1.. SPDX-License-Identifier: GPL-2.0 2 3==================== 4The /proc Filesystem 5==================== 6 7===================== ======================================= ================ 8/proc/sys Terrehon Bowden <terrehon@pacbell.net>, October 7 1999 9 Bodo Bauer <bb@ricochet.net> 102.4.x update Jorge Nerin <comandante@zaralinux.com> November 14 2000 11move /proc/sys Shen Feng <shen@cn.fujitsu.com> April 1 2009 12fixes/update part 1.1 Stefani Seibold <stefani@seibold.net> June 9 2009 13===================== ======================================= ================ 14 15 16 17.. Table of Contents 18 19 0 Preface 20 0.1 Introduction/Credits 21 0.2 Legal Stuff 22 23 1 Collecting System Information 24 1.1 Process-Specific Subdirectories 25 1.2 Kernel data 26 1.3 IDE devices in /proc/ide 27 1.4 Networking info in /proc/net 28 1.5 SCSI info 29 1.6 Parallel port info in /proc/parport 30 1.7 TTY info in /proc/tty 31 1.8 Miscellaneous kernel statistics in /proc/stat 32 1.9 Ext4 file system parameters 33 34 2 Modifying System Parameters 35 36 3 Per-Process Parameters 37 3.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj - Adjust the oom-killer 38 score 39 3.2 /proc/<pid>/oom_score - Display current oom-killer score 40 3.3 /proc/<pid>/io - Display the IO accounting fields 41 3.4 /proc/<pid>/coredump_filter - Core dump filtering settings 42 3.5 /proc/<pid>/mountinfo - Information about mounts 43 3.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm 44 3.7 /proc/<pid>/task/<tid>/children - Information about task children 45 3.8 /proc/<pid>/fdinfo/<fd> - Information about opened file 46 3.9 /proc/<pid>/map_files - Information about memory mapped files 47 3.10 /proc/<pid>/timerslack_ns - Task timerslack value 48 3.11 /proc/<pid>/patch_state - Livepatch patch operation state 49 3.12 /proc/<pid>/arch_status - Task architecture specific information 50 51 4 Configuring procfs 52 4.1 Mount options 53 54 5 Filesystem behavior 55 56Preface 57======= 58 590.1 Introduction/Credits 60------------------------ 61 62This documentation is part of a soon (or so we hope) to be released book on 63the SuSE Linux distribution. As there is no complete documentation for the 64/proc file system and we've used many freely available sources to write these 65chapters, it seems only fair to give the work back to the Linux community. 66This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm 67afraid it's still far from complete, but we hope it will be useful. As far as 68we know, it is the first 'all-in-one' document about the /proc file system. It 69is focused on the Intel x86 hardware, so if you are looking for PPC, ARM, 70SPARC, AXP, etc., features, you probably won't find what you are looking for. 71It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But 72additions and patches are welcome and will be added to this document if you 73mail them to Bodo. 74 75We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of 76other people for help compiling this documentation. We'd also like to extend a 77special thank you to Andi Kleen for documentation, which we relied on heavily 78to create this document, as well as the additional information he provided. 79Thanks to everybody else who contributed source or docs to the Linux kernel 80and helped create a great piece of software... :) 81 82If you have any comments, corrections or additions, please don't hesitate to 83contact Bodo Bauer at bb@ricochet.net. We'll be happy to add them to this 84document. 85 86The latest version of this document is available online at 87http://tldp.org/LDP/Linux-Filesystem-Hierarchy/html/proc.html 88 89If the above direction does not works for you, you could try the kernel 90mailing list at linux-kernel@vger.kernel.org and/or try to reach me at 91comandante@zaralinux.com. 92 930.2 Legal Stuff 94--------------- 95 96We don't guarantee the correctness of this document, and if you come to us 97complaining about how you screwed up your system because of incorrect 98documentation, we won't feel responsible... 99 100Chapter 1: Collecting System Information 101======================================== 102 103In This Chapter 104--------------- 105* Investigating the properties of the pseudo file system /proc and its 106 ability to provide information on the running Linux system 107* Examining /proc's structure 108* Uncovering various information about the kernel and the processes running 109 on the system 110 111------------------------------------------------------------------------------ 112 113The proc file system acts as an interface to internal data structures in the 114kernel. It can be used to obtain information about the system and to change 115certain kernel parameters at runtime (sysctl). 116 117First, we'll take a look at the read-only parts of /proc. In Chapter 2, we 118show you how you can use /proc/sys to change settings. 119 1201.1 Process-Specific Subdirectories 121----------------------------------- 122 123The directory /proc contains (among other things) one subdirectory for each 124process running on the system, which is named after the process ID (PID). 125 126The link 'self' points to the process reading the file system. Each process 127subdirectory has the entries listed in Table 1-1. 128 129Note that an open file descriptor to /proc/<pid> or to any of its 130contained files or subdirectories does not prevent <pid> being reused 131for some other process in the event that <pid> exits. Operations on 132open /proc/<pid> file descriptors corresponding to dead processes 133never act on any new process that the kernel may, through chance, have 134also assigned the process ID <pid>. Instead, operations on these FDs 135usually fail with ESRCH. 136 137.. table:: Table 1-1: Process specific entries in /proc 138 139 ============= =============================================================== 140 File Content 141 ============= =============================================================== 142 clear_refs Clears page referenced bits shown in smaps output 143 cmdline Command line arguments 144 cpu Current and last cpu in which it was executed (2.4)(smp) 145 cwd Link to the current working directory 146 environ Values of environment variables 147 exe Link to the executable of this process 148 fd Directory, which contains all file descriptors 149 maps Memory maps to executables and library files (2.4) 150 mem Memory held by this process 151 root Link to the root directory of this process 152 stat Process status 153 statm Process memory status information 154 status Process status in human readable form 155 wchan Present with CONFIG_KALLSYMS=y: it shows the kernel function 156 symbol the task is blocked in - or "0" if not blocked. 157 pagemap Page table 158 stack Report full stack trace, enable via CONFIG_STACKTRACE 159 smaps An extension based on maps, showing the memory consumption of 160 each mapping and flags associated with it 161 smaps_rollup Accumulated smaps stats for all mappings of the process. This 162 can be derived from smaps, but is faster and more convenient 163 numa_maps An extension based on maps, showing the memory locality and 164 binding policy as well as mem usage (in pages) of each mapping. 165 ============= =============================================================== 166 167For example, to get the status information of a process, all you have to do is 168read the file /proc/PID/status:: 169 170 >cat /proc/self/status 171 Name: cat 172 State: R (running) 173 Tgid: 5452 174 Pid: 5452 175 PPid: 743 176 TracerPid: 0 (2.4) 177 Uid: 501 501 501 501 178 Gid: 100 100 100 100 179 FDSize: 256 180 Groups: 100 14 16 181 VmPeak: 5004 kB 182 VmSize: 5004 kB 183 VmLck: 0 kB 184 VmHWM: 476 kB 185 VmRSS: 476 kB 186 RssAnon: 352 kB 187 RssFile: 120 kB 188 RssShmem: 4 kB 189 VmData: 156 kB 190 VmStk: 88 kB 191 VmExe: 68 kB 192 VmLib: 1412 kB 193 VmPTE: 20 kb 194 VmSwap: 0 kB 195 HugetlbPages: 0 kB 196 CoreDumping: 0 197 THP_enabled: 1 198 Threads: 1 199 SigQ: 0/28578 200 SigPnd: 0000000000000000 201 ShdPnd: 0000000000000000 202 SigBlk: 0000000000000000 203 SigIgn: 0000000000000000 204 SigCgt: 0000000000000000 205 CapInh: 00000000fffffeff 206 CapPrm: 0000000000000000 207 CapEff: 0000000000000000 208 CapBnd: ffffffffffffffff 209 CapAmb: 0000000000000000 210 NoNewPrivs: 0 211 Seccomp: 0 212 Speculation_Store_Bypass: thread vulnerable 213 voluntary_ctxt_switches: 0 214 nonvoluntary_ctxt_switches: 1 215 216This shows you nearly the same information you would get if you viewed it with 217the ps command. In fact, ps uses the proc file system to obtain its 218information. But you get a more detailed view of the process by reading the 219file /proc/PID/status. It fields are described in table 1-2. 220 221The statm file contains more detailed information about the process 222memory usage. Its seven fields are explained in Table 1-3. The stat file 223contains detailed information about the process itself. Its fields are 224explained in Table 1-4. 225 226(for SMP CONFIG users) 227 228For making accounting scalable, RSS related information are handled in an 229asynchronous manner and the value may not be very precise. To see a precise 230snapshot of a moment, you can see /proc/<pid>/smaps file and scan page table. 231It's slow but very precise. 232 233.. table:: Table 1-2: Contents of the status files (as of 4.19) 234 235 ========================== =================================================== 236 Field Content 237 ========================== =================================================== 238 Name filename of the executable 239 Umask file mode creation mask 240 State state (R is running, S is sleeping, D is sleeping 241 in an uninterruptible wait, Z is zombie, 242 T is traced or stopped) 243 Tgid thread group ID 244 Ngid NUMA group ID (0 if none) 245 Pid process id 246 PPid process id of the parent process 247 TracerPid PID of process tracing this process (0 if not) 248 Uid Real, effective, saved set, and file system UIDs 249 Gid Real, effective, saved set, and file system GIDs 250 FDSize number of file descriptor slots currently allocated 251 Groups supplementary group list 252 NStgid descendant namespace thread group ID hierarchy 253 NSpid descendant namespace process ID hierarchy 254 NSpgid descendant namespace process group ID hierarchy 255 NSsid descendant namespace session ID hierarchy 256 VmPeak peak virtual memory size 257 VmSize total program size 258 VmLck locked memory size 259 VmPin pinned memory size 260 VmHWM peak resident set size ("high water mark") 261 VmRSS size of memory portions. It contains the three 262 following parts 263 (VmRSS = RssAnon + RssFile + RssShmem) 264 RssAnon size of resident anonymous memory 265 RssFile size of resident file mappings 266 RssShmem size of resident shmem memory (includes SysV shm, 267 mapping of tmpfs and shared anonymous mappings) 268 VmData size of private data segments 269 VmStk size of stack segments 270 VmExe size of text segment 271 VmLib size of shared library code 272 VmPTE size of page table entries 273 VmSwap amount of swap used by anonymous private data 274 (shmem swap usage is not included) 275 HugetlbPages size of hugetlb memory portions 276 CoreDumping process's memory is currently being dumped 277 (killing the process may lead to a corrupted core) 278 THP_enabled process is allowed to use THP (returns 0 when 279 PR_SET_THP_DISABLE is set on the process 280 Threads number of threads 281 SigQ number of signals queued/max. number for queue 282 SigPnd bitmap of pending signals for the thread 283 ShdPnd bitmap of shared pending signals for the process 284 SigBlk bitmap of blocked signals 285 SigIgn bitmap of ignored signals 286 SigCgt bitmap of caught signals 287 CapInh bitmap of inheritable capabilities 288 CapPrm bitmap of permitted capabilities 289 CapEff bitmap of effective capabilities 290 CapBnd bitmap of capabilities bounding set 291 CapAmb bitmap of ambient capabilities 292 NoNewPrivs no_new_privs, like prctl(PR_GET_NO_NEW_PRIV, ...) 293 Seccomp seccomp mode, like prctl(PR_GET_SECCOMP, ...) 294 Speculation_Store_Bypass speculative store bypass mitigation status 295 Cpus_allowed mask of CPUs on which this process may run 296 Cpus_allowed_list Same as previous, but in "list format" 297 Mems_allowed mask of memory nodes allowed to this process 298 Mems_allowed_list Same as previous, but in "list format" 299 voluntary_ctxt_switches number of voluntary context switches 300 nonvoluntary_ctxt_switches number of non voluntary context switches 301 ========================== =================================================== 302 303 304.. table:: Table 1-3: Contents of the statm files (as of 2.6.8-rc3) 305 306 ======== =============================== ============================== 307 Field Content 308 ======== =============================== ============================== 309 size total program size (pages) (same as VmSize in status) 310 resident size of memory portions (pages) (same as VmRSS in status) 311 shared number of pages that are shared (i.e. backed by a file, same 312 as RssFile+RssShmem in status) 313 trs number of pages that are 'code' (not including libs; broken, 314 includes data segment) 315 lrs number of pages of library (always 0 on 2.6) 316 drs number of pages of data/stack (including libs; broken, 317 includes library text) 318 dt number of dirty pages (always 0 on 2.6) 319 ======== =============================== ============================== 320 321 322.. table:: Table 1-4: Contents of the stat files (as of 2.6.30-rc7) 323 324 ============= =============================================================== 325 Field Content 326 ============= =============================================================== 327 pid process id 328 tcomm filename of the executable 329 state state (R is running, S is sleeping, D is sleeping in an 330 uninterruptible wait, Z is zombie, T is traced or stopped) 331 ppid process id of the parent process 332 pgrp pgrp of the process 333 sid session id 334 tty_nr tty the process uses 335 tty_pgrp pgrp of the tty 336 flags task flags 337 min_flt number of minor faults 338 cmin_flt number of minor faults with child's 339 maj_flt number of major faults 340 cmaj_flt number of major faults with child's 341 utime user mode jiffies 342 stime kernel mode jiffies 343 cutime user mode jiffies with child's 344 cstime kernel mode jiffies with child's 345 priority priority level 346 nice nice level 347 num_threads number of threads 348 it_real_value (obsolete, always 0) 349 start_time time the process started after system boot 350 vsize virtual memory size 351 rss resident set memory size 352 rsslim current limit in bytes on the rss 353 start_code address above which program text can run 354 end_code address below which program text can run 355 start_stack address of the start of the main process stack 356 esp current value of ESP 357 eip current value of EIP 358 pending bitmap of pending signals 359 blocked bitmap of blocked signals 360 sigign bitmap of ignored signals 361 sigcatch bitmap of caught signals 362 0 (place holder, used to be the wchan address, 363 use /proc/PID/wchan instead) 364 0 (place holder) 365 0 (place holder) 366 exit_signal signal to send to parent thread on exit 367 task_cpu which CPU the task is scheduled on 368 rt_priority realtime priority 369 policy scheduling policy (man sched_setscheduler) 370 blkio_ticks time spent waiting for block IO 371 gtime guest time of the task in jiffies 372 cgtime guest time of the task children in jiffies 373 start_data address above which program data+bss is placed 374 end_data address below which program data+bss is placed 375 start_brk address above which program heap can be expanded with brk() 376 arg_start address above which program command line is placed 377 arg_end address below which program command line is placed 378 env_start address above which program environment is placed 379 env_end address below which program environment is placed 380 exit_code the thread's exit_code in the form reported by the waitpid 381 system call 382 ============= =============================================================== 383 384The /proc/PID/maps file contains the currently mapped memory regions and 385their access permissions. 386 387The format is:: 388 389 address perms offset dev inode pathname 390 391 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test 392 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test 393 0804a000-0806b000 rw-p 00000000 00:00 0 [heap] 394 a7cb1000-a7cb2000 ---p 00000000 00:00 0 395 a7cb2000-a7eb2000 rw-p 00000000 00:00 0 396 a7eb2000-a7eb3000 ---p 00000000 00:00 0 397 a7eb3000-a7ed5000 rw-p 00000000 00:00 0 398 a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6 399 a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6 400 a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6 401 a800b000-a800e000 rw-p 00000000 00:00 0 402 a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0 403 a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0 404 a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0 405 a8024000-a8027000 rw-p 00000000 00:00 0 406 a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2 407 a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2 408 a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2 409 aff35000-aff4a000 rw-p 00000000 00:00 0 [stack] 410 ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso] 411 412where "address" is the address space in the process that it occupies, "perms" 413is a set of permissions:: 414 415 r = read 416 w = write 417 x = execute 418 s = shared 419 p = private (copy on write) 420 421"offset" is the offset into the mapping, "dev" is the device (major:minor), and 422"inode" is the inode on that device. 0 indicates that no inode is associated 423with the memory region, as the case would be with BSS (uninitialized data). 424The "pathname" shows the name associated file for this mapping. If the mapping 425is not associated with a file: 426 427 ======= ==================================== 428 [heap] the heap of the program 429 [stack] the stack of the main process 430 [vdso] the "virtual dynamic shared object", 431 the kernel system call handler 432 [anon:<name>] an anonymous mapping that has been 433 named by userspace 434 ======= ==================================== 435 436 or if empty, the mapping is anonymous. 437 438The /proc/PID/smaps is an extension based on maps, showing the memory 439consumption for each of the process's mappings. For each mapping (aka Virtual 440Memory Area, or VMA) there is a series of lines such as the following:: 441 442 08048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash 443 444 Size: 1084 kB 445 KernelPageSize: 4 kB 446 MMUPageSize: 4 kB 447 Rss: 892 kB 448 Pss: 374 kB 449 Shared_Clean: 892 kB 450 Shared_Dirty: 0 kB 451 Private_Clean: 0 kB 452 Private_Dirty: 0 kB 453 Referenced: 892 kB 454 Anonymous: 0 kB 455 LazyFree: 0 kB 456 AnonHugePages: 0 kB 457 ShmemPmdMapped: 0 kB 458 Shared_Hugetlb: 0 kB 459 Private_Hugetlb: 0 kB 460 Swap: 0 kB 461 SwapPss: 0 kB 462 KernelPageSize: 4 kB 463 MMUPageSize: 4 kB 464 Locked: 0 kB 465 THPeligible: 0 466 VmFlags: rd ex mr mw me dw 467 Name: name from userspace 468 469The first of these lines shows the same information as is displayed for the 470mapping in /proc/PID/maps. Following lines show the size of the mapping 471(size); the size of each page allocated when backing a VMA (KernelPageSize), 472which is usually the same as the size in the page table entries; the page size 473used by the MMU when backing a VMA (in most cases, the same as KernelPageSize); 474the amount of the mapping that is currently resident in RAM (RSS); the 475process' proportional share of this mapping (PSS); and the number of clean and 476dirty shared and private pages in the mapping. 477 478The "proportional set size" (PSS) of a process is the count of pages it has 479in memory, where each page is divided by the number of processes sharing it. 480So if a process has 1000 pages all to itself, and 1000 shared with one other 481process, its PSS will be 1500. 482 483Note that even a page which is part of a MAP_SHARED mapping, but has only 484a single pte mapped, i.e. is currently used by only one process, is accounted 485as private and not as shared. 486 487"Referenced" indicates the amount of memory currently marked as referenced or 488accessed. 489 490"Anonymous" shows the amount of memory that does not belong to any file. Even 491a mapping associated with a file may contain anonymous pages: when MAP_PRIVATE 492and a page is modified, the file page is replaced by a private anonymous copy. 493 494"LazyFree" shows the amount of memory which is marked by madvise(MADV_FREE). 495The memory isn't freed immediately with madvise(). It's freed in memory 496pressure if the memory is clean. Please note that the printed value might 497be lower than the real value due to optimizations used in the current 498implementation. If this is not desirable please file a bug report. 499 500"AnonHugePages" shows the ammount of memory backed by transparent hugepage. 501 502"ShmemPmdMapped" shows the ammount of shared (shmem/tmpfs) memory backed by 503huge pages. 504 505"Shared_Hugetlb" and "Private_Hugetlb" show the ammounts of memory backed by 506hugetlbfs page which is *not* counted in "RSS" or "PSS" field for historical 507reasons. And these are not included in {Shared,Private}_{Clean,Dirty} field. 508 509"Swap" shows how much would-be-anonymous memory is also used, but out on swap. 510 511For shmem mappings, "Swap" includes also the size of the mapped (and not 512replaced by copy-on-write) part of the underlying shmem object out on swap. 513"SwapPss" shows proportional swap share of this mapping. Unlike "Swap", this 514does not take into account swapped out page of underlying shmem objects. 515"Locked" indicates whether the mapping is locked in memory or not. 516"THPeligible" indicates whether the mapping is eligible for allocating THP 517pages - 1 if true, 0 otherwise. It just shows the current status. 518 519"VmFlags" field deserves a separate description. This member represents the 520kernel flags associated with the particular virtual memory area in two letter 521encoded manner. The codes are the following: 522 523 == ======================================= 524 rd readable 525 wr writeable 526 ex executable 527 sh shared 528 mr may read 529 mw may write 530 me may execute 531 ms may share 532 gd stack segment growns down 533 pf pure PFN range 534 dw disabled write to the mapped file 535 lo pages are locked in memory 536 io memory mapped I/O area 537 sr sequential read advise provided 538 rr random read advise provided 539 dc do not copy area on fork 540 de do not expand area on remapping 541 ac area is accountable 542 nr swap space is not reserved for the area 543 ht area uses huge tlb pages 544 ar architecture specific flag 545 dd do not include area into core dump 546 sd soft dirty flag 547 mm mixed map area 548 hg huge page advise flag 549 nh no huge page advise flag 550 mg mergable advise flag 551 bt arm64 BTI guarded page 552 mt arm64 MTE allocation tags are enabled 553 == ======================================= 554 555Note that there is no guarantee that every flag and associated mnemonic will 556be present in all further kernel releases. Things get changed, the flags may 557be vanished or the reverse -- new added. Interpretation of their meaning 558might change in future as well. So each consumer of these flags has to 559follow each specific kernel version for the exact semantic. 560 561The "Name" field will only be present on a mapping that has been named by 562userspace, and will show the name passed in by userspace. 563 564This file is only present if the CONFIG_MMU kernel configuration option is 565enabled. 566 567Note: reading /proc/PID/maps or /proc/PID/smaps is inherently racy (consistent 568output can be achieved only in the single read call). 569 570This typically manifests when doing partial reads of these files while the 571memory map is being modified. Despite the races, we do provide the following 572guarantees: 573 5741) The mapped addresses never go backwards, which implies no two 575 regions will ever overlap. 5762) If there is something at a given vaddr during the entirety of the 577 life of the smaps/maps walk, there will be some output for it. 578 579The /proc/PID/smaps_rollup file includes the same fields as /proc/PID/smaps, 580but their values are the sums of the corresponding values for all mappings of 581the process. Additionally, it contains these fields: 582 583- Pss_Anon 584- Pss_File 585- Pss_Shmem 586 587They represent the proportional shares of anonymous, file, and shmem pages, as 588described for smaps above. These fields are omitted in smaps since each 589mapping identifies the type (anon, file, or shmem) of all pages it contains. 590Thus all information in smaps_rollup can be derived from smaps, but at a 591significantly higher cost. 592 593The /proc/PID/clear_refs is used to reset the PG_Referenced and ACCESSED/YOUNG 594bits on both physical and virtual pages associated with a process, and the 595soft-dirty bit on pte (see Documentation/admin-guide/mm/soft-dirty.rst 596for details). 597To clear the bits for all the pages associated with the process:: 598 599 > echo 1 > /proc/PID/clear_refs 600 601To clear the bits for the anonymous pages associated with the process:: 602 603 > echo 2 > /proc/PID/clear_refs 604 605To clear the bits for the file mapped pages associated with the process:: 606 607 > echo 3 > /proc/PID/clear_refs 608 609To clear the soft-dirty bit:: 610 611 > echo 4 > /proc/PID/clear_refs 612 613To reset the peak resident set size ("high water mark") to the process's 614current value:: 615 616 > echo 5 > /proc/PID/clear_refs 617 618Any other value written to /proc/PID/clear_refs will have no effect. 619 620The /proc/pid/pagemap gives the PFN, which can be used to find the pageflags 621using /proc/kpageflags and number of times a page is mapped using 622/proc/kpagecount. For detailed explanation, see 623Documentation/admin-guide/mm/pagemap.rst. 624 625The /proc/pid/numa_maps is an extension based on maps, showing the memory 626locality and binding policy, as well as the memory usage (in pages) of 627each mapping. The output follows a general format where mapping details get 628summarized separated by blank spaces, one mapping per each file line:: 629 630 address policy mapping details 631 632 00400000 default file=/usr/local/bin/app mapped=1 active=0 N3=1 kernelpagesize_kB=4 633 00600000 default file=/usr/local/bin/app anon=1 dirty=1 N3=1 kernelpagesize_kB=4 634 3206000000 default file=/lib64/ld-2.12.so mapped=26 mapmax=6 N0=24 N3=2 kernelpagesize_kB=4 635 320621f000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 636 3206220000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 637 3206221000 default anon=1 dirty=1 N3=1 kernelpagesize_kB=4 638 3206800000 default file=/lib64/libc-2.12.so mapped=59 mapmax=21 active=55 N0=41 N3=18 kernelpagesize_kB=4 639 320698b000 default file=/lib64/libc-2.12.so 640 3206b8a000 default file=/lib64/libc-2.12.so anon=2 dirty=2 N3=2 kernelpagesize_kB=4 641 3206b8e000 default file=/lib64/libc-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 642 3206b8f000 default anon=3 dirty=3 active=1 N3=3 kernelpagesize_kB=4 643 7f4dc10a2000 default anon=3 dirty=3 N3=3 kernelpagesize_kB=4 644 7f4dc10b4000 default anon=2 dirty=2 active=1 N3=2 kernelpagesize_kB=4 645 7f4dc1200000 default file=/anon_hugepage\040(deleted) huge anon=1 dirty=1 N3=1 kernelpagesize_kB=2048 646 7fff335f0000 default stack anon=3 dirty=3 N3=3 kernelpagesize_kB=4 647 7fff3369d000 default mapped=1 mapmax=35 active=0 N3=1 kernelpagesize_kB=4 648 649Where: 650 651"address" is the starting address for the mapping; 652 653"policy" reports the NUMA memory policy set for the mapping (see Documentation/admin-guide/mm/numa_memory_policy.rst); 654 655"mapping details" summarizes mapping data such as mapping type, page usage counters, 656node locality page counters (N0 == node0, N1 == node1, ...) and the kernel page 657size, in KB, that is backing the mapping up. 658 6591.2 Kernel data 660--------------- 661 662Similar to the process entries, the kernel data files give information about 663the running kernel. The files used to obtain this information are contained in 664/proc and are listed in Table 1-5. Not all of these will be present in your 665system. It depends on the kernel configuration and the loaded modules, which 666files are there, and which are missing. 667 668.. table:: Table 1-5: Kernel info in /proc 669 670 ============ =============================================================== 671 File Content 672 ============ =============================================================== 673 apm Advanced power management info 674 buddyinfo Kernel memory allocator information (see text) (2.5) 675 bus Directory containing bus specific information 676 cmdline Kernel command line 677 cpuinfo Info about the CPU 678 devices Available devices (block and character) 679 dma Used DMS channels 680 filesystems Supported filesystems 681 driver Various drivers grouped here, currently rtc (2.4) 682 execdomains Execdomains, related to security (2.4) 683 fb Frame Buffer devices (2.4) 684 fs File system parameters, currently nfs/exports (2.4) 685 ide Directory containing info about the IDE subsystem 686 interrupts Interrupt usage 687 iomem Memory map (2.4) 688 ioports I/O port usage 689 irq Masks for irq to cpu affinity (2.4)(smp?) 690 isapnp ISA PnP (Plug&Play) Info (2.4) 691 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4)) 692 kmsg Kernel messages 693 ksyms Kernel symbol table 694 loadavg Load average of last 1, 5 & 15 minutes 695 locks Kernel locks 696 meminfo Memory info 697 misc Miscellaneous 698 modules List of loaded modules 699 mounts Mounted filesystems 700 net Networking info (see text) 701 pagetypeinfo Additional page allocator information (see text) (2.5) 702 partitions Table of partitions known to the system 703 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/, 704 decoupled by lspci (2.4) 705 rtc Real time clock 706 scsi SCSI info (see text) 707 slabinfo Slab pool info 708 softirqs softirq usage 709 stat Overall statistics 710 swaps Swap space utilization 711 sys See chapter 2 712 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4) 713 tty Info of tty drivers 714 uptime Wall clock since boot, combined idle time of all cpus 715 version Kernel version 716 video bttv info of video resources (2.4) 717 vmallocinfo Show vmalloced areas 718 ============ =============================================================== 719 720You can, for example, check which interrupts are currently in use and what 721they are used for by looking in the file /proc/interrupts:: 722 723 > cat /proc/interrupts 724 CPU0 725 0: 8728810 XT-PIC timer 726 1: 895 XT-PIC keyboard 727 2: 0 XT-PIC cascade 728 3: 531695 XT-PIC aha152x 729 4: 2014133 XT-PIC serial 730 5: 44401 XT-PIC pcnet_cs 731 8: 2 XT-PIC rtc 732 11: 8 XT-PIC i82365 733 12: 182918 XT-PIC PS/2 Mouse 734 13: 1 XT-PIC fpu 735 14: 1232265 XT-PIC ide0 736 15: 7 XT-PIC ide1 737 NMI: 0 738 739In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the 740output of a SMP machine):: 741 742 > cat /proc/interrupts 743 744 CPU0 CPU1 745 0: 1243498 1214548 IO-APIC-edge timer 746 1: 8949 8958 IO-APIC-edge keyboard 747 2: 0 0 XT-PIC cascade 748 5: 11286 10161 IO-APIC-edge soundblaster 749 8: 1 0 IO-APIC-edge rtc 750 9: 27422 27407 IO-APIC-edge 3c503 751 12: 113645 113873 IO-APIC-edge PS/2 Mouse 752 13: 0 0 XT-PIC fpu 753 14: 22491 24012 IO-APIC-edge ide0 754 15: 2183 2415 IO-APIC-edge ide1 755 17: 30564 30414 IO-APIC-level eth0 756 18: 177 164 IO-APIC-level bttv 757 NMI: 2457961 2457959 758 LOC: 2457882 2457881 759 ERR: 2155 760 761NMI is incremented in this case because every timer interrupt generates a NMI 762(Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups. 763 764LOC is the local interrupt counter of the internal APIC of every CPU. 765 766ERR is incremented in the case of errors in the IO-APIC bus (the bus that 767connects the CPUs in a SMP system. This means that an error has been detected, 768the IO-APIC automatically retry the transmission, so it should not be a big 769problem, but you should read the SMP-FAQ. 770 771In 2.6.2* /proc/interrupts was expanded again. This time the goal was for 772/proc/interrupts to display every IRQ vector in use by the system, not 773just those considered 'most important'. The new vectors are: 774 775THR 776 interrupt raised when a machine check threshold counter 777 (typically counting ECC corrected errors of memory or cache) exceeds 778 a configurable threshold. Only available on some systems. 779 780TRM 781 a thermal event interrupt occurs when a temperature threshold 782 has been exceeded for the CPU. This interrupt may also be generated 783 when the temperature drops back to normal. 784 785SPU 786 a spurious interrupt is some interrupt that was raised then lowered 787 by some IO device before it could be fully processed by the APIC. Hence 788 the APIC sees the interrupt but does not know what device it came from. 789 For this case the APIC will generate the interrupt with a IRQ vector 790 of 0xff. This might also be generated by chipset bugs. 791 792RES, CAL, TLB 793 rescheduling, call and TLB flush interrupts are 794 sent from one CPU to another per the needs of the OS. Typically, 795 their statistics are used by kernel developers and interested users to 796 determine the occurrence of interrupts of the given type. 797 798The above IRQ vectors are displayed only when relevant. For example, 799the threshold vector does not exist on x86_64 platforms. Others are 800suppressed when the system is a uniprocessor. As of this writing, only 801i386 and x86_64 platforms support the new IRQ vector displays. 802 803Of some interest is the introduction of the /proc/irq directory to 2.4. 804It could be used to set IRQ to CPU affinity. This means that you can "hook" an 805IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the 806irq subdir is one subdir for each IRQ, and two files; default_smp_affinity and 807prof_cpu_mask. 808 809For example:: 810 811 > ls /proc/irq/ 812 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask 813 1 11 13 15 17 19 3 5 7 9 default_smp_affinity 814 > ls /proc/irq/0/ 815 smp_affinity 816 817smp_affinity is a bitmask, in which you can specify which CPUs can handle the 818IRQ. You can set it by doing:: 819 820 > echo 1 > /proc/irq/10/smp_affinity 821 822This means that only the first CPU will handle the IRQ, but you can also echo 8235 which means that only the first and third CPU can handle the IRQ. 824 825The contents of each smp_affinity file is the same by default:: 826 827 > cat /proc/irq/0/smp_affinity 828 ffffffff 829 830There is an alternate interface, smp_affinity_list which allows specifying 831a CPU range instead of a bitmask:: 832 833 > cat /proc/irq/0/smp_affinity_list 834 1024-1031 835 836The default_smp_affinity mask applies to all non-active IRQs, which are the 837IRQs which have not yet been allocated/activated, and hence which lack a 838/proc/irq/[0-9]* directory. 839 840The node file on an SMP system shows the node to which the device using the IRQ 841reports itself as being attached. This hardware locality information does not 842include information about any possible driver locality preference. 843 844prof_cpu_mask specifies which CPUs are to be profiled by the system wide 845profiler. Default value is ffffffff (all CPUs if there are only 32 of them). 846 847The way IRQs are routed is handled by the IO-APIC, and it's Round Robin 848between all the CPUs which are allowed to handle it. As usual the kernel has 849more info than you and does a better job than you, so the defaults are the 850best choice for almost everyone. [Note this applies only to those IO-APIC's 851that support "Round Robin" interrupt distribution.] 852 853There are three more important subdirectories in /proc: net, scsi, and sys. 854The general rule is that the contents, or even the existence of these 855directories, depend on your kernel configuration. If SCSI is not enabled, the 856directory scsi may not exist. The same is true with the net, which is there 857only when networking support is present in the running kernel. 858 859The slabinfo file gives information about memory usage at the slab level. 860Linux uses slab pools for memory management above page level in version 2.2. 861Commonly used objects have their own slab pool (such as network buffers, 862directory cache, and so on). 863 864:: 865 866 > cat /proc/buddyinfo 867 868 Node 0, zone DMA 0 4 5 4 4 3 ... 869 Node 0, zone Normal 1 0 0 1 101 8 ... 870 Node 0, zone HighMem 2 0 0 1 1 0 ... 871 872External fragmentation is a problem under some workloads, and buddyinfo is a 873useful tool for helping diagnose these problems. Buddyinfo will give you a 874clue as to how big an area you can safely allocate, or why a previous 875allocation failed. 876 877Each column represents the number of pages of a certain order which are 878available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in 879ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE 880available in ZONE_NORMAL, etc... 881 882More information relevant to external fragmentation can be found in 883pagetypeinfo:: 884 885 > cat /proc/pagetypeinfo 886 Page block order: 9 887 Pages per block: 512 888 889 Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 890 Node 0, zone DMA, type Unmovable 0 0 0 1 1 1 1 1 1 1 0 891 Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 892 Node 0, zone DMA, type Movable 1 1 2 1 2 1 1 0 1 0 2 893 Node 0, zone DMA, type Reserve 0 0 0 0 0 0 0 0 0 1 0 894 Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 895 Node 0, zone DMA32, type Unmovable 103 54 77 1 1 1 11 8 7 1 9 896 Node 0, zone DMA32, type Reclaimable 0 0 2 1 0 0 0 0 1 0 0 897 Node 0, zone DMA32, type Movable 169 152 113 91 77 54 39 13 6 1 452 898 Node 0, zone DMA32, type Reserve 1 2 2 2 2 0 1 1 1 1 0 899 Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 900 901 Number of blocks type Unmovable Reclaimable Movable Reserve Isolate 902 Node 0, zone DMA 2 0 5 1 0 903 Node 0, zone DMA32 41 6 967 2 0 904 905Fragmentation avoidance in the kernel works by grouping pages of different 906migrate types into the same contiguous regions of memory called page blocks. 907A page block is typically the size of the default hugepage size, e.g. 2MB on 908X86-64. By keeping pages grouped based on their ability to move, the kernel 909can reclaim pages within a page block to satisfy a high-order allocation. 910 911The pagetypinfo begins with information on the size of a page block. It 912then gives the same type of information as buddyinfo except broken down 913by migrate-type and finishes with details on how many page blocks of each 914type exist. 915 916If min_free_kbytes has been tuned correctly (recommendations made by hugeadm 917from libhugetlbfs https://github.com/libhugetlbfs/libhugetlbfs/), one can 918make an estimate of the likely number of huge pages that can be allocated 919at a given point in time. All the "Movable" blocks should be allocatable 920unless memory has been mlock()'d. Some of the Reclaimable blocks should 921also be allocatable although a lot of filesystem metadata may have to be 922reclaimed to achieve this. 923 924 925meminfo 926~~~~~~~ 927 928Provides information about distribution and utilization of memory. This 929varies by architecture and compile options. The following is from a 93016GB PIII, which has highmem enabled. You may not have all of these fields. 931 932:: 933 934 > cat /proc/meminfo 935 936 MemTotal: 16344972 kB 937 MemFree: 13634064 kB 938 MemAvailable: 14836172 kB 939 Buffers: 3656 kB 940 Cached: 1195708 kB 941 SwapCached: 0 kB 942 Active: 891636 kB 943 Inactive: 1077224 kB 944 HighTotal: 15597528 kB 945 HighFree: 13629632 kB 946 LowTotal: 747444 kB 947 LowFree: 4432 kB 948 SwapTotal: 0 kB 949 SwapFree: 0 kB 950 Dirty: 968 kB 951 Writeback: 0 kB 952 AnonPages: 861800 kB 953 Mapped: 280372 kB 954 Shmem: 644 kB 955 KReclaimable: 168048 kB 956 Slab: 284364 kB 957 SReclaimable: 159856 kB 958 SUnreclaim: 124508 kB 959 PageTables: 24448 kB 960 NFS_Unstable: 0 kB 961 Bounce: 0 kB 962 WritebackTmp: 0 kB 963 CommitLimit: 7669796 kB 964 Committed_AS: 100056 kB 965 VmallocTotal: 112216 kB 966 VmallocUsed: 428 kB 967 VmallocChunk: 111088 kB 968 Percpu: 62080 kB 969 HardwareCorrupted: 0 kB 970 AnonHugePages: 49152 kB 971 ShmemHugePages: 0 kB 972 ShmemPmdMapped: 0 kB 973 974MemTotal 975 Total usable RAM (i.e. physical RAM minus a few reserved 976 bits and the kernel binary code) 977MemFree 978 The sum of LowFree+HighFree 979MemAvailable 980 An estimate of how much memory is available for starting new 981 applications, without swapping. Calculated from MemFree, 982 SReclaimable, the size of the file LRU lists, and the low 983 watermarks in each zone. 984 The estimate takes into account that the system needs some 985 page cache to function well, and that not all reclaimable 986 slab will be reclaimable, due to items being in use. The 987 impact of those factors will vary from system to system. 988Buffers 989 Relatively temporary storage for raw disk blocks 990 shouldn't get tremendously large (20MB or so) 991Cached 992 in-memory cache for files read from the disk (the 993 pagecache). Doesn't include SwapCached 994SwapCached 995 Memory that once was swapped out, is swapped back in but 996 still also is in the swapfile (if memory is needed it 997 doesn't need to be swapped out AGAIN because it is already 998 in the swapfile. This saves I/O) 999Active 1000 Memory that has been used more recently and usually not 1001 reclaimed unless absolutely necessary. 1002Inactive 1003 Memory which has been less recently used. It is more 1004 eligible to be reclaimed for other purposes 1005HighTotal, HighFree 1006 Highmem is all memory above ~860MB of physical memory. 1007 Highmem areas are for use by userspace programs, or 1008 for the pagecache. The kernel must use tricks to access 1009 this memory, making it slower to access than lowmem. 1010LowTotal, LowFree 1011 Lowmem is memory which can be used for everything that 1012 highmem can be used for, but it is also available for the 1013 kernel's use for its own data structures. Among many 1014 other things, it is where everything from the Slab is 1015 allocated. Bad things happen when you're out of lowmem. 1016SwapTotal 1017 total amount of swap space available 1018SwapFree 1019 Memory which has been evicted from RAM, and is temporarily 1020 on the disk 1021Dirty 1022 Memory which is waiting to get written back to the disk 1023Writeback 1024 Memory which is actively being written back to the disk 1025AnonPages 1026 Non-file backed pages mapped into userspace page tables 1027HardwareCorrupted 1028 The amount of RAM/memory in KB, the kernel identifies as 1029 corrupted. 1030AnonHugePages 1031 Non-file backed huge pages mapped into userspace page tables 1032Mapped 1033 files which have been mmaped, such as libraries 1034Shmem 1035 Total memory used by shared memory (shmem) and tmpfs 1036ShmemHugePages 1037 Memory used by shared memory (shmem) and tmpfs allocated 1038 with huge pages 1039ShmemPmdMapped 1040 Shared memory mapped into userspace with huge pages 1041KReclaimable 1042 Kernel allocations that the kernel will attempt to reclaim 1043 under memory pressure. Includes SReclaimable (below), and other 1044 direct allocations with a shrinker. 1045Slab 1046 in-kernel data structures cache 1047SReclaimable 1048 Part of Slab, that might be reclaimed, such as caches 1049SUnreclaim 1050 Part of Slab, that cannot be reclaimed on memory pressure 1051PageTables 1052 amount of memory dedicated to the lowest level of page 1053 tables. 1054NFS_Unstable 1055 Always zero. Previous counted pages which had been written to 1056 the server, but has not been committed to stable storage. 1057Bounce 1058 Memory used for block device "bounce buffers" 1059WritebackTmp 1060 Memory used by FUSE for temporary writeback buffers 1061CommitLimit 1062 Based on the overcommit ratio ('vm.overcommit_ratio'), 1063 this is the total amount of memory currently available to 1064 be allocated on the system. This limit is only adhered to 1065 if strict overcommit accounting is enabled (mode 2 in 1066 'vm.overcommit_memory'). 1067 1068 The CommitLimit is calculated with the following formula:: 1069 1070 CommitLimit = ([total RAM pages] - [total huge TLB pages]) * 1071 overcommit_ratio / 100 + [total swap pages] 1072 1073 For example, on a system with 1G of physical RAM and 7G 1074 of swap with a `vm.overcommit_ratio` of 30 it would 1075 yield a CommitLimit of 7.3G. 1076 1077 For more details, see the memory overcommit documentation 1078 in vm/overcommit-accounting. 1079Committed_AS 1080 The amount of memory presently allocated on the system. 1081 The committed memory is a sum of all of the memory which 1082 has been allocated by processes, even if it has not been 1083 "used" by them as of yet. A process which malloc()'s 1G 1084 of memory, but only touches 300M of it will show up as 1085 using 1G. This 1G is memory which has been "committed" to 1086 by the VM and can be used at any time by the allocating 1087 application. With strict overcommit enabled on the system 1088 (mode 2 in 'vm.overcommit_memory'), allocations which would 1089 exceed the CommitLimit (detailed above) will not be permitted. 1090 This is useful if one needs to guarantee that processes will 1091 not fail due to lack of memory once that memory has been 1092 successfully allocated. 1093VmallocTotal 1094 total size of vmalloc memory area 1095VmallocUsed 1096 amount of vmalloc area which is used 1097VmallocChunk 1098 largest contiguous block of vmalloc area which is free 1099Percpu 1100 Memory allocated to the percpu allocator used to back percpu 1101 allocations. This stat excludes the cost of metadata. 1102 1103vmallocinfo 1104~~~~~~~~~~~ 1105 1106Provides information about vmalloced/vmaped areas. One line per area, 1107containing the virtual address range of the area, size in bytes, 1108caller information of the creator, and optional information depending 1109on the kind of area: 1110 1111 ========== =================================================== 1112 pages=nr number of pages 1113 phys=addr if a physical address was specified 1114 ioremap I/O mapping (ioremap() and friends) 1115 vmalloc vmalloc() area 1116 vmap vmap()ed pages 1117 user VM_USERMAP area 1118 vpages buffer for pages pointers was vmalloced (huge area) 1119 N<node>=nr (Only on NUMA kernels) 1120 Number of pages allocated on memory node <node> 1121 ========== =================================================== 1122 1123:: 1124 1125 > cat /proc/vmallocinfo 1126 0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ... 1127 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128 1128 0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ... 1129 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64 1130 0xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f... 1131 phys=7fee8000 ioremap 1132 0xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f... 1133 phys=7fee7000 ioremap 1134 0xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210 1135 0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ... 1136 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3 1137 0xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ... 1138 pages=2 vmalloc N1=2 1139 0xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ... 1140 /0x130 [x_tables] pages=4 vmalloc N0=4 1141 0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ... 1142 pages=14 vmalloc N2=14 1143 0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ... 1144 pages=4 vmalloc N1=4 1145 0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ... 1146 pages=2 vmalloc N1=2 1147 0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ... 1148 pages=10 vmalloc N0=10 1149 1150 1151softirqs 1152~~~~~~~~ 1153 1154Provides counts of softirq handlers serviced since boot time, for each CPU. 1155 1156:: 1157 1158 > cat /proc/softirqs 1159 CPU0 CPU1 CPU2 CPU3 1160 HI: 0 0 0 0 1161 TIMER: 27166 27120 27097 27034 1162 NET_TX: 0 0 0 17 1163 NET_RX: 42 0 0 39 1164 BLOCK: 0 0 107 1121 1165 TASKLET: 0 0 0 290 1166 SCHED: 27035 26983 26971 26746 1167 HRTIMER: 0 0 0 0 1168 RCU: 1678 1769 2178 2250 1169 1170 11711.3 IDE devices in /proc/ide 1172---------------------------- 1173 1174The subdirectory /proc/ide contains information about all IDE devices of which 1175the kernel is aware. There is one subdirectory for each IDE controller, the 1176file drivers and a link for each IDE device, pointing to the device directory 1177in the controller specific subtree. 1178 1179The file 'drivers' contains general information about the drivers used for the 1180IDE devices:: 1181 1182 > cat /proc/ide/drivers 1183 ide-cdrom version 4.53 1184 ide-disk version 1.08 1185 1186More detailed information can be found in the controller specific 1187subdirectories. These are named ide0, ide1 and so on. Each of these 1188directories contains the files shown in table 1-6. 1189 1190 1191.. table:: Table 1-6: IDE controller info in /proc/ide/ide? 1192 1193 ======= ======================================= 1194 File Content 1195 ======= ======================================= 1196 channel IDE channel (0 or 1) 1197 config Configuration (only for PCI/IDE bridge) 1198 mate Mate name 1199 model Type/Chipset of IDE controller 1200 ======= ======================================= 1201 1202Each device connected to a controller has a separate subdirectory in the 1203controllers directory. The files listed in table 1-7 are contained in these 1204directories. 1205 1206 1207.. table:: Table 1-7: IDE device information 1208 1209 ================ ========================================== 1210 File Content 1211 ================ ========================================== 1212 cache The cache 1213 capacity Capacity of the medium (in 512Byte blocks) 1214 driver driver and version 1215 geometry physical and logical geometry 1216 identify device identify block 1217 media media type 1218 model device identifier 1219 settings device setup 1220 smart_thresholds IDE disk management thresholds 1221 smart_values IDE disk management values 1222 ================ ========================================== 1223 1224The most interesting file is ``settings``. This file contains a nice 1225overview of the drive parameters:: 1226 1227 # cat /proc/ide/ide0/hda/settings 1228 name value min max mode 1229 ---- ----- --- --- ---- 1230 bios_cyl 526 0 65535 rw 1231 bios_head 255 0 255 rw 1232 bios_sect 63 0 63 rw 1233 breada_readahead 4 0 127 rw 1234 bswap 0 0 1 r 1235 file_readahead 72 0 2097151 rw 1236 io_32bit 0 0 3 rw 1237 keepsettings 0 0 1 rw 1238 max_kb_per_request 122 1 127 rw 1239 multcount 0 0 8 rw 1240 nice1 1 0 1 rw 1241 nowerr 0 0 1 rw 1242 pio_mode write-only 0 255 w 1243 slow 0 0 1 rw 1244 unmaskirq 0 0 1 rw 1245 using_dma 0 0 1 rw 1246 1247 12481.4 Networking info in /proc/net 1249-------------------------------- 1250 1251The subdirectory /proc/net follows the usual pattern. Table 1-8 shows the 1252additional values you get for IP version 6 if you configure the kernel to 1253support this. Table 1-9 lists the files and their meaning. 1254 1255 1256.. table:: Table 1-8: IPv6 info in /proc/net 1257 1258 ========== ===================================================== 1259 File Content 1260 ========== ===================================================== 1261 udp6 UDP sockets (IPv6) 1262 tcp6 TCP sockets (IPv6) 1263 raw6 Raw device statistics (IPv6) 1264 igmp6 IP multicast addresses, which this host joined (IPv6) 1265 if_inet6 List of IPv6 interface addresses 1266 ipv6_route Kernel routing table for IPv6 1267 rt6_stats Global IPv6 routing tables statistics 1268 sockstat6 Socket statistics (IPv6) 1269 snmp6 Snmp data (IPv6) 1270 ========== ===================================================== 1271 1272.. table:: Table 1-9: Network info in /proc/net 1273 1274 ============= ================================================================ 1275 File Content 1276 ============= ================================================================ 1277 arp Kernel ARP table 1278 dev network devices with statistics 1279 dev_mcast the Layer2 multicast groups a device is listening too 1280 (interface index, label, number of references, number of bound 1281 addresses). 1282 dev_stat network device status 1283 ip_fwchains Firewall chain linkage 1284 ip_fwnames Firewall chain names 1285 ip_masq Directory containing the masquerading tables 1286 ip_masquerade Major masquerading table 1287 netstat Network statistics 1288 raw raw device statistics 1289 route Kernel routing table 1290 rpc Directory containing rpc info 1291 rt_cache Routing cache 1292 snmp SNMP data 1293 sockstat Socket statistics 1294 tcp TCP sockets 1295 udp UDP sockets 1296 unix UNIX domain sockets 1297 wireless Wireless interface data (Wavelan etc) 1298 igmp IP multicast addresses, which this host joined 1299 psched Global packet scheduler parameters. 1300 netlink List of PF_NETLINK sockets 1301 ip_mr_vifs List of multicast virtual interfaces 1302 ip_mr_cache List of multicast routing cache 1303 ============= ================================================================ 1304 1305You can use this information to see which network devices are available in 1306your system and how much traffic was routed over those devices:: 1307 1308 > cat /proc/net/dev 1309 Inter-|Receive |[... 1310 face |bytes packets errs drop fifo frame compressed multicast|[... 1311 lo: 908188 5596 0 0 0 0 0 0 [... 1312 ppp0:15475140 20721 410 0 0 410 0 0 [... 1313 eth0: 614530 7085 0 0 0 0 0 1 [... 1314 1315 ...] Transmit 1316 ...] bytes packets errs drop fifo colls carrier compressed 1317 ...] 908188 5596 0 0 0 0 0 0 1318 ...] 1375103 17405 0 0 0 0 0 0 1319 ...] 1703981 5535 0 0 0 3 0 0 1320 1321In addition, each Channel Bond interface has its own directory. For 1322example, the bond0 device will have a directory called /proc/net/bond0/. 1323It will contain information that is specific to that bond, such as the 1324current slaves of the bond, the link status of the slaves, and how 1325many times the slaves link has failed. 1326 13271.5 SCSI info 1328------------- 1329 1330If you have a SCSI host adapter in your system, you'll find a subdirectory 1331named after the driver for this adapter in /proc/scsi. You'll also see a list 1332of all recognized SCSI devices in /proc/scsi:: 1333 1334 >cat /proc/scsi/scsi 1335 Attached devices: 1336 Host: scsi0 Channel: 00 Id: 00 Lun: 00 1337 Vendor: IBM Model: DGHS09U Rev: 03E0 1338 Type: Direct-Access ANSI SCSI revision: 03 1339 Host: scsi0 Channel: 00 Id: 06 Lun: 00 1340 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04 1341 Type: CD-ROM ANSI SCSI revision: 02 1342 1343 1344The directory named after the driver has one file for each adapter found in 1345the system. These files contain information about the controller, including 1346the used IRQ and the IO address range. The amount of information shown is 1347dependent on the adapter you use. The example shows the output for an Adaptec 1348AHA-2940 SCSI adapter:: 1349 1350 > cat /proc/scsi/aic7xxx/0 1351 1352 Adaptec AIC7xxx driver version: 5.1.19/3.2.4 1353 Compile Options: 1354 TCQ Enabled By Default : Disabled 1355 AIC7XXX_PROC_STATS : Disabled 1356 AIC7XXX_RESET_DELAY : 5 1357 Adapter Configuration: 1358 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter 1359 Ultra Wide Controller 1360 PCI MMAPed I/O Base: 0xeb001000 1361 Adapter SEEPROM Config: SEEPROM found and used. 1362 Adaptec SCSI BIOS: Enabled 1363 IRQ: 10 1364 SCBs: Active 0, Max Active 2, 1365 Allocated 15, HW 16, Page 255 1366 Interrupts: 160328 1367 BIOS Control Word: 0x18b6 1368 Adapter Control Word: 0x005b 1369 Extended Translation: Enabled 1370 Disconnect Enable Flags: 0xffff 1371 Ultra Enable Flags: 0x0001 1372 Tag Queue Enable Flags: 0x0000 1373 Ordered Queue Tag Flags: 0x0000 1374 Default Tag Queue Depth: 8 1375 Tagged Queue By Device array for aic7xxx host instance 0: 1376 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255} 1377 Actual queue depth per device for aic7xxx host instance 0: 1378 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1} 1379 Statistics: 1380 (scsi0:0:0:0) 1381 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8 1382 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0) 1383 Total transfers 160151 (74577 reads and 85574 writes) 1384 (scsi0:0:6:0) 1385 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15 1386 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0) 1387 Total transfers 0 (0 reads and 0 writes) 1388 1389 13901.6 Parallel port info in /proc/parport 1391--------------------------------------- 1392 1393The directory /proc/parport contains information about the parallel ports of 1394your system. It has one subdirectory for each port, named after the port 1395number (0,1,2,...). 1396 1397These directories contain the four files shown in Table 1-10. 1398 1399 1400.. table:: Table 1-10: Files in /proc/parport 1401 1402 ========= ==================================================================== 1403 File Content 1404 ========= ==================================================================== 1405 autoprobe Any IEEE-1284 device ID information that has been acquired. 1406 devices list of the device drivers using that port. A + will appear by the 1407 name of the device currently using the port (it might not appear 1408 against any). 1409 hardware Parallel port's base address, IRQ line and DMA channel. 1410 irq IRQ that parport is using for that port. This is in a separate 1411 file to allow you to alter it by writing a new value in (IRQ 1412 number or none). 1413 ========= ==================================================================== 1414 14151.7 TTY info in /proc/tty 1416------------------------- 1417 1418Information about the available and actually used tty's can be found in the 1419directory /proc/tty. You'll find entries for drivers and line disciplines in 1420this directory, as shown in Table 1-11. 1421 1422 1423.. table:: Table 1-11: Files in /proc/tty 1424 1425 ============= ============================================== 1426 File Content 1427 ============= ============================================== 1428 drivers list of drivers and their usage 1429 ldiscs registered line disciplines 1430 driver/serial usage statistic and status of single tty lines 1431 ============= ============================================== 1432 1433To see which tty's are currently in use, you can simply look into the file 1434/proc/tty/drivers:: 1435 1436 > cat /proc/tty/drivers 1437 pty_slave /dev/pts 136 0-255 pty:slave 1438 pty_master /dev/ptm 128 0-255 pty:master 1439 pty_slave /dev/ttyp 3 0-255 pty:slave 1440 pty_master /dev/pty 2 0-255 pty:master 1441 serial /dev/cua 5 64-67 serial:callout 1442 serial /dev/ttyS 4 64-67 serial 1443 /dev/tty0 /dev/tty0 4 0 system:vtmaster 1444 /dev/ptmx /dev/ptmx 5 2 system 1445 /dev/console /dev/console 5 1 system:console 1446 /dev/tty /dev/tty 5 0 system:/dev/tty 1447 unknown /dev/tty 4 1-63 console 1448 1449 14501.8 Miscellaneous kernel statistics in /proc/stat 1451------------------------------------------------- 1452 1453Various pieces of information about kernel activity are available in the 1454/proc/stat file. All of the numbers reported in this file are aggregates 1455since the system first booted. For a quick look, simply cat the file:: 1456 1457 > cat /proc/stat 1458 cpu 2255 34 2290 22625563 6290 127 456 0 0 0 1459 cpu0 1132 34 1441 11311718 3675 127 438 0 0 0 1460 cpu1 1123 0 849 11313845 2614 0 18 0 0 0 1461 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...] 1462 ctxt 1990473 1463 btime 1062191376 1464 processes 2915 1465 procs_running 1 1466 procs_blocked 0 1467 softirq 183433 0 21755 12 39 1137 231 21459 2263 1468 1469The very first "cpu" line aggregates the numbers in all of the other "cpuN" 1470lines. These numbers identify the amount of time the CPU has spent performing 1471different kinds of work. Time units are in USER_HZ (typically hundredths of a 1472second). The meanings of the columns are as follows, from left to right: 1473 1474- user: normal processes executing in user mode 1475- nice: niced processes executing in user mode 1476- system: processes executing in kernel mode 1477- idle: twiddling thumbs 1478- iowait: In a word, iowait stands for waiting for I/O to complete. But there 1479 are several problems: 1480 1481 1. CPU will not wait for I/O to complete, iowait is the time that a task is 1482 waiting for I/O to complete. When CPU goes into idle state for 1483 outstanding task I/O, another task will be scheduled on this CPU. 1484 2. In a multi-core CPU, the task waiting for I/O to complete is not running 1485 on any CPU, so the iowait of each CPU is difficult to calculate. 1486 3. The value of iowait field in /proc/stat will decrease in certain 1487 conditions. 1488 1489 So, the iowait is not reliable by reading from /proc/stat. 1490- irq: servicing interrupts 1491- softirq: servicing softirqs 1492- steal: involuntary wait 1493- guest: running a normal guest 1494- guest_nice: running a niced guest 1495 1496The "intr" line gives counts of interrupts serviced since boot time, for each 1497of the possible system interrupts. The first column is the total of all 1498interrupts serviced including unnumbered architecture specific interrupts; 1499each subsequent column is the total for that particular numbered interrupt. 1500Unnumbered interrupts are not shown, only summed into the total. 1501 1502The "ctxt" line gives the total number of context switches across all CPUs. 1503 1504The "btime" line gives the time at which the system booted, in seconds since 1505the Unix epoch. 1506 1507The "processes" line gives the number of processes and threads created, which 1508includes (but is not limited to) those created by calls to the fork() and 1509clone() system calls. 1510 1511The "procs_running" line gives the total number of threads that are 1512running or ready to run (i.e., the total number of runnable threads). 1513 1514The "procs_blocked" line gives the number of processes currently blocked, 1515waiting for I/O to complete. 1516 1517The "softirq" line gives counts of softirqs serviced since boot time, for each 1518of the possible system softirqs. The first column is the total of all 1519softirqs serviced; each subsequent column is the total for that particular 1520softirq. 1521 1522 15231.9 Ext4 file system parameters 1524------------------------------- 1525 1526Information about mounted ext4 file systems can be found in 1527/proc/fs/ext4. Each mounted filesystem will have a directory in 1528/proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or 1529/proc/fs/ext4/dm-0). The files in each per-device directory are shown 1530in Table 1-12, below. 1531 1532.. table:: Table 1-12: Files in /proc/fs/ext4/<devname> 1533 1534 ============== ========================================================== 1535 File Content 1536 mb_groups details of multiblock allocator buddy cache of free blocks 1537 ============== ========================================================== 1538 15391.10 /proc/consoles 1540------------------- 1541Shows registered system console lines. 1542 1543To see which character device lines are currently used for the system console 1544/dev/console, you may simply look into the file /proc/consoles:: 1545 1546 > cat /proc/consoles 1547 tty0 -WU (ECp) 4:7 1548 ttyS0 -W- (Ep) 4:64 1549 1550The columns are: 1551 1552+--------------------+-------------------------------------------------------+ 1553| device | name of the device | 1554+====================+=======================================================+ 1555| operations | * R = can do read operations | 1556| | * W = can do write operations | 1557| | * U = can do unblank | 1558+--------------------+-------------------------------------------------------+ 1559| flags | * E = it is enabled | 1560| | * C = it is preferred console | 1561| | * B = it is primary boot console | 1562| | * p = it is used for printk buffer | 1563| | * b = it is not a TTY but a Braille device | 1564| | * a = it is safe to use when cpu is offline | 1565+--------------------+-------------------------------------------------------+ 1566| major:minor | major and minor number of the device separated by a | 1567| | colon | 1568+--------------------+-------------------------------------------------------+ 1569 1570Summary 1571------- 1572 1573The /proc file system serves information about the running system. It not only 1574allows access to process data but also allows you to request the kernel status 1575by reading files in the hierarchy. 1576 1577The directory structure of /proc reflects the types of information and makes 1578it easy, if not obvious, where to look for specific data. 1579 1580Chapter 2: Modifying System Parameters 1581====================================== 1582 1583In This Chapter 1584--------------- 1585 1586* Modifying kernel parameters by writing into files found in /proc/sys 1587* Exploring the files which modify certain parameters 1588* Review of the /proc/sys file tree 1589 1590------------------------------------------------------------------------------ 1591 1592A very interesting part of /proc is the directory /proc/sys. This is not only 1593a source of information, it also allows you to change parameters within the 1594kernel. Be very careful when attempting this. You can optimize your system, 1595but you can also cause it to crash. Never alter kernel parameters on a 1596production system. Set up a development machine and test to make sure that 1597everything works the way you want it to. You may have no alternative but to 1598reboot the machine once an error has been made. 1599 1600To change a value, simply echo the new value into the file. 1601You need to be root to do this. You can create your own boot script 1602to perform this every time your system boots. 1603 1604The files in /proc/sys can be used to fine tune and monitor miscellaneous and 1605general things in the operation of the Linux kernel. Since some of the files 1606can inadvertently disrupt your system, it is advisable to read both 1607documentation and source before actually making adjustments. In any case, be 1608very careful when writing to any of these files. The entries in /proc may 1609change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt 1610review the kernel documentation in the directory /usr/src/linux/Documentation. 1611This chapter is heavily based on the documentation included in the pre 2.2 1612kernels, and became part of it in version 2.2.1 of the Linux kernel. 1613 1614Please see: Documentation/admin-guide/sysctl/ directory for descriptions of these 1615entries. 1616 1617Summary 1618------- 1619 1620Certain aspects of kernel behavior can be modified at runtime, without the 1621need to recompile the kernel, or even to reboot the system. The files in the 1622/proc/sys tree can not only be read, but also modified. You can use the echo 1623command to write value into these files, thereby changing the default settings 1624of the kernel. 1625 1626 1627Chapter 3: Per-process Parameters 1628================================= 1629 16303.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj- Adjust the oom-killer score 1631-------------------------------------------------------------------------------- 1632 1633These files can be used to adjust the badness heuristic used to select which 1634process gets killed in out of memory (oom) conditions. 1635 1636The badness heuristic assigns a value to each candidate task ranging from 0 1637(never kill) to 1000 (always kill) to determine which process is targeted. The 1638units are roughly a proportion along that range of allowed memory the process 1639may allocate from based on an estimation of its current memory and swap use. 1640For example, if a task is using all allowed memory, its badness score will be 16411000. If it is using half of its allowed memory, its score will be 500. 1642 1643The amount of "allowed" memory depends on the context in which the oom killer 1644was called. If it is due to the memory assigned to the allocating task's cpuset 1645being exhausted, the allowed memory represents the set of mems assigned to that 1646cpuset. If it is due to a mempolicy's node(s) being exhausted, the allowed 1647memory represents the set of mempolicy nodes. If it is due to a memory 1648limit (or swap limit) being reached, the allowed memory is that configured 1649limit. Finally, if it is due to the entire system being out of memory, the 1650allowed memory represents all allocatable resources. 1651 1652The value of /proc/<pid>/oom_score_adj is added to the badness score before it 1653is used to determine which task to kill. Acceptable values range from -1000 1654(OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_MAX). This allows userspace to 1655polarize the preference for oom killing either by always preferring a certain 1656task or completely disabling it. The lowest possible value, -1000, is 1657equivalent to disabling oom killing entirely for that task since it will always 1658report a badness score of 0. 1659 1660Consequently, it is very simple for userspace to define the amount of memory to 1661consider for each task. Setting a /proc/<pid>/oom_score_adj value of +500, for 1662example, is roughly equivalent to allowing the remainder of tasks sharing the 1663same system, cpuset, mempolicy, or memory controller resources to use at least 166450% more memory. A value of -500, on the other hand, would be roughly 1665equivalent to discounting 50% of the task's allowed memory from being considered 1666as scoring against the task. 1667 1668For backwards compatibility with previous kernels, /proc/<pid>/oom_adj may also 1669be used to tune the badness score. Its acceptable values range from -16 1670(OOM_ADJUST_MIN) to +15 (OOM_ADJUST_MAX) and a special value of -17 1671(OOM_DISABLE) to disable oom killing entirely for that task. Its value is 1672scaled linearly with /proc/<pid>/oom_score_adj. 1673 1674The value of /proc/<pid>/oom_score_adj may be reduced no lower than the last 1675value set by a CAP_SYS_RESOURCE process. To reduce the value any lower 1676requires CAP_SYS_RESOURCE. 1677 1678 16793.2 /proc/<pid>/oom_score - Display current oom-killer score 1680------------------------------------------------------------- 1681 1682This file can be used to check the current score used by the oom-killer for 1683any given <pid>. Use it together with /proc/<pid>/oom_score_adj to tune which 1684process should be killed in an out-of-memory situation. 1685 1686Please note that the exported value includes oom_score_adj so it is 1687effectively in range [0,2000]. 1688 1689 16903.3 /proc/<pid>/io - Display the IO accounting fields 1691------------------------------------------------------- 1692 1693This file contains IO statistics for each running process. 1694 1695Example 1696~~~~~~~ 1697 1698:: 1699 1700 test:/tmp # dd if=/dev/zero of=/tmp/test.dat & 1701 [1] 3828 1702 1703 test:/tmp # cat /proc/3828/io 1704 rchar: 323934931 1705 wchar: 323929600 1706 syscr: 632687 1707 syscw: 632675 1708 read_bytes: 0 1709 write_bytes: 323932160 1710 cancelled_write_bytes: 0 1711 1712 1713Description 1714~~~~~~~~~~~ 1715 1716rchar 1717^^^^^ 1718 1719I/O counter: chars read 1720The number of bytes which this task has caused to be read from storage. This 1721is simply the sum of bytes which this process passed to read() and pread(). 1722It includes things like tty IO and it is unaffected by whether or not actual 1723physical disk IO was required (the read might have been satisfied from 1724pagecache). 1725 1726 1727wchar 1728^^^^^ 1729 1730I/O counter: chars written 1731The number of bytes which this task has caused, or shall cause to be written 1732to disk. Similar caveats apply here as with rchar. 1733 1734 1735syscr 1736^^^^^ 1737 1738I/O counter: read syscalls 1739Attempt to count the number of read I/O operations, i.e. syscalls like read() 1740and pread(). 1741 1742 1743syscw 1744^^^^^ 1745 1746I/O counter: write syscalls 1747Attempt to count the number of write I/O operations, i.e. syscalls like 1748write() and pwrite(). 1749 1750 1751read_bytes 1752^^^^^^^^^^ 1753 1754I/O counter: bytes read 1755Attempt to count the number of bytes which this process really did cause to 1756be fetched from the storage layer. Done at the submit_bio() level, so it is 1757accurate for block-backed filesystems. <please add status regarding NFS and 1758CIFS at a later time> 1759 1760 1761write_bytes 1762^^^^^^^^^^^ 1763 1764I/O counter: bytes written 1765Attempt to count the number of bytes which this process caused to be sent to 1766the storage layer. This is done at page-dirtying time. 1767 1768 1769cancelled_write_bytes 1770^^^^^^^^^^^^^^^^^^^^^ 1771 1772The big inaccuracy here is truncate. If a process writes 1MB to a file and 1773then deletes the file, it will in fact perform no writeout. But it will have 1774been accounted as having caused 1MB of write. 1775In other words: The number of bytes which this process caused to not happen, 1776by truncating pagecache. A task can cause "negative" IO too. If this task 1777truncates some dirty pagecache, some IO which another task has been accounted 1778for (in its write_bytes) will not be happening. We _could_ just subtract that 1779from the truncating task's write_bytes, but there is information loss in doing 1780that. 1781 1782 1783.. Note:: 1784 1785 At its current implementation state, this is a bit racy on 32-bit machines: 1786 if process A reads process B's /proc/pid/io while process B is updating one 1787 of those 64-bit counters, process A could see an intermediate result. 1788 1789 1790More information about this can be found within the taskstats documentation in 1791Documentation/accounting. 1792 17933.4 /proc/<pid>/coredump_filter - Core dump filtering settings 1794--------------------------------------------------------------- 1795When a process is dumped, all anonymous memory is written to a core file as 1796long as the size of the core file isn't limited. But sometimes we don't want 1797to dump some memory segments, for example, huge shared memory or DAX. 1798Conversely, sometimes we want to save file-backed memory segments into a core 1799file, not only the individual files. 1800 1801/proc/<pid>/coredump_filter allows you to customize which memory segments 1802will be dumped when the <pid> process is dumped. coredump_filter is a bitmask 1803of memory types. If a bit of the bitmask is set, memory segments of the 1804corresponding memory type are dumped, otherwise they are not dumped. 1805 1806The following 9 memory types are supported: 1807 1808 - (bit 0) anonymous private memory 1809 - (bit 1) anonymous shared memory 1810 - (bit 2) file-backed private memory 1811 - (bit 3) file-backed shared memory 1812 - (bit 4) ELF header pages in file-backed private memory areas (it is 1813 effective only if the bit 2 is cleared) 1814 - (bit 5) hugetlb private memory 1815 - (bit 6) hugetlb shared memory 1816 - (bit 7) DAX private memory 1817 - (bit 8) DAX shared memory 1818 1819 Note that MMIO pages such as frame buffer are never dumped and vDSO pages 1820 are always dumped regardless of the bitmask status. 1821 1822 Note that bits 0-4 don't affect hugetlb or DAX memory. hugetlb memory is 1823 only affected by bit 5-6, and DAX is only affected by bits 7-8. 1824 1825The default value of coredump_filter is 0x33; this means all anonymous memory 1826segments, ELF header pages and hugetlb private memory are dumped. 1827 1828If you don't want to dump all shared memory segments attached to pid 1234, 1829write 0x31 to the process's proc file:: 1830 1831 $ echo 0x31 > /proc/1234/coredump_filter 1832 1833When a new process is created, the process inherits the bitmask status from its 1834parent. It is useful to set up coredump_filter before the program runs. 1835For example:: 1836 1837 $ echo 0x7 > /proc/self/coredump_filter 1838 $ ./some_program 1839 18403.5 /proc/<pid>/mountinfo - Information about mounts 1841-------------------------------------------------------- 1842 1843This file contains lines of the form:: 1844 1845 36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue 1846 (1)(2)(3) (4) (5) (6) (7) (8) (9) (10) (11) 1847 1848 (1) mount ID: unique identifier of the mount (may be reused after umount) 1849 (2) parent ID: ID of parent (or of self for the top of the mount tree) 1850 (3) major:minor: value of st_dev for files on filesystem 1851 (4) root: root of the mount within the filesystem 1852 (5) mount point: mount point relative to the process's root 1853 (6) mount options: per mount options 1854 (7) optional fields: zero or more fields of the form "tag[:value]" 1855 (8) separator: marks the end of the optional fields 1856 (9) filesystem type: name of filesystem of the form "type[.subtype]" 1857 (10) mount source: filesystem specific information or "none" 1858 (11) super options: per super block options 1859 1860Parsers should ignore all unrecognised optional fields. Currently the 1861possible optional fields are: 1862 1863================ ============================================================== 1864shared:X mount is shared in peer group X 1865master:X mount is slave to peer group X 1866propagate_from:X mount is slave and receives propagation from peer group X [#]_ 1867unbindable mount is unbindable 1868================ ============================================================== 1869 1870.. [#] X is the closest dominant peer group under the process's root. If 1871 X is the immediate master of the mount, or if there's no dominant peer 1872 group under the same root, then only the "master:X" field is present 1873 and not the "propagate_from:X" field. 1874 1875For more information on mount propagation see: 1876 1877 Documentation/filesystems/sharedsubtree.rst 1878 1879 18803.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm 1881-------------------------------------------------------- 1882These files provide a method to access a task's comm value. It also allows for 1883a task to set its own or one of its thread siblings comm value. The comm value 1884is limited in size compared to the cmdline value, so writing anything longer 1885then the kernel's TASK_COMM_LEN (currently 16 chars) will result in a truncated 1886comm value. 1887 1888 18893.7 /proc/<pid>/task/<tid>/children - Information about task children 1890------------------------------------------------------------------------- 1891This file provides a fast way to retrieve first level children pids 1892of a task pointed by <pid>/<tid> pair. The format is a space separated 1893stream of pids. 1894 1895Note the "first level" here -- if a child has its own children they will 1896not be listed here; one needs to read /proc/<children-pid>/task/<tid>/children 1897to obtain the descendants. 1898 1899Since this interface is intended to be fast and cheap it doesn't 1900guarantee to provide precise results and some children might be 1901skipped, especially if they've exited right after we printed their 1902pids, so one needs to either stop or freeze processes being inspected 1903if precise results are needed. 1904 1905 19063.8 /proc/<pid>/fdinfo/<fd> - Information about opened file 1907--------------------------------------------------------------- 1908This file provides information associated with an opened file. The regular 1909files have at least four fields -- 'pos', 'flags', 'mnt_id' and 'ino'. 1910The 'pos' represents the current offset of the opened file in decimal 1911form [see lseek(2) for details], 'flags' denotes the octal O_xxx mask the 1912file has been created with [see open(2) for details] and 'mnt_id' represents 1913mount ID of the file system containing the opened file [see 3.5 1914/proc/<pid>/mountinfo for details]. 'ino' represents the inode number of 1915the file. 1916 1917A typical output is:: 1918 1919 pos: 0 1920 flags: 0100002 1921 mnt_id: 19 1922 ino: 63107 1923 1924All locks associated with a file descriptor are shown in its fdinfo too:: 1925 1926 lock: 1: FLOCK ADVISORY WRITE 359 00:13:11691 0 EOF 1927 1928The files such as eventfd, fsnotify, signalfd, epoll among the regular pos/flags 1929pair provide additional information particular to the objects they represent. 1930 1931Eventfd files 1932~~~~~~~~~~~~~ 1933 1934:: 1935 1936 pos: 0 1937 flags: 04002 1938 mnt_id: 9 1939 ino: 63107 1940 eventfd-count: 5a 1941 1942where 'eventfd-count' is hex value of a counter. 1943 1944Signalfd files 1945~~~~~~~~~~~~~~ 1946 1947:: 1948 1949 pos: 0 1950 flags: 04002 1951 mnt_id: 9 1952 ino: 63107 1953 sigmask: 0000000000000200 1954 1955where 'sigmask' is hex value of the signal mask associated 1956with a file. 1957 1958Epoll files 1959~~~~~~~~~~~ 1960 1961:: 1962 1963 pos: 0 1964 flags: 02 1965 mnt_id: 9 1966 ino: 63107 1967 tfd: 5 events: 1d data: ffffffffffffffff pos:0 ino:61af sdev:7 1968 1969where 'tfd' is a target file descriptor number in decimal form, 1970'events' is events mask being watched and the 'data' is data 1971associated with a target [see epoll(7) for more details]. 1972 1973The 'pos' is current offset of the target file in decimal form 1974[see lseek(2)], 'ino' and 'sdev' are inode and device numbers 1975where target file resides, all in hex format. 1976 1977Fsnotify files 1978~~~~~~~~~~~~~~ 1979For inotify files the format is the following:: 1980 1981 pos: 0 1982 flags: 02000000 1983 mnt_id: 9 1984 ino: 63107 1985 inotify wd:3 ino:9e7e sdev:800013 mask:800afce ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:7e9e0000640d1b6d 1986 1987where 'wd' is a watch descriptor in decimal form, i.e. a target file 1988descriptor number, 'ino' and 'sdev' are inode and device where the 1989target file resides and the 'mask' is the mask of events, all in hex 1990form [see inotify(7) for more details]. 1991 1992If the kernel was built with exportfs support, the path to the target 1993file is encoded as a file handle. The file handle is provided by three 1994fields 'fhandle-bytes', 'fhandle-type' and 'f_handle', all in hex 1995format. 1996 1997If the kernel is built without exportfs support the file handle won't be 1998printed out. 1999 2000If there is no inotify mark attached yet the 'inotify' line will be omitted. 2001 2002For fanotify files the format is:: 2003 2004 pos: 0 2005 flags: 02 2006 mnt_id: 9 2007 ino: 63107 2008 fanotify flags:10 event-flags:0 2009 fanotify mnt_id:12 mflags:40 mask:38 ignored_mask:40000003 2010 fanotify ino:4f969 sdev:800013 mflags:0 mask:3b ignored_mask:40000000 fhandle-bytes:8 fhandle-type:1 f_handle:69f90400c275b5b4 2011 2012where fanotify 'flags' and 'event-flags' are values used in fanotify_init 2013call, 'mnt_id' is the mount point identifier, 'mflags' is the value of 2014flags associated with mark which are tracked separately from events 2015mask. 'ino' and 'sdev' are target inode and device, 'mask' is the events 2016mask and 'ignored_mask' is the mask of events which are to be ignored. 2017All are in hex format. Incorporation of 'mflags', 'mask' and 'ignored_mask' 2018provide information about flags and mask used in fanotify_mark 2019call [see fsnotify manpage for details]. 2020 2021While the first three lines are mandatory and always printed, the rest is 2022optional and may be omitted if no marks created yet. 2023 2024Timerfd files 2025~~~~~~~~~~~~~ 2026 2027:: 2028 2029 pos: 0 2030 flags: 02 2031 mnt_id: 9 2032 ino: 63107 2033 clockid: 0 2034 ticks: 0 2035 settime flags: 01 2036 it_value: (0, 49406829) 2037 it_interval: (1, 0) 2038 2039where 'clockid' is the clock type and 'ticks' is the number of the timer expirations 2040that have occurred [see timerfd_create(2) for details]. 'settime flags' are 2041flags in octal form been used to setup the timer [see timerfd_settime(2) for 2042details]. 'it_value' is remaining time until the timer expiration. 2043'it_interval' is the interval for the timer. Note the timer might be set up 2044with TIMER_ABSTIME option which will be shown in 'settime flags', but 'it_value' 2045still exhibits timer's remaining time. 2046 2047DMA Buffer files 2048~~~~~~~~~~~~~~~~ 2049 2050:: 2051 2052 pos: 0 2053 flags: 04002 2054 mnt_id: 9 2055 ino: 63107 2056 size: 32768 2057 count: 2 2058 exp_name: system-heap 2059 2060where 'size' is the size of the DMA buffer in bytes. 'count' is the file count of 2061the DMA buffer file. 'exp_name' is the name of the DMA buffer exporter. 2062 20633.9 /proc/<pid>/map_files - Information about memory mapped files 2064--------------------------------------------------------------------- 2065This directory contains symbolic links which represent memory mapped files 2066the process is maintaining. Example output:: 2067 2068 | lr-------- 1 root root 64 Jan 27 11:24 333c600000-333c620000 -> /usr/lib64/ld-2.18.so 2069 | lr-------- 1 root root 64 Jan 27 11:24 333c81f000-333c820000 -> /usr/lib64/ld-2.18.so 2070 | lr-------- 1 root root 64 Jan 27 11:24 333c820000-333c821000 -> /usr/lib64/ld-2.18.so 2071 | ... 2072 | lr-------- 1 root root 64 Jan 27 11:24 35d0421000-35d0422000 -> /usr/lib64/libselinux.so.1 2073 | lr-------- 1 root root 64 Jan 27 11:24 400000-41a000 -> /usr/bin/ls 2074 2075The name of a link represents the virtual memory bounds of a mapping, i.e. 2076vm_area_struct::vm_start-vm_area_struct::vm_end. 2077 2078The main purpose of the map_files is to retrieve a set of memory mapped 2079files in a fast way instead of parsing /proc/<pid>/maps or 2080/proc/<pid>/smaps, both of which contain many more records. At the same 2081time one can open(2) mappings from the listings of two processes and 2082comparing their inode numbers to figure out which anonymous memory areas 2083are actually shared. 2084 20853.10 /proc/<pid>/timerslack_ns - Task timerslack value 2086--------------------------------------------------------- 2087This file provides the value of the task's timerslack value in nanoseconds. 2088This value specifies an amount of time that normal timers may be deferred 2089in order to coalesce timers and avoid unnecessary wakeups. 2090 2091This allows a task's interactivity vs power consumption tradeoff to be 2092adjusted. 2093 2094Writing 0 to the file will set the task's timerslack to the default value. 2095 2096Valid values are from 0 - ULLONG_MAX 2097 2098An application setting the value must have PTRACE_MODE_ATTACH_FSCREDS level 2099permissions on the task specified to change its timerslack_ns value. 2100 21013.11 /proc/<pid>/patch_state - Livepatch patch operation state 2102----------------------------------------------------------------- 2103When CONFIG_LIVEPATCH is enabled, this file displays the value of the 2104patch state for the task. 2105 2106A value of '-1' indicates that no patch is in transition. 2107 2108A value of '0' indicates that a patch is in transition and the task is 2109unpatched. If the patch is being enabled, then the task hasn't been 2110patched yet. If the patch is being disabled, then the task has already 2111been unpatched. 2112 2113A value of '1' indicates that a patch is in transition and the task is 2114patched. If the patch is being enabled, then the task has already been 2115patched. If the patch is being disabled, then the task hasn't been 2116unpatched yet. 2117 21183.12 /proc/<pid>/arch_status - task architecture specific status 2119------------------------------------------------------------------- 2120When CONFIG_PROC_PID_ARCH_STATUS is enabled, this file displays the 2121architecture specific status of the task. 2122 2123Example 2124~~~~~~~ 2125 2126:: 2127 2128 $ cat /proc/6753/arch_status 2129 AVX512_elapsed_ms: 8 2130 2131Description 2132~~~~~~~~~~~ 2133 2134x86 specific entries 2135~~~~~~~~~~~~~~~~~~~~~ 2136 2137AVX512_elapsed_ms 2138^^^^^^^^^^^^^^^^^^ 2139 2140 If AVX512 is supported on the machine, this entry shows the milliseconds 2141 elapsed since the last time AVX512 usage was recorded. The recording 2142 happens on a best effort basis when a task is scheduled out. This means 2143 that the value depends on two factors: 2144 2145 1) The time which the task spent on the CPU without being scheduled 2146 out. With CPU isolation and a single runnable task this can take 2147 several seconds. 2148 2149 2) The time since the task was scheduled out last. Depending on the 2150 reason for being scheduled out (time slice exhausted, syscall ...) 2151 this can be arbitrary long time. 2152 2153 As a consequence the value cannot be considered precise and authoritative 2154 information. The application which uses this information has to be aware 2155 of the overall scenario on the system in order to determine whether a 2156 task is a real AVX512 user or not. Precise information can be obtained 2157 with performance counters. 2158 2159 A special value of '-1' indicates that no AVX512 usage was recorded, thus 2160 the task is unlikely an AVX512 user, but depends on the workload and the 2161 scheduling scenario, it also could be a false negative mentioned above. 2162 2163Chapter 4: Configuring procfs 2164============================= 2165 21664.1 Mount options 2167--------------------- 2168 2169The following mount options are supported: 2170 2171 ========= ======================================================== 2172 hidepid= Set /proc/<pid>/ access mode. 2173 gid= Set the group authorized to learn processes information. 2174 subset= Show only the specified subset of procfs. 2175 ========= ======================================================== 2176 2177hidepid=off or hidepid=0 means classic mode - everybody may access all 2178/proc/<pid>/ directories (default). 2179 2180hidepid=noaccess or hidepid=1 means users may not access any /proc/<pid>/ 2181directories but their own. Sensitive files like cmdline, sched*, status are now 2182protected against other users. This makes it impossible to learn whether any 2183user runs specific program (given the program doesn't reveal itself by its 2184behaviour). As an additional bonus, as /proc/<pid>/cmdline is unaccessible for 2185other users, poorly written programs passing sensitive information via program 2186arguments are now protected against local eavesdroppers. 2187 2188hidepid=invisible or hidepid=2 means hidepid=1 plus all /proc/<pid>/ will be 2189fully invisible to other users. It doesn't mean that it hides a fact whether a 2190process with a specific pid value exists (it can be learned by other means, e.g. 2191by "kill -0 $PID"), but it hides process' uid and gid, which may be learned by 2192stat()'ing /proc/<pid>/ otherwise. It greatly complicates an intruder's task of 2193gathering information about running processes, whether some daemon runs with 2194elevated privileges, whether other user runs some sensitive program, whether 2195other users run any program at all, etc. 2196 2197hidepid=ptraceable or hidepid=4 means that procfs should only contain 2198/proc/<pid>/ directories that the caller can ptrace. 2199 2200gid= defines a group authorized to learn processes information otherwise 2201prohibited by hidepid=. If you use some daemon like identd which needs to learn 2202information about processes information, just add identd to this group. 2203 2204subset=pid hides all top level files and directories in the procfs that 2205are not related to tasks. 2206 2207Chapter 5: Filesystem behavior 2208============================== 2209 2210Originally, before the advent of pid namepsace, procfs was a global file 2211system. It means that there was only one procfs instance in the system. 2212 2213When pid namespace was added, a separate procfs instance was mounted in 2214each pid namespace. So, procfs mount options are global among all 2215mountpoints within the same namespace:: 2216 2217 # grep ^proc /proc/mounts 2218 proc /proc proc rw,relatime,hidepid=2 0 0 2219 2220 # strace -e mount mount -o hidepid=1 -t proc proc /tmp/proc 2221 mount("proc", "/tmp/proc", "proc", 0, "hidepid=1") = 0 2222 +++ exited with 0 +++ 2223 2224 # grep ^proc /proc/mounts 2225 proc /proc proc rw,relatime,hidepid=2 0 0 2226 proc /tmp/proc proc rw,relatime,hidepid=2 0 0 2227 2228and only after remounting procfs mount options will change at all 2229mountpoints:: 2230 2231 # mount -o remount,hidepid=1 -t proc proc /tmp/proc 2232 2233 # grep ^proc /proc/mounts 2234 proc /proc proc rw,relatime,hidepid=1 0 0 2235 proc /tmp/proc proc rw,relatime,hidepid=1 0 0 2236 2237This behavior is different from the behavior of other filesystems. 2238 2239The new procfs behavior is more like other filesystems. Each procfs mount 2240creates a new procfs instance. Mount options affect own procfs instance. 2241It means that it became possible to have several procfs instances 2242displaying tasks with different filtering options in one pid namespace:: 2243 2244 # mount -o hidepid=invisible -t proc proc /proc 2245 # mount -o hidepid=noaccess -t proc proc /tmp/proc 2246 # grep ^proc /proc/mounts 2247 proc /proc proc rw,relatime,hidepid=invisible 0 0 2248 proc /tmp/proc proc rw,relatime,hidepid=noaccess 0 0 2249