Based on kernel version 2.6.32. Page generated on 2009-12-11 16:22 EST.
1 ------------------------------------------------------------------------------ 2 T H E /proc F I L E S Y S T E M 3 ------------------------------------------------------------------------------ 4 /proc/sys Terrehon Bowden <terrehon[AT]pacbell[DOT]net> October 7 1999 5 Bodo Bauer <bb[AT]ricochet[DOT]net> 6 7 2.4.x update Jorge Nerin <comandante[AT]zaralinux[DOT]com> November 14 2000 8 move /proc/sys Shen Feng <shen[AT]cn.fujitsu[DOT]com> April 1 2009 9 ------------------------------------------------------------------------------ 10 Version 1.3 Kernel version 2.2.12 11 Kernel version 2.4.0-test11-pre4 12 ------------------------------------------------------------------------------ 13 fixes/update part 1.1 Stefani Seibold <stefani[AT]seibold[DOT]net> June 9 2009 14 15 Table 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 - Adjust the oom-killer score 37 3.2 /proc/<pid>/oom_score - Display current oom-killer score 38 3.3 /proc/<pid>/io - Display the IO accounting fields 39 3.4 /proc/<pid>/coredump_filter - Core dump filtering settings 40 3.5 /proc/<pid>/mountinfo - Information about mounts 41 42 43 ------------------------------------------------------------------------------ 44 Preface 45 ------------------------------------------------------------------------------ 46 47 0.1 Introduction/Credits 48 ------------------------ 49 50 This documentation is part of a soon (or so we hope) to be released book on 51 the SuSE Linux distribution. As there is no complete documentation for the 52 /proc file system and we've used many freely available sources to write these 53 chapters, it seems only fair to give the work back to the Linux community. 54 This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm 55 afraid it's still far from complete, but we hope it will be useful. As far as 56 we know, it is the first 'all-in-one' document about the /proc file system. It 57 is focused on the Intel x86 hardware, so if you are looking for PPC, ARM, 58 SPARC, AXP, etc., features, you probably won't find what you are looking for. 59 It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But 60 additions and patches are welcome and will be added to this document if you 61 mail them to Bodo. 62 63 We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of 64 other people for help compiling this documentation. We'd also like to extend a 65 special thank you to Andi Kleen for documentation, which we relied on heavily 66 to create this document, as well as the additional information he provided. 67 Thanks to everybody else who contributed source or docs to the Linux kernel 68 and helped create a great piece of software... :) 69 70 If you have any comments, corrections or additions, please don't hesitate to 71 contact Bodo Bauer at bb[AT]ricochet.net[DOT] We'll be happy to add them to this 72 document. 73 74 The latest version of this document is available online at 75 http://skaro.nightcrawler.com/~bb/Docs/Proc as HTML version. 76 77 If the above direction does not works for you, ypu could try the kernel 78 mailing list at linux-kernel[AT]vger.kernel[DOT]org and/or try to reach me at 79 comandante[AT]zaralinux.com[DOT] 80 81 0.2 Legal Stuff 82 --------------- 83 84 We don't guarantee the correctness of this document, and if you come to us 85 complaining about how you screwed up your system because of incorrect 86 documentation, we won't feel responsible... 87 88 ------------------------------------------------------------------------------ 89 CHAPTER 1: COLLECTING SYSTEM INFORMATION 90 ------------------------------------------------------------------------------ 91 92 ------------------------------------------------------------------------------ 93 In This Chapter 94 ------------------------------------------------------------------------------ 95 * Investigating the properties of the pseudo file system /proc and its 96 ability to provide information on the running Linux system 97 * Examining /proc's structure 98 * Uncovering various information about the kernel and the processes running 99 on the system 100 ------------------------------------------------------------------------------ 101 102 103 The proc file system acts as an interface to internal data structures in the 104 kernel. It can be used to obtain information about the system and to change 105 certain kernel parameters at runtime (sysctl). 106 107 First, we'll take a look at the read-only parts of /proc. In Chapter 2, we 108 show you how you can use /proc/sys to change settings. 109 110 1.1 Process-Specific Subdirectories 111 ----------------------------------- 112 113 The directory /proc contains (among other things) one subdirectory for each 114 process running on the system, which is named after the process ID (PID). 115 116 The link self points to the process reading the file system. Each process 117 subdirectory has the entries listed in Table 1-1. 118 119 120 Table 1-1: Process specific entries in /proc 121 .............................................................................. 122 File Content 123 clear_refs Clears page referenced bits shown in smaps output 124 cmdline Command line arguments 125 cpu Current and last cpu in which it was executed (2.4)(smp) 126 cwd Link to the current working directory 127 environ Values of environment variables 128 exe Link to the executable of this process 129 fd Directory, which contains all file descriptors 130 maps Memory maps to executables and library files (2.4) 131 mem Memory held by this process 132 root Link to the root directory of this process 133 stat Process status 134 statm Process memory status information 135 status Process status in human readable form 136 wchan If CONFIG_KALLSYMS is set, a pre-decoded wchan 137 stack Report full stack trace, enable via CONFIG_STACKTRACE 138 smaps a extension based on maps, showing the memory consumption of 139 each mapping 140 .............................................................................. 141 142 For example, to get the status information of a process, all you have to do is 143 read the file /proc/PID/status: 144 145 >cat /proc/self/status 146 Name: cat 147 State: R (running) 148 Tgid: 5452 149 Pid: 5452 150 PPid: 743 151 TracerPid: 0 (2.4) 152 Uid: 501 501 501 501 153 Gid: 100 100 100 100 154 FDSize: 256 155 Groups: 100 14 16 156 VmPeak: 5004 kB 157 VmSize: 5004 kB 158 VmLck: 0 kB 159 VmHWM: 476 kB 160 VmRSS: 476 kB 161 VmData: 156 kB 162 VmStk: 88 kB 163 VmExe: 68 kB 164 VmLib: 1412 kB 165 VmPTE: 20 kb 166 Threads: 1 167 SigQ: 0/28578 168 SigPnd: 0000000000000000 169 ShdPnd: 0000000000000000 170 SigBlk: 0000000000000000 171 SigIgn: 0000000000000000 172 SigCgt: 0000000000000000 173 CapInh: 00000000fffffeff 174 CapPrm: 0000000000000000 175 CapEff: 0000000000000000 176 CapBnd: ffffffffffffffff 177 voluntary_ctxt_switches: 0 178 nonvoluntary_ctxt_switches: 1 179 Stack usage: 12 kB 180 181 This shows you nearly the same information you would get if you viewed it with 182 the ps command. In fact, ps uses the proc file system to obtain its 183 information. But you get a more detailed view of the process by reading the 184 file /proc/PID/status. It fields are described in table 1-2. 185 186 The statm file contains more detailed information about the process 187 memory usage. Its seven fields are explained in Table 1-3. The stat file 188 contains details information about the process itself. Its fields are 189 explained in Table 1-4. 190 191 Table 1-2: Contents of the statm files (as of 2.6.30-rc7) 192 .............................................................................. 193 Field Content 194 Name filename of the executable 195 State state (R is running, S is sleeping, D is sleeping 196 in an uninterruptible wait, Z is zombie, 197 T is traced or stopped) 198 Tgid thread group ID 199 Pid process id 200 PPid process id of the parent process 201 TracerPid PID of process tracing this process (0 if not) 202 Uid Real, effective, saved set, and file system UIDs 203 Gid Real, effective, saved set, and file system GIDs 204 FDSize number of file descriptor slots currently allocated 205 Groups supplementary group list 206 VmPeak peak virtual memory size 207 VmSize total program size 208 VmLck locked memory size 209 VmHWM peak resident set size ("high water mark") 210 VmRSS size of memory portions 211 VmData size of data, stack, and text segments 212 VmStk size of data, stack, and text segments 213 VmExe size of text segment 214 VmLib size of shared library code 215 VmPTE size of page table entries 216 Threads number of threads 217 SigQ number of signals queued/max. number for queue 218 SigPnd bitmap of pending signals for the thread 219 ShdPnd bitmap of shared pending signals for the process 220 SigBlk bitmap of blocked signals 221 SigIgn bitmap of ignored signals 222 SigCgt bitmap of catched signals 223 CapInh bitmap of inheritable capabilities 224 CapPrm bitmap of permitted capabilities 225 CapEff bitmap of effective capabilities 226 CapBnd bitmap of capabilities bounding set 227 Cpus_allowed mask of CPUs on which this process may run 228 Cpus_allowed_list Same as previous, but in "list format" 229 Mems_allowed mask of memory nodes allowed to this process 230 Mems_allowed_list Same as previous, but in "list format" 231 voluntary_ctxt_switches number of voluntary context switches 232 nonvoluntary_ctxt_switches number of non voluntary context switches 233 Stack usage: stack usage high water mark (round up to page size) 234 .............................................................................. 235 236 Table 1-3: Contents of the statm files (as of 2.6.8-rc3) 237 .............................................................................. 238 Field Content 239 size total program size (pages) (same as VmSize in status) 240 resident size of memory portions (pages) (same as VmRSS in status) 241 shared number of pages that are shared (i.e. backed by a file) 242 trs number of pages that are 'code' (not including libs; broken, 243 includes data segment) 244 lrs number of pages of library (always 0 on 2.6) 245 drs number of pages of data/stack (including libs; broken, 246 includes library text) 247 dt number of dirty pages (always 0 on 2.6) 248 .............................................................................. 249 250 251 Table 1-4: Contents of the stat files (as of 2.6.30-rc7) 252 .............................................................................. 253 Field Content 254 pid process id 255 tcomm filename of the executable 256 state state (R is running, S is sleeping, D is sleeping in an 257 uninterruptible wait, Z is zombie, T is traced or stopped) 258 ppid process id of the parent process 259 pgrp pgrp of the process 260 sid session id 261 tty_nr tty the process uses 262 tty_pgrp pgrp of the tty 263 flags task flags 264 min_flt number of minor faults 265 cmin_flt number of minor faults with child's 266 maj_flt number of major faults 267 cmaj_flt number of major faults with child's 268 utime user mode jiffies 269 stime kernel mode jiffies 270 cutime user mode jiffies with child's 271 cstime kernel mode jiffies with child's 272 priority priority level 273 nice nice level 274 num_threads number of threads 275 it_real_value (obsolete, always 0) 276 start_time time the process started after system boot 277 vsize virtual memory size 278 rss resident set memory size 279 rsslim current limit in bytes on the rss 280 start_code address above which program text can run 281 end_code address below which program text can run 282 start_stack address of the start of the stack 283 esp current value of ESP 284 eip current value of EIP 285 pending bitmap of pending signals 286 blocked bitmap of blocked signals 287 sigign bitmap of ignored signals 288 sigcatch bitmap of catched signals 289 wchan address where process went to sleep 290 0 (place holder) 291 0 (place holder) 292 exit_signal signal to send to parent thread on exit 293 task_cpu which CPU the task is scheduled on 294 rt_priority realtime priority 295 policy scheduling policy (man sched_setscheduler) 296 blkio_ticks time spent waiting for block IO 297 gtime guest time of the task in jiffies 298 cgtime guest time of the task children in jiffies 299 .............................................................................. 300 301 The /proc/PID/map file containing the currently mapped memory regions and 302 their access permissions. 303 304 The format is: 305 306 address perms offset dev inode pathname 307 308 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test 309 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test 310 0804a000-0806b000 rw-p 00000000 00:00 0 [heap] 311 a7cb1000-a7cb2000 ---p 00000000 00:00 0 312 a7cb2000-a7eb2000 rw-p 00000000 00:00 0 [threadstack:001ff4b4] 313 a7eb2000-a7eb3000 ---p 00000000 00:00 0 314 a7eb3000-a7ed5000 rw-p 00000000 00:00 0 315 a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6 316 a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6 317 a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6 318 a800b000-a800e000 rw-p 00000000 00:00 0 319 a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0 320 a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0 321 a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0 322 a8024000-a8027000 rw-p 00000000 00:00 0 323 a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2 324 a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2 325 a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2 326 aff35000-aff4a000 rw-p 00000000 00:00 0 [stack] 327 ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso] 328 329 where "address" is the address space in the process that it occupies, "perms" 330 is a set of permissions: 331 332 r = read 333 w = write 334 x = execute 335 s = shared 336 p = private (copy on write) 337 338 "offset" is the offset into the mapping, "dev" is the device (major:minor), and 339 "inode" is the inode on that device. 0 indicates that no inode is associated 340 with the memory region, as the case would be with BSS (uninitialized data). 341 The "pathname" shows the name associated file for this mapping. If the mapping 342 is not associated with a file: 343 344 [heap] = the heap of the program 345 [stack] = the stack of the main process 346 [vdso] = the "virtual dynamic shared object", 347 the kernel system call handler 348 [threadstack:xxxxxxxx] = the stack of the thread, xxxxxxxx is the stack size 349 350 or if empty, the mapping is anonymous. 351 352 353 The /proc/PID/smaps is an extension based on maps, showing the memory 354 consumption for each of the process's mappings. For each of mappings there 355 is a series of lines such as the following: 356 357 08048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash 358 Size: 1084 kB 359 Rss: 892 kB 360 Pss: 374 kB 361 Shared_Clean: 892 kB 362 Shared_Dirty: 0 kB 363 Private_Clean: 0 kB 364 Private_Dirty: 0 kB 365 Referenced: 892 kB 366 Swap: 0 kB 367 KernelPageSize: 4 kB 368 MMUPageSize: 4 kB 369 370 The first of these lines shows the same information as is displayed for the 371 mapping in /proc/PID/maps. The remaining lines show the size of the mapping, 372 the amount of the mapping that is currently resident in RAM, the "proportional 373 set sizeâ (divide each shared page by the number of processes sharing it), the 374 number of clean and dirty shared pages in the mapping, and the number of clean 375 and dirty private pages in the mapping. The "Referenced" indicates the amount 376 of memory currently marked as referenced or accessed. 377 378 This file is only present if the CONFIG_MMU kernel configuration option is 379 enabled. 380 381 The /proc/PID/clear_refs is used to reset the PG_Referenced and ACCESSED/YOUNG 382 bits on both physical and virtual pages associated with a process. 383 To clear the bits for all the pages associated with the process 384 > echo 1 > /proc/PID/clear_refs 385 386 To clear the bits for the anonymous pages associated with the process 387 > echo 2 > /proc/PID/clear_refs 388 389 To clear the bits for the file mapped pages associated with the process 390 > echo 3 > /proc/PID/clear_refs 391 Any other value written to /proc/PID/clear_refs will have no effect. 392 393 394 1.2 Kernel data 395 --------------- 396 397 Similar to the process entries, the kernel data files give information about 398 the running kernel. The files used to obtain this information are contained in 399 /proc and are listed in Table 1-5. Not all of these will be present in your 400 system. It depends on the kernel configuration and the loaded modules, which 401 files are there, and which are missing. 402 403 Table 1-5: Kernel info in /proc 404 .............................................................................. 405 File Content 406 apm Advanced power management info 407 buddyinfo Kernel memory allocator information (see text) (2.5) 408 bus Directory containing bus specific information 409 cmdline Kernel command line 410 cpuinfo Info about the CPU 411 devices Available devices (block and character) 412 dma Used DMS channels 413 filesystems Supported filesystems 414 driver Various drivers grouped here, currently rtc (2.4) 415 execdomains Execdomains, related to security (2.4) 416 fb Frame Buffer devices (2.4) 417 fs File system parameters, currently nfs/exports (2.4) 418 ide Directory containing info about the IDE subsystem 419 interrupts Interrupt usage 420 iomem Memory map (2.4) 421 ioports I/O port usage 422 irq Masks for irq to cpu affinity (2.4)(smp?) 423 isapnp ISA PnP (Plug&Play) Info (2.4) 424 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4)) 425 kmsg Kernel messages 426 ksyms Kernel symbol table 427 loadavg Load average of last 1, 5 & 15 minutes 428 locks Kernel locks 429 meminfo Memory info 430 misc Miscellaneous 431 modules List of loaded modules 432 mounts Mounted filesystems 433 net Networking info (see text) 434 partitions Table of partitions known to the system 435 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/, 436 decoupled by lspci (2.4) 437 rtc Real time clock 438 scsi SCSI info (see text) 439 slabinfo Slab pool info 440 softirqs softirq usage 441 stat Overall statistics 442 swaps Swap space utilization 443 sys See chapter 2 444 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4) 445 tty Info of tty drivers 446 uptime System uptime 447 version Kernel version 448 video bttv info of video resources (2.4) 449 vmallocinfo Show vmalloced areas 450 .............................................................................. 451 452 You can, for example, check which interrupts are currently in use and what 453 they are used for by looking in the file /proc/interrupts: 454 455 > cat /proc/interrupts 456 CPU0 457 0: 8728810 XT-PIC timer 458 1: 895 XT-PIC keyboard 459 2: 0 XT-PIC cascade 460 3: 531695 XT-PIC aha152x 461 4: 2014133 XT-PIC serial 462 5: 44401 XT-PIC pcnet_cs 463 8: 2 XT-PIC rtc 464 11: 8 XT-PIC i82365 465 12: 182918 XT-PIC PS/2 Mouse 466 13: 1 XT-PIC fpu 467 14: 1232265 XT-PIC ide0 468 15: 7 XT-PIC ide1 469 NMI: 0 470 471 In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the 472 output of a SMP machine): 473 474 > cat /proc/interrupts 475 476 CPU0 CPU1 477 0: 1243498 1214548 IO-APIC-edge timer 478 1: 8949 8958 IO-APIC-edge keyboard 479 2: 0 0 XT-PIC cascade 480 5: 11286 10161 IO-APIC-edge soundblaster 481 8: 1 0 IO-APIC-edge rtc 482 9: 27422 27407 IO-APIC-edge 3c503 483 12: 113645 113873 IO-APIC-edge PS/2 Mouse 484 13: 0 0 XT-PIC fpu 485 14: 22491 24012 IO-APIC-edge ide0 486 15: 2183 2415 IO-APIC-edge ide1 487 17: 30564 30414 IO-APIC-level eth0 488 18: 177 164 IO-APIC-level bttv 489 NMI: 2457961 2457959 490 LOC: 2457882 2457881 491 ERR: 2155 492 493 NMI is incremented in this case because every timer interrupt generates a NMI 494 (Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups. 495 496 LOC is the local interrupt counter of the internal APIC of every CPU. 497 498 ERR is incremented in the case of errors in the IO-APIC bus (the bus that 499 connects the CPUs in a SMP system. This means that an error has been detected, 500 the IO-APIC automatically retry the transmission, so it should not be a big 501 problem, but you should read the SMP-FAQ. 502 503 In 2.6.2* /proc/interrupts was expanded again. This time the goal was for 504 /proc/interrupts to display every IRQ vector in use by the system, not 505 just those considered 'most important'. The new vectors are: 506 507 THR -- interrupt raised when a machine check threshold counter 508 (typically counting ECC corrected errors of memory or cache) exceeds 509 a configurable threshold. Only available on some systems. 510 511 TRM -- a thermal event interrupt occurs when a temperature threshold 512 has been exceeded for the CPU. This interrupt may also be generated 513 when the temperature drops back to normal. 514 515 SPU -- a spurious interrupt is some interrupt that was raised then lowered 516 by some IO device before it could be fully processed by the APIC. Hence 517 the APIC sees the interrupt but does not know what device it came from. 518 For this case the APIC will generate the interrupt with a IRQ vector 519 of 0xff. This might also be generated by chipset bugs. 520 521 RES, CAL, TLB -- rescheduling, call and TLB flush interrupts are 522 sent from one CPU to another per the needs of the OS. Typically, 523 their statistics are used by kernel developers and interested users to 524 determine the occurrence of interrupts of the given type. 525 526 The above IRQ vectors are displayed only when relevent. For example, 527 the threshold vector does not exist on x86_64 platforms. Others are 528 suppressed when the system is a uniprocessor. As of this writing, only 529 i386 and x86_64 platforms support the new IRQ vector displays. 530 531 Of some interest is the introduction of the /proc/irq directory to 2.4. 532 It could be used to set IRQ to CPU affinity, this means that you can "hook" an 533 IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the 534 irq subdir is one subdir for each IRQ, and two files; default_smp_affinity and 535 prof_cpu_mask. 536 537 For example 538 > ls /proc/irq/ 539 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask 540 1 11 13 15 17 19 3 5 7 9 default_smp_affinity 541 > ls /proc/irq/0/ 542 smp_affinity 543 544 smp_affinity is a bitmask, in which you can specify which CPUs can handle the 545 IRQ, you can set it by doing: 546 547 > echo 1 > /proc/irq/10/smp_affinity 548 549 This means that only the first CPU will handle the IRQ, but you can also echo 550 5 which means that only the first and fourth CPU can handle the IRQ. 551 552 The contents of each smp_affinity file is the same by default: 553 554 > cat /proc/irq/0/smp_affinity 555 ffffffff 556 557 The default_smp_affinity mask applies to all non-active IRQs, which are the 558 IRQs which have not yet been allocated/activated, and hence which lack a 559 /proc/irq/[0-9]* directory. 560 561 prof_cpu_mask specifies which CPUs are to be profiled by the system wide 562 profiler. Default value is ffffffff (all cpus). 563 564 The way IRQs are routed is handled by the IO-APIC, and it's Round Robin 565 between all the CPUs which are allowed to handle it. As usual the kernel has 566 more info than you and does a better job than you, so the defaults are the 567 best choice for almost everyone. 568 569 There are three more important subdirectories in /proc: net, scsi, and sys. 570 The general rule is that the contents, or even the existence of these 571 directories, depend on your kernel configuration. If SCSI is not enabled, the 572 directory scsi may not exist. The same is true with the net, which is there 573 only when networking support is present in the running kernel. 574 575 The slabinfo file gives information about memory usage at the slab level. 576 Linux uses slab pools for memory management above page level in version 2.2. 577 Commonly used objects have their own slab pool (such as network buffers, 578 directory cache, and so on). 579 580 .............................................................................. 581 582 > cat /proc/buddyinfo 583 584 Node 0, zone DMA 0 4 5 4 4 3 ... 585 Node 0, zone Normal 1 0 0 1 101 8 ... 586 Node 0, zone HighMem 2 0 0 1 1 0 ... 587 588 Memory fragmentation is a problem under some workloads, and buddyinfo is a 589 useful tool for helping diagnose these problems. Buddyinfo will give you a 590 clue as to how big an area you can safely allocate, or why a previous 591 allocation failed. 592 593 Each column represents the number of pages of a certain order which are 594 available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in 595 ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE 596 available in ZONE_NORMAL, etc... 597 598 .............................................................................. 599 600 meminfo: 601 602 Provides information about distribution and utilization of memory. This 603 varies by architecture and compile options. The following is from a 604 16GB PIII, which has highmem enabled. You may not have all of these fields. 605 606 > cat /proc/meminfo 607 608 609 MemTotal: 16344972 kB 610 MemFree: 13634064 kB 611 Buffers: 3656 kB 612 Cached: 1195708 kB 613 SwapCached: 0 kB 614 Active: 891636 kB 615 Inactive: 1077224 kB 616 HighTotal: 15597528 kB 617 HighFree: 13629632 kB 618 LowTotal: 747444 kB 619 LowFree: 4432 kB 620 SwapTotal: 0 kB 621 SwapFree: 0 kB 622 Dirty: 968 kB 623 Writeback: 0 kB 624 AnonPages: 861800 kB 625 Mapped: 280372 kB 626 Slab: 284364 kB 627 SReclaimable: 159856 kB 628 SUnreclaim: 124508 kB 629 PageTables: 24448 kB 630 NFS_Unstable: 0 kB 631 Bounce: 0 kB 632 WritebackTmp: 0 kB 633 CommitLimit: 7669796 kB 634 Committed_AS: 100056 kB 635 VmallocTotal: 112216 kB 636 VmallocUsed: 428 kB 637 VmallocChunk: 111088 kB 638 639 MemTotal: Total usable ram (i.e. physical ram minus a few reserved 640 bits and the kernel binary code) 641 MemFree: The sum of LowFree+HighFree 642 Buffers: Relatively temporary storage for raw disk blocks 643 shouldn't get tremendously large (20MB or so) 644 Cached: in-memory cache for files read from the disk (the 645 pagecache). Doesn't include SwapCached 646 SwapCached: Memory that once was swapped out, is swapped back in but 647 still also is in the swapfile (if memory is needed it 648 doesn't need to be swapped out AGAIN because it is already 649 in the swapfile. This saves I/O) 650 Active: Memory that has been used more recently and usually not 651 reclaimed unless absolutely necessary. 652 Inactive: Memory which has been less recently used. It is more 653 eligible to be reclaimed for other purposes 654 HighTotal: 655 HighFree: Highmem is all memory above ~860MB of physical memory 656 Highmem areas are for use by userspace programs, or 657 for the pagecache. The kernel must use tricks to access 658 this memory, making it slower to access than lowmem. 659 LowTotal: 660 LowFree: Lowmem is memory which can be used for everything that 661 highmem can be used for, but it is also available for the 662 kernel's use for its own data structures. Among many 663 other things, it is where everything from the Slab is 664 allocated. Bad things happen when you're out of lowmem. 665 SwapTotal: total amount of swap space available 666 SwapFree: Memory which has been evicted from RAM, and is temporarily 667 on the disk 668 Dirty: Memory which is waiting to get written back to the disk 669 Writeback: Memory which is actively being written back to the disk 670 AnonPages: Non-file backed pages mapped into userspace page tables 671 Mapped: files which have been mmaped, such as libraries 672 Slab: in-kernel data structures cache 673 SReclaimable: Part of Slab, that might be reclaimed, such as caches 674 SUnreclaim: Part of Slab, that cannot be reclaimed on memory pressure 675 PageTables: amount of memory dedicated to the lowest level of page 676 tables. 677 NFS_Unstable: NFS pages sent to the server, but not yet committed to stable 678 storage 679 Bounce: Memory used for block device "bounce buffers" 680 WritebackTmp: Memory used by FUSE for temporary writeback buffers 681 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'), 682 this is the total amount of memory currently available to 683 be allocated on the system. This limit is only adhered to 684 if strict overcommit accounting is enabled (mode 2 in 685 'vm.overcommit_memory'). 686 The CommitLimit is calculated with the following formula: 687 CommitLimit = ('vm.overcommit_ratio' * Physical RAM) + Swap 688 For example, on a system with 1G of physical RAM and 7G 689 of swap with a `vm.overcommit_ratio` of 30 it would 690 yield a CommitLimit of 7.3G. 691 For more details, see the memory overcommit documentation 692 in vm/overcommit-accounting. 693 Committed_AS: The amount of memory presently allocated on the system. 694 The committed memory is a sum of all of the memory which 695 has been allocated by processes, even if it has not been 696 "used" by them as of yet. A process which malloc()'s 1G 697 of memory, but only touches 300M of it will only show up 698 as using 300M of memory even if it has the address space 699 allocated for the entire 1G. This 1G is memory which has 700 been "committed" to by the VM and can be used at any time 701 by the allocating application. With strict overcommit 702 enabled on the system (mode 2 in 'vm.overcommit_memory'), 703 allocations which would exceed the CommitLimit (detailed 704 above) will not be permitted. This is useful if one needs 705 to guarantee that processes will not fail due to lack of 706 memory once that memory has been successfully allocated. 707 VmallocTotal: total size of vmalloc memory area 708 VmallocUsed: amount of vmalloc area which is used 709 VmallocChunk: largest contiguous block of vmalloc area which is free 710 711 .............................................................................. 712 713 vmallocinfo: 714 715 Provides information about vmalloced/vmaped areas. One line per area, 716 containing the virtual address range of the area, size in bytes, 717 caller information of the creator, and optional information depending 718 on the kind of area : 719 720 pages=nr number of pages 721 phys=addr if a physical address was specified 722 ioremap I/O mapping (ioremap() and friends) 723 vmalloc vmalloc() area 724 vmap vmap()ed pages 725 user VM_USERMAP area 726 vpages buffer for pages pointers was vmalloced (huge area) 727 N<node>=nr (Only on NUMA kernels) 728 Number of pages allocated on memory node <node> 729 730 > cat /proc/vmallocinfo 731 0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ... 732 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128 733 0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ... 734 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64 735 0xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f... 736 phys=7fee8000 ioremap 737 0xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f... 738 phys=7fee7000 ioremap 739 0xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210 740 0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ... 741 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3 742 0xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ... 743 pages=2 vmalloc N1=2 744 0xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ... 745 /0x130 [x_tables] pages=4 vmalloc N0=4 746 0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ... 747 pages=14 vmalloc N2=14 748 0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ... 749 pages=4 vmalloc N1=4 750 0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ... 751 pages=2 vmalloc N1=2 752 0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ... 753 pages=10 vmalloc N0=10 754 755 .............................................................................. 756 757 softirqs: 758 759 Provides counts of softirq handlers serviced since boot time, for each cpu. 760 761 > cat /proc/softirqs 762 CPU0 CPU1 CPU2 CPU3 763 HI: 0 0 0 0 764 TIMER: 27166 27120 27097 27034 765 NET_TX: 0 0 0 17 766 NET_RX: 42 0 0 39 767 BLOCK: 0 0 107 1121 768 TASKLET: 0 0 0 290 769 SCHED: 27035 26983 26971 26746 770 HRTIMER: 0 0 0 0 771 RCU: 1678 1769 2178 2250 772 773 774 1.3 IDE devices in /proc/ide 775 ---------------------------- 776 777 The subdirectory /proc/ide contains information about all IDE devices of which 778 the kernel is aware. There is one subdirectory for each IDE controller, the 779 file drivers and a link for each IDE device, pointing to the device directory 780 in the controller specific subtree. 781 782 The file drivers contains general information about the drivers used for the 783 IDE devices: 784 785 > cat /proc/ide/drivers 786 ide-cdrom version 4.53 787 ide-disk version 1.08 788 789 More detailed information can be found in the controller specific 790 subdirectories. These are named ide0, ide1 and so on. Each of these 791 directories contains the files shown in table 1-6. 792 793 794 Table 1-6: IDE controller info in /proc/ide/ide? 795 .............................................................................. 796 File Content 797 channel IDE channel (0 or 1) 798 config Configuration (only for PCI/IDE bridge) 799 mate Mate name 800 model Type/Chipset of IDE controller 801 .............................................................................. 802 803 Each device connected to a controller has a separate subdirectory in the 804 controllers directory. The files listed in table 1-7 are contained in these 805 directories. 806 807 808 Table 1-7: IDE device information 809 .............................................................................. 810 File Content 811 cache The cache 812 capacity Capacity of the medium (in 512Byte blocks) 813 driver driver and version 814 geometry physical and logical geometry 815 identify device identify block 816 media media type 817 model device identifier 818 settings device setup 819 smart_thresholds IDE disk management thresholds 820 smart_values IDE disk management values 821 .............................................................................. 822 823 The most interesting file is settings. This file contains a nice overview of 824 the drive parameters: 825 826 # cat /proc/ide/ide0/hda/settings 827 name value min max mode 828 ---- ----- --- --- ---- 829 bios_cyl 526 0 65535 rw 830 bios_head 255 0 255 rw 831 bios_sect 63 0 63 rw 832 breada_readahead 4 0 127 rw 833 bswap 0 0 1 r 834 file_readahead 72 0 2097151 rw 835 io_32bit 0 0 3 rw 836 keepsettings 0 0 1 rw 837 max_kb_per_request 122 1 127 rw 838 multcount 0 0 8 rw 839 nice1 1 0 1 rw 840 nowerr 0 0 1 rw 841 pio_mode write-only 0 255 w 842 slow 0 0 1 rw 843 unmaskirq 0 0 1 rw 844 using_dma 0 0 1 rw 845 846 847 1.4 Networking info in /proc/net 848 -------------------------------- 849 850 The subdirectory /proc/net follows the usual pattern. Table 1-8 shows the 851 additional values you get for IP version 6 if you configure the kernel to 852 support this. Table 1-9 lists the files and their meaning. 853 854 855 Table 1-8: IPv6 info in /proc/net 856 .............................................................................. 857 File Content 858 udp6 UDP sockets (IPv6) 859 tcp6 TCP sockets (IPv6) 860 raw6 Raw device statistics (IPv6) 861 igmp6 IP multicast addresses, which this host joined (IPv6) 862 if_inet6 List of IPv6 interface addresses 863 ipv6_route Kernel routing table for IPv6 864 rt6_stats Global IPv6 routing tables statistics 865 sockstat6 Socket statistics (IPv6) 866 snmp6 Snmp data (IPv6) 867 .............................................................................. 868 869 870 Table 1-9: Network info in /proc/net 871 .............................................................................. 872 File Content 873 arp Kernel ARP table 874 dev network devices with statistics 875 dev_mcast the Layer2 multicast groups a device is listening too 876 (interface index, label, number of references, number of bound 877 addresses). 878 dev_stat network device status 879 ip_fwchains Firewall chain linkage 880 ip_fwnames Firewall chain names 881 ip_masq Directory containing the masquerading tables 882 ip_masquerade Major masquerading table 883 netstat Network statistics 884 raw raw device statistics 885 route Kernel routing table 886 rpc Directory containing rpc info 887 rt_cache Routing cache 888 snmp SNMP data 889 sockstat Socket statistics 890 tcp TCP sockets 891 tr_rif Token ring RIF routing table 892 udp UDP sockets 893 unix UNIX domain sockets 894 wireless Wireless interface data (Wavelan etc) 895 igmp IP multicast addresses, which this host joined 896 psched Global packet scheduler parameters. 897 netlink List of PF_NETLINK sockets 898 ip_mr_vifs List of multicast virtual interfaces 899 ip_mr_cache List of multicast routing cache 900 .............................................................................. 901 902 You can use this information to see which network devices are available in 903 your system and how much traffic was routed over those devices: 904 905 > cat /proc/net/dev 906 Inter-|Receive |[... 907 face |bytes packets errs drop fifo frame compressed multicast|[... 908 lo: 908188 5596 0 0 0 0 0 0 [... 909 ppp0:15475140 20721 410 0 0 410 0 0 [... 910 eth0: 614530 7085 0 0 0 0 0 1 [... 911 912 ...] Transmit 913 ...] bytes packets errs drop fifo colls carrier compressed 914 ...] 908188 5596 0 0 0 0 0 0 915 ...] 1375103 17405 0 0 0 0 0 0 916 ...] 1703981 5535 0 0 0 3 0 0 917 918 In addition, each Channel Bond interface has it's own directory. For 919 example, the bond0 device will have a directory called /proc/net/bond0/. 920 It will contain information that is specific to that bond, such as the 921 current slaves of the bond, the link status of the slaves, and how 922 many times the slaves link has failed. 923 924 1.5 SCSI info 925 ------------- 926 927 If you have a SCSI host adapter in your system, you'll find a subdirectory 928 named after the driver for this adapter in /proc/scsi. You'll also see a list 929 of all recognized SCSI devices in /proc/scsi: 930 931 >cat /proc/scsi/scsi 932 Attached devices: 933 Host: scsi0 Channel: 00 Id: 00 Lun: 00 934 Vendor: IBM Model: DGHS09U Rev: 03E0 935 Type: Direct-Access ANSI SCSI revision: 03 936 Host: scsi0 Channel: 00 Id: 06 Lun: 00 937 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04 938 Type: CD-ROM ANSI SCSI revision: 02 939 940 941 The directory named after the driver has one file for each adapter found in 942 the system. These files contain information about the controller, including 943 the used IRQ and the IO address range. The amount of information shown is 944 dependent on the adapter you use. The example shows the output for an Adaptec 945 AHA-2940 SCSI adapter: 946 947 > cat /proc/scsi/aic7xxx/0 948 949 Adaptec AIC7xxx driver version: 5.1.19/3.2.4 950 Compile Options: 951 TCQ Enabled By Default : Disabled 952 AIC7XXX_PROC_STATS : Disabled 953 AIC7XXX_RESET_DELAY : 5 954 Adapter Configuration: 955 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter 956 Ultra Wide Controller 957 PCI MMAPed I/O Base: 0xeb001000 958 Adapter SEEPROM Config: SEEPROM found and used. 959 Adaptec SCSI BIOS: Enabled 960 IRQ: 10 961 SCBs: Active 0, Max Active 2, 962 Allocated 15, HW 16, Page 255 963 Interrupts: 160328 964 BIOS Control Word: 0x18b6 965 Adapter Control Word: 0x005b 966 Extended Translation: Enabled 967 Disconnect Enable Flags: 0xffff 968 Ultra Enable Flags: 0x0001 969 Tag Queue Enable Flags: 0x0000 970 Ordered Queue Tag Flags: 0x0000 971 Default Tag Queue Depth: 8 972 Tagged Queue By Device array for aic7xxx host instance 0: 973 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255} 974 Actual queue depth per device for aic7xxx host instance 0: 975 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1} 976 Statistics: 977 (scsi0:0:0:0) 978 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8 979 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0) 980 Total transfers 160151 (74577 reads and 85574 writes) 981 (scsi0:0:6:0) 982 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15 983 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0) 984 Total transfers 0 (0 reads and 0 writes) 985 986 987 1.6 Parallel port info in /proc/parport 988 --------------------------------------- 989 990 The directory /proc/parport contains information about the parallel ports of 991 your system. It has one subdirectory for each port, named after the port 992 number (0,1,2,...). 993 994 These directories contain the four files shown in Table 1-10. 995 996 997 Table 1-10: Files in /proc/parport 998 .............................................................................. 999 File Content 1000 autoprobe Any IEEE-1284 device ID information that has been acquired. 1001 devices list of the device drivers using that port. A + will appear by the 1002 name of the device currently using the port (it might not appear 1003 against any). 1004 hardware Parallel port's base address, IRQ line and DMA channel. 1005 irq IRQ that parport is using for that port. This is in a separate 1006 file to allow you to alter it by writing a new value in (IRQ 1007 number or none). 1008 .............................................................................. 1009 1010 1.7 TTY info in /proc/tty 1011 ------------------------- 1012 1013 Information about the available and actually used tty's can be found in the 1014 directory /proc/tty.You'll find entries for drivers and line disciplines in 1015 this directory, as shown in Table 1-11. 1016 1017 1018 Table 1-11: Files in /proc/tty 1019 .............................................................................. 1020 File Content 1021 drivers list of drivers and their usage 1022 ldiscs registered line disciplines 1023 driver/serial usage statistic and status of single tty lines 1024 .............................................................................. 1025 1026 To see which tty's are currently in use, you can simply look into the file 1027 /proc/tty/drivers: 1028 1029 > cat /proc/tty/drivers 1030 pty_slave /dev/pts 136 0-255 pty:slave 1031 pty_master /dev/ptm 128 0-255 pty:master 1032 pty_slave /dev/ttyp 3 0-255 pty:slave 1033 pty_master /dev/pty 2 0-255 pty:master 1034 serial /dev/cua 5 64-67 serial:callout 1035 serial /dev/ttyS 4 64-67 serial 1036 /dev/tty0 /dev/tty0 4 0 system:vtmaster 1037 /dev/ptmx /dev/ptmx 5 2 system 1038 /dev/console /dev/console 5 1 system:console 1039 /dev/tty /dev/tty 5 0 system:/dev/tty 1040 unknown /dev/tty 4 1-63 console 1041 1042 1043 1.8 Miscellaneous kernel statistics in /proc/stat 1044 ------------------------------------------------- 1045 1046 Various pieces of information about kernel activity are available in the 1047 /proc/stat file. All of the numbers reported in this file are aggregates 1048 since the system first booted. For a quick look, simply cat the file: 1049 1050 > cat /proc/stat 1051 cpu 2255 34 2290 22625563 6290 127 456 0 0 1052 cpu0 1132 34 1441 11311718 3675 127 438 0 0 1053 cpu1 1123 0 849 11313845 2614 0 18 0 0 1054 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...] 1055 ctxt 1990473 1056 btime 1062191376 1057 processes 2915 1058 procs_running 1 1059 procs_blocked 0 1060 softirq 183433 0 21755 12 39 1137 231 21459 2263 1061 1062 The very first "cpu" line aggregates the numbers in all of the other "cpuN" 1063 lines. These numbers identify the amount of time the CPU has spent performing 1064 different kinds of work. Time units are in USER_HZ (typically hundredths of a 1065 second). The meanings of the columns are as follows, from left to right: 1066 1067 - user: normal processes executing in user mode 1068 - nice: niced processes executing in user mode 1069 - system: processes executing in kernel mode 1070 - idle: twiddling thumbs 1071 - iowait: waiting for I/O to complete 1072 - irq: servicing interrupts 1073 - softirq: servicing softirqs 1074 - steal: involuntary wait 1075 - guest: running a guest 1076 1077 The "intr" line gives counts of interrupts serviced since boot time, for each 1078 of the possible system interrupts. The first column is the total of all 1079 interrupts serviced; each subsequent column is the total for that particular 1080 interrupt. 1081 1082 The "ctxt" line gives the total number of context switches across all CPUs. 1083 1084 The "btime" line gives the time at which the system booted, in seconds since 1085 the Unix epoch. 1086 1087 The "processes" line gives the number of processes and threads created, which 1088 includes (but is not limited to) those created by calls to the fork() and 1089 clone() system calls. 1090 1091 The "procs_running" line gives the number of processes currently running on 1092 CPUs. 1093 1094 The "procs_blocked" line gives the number of processes currently blocked, 1095 waiting for I/O to complete. 1096 1097 The "softirq" line gives counts of softirqs serviced since boot time, for each 1098 of the possible system softirqs. The first column is the total of all 1099 softirqs serviced; each subsequent column is the total for that particular 1100 softirq. 1101 1102 1103 1.9 Ext4 file system parameters 1104 ------------------------------ 1105 1106 Information about mounted ext4 file systems can be found in 1107 /proc/fs/ext4. Each mounted filesystem will have a directory in 1108 /proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or 1109 /proc/fs/ext4/dm-0). The files in each per-device directory are shown 1110 in Table 1-12, below. 1111 1112 Table 1-12: Files in /proc/fs/ext4/<devname> 1113 .............................................................................. 1114 File Content 1115 mb_groups details of multiblock allocator buddy cache of free blocks 1116 .............................................................................. 1117 1118 1119 ------------------------------------------------------------------------------ 1120 Summary 1121 ------------------------------------------------------------------------------ 1122 The /proc file system serves information about the running system. It not only 1123 allows access to process data but also allows you to request the kernel status 1124 by reading files in the hierarchy. 1125 1126 The directory structure of /proc reflects the types of information and makes 1127 it easy, if not obvious, where to look for specific data. 1128 ------------------------------------------------------------------------------ 1129 1130 ------------------------------------------------------------------------------ 1131 CHAPTER 2: MODIFYING SYSTEM PARAMETERS 1132 ------------------------------------------------------------------------------ 1133 1134 ------------------------------------------------------------------------------ 1135 In This Chapter 1136 ------------------------------------------------------------------------------ 1137 * Modifying kernel parameters by writing into files found in /proc/sys 1138 * Exploring the files which modify certain parameters 1139 * Review of the /proc/sys file tree 1140 ------------------------------------------------------------------------------ 1141 1142 1143 A very interesting part of /proc is the directory /proc/sys. This is not only 1144 a source of information, it also allows you to change parameters within the 1145 kernel. Be very careful when attempting this. You can optimize your system, 1146 but you can also cause it to crash. Never alter kernel parameters on a 1147 production system. Set up a development machine and test to make sure that 1148 everything works the way you want it to. You may have no alternative but to 1149 reboot the machine once an error has been made. 1150 1151 To change a value, simply echo the new value into the file. An example is 1152 given below in the section on the file system data. You need to be root to do 1153 this. You can create your own boot script to perform this every time your 1154 system boots. 1155 1156 The files in /proc/sys can be used to fine tune and monitor miscellaneous and 1157 general things in the operation of the Linux kernel. Since some of the files 1158 can inadvertently disrupt your system, it is advisable to read both 1159 documentation and source before actually making adjustments. In any case, be 1160 very careful when writing to any of these files. The entries in /proc may 1161 change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt 1162 review the kernel documentation in the directory /usr/src/linux/Documentation. 1163 This chapter is heavily based on the documentation included in the pre 2.2 1164 kernels, and became part of it in version 2.2.1 of the Linux kernel. 1165 1166 Please see: Documentation/sysctls/ directory for descriptions of these 1167 entries. 1168 1169 ------------------------------------------------------------------------------ 1170 Summary 1171 ------------------------------------------------------------------------------ 1172 Certain aspects of kernel behavior can be modified at runtime, without the 1173 need to recompile the kernel, or even to reboot the system. The files in the 1174 /proc/sys tree can not only be read, but also modified. You can use the echo 1175 command to write value into these files, thereby changing the default settings 1176 of the kernel. 1177 ------------------------------------------------------------------------------ 1178 1179 ------------------------------------------------------------------------------ 1180 CHAPTER 3: PER-PROCESS PARAMETERS 1181 ------------------------------------------------------------------------------ 1182 1183 3.1 /proc/<pid>/oom_adj - Adjust the oom-killer score 1184 ------------------------------------------------------ 1185 1186 This file can be used to adjust the score used to select which processes 1187 should be killed in an out-of-memory situation. Giving it a high score will 1188 increase the likelihood of this process being killed by the oom-killer. Valid 1189 values are in the range -16 to +15, plus the special value -17, which disables 1190 oom-killing altogether for this process. 1191 1192 The process to be killed in an out-of-memory situation is selected among all others 1193 based on its badness score. This value equals the original memory size of the process 1194 and is then updated according to its CPU time (utime + stime) and the 1195 run time (uptime - start time). The longer it runs the smaller is the score. 1196 Badness score is divided by the square root of the CPU time and then by 1197 the double square root of the run time. 1198 1199 Swapped out tasks are killed first. Half of each child's memory size is added to 1200 the parent's score if they do not share the same memory. Thus forking servers 1201 are the prime candidates to be killed. Having only one 'hungry' child will make 1202 parent less preferable than the child. 1203 1204 /proc/<pid>/oom_score shows process' current badness score. 1205 1206 The following heuristics are then applied: 1207 * if the task was reniced, its score doubles 1208 * superuser or direct hardware access tasks (CAP_SYS_ADMIN, CAP_SYS_RESOURCE 1209 or CAP_SYS_RAWIO) have their score divided by 4 1210 * if oom condition happened in one cpuset and checked process does not belong 1211 to it, its score is divided by 8 1212 * the resulting score is multiplied by two to the power of oom_adj, i.e. 1213 points <<= oom_adj when it is positive and 1214 points >>= -(oom_adj) otherwise 1215 1216 The task with the highest badness score is then selected and its children 1217 are killed, process itself will be killed in an OOM situation when it does 1218 not have children or some of them disabled oom like described above. 1219 1220 3.2 /proc/<pid>/oom_score - Display current oom-killer score 1221 ------------------------------------------------------------- 1222 1223 This file can be used to check the current score used by the oom-killer is for 1224 any given <pid>. Use it together with /proc/<pid>/oom_adj to tune which 1225 process should be killed in an out-of-memory situation. 1226 1227 1228 3.3 /proc/<pid>/io - Display the IO accounting fields 1229 ------------------------------------------------------- 1230 1231 This file contains IO statistics for each running process 1232 1233 Example 1234 ------- 1235 1236 test:/tmp # dd if=/dev/zero of=/tmp/test.dat & 1237 [1] 3828 1238 1239 test:/tmp # cat /proc/3828/io 1240 rchar: 323934931 1241 wchar: 323929600 1242 syscr: 632687 1243 syscw: 632675 1244 read_bytes: 0 1245 write_bytes: 323932160 1246 cancelled_write_bytes: 0 1247 1248 1249 Description 1250 ----------- 1251 1252 rchar 1253 ----- 1254 1255 I/O counter: chars read 1256 The number of bytes which this task has caused to be read from storage. This 1257 is simply the sum of bytes which this process passed to read() and pread(). 1258 It includes things like tty IO and it is unaffected by whether or not actual 1259 physical disk IO was required (the read might have been satisfied from 1260 pagecache) 1261 1262 1263 wchar 1264 ----- 1265 1266 I/O counter: chars written 1267 The number of bytes which this task has caused, or shall cause to be written 1268 to disk. Similar caveats apply here as with rchar. 1269 1270 1271 syscr 1272 ----- 1273 1274 I/O counter: read syscalls 1275 Attempt to count the number of read I/O operations, i.e. syscalls like read() 1276 and pread(). 1277 1278 1279 syscw 1280 ----- 1281 1282 I/O counter: write syscalls 1283 Attempt to count the number of write I/O operations, i.e. syscalls like 1284 write() and pwrite(). 1285 1286 1287 read_bytes 1288 ---------- 1289 1290 I/O counter: bytes read 1291 Attempt to count the number of bytes which this process really did cause to 1292 be fetched from the storage layer. Done at the submit_bio() level, so it is 1293 accurate for block-backed filesystems. <please add status regarding NFS and 1294 CIFS at a later time> 1295 1296 1297 write_bytes 1298 ----------- 1299 1300 I/O counter: bytes written 1301 Attempt to count the number of bytes which this process caused to be sent to 1302 the storage layer. This is done at page-dirtying time. 1303 1304 1305 cancelled_write_bytes 1306 --------------------- 1307 1308 The big inaccuracy here is truncate. If a process writes 1MB to a file and 1309 then deletes the file, it will in fact perform no writeout. But it will have 1310 been accounted as having caused 1MB of write. 1311 In other words: The number of bytes which this process caused to not happen, 1312 by truncating pagecache. A task can cause "negative" IO too. If this task 1313 truncates some dirty pagecache, some IO which another task has been accounted 1314 for (in it's write_bytes) will not be happening. We _could_ just subtract that 1315 from the truncating task's write_bytes, but there is information loss in doing 1316 that. 1317 1318 1319 Note 1320 ---- 1321 1322 At its current implementation state, this is a bit racy on 32-bit machines: if 1323 process A reads process B's /proc/pid/io while process B is updating one of 1324 those 64-bit counters, process A could see an intermediate result. 1325 1326 1327 More information about this can be found within the taskstats documentation in 1328 Documentation/accounting. 1329 1330 3.4 /proc/<pid>/coredump_filter - Core dump filtering settings 1331 --------------------------------------------------------------- 1332 When a process is dumped, all anonymous memory is written to a core file as 1333 long as the size of the core file isn't limited. But sometimes we don't want 1334 to dump some memory segments, for example, huge shared memory. Conversely, 1335 sometimes we want to save file-backed memory segments into a core file, not 1336 only the individual files. 1337 1338 /proc/<pid>/coredump_filter allows you to customize which memory segments 1339 will be dumped when the <pid> process is dumped. coredump_filter is a bitmask 1340 of memory types. If a bit of the bitmask is set, memory segments of the 1341 corresponding memory type are dumped, otherwise they are not dumped. 1342 1343 The following 7 memory types are supported: 1344 - (bit 0) anonymous private memory 1345 - (bit 1) anonymous shared memory 1346 - (bit 2) file-backed private memory 1347 - (bit 3) file-backed shared memory 1348 - (bit 4) ELF header pages in file-backed private memory areas (it is 1349 effective only if the bit 2 is cleared) 1350 - (bit 5) hugetlb private memory 1351 - (bit 6) hugetlb shared memory 1352 1353 Note that MMIO pages such as frame buffer are never dumped and vDSO pages 1354 are always dumped regardless of the bitmask status. 1355 1356 Note bit 0-4 doesn't effect any hugetlb memory. hugetlb memory are only 1357 effected by bit 5-6. 1358 1359 Default value of coredump_filter is 0x23; this means all anonymous memory 1360 segments and hugetlb private memory are dumped. 1361 1362 If you don't want to dump all shared memory segments attached to pid 1234, 1363 write 0x21 to the process's proc file. 1364 1365 $ echo 0x21 > /proc/1234/coredump_filter 1366 1367 When a new process is created, the process inherits the bitmask status from its 1368 parent. It is useful to set up coredump_filter before the program runs. 1369 For example: 1370 1371 $ echo 0x7 > /proc/self/coredump_filter 1372 $ ./some_program 1373 1374 3.5 /proc/<pid>/mountinfo - Information about mounts 1375 -------------------------------------------------------- 1376 1377 This file contains lines of the form: 1378 1379 36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue 1380 (1)(2)(3) (4) (5) (6) (7) (8) (9) (10) (11) 1381 1382 (1) mount ID: unique identifier of the mount (may be reused after umount) 1383 (2) parent ID: ID of parent (or of self for the top of the mount tree) 1384 (3) major:minor: value of st_dev for files on filesystem 1385 (4) root: root of the mount within the filesystem 1386 (5) mount point: mount point relative to the process's root 1387 (6) mount options: per mount options 1388 (7) optional fields: zero or more fields of the form "tag[:value]" 1389 (8) separator: marks the end of the optional fields 1390 (9) filesystem type: name of filesystem of the form "type[.subtype]" 1391 (10) mount source: filesystem specific information or "none" 1392 (11) super options: per super block options 1393 1394 Parsers should ignore all unrecognised optional fields. Currently the 1395 possible optional fields are: 1396 1397 shared:X mount is shared in peer group X 1398 master:X mount is slave to peer group X 1399 propagate_from:X mount is slave and receives propagation from peer group X (*) 1400 unbindable mount is unbindable 1401 1402 (*) X is the closest dominant peer group under the process's root. If 1403 X is the immediate master of the mount, or if there's no dominant peer 1404 group under the same root, then only the "master:X" field is present 1405 and not the "propagate_from:X" field. 1406 1407 For more information on mount propagation see: 1408 1409 Documentation/filesystems/sharedsubtree.txt