1------------------------------------------------------------------------------ 2 T H E /proc F I L E S Y S T E M 3------------------------------------------------------------------------------ 4/proc/sys Terrehon Bowden <terrehon@pacbell.net> October 7 1999 5 Bodo Bauer <bb@ricochet.net> 6 72.4.x update Jorge Nerin <comandante@zaralinux.com> November 14 2000 8move /proc/sys Shen Feng <shen@cn.fujitsu.com> April 1 2009 9------------------------------------------------------------------------------ 10Version 1.3 Kernel version 2.2.12 11 Kernel version 2.4.0-test11-pre4 12------------------------------------------------------------------------------ 13fixes/update part 1.1 Stefani Seibold <stefani@seibold.net> June 9 2009 14 15Table of Contents 16----------------- 17 18 0 Preface 19 0.1 Introduction/Credits 20 0.2 Legal Stuff 21 22 1 Collecting System Information 23 1.1 Process-Specific Subdirectories 24 1.2 Kernel data 25 1.3 IDE devices in /proc/ide 26 1.4 Networking info in /proc/net 27 1.5 SCSI info 28 1.6 Parallel port info in /proc/parport 29 1.7 TTY info in /proc/tty 30 1.8 Miscellaneous kernel statistics in /proc/stat 31 1.9 Ext4 file system parameters 32 33 2 Modifying System Parameters 34 35 3 Per-Process Parameters 36 3.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj - Adjust the oom-killer 37 score 38 3.2 /proc/<pid>/oom_score - Display current oom-killer score 39 3.3 /proc/<pid>/io - Display the IO accounting fields 40 3.4 /proc/<pid>/coredump_filter - Core dump filtering settings 41 3.5 /proc/<pid>/mountinfo - Information about mounts 42 3.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm 43 3.7 /proc/<pid>/task/<tid>/children - Information about task children 44 3.8 /proc/<pid>/fdinfo/<fd> - Information about opened file 45 46 4 Configuring procfs 47 4.1 Mount options 48 49------------------------------------------------------------------------------ 50Preface 51------------------------------------------------------------------------------ 52 530.1 Introduction/Credits 54------------------------ 55 56This documentation is part of a soon (or so we hope) to be released book on 57the SuSE Linux distribution. As there is no complete documentation for the 58/proc file system and we've used many freely available sources to write these 59chapters, it seems only fair to give the work back to the Linux community. 60This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm 61afraid it's still far from complete, but we hope it will be useful. As far as 62we know, it is the first 'all-in-one' document about the /proc file system. It 63is focused on the Intel x86 hardware, so if you are looking for PPC, ARM, 64SPARC, AXP, etc., features, you probably won't find what you are looking for. 65It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But 66additions and patches are welcome and will be added to this document if you 67mail them to Bodo. 68 69We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of 70other people for help compiling this documentation. We'd also like to extend a 71special thank you to Andi Kleen for documentation, which we relied on heavily 72to create this document, as well as the additional information he provided. 73Thanks to everybody else who contributed source or docs to the Linux kernel 74and helped create a great piece of software... :) 75 76If you have any comments, corrections or additions, please don't hesitate to 77contact Bodo Bauer at bb@ricochet.net. We'll be happy to add them to this 78document. 79 80The latest version of this document is available online at 81http://tldp.org/LDP/Linux-Filesystem-Hierarchy/html/proc.html 82 83If the above direction does not works for you, you could try the kernel 84mailing list at linux-kernel@vger.kernel.org and/or try to reach me at 85comandante@zaralinux.com. 86 870.2 Legal Stuff 88--------------- 89 90We don't guarantee the correctness of this document, and if you come to us 91complaining about how you screwed up your system because of incorrect 92documentation, we won't feel responsible... 93 94------------------------------------------------------------------------------ 95CHAPTER 1: COLLECTING SYSTEM INFORMATION 96------------------------------------------------------------------------------ 97 98------------------------------------------------------------------------------ 99In This Chapter 100------------------------------------------------------------------------------ 101* Investigating the properties of the pseudo file system /proc and its 102 ability to provide information on the running Linux system 103* Examining /proc's structure 104* Uncovering various information about the kernel and the processes running 105 on the system 106------------------------------------------------------------------------------ 107 108 109The proc file system acts as an interface to internal data structures in the 110kernel. It can be used to obtain information about the system and to change 111certain kernel parameters at runtime (sysctl). 112 113First, we'll take a look at the read-only parts of /proc. In Chapter 2, we 114show you how you can use /proc/sys to change settings. 115 1161.1 Process-Specific Subdirectories 117----------------------------------- 118 119The directory /proc contains (among other things) one subdirectory for each 120process running on the system, which is named after the process ID (PID). 121 122The link self points to the process reading the file system. Each process 123subdirectory has the entries listed in Table 1-1. 124 125 126Table 1-1: Process specific entries in /proc 127.............................................................................. 128 File Content 129 clear_refs Clears page referenced bits shown in smaps output 130 cmdline Command line arguments 131 cpu Current and last cpu in which it was executed (2.4)(smp) 132 cwd Link to the current working directory 133 environ Values of environment variables 134 exe Link to the executable of this process 135 fd Directory, which contains all file descriptors 136 maps Memory maps to executables and library files (2.4) 137 mem Memory held by this process 138 root Link to the root directory of this process 139 stat Process status 140 statm Process memory status information 141 status Process status in human readable form 142 wchan If CONFIG_KALLSYMS is set, a pre-decoded wchan 143 pagemap Page table 144 stack Report full stack trace, enable via CONFIG_STACKTRACE 145 smaps a extension based on maps, showing the memory consumption of 146 each mapping and flags associated with it 147.............................................................................. 148 149For example, to get the status information of a process, all you have to do is 150read the file /proc/PID/status: 151 152 >cat /proc/self/status 153 Name: cat 154 State: R (running) 155 Tgid: 5452 156 Pid: 5452 157 PPid: 743 158 TracerPid: 0 (2.4) 159 Uid: 501 501 501 501 160 Gid: 100 100 100 100 161 FDSize: 256 162 Groups: 100 14 16 163 VmPeak: 5004 kB 164 VmSize: 5004 kB 165 VmLck: 0 kB 166 VmHWM: 476 kB 167 VmRSS: 476 kB 168 VmData: 156 kB 169 VmStk: 88 kB 170 VmExe: 68 kB 171 VmLib: 1412 kB 172 VmPTE: 20 kb 173 VmSwap: 0 kB 174 Threads: 1 175 SigQ: 0/28578 176 SigPnd: 0000000000000000 177 ShdPnd: 0000000000000000 178 SigBlk: 0000000000000000 179 SigIgn: 0000000000000000 180 SigCgt: 0000000000000000 181 CapInh: 00000000fffffeff 182 CapPrm: 0000000000000000 183 CapEff: 0000000000000000 184 CapBnd: ffffffffffffffff 185 Seccomp: 0 186 voluntary_ctxt_switches: 0 187 nonvoluntary_ctxt_switches: 1 188 189This shows you nearly the same information you would get if you viewed it with 190the ps command. In fact, ps uses the proc file system to obtain its 191information. But you get a more detailed view of the process by reading the 192file /proc/PID/status. It fields are described in table 1-2. 193 194The statm file contains more detailed information about the process 195memory usage. Its seven fields are explained in Table 1-3. The stat file 196contains details information about the process itself. Its fields are 197explained in Table 1-4. 198 199(for SMP CONFIG users) 200For making accounting scalable, RSS related information are handled in 201asynchronous manner and the vaule may not be very precise. To see a precise 202snapshot of a moment, you can see /proc/<pid>/smaps file and scan page table. 203It's slow but very precise. 204 205Table 1-2: Contents of the status files (as of 2.6.30-rc7) 206.............................................................................. 207 Field Content 208 Name filename of the executable 209 State state (R is running, S is sleeping, D is sleeping 210 in an uninterruptible wait, Z is zombie, 211 T is traced or stopped) 212 Tgid thread group ID 213 Pid process id 214 PPid process id of the parent process 215 TracerPid PID of process tracing this process (0 if not) 216 Uid Real, effective, saved set, and file system UIDs 217 Gid Real, effective, saved set, and file system GIDs 218 FDSize number of file descriptor slots currently allocated 219 Groups supplementary group list 220 VmPeak peak virtual memory size 221 VmSize total program size 222 VmLck locked memory size 223 VmHWM peak resident set size ("high water mark") 224 VmRSS size of memory portions 225 VmData size of data, stack, and text segments 226 VmStk size of data, stack, and text segments 227 VmExe size of text segment 228 VmLib size of shared library code 229 VmPTE size of page table entries 230 VmSwap size of swap usage (the number of referred swapents) 231 Threads number of threads 232 SigQ number of signals queued/max. number for queue 233 SigPnd bitmap of pending signals for the thread 234 ShdPnd bitmap of shared pending signals for the process 235 SigBlk bitmap of blocked signals 236 SigIgn bitmap of ignored signals 237 SigCgt bitmap of catched signals 238 CapInh bitmap of inheritable capabilities 239 CapPrm bitmap of permitted capabilities 240 CapEff bitmap of effective capabilities 241 CapBnd bitmap of capabilities bounding set 242 Seccomp seccomp mode, like prctl(PR_GET_SECCOMP, ...) 243 Cpus_allowed mask of CPUs on which this process may run 244 Cpus_allowed_list Same as previous, but in "list format" 245 Mems_allowed mask of memory nodes allowed to this process 246 Mems_allowed_list Same as previous, but in "list format" 247 voluntary_ctxt_switches number of voluntary context switches 248 nonvoluntary_ctxt_switches number of non voluntary context switches 249.............................................................................. 250 251Table 1-3: Contents of the statm files (as of 2.6.8-rc3) 252.............................................................................. 253 Field Content 254 size total program size (pages) (same as VmSize in status) 255 resident size of memory portions (pages) (same as VmRSS in status) 256 shared number of pages that are shared (i.e. backed by a file) 257 trs number of pages that are 'code' (not including libs; broken, 258 includes data segment) 259 lrs number of pages of library (always 0 on 2.6) 260 drs number of pages of data/stack (including libs; broken, 261 includes library text) 262 dt number of dirty pages (always 0 on 2.6) 263.............................................................................. 264 265 266Table 1-4: Contents of the stat files (as of 2.6.30-rc7) 267.............................................................................. 268 Field Content 269 pid process id 270 tcomm filename of the executable 271 state state (R is running, S is sleeping, D is sleeping in an 272 uninterruptible wait, Z is zombie, T is traced or stopped) 273 ppid process id of the parent process 274 pgrp pgrp of the process 275 sid session id 276 tty_nr tty the process uses 277 tty_pgrp pgrp of the tty 278 flags task flags 279 min_flt number of minor faults 280 cmin_flt number of minor faults with child's 281 maj_flt number of major faults 282 cmaj_flt number of major faults with child's 283 utime user mode jiffies 284 stime kernel mode jiffies 285 cutime user mode jiffies with child's 286 cstime kernel mode jiffies with child's 287 priority priority level 288 nice nice level 289 num_threads number of threads 290 it_real_value (obsolete, always 0) 291 start_time time the process started after system boot 292 vsize virtual memory size 293 rss resident set memory size 294 rsslim current limit in bytes on the rss 295 start_code address above which program text can run 296 end_code address below which program text can run 297 start_stack address of the start of the main process stack 298 esp current value of ESP 299 eip current value of EIP 300 pending bitmap of pending signals 301 blocked bitmap of blocked signals 302 sigign bitmap of ignored signals 303 sigcatch bitmap of catched signals 304 wchan address where process went to sleep 305 0 (place holder) 306 0 (place holder) 307 exit_signal signal to send to parent thread on exit 308 task_cpu which CPU the task is scheduled on 309 rt_priority realtime priority 310 policy scheduling policy (man sched_setscheduler) 311 blkio_ticks time spent waiting for block IO 312 gtime guest time of the task in jiffies 313 cgtime guest time of the task children in jiffies 314 start_data address above which program data+bss is placed 315 end_data address below which program data+bss is placed 316 start_brk address above which program heap can be expanded with brk() 317 arg_start address above which program command line is placed 318 arg_end address below which program command line is placed 319 env_start address above which program environment is placed 320 env_end address below which program environment is placed 321 exit_code the thread's exit_code in the form reported by the waitpid system call 322.............................................................................. 323 324The /proc/PID/maps file containing the currently mapped memory regions and 325their access permissions. 326 327The format is: 328 329address perms offset dev inode pathname 330 33108048000-08049000 r-xp 00000000 03:00 8312 /opt/test 33208049000-0804a000 rw-p 00001000 03:00 8312 /opt/test 3330804a000-0806b000 rw-p 00000000 00:00 0 [heap] 334a7cb1000-a7cb2000 ---p 00000000 00:00 0 335a7cb2000-a7eb2000 rw-p 00000000 00:00 0 336a7eb2000-a7eb3000 ---p 00000000 00:00 0 337a7eb3000-a7ed5000 rw-p 00000000 00:00 0 [stack:1001] 338a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6 339a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6 340a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6 341a800b000-a800e000 rw-p 00000000 00:00 0 342a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0 343a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0 344a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0 345a8024000-a8027000 rw-p 00000000 00:00 0 346a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2 347a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2 348a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2 349aff35000-aff4a000 rw-p 00000000 00:00 0 [stack] 350ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso] 351 352where "address" is the address space in the process that it occupies, "perms" 353is a set of permissions: 354 355 r = read 356 w = write 357 x = execute 358 s = shared 359 p = private (copy on write) 360 361"offset" is the offset into the mapping, "dev" is the device (major:minor), and 362"inode" is the inode on that device. 0 indicates that no inode is associated 363with the memory region, as the case would be with BSS (uninitialized data). 364The "pathname" shows the name associated file for this mapping. If the mapping 365is not associated with a file: 366 367 [heap] = the heap of the program 368 [stack] = the stack of the main process 369 [stack:1001] = the stack of the thread with tid 1001 370 [vdso] = the "virtual dynamic shared object", 371 the kernel system call handler 372 [anon:<name>] = an anonymous mapping that has been 373 named by userspace 374 375 or if empty, the mapping is anonymous. 376 377The /proc/PID/task/TID/maps is a view of the virtual memory from the viewpoint 378of the individual tasks of a process. In this file you will see a mapping marked 379as [stack] if that task sees it as a stack. This is a key difference from the 380content of /proc/PID/maps, where you will see all mappings that are being used 381as stack by all of those tasks. Hence, for the example above, the task-level 382map, i.e. /proc/PID/task/TID/maps for thread 1001 will look like this: 383 38408048000-08049000 r-xp 00000000 03:00 8312 /opt/test 38508049000-0804a000 rw-p 00001000 03:00 8312 /opt/test 3860804a000-0806b000 rw-p 00000000 00:00 0 [heap] 387a7cb1000-a7cb2000 ---p 00000000 00:00 0 388a7cb2000-a7eb2000 rw-p 00000000 00:00 0 389a7eb2000-a7eb3000 ---p 00000000 00:00 0 390a7eb3000-a7ed5000 rw-p 00000000 00:00 0 [stack] 391a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6 392a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6 393a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6 394a800b000-a800e000 rw-p 00000000 00:00 0 395a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0 396a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0 397a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0 398a8024000-a8027000 rw-p 00000000 00:00 0 399a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2 400a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2 401a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2 402aff35000-aff4a000 rw-p 00000000 00:00 0 403ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso] 404 405The /proc/PID/smaps is an extension based on maps, showing the memory 406consumption for each of the process's mappings. For each of mappings there 407is a series of lines such as the following: 408 40908048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash 410Size: 1084 kB 411Rss: 892 kB 412Pss: 374 kB 413Shared_Clean: 892 kB 414Shared_Dirty: 0 kB 415Private_Clean: 0 kB 416Private_Dirty: 0 kB 417Referenced: 892 kB 418Anonymous: 0 kB 419Swap: 0 kB 420SwapPss: 0 kB 421KernelPageSize: 4 kB 422MMUPageSize: 4 kB 423Locked: 374 kB 424VmFlags: rd ex mr mw me de 425Name: name from userspace 426 427the first of these lines shows the same information as is displayed for the 428mapping in /proc/PID/maps. The remaining lines show the size of the mapping 429(size), the amount of the mapping that is currently resident in RAM (RSS), the 430process' proportional share of this mapping (PSS), the number of clean and 431dirty private pages in the mapping. 432 433The "proportional set size" (PSS) of a process is the count of pages it has 434in memory, where each page is divided by the number of processes sharing it. 435So if a process has 1000 pages all to itself, and 1000 shared with one other 436process, its PSS will be 1500. 437Note that even a page which is part of a MAP_SHARED mapping, but has only 438a single pte mapped, i.e. is currently used by only one process, is accounted 439as private and not as shared. 440"Referenced" indicates the amount of memory currently marked as referenced or 441accessed. 442"Anonymous" shows the amount of memory that does not belong to any file. Even 443a mapping associated with a file may contain anonymous pages: when MAP_PRIVATE 444and a page is modified, the file page is replaced by a private anonymous copy. 445"Swap" shows how much would-be-anonymous memory is also used, but out on 446swap. 447"SwapPss" shows proportional swap share of this mapping. 448"VmFlags" field deserves a separate description. This member represents the kernel 449flags associated with the particular virtual memory area in two letter encoded 450manner. The codes are the following: 451 rd - readable 452 wr - writeable 453 ex - executable 454 sh - shared 455 mr - may read 456 mw - may write 457 me - may execute 458 ms - may share 459 gd - stack segment growns down 460 pf - pure PFN range 461 dw - disabled write to the mapped file 462 lo - pages are locked in memory 463 io - memory mapped I/O area 464 sr - sequential read advise provided 465 rr - random read advise provided 466 dc - do not copy area on fork 467 de - do not expand area on remapping 468 ac - area is accountable 469 nr - swap space is not reserved for the area 470 ht - area uses huge tlb pages 471 nl - non-linear mapping 472 ar - architecture specific flag 473 dd - do not include area into core dump 474 mm - mixed map area 475 hg - huge page advise flag 476 nh - no-huge page advise flag 477 mg - mergable advise flag 478 479Note that there is no guarantee that every flag and associated mnemonic will 480be present in all further kernel releases. Things get changed, the flags may 481be vanished or the reverse -- new added. 482 483The "Name" field will only be present on a mapping that has been named by 484userspace, and will show the name passed in by userspace. 485 486This file is only present if the CONFIG_MMU kernel configuration option is 487enabled. 488 489The /proc/PID/clear_refs is used to reset the PG_Referenced and ACCESSED/YOUNG 490bits on both physical and virtual pages associated with a process. 491To clear the bits for all the pages associated with the process 492 > echo 1 > /proc/PID/clear_refs 493 494To clear the bits for the anonymous pages associated with the process 495 > echo 2 > /proc/PID/clear_refs 496 497To clear the bits for the file mapped pages associated with the process 498 > echo 3 > /proc/PID/clear_refs 499Any other value written to /proc/PID/clear_refs will have no effect. 500 501To reset the peak resident set size ("high water mark") to the process's 502current value: 503 > echo 5 > /proc/PID/clear_refs 504 505The /proc/pid/pagemap gives the PFN, which can be used to find the pageflags 506using /proc/kpageflags and number of times a page is mapped using 507/proc/kpagecount. For detailed explanation, see Documentation/vm/pagemap.txt. 508 5091.2 Kernel data 510--------------- 511 512Similar to the process entries, the kernel data files give information about 513the running kernel. The files used to obtain this information are contained in 514/proc and are listed in Table 1-5. Not all of these will be present in your 515system. It depends on the kernel configuration and the loaded modules, which 516files are there, and which are missing. 517 518Table 1-5: Kernel info in /proc 519.............................................................................. 520 File Content 521 apm Advanced power management info 522 buddyinfo Kernel memory allocator information (see text) (2.5) 523 bus Directory containing bus specific information 524 cmdline Kernel command line 525 cpuinfo Info about the CPU 526 devices Available devices (block and character) 527 dma Used DMS channels 528 filesystems Supported filesystems 529 driver Various drivers grouped here, currently rtc (2.4) 530 execdomains Execdomains, related to security (2.4) 531 fb Frame Buffer devices (2.4) 532 fs File system parameters, currently nfs/exports (2.4) 533 ide Directory containing info about the IDE subsystem 534 interrupts Interrupt usage 535 iomem Memory map (2.4) 536 ioports I/O port usage 537 irq Masks for irq to cpu affinity (2.4)(smp?) 538 isapnp ISA PnP (Plug&Play) Info (2.4) 539 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4)) 540 kmsg Kernel messages 541 ksyms Kernel symbol table 542 loadavg Load average of last 1, 5 & 15 minutes 543 locks Kernel locks 544 meminfo Memory info 545 misc Miscellaneous 546 modules List of loaded modules 547 mounts Mounted filesystems 548 net Networking info (see text) 549 pagetypeinfo Additional page allocator information (see text) (2.5) 550 partitions Table of partitions known to the system 551 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/, 552 decoupled by lspci (2.4) 553 rtc Real time clock 554 scsi SCSI info (see text) 555 slabinfo Slab pool info 556 softirqs softirq usage 557 stat Overall statistics 558 swaps Swap space utilization 559 sys See chapter 2 560 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4) 561 tty Info of tty drivers 562 uptime System uptime 563 version Kernel version 564 video bttv info of video resources (2.4) 565 vmallocinfo Show vmalloced areas 566.............................................................................. 567 568You can, for example, check which interrupts are currently in use and what 569they are used for by looking in the file /proc/interrupts: 570 571 > cat /proc/interrupts 572 CPU0 573 0: 8728810 XT-PIC timer 574 1: 895 XT-PIC keyboard 575 2: 0 XT-PIC cascade 576 3: 531695 XT-PIC aha152x 577 4: 2014133 XT-PIC serial 578 5: 44401 XT-PIC pcnet_cs 579 8: 2 XT-PIC rtc 580 11: 8 XT-PIC i82365 581 12: 182918 XT-PIC PS/2 Mouse 582 13: 1 XT-PIC fpu 583 14: 1232265 XT-PIC ide0 584 15: 7 XT-PIC ide1 585 NMI: 0 586 587In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the 588output of a SMP machine): 589 590 > cat /proc/interrupts 591 592 CPU0 CPU1 593 0: 1243498 1214548 IO-APIC-edge timer 594 1: 8949 8958 IO-APIC-edge keyboard 595 2: 0 0 XT-PIC cascade 596 5: 11286 10161 IO-APIC-edge soundblaster 597 8: 1 0 IO-APIC-edge rtc 598 9: 27422 27407 IO-APIC-edge 3c503 599 12: 113645 113873 IO-APIC-edge PS/2 Mouse 600 13: 0 0 XT-PIC fpu 601 14: 22491 24012 IO-APIC-edge ide0 602 15: 2183 2415 IO-APIC-edge ide1 603 17: 30564 30414 IO-APIC-level eth0 604 18: 177 164 IO-APIC-level bttv 605 NMI: 2457961 2457959 606 LOC: 2457882 2457881 607 ERR: 2155 608 609NMI is incremented in this case because every timer interrupt generates a NMI 610(Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups. 611 612LOC is the local interrupt counter of the internal APIC of every CPU. 613 614ERR is incremented in the case of errors in the IO-APIC bus (the bus that 615connects the CPUs in a SMP system. This means that an error has been detected, 616the IO-APIC automatically retry the transmission, so it should not be a big 617problem, but you should read the SMP-FAQ. 618 619In 2.6.2* /proc/interrupts was expanded again. This time the goal was for 620/proc/interrupts to display every IRQ vector in use by the system, not 621just those considered 'most important'. The new vectors are: 622 623 THR -- interrupt raised when a machine check threshold counter 624 (typically counting ECC corrected errors of memory or cache) exceeds 625 a configurable threshold. Only available on some systems. 626 627 TRM -- a thermal event interrupt occurs when a temperature threshold 628 has been exceeded for the CPU. This interrupt may also be generated 629 when the temperature drops back to normal. 630 631 SPU -- a spurious interrupt is some interrupt that was raised then lowered 632 by some IO device before it could be fully processed by the APIC. Hence 633 the APIC sees the interrupt but does not know what device it came from. 634 For this case the APIC will generate the interrupt with a IRQ vector 635 of 0xff. This might also be generated by chipset bugs. 636 637 RES, CAL, TLB -- rescheduling, call and TLB flush interrupts are 638 sent from one CPU to another per the needs of the OS. Typically, 639 their statistics are used by kernel developers and interested users to 640 determine the occurrence of interrupts of the given type. 641 642The above IRQ vectors are displayed only when relevant. For example, 643the threshold vector does not exist on x86_64 platforms. Others are 644suppressed when the system is a uniprocessor. As of this writing, only 645i386 and x86_64 platforms support the new IRQ vector displays. 646 647Of some interest is the introduction of the /proc/irq directory to 2.4. 648It could be used to set IRQ to CPU affinity, this means that you can "hook" an 649IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the 650irq subdir is one subdir for each IRQ, and two files; default_smp_affinity and 651prof_cpu_mask. 652 653For example 654 > ls /proc/irq/ 655 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask 656 1 11 13 15 17 19 3 5 7 9 default_smp_affinity 657 > ls /proc/irq/0/ 658 smp_affinity 659 660smp_affinity is a bitmask, in which you can specify which CPUs can handle the 661IRQ, you can set it by doing: 662 663 > echo 1 > /proc/irq/10/smp_affinity 664 665This means that only the first CPU will handle the IRQ, but you can also echo 6665 which means that only the first and fourth CPU can handle the IRQ. 667 668The contents of each smp_affinity file is the same by default: 669 670 > cat /proc/irq/0/smp_affinity 671 ffffffff 672 673There is an alternate interface, smp_affinity_list which allows specifying 674a cpu range instead of a bitmask: 675 676 > cat /proc/irq/0/smp_affinity_list 677 1024-1031 678 679The default_smp_affinity mask applies to all non-active IRQs, which are the 680IRQs which have not yet been allocated/activated, and hence which lack a 681/proc/irq/[0-9]* directory. 682 683The node file on an SMP system shows the node to which the device using the IRQ 684reports itself as being attached. This hardware locality information does not 685include information about any possible driver locality preference. 686 687prof_cpu_mask specifies which CPUs are to be profiled by the system wide 688profiler. Default value is ffffffff (all cpus if there are only 32 of them). 689 690The way IRQs are routed is handled by the IO-APIC, and it's Round Robin 691between all the CPUs which are allowed to handle it. As usual the kernel has 692more info than you and does a better job than you, so the defaults are the 693best choice for almost everyone. [Note this applies only to those IO-APIC's 694that support "Round Robin" interrupt distribution.] 695 696There are three more important subdirectories in /proc: net, scsi, and sys. 697The general rule is that the contents, or even the existence of these 698directories, depend on your kernel configuration. If SCSI is not enabled, the 699directory scsi may not exist. The same is true with the net, which is there 700only when networking support is present in the running kernel. 701 702The slabinfo file gives information about memory usage at the slab level. 703Linux uses slab pools for memory management above page level in version 2.2. 704Commonly used objects have their own slab pool (such as network buffers, 705directory cache, and so on). 706 707.............................................................................. 708 709> cat /proc/buddyinfo 710 711Node 0, zone DMA 0 4 5 4 4 3 ... 712Node 0, zone Normal 1 0 0 1 101 8 ... 713Node 0, zone HighMem 2 0 0 1 1 0 ... 714 715External fragmentation is a problem under some workloads, and buddyinfo is a 716useful tool for helping diagnose these problems. Buddyinfo will give you a 717clue as to how big an area you can safely allocate, or why a previous 718allocation failed. 719 720Each column represents the number of pages of a certain order which are 721available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in 722ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE 723available in ZONE_NORMAL, etc... 724 725More information relevant to external fragmentation can be found in 726pagetypeinfo. 727 728> cat /proc/pagetypeinfo 729Page block order: 9 730Pages per block: 512 731 732Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 733Node 0, zone DMA, type Unmovable 0 0 0 1 1 1 1 1 1 1 0 734Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 735Node 0, zone DMA, type Movable 1 1 2 1 2 1 1 0 1 0 2 736Node 0, zone DMA, type Reserve 0 0 0 0 0 0 0 0 0 1 0 737Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 738Node 0, zone DMA32, type Unmovable 103 54 77 1 1 1 11 8 7 1 9 739Node 0, zone DMA32, type Reclaimable 0 0 2 1 0 0 0 0 1 0 0 740Node 0, zone DMA32, type Movable 169 152 113 91 77 54 39 13 6 1 452 741Node 0, zone DMA32, type Reserve 1 2 2 2 2 0 1 1 1 1 0 742Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 743 744Number of blocks type Unmovable Reclaimable Movable Reserve Isolate 745Node 0, zone DMA 2 0 5 1 0 746Node 0, zone DMA32 41 6 967 2 0 747 748Fragmentation avoidance in the kernel works by grouping pages of different 749migrate types into the same contiguous regions of memory called page blocks. 750A page block is typically the size of the default hugepage size e.g. 2MB on 751X86-64. By keeping pages grouped based on their ability to move, the kernel 752can reclaim pages within a page block to satisfy a high-order allocation. 753 754The pagetypinfo begins with information on the size of a page block. It 755then gives the same type of information as buddyinfo except broken down 756by migrate-type and finishes with details on how many page blocks of each 757type exist. 758 759If min_free_kbytes has been tuned correctly (recommendations made by hugeadm 760from libhugetlbfs http://sourceforge.net/projects/libhugetlbfs/), one can 761make an estimate of the likely number of huge pages that can be allocated 762at a given point in time. All the "Movable" blocks should be allocatable 763unless memory has been mlock()'d. Some of the Reclaimable blocks should 764also be allocatable although a lot of filesystem metadata may have to be 765reclaimed to achieve this. 766 767.............................................................................. 768 769meminfo: 770 771Provides information about distribution and utilization of memory. This 772varies by architecture and compile options. The following is from a 77316GB PIII, which has highmem enabled. You may not have all of these fields. 774 775> cat /proc/meminfo 776 777The "Locked" indicates whether the mapping is locked in memory or not. 778 779 780MemTotal: 16344972 kB 781MemFree: 13634064 kB 782Buffers: 3656 kB 783Cached: 1195708 kB 784SwapCached: 0 kB 785Active: 891636 kB 786Inactive: 1077224 kB 787HighTotal: 15597528 kB 788HighFree: 13629632 kB 789LowTotal: 747444 kB 790LowFree: 4432 kB 791SwapTotal: 0 kB 792SwapFree: 0 kB 793Dirty: 968 kB 794Writeback: 0 kB 795AnonPages: 861800 kB 796Mapped: 280372 kB 797Slab: 284364 kB 798SReclaimable: 159856 kB 799SUnreclaim: 124508 kB 800PageTables: 24448 kB 801NFS_Unstable: 0 kB 802Bounce: 0 kB 803WritebackTmp: 0 kB 804CommitLimit: 7669796 kB 805Committed_AS: 100056 kB 806VmallocTotal: 112216 kB 807VmallocUsed: 428 kB 808VmallocChunk: 111088 kB 809AnonHugePages: 49152 kB 810 811 MemTotal: Total usable ram (i.e. physical ram minus a few reserved 812 bits and the kernel binary code) 813 MemFree: The sum of LowFree+HighFree 814 Buffers: Relatively temporary storage for raw disk blocks 815 shouldn't get tremendously large (20MB or so) 816 Cached: in-memory cache for files read from the disk (the 817 pagecache). Doesn't include SwapCached 818 SwapCached: Memory that once was swapped out, is swapped back in but 819 still also is in the swapfile (if memory is needed it 820 doesn't need to be swapped out AGAIN because it is already 821 in the swapfile. This saves I/O) 822 Active: Memory that has been used more recently and usually not 823 reclaimed unless absolutely necessary. 824 Inactive: Memory which has been less recently used. It is more 825 eligible to be reclaimed for other purposes 826 HighTotal: 827 HighFree: Highmem is all memory above ~860MB of physical memory 828 Highmem areas are for use by userspace programs, or 829 for the pagecache. The kernel must use tricks to access 830 this memory, making it slower to access than lowmem. 831 LowTotal: 832 LowFree: Lowmem is memory which can be used for everything that 833 highmem can be used for, but it is also available for the 834 kernel's use for its own data structures. Among many 835 other things, it is where everything from the Slab is 836 allocated. Bad things happen when you're out of lowmem. 837 SwapTotal: total amount of swap space available 838 SwapFree: Memory which has been evicted from RAM, and is temporarily 839 on the disk 840 Dirty: Memory which is waiting to get written back to the disk 841 Writeback: Memory which is actively being written back to the disk 842 AnonPages: Non-file backed pages mapped into userspace page tables 843AnonHugePages: Non-file backed huge pages mapped into userspace page tables 844 Mapped: files which have been mmaped, such as libraries 845 Slab: in-kernel data structures cache 846SReclaimable: Part of Slab, that might be reclaimed, such as caches 847 SUnreclaim: Part of Slab, that cannot be reclaimed on memory pressure 848 PageTables: amount of memory dedicated to the lowest level of page 849 tables. 850NFS_Unstable: NFS pages sent to the server, but not yet committed to stable 851 storage 852 Bounce: Memory used for block device "bounce buffers" 853WritebackTmp: Memory used by FUSE for temporary writeback buffers 854 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'), 855 this is the total amount of memory currently available to 856 be allocated on the system. This limit is only adhered to 857 if strict overcommit accounting is enabled (mode 2 in 858 'vm.overcommit_memory'). 859 The CommitLimit is calculated with the following formula: 860 CommitLimit = ('vm.overcommit_ratio' * Physical RAM) + Swap 861 For example, on a system with 1G of physical RAM and 7G 862 of swap with a `vm.overcommit_ratio` of 30 it would 863 yield a CommitLimit of 7.3G. 864 For more details, see the memory overcommit documentation 865 in vm/overcommit-accounting. 866Committed_AS: The amount of memory presently allocated on the system. 867 The committed memory is a sum of all of the memory which 868 has been allocated by processes, even if it has not been 869 "used" by them as of yet. A process which malloc()'s 1G 870 of memory, but only touches 300M of it will only show up 871 as using 300M of memory even if it has the address space 872 allocated for the entire 1G. This 1G is memory which has 873 been "committed" to by the VM and can be used at any time 874 by the allocating application. With strict overcommit 875 enabled on the system (mode 2 in 'vm.overcommit_memory'), 876 allocations which would exceed the CommitLimit (detailed 877 above) will not be permitted. This is useful if one needs 878 to guarantee that processes will not fail due to lack of 879 memory once that memory has been successfully allocated. 880VmallocTotal: total size of vmalloc memory area 881 VmallocUsed: amount of vmalloc area which is used 882VmallocChunk: largest contiguous block of vmalloc area which is free 883 884.............................................................................. 885 886vmallocinfo: 887 888Provides information about vmalloced/vmaped areas. One line per area, 889containing the virtual address range of the area, size in bytes, 890caller information of the creator, and optional information depending 891on the kind of area : 892 893 pages=nr number of pages 894 phys=addr if a physical address was specified 895 ioremap I/O mapping (ioremap() and friends) 896 vmalloc vmalloc() area 897 vmap vmap()ed pages 898 user VM_USERMAP area 899 vpages buffer for pages pointers was vmalloced (huge area) 900 N<node>=nr (Only on NUMA kernels) 901 Number of pages allocated on memory node <node> 902 903> cat /proc/vmallocinfo 9040xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ... 905 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128 9060xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ... 907 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64 9080xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f... 909 phys=7fee8000 ioremap 9100xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f... 911 phys=7fee7000 ioremap 9120xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210 9130xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ... 914 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3 9150xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ... 916 pages=2 vmalloc N1=2 9170xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ... 918 /0x130 [x_tables] pages=4 vmalloc N0=4 9190xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ... 920 pages=14 vmalloc N2=14 9210xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ... 922 pages=4 vmalloc N1=4 9230xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ... 924 pages=2 vmalloc N1=2 9250xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ... 926 pages=10 vmalloc N0=10 927 928.............................................................................. 929 930softirqs: 931 932Provides counts of softirq handlers serviced since boot time, for each cpu. 933 934> cat /proc/softirqs 935 CPU0 CPU1 CPU2 CPU3 936 HI: 0 0 0 0 937 TIMER: 27166 27120 27097 27034 938 NET_TX: 0 0 0 17 939 NET_RX: 42 0 0 39 940 BLOCK: 0 0 107 1121 941 TASKLET: 0 0 0 290 942 SCHED: 27035 26983 26971 26746 943 HRTIMER: 0 0 0 0 944 RCU: 1678 1769 2178 2250 945 946 9471.3 IDE devices in /proc/ide 948---------------------------- 949 950The subdirectory /proc/ide contains information about all IDE devices of which 951the kernel is aware. There is one subdirectory for each IDE controller, the 952file drivers and a link for each IDE device, pointing to the device directory 953in the controller specific subtree. 954 955The file drivers contains general information about the drivers used for the 956IDE devices: 957 958 > cat /proc/ide/drivers 959 ide-cdrom version 4.53 960 ide-disk version 1.08 961 962More detailed information can be found in the controller specific 963subdirectories. These are named ide0, ide1 and so on. Each of these 964directories contains the files shown in table 1-6. 965 966 967Table 1-6: IDE controller info in /proc/ide/ide? 968.............................................................................. 969 File Content 970 channel IDE channel (0 or 1) 971 config Configuration (only for PCI/IDE bridge) 972 mate Mate name 973 model Type/Chipset of IDE controller 974.............................................................................. 975 976Each device connected to a controller has a separate subdirectory in the 977controllers directory. The files listed in table 1-7 are contained in these 978directories. 979 980 981Table 1-7: IDE device information 982.............................................................................. 983 File Content 984 cache The cache 985 capacity Capacity of the medium (in 512Byte blocks) 986 driver driver and version 987 geometry physical and logical geometry 988 identify device identify block 989 media media type 990 model device identifier 991 settings device setup 992 smart_thresholds IDE disk management thresholds 993 smart_values IDE disk management values 994.............................................................................. 995 996The most interesting file is settings. This file contains a nice overview of 997the drive parameters: 998 999 # cat /proc/ide/ide0/hda/settings 1000 name value min max mode 1001 ---- ----- --- --- ---- 1002 bios_cyl 526 0 65535 rw 1003 bios_head 255 0 255 rw 1004 bios_sect 63 0 63 rw 1005 breada_readahead 4 0 127 rw 1006 bswap 0 0 1 r 1007 file_readahead 72 0 2097151 rw 1008 io_32bit 0 0 3 rw 1009 keepsettings 0 0 1 rw 1010 max_kb_per_request 122 1 127 rw 1011 multcount 0 0 8 rw 1012 nice1 1 0 1 rw 1013 nowerr 0 0 1 rw 1014 pio_mode write-only 0 255 w 1015 slow 0 0 1 rw 1016 unmaskirq 0 0 1 rw 1017 using_dma 0 0 1 rw 1018 1019 10201.4 Networking info in /proc/net 1021-------------------------------- 1022 1023The subdirectory /proc/net follows the usual pattern. Table 1-8 shows the 1024additional values you get for IP version 6 if you configure the kernel to 1025support this. Table 1-9 lists the files and their meaning. 1026 1027 1028Table 1-8: IPv6 info in /proc/net 1029.............................................................................. 1030 File Content 1031 udp6 UDP sockets (IPv6) 1032 tcp6 TCP sockets (IPv6) 1033 raw6 Raw device statistics (IPv6) 1034 igmp6 IP multicast addresses, which this host joined (IPv6) 1035 if_inet6 List of IPv6 interface addresses 1036 ipv6_route Kernel routing table for IPv6 1037 rt6_stats Global IPv6 routing tables statistics 1038 sockstat6 Socket statistics (IPv6) 1039 snmp6 Snmp data (IPv6) 1040.............................................................................. 1041 1042 1043Table 1-9: Network info in /proc/net 1044.............................................................................. 1045 File Content 1046 arp Kernel ARP table 1047 dev network devices with statistics 1048 dev_mcast the Layer2 multicast groups a device is listening too 1049 (interface index, label, number of references, number of bound 1050 addresses). 1051 dev_stat network device status 1052 ip_fwchains Firewall chain linkage 1053 ip_fwnames Firewall chain names 1054 ip_masq Directory containing the masquerading tables 1055 ip_masquerade Major masquerading table 1056 netstat Network statistics 1057 raw raw device statistics 1058 route Kernel routing table 1059 rpc Directory containing rpc info 1060 rt_cache Routing cache 1061 snmp SNMP data 1062 sockstat Socket statistics 1063 tcp TCP sockets 1064 udp UDP sockets 1065 unix UNIX domain sockets 1066 wireless Wireless interface data (Wavelan etc) 1067 igmp IP multicast addresses, which this host joined 1068 psched Global packet scheduler parameters. 1069 netlink List of PF_NETLINK sockets 1070 ip_mr_vifs List of multicast virtual interfaces 1071 ip_mr_cache List of multicast routing cache 1072.............................................................................. 1073 1074You can use this information to see which network devices are available in 1075your system and how much traffic was routed over those devices: 1076 1077 > cat /proc/net/dev 1078 Inter-|Receive |[... 1079 face |bytes packets errs drop fifo frame compressed multicast|[... 1080 lo: 908188 5596 0 0 0 0 0 0 [... 1081 ppp0:15475140 20721 410 0 0 410 0 0 [... 1082 eth0: 614530 7085 0 0 0 0 0 1 [... 1083 1084 ...] Transmit 1085 ...] bytes packets errs drop fifo colls carrier compressed 1086 ...] 908188 5596 0 0 0 0 0 0 1087 ...] 1375103 17405 0 0 0 0 0 0 1088 ...] 1703981 5535 0 0 0 3 0 0 1089 1090In addition, each Channel Bond interface has its own directory. For 1091example, the bond0 device will have a directory called /proc/net/bond0/. 1092It will contain information that is specific to that bond, such as the 1093current slaves of the bond, the link status of the slaves, and how 1094many times the slaves link has failed. 1095 10961.5 SCSI info 1097------------- 1098 1099If you have a SCSI host adapter in your system, you'll find a subdirectory 1100named after the driver for this adapter in /proc/scsi. You'll also see a list 1101of all recognized SCSI devices in /proc/scsi: 1102 1103 >cat /proc/scsi/scsi 1104 Attached devices: 1105 Host: scsi0 Channel: 00 Id: 00 Lun: 00 1106 Vendor: IBM Model: DGHS09U Rev: 03E0 1107 Type: Direct-Access ANSI SCSI revision: 03 1108 Host: scsi0 Channel: 00 Id: 06 Lun: 00 1109 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04 1110 Type: CD-ROM ANSI SCSI revision: 02 1111 1112 1113The directory named after the driver has one file for each adapter found in 1114the system. These files contain information about the controller, including 1115the used IRQ and the IO address range. The amount of information shown is 1116dependent on the adapter you use. The example shows the output for an Adaptec 1117AHA-2940 SCSI adapter: 1118 1119 > cat /proc/scsi/aic7xxx/0 1120 1121 Adaptec AIC7xxx driver version: 5.1.19/3.2.4 1122 Compile Options: 1123 TCQ Enabled By Default : Disabled 1124 AIC7XXX_PROC_STATS : Disabled 1125 AIC7XXX_RESET_DELAY : 5 1126 Adapter Configuration: 1127 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter 1128 Ultra Wide Controller 1129 PCI MMAPed I/O Base: 0xeb001000 1130 Adapter SEEPROM Config: SEEPROM found and used. 1131 Adaptec SCSI BIOS: Enabled 1132 IRQ: 10 1133 SCBs: Active 0, Max Active 2, 1134 Allocated 15, HW 16, Page 255 1135 Interrupts: 160328 1136 BIOS Control Word: 0x18b6 1137 Adapter Control Word: 0x005b 1138 Extended Translation: Enabled 1139 Disconnect Enable Flags: 0xffff 1140 Ultra Enable Flags: 0x0001 1141 Tag Queue Enable Flags: 0x0000 1142 Ordered Queue Tag Flags: 0x0000 1143 Default Tag Queue Depth: 8 1144 Tagged Queue By Device array for aic7xxx host instance 0: 1145 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255} 1146 Actual queue depth per device for aic7xxx host instance 0: 1147 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1} 1148 Statistics: 1149 (scsi0:0:0:0) 1150 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8 1151 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0) 1152 Total transfers 160151 (74577 reads and 85574 writes) 1153 (scsi0:0:6:0) 1154 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15 1155 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0) 1156 Total transfers 0 (0 reads and 0 writes) 1157 1158 11591.6 Parallel port info in /proc/parport 1160--------------------------------------- 1161 1162The directory /proc/parport contains information about the parallel ports of 1163your system. It has one subdirectory for each port, named after the port 1164number (0,1,2,...). 1165 1166These directories contain the four files shown in Table 1-10. 1167 1168 1169Table 1-10: Files in /proc/parport 1170.............................................................................. 1171 File Content 1172 autoprobe Any IEEE-1284 device ID information that has been acquired. 1173 devices list of the device drivers using that port. A + will appear by the 1174 name of the device currently using the port (it might not appear 1175 against any). 1176 hardware Parallel port's base address, IRQ line and DMA channel. 1177 irq IRQ that parport is using for that port. This is in a separate 1178 file to allow you to alter it by writing a new value in (IRQ 1179 number or none). 1180.............................................................................. 1181 11821.7 TTY info in /proc/tty 1183------------------------- 1184 1185Information about the available and actually used tty's can be found in the 1186directory /proc/tty.You'll find entries for drivers and line disciplines in 1187this directory, as shown in Table 1-11. 1188 1189 1190Table 1-11: Files in /proc/tty 1191.............................................................................. 1192 File Content 1193 drivers list of drivers and their usage 1194 ldiscs registered line disciplines 1195 driver/serial usage statistic and status of single tty lines 1196.............................................................................. 1197 1198To see which tty's are currently in use, you can simply look into the file 1199/proc/tty/drivers: 1200 1201 > cat /proc/tty/drivers 1202 pty_slave /dev/pts 136 0-255 pty:slave 1203 pty_master /dev/ptm 128 0-255 pty:master 1204 pty_slave /dev/ttyp 3 0-255 pty:slave 1205 pty_master /dev/pty 2 0-255 pty:master 1206 serial /dev/cua 5 64-67 serial:callout 1207 serial /dev/ttyS 4 64-67 serial 1208 /dev/tty0 /dev/tty0 4 0 system:vtmaster 1209 /dev/ptmx /dev/ptmx 5 2 system 1210 /dev/console /dev/console 5 1 system:console 1211 /dev/tty /dev/tty 5 0 system:/dev/tty 1212 unknown /dev/tty 4 1-63 console 1213 1214 12151.8 Miscellaneous kernel statistics in /proc/stat 1216------------------------------------------------- 1217 1218Various pieces of information about kernel activity are available in the 1219/proc/stat file. All of the numbers reported in this file are aggregates 1220since the system first booted. For a quick look, simply cat the file: 1221 1222 > cat /proc/stat 1223 cpu 2255 34 2290 22625563 6290 127 456 0 0 1224 cpu0 1132 34 1441 11311718 3675 127 438 0 0 1225 cpu1 1123 0 849 11313845 2614 0 18 0 0 1226 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...] 1227 ctxt 1990473 1228 btime 1062191376 1229 processes 2915 1230 procs_running 1 1231 procs_blocked 0 1232 softirq 183433 0 21755 12 39 1137 231 21459 2263 1233 1234The very first "cpu" line aggregates the numbers in all of the other "cpuN" 1235lines. These numbers identify the amount of time the CPU has spent performing 1236different kinds of work. Time units are in USER_HZ (typically hundredths of a 1237second). The meanings of the columns are as follows, from left to right: 1238 1239- user: normal processes executing in user mode 1240- nice: niced processes executing in user mode 1241- system: processes executing in kernel mode 1242- idle: twiddling thumbs 1243- iowait: waiting for I/O to complete 1244- irq: servicing interrupts 1245- softirq: servicing softirqs 1246- steal: involuntary wait 1247- guest: running a normal guest 1248- guest_nice: running a niced guest 1249 1250The "intr" line gives counts of interrupts serviced since boot time, for each 1251of the possible system interrupts. The first column is the total of all 1252interrupts serviced; each subsequent column is the total for that particular 1253interrupt. 1254 1255The "ctxt" line gives the total number of context switches across all CPUs. 1256 1257The "btime" line gives the time at which the system booted, in seconds since 1258the Unix epoch. 1259 1260The "processes" line gives the number of processes and threads created, which 1261includes (but is not limited to) those created by calls to the fork() and 1262clone() system calls. 1263 1264The "procs_running" line gives the total number of threads that are 1265running or ready to run (i.e., the total number of runnable threads). 1266 1267The "procs_blocked" line gives the number of processes currently blocked, 1268waiting for I/O to complete. 1269 1270The "softirq" line gives counts of softirqs serviced since boot time, for each 1271of the possible system softirqs. The first column is the total of all 1272softirqs serviced; each subsequent column is the total for that particular 1273softirq. 1274 1275 12761.9 Ext4 file system parameters 1277------------------------------ 1278 1279Information about mounted ext4 file systems can be found in 1280/proc/fs/ext4. Each mounted filesystem will have a directory in 1281/proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or 1282/proc/fs/ext4/dm-0). The files in each per-device directory are shown 1283in Table 1-12, below. 1284 1285Table 1-12: Files in /proc/fs/ext4/<devname> 1286.............................................................................. 1287 File Content 1288 mb_groups details of multiblock allocator buddy cache of free blocks 1289.............................................................................. 1290 12912.0 /proc/consoles 1292------------------ 1293Shows registered system console lines. 1294 1295To see which character device lines are currently used for the system console 1296/dev/console, you may simply look into the file /proc/consoles: 1297 1298 > cat /proc/consoles 1299 tty0 -WU (ECp) 4:7 1300 ttyS0 -W- (Ep) 4:64 1301 1302The columns are: 1303 1304 device name of the device 1305 operations R = can do read operations 1306 W = can do write operations 1307 U = can do unblank 1308 flags E = it is enabled 1309 C = it is preferred console 1310 B = it is primary boot console 1311 p = it is used for printk buffer 1312 b = it is not a TTY but a Braille device 1313 a = it is safe to use when cpu is offline 1314 major:minor major and minor number of the device separated by a colon 1315 1316------------------------------------------------------------------------------ 1317Summary 1318------------------------------------------------------------------------------ 1319The /proc file system serves information about the running system. It not only 1320allows access to process data but also allows you to request the kernel status 1321by reading files in the hierarchy. 1322 1323The directory structure of /proc reflects the types of information and makes 1324it easy, if not obvious, where to look for specific data. 1325------------------------------------------------------------------------------ 1326 1327------------------------------------------------------------------------------ 1328CHAPTER 2: MODIFYING SYSTEM PARAMETERS 1329------------------------------------------------------------------------------ 1330 1331------------------------------------------------------------------------------ 1332In This Chapter 1333------------------------------------------------------------------------------ 1334* Modifying kernel parameters by writing into files found in /proc/sys 1335* Exploring the files which modify certain parameters 1336* Review of the /proc/sys file tree 1337------------------------------------------------------------------------------ 1338 1339 1340A very interesting part of /proc is the directory /proc/sys. This is not only 1341a source of information, it also allows you to change parameters within the 1342kernel. Be very careful when attempting this. You can optimize your system, 1343but you can also cause it to crash. Never alter kernel parameters on a 1344production system. Set up a development machine and test to make sure that 1345everything works the way you want it to. You may have no alternative but to 1346reboot the machine once an error has been made. 1347 1348To change a value, simply echo the new value into the file. An example is 1349given below in the section on the file system data. You need to be root to do 1350this. You can create your own boot script to perform this every time your 1351system boots. 1352 1353The files in /proc/sys can be used to fine tune and monitor miscellaneous and 1354general things in the operation of the Linux kernel. Since some of the files 1355can inadvertently disrupt your system, it is advisable to read both 1356documentation and source before actually making adjustments. In any case, be 1357very careful when writing to any of these files. The entries in /proc may 1358change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt 1359review the kernel documentation in the directory /usr/src/linux/Documentation. 1360This chapter is heavily based on the documentation included in the pre 2.2 1361kernels, and became part of it in version 2.2.1 of the Linux kernel. 1362 1363Please see: Documentation/sysctl/ directory for descriptions of these 1364entries. 1365 1366------------------------------------------------------------------------------ 1367Summary 1368------------------------------------------------------------------------------ 1369Certain aspects of kernel behavior can be modified at runtime, without the 1370need to recompile the kernel, or even to reboot the system. The files in the 1371/proc/sys tree can not only be read, but also modified. You can use the echo 1372command to write value into these files, thereby changing the default settings 1373of the kernel. 1374------------------------------------------------------------------------------ 1375 1376------------------------------------------------------------------------------ 1377CHAPTER 3: PER-PROCESS PARAMETERS 1378------------------------------------------------------------------------------ 1379 13803.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj- Adjust the oom-killer score 1381-------------------------------------------------------------------------------- 1382 1383These file can be used to adjust the badness heuristic used to select which 1384process gets killed in out of memory conditions. 1385 1386The badness heuristic assigns a value to each candidate task ranging from 0 1387(never kill) to 1000 (always kill) to determine which process is targeted. The 1388units are roughly a proportion along that range of allowed memory the process 1389may allocate from based on an estimation of its current memory and swap use. 1390For example, if a task is using all allowed memory, its badness score will be 13911000. If it is using half of its allowed memory, its score will be 500. 1392 1393There is an additional factor included in the badness score: root 1394processes are given 3% extra memory over other tasks. 1395 1396The amount of "allowed" memory depends on the context in which the oom killer 1397was called. If it is due to the memory assigned to the allocating task's cpuset 1398being exhausted, the allowed memory represents the set of mems assigned to that 1399cpuset. If it is due to a mempolicy's node(s) being exhausted, the allowed 1400memory represents the set of mempolicy nodes. If it is due to a memory 1401limit (or swap limit) being reached, the allowed memory is that configured 1402limit. Finally, if it is due to the entire system being out of memory, the 1403allowed memory represents all allocatable resources. 1404 1405The value of /proc/<pid>/oom_score_adj is added to the badness score before it 1406is used to determine which task to kill. Acceptable values range from -1000 1407(OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_MAX). This allows userspace to 1408polarize the preference for oom killing either by always preferring a certain 1409task or completely disabling it. The lowest possible value, -1000, is 1410equivalent to disabling oom killing entirely for that task since it will always 1411report a badness score of 0. 1412 1413Consequently, it is very simple for userspace to define the amount of memory to 1414consider for each task. Setting a /proc/<pid>/oom_score_adj value of +500, for 1415example, is roughly equivalent to allowing the remainder of tasks sharing the 1416same system, cpuset, mempolicy, or memory controller resources to use at least 141750% more memory. A value of -500, on the other hand, would be roughly 1418equivalent to discounting 50% of the task's allowed memory from being considered 1419as scoring against the task. 1420 1421For backwards compatibility with previous kernels, /proc/<pid>/oom_adj may also 1422be used to tune the badness score. Its acceptable values range from -16 1423(OOM_ADJUST_MIN) to +15 (OOM_ADJUST_MAX) and a special value of -17 1424(OOM_DISABLE) to disable oom killing entirely for that task. Its value is 1425scaled linearly with /proc/<pid>/oom_score_adj. 1426 1427The value of /proc/<pid>/oom_score_adj may be reduced no lower than the last 1428value set by a CAP_SYS_RESOURCE process. To reduce the value any lower 1429requires CAP_SYS_RESOURCE. 1430 1431Caveat: when a parent task is selected, the oom killer will sacrifice any first 1432generation children with separate address spaces instead, if possible. This 1433avoids servers and important system daemons from being killed and loses the 1434minimal amount of work. 1435 1436 14373.2 /proc/<pid>/oom_score - Display current oom-killer score 1438------------------------------------------------------------- 1439 1440This file can be used to check the current score used by the oom-killer is for 1441any given <pid>. Use it together with /proc/<pid>/oom_score_adj to tune which 1442process should be killed in an out-of-memory situation. 1443 1444 14453.3 /proc/<pid>/io - Display the IO accounting fields 1446------------------------------------------------------- 1447 1448This file contains IO statistics for each running process 1449 1450Example 1451------- 1452 1453test:/tmp # dd if=/dev/zero of=/tmp/test.dat & 1454[1] 3828 1455 1456test:/tmp # cat /proc/3828/io 1457rchar: 323934931 1458wchar: 323929600 1459syscr: 632687 1460syscw: 632675 1461read_bytes: 0 1462write_bytes: 323932160 1463cancelled_write_bytes: 0 1464 1465 1466Description 1467----------- 1468 1469rchar 1470----- 1471 1472I/O counter: chars read 1473The number of bytes which this task has caused to be read from storage. This 1474is simply the sum of bytes which this process passed to read() and pread(). 1475It includes things like tty IO and it is unaffected by whether or not actual 1476physical disk IO was required (the read might have been satisfied from 1477pagecache) 1478 1479 1480wchar 1481----- 1482 1483I/O counter: chars written 1484The number of bytes which this task has caused, or shall cause to be written 1485to disk. Similar caveats apply here as with rchar. 1486 1487 1488syscr 1489----- 1490 1491I/O counter: read syscalls 1492Attempt to count the number of read I/O operations, i.e. syscalls like read() 1493and pread(). 1494 1495 1496syscw 1497----- 1498 1499I/O counter: write syscalls 1500Attempt to count the number of write I/O operations, i.e. syscalls like 1501write() and pwrite(). 1502 1503 1504read_bytes 1505---------- 1506 1507I/O counter: bytes read 1508Attempt to count the number of bytes which this process really did cause to 1509be fetched from the storage layer. Done at the submit_bio() level, so it is 1510accurate for block-backed filesystems. <please add status regarding NFS and 1511CIFS at a later time> 1512 1513 1514write_bytes 1515----------- 1516 1517I/O counter: bytes written 1518Attempt to count the number of bytes which this process caused to be sent to 1519the storage layer. This is done at page-dirtying time. 1520 1521 1522cancelled_write_bytes 1523--------------------- 1524 1525The big inaccuracy here is truncate. If a process writes 1MB to a file and 1526then deletes the file, it will in fact perform no writeout. But it will have 1527been accounted as having caused 1MB of write. 1528In other words: The number of bytes which this process caused to not happen, 1529by truncating pagecache. A task can cause "negative" IO too. If this task 1530truncates some dirty pagecache, some IO which another task has been accounted 1531for (in its write_bytes) will not be happening. We _could_ just subtract that 1532from the truncating task's write_bytes, but there is information loss in doing 1533that. 1534 1535 1536Note 1537---- 1538 1539At its current implementation state, this is a bit racy on 32-bit machines: if 1540process A reads process B's /proc/pid/io while process B is updating one of 1541those 64-bit counters, process A could see an intermediate result. 1542 1543 1544More information about this can be found within the taskstats documentation in 1545Documentation/accounting. 1546 15473.4 /proc/<pid>/coredump_filter - Core dump filtering settings 1548--------------------------------------------------------------- 1549When a process is dumped, all anonymous memory is written to a core file as 1550long as the size of the core file isn't limited. But sometimes we don't want 1551to dump some memory segments, for example, huge shared memory. Conversely, 1552sometimes we want to save file-backed memory segments into a core file, not 1553only the individual files. 1554 1555/proc/<pid>/coredump_filter allows you to customize which memory segments 1556will be dumped when the <pid> process is dumped. coredump_filter is a bitmask 1557of memory types. If a bit of the bitmask is set, memory segments of the 1558corresponding memory type are dumped, otherwise they are not dumped. 1559 1560The following 7 memory types are supported: 1561 - (bit 0) anonymous private memory 1562 - (bit 1) anonymous shared memory 1563 - (bit 2) file-backed private memory 1564 - (bit 3) file-backed shared memory 1565 - (bit 4) ELF header pages in file-backed private memory areas (it is 1566 effective only if the bit 2 is cleared) 1567 - (bit 5) hugetlb private memory 1568 - (bit 6) hugetlb shared memory 1569 1570 Note that MMIO pages such as frame buffer are never dumped and vDSO pages 1571 are always dumped regardless of the bitmask status. 1572 1573 Note bit 0-4 doesn't effect any hugetlb memory. hugetlb memory are only 1574 effected by bit 5-6. 1575 1576Default value of coredump_filter is 0x23; this means all anonymous memory 1577segments and hugetlb private memory are dumped. 1578 1579If you don't want to dump all shared memory segments attached to pid 1234, 1580write 0x21 to the process's proc file. 1581 1582 $ echo 0x21 > /proc/1234/coredump_filter 1583 1584When a new process is created, the process inherits the bitmask status from its 1585parent. It is useful to set up coredump_filter before the program runs. 1586For example: 1587 1588 $ echo 0x7 > /proc/self/coredump_filter 1589 $ ./some_program 1590 15913.5 /proc/<pid>/mountinfo - Information about mounts 1592-------------------------------------------------------- 1593 1594This file contains lines of the form: 1595 159636 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue 1597(1)(2)(3) (4) (5) (6) (7) (8) (9) (10) (11) 1598 1599(1) mount ID: unique identifier of the mount (may be reused after umount) 1600(2) parent ID: ID of parent (or of self for the top of the mount tree) 1601(3) major:minor: value of st_dev for files on filesystem 1602(4) root: root of the mount within the filesystem 1603(5) mount point: mount point relative to the process's root 1604(6) mount options: per mount options 1605(7) optional fields: zero or more fields of the form "tag[:value]" 1606(8) separator: marks the end of the optional fields 1607(9) filesystem type: name of filesystem of the form "type[.subtype]" 1608(10) mount source: filesystem specific information or "none" 1609(11) super options: per super block options 1610 1611Parsers should ignore all unrecognised optional fields. Currently the 1612possible optional fields are: 1613 1614shared:X mount is shared in peer group X 1615master:X mount is slave to peer group X 1616propagate_from:X mount is slave and receives propagation from peer group X (*) 1617unbindable mount is unbindable 1618 1619(*) X is the closest dominant peer group under the process's root. If 1620X is the immediate master of the mount, or if there's no dominant peer 1621group under the same root, then only the "master:X" field is present 1622and not the "propagate_from:X" field. 1623 1624For more information on mount propagation see: 1625 1626 Documentation/filesystems/sharedsubtree.txt 1627 1628 16293.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm 1630-------------------------------------------------------- 1631These files provide a method to access a tasks comm value. It also allows for 1632a task to set its own or one of its thread siblings comm value. The comm value 1633is limited in size compared to the cmdline value, so writing anything longer 1634then the kernel's TASK_COMM_LEN (currently 16 chars) will result in a truncated 1635comm value. 1636 1637 16383.7 /proc/<pid>/task/<tid>/children - Information about task children 1639------------------------------------------------------------------------- 1640This file provides a fast way to retrieve first level children pids 1641of a task pointed by <pid>/<tid> pair. The format is a space separated 1642stream of pids. 1643 1644Note the "first level" here -- if a child has own children they will 1645not be listed here, one needs to read /proc/<children-pid>/task/<tid>/children 1646to obtain the descendants. 1647 1648Since this interface is intended to be fast and cheap it doesn't 1649guarantee to provide precise results and some children might be 1650skipped, especially if they've exited right after we printed their 1651pids, so one need to either stop or freeze processes being inspected 1652if precise results are needed. 1653 1654 16553.7 /proc/<pid>/fdinfo/<fd> - Information about opened file 1656--------------------------------------------------------------- 1657This file provides information associated with an opened file. The regular 1658files have at least two fields -- 'pos' and 'flags'. The 'pos' represents 1659the current offset of the opened file in decimal form [see lseek(2) for 1660details] and 'flags' denotes the octal O_xxx mask the file has been 1661created with [see open(2) for details]. 1662 1663A typical output is 1664 1665 pos: 0 1666 flags: 0100002 1667 1668The files such as eventfd, fsnotify, signalfd, epoll among the regular pos/flags 1669pair provide additional information particular to the objects they represent. 1670 1671 Eventfd files 1672 ~~~~~~~~~~~~~ 1673 pos: 0 1674 flags: 04002 1675 eventfd-count: 5a 1676 1677 where 'eventfd-count' is hex value of a counter. 1678 1679 Signalfd files 1680 ~~~~~~~~~~~~~~ 1681 pos: 0 1682 flags: 04002 1683 sigmask: 0000000000000200 1684 1685 where 'sigmask' is hex value of the signal mask associated 1686 with a file. 1687 1688 Epoll files 1689 ~~~~~~~~~~~ 1690 pos: 0 1691 flags: 02 1692 tfd: 5 events: 1d data: ffffffffffffffff 1693 1694 where 'tfd' is a target file descriptor number in decimal form, 1695 'events' is events mask being watched and the 'data' is data 1696 associated with a target [see epoll(7) for more details]. 1697 1698 Fsnotify files 1699 ~~~~~~~~~~~~~~ 1700 For inotify files the format is the following 1701 1702 pos: 0 1703 flags: 02000000 1704 inotify wd:3 ino:9e7e sdev:800013 mask:800afce ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:7e9e0000640d1b6d 1705 1706 where 'wd' is a watch descriptor in decimal form, ie a target file 1707 descriptor number, 'ino' and 'sdev' are inode and device where the 1708 target file resides and the 'mask' is the mask of events, all in hex 1709 form [see inotify(7) for more details]. 1710 1711 If the kernel was built with exportfs support, the path to the target 1712 file is encoded as a file handle. The file handle is provided by three 1713 fields 'fhandle-bytes', 'fhandle-type' and 'f_handle', all in hex 1714 format. 1715 1716 If the kernel is built without exportfs support the file handle won't be 1717 printed out. 1718 1719 If there is no inotify mark attached yet the 'inotify' line will be omitted. 1720 1721 For fanotify files the format is 1722 1723 pos: 0 1724 flags: 02 1725 fanotify flags:10 event-flags:0 1726 fanotify mnt_id:12 mflags:40 mask:38 ignored_mask:40000003 1727 fanotify ino:4f969 sdev:800013 mflags:0 mask:3b ignored_mask:40000000 fhandle-bytes:8 fhandle-type:1 f_handle:69f90400c275b5b4 1728 1729 where fanotify 'flags' and 'event-flags' are values used in fanotify_init 1730 call, 'mnt_id' is the mount point identifier, 'mflags' is the value of 1731 flags associated with mark which are tracked separately from events 1732 mask. 'ino', 'sdev' are target inode and device, 'mask' is the events 1733 mask and 'ignored_mask' is the mask of events which are to be ignored. 1734 All in hex format. Incorporation of 'mflags', 'mask' and 'ignored_mask' 1735 does provide information about flags and mask used in fanotify_mark 1736 call [see fsnotify manpage for details]. 1737 1738 While the first three lines are mandatory and always printed, the rest is 1739 optional and may be omitted if no marks created yet. 1740 1741 1742------------------------------------------------------------------------------ 1743Configuring procfs 1744------------------------------------------------------------------------------ 1745 17464.1 Mount options 1747--------------------- 1748 1749The following mount options are supported: 1750 1751 hidepid= Set /proc/<pid>/ access mode. 1752 gid= Set the group authorized to learn processes information. 1753 1754hidepid=0 means classic mode - everybody may access all /proc/<pid>/ directories 1755(default). 1756 1757hidepid=1 means users may not access any /proc/<pid>/ directories but their 1758own. Sensitive files like cmdline, sched*, status are now protected against 1759other users. This makes it impossible to learn whether any user runs 1760specific program (given the program doesn't reveal itself by its behaviour). 1761As an additional bonus, as /proc/<pid>/cmdline is unaccessible for other users, 1762poorly written programs passing sensitive information via program arguments are 1763now protected against local eavesdroppers. 1764 1765hidepid=2 means hidepid=1 plus all /proc/<pid>/ will be fully invisible to other 1766users. It doesn't mean that it hides a fact whether a process with a specific 1767pid value exists (it can be learned by other means, e.g. by "kill -0 $PID"), 1768but it hides process' uid and gid, which may be learned by stat()'ing 1769/proc/<pid>/ otherwise. It greatly complicates an intruder's task of gathering 1770information about running processes, whether some daemon runs with elevated 1771privileges, whether other user runs some sensitive program, whether other users 1772run any program at all, etc. 1773 1774gid= defines a group authorized to learn processes information otherwise 1775prohibited by hidepid=. If you use some daemon like identd which needs to learn 1776information about processes information, just add identd to this group. 1777