uniprocessor. As of this writing, only
752i386 and x86_64 platforms support the new IRQ vector displays.
753
754Of some interest is the introduction of the /proc/irq directory to 2.4.
755It could be used to set IRQ to CPU affinity, this means that you can "hook" an
756IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the
757irq subdir is one subdir for each IRQ, and two files; default_smp_affinity and
758prof_cpu_mask.
759
760For example
761 > ls /proc/irq/
762 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask
763 1 11 13 15 17 19 3 5 7 9 default_smp_affinity
764 > ls /proc/irq/0/
765 smp_affinity
766
767smp_affinity is a bitmask, in which you can specify which CPUs can handle the
768IRQ, you can set it by doing:
769
770 > echo 1 > /proc/irq/10/smp_affinity
771
772This means that only the first CPU will handle the IRQ, but you can also echo
7735 which means that only the first and third CPU can handle the IRQ.
774
775The contents of each smp_affinity file is the same by default:
776
777 > cat /proc/irq/0/smp_affinity
778 ffffffff
779
780There is an alternate interface, smp_affinity_list which allows specifying
781a cpu range instead of a bitmask:
782
783 > cat /proc/irq/0/smp_affinity_list
784 1024-1031
785
786The default_smp_affinity mask applies to all non-active IRQs, which are the
787IRQs which have not yet been allocated/activated, and hence which lack a
788/proc/irq/[0-9]* directory.
789
790The node file on an SMP system shows the node to which the device using the IRQ
791reports itself as being attached. This hardware locality information does not
792include information about any possible driver locality preference.
793
794prof_cpu_mask specifies which CPUs are to be profiled by the system wide
795profiler. Default value is ffffffff (all cpus if there are only 32 of them).
796
797The way IRQs are routed is handled by the IO-APIC, and it's Round Robin
798between all the CPUs which are allowed to handle it. As usual the kernel has
799more info than you and does a better job than you, so the defaults are the
800best choice for almost everyone. [Note this applies only to those IO-APIC's
801that support "Round Robin" interrupt distribution.]
802
803There are three more important subdirectories in /proc: net, scsi, and sys.
804The general rule is that the contents, or even the existence of these
805directories, depend on your kernel configuration. If SCSI is not enabled, the
806directory scsi may not exist. The same is true with the net, which is there
807only when networking support is present in the running kernel.
808
809The slabinfo file gives information about memory usage at the slab level.
810Linux uses slab pools for memory management above page level in version 2.2.
811Commonly used objects have their own slab pool (such as network buffers,
812directory cache, and so on).
813
814..............................................................................
815
816> cat /proc/buddyinfo
817
818Node 0, zone DMA 0 4 5 4 4 3 ...
819Node 0, zone Normal 1 0 0 1 101 8 ...
820Node 0, zone HighMem 2 0 0 1 1 0 ...
821
822External fragmentation is a problem under some workloads, and buddyinfo is a
823useful tool for helping diagnose these problems. Buddyinfo will give you a
824clue as to how big an area you can safely allocate, or why a previous
825allocation failed.
826
827Each column represents the number of pages of a certain order which are
828available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in
829ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE
830available in ZONE_NORMAL, etc...
831
832More information relevant to external fragmentation can be found in
833pagetypeinfo.
834
835> cat /proc/pagetypeinfo
836Page block order: 9
837Pages per block: 512
838
839Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
840Node 0, zone DMA, type Unmovable 0 0 0 1 1 1 1 1 1 1 0
841Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0
842Node 0, zone DMA, type Movable 1 1 2 1 2 1 1 0 1 0 2
843Node 0, zone DMA, type Reserve 0 0 0 0 0 0 0 0 0 1 0
844Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0
845Node 0, zone DMA32, type Unmovable 103 54 77 1 1 1 11 8 7 1 9
846Node 0, zone DMA32, type Reclaimable 0 0 2 1 0 0 0 0 1 0 0
847Node 0, zone DMA32, type Movable 169 152 113 91 77 54 39 13 6 1 452
848Node 0, zone DMA32, type Reserve 1 2 2 2 2 0 1 1 1 1 0
849Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0
850
851Number of blocks type Unmovable Reclaimable Movable Reserve Isolate
852Node 0, zone DMA 2 0 5 1 0
853Node 0, zone DMA32 41 6 967 2 0
854
855Fragmentation avoidance in the kernel works by grouping pages of different
856migrate types into the same contiguous regions of memory called page blocks.
857A page block is typically the size of the default hugepage size e.g. 2MB on
858X86-64. By keeping pages grouped based on their ability to move, the kernel
859can reclaim pages within a page block to satisfy a high-order allocation.
860
861The pagetypinfo begins with information on the size of a page block. It
862then gives the same type of information as buddyinfo except broken down
863by migrate-type and finishes with details on how many page blocks of each
864type exist.
865
866If min_free_kbytes has been tuned correctly (recommendations made by hugeadm
867from libhugetlbfs https://github.com/libhugetlbfs/libhugetlbfs/), one can
868make an estimate of the likely number of huge pages that can be allocated
869at a given point in time. All the "Movable" blocks should be allocatable
870unless memory has been mlock()'d. Some of the Reclaimable blocks should
871also be allocatable although a lot of filesystem metadata may have to be
872reclaimed to achieve this.
873
874..............................................................................
875
876meminfo:
877
878Provides information about distribution and utilization of memory. This
879varies by architecture and compile options. The following is from a
88016GB PIII, which has highmem enabled. You may not have all of these fields.
881
882> cat /proc/meminfo
883
884MemTotal: 16344972 kB
885MemFree: 13634064 kB
886MemAvailable: 14836172 kB
887Buffers: 3656 kB
888Cached: 1195708 kB
889SwapCached: 0 kB
890Active: 891636 kB
891Inactive: 1077224 kB
892HighTotal: 15597528 kB
893HighFree: 13629632 kB
894LowTotal: 747444 kB
895LowFree: 4432 kB
896SwapTotal: 0 kB
897SwapFree: 0 kB
898Dirty: 968 kB
899Writeback: 0 kB
900AnonPages: 861800 kB
901Mapped: 280372 kB
902Shmem: 644 kB
903KReclaimable: 168048 kB
904Slab: 284364 kB
905SReclaimable: 159856 kB
906SUnreclaim: 124508 kB
907PageTables: 24448 kB
908NFS_Unstable: 0 kB
909Bounce: 0 kB
910WritebackTmp: 0 kB
911CommitLimit: 7669796 kB
912Committed_AS: 100056 kB
913VmallocTotal: 112216 kB
914VmallocUsed: 428 kB
915VmallocChunk: 111088 kB
916Percpu: 62080 kB
917HardwareCorrupted: 0 kB
918AnonHugePages: 49152 kB
919ShmemHugePages: 0 kB
920ShmemPmdMapped: 0 kB
921
922
923 MemTotal: Total usable ram (i.e. physical ram minus a few reserved
924 bits and the kernel binary code)
925 MemFree: The sum of LowFree+HighFree
926MemAvailable: An estimate of how much memory is available for starting new
927 applications, without swapping. Calculated from MemFree,
928 SReclaimable, the size of the file LRU lists, and the low
929 watermarks in each zone.
930 The estimate takes into account that the system needs some
931 page cache to function well, and that not all reclaimable
932 slab will be reclaimable, due to items being in use. The
933 impact of those factors will vary from system to system.
934 Buffers: Relatively temporary storage for raw disk blocks
935 shouldn't get tremendously large (20MB or so)
936 Cached: in-memory cache for files read from the disk (the
937 pagecache). Doesn't include SwapCached
938 SwapCached: Memory that once was swapped out, is swapped back in but
939 still also is in the swapfile (if memory is needed it
940 doesn't need to be swapped out AGAIN because it is already
941 in the swapfile. This saves I/O)
942 Active: Memory that has been used more recently and usually not
943 reclaimed unless absolutely necessary.
944 Inactive: Memory which has been less recently used. It is more
945 eligible to be reclaimed for other purposes
946 HighTotal:
947 HighFree: Highmem is all memory above ~860MB of physical memory
948 Highmem areas are for use by userspace programs, or
949 for the pagecache. The kernel must use tricks to access
950 this memory, making it slower to access than lowmem.
951 LowTotal:
952 LowFree: Lowmem is memory which can be used for everything that
953 highmem can be used for, but it is also available for the
954 kernel's use for its own data structures. Among many
955 other things, it is where everything from the Slab is
956 allocated. Bad things happen when you're out of lowmem.
957 SwapTotal: total amount of swap space available
958 SwapFree: Memory which has been evicted from RAM, and is temporarily
959 on the disk
960 Dirty: Memory which is waiting to get written back to the disk
961 Writeback: Memory which is actively being written back to the disk
962 AnonPages: Non-file backed pages mapped into userspace page tables
963HardwareCorrupted: The amount of RAM/memory in KB, the kernel identifies as
964 corrupted.
965AnonHugePages: Non-file backed huge pages mapped into userspace page tables
966 Mapped: files which have been mmaped, such as libraries
967 Shmem: Total memory used by shared memory (shmem) and tmpfs
968ShmemHugePages: Memory used by shared memory (shmem) and tmpfs allocated
969 with huge pages
970ShmemPmdMapped: Shared memory mapped into userspace with huge pages
971KReclaimable: Kernel allocations that the kernel will attempt to reclaim
972 under memory pressure. Includes SReclaimable (below), and other
973 direct allocations with a shrinker.
974 Slab: in-kernel data structures cache
975SReclaimable: Part of Slab, that might be reclaimed, such as caches
976 SUnreclaim: Part of Slab, that cannot be reclaimed on memory pressure
977 PageTables: amount of memory dedicated to the lowest level of page
978 tables.
979NFS_Unstable: NFS pages sent to the server, but not yet committed to stable
980 storage
981 Bounce: Memory used for block device "bounce buffers"
982WritebackTmp: Memory used by FUSE for temporary writeback buffers
983 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'),
984 this is the total amount of memory currently available to
985 be allocated on the system. This limit is only adhered to
986 if strict overcommit accounting is enabled (mode 2 in
987 'vm.overcommit_memory').
988 The CommitLimit is calculated with the following formula:
989 CommitLimit = ([total RAM pages] - [total huge TLB pages]) *
990 overcommit_ratio / 100 + [total swap pages]
991 For example, on a system with 1G of physical RAM and 7G
992 of swap with a `vm.overcommit_ratio` of 30 it would
993 yield a CommitLimit of 7.3G.
994 For more details, see the memory overcommit documentation
995 in vm/overcommit-accounting.
996Committed_AS: The amount of memory presently allocated on the system.
997 The committed memory is a sum of all of the memory which
998 has been allocated by processes, even if it has not been
999 "used" by them as of yet. A process which malloc()'s 1G
1000 of memory, but only touches 300M of it will show up as
1001 using 1G. This 1G is memory which has been "committed" to
1002 by the VM and can be used at any time by the allocating
1003 application. With strict overcommit enabled on the system
1004 (mode 2 in 'vm.overcommit_memory'),allocations which would
1005 exceed the CommitLimit (detailed above) will not be permitted.
1006 This is useful if one needs to guarantee that processes will
1007 not fail due to lack of memory once that memory has been
1008 successfully allocated.
1009VmallocTotal: total size of vmalloc memory area
1010 VmallocUsed: amount of vmalloc area which is used
1011VmallocChunk: largest contiguous block of vmalloc area which is free
1012 Percpu: Memory allocated to the percpu allocator used to back percpu
1013 allocations. This stat excludes the cost of metadata.
1014
1015..............................................................................
1016
1017vmallocinfo:
1018
1019Provides information about vmalloced/vmaped areas. One line per area,
1020containing the virtual address range of the area, size in bytes,
1021caller information of the creator, and optional information depending
1022on the kind of area :
1023
1024 pages=nr number of pages
1025 phys=addr if a physical address was specified
1026 ioremap I/O mapping (ioremap() and friends)
1027 vmalloc vmalloc() area
1028 vmap vmap()ed pages
1029 user VM_USERMAP area
1030 vpages buffer for pages pointers was vmalloced (huge area)
1031 N<node>=nr (Only on NUMA kernels)
1032 Number of pages allocated on memory node <node>
1033
1034> cat /proc/vmallocinfo
10350xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ...
1036 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128
10370xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ...
1038 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64
10390xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f...
1040 phys=7fee8000 ioremap
10410xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f...
1042 phys=7fee7000 ioremap
10430xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210
10440xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ...
1045 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3
10460xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ...
1047 pages=2 vmalloc N1=2
10480xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ...
1049 /0x130 [x_tables] pages=4 vmalloc N0=4
10500xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ...
1051 pages=14 vmalloc N2=14
10520xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ...
1053 pages=4 vmalloc N1=4
10540xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ...
1055 pages=2 vmalloc N1=2
10560xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ...
1057 pages=10 vmalloc N0=10
1058
1059..............................................................................
1060
1061softirqs:
1062
1063Provides counts of softirq handlers serviced since boot time, for each cpu.
1064
1065> cat /proc/softirqs
1066 CPU0 CPU1 CPU2 CPU3
1067 HI: 0 0 0 0
1068 TIMER: 27166 27120 27097 27034
1069 NET_TX: 0 0 0 17
1070 NET_RX: 42 0 0 39
1071 BLOCK: 0 0 107 1121
1072 TASKLET: 0 0 0 290
1073 SCHED: 27035 26983 26971 26746
1074 HRTIMER: 0 0 0 0
1075 RCU: 1678 1769 2178 2250
1076
1077
10781.3 IDE devices in /proc/ide
1079----------------------------
1080
1081The subdirectory /proc/ide contains information about all IDE devices of which
1082the kernel is aware. There is one subdirectory for each IDE controller, the
1083file drivers and a link for each IDE device, pointing to the device directory
1084in the controller specific subtree.
1085
1086The file drivers contains general information about the drivers used for the
1087IDE devices:
1088
1089 > cat /proc/ide/drivers
1090 ide-cdrom version 4.53
1091 ide-disk version 1.08
1092
1093More detailed information can be found in the controller specific
1094subdirectories. These are named ide0, ide1 and so on. Each of these
1095directories contains the files shown in table 1-6.
1096
1097
1098Table 1-6: IDE controller info in /proc/ide/ide?
1099..............................................................................
1100 File Content
1101 channel IDE channel (0 or 1)
1102 config Configuration (only for PCI/IDE bridge)
1103 mate Mate name
1104 model Type/Chipset of IDE controller
1105..............................................................................
1106
1107Each device connected to a controller has a separate subdirectory in the
1108controllers directory. The files listed in table 1-7 are contained in these
1109directories.
1110
1111
1112Table 1-7: IDE device information
1113..............................................................................
1114 File Content
1115 cache The cache
1116 capacity Capacity of the medium (in 512Byte blocks)
1117 driver driver and version
1118 geometry physical and logical geometry
1119 identify device identify block
1120 media media type
1121 model device identifier
1122 settings device setup
1123 smart_thresholds IDE disk management thresholds
1124 smart_values IDE disk management values
1125..............................................................................
1126
1127The most interesting file is settings. This file contains a nice overview of
1128the drive parameters:
1129
1130 # cat /proc/ide/ide0/hda/settings
1131 name value min max mode
1132 ---- ----- --- --- ----
1133 bios_cyl 526 0 65535 rw
1134 bios_head 255 0 255 rw
1135 bios_sect 63 0 63 rw
1136 breada_readahead 4 0 127 rw
1137 bswap 0 0 1 r
1138 file_readahead 72 0 2097151 rw
1139 io_32bit 0 0 3 rw
1140 keepsettings 0 0 1 rw
1141 max_kb_per_request 122 1 127 rw
1142 multcount 0 0 8 rw
1143 nice1 1 0 1 rw
1144 nowerr 0 0 1 rw
1145 pio_mode write-only 0 255 w
1146 slow 0 0 1 rw
1147 unmaskirq 0 0 1 rw
1148 using_dma 0 0 1 rw
1149
1150
11511.4 Networking info in /proc/net
1152--------------------------------
1153
1154The subdirectory /proc/net follows the usual pattern. Table 1-8 shows the
1155additional values you get for IP version 6 if you configure the kernel to
1156support this. Table 1-9 lists the files and their meaning.
1157
1158
1159Table 1-8: IPv6 info in /proc/net
1160..............................................................................
1161 File Content
1162 udp6 UDP sockets (IPv6)
1163 tcp6 TCP sockets (IPv6)
1164 raw6 Raw device statistics (IPv6)
1165 igmp6 IP multicast addresses, which this host joined (IPv6)
1166 if_inet6 List of IPv6 interface addresses
1167 ipv6_route Kernel routing table for IPv6
1168 rt6_stats Global IPv6 routing tables statistics
1169 sockstat6 Socket statistics (IPv6)
1170 snmp6 Snmp data (IPv6)
1171..............................................................................
1172
1173
1174Table 1-9: Network info in /proc/net
1175..............................................................................
1176 File Content
1177 arp Kernel ARP table
1178 dev network devices with statistics
1179 dev_mcast the Layer2 multicast groups a device is listening too
1180 (interface index, label, number of references, number of bound
1181 addresses).
1182 dev_stat network device status
1183 ip_fwchains Firewall chain linkage
1184 ip_fwnames Firewall chain names
1185 ip_masq Directory containing the masquerading tables
1186 ip_masquerade Major masquerading table
1187 netstat Network statistics
1188 raw raw device statistics
1189 route Kernel routing table
1190 rpc Directory containing rpc info
1191 rt_cache Routing cache
1192 snmp SNMP data
1193 sockstat Socket statistics
1194 tcp TCP sockets
1195 udp UDP sockets
1196 unix UNIX domain sockets
1197 wireless Wireless interface data (Wavelan etc)
1198 igmp IP multicast addresses, which this host joined
1199 psched Global packet scheduler parameters.
1200 netlink List of PF_NETLINK sockets
1201 ip_mr_vifs List of multicast virtual interfaces
1202 ip_mr_cache List of multicast routing cache
1203..............................................................................
1204
1205You can use this information to see which network devices are available in
1206your system and how much traffic was routed over those devices:
1207
1208 > cat /proc/net/dev
1209 Inter-|Receive |[...
1210 face |bytes packets errs drop fifo frame compressed multicast|[...
1211 lo: 908188 5596 0 0 0 0 0 0 [...
1212 ppp0:15475140 20721 410 0 0 410 0 0 [...
1213 eth0: 614530 7085 0 0 0 0 0 1 [...
1214
1215 ...] Transmit
1216 ...] bytes packets errs drop fifo colls carrier compressed
1217 ...] 908188 5596 0 0 0 0 0 0
1218 ...] 1375103 17405 0 0 0 0 0 0
1219 ...] 1703981 5535 0 0 0 3 0 0
1220
1221In addition, each Channel Bond interface has its own directory. For
1222example, the bond0 device will have a directory called /proc/net/bond0/.
1223It will contain information that is specific to that bond, such as the
1224current slaves of the bond, the link status of the slaves, and how
1225many times the slaves link has failed.
1226
12271.5 SCSI info
1228-------------
1229
1230If you have a SCSI host adapter in your system, you'll find a subdirectory
1231named after the driver for this adapter in /proc/scsi. You'll also see a list
1232of all recognized SCSI devices in /proc/scsi:
1233
1234 >cat /proc/scsi/scsi
1235 Attached devices:
1236 Host: scsi0 Channel: 00 Id: 00 Lun: 00
1237 Vendor: IBM Model: DGHS09U Rev: 03E0
1238 Type: Direct-Access ANSI SCSI revision: 03
1239 Host: scsi0 Channel: 00 Id: 06 Lun: 00
1240 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04
1241 Type: CD-ROM ANSI SCSI revision: 02
1242
1243
1244The directory named after the driver has one file for each adapter found in
1245the system. These files contain information about the controller, including
1246the used IRQ and the IO address range. The amount of information shown is
1247dependent on the adapter you use. The example shows the output for an Adaptec
1248AHA-2940 SCSI adapter:
1249
1250 > cat /proc/scsi/aic7xxx/0
1251
1252 Adaptec AIC7xxx driver version: 5.1.19/3.2.4
1253 Compile Options:
1254 TCQ Enabled By Default : Disabled
1255 AIC7XXX_PROC_STATS : Disabled
1256 AIC7XXX_RESET_DELAY : 5
1257 Adapter Configuration:
1258 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter
1259 Ultra Wide Controller
1260 PCI MMAPed I/O Base: 0xeb001000
1261 Adapter SEEPROM Config: SEEPROM found and used.
1262 Adaptec SCSI BIOS: Enabled
1263 IRQ: 10
1264 SCBs: Active 0, Max Active 2,
1265 Allocated 15, HW 16, Page 255
1266 Interrupts: 160328
1267 BIOS Control Word: 0x18b6
1268 Adapter Control Word: 0x005b
1269 Extended Translation: Enabled
1270 Disconnect Enable Flags: 0xffff
1271 Ultra Enable Flags: 0x0001
1272 Tag Queue Enable Flags: 0x0000
1273 Ordered Queue Tag Flags: 0x0000
1274 Default Tag Queue Depth: 8
1275 Tagged Queue By Device array for aic7xxx host instance 0:
1276 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255}
1277 Actual queue depth per device for aic7xxx host instance 0:
1278 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}
1279 Statistics:
1280 (scsi0:0:0:0)
1281 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8
1282 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0)
1283 Total transfers 160151 (74577 reads and 85574 writes)
1284 (scsi0:0:6:0)
1285 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15
1286 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0)
1287 Total transfers 0 (0 reads and 0 writes)
1288
1289
12901.6 Parallel port info in /proc/parport
1291---------------------------------------
1292
1293The directory /proc/parport contains information about the parallel ports of
1294your system. It has one subdirectory for each port, named after the port
1295number (0,1,2,...).
1296
1297These directories contain the four files shown in Table 1-10.
1298
1299
1300Table 1-10: Files in /proc/parport
1301..............................................................................
1302 File Content
1303 autoprobe Any IEEE-1284 device ID information that has been acquired.
1304 devices list of the device drivers using that port. A + will appear by the
1305 name of the device currently using the port (it might not appear
1306 against any).
1307 hardware Parallel port's base address, IRQ line and DMA channel.
1308 irq IRQ that parport is using for that port. This is in a separate
1309 file to allow you to alter it by writing a new value in (IRQ
1310 number or none).
1311..............................................................................
1312
13131.7 TTY info in /proc/tty
1314-------------------------
1315
1316Information about the available and actually used tty's can be found in the
1317directory /proc/tty.You'll find entries for drivers and line disciplines in
1318this directory, as shown in Table 1-11.
1319
1320
1321Table 1-11: Files in /proc/tty
1322..............................................................................
1323 File Content
1324 drivers list of drivers and their usage
1325 ldiscs registered line disciplines
1326 driver/serial usage statistic and status of single tty lines
1327..............................................................................
1328
1329To see which tty's are currently in use, you can simply look into the file
1330/proc/tty/drivers:
1331
1332 > cat /proc/tty/drivers
1333 pty_slave /dev/pts 136 0-255 pty:slave
1334 pty_master /dev/ptm 128 0-255 pty:master
1335 pty_slave /dev/ttyp 3 0-255 pty:slave
1336 pty_master /dev/pty 2 0-255 pty:master
1337 serial /dev/cua 5 64-67 serial:callout
1338 serial /dev/ttyS 4 64-67 serial
1339 /dev/tty0 /dev/tty0 4 0 system:vtmaster
1340 /dev/ptmx /dev/ptmx 5 2 system
1341 /dev/console /dev/console 5 1 system:console
1342 /dev/tty /dev/tty 5 0 system:/dev/tty
1343 unknown /dev/tty 4 1-63 console
1344
1345
13461.8 Miscellaneous kernel statistics in /proc/stat
1347-------------------------------------------------
1348
1349Various pieces of information about kernel activity are available in the
1350/proc/stat file. All of the numbers reported in this file are aggregates
1351since the system first booted. For a quick look, simply cat the file:
1352
1353 > cat /proc/stat
1354 cpu 2255 34 2290 22625563 6290 127 456 0 0 0
1355 cpu0 1132 34 1441 11311718 3675 127 438 0 0 0
1356 cpu1 1123 0 849 11313845 2614 0 18 0 0 0
1357 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...]
1358 ctxt 1990473
1359 btime 1062191376
1360 processes 2915
1361 procs_running 1
1362 procs_blocked 0
1363 softirq 183433 0 21755 12 39 1137 231 21459 2263
1364
1365The very first "cpu" line aggregates the numbers in all of the other "cpuN"
1366lines. These numbers identify the amount of time the CPU has spent performing
1367different kinds of work. Time units are in USER_HZ (typically hundredths of a
1368second). The meanings of the columns are as follows, from left to right:
1369
1370- user: normal processes executing in user mode
1371- nice: niced processes executing in user mode
1372- system: processes executing in kernel mode
1373- idle: twiddling thumbs
1374- iowait: In a word, iowait stands for waiting for I/O to complete. But there
1375 are several problems:
1376 1. Cpu will not wait for I/O to complete, iowait is the time that a task is
1377 waiting for I/O to complete. When cpu goes into idle state for
1378 outstanding task io, another task will be scheduled on this CPU.
1379 2. In a multi-core CPU, the task waiting for I/O to complete is not running
1380 on any CPU, so the iowait of each CPU is difficult to calculate.
1381 3. The value of iowait field in /proc/stat will decrease in certain
1382 conditions.
1383 So, the iowait is not reliable by reading from /proc/stat.
1384- irq: servicing interrupts
1385- softirq: servicing softirqs
1386- steal: involuntary wait
1387- guest: running a normal guest
1388- guest_nice: running a niced guest
1389
1390The "intr" line gives counts of interrupts serviced since boot time, for each
1391of the possible system interrupts. The first column is the total of all
1392interrupts serviced including unnumbered architecture specific interrupts;
1393each subsequent column is the total for that particular numbered interrupt.
1394Unnumbered interrupts are not shown, only summed into the total.
1395
1396The "ctxt" line gives the total number of context switches across all CPUs.
1397
1398The "btime" line gives the time at which the system booted, in seconds since
1399the Unix epoch.
1400
1401The "processes" line gives the number of processes and threads created, which
1402includes (but is not limited to) those created by calls to the fork() and
1403clone() system calls.
1404
1405The "procs_running" line gives the total number of threads that are
1406running or ready to run (i.e., the total number of runnable threads).
1407
1408The "procs_blocked" line gives the number of processes currently blocked,
1409waiting for I/O to complete.
1410
1411The "softirq" line gives counts of softirqs serviced since boot time, for each
1412of the possible system softirqs. The first column is the total of all
1413softirqs serviced; each subsequent column is the total for that particular
1414softirq.
1415
1416
14171.9 Ext4 file system parameters
1418-------------------------------
1419
1420Information about mounted ext4 file systems can be found in
1421/proc/fs/ext4. Each mounted filesystem will have a directory in
1422/proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or
1423/proc/fs/ext4/dm-0). The files in each per-device directory are shown
1424in Table 1-12, below.
1425
1426Table 1-12: Files in /proc/fs/ext4/<devname>
1427..............................................................................
1428 File Content
1429 mb_groups details of multiblock allocator buddy cache of free blocks
1430..............................................................................
1431
14322.0 /proc/consoles
1433------------------
1434Shows registered system console lines.
1435
1436To see which character device lines are currently used for the system console
1437/dev/console, you may simply look into the file /proc/consoles:
1438
1439 > cat /proc/consoles
1440 tty0 -WU (ECp) 4:7
1441 ttyS0 -W- (Ep) 4:64
1442
1443The columns are:
1444
1445 device name of the device
1446 operations R = can do read operations
1447 W = can do write operations
1448 U = can do unblank
1449 flags E = it is enabled
1450 C = it is preferred console
1451 B = it is primary boot console
1452 p = it is used for printk buffer
1453 b = it is not a TTY but a Braille device
1454 a = it is safe to use when cpu is offline
1455 major:minor major and minor number of the device separated by a colon
1456
1457------------------------------------------------------------------------------
1458Summary
1459------------------------------------------------------------------------------
1460The /proc file system serves information about the running system. It not only
1461allows access to process data but also allows you to request the kernel status
1462by reading files in the hierarchy.
1463
1464The directory structure of /proc reflects the types of information and makes
1465it easy, if not obvious, where to look for specific data.
1466------------------------------------------------------------------------------
1467
1468------------------------------------------------------------------------------
1469CHAPTER 2: MODIFYING SYSTEM PARAMETERS
1470------------------------------------------------------------------------------
1471
1472------------------------------------------------------------------------------
1473In This Chapter
1474------------------------------------------------------------------------------
1475* Modifying kernel parameters by writing into files found in /proc/sys
1476* Exploring the files which modify certain parameters
1477* Review of the /proc/sys file tree
1478------------------------------------------------------------------------------
1479
1480
1481A very interesting part of /proc is the directory /proc/sys. This is not only
1482a source of information, it also allows you to change parameters within the
1483kernel. Be very careful when attempting this. You can optimize your system,
1484but you can also cause it to crash. Never alter kernel parameters on a
1485production system. Set up a development machine and test to make sure that
1486everything works the way you want it to. You may have no alternative but to
1487reboot the machine once an error has been made.
1488
1489To change a value, simply echo the new value into the file. An example is
1490given below in the section on the file system data. You need to be root to do
1491this. You can create your own boot script to perform this every time your
1492system boots.
1493
1494The files in /proc/sys can be used to fine tune and monitor miscellaneous and
1495general things in the operation of the Linux kernel. Since some of the files
1496can inadvertently disrupt your system, it is advisable to read both
1497documentation and source before actually making adjustments. In any case, be
1498very careful when writing to any of these files. The entries in /proc may
1499change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt
1500review the kernel documentation in the directory /usr/src/linux/Documentation.
1501This chapter is heavily based on the documentation included in the pre 2.2
1502kernels, and became part of it in version 2.2.1 of the Linux kernel.
1503
1504Please see: Documentation/admin-guide/sysctl/ directory for descriptions of these
1505entries.
1506
1507------------------------------------------------------------------------------
1508Summary
1509------------------------------------------------------------------------------
1510Certain aspects of kernel behavior can be modified at runtime, without the
1511need to recompile the kernel, or even to reboot the system. The files in the
1512/proc/sys tree can not only be read, but also modified. You can use the echo
1513command to write value into these files, thereby changing the default settings
1514of the kernel.
1515------------------------------------------------------------------------------
1516
1517------------------------------------------------------------------------------
1518CHAPTER 3: PER-PROCESS PARAMETERS
1519------------------------------------------------------------------------------
1520
15213.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj- Adjust the oom-killer score
1522--------------------------------------------------------------------------------
1523
1524These file can be used to adjust the badness heuristic used to select which
1525process gets killed in out of memory conditions.
1526
1527The badness heuristic assigns a value to each candidate task ranging from 0
1528(never kill) to 1000 (always kill) to determine which process is targeted. The
1529units are roughly a proportion along that range of allowed memory the process
1530may allocate from based on an estimation of its current memory and swap use.
1531For example, if a task is using all allowed memory, its badness score will be
15321000. If it is using half of its allowed memory, its score will be 500.
1533
1534There is an additional factor included in the badness score: the current memory
1535and swap usage is discounted by 3% for root processes.
1536
1537The amount of "allowed" memory depends on the context in which the oom killer
1538was called. If it is due to the memory assigned to the allocating task's cpuset
1539being exhausted, the allowed memory represents the set of mems assigned to that
1540cpuset. If it is due to a mempolicy's node(s) being exhausted, the allowed
1541memory represents the set of mempolicy nodes. If it is due to a memory
1542limit (or swap limit) being reached, the allowed memory is that configured
1543limit. Finally, if it is due to the entire system being out of memory, the
1544allowed memory represents all allocatable resources.
1545
1546The value of /proc/<pid>/oom_score_adj is added to the badness score before it
1547is used to determine which task to kill. Acceptable values range from -1000
1548(OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_MAX). This allows userspace to
1549polarize the preference for oom killing either by always preferring a certain
1550task or completely disabling it. The lowest possible value, -1000, is
1551equivalent to disabling oom killing entirely for that task since it will always
1552report a badness score of 0.
1553
1554Consequently, it is very simple for userspace to define the amount of memory to
1555consider for each task. Setting a /proc/<pid>/oom_score_adj value of +500, for
1556example, is roughly equivalent to allowing the remainder of tasks sharing the
1557same system, cpuset, mempolicy, or memory controller resources to use at least
155850% more memory. A value of -500, on the other hand, would be roughly
1559equivalent to discounting 50% of the task's allowed memory from being considered
1560as scoring against the task.
1561
1562For backwards compatibility with previous kernels, /proc/<pid>/oom_adj may also
1563be used to tune the badness score. Its acceptable values range from -16
1564(OOM_ADJUST_MIN) to +15 (OOM_ADJUST_MAX) and a special value of -17
1565(OOM_DISABLE) to disable oom killing entirely for that task. Its value is
1566scaled linearly with /proc/<pid>/oom_score_adj.
1567
1568The value of /proc/<pid>/oom_score_adj may be reduced no lower than the last
1569value set by a CAP_SYS_RESOURCE process. To reduce the value any lower
1570requires CAP_SYS_RESOURCE.
1571
1572Caveat: when a parent task is selected, the oom killer will sacrifice any first
1573generation children with separate address spaces instead, if possible. This
1574avoids servers and important system daemons from being killed and loses the
1575minimal amount of work.
1576
1577
15783.2 /proc/<pid>/oom_score - Display current oom-killer score
1579-------------------------------------------------------------
1580
1581This file can be used to check the current score used by the oom-killer is for
1582any given <pid>. Use it together with /proc/<pid>/oom_score_adj to tune which
1583process should be killed in an out-of-memory situation.
1584
1585
15863.3 /proc/<pid>/io - Display the IO accounting fields
1587-------------------------------------------------------
1588
1589This file contains IO statistics for each running process
1590
1591Example
1592-------
1593
1594test:/tmp # dd if=/dev/zero of=/tmp/test.dat &
1595[1] 3828
1596
1597test:/tmp # cat /proc/3828/io
1598rchar: 323934931
1599wchar: 323929600
1600syscr: 632687
1601syscw: 632675
1602read_bytes: 0
1603write_bytes: 323932160
1604cancelled_write_bytes: 0
1605
1606
1607Description
1608-----------
1609
1610rchar
1611-----
1612
1613I/O counter: chars read
1614The number of bytes which this task has caused to be read from storage. This
1615is simply the sum of bytes which this process passed to read() and pread().
1616It includes things like tty IO and it is unaffected by whether or not actual
1617physical disk IO was required (the read might have been satisfied from
1618pagecache)
1619
1620
1621wchar
1622-----
1623
1624I/O counter: chars written
1625The number of bytes which this task has caused, or shall cause to be written
1626to disk. Similar caveats apply here as with rchar.
1627
1628
1629syscr
1630-----
1631
1632I/O counter: read syscalls
1633Attempt to count the number of read I/O operations, i.e. syscalls like read()
1634and pread().
1635
1636
1637syscw
1638-----
1639
1640I/O counter: write syscalls
1641Attempt to count the number of write I/O operations, i.e. syscalls like
1642write() and pwrite().
1643
1644
1645read_bytes
1646----------
1647
1648I/O counter: bytes read
1649Attempt to count the number of bytes which this process really did cause to
1650be fetched from the storage layer. Done at the submit_bio() level, so it is
1651accurate for block-backed filesystems. <please add status regarding NFS and
1652CIFS at a later time>
1653
1654
1655write_bytes
1656-----------
1657
1658I/O counter: bytes written
1659Attempt to count the number of bytes which this process caused to be sent to
1660the storage layer. This is done at page-dirtying time.
1661
1662
1663cancelled_write_bytes
1664---------------------
1665
1666The big inaccuracy here is truncate. If a process writes 1MB to a file and
1667then deletes the file, it will in fact perform no writeout. But it will have
1668been accounted as having caused 1MB of write.
1669In other words: The number of bytes which this process caused to not happen,
1670by truncating pagecache. A task can cause "negative" IO too. If this task
1671truncates some dirty pagecache, some IO which another task has been accounted
1672for (in its write_bytes) will not be happening. We _could_ just subtract that
1673from the truncating task's write_bytes, but there is information loss in doing
1674that.
1675
1676
1677Note
1678----
1679
1680At its current implementation state, this is a bit racy on 32-bit machines: if
1681process A reads process B's /proc/pid/io while process B is updating one of
1682those 64-bit counters, process A could see an intermediate result.
1683
1684
1685More information about this can be found within the taskstats documentation in
1686Documentation/accounting.
1687
16883.4 /proc/<pid>/coredump_filter - Core dump filtering settings
1689---------------------------------------------------------------
1690When a process is dumped, all anonymous memory is written to a core file as
1691long as the size of the core file isn't limited. But sometimes we don't want
1692to dump some memory segments, for example, huge shared memory or DAX.
1693Conversely, sometimes we want to save file-backed memory segments into a core
1694file, not only the individual files.
1695
1696/proc/<pid>/coredump_filter allows you to customize which memory segments
1697will be dumped when the <pid> process is dumped. coredump_filter is a bitmask
1698of memory types. If a bit of the bitmask is set, memory segments of the
1699corresponding memory type are dumped, otherwise they are not dumped.
1700
1701The following 9 memory types are supported:
1702 - (bit 0) anonymous private memory
1703 - (bit 1) anonymous shared memory
1704 - (bit 2) file-backed private memory
1705 - (bit 3) file-backed shared memory
1706 - (bit 4) ELF header pages in file-backed private memory areas (it is
1707 effective only if the bit 2 is cleared)
1708 - (bit 5) hugetlb private memory
1709 - (bit 6) hugetlb shared memory
1710 - (bit 7) DAX private memory
1711 - (bit 8) DAX shared memory
1712
1713 Note that MMIO pages such as frame buffer are never dumped and vDSO pages
1714 are always dumped regardless of the bitmask status.
1715
1716 Note that bits 0-4 don't affect hugetlb or DAX memory. hugetlb memory is
1717 only affected by bit 5-6, and DAX is only affected by bits 7-8.
1718
1719The default value of coredump_filter is 0x33; this means all anonymous memory
1720segments, ELF header pages and hugetlb private memory are dumped.
1721
1722If you don't want to dump all shared memory segments attached to pid 1234,
1723write 0x31 to the process's proc file.
1724
1725 $ echo 0x31 > /proc/1234/coredump_filter
1726
1727When a new process is created, the process inherits the bitmask status from its
1728parent. It is useful to set up coredump_filter before the program runs.
1729For example:
1730
1731 $ echo 0x7 > /proc/self/coredump_filter
1732 $ ./some_program
1733
17343.5 /proc/<pid>/mountinfo - Information about mounts
1735--------------------------------------------------------
1736
1737This file contains lines of the form:
1738
173936 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue
1740(1)(2)(3) (4) (5) (6) (7) (8) (9) (10) (11)
1741
1742(1) mount ID: unique identifier of the mount (may be reused after umount)
1743(2) parent ID: ID of parent (or of self for the top of the mount tree)
1744(3) major:minor: value of st_dev for files on filesystem
1745(4) root: root of the mount within the filesystem
1746(5) mount point: mount point relative to the process's root
1747(6) mount options: per mount options
1748(7) optional fields: zero or more fields of the form "tag[:value]"
1749(8) separator: marks the end of the optional fields
1750(9) filesystem type: name of filesystem of the form "type[.subtype]"
1751(10) mount source: filesystem specific information or "none"
1752(11) super options: per super block options
1753
1754Parsers should ignore all unrecognised optional fields. Currently the
1755possible optional fields are:
1756
1757shared:X mount is shared in peer group X
1758master:X mount is slave to peer group X
1759propagate_from:X mount is slave and receives propagation from peer group X (*)
1760unbindable mount is unbindable
1761
1762(*) X is the closest dominant peer group under the process's root. If
1763X is the immediate master of the mount, or if there's no dominant peer
1764group under the same root, then only the "master:X" field is present
1765and not the "propagate_from:X" field.
1766
1767For more information on mount propagation see:
1768
1769 Documentation/filesystems/sharedsubtree.txt
1770
1771
17723.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm
1773--------------------------------------------------------
1774These files provide a method to access a tasks comm value. It also allows for
1775a task to set its own or one of its thread siblings comm value. The comm value
1776is limited in size compared to the cmdline value, so writing anything longer
1777then the kernel's TASK_COMM_LEN (currently 16 chars) will result in a truncated
1778comm value.
1779
1780
17813.7 /proc/<pid>/task/<tid>/children - Information about task children
1782-------------------------------------------------------------------------
1783This file provides a fast way to retrieve first level children pids
1784of a task pointed by <pid>/<tid> pair. The format is a space separated
1785stream of pids.
1786
1787Note the "first level" here -- if a child has own children they will
1788not be listed here, one needs to read /proc/<children-pid>/task/<tid>/children
1789to obtain the descendants.
1790
1791Since this interface is intended to be fast and cheap it doesn't
1792guarantee to provide precise results and some children might be
1793skipped, especially if they've exited right after we printed their
1794pids, so one need to either stop or freeze processes being inspected
1795if precise results are needed.
1796
1797
17983.8 /proc/<pid>/fdinfo/<fd> - Information about opened file
1799---------------------------------------------------------------
1800This file provides information associated with an opened file. The regular
1801files have at least three fields -- 'pos', 'flags' and mnt_id. The 'pos'
1802represents the current offset of the opened file in decimal form [see lseek(2)
1803for details], 'flags' denotes the octal O_xxx mask the file has been
1804created with [see open(2) for details] and 'mnt_id' represents mount ID of
1805the file system containing the opened file [see 3.5 /proc/<pid>/mountinfo
1806for details].
1807
1808A typical output is
1809
1810 pos: 0
1811 flags: 0100002
1812 mnt_id: 19
1813
1814All locks associated with a file descriptor are shown in its fdinfo too.
1815
1816lock: 1: FLOCK ADVISORY WRITE 359 00:13:11691 0 EOF
1817
1818The files such as eventfd, fsnotify, signalfd, epoll among the regular pos/flags
1819pair provide additional information particular to the objects they represent.
1820
1821 Eventfd files
1822 ~~~~~~~~~~~~~
1823 pos: 0
1824 flags: 04002
1825 mnt_id: 9
1826 eventfd-count: 5a
1827
1828 where 'eventfd-count' is hex value of a counter.
1829
1830 Signalfd files
1831 ~~~~~~~~~~~~~~
1832 pos: 0
1833 flags: 04002
1834 mnt_id: 9
1835 sigmask: 0000000000000200
1836
1837 where 'sigmask' is hex value of the signal mask associated
1838 with a file.
1839
1840 Epoll files
1841 ~~~~~~~~~~~
1842 pos: 0
1843 flags: 02
1844 mnt_id: 9
1845 tfd: 5 events: 1d data: ffffffffffffffff pos:0 ino:61af sdev:7
1846
1847 where 'tfd' is a target file descriptor number in decimal form,
1848 'events' is events mask being watched and the 'data' is data
1849 associated with a target [see epoll(7) for more details].
1850
1851 The 'pos' is current offset of the target file in decimal form
1852 [see lseek(2)], 'ino' and 'sdev' are inode and device numbers
1853 where target file resides, all in hex format.
1854
1855 Fsnotify files
1856 ~~~~~~~~~~~~~~
1857 For inotify files the format is the following
1858
1859 pos: 0
1860 flags: 02000000
1861 inotify wd:3 ino:9e7e sdev:800013 mask:800afce ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:7e9e0000640d1b6d
1862
1863 where 'wd' is a watch descriptor in decimal form, ie a target file
1864 descriptor number, 'ino' and 'sdev' are inode and device where the
1865 target file resides and the 'mask' is the mask of events, all in hex
1866 form [see inotify(7) for more details].
1867
1868 If the kernel was built with exportfs support, the path to the target
1869 file is encoded as a file handle. The file handle is provided by three
1870 fields 'fhandle-bytes', 'fhandle-type' and 'f_handle', all in hex
1871 format.
1872
1873 If the kernel is built without exportfs support the file handle won't be
1874 printed out.
1875
1876 If there is no inotify mark attached yet the 'inotify' line will be omitted.
1877
1878 For fanotify files the format is
1879
1880 pos: 0
1881 flags: 02
1882 mnt_id: 9
1883 fanotify flags:10 event-flags:0
1884 fanotify mnt_id:12 mflags:40 mask:38 ignored_mask:40000003
1885 fanotify ino:4f969 sdev:800013 mflags:0 mask:3b ignored_mask:40000000 fhandle-bytes:8 fhandle-type:1 f_handle:69f90400c275b5b4
1886
1887 where fanotify 'flags' and 'event-flags' are values used in fanotify_init
1888 call, 'mnt_id' is the mount point identifier, 'mflags' is the value of
1889 flags associated with mark which are tracked separately from events
1890 mask. 'ino', 'sdev' are target inode and device, 'mask' is the events
1891 mask and 'ignored_mask' is the mask of events which are to be ignored.
1892 All in hex format. Incorporation of 'mflags', 'mask' and 'ignored_mask'
1893 does provide information about flags and mask used in fanotify_mark
1894 call [see fsnotify manpage for details].
1895
1896 While the first three lines are mandatory and always printed, the rest is
1897 optional and may be omitted if no marks created yet.
1898
1899 Timerfd files
1900 ~~~~~~~~~~~~~
1901
1902 pos: 0
1903 flags: 02
1904 mnt_id: 9
1905 clockid: 0
1906 ticks: 0
1907 settime flags: 01
1908 it_value: (0, 49406829)
1909 it_interval: (1, 0)
1910
1911 where 'clockid' is the clock type and 'ticks' is the number of the timer expirations
1912 that have occurred [see timerfd_create(2) for details]. 'settime flags' are
1913 flags in octal form been used to setup the timer [see timerfd_settime(2) for
1914 details]. 'it_value' is remaining time until the timer exiration.
1915 'it_interval' is the interval for the timer. Note the timer might be set up
1916 with TIMER_ABSTIME option which will be shown in 'settime flags', but 'it_value'
1917 still exhibits timer's remaining time.
1918
19193.9 /proc/<pid>/map_files - Information about memory mapped files
1920---------------------------------------------------------------------
1921This directory contains symbolic links which represent memory mapped files
1922the process is maintaining. Example output:
1923
1924 | lr-------- 1 root root 64 Jan 27 11:24 333c600000-333c620000 -> /usr/lib64/ld-2.18.so
1925 | lr-------- 1 root root 64 Jan 27 11:24 333c81f000-333c820000 -> /usr/lib64/ld-2.18.so
1926 | lr-------- 1 root root 64 Jan 27 11:24 333c820000-333c821000 -> /usr/lib64/ld-2.18.so
1927 | ...
1928 | lr-------- 1 root root 64 Jan 27 11:24 35d0421000-35d0422000 -> /usr/lib64/libselinux.so.1
1929 | lr-------- 1 root root 64 Jan 27 11:24 400000-41a000 -> /usr/bin/ls
1930
1931The name of a link represents the virtual memory bounds of a mapping, i.e.
1932vm_area_struct::vm_start-vm_area_struct::vm_end.
1933
1934The main purpose of the map_files is to retrieve a set of memory mapped
1935files in a fast way instead of parsing /proc/<pid>/maps or
1936/proc/<pid>/smaps, both of which contain many more records. At the same
1937time one can open(2) mappings from the listings of two processes and
1938comparing their inode numbers to figure out which anonymous memory areas
1939are actually shared.
1940
19413.10 /proc/<pid>/timerslack_ns - Task timerslack value
1942---------------------------------------------------------
1943This file provides the value of the task's timerslack value in nanoseconds.
1944This value specifies a amount of time that normal timers may be deferred
1945in order to coalesce timers and avoid unnecessary wakeups.
1946
1947This allows a task's interactivity vs power consumption trade off to be
1948adjusted.
1949
1950Writing 0 to the file will set the tasks timerslack to the default value.
1951
1952Valid values are from 0 - ULLONG_MAX
1953
1954An application setting the value must have PTRACE_MODE_ATTACH_FSCREDS level
1955permissions on the task specified to change its timerslack_ns value.
1956
19573.11 /proc/<pid>/patch_state - Livepatch patch operation state
1958-----------------------------------------------------------------
1959When CONFIG_LIVEPATCH is enabled, this file displays the value of the
1960patch state for the task.
1961
1962A value of '-1' indicates that no patch is in transition.
1963
1964A value of '0' indicates that a patch is in transition and the task is
1965unpatched. If the patch is being enabled, then the task hasn't been
1966patched yet. If the patch is being disabled, then the task has already
1967been unpatched.
1968
1969A value of '1' indicates that a patch is in transition and the task is
1970patched. If the patch is being enabled, then the task has already been
1971patched. If the patch is being disabled, then the task hasn't been
1972unpatched yet.
1973
19743.12 /proc/<pid>/arch_status - task architecture specific status
1975-------------------------------------------------------------------
1976When CONFIG_PROC_PID_ARCH_STATUS is enabled, this file displays the
1977architecture specific status of the task.
1978
1979Example
1980-------
1981 $ cat /proc/6753/arch_status
1982 AVX512_elapsed_ms: 8
1983
1984Description
1985-----------
1986
1987x86 specific entries:
1988---------------------
1989 AVX512_elapsed_ms:
1990 ------------------
1991 If AVX512 is supported on the machine, this entry shows the milliseconds
1992 elapsed since the last time AVX512 usage was recorded. The recording
1993 happens on a best effort basis when a task is scheduled out. This means
1994 that the value depends on two factors:
1995
1996 1) The time which the task spent on the CPU without being scheduled
1997 out. With CPU isolation and a single runnable task this can take
1998 several seconds.
1999
2000 2) The time since the task was scheduled out last. Depending on the
2001 reason for being scheduled out (time slice exhausted, syscall ...)
2002 this can be arbitrary long time.
2003
2004 As a consequence the value cannot be considered precise and authoritative
2005 information. The application which uses this information has to be aware
2006 of the overall scenario on the system in order to determine whether a
2007 task is a real AVX512 user or not. Precise information can be obtained
2008 with performance counters.
2009
2010 A special value of '-1' indicates that no AVX512 usage was recorded, thus
2011 the task is unlikely an AVX512 user, but depends on the workload and the
2012 scheduling scenario, it also could be a false negative mentioned above.
2013
2014------------------------------------------------------------------------------
2015Configuring procfs
2016------------------------------------------------------------------------------
2017
20184.1 Mount options
2019---------------------
2020
2021The following mount options are supported:
2022
2023 hidepid= Set /proc/<pid>/ access mode.
2024 gid= Set the group authorized to learn processes information.
2025
2026hidepid=0 means classic mode - everybody may access all /proc/<pid>/ directories
2027(default).
2028
2029hidepid=1 means users may not access any /proc/<pid>/ directories but their
2030own. Sensitive files like cmdline, sched*, status are now protected against
2031other users. This makes it impossible to learn whether any user runs
2032specific program (given the program doesn't reveal itself by its behaviour).
2033As an additional bonus, as /proc/<pid>/cmdline is unaccessible for other users,
2034poorly written programs passing sensitive information via program arguments are
2035now protected against local eavesdroppers.
2036
2037hidepid=2 means hidepid=1 plus all /proc/<pid>/ will be fully invisible to other
2038users. It doesn't mean that it hides a fact whether a process with a specific
2039pid value exists (it can be learned by other means, e.g. by "kill -0 $PID"),
2040but it hides process' uid and gid, which may be learned by stat()'ing
2041/proc/<pid>/ otherwise. It greatly complicates an intruder's task of gathering
2042information about running processes, whether some daemon runs with elevated
2043privileges, whether other user runs some sensitive program, whether other users
2044run any program at all, etc.
2045
2046gid= defines a group authorized to learn processes information otherwise
2047prohibited by hidepid=. If you use some daemon like identd which needs to learn
2048information about processes information, just add identd to this group.