| blob | 7aaf09b86a552210584f6e346ad903b10a9cb0b6 |
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>
7 2.4.x update Jorge Nerin <comandante@zaralinux.com> November 14 2000
8 ------------------------------------------------------------------------------
9 Version 1.3 Kernel version 2.2.12
10 Kernel version 2.4.0-test11-pre4
11 ------------------------------------------------------------------------------
13 Table of Contents
14 -----------------
16 0 Preface
17 0.1 Introduction/Credits
18 0.2 Legal Stuff
20 1 Collecting System Information
21 1.1 Process-Specific Subdirectories
22 1.2 Kernel data
23 1.3 IDE devices in /proc/ide
24 1.4 Networking info in /proc/net
25 1.5 SCSI info
26 1.6 Parallel port info in /proc/parport
27 1.7 TTY info in /proc/tty
28 1.8 Miscellaneous kernel statistics in /proc/stat
30 2 Modifying System Parameters
31 2.1 /proc/sys/fs - File system data
32 2.2 /proc/sys/fs/binfmt_misc - Miscellaneous binary formats
33 2.3 /proc/sys/kernel - general kernel parameters
34 2.4 /proc/sys/vm - The virtual memory subsystem
35 2.5 /proc/sys/dev - Device specific parameters
36 2.6 /proc/sys/sunrpc - Remote procedure calls
37 2.7 /proc/sys/net - Networking stuff
38 2.8 /proc/sys/net/ipv4 - IPV4 settings
39 2.9 Appletalk
40 2.10 IPX
41 2.11 /proc/sys/fs/mqueue - POSIX message queues filesystem
42 2.12 /proc/<pid>/oom_adj - Adjust the oom-killer score
43 2.13 /proc/<pid>/oom_score - Display current oom-killer score
44 2.14 /proc/<pid>/io - Display the IO accounting fields
46 ------------------------------------------------------------------------------
47 Preface
48 ------------------------------------------------------------------------------
50 0.1 Introduction/Credits
51 ------------------------
53 This documentation is part of a soon (or so we hope) to be released book on
54 the SuSE Linux distribution. As there is no complete documentation for the
55 /proc file system and we've used many freely available sources to write these
56 chapters, it seems only fair to give the work back to the Linux community.
57 This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm
58 afraid it's still far from complete, but we hope it will be useful. As far as
59 we know, it is the first 'all-in-one' document about the /proc file system. It
60 is focused on the Intel x86 hardware, so if you are looking for PPC, ARM,
61 SPARC, AXP, etc., features, you probably won't find what you are looking for.
62 It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But
63 additions and patches are welcome and will be added to this document if you
64 mail them to Bodo.
66 We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of
67 other people for help compiling this documentation. We'd also like to extend a
68 special thank you to Andi Kleen for documentation, which we relied on heavily
69 to create this document, as well as the additional information he provided.
70 Thanks to everybody else who contributed source or docs to the Linux kernel
71 and helped create a great piece of software... :)
73 If you have any comments, corrections or additions, please don't hesitate to
74 contact Bodo Bauer at bb@ricochet.net. We'll be happy to add them to this
75 document.
77 The latest version of this document is available online at
78 http://skaro.nightcrawler.com/~bb/Docs/Proc as HTML version.
80 If the above direction does not works for you, ypu could try the kernel
81 mailing list at linux-kernel@vger.kernel.org and/or try to reach me at
82 comandante@zaralinux.com.
84 0.2 Legal Stuff
85 ---------------
87 We don't guarantee the correctness of this document, and if you come to us
88 complaining about how you screwed up your system because of incorrect
89 documentation, we won't feel responsible...
91 ------------------------------------------------------------------------------
92 CHAPTER 1: COLLECTING SYSTEM INFORMATION
93 ------------------------------------------------------------------------------
95 ------------------------------------------------------------------------------
96 In This Chapter
97 ------------------------------------------------------------------------------
98 * Investigating the properties of the pseudo file system /proc and its
99 ability to provide information on the running Linux system
100 * Examining /proc's structure
101 * Uncovering various information about the kernel and the processes running
102 on the system
103 ------------------------------------------------------------------------------
106 The proc file system acts as an interface to internal data structures in the
107 kernel. It can be used to obtain information about the system and to change
108 certain kernel parameters at runtime (sysctl).
110 First, we'll take a look at the read-only parts of /proc. In Chapter 2, we
111 show you how you can use /proc/sys to change settings.
113 1.1 Process-Specific Subdirectories
114 -----------------------------------
116 The directory /proc contains (among other things) one subdirectory for each
117 process running on the system, which is named after the process ID (PID).
119 The link self points to the process reading the file system. Each process
120 subdirectory has the entries listed in Table 1-1.
123 Table 1-1: Process specific entries in /proc
124 ..............................................................................
125 File Content
126 cmdline Command line arguments
127 cpu Current and last cpu in which it was executed (2.4)(smp)
128 cwd Link to the current working directory
129 environ Values of environment variables
130 exe Link to the executable of this process
131 fd Directory, which contains all file descriptors
132 maps Memory maps to executables and library files (2.4)
133 mem Memory held by this process
134 root Link to the root directory of this process
135 stat Process status
136 statm Process memory status information
137 status Process status in human readable form
138 wchan If CONFIG_KALLSYMS is set, a pre-decoded wchan
139 smaps Extension based on maps, presenting the rss size for each mapped file
140 ..............................................................................
142 For example, to get the status information of a process, all you have to do is
143 read the file /proc/PID/status:
145 >cat /proc/self/status
146 Name: cat
147 State: R (running)
148 Pid: 5452
149 PPid: 743
150 TracerPid: 0 (2.4)
151 Uid: 501 501 501 501
152 Gid: 100 100 100 100
153 Groups: 100 14 16
154 VmSize: 1112 kB
155 VmLck: 0 kB
156 VmRSS: 348 kB
157 VmData: 24 kB
158 VmStk: 12 kB
159 VmExe: 8 kB
160 VmLib: 1044 kB
161 SigPnd: 0000000000000000
162 SigBlk: 0000000000000000
163 SigIgn: 0000000000000000
164 SigCgt: 0000000000000000
165 CapInh: 00000000fffffeff
166 CapPrm: 0000000000000000
167 CapEff: 0000000000000000
170 This shows you nearly the same information you would get if you viewed it with
171 the ps command. In fact, ps uses the proc file system to obtain its
172 information. The statm file contains more detailed information about the
173 process memory usage. Its seven fields are explained in Table 1-2.
176 Table 1-2: Contents of the statm files (as of 2.6.8-rc3)
177 ..............................................................................
178 Field Content
179 size total program size (pages) (same as VmSize in status)
180 resident size of memory portions (pages) (same as VmRSS in status)
181 shared number of pages that are shared (i.e. backed by a file)
182 trs number of pages that are 'code' (not including libs; broken,
183 includes data segment)
184 lrs number of pages of library (always 0 on 2.6)
185 drs number of pages of data/stack (including libs; broken,
186 includes library text)
187 dt number of dirty pages (always 0 on 2.6)
188 ..............................................................................
190 1.2 Kernel data
191 ---------------
193 Similar to the process entries, the kernel data files give information about
194 the running kernel. The files used to obtain this information are contained in
195 /proc and are listed in Table 1-3. Not all of these will be present in your
196 system. It depends on the kernel configuration and the loaded modules, which
197 files are there, and which are missing.
199 Table 1-3: Kernel info in /proc
200 ..............................................................................
201 File Content
202 apm Advanced power management info
203 buddyinfo Kernel memory allocator information (see text) (2.5)
204 bus Directory containing bus specific information
205 cmdline Kernel command line
206 cpuinfo Info about the CPU
207 devices Available devices (block and character)
208 dma Used DMS channels
209 filesystems Supported filesystems
210 driver Various drivers grouped here, currently rtc (2.4)
211 execdomains Execdomains, related to security (2.4)
212 fb Frame Buffer devices (2.4)
213 fs File system parameters, currently nfs/exports (2.4)
214 ide Directory containing info about the IDE subsystem
215 interrupts Interrupt usage
216 iomem Memory map (2.4)
217 ioports I/O port usage
218 irq Masks for irq to cpu affinity (2.4)(smp?)
219 isapnp ISA PnP (Plug&Play) Info (2.4)
220 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4))
221 kmsg Kernel messages
222 ksyms Kernel symbol table
223 loadavg Load average of last 1, 5 & 15 minutes
224 locks Kernel locks
225 meminfo Memory info
226 misc Miscellaneous
227 modules List of loaded modules
228 mounts Mounted filesystems
229 net Networking info (see text)
230 partitions Table of partitions known to the system
231 pci Depreciated info of PCI bus (new way -> /proc/bus/pci/,
232 decoupled by lspci (2.4)
233 rtc Real time clock
234 scsi SCSI info (see text)
235 slabinfo Slab pool info
236 stat Overall statistics
237 swaps Swap space utilization
238 sys See chapter 2
239 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4)
240 tty Info of tty drivers
241 uptime System uptime
242 version Kernel version
243 video bttv info of video resources (2.4)
244 ..............................................................................
246 You can, for example, check which interrupts are currently in use and what
247 they are used for by looking in the file /proc/interrupts:
249 > cat /proc/interrupts
250 CPU0
251 0: 8728810 XT-PIC timer
252 1: 895 XT-PIC keyboard
253 2: 0 XT-PIC cascade
254 3: 531695 XT-PIC aha152x
255 4: 2014133 XT-PIC serial
256 5: 44401 XT-PIC pcnet_cs
257 8: 2 XT-PIC rtc
258 11: 8 XT-PIC i82365
259 12: 182918 XT-PIC PS/2 Mouse
260 13: 1 XT-PIC fpu
261 14: 1232265 XT-PIC ide0
262 15: 7 XT-PIC ide1
263 NMI: 0
265 In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the
266 output of a SMP machine):
268 > cat /proc/interrupts
270 CPU0 CPU1
271 0: 1243498 1214548 IO-APIC-edge timer
272 1: 8949 8958 IO-APIC-edge keyboard
273 2: 0 0 XT-PIC cascade
274 5: 11286 10161 IO-APIC-edge soundblaster
275 8: 1 0 IO-APIC-edge rtc
276 9: 27422 27407 IO-APIC-edge 3c503
277 12: 113645 113873 IO-APIC-edge PS/2 Mouse
278 13: 0 0 XT-PIC fpu
279 14: 22491 24012 IO-APIC-edge ide0
280 15: 2183 2415 IO-APIC-edge ide1
281 17: 30564 30414 IO-APIC-level eth0
282 18: 177 164 IO-APIC-level bttv
283 NMI: 2457961 2457959
284 LOC: 2457882 2457881
285 ERR: 2155
287 NMI is incremented in this case because every timer interrupt generates a NMI
288 (Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups.
290 LOC is the local interrupt counter of the internal APIC of every CPU.
292 ERR is incremented in the case of errors in the IO-APIC bus (the bus that
293 connects the CPUs in a SMP system. This means that an error has been detected,
294 the IO-APIC automatically retry the transmission, so it should not be a big
295 problem, but you should read the SMP-FAQ.
297 In this context it could be interesting to note the new irq directory in 2.4.
298 It could be used to set IRQ to CPU affinity, this means that you can "hook" an
299 IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the
300 irq subdir is one subdir for each IRQ, and one file; prof_cpu_mask
302 For example
303 > ls /proc/irq/
304 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask
305 1 11 13 15 17 19 3 5 7 9
306 > ls /proc/irq/0/
307 smp_affinity
309 The contents of the prof_cpu_mask file and each smp_affinity file for each IRQ
310 is the same by default:
312 > cat /proc/irq/0/smp_affinity
313 ffffffff
315 It's a bitmask, in which you can specify which CPUs can handle the IRQ, you can
316 set it by doing:
318 > echo 1 > /proc/irq/prof_cpu_mask
320 This means that only the first CPU will handle the IRQ, but you can also echo 5
321 which means that only the first and fourth CPU can handle the IRQ.
323 The way IRQs are routed is handled by the IO-APIC, and it's Round Robin
324 between all the CPUs which are allowed to handle it. As usual the kernel has
325 more info than you and does a better job than you, so the defaults are the
326 best choice for almost everyone.
328 There are three more important subdirectories in /proc: net, scsi, and sys.
329 The general rule is that the contents, or even the existence of these
330 directories, depend on your kernel configuration. If SCSI is not enabled, the
331 directory scsi may not exist. The same is true with the net, which is there
332 only when networking support is present in the running kernel.
334 The slabinfo file gives information about memory usage at the slab level.
335 Linux uses slab pools for memory management above page level in version 2.2.
336 Commonly used objects have their own slab pool (such as network buffers,
337 directory cache, and so on).
339 ..............................................................................
341 > cat /proc/buddyinfo
343 Node 0, zone DMA 0 4 5 4 4 3 ...
344 Node 0, zone Normal 1 0 0 1 101 8 ...
345 Node 0, zone HighMem 2 0 0 1 1 0 ...
347 Memory fragmentation is a problem under some workloads, and buddyinfo is a
348 useful tool for helping diagnose these problems. Buddyinfo will give you a
349 clue as to how big an area you can safely allocate, or why a previous
350 allocation failed.
352 Each column represents the number of pages of a certain order which are
353 available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in
354 ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE
355 available in ZONE_NORMAL, etc...
357 ..............................................................................
359 meminfo:
361 Provides information about distribution and utilization of memory. This
362 varies by architecture and compile options. The following is from a
363 16GB PIII, which has highmem enabled. You may not have all of these fields.
365 > cat /proc/meminfo
368 MemTotal: 16344972 kB
369 MemFree: 13634064 kB
370 Buffers: 3656 kB
371 Cached: 1195708 kB
372 SwapCached: 0 kB
373 Active: 891636 kB
374 Inactive: 1077224 kB
375 HighTotal: 15597528 kB
376 HighFree: 13629632 kB
377 LowTotal: 747444 kB
378 LowFree: 4432 kB
379 SwapTotal: 0 kB
380 SwapFree: 0 kB
381 Dirty: 968 kB
382 Writeback: 0 kB
383 Mapped: 280372 kB
384 Slab: 684068 kB
385 CommitLimit: 7669796 kB
386 Committed_AS: 100056 kB
387 PageTables: 24448 kB
388 VmallocTotal: 112216 kB
389 VmallocUsed: 428 kB
390 VmallocChunk: 111088 kB
392 MemTotal: Total usable ram (i.e. physical ram minus a few reserved
393 bits and the kernel binary code)
394 MemFree: The sum of LowFree+HighFree
395 Buffers: Relatively temporary storage for raw disk blocks
396 shouldn't get tremendously large (20MB or so)
397 Cached: in-memory cache for files read from the disk (the
398 pagecache). Doesn't include SwapCached
399 SwapCached: Memory that once was swapped out, is swapped back in but
400 still also is in the swapfile (if memory is needed it
401 doesn't need to be swapped out AGAIN because it is already
402 in the swapfile. This saves I/O)
403 Active: Memory that has been used more recently and usually not
404 reclaimed unless absolutely necessary.
405 Inactive: Memory which has been less recently used. It is more
406 eligible to be reclaimed for other purposes
407 HighTotal:
408 HighFree: Highmem is all memory above ~860MB of physical memory
409 Highmem areas are for use by userspace programs, or
410 for the pagecache. The kernel must use tricks to access
411 this memory, making it slower to access than lowmem.
412 LowTotal:
413 LowFree: Lowmem is memory which can be used for everything that
414 highmem can be used for, but it is also available for the
415 kernel's use for its own data structures. Among many
416 other things, it is where everything from the Slab is
417 allocated. Bad things happen when you're out of lowmem.
418 SwapTotal: total amount of swap space available
419 SwapFree: Memory which has been evicted from RAM, and is temporarily
420 on the disk
421 Dirty: Memory which is waiting to get written back to the disk
422 Writeback: Memory which is actively being written back to the disk
423 Mapped: files which have been mmaped, such as libraries
424 Slab: in-kernel data structures cache
425 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'),
426 this is the total amount of memory currently available to
427 be allocated on the system. This limit is only adhered to
428 if strict overcommit accounting is enabled (mode 2 in
429 'vm.overcommit_memory').
430 The CommitLimit is calculated with the following formula:
431 CommitLimit = ('vm.overcommit_ratio' * Physical RAM) + Swap
432 For example, on a system with 1G of physical RAM and 7G
433 of swap with a `vm.overcommit_ratio` of 30 it would
434 yield a CommitLimit of 7.3G.
435 For more details, see the memory overcommit documentation
436 in vm/overcommit-accounting.
437 Committed_AS: The amount of memory presently allocated on the system.
438 The committed memory is a sum of all of the memory which
439 has been allocated by processes, even if it has not been
440 "used" by them as of yet. A process which malloc()'s 1G
441 of memory, but only touches 300M of it will only show up
442 as using 300M of memory even if it has the address space
443 allocated for the entire 1G. This 1G is memory which has
444 been "committed" to by the VM and can be used at any time
445 by the allocating application. With strict overcommit
446 enabled on the system (mode 2 in 'vm.overcommit_memory'),
447 allocations which would exceed the CommitLimit (detailed
448 above) will not be permitted. This is useful if one needs
449 to guarantee that processes will not fail due to lack of
450 memory once that memory has been successfully allocated.
451 PageTables: amount of memory dedicated to the lowest level of page
452 tables.
453 VmallocTotal: total size of vmalloc memory area
454 VmallocUsed: amount of vmalloc area which is used
455 VmallocChunk: largest contigious block of vmalloc area which is free
458 1.3 IDE devices in /proc/ide
459 ----------------------------
461 The subdirectory /proc/ide contains information about all IDE devices of which
462 the kernel is aware. There is one subdirectory for each IDE controller, the
463 file drivers and a link for each IDE device, pointing to the device directory
464 in the controller specific subtree.
466 The file drivers contains general information about the drivers used for the
467 IDE devices:
469 > cat /proc/ide/drivers
470 ide-cdrom version 4.53
471 ide-disk version 1.08
473 More detailed information can be found in the controller specific
474 subdirectories. These are named ide0, ide1 and so on. Each of these
475 directories contains the files shown in table 1-4.
478 Table 1-4: IDE controller info in /proc/ide/ide?
479 ..............................................................................
480 File Content
481 channel IDE channel (0 or 1)
482 config Configuration (only for PCI/IDE bridge)
483 mate Mate name
484 model Type/Chipset of IDE controller
485 ..............................................................................
487 Each device connected to a controller has a separate subdirectory in the
488 controllers directory. The files listed in table 1-5 are contained in these
489 directories.
492 Table 1-5: IDE device information
493 ..............................................................................
494 File Content
495 cache The cache
496 capacity Capacity of the medium (in 512Byte blocks)
497 driver driver and version
498 geometry physical and logical geometry
499 identify device identify block
500 media media type
501 model device identifier
502 settings device setup
503 smart_thresholds IDE disk management thresholds
504 smart_values IDE disk management values
505 ..............................................................................
507 The most interesting file is settings. This file contains a nice overview of
508 the drive parameters:
510 # cat /proc/ide/ide0/hda/settings
511 name value min max mode
512 ---- ----- --- --- ----
513 bios_cyl 526 0 65535 rw
514 bios_head 255 0 255 rw
515 bios_sect 63 0 63 rw
516 breada_readahead 4 0 127 rw
517 bswap 0 0 1 r
518 file_readahead 72 0 2097151 rw
519 io_32bit 0 0 3 rw
520 keepsettings 0 0 1 rw
521 max_kb_per_request 122 1 127 rw
522 multcount 0 0 8 rw
523 nice1 1 0 1 rw
524 nowerr 0 0 1 rw
525 pio_mode write-only 0 255 w
526 slow 0 0 1 rw
527 unmaskirq 0 0 1 rw
528 using_dma 0 0 1 rw
531 1.4 Networking info in /proc/net
532 --------------------------------
534 The subdirectory /proc/net follows the usual pattern. Table 1-6 shows the
535 additional values you get for IP version 6 if you configure the kernel to
536 support this. Table 1-7 lists the files and their meaning.
539 Table 1-6: IPv6 info in /proc/net
540 ..............................................................................
541 File Content
542 udp6 UDP sockets (IPv6)
543 tcp6 TCP sockets (IPv6)
544 raw6 Raw device statistics (IPv6)
545 igmp6 IP multicast addresses, which this host joined (IPv6)
546 if_inet6 List of IPv6 interface addresses
547 ipv6_route Kernel routing table for IPv6
548 rt6_stats Global IPv6 routing tables statistics
549 sockstat6 Socket statistics (IPv6)
550 snmp6 Snmp data (IPv6)
551 ..............................................................................
554 Table 1-7: Network info in /proc/net
555 ..............................................................................
556 File Content
557 arp Kernel ARP table
558 dev network devices with statistics
559 dev_mcast the Layer2 multicast groups a device is listening too
560 (interface index, label, number of references, number of bound
561 addresses).
562 dev_stat network device status
563 ip_fwchains Firewall chain linkage
564 ip_fwnames Firewall chain names
565 ip_masq Directory containing the masquerading tables
566 ip_masquerade Major masquerading table
567 netstat Network statistics
568 raw raw device statistics
569 route Kernel routing table
570 rpc Directory containing rpc info
571 rt_cache Routing cache
572 snmp SNMP data
573 sockstat Socket statistics
574 tcp TCP sockets
575 tr_rif Token ring RIF routing table
576 udp UDP sockets
577 unix UNIX domain sockets
578 wireless Wireless interface data (Wavelan etc)
579 igmp IP multicast addresses, which this host joined
580 psched Global packet scheduler parameters.
581 netlink List of PF_NETLINK sockets
582 ip_mr_vifs List of multicast virtual interfaces
583 ip_mr_cache List of multicast routing cache
584 ..............................................................................
586 You can use this information to see which network devices are available in
587 your system and how much traffic was routed over those devices:
589 > cat /proc/net/dev
590 Inter-|Receive |[...
591 face |bytes packets errs drop fifo frame compressed multicast|[...
592 lo: 908188 5596 0 0 0 0 0 0 [...
593 ppp0:15475140 20721 410 0 0 410 0 0 [...
594 eth0: 614530 7085 0 0 0 0 0 1 [...
596 ...] Transmit
597 ...] bytes packets errs drop fifo colls carrier compressed
598 ...] 908188 5596 0 0 0 0 0 0
599 ...] 1375103 17405 0 0 0 0 0 0
600 ...] 1703981 5535 0 0 0 3 0 0
602 In addition, each Channel Bond interface has it's own directory. For
603 example, the bond0 device will have a directory called /proc/net/bond0/.
604 It will contain information that is specific to that bond, such as the
605 current slaves of the bond, the link status of the slaves, and how
606 many times the slaves link has failed.
608 1.5 SCSI info
609 -------------
611 If you have a SCSI host adapter in your system, you'll find a subdirectory
612 named after the driver for this adapter in /proc/scsi. You'll also see a list
613 of all recognized SCSI devices in /proc/scsi:
615 >cat /proc/scsi/scsi
616 Attached devices:
617 Host: scsi0 Channel: 00 Id: 00 Lun: 00
618 Vendor: IBM Model: DGHS09U Rev: 03E0
619 Type: Direct-Access ANSI SCSI revision: 03
620 Host: scsi0 Channel: 00 Id: 06 Lun: 00
621 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04
622 Type: CD-ROM ANSI SCSI revision: 02
625 The directory named after the driver has one file for each adapter found in
626 the system. These files contain information about the controller, including
627 the used IRQ and the IO address range. The amount of information shown is
628 dependent on the adapter you use. The example shows the output for an Adaptec
629 AHA-2940 SCSI adapter:
631 > cat /proc/scsi/aic7xxx/0
633 Adaptec AIC7xxx driver version: 5.1.19/3.2.4
634 Compile Options:
635 TCQ Enabled By Default : Disabled
636 AIC7XXX_PROC_STATS : Disabled
637 AIC7XXX_RESET_DELAY : 5
638 Adapter Configuration:
639 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter
640 Ultra Wide Controller
641 PCI MMAPed I/O Base: 0xeb001000
642 Adapter SEEPROM Config: SEEPROM found and used.
643 Adaptec SCSI BIOS: Enabled
644 IRQ: 10
645 SCBs: Active 0, Max Active 2,
646 Allocated 15, HW 16, Page 255
647 Interrupts: 160328
648 BIOS Control Word: 0x18b6
649 Adapter Control Word: 0x005b
650 Extended Translation: Enabled
651 Disconnect Enable Flags: 0xffff
652 Ultra Enable Flags: 0x0001
653 Tag Queue Enable Flags: 0x0000
654 Ordered Queue Tag Flags: 0x0000
655 Default Tag Queue Depth: 8
656 Tagged Queue By Device array for aic7xxx host instance 0:
657 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255}
658 Actual queue depth per device for aic7xxx host instance 0:
659 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}
660 Statistics:
661 (scsi0:0:0:0)
662 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8
663 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0)
664 Total transfers 160151 (74577 reads and 85574 writes)
665 (scsi0:0:6:0)
666 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15
667 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0)
668 Total transfers 0 (0 reads and 0 writes)
671 1.6 Parallel port info in /proc/parport
672 ---------------------------------------
674 The directory /proc/parport contains information about the parallel ports of
675 your system. It has one subdirectory for each port, named after the port
676 number (0,1,2,...).
678 These directories contain the four files shown in Table 1-8.
681 Table 1-8: Files in /proc/parport
682 ..............................................................................
683 File Content
684 autoprobe Any IEEE-1284 device ID information that has been acquired.
685 devices list of the device drivers using that port. A + will appear by the
686 name of the device currently using the port (it might not appear
687 against any).
688 hardware Parallel port's base address, IRQ line and DMA channel.
689 irq IRQ that parport is using for that port. This is in a separate
690 file to allow you to alter it by writing a new value in (IRQ
691 number or none).
692 ..............................................................................
694 1.7 TTY info in /proc/tty
695 -------------------------
697 Information about the available and actually used tty's can be found in the
698 directory /proc/tty.You'll find entries for drivers and line disciplines in
699 this directory, as shown in Table 1-9.
702 Table 1-9: Files in /proc/tty
703 ..............................................................................
704 File Content
705 drivers list of drivers and their usage
706 ldiscs registered line disciplines
707 driver/serial usage statistic and status of single tty lines
708 ..............................................................................
710 To see which tty's are currently in use, you can simply look into the file
711 /proc/tty/drivers:
713 > cat /proc/tty/drivers
714 pty_slave /dev/pts 136 0-255 pty:slave
715 pty_master /dev/ptm 128 0-255 pty:master
716 pty_slave /dev/ttyp 3 0-255 pty:slave
717 pty_master /dev/pty 2 0-255 pty:master
718 serial /dev/cua 5 64-67 serial:callout
719 serial /dev/ttyS 4 64-67 serial
720 /dev/tty0 /dev/tty0 4 0 system:vtmaster
721 /dev/ptmx /dev/ptmx 5 2 system
722 /dev/console /dev/console 5 1 system:console
723 /dev/tty /dev/tty 5 0 system:/dev/tty
724 unknown /dev/tty 4 1-63 console
727 1.8 Miscellaneous kernel statistics in /proc/stat
728 -------------------------------------------------
730 Various pieces of information about kernel activity are available in the
731 /proc/stat file. All of the numbers reported in this file are aggregates
732 since the system first booted. For a quick look, simply cat the file:
734 > cat /proc/stat
735 cpu 2255 34 2290 22625563 6290 127 456
736 cpu0 1132 34 1441 11311718 3675 127 438
737 cpu1 1123 0 849 11313845 2614 0 18
738 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...]
739 ctxt 1990473
740 btime 1062191376
741 processes 2915
742 procs_running 1
743 procs_blocked 0
745 The very first "cpu" line aggregates the numbers in all of the other "cpuN"
746 lines. These numbers identify the amount of time the CPU has spent performing
747 different kinds of work. Time units are in USER_HZ (typically hundredths of a
748 second). The meanings of the columns are as follows, from left to right:
750 - user: normal processes executing in user mode
751 - nice: niced processes executing in user mode
752 - system: processes executing in kernel mode
753 - idle: twiddling thumbs
754 - iowait: waiting for I/O to complete
755 - irq: servicing interrupts
756 - softirq: servicing softirqs
758 The "intr" line gives counts of interrupts serviced since boot time, for each
759 of the possible system interrupts. The first column is the total of all
760 interrupts serviced; each subsequent column is the total for that particular
761 interrupt.
763 The "ctxt" line gives the total number of context switches across all CPUs.
765 The "btime" line gives the time at which the system booted, in seconds since
766 the Unix epoch.
768 The "processes" line gives the number of processes and threads created, which
769 includes (but is not limited to) those created by calls to the fork() and
770 clone() system calls.
772 The "procs_running" line gives the number of processes currently running on
773 CPUs.
775 The "procs_blocked" line gives the number of processes currently blocked,
776 waiting for I/O to complete.
779 ------------------------------------------------------------------------------
780 Summary
781 ------------------------------------------------------------------------------
782 The /proc file system serves information about the running system. It not only
783 allows access to process data but also allows you to request the kernel status
784 by reading files in the hierarchy.
786 The directory structure of /proc reflects the types of information and makes
787 it easy, if not obvious, where to look for specific data.
788 ------------------------------------------------------------------------------
790 ------------------------------------------------------------------------------
791 CHAPTER 2: MODIFYING SYSTEM PARAMETERS
792 ------------------------------------------------------------------------------
794 ------------------------------------------------------------------------------
795 In This Chapter
796 ------------------------------------------------------------------------------
797 * Modifying kernel parameters by writing into files found in /proc/sys
798 * Exploring the files which modify certain parameters
799 * Review of the /proc/sys file tree
800 ------------------------------------------------------------------------------
803 A very interesting part of /proc is the directory /proc/sys. This is not only
804 a source of information, it also allows you to change parameters within the
805 kernel. Be very careful when attempting this. You can optimize your system,
806 but you can also cause it to crash. Never alter kernel parameters on a
807 production system. Set up a development machine and test to make sure that
808 everything works the way you want it to. You may have no alternative but to
809 reboot the machine once an error has been made.
811 To change a value, simply echo the new value into the file. An example is
812 given below in the section on the file system data. You need to be root to do
813 this. You can create your own boot script to perform this every time your
814 system boots.
816 The files in /proc/sys can be used to fine tune and monitor miscellaneous and
817 general things in the operation of the Linux kernel. Since some of the files
818 can inadvertently disrupt your system, it is advisable to read both
819 documentation and source before actually making adjustments. In any case, be
820 very careful when writing to any of these files. The entries in /proc may
821 change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt
822 review the kernel documentation in the directory /usr/src/linux/Documentation.
823 This chapter is heavily based on the documentation included in the pre 2.2
824 kernels, and became part of it in version 2.2.1 of the Linux kernel.
826 2.1 /proc/sys/fs - File system data
827 -----------------------------------
829 This subdirectory contains specific file system, file handle, inode, dentry
830 and quota information.
832 Currently, these files are in /proc/sys/fs:
834 dentry-state
835 ------------
837 Status of the directory cache. Since directory entries are dynamically
838 allocated and deallocated, this file indicates the current status. It holds
839 six values, in which the last two are not used and are always zero. The others
840 are listed in table 2-1.
843 Table 2-1: Status files of the directory cache
844 ..............................................................................
845 File Content
846 nr_dentry Almost always zero
847 nr_unused Number of unused cache entries
848 age_limit
849 in seconds after the entry may be reclaimed, when memory is short
850 want_pages internally
851 ..............................................................................
853 dquot-nr and dquot-max
854 ----------------------
856 The file dquot-max shows the maximum number of cached disk quota entries.
858 The file dquot-nr shows the number of allocated disk quota entries and the
859 number of free disk quota entries.
861 If the number of available cached disk quotas is very low and you have a large
862 number of simultaneous system users, you might want to raise the limit.
864 file-nr and file-max
865 --------------------
867 The kernel allocates file handles dynamically, but doesn't free them again at
868 this time.
870 The value in file-max denotes the maximum number of file handles that the
871 Linux kernel will allocate. When you get a lot of error messages about running
872 out of file handles, you might want to raise this limit. The default value is
873 10% of RAM in kilobytes. To change it, just write the new number into the
874 file:
876 # cat /proc/sys/fs/file-max
877 4096
878 # echo 8192 > /proc/sys/fs/file-max
879 # cat /proc/sys/fs/file-max
880 8192
883 This method of revision is useful for all customizable parameters of the
884 kernel - simply echo the new value to the corresponding file.
886 Historically, the three values in file-nr denoted the number of allocated file
887 handles, the number of allocated but unused file handles, and the maximum
888 number of file handles. Linux 2.6 always reports 0 as the number of free file
889 handles -- this is not an error, it just means that the number of allocated
890 file handles exactly matches the number of used file handles.
892 Attempts to allocate more file descriptors than file-max are reported with
893 printk, look for "VFS: file-max limit <number> reached".
895 inode-state and inode-nr
896 ------------------------
898 The file inode-nr contains the first two items from inode-state, so we'll skip
899 to that file...
901 inode-state contains two actual numbers and five dummy values. The numbers
902 are nr_inodes and nr_free_inodes (in order of appearance).
904 nr_inodes
905 ~~~~~~~~~
907 Denotes the number of inodes the system has allocated. This number will
908 grow and shrink dynamically.
910 nr_free_inodes
911 --------------
913 Represents the number of free inodes. Ie. The number of inuse inodes is
914 (nr_inodes - nr_free_inodes).
916 aio-nr and aio-max-nr
917 ---------------------
919 aio-nr is the running total of the number of events specified on the
920 io_setup system call for all currently active aio contexts. If aio-nr
921 reaches aio-max-nr then io_setup will fail with EAGAIN. Note that
922 raising aio-max-nr does not result in the pre-allocation or re-sizing
923 of any kernel data structures.
925 2.2 /proc/sys/fs/binfmt_misc - Miscellaneous binary formats
926 -----------------------------------------------------------
928 Besides these files, there is the subdirectory /proc/sys/fs/binfmt_misc. This
929 handles the kernel support for miscellaneous binary formats.
931 Binfmt_misc provides the ability to register additional binary formats to the
932 Kernel without compiling an additional module/kernel. Therefore, binfmt_misc
933 needs to know magic numbers at the beginning or the filename extension of the
934 binary.
936 It works by maintaining a linked list of structs that contain a description of
937 a binary format, including a magic with size (or the filename extension),
938 offset and mask, and the interpreter name. On request it invokes the given
939 interpreter with the original program as argument, as binfmt_java and
940 binfmt_em86 and binfmt_mz do. Since binfmt_misc does not define any default
941 binary-formats, you have to register an additional binary-format.
943 There are two general files in binfmt_misc and one file per registered format.
944 The two general files are register and status.
946 Registering a new binary format
947 -------------------------------
949 To register a new binary format you have to issue the command
951 echo :name:type:offset:magic:mask:interpreter: > /proc/sys/fs/binfmt_misc/register
955 with appropriate name (the name for the /proc-dir entry), offset (defaults to
956 0, if omitted), magic, mask (which can be omitted, defaults to all 0xff) and
957 last but not least, the interpreter that is to be invoked (for example and
958 testing /bin/echo). Type can be M for usual magic matching or E for filename
959 extension matching (give extension in place of magic).
961 Check or reset the status of the binary format handler
962 ------------------------------------------------------
964 If you do a cat on the file /proc/sys/fs/binfmt_misc/status, you will get the
965 current status (enabled/disabled) of binfmt_misc. Change the status by echoing
966 0 (disables) or 1 (enables) or -1 (caution: this clears all previously
967 registered binary formats) to status. For example echo 0 > status to disable
968 binfmt_misc (temporarily).
970 Status of a single handler
971 --------------------------
973 Each registered handler has an entry in /proc/sys/fs/binfmt_misc. These files
974 perform the same function as status, but their scope is limited to the actual
975 binary format. By cating this file, you also receive all related information
976 about the interpreter/magic of the binfmt.
978 Example usage of binfmt_misc (emulate binfmt_java)
979 --------------------------------------------------
981 cd /proc/sys/fs/binfmt_misc
982 echo ':Java:M::\xca\xfe\xba\xbe::/usr/local/java/bin/javawrapper:' > register
983 echo ':HTML:E::html::/usr/local/java/bin/appletviewer:' > register
984 echo ':Applet:M::<!--applet::/usr/local/java/bin/appletviewer:' > register
985 echo ':DEXE:M::\x0eDEX::/usr/bin/dosexec:' > register
988 These four lines add support for Java executables and Java applets (like
989 binfmt_java, additionally recognizing the .html extension with no need to put
990 <!--applet> to every applet file). You have to install the JDK and the
991 shell-script /usr/local/java/bin/javawrapper too. It works around the
992 brokenness of the Java filename handling. To add a Java binary, just create a
993 link to the class-file somewhere in the path.
995 2.3 /proc/sys/kernel - general kernel parameters
996 ------------------------------------------------
998 This directory reflects general kernel behaviors. As I've said before, the
999 contents depend on your configuration. Here you'll find the most important
1000 files, along with descriptions of what they mean and how to use them.
1002 acct
1003 ----
1005 The file contains three values; highwater, lowwater, and frequency.
1007 It exists only when BSD-style process accounting is enabled. These values
1008 control its behavior. If the free space on the file system where the log lives
1009 goes below lowwater percentage, accounting suspends. If it goes above
1010 highwater percentage, accounting resumes. Frequency determines how often you
1011 check the amount of free space (value is in seconds). Default settings are: 4,
1012 2, and 30. That is, suspend accounting if there is less than 2 percent free;
1013 resume it if we have a value of 3 or more percent; consider information about
1014 the amount of free space valid for 30 seconds
1016 ctrl-alt-del
1017 ------------
1019 When the value in this file is 0, ctrl-alt-del is trapped and sent to the init
1020 program to handle a graceful restart. However, when the value is greater that
1021 zero, Linux's reaction to this key combination will be an immediate reboot,
1022 without syncing its dirty buffers.
1024 [NOTE]
1025 When a program (like dosemu) has the keyboard in raw mode, the
1026 ctrl-alt-del is intercepted by the program before it ever reaches the
1027 kernel tty layer, and it is up to the program to decide what to do with
1028 it.
1030 domainname and hostname
1031 -----------------------
1033 These files can be controlled to set the NIS domainname and hostname of your
1034 box. For the classic darkstar.frop.org a simple:
1036 # echo "darkstar" > /proc/sys/kernel/hostname
1037 # echo "frop.org" > /proc/sys/kernel/domainname
1040 would suffice to set your hostname and NIS domainname.
1042 osrelease, ostype and version
1043 -----------------------------
1045 The names make it pretty obvious what these fields contain:
1047 > cat /proc/sys/kernel/osrelease
1048 2.2.12
1050 > cat /proc/sys/kernel/ostype
1051 Linux
1053 > cat /proc/sys/kernel/version
1054 #4 Fri Oct 1 12:41:14 PDT 1999
1057 The files osrelease and ostype should be clear enough. Version needs a little
1058 more clarification. The #4 means that this is the 4th kernel built from this
1059 source base and the date after it indicates the time the kernel was built. The
1060 only way to tune these values is to rebuild the kernel.
1062 panic
1063 -----
1065 The value in this file represents the number of seconds the kernel waits
1066 before rebooting on a panic. When you use the software watchdog, the
1067 recommended setting is 60. If set to 0, the auto reboot after a kernel panic
1068 is disabled, which is the default setting.
1070 printk
1071 ------
1073 The four values in printk denote
1074 * console_loglevel,
1075 * default_message_loglevel,
1076 * minimum_console_loglevel and
1077 * default_console_loglevel
1078 respectively.
1080 These values influence printk() behavior when printing or logging error
1081 messages, which come from inside the kernel. See syslog(2) for more
1082 information on the different log levels.
1084 console_loglevel
1085 ----------------
1087 Messages with a higher priority than this will be printed to the console.
1089 default_message_level
1090 ---------------------
1092 Messages without an explicit priority will be printed with this priority.
1094 minimum_console_loglevel
1095 ------------------------
1097 Minimum (highest) value to which the console_loglevel can be set.
1099 default_console_loglevel
1100 ------------------------
1102 Default value for console_loglevel.
1104 sg-big-buff
1105 -----------
1107 This file shows the size of the generic SCSI (sg) buffer. At this point, you
1108 can't tune it yet, but you can change it at compile time by editing
1109 include/scsi/sg.h and changing the value of SG_BIG_BUFF.
1111 If you use a scanner with SANE (Scanner Access Now Easy) you might want to set
1112 this to a higher value. Refer to the SANE documentation on this issue.
1114 modprobe
1115 --------
1117 The location where the modprobe binary is located. The kernel uses this
1118 program to load modules on demand.
1120 unknown_nmi_panic
1121 -----------------
1123 The value in this file affects behavior of handling NMI. When the value is
1124 non-zero, unknown NMI is trapped and then panic occurs. At that time, kernel
1125 debugging information is displayed on console.
1127 NMI switch that most IA32 servers have fires unknown NMI up, for example.
1128 If a system hangs up, try pressing the NMI switch.
1130 nmi_watchdog
1131 ------------
1133 Enables/Disables the NMI watchdog on x86 systems. When the value is non-zero
1134 the NMI watchdog is enabled and will continuously test all online cpus to
1135 determine whether or not they are still functioning properly.
1137 Because the NMI watchdog shares registers with oprofile, by disabling the NMI
1138 watchdog, oprofile may have more registers to utilize.
1141 2.4 /proc/sys/vm - The virtual memory subsystem
1142 -----------------------------------------------
1144 The files in this directory can be used to tune the operation of the virtual
1145 memory (VM) subsystem of the Linux kernel.
1147 vfs_cache_pressure
1148 ------------------
1150 Controls the tendency of the kernel to reclaim the memory which is used for
1151 caching of directory and inode objects.
1153 At the default value of vfs_cache_pressure=100 the kernel will attempt to
1154 reclaim dentries and inodes at a "fair" rate with respect to pagecache and
1155 swapcache reclaim. Decreasing vfs_cache_pressure causes the kernel to prefer
1156 to retain dentry and inode caches. Increasing vfs_cache_pressure beyond 100
1157 causes the kernel to prefer to reclaim dentries and inodes.
1159 dirty_background_ratio
1160 ----------------------
1162 Contains, as a percentage of total system memory, the number of pages at which
1163 the pdflush background writeback daemon will start writing out dirty data.
1165 dirty_ratio
1166 -----------------
1168 Contains, as a percentage of total system memory, the number of pages at which
1169 a process which is generating disk writes will itself start writing out dirty
1170 data.
1172 dirty_writeback_centisecs
1173 -------------------------
1175 The pdflush writeback daemons will periodically wake up and write `old' data
1176 out to disk. This tunable expresses the interval between those wakeups, in
1177 100'ths of a second.
1179 Setting this to zero disables periodic writeback altogether.
1181 dirty_expire_centisecs
1182 ----------------------
1184 This tunable is used to define when dirty data is old enough to be eligible
1185 for writeout by the pdflush daemons. It is expressed in 100'ths of a second.
1186 Data which has been dirty in-memory for longer than this interval will be
1187 written out next time a pdflush daemon wakes up.
1189 legacy_va_layout
1190 ----------------
1192 If non-zero, this sysctl disables the new 32-bit mmap mmap layout - the kernel
1193 will use the legacy (2.4) layout for all processes.
1195 lower_zone_protection
1196 ---------------------
1198 For some specialised workloads on highmem machines it is dangerous for
1199 the kernel to allow process memory to be allocated from the "lowmem"
1200 zone. This is because that memory could then be pinned via the mlock()
1201 system call, or by unavailability of swapspace.
1203 And on large highmem machines this lack of reclaimable lowmem memory
1204 can be fatal.
1206 So the Linux page allocator has a mechanism which prevents allocations
1207 which _could_ use highmem from using too much lowmem. This means that
1208 a certain amount of lowmem is defended from the possibility of being
1209 captured into pinned user memory.
1211 (The same argument applies to the old 16 megabyte ISA DMA region. This
1212 mechanism will also defend that region from allocations which could use
1213 highmem or lowmem).
1215 The `lower_zone_protection' tunable determines how aggressive the kernel is
1216 in defending these lower zones. The default value is zero - no
1217 protection at all.
1219 If you have a machine which uses highmem or ISA DMA and your
1220 applications are using mlock(), or if you are running with no swap then
1221 you probably should increase the lower_zone_protection setting.
1223 The units of this tunable are fairly vague. It is approximately equal
1224 to "megabytes," so setting lower_zone_protection=100 will protect around 100
1225 megabytes of the lowmem zone from user allocations. It will also make
1226 those 100 megabytes unavailable for use by applications and by
1227 pagecache, so there is a cost.
1229 The effects of this tunable may be observed by monitoring
1230 /proc/meminfo:LowFree. Write a single huge file and observe the point
1231 at which LowFree ceases to fall.
1233 A reasonable value for lower_zone_protection is 100.
1235 page-cluster
1236 ------------
1238 page-cluster controls the number of pages which are written to swap in
1239 a single attempt. The swap I/O size.
1241 It is a logarithmic value - setting it to zero means "1 page", setting
1242 it to 1 means "2 pages", setting it to 2 means "4 pages", etc.
1244 The default value is three (eight pages at a time). There may be some
1245 small benefits in tuning this to a different value if your workload is
1246 swap-intensive.
1248 overcommit_memory
1249 -----------------
1251 Controls overcommit of system memory, possibly allowing processes
1252 to allocate (but not use) more memory than is actually available.
1255 0 - Heuristic overcommit handling. Obvious overcommits of
1256 address space are refused. Used for a typical system. It
1257 ensures a seriously wild allocation fails while allowing
1258 overcommit to reduce swap usage. root is allowed to
1259 allocate slightly more memory in this mode. This is the
1260 default.
1262 1 - Always overcommit. Appropriate for some scientific
1263 applications.
1265 2 - Don't overcommit. The total address space commit
1266 for the system is not permitted to exceed swap plus a
1267 configurable percentage (default is 50) of physical RAM.
1268 Depending on the percentage you use, in most situations
1269 this means a process will not be killed while attempting
1270 to use already-allocated memory but will receive errors
1271 on memory allocation as appropriate.
1273 overcommit_ratio
1274 ----------------
1276 Percentage of physical memory size to include in overcommit calculations
1277 (see above.)
1279 Memory allocation limit = swapspace + physmem * (overcommit_ratio / 100)
1281 swapspace = total size of all swap areas
1282 physmem = size of physical memory in system
1284 nr_hugepages and hugetlb_shm_group
1285 ----------------------------------
1287 nr_hugepages configures number of hugetlb page reserved for the system.
1289 hugetlb_shm_group contains group id that is allowed to create SysV shared
1290 memory segment using hugetlb page.
1292 laptop_mode
1293 -----------
1295 laptop_mode is a knob that controls "laptop mode". All the things that are
1296 controlled by this knob are discussed in Documentation/laptop-mode.txt.
1298 block_dump
1299 ----------
1301 block_dump enables block I/O debugging when set to a nonzero value. More
1302 information on block I/O debugging is in Documentation/laptop-mode.txt.
1304 swap_token_timeout
1305 ------------------
1307 This file contains valid hold time of swap out protection token. The Linux
1308 VM has token based thrashing control mechanism and uses the token to prevent
1309 unnecessary page faults in thrashing situation. The unit of the value is
1310 second. The value would be useful to tune thrashing behavior.
1312 drop_caches
1313 -----------
1315 Writing to this will cause the kernel to drop clean caches, dentries and
1316 inodes from memory, causing that memory to become free.
1318 To free pagecache:
1319 echo 1 > /proc/sys/vm/drop_caches
1320 To free dentries and inodes:
1321 echo 2 > /proc/sys/vm/drop_caches
1322 To free pagecache, dentries and inodes:
1323 echo 3 > /proc/sys/vm/drop_caches
1325 As this is a non-destructive operation and dirty objects are not freeable, the
1326 user should run `sync' first.
1329 2.5 /proc/sys/dev - Device specific parameters
1330 ----------------------------------------------
1332 Currently there is only support for CDROM drives, and for those, there is only
1333 one read-only file containing information about the CD-ROM drives attached to
1334 the system:
1336 >cat /proc/sys/dev/cdrom/info
1337 CD-ROM information, Id: cdrom.c 2.55 1999/04/25
1339 drive name: sr0 hdb
1340 drive speed: 32 40
1341 drive # of slots: 1 0
1342 Can close tray: 1 1
1343 Can open tray: 1 1
1344 Can lock tray: 1 1
1345 Can change speed: 1 1
1346 Can select disk: 0 1
1347 Can read multisession: 1 1
1348 Can read MCN: 1 1
1349 Reports media changed: 1 1
1350 Can play audio: 1 1
1353 You see two drives, sr0 and hdb, along with a list of their features.
1355 2.6 /proc/sys/sunrpc - Remote procedure calls
1356 ---------------------------------------------
1358 This directory contains four files, which enable or disable debugging for the
1359 RPC functions NFS, NFS-daemon, RPC and NLM. The default values are 0. They can
1360 be set to one to turn debugging on. (The default value is 0 for each)
1362 2.7 /proc/sys/net - Networking stuff
1363 ------------------------------------
1365 The interface to the networking parts of the kernel is located in
1366 /proc/sys/net. Table 2-3 shows all possible subdirectories. You may see only
1367 some of them, depending on your kernel's configuration.
1370 Table 2-3: Subdirectories in /proc/sys/net
1371 ..............................................................................
1372 Directory Content Directory Content
1373 core General parameter appletalk Appletalk protocol
1374 unix Unix domain sockets netrom NET/ROM
1375 802 E802 protocol ax25 AX25
1376 ethernet Ethernet protocol rose X.25 PLP layer
1377 ipv4 IP version 4 x25 X.25 protocol
1378 ipx IPX token-ring IBM token ring
1379 bridge Bridging decnet DEC net
1380 ipv6 IP version 6
1381 ..............................................................................
1383 We will concentrate on IP networking here. Since AX15, X.25, and DEC Net are
1384 only minor players in the Linux world, we'll skip them in this chapter. You'll
1385 find some short info on Appletalk and IPX further on in this chapter. Review
1386 the online documentation and the kernel source to get a detailed view of the
1387 parameters for those protocols. In this section we'll discuss the
1388 subdirectories printed in bold letters in the table above. As default values
1389 are suitable for most needs, there is no need to change these values.
1391 /proc/sys/net/core - Network core options
1392 -----------------------------------------
1394 rmem_default
1395 ------------
1397 The default setting of the socket receive buffer in bytes.
1399 rmem_max
1400 --------
1402 The maximum receive socket buffer size in bytes.
1404 wmem_default
1405 ------------
1407 The default setting (in bytes) of the socket send buffer.
1409 wmem_max
1410 --------
1412 The maximum send socket buffer size in bytes.
1414 message_burst and message_cost
1415 ------------------------------
1417 These parameters are used to limit the warning messages written to the kernel
1418 log from the networking code. They enforce a rate limit to make a
1419 denial-of-service attack impossible. A higher message_cost factor, results in
1420 fewer messages that will be written. Message_burst controls when messages will
1421 be dropped. The default settings limit warning messages to one every five
1422 seconds.
1424 warnings
1425 --------
1427 This controls console messages from the networking stack that can occur because
1428 of problems on the network like duplicate address or bad checksums. Normally,
1429 this should be enabled, but if the problem persists the messages can be
1430 disabled.
1433 netdev_max_backlog
1434 ------------------
1436 Maximum number of packets, queued on the INPUT side, when the interface
1437 receives packets faster than kernel can process them.
1439 optmem_max
1440 ----------
1442 Maximum ancillary buffer size allowed per socket. Ancillary data is a sequence
1443 of struct cmsghdr structures with appended data.
1445 /proc/sys/net/unix - Parameters for Unix domain sockets
1446 -------------------------------------------------------
1448 There are only two files in this subdirectory. They control the delays for
1449 deleting and destroying socket descriptors.
1451 2.8 /proc/sys/net/ipv4 - IPV4 settings
1452 --------------------------------------
1454 IP version 4 is still the most used protocol in Unix networking. It will be
1455 replaced by IP version 6 in the next couple of years, but for the moment it's
1456 the de facto standard for the internet and is used in most networking
1457 environments around the world. Because of the importance of this protocol,
1458 we'll have a deeper look into the subtree controlling the behavior of the IPv4
1459 subsystem of the Linux kernel.
1461 Let's start with the entries in /proc/sys/net/ipv4.
1463 ICMP settings
1464 -------------
1466 icmp_echo_ignore_all and icmp_echo_ignore_broadcasts
1467 ----------------------------------------------------
1469 Turn on (1) or off (0), if the kernel should ignore all ICMP ECHO requests, or
1470 just those to broadcast and multicast addresses.
1472 Please note that if you accept ICMP echo requests with a broadcast/multi\-cast
1473 destination address your network may be used as an exploder for denial of
1474 service packet flooding attacks to other hosts.
1476 icmp_destunreach_rate, icmp_echoreply_rate, icmp_paramprob_rate and icmp_timeexeed_rate
1477 ---------------------------------------------------------------------------------------
1479 Sets limits for sending ICMP packets to specific targets. A value of zero
1480 disables all limiting. Any positive value sets the maximum package rate in
1481 hundredth of a second (on Intel systems).
1483 IP settings
1484 -----------
1486 ip_autoconfig
1487 -------------
1489 This file contains the number one if the host received its IP configuration by
1490 RARP, BOOTP, DHCP or a similar mechanism. Otherwise it is zero.
1492 ip_default_ttl
1493 --------------
1495 TTL (Time To Live) for IPv4 interfaces. This is simply the maximum number of
1496 hops a packet may travel.
1498 ip_dynaddr
1499 ----------
1501 Enable dynamic socket address rewriting on interface address change. This is
1502 useful for dialup interface with changing IP addresses.
1504 ip_forward
1505 ----------
1507 Enable or disable forwarding of IP packages between interfaces. Changing this
1508 value resets all other parameters to their default values. They differ if the
1509 kernel is configured as host or router.
1511 ip_local_port_range
1512 -------------------
1514 Range of ports used by TCP and UDP to choose the local port. Contains two
1515 numbers, the first number is the lowest port, the second number the highest
1516 local port. Default is 1024-4999. Should be changed to 32768-61000 for
1517 high-usage systems.
1519 ip_no_pmtu_disc
1520 ---------------
1522 Global switch to turn path MTU discovery off. It can also be set on a per
1523 socket basis by the applications or on a per route basis.
1525 ip_masq_debug
1526 -------------
1528 Enable/disable debugging of IP masquerading.
1530 IP fragmentation settings
1531 -------------------------
1533 ipfrag_high_trash and ipfrag_low_trash
1534 --------------------------------------
1536 Maximum memory used to reassemble IP fragments. When ipfrag_high_thresh bytes
1537 of memory is allocated for this purpose, the fragment handler will toss
1538 packets until ipfrag_low_thresh is reached.
1540 ipfrag_time
1541 -----------
1543 Time in seconds to keep an IP fragment in memory.
1545 TCP settings
1546 ------------
1548 tcp_ecn
1549 -------
1551 This file controls the use of the ECN bit in the IPv4 headers. This is a new
1552 feature about Explicit Congestion Notification, but some routers and firewalls
1553 block traffic that has this bit set, so it could be necessary to echo 0 to
1554 /proc/sys/net/ipv4/tcp_ecn if you want to talk to these sites. For more info
1555 you could read RFC2481.
1557 tcp_retrans_collapse
1558 --------------------
1560 Bug-to-bug compatibility with some broken printers. On retransmit, try to send
1561 larger packets to work around bugs in certain TCP stacks. Can be turned off by
1562 setting it to zero.
1564 tcp_keepalive_probes
1565 --------------------
1567 Number of keep alive probes TCP sends out, until it decides that the
1568 connection is broken.
1570 tcp_keepalive_time
1571 ------------------
1573 How often TCP sends out keep alive messages, when keep alive is enabled. The
1574 default is 2 hours.
1576 tcp_syn_retries
1577 ---------------
1579 Number of times initial SYNs for a TCP connection attempt will be
1580 retransmitted. Should not be higher than 255. This is only the timeout for
1581 outgoing connections, for incoming connections the number of retransmits is
1582 defined by tcp_retries1.
1584 tcp_sack
1585 --------
1587 Enable select acknowledgments after RFC2018.
1589 tcp_timestamps
1590 --------------
1592 Enable timestamps as defined in RFC1323.
1594 tcp_stdurg
1595 ----------
1597 Enable the strict RFC793 interpretation of the TCP urgent pointer field. The
1598 default is to use the BSD compatible interpretation of the urgent pointer
1599 pointing to the first byte after the urgent data. The RFC793 interpretation is
1600 to have it point to the last byte of urgent data. Enabling this option may
1601 lead to interoperability problems. Disabled by default.
1603 tcp_syncookies
1604 --------------
1606 Only valid when the kernel was compiled with CONFIG_SYNCOOKIES. Send out
1607 syncookies when the syn backlog queue of a socket overflows. This is to ward
1608 off the common 'syn flood attack'. Disabled by default.
1610 Note that the concept of a socket backlog is abandoned. This means the peer
1611 may not receive reliable error messages from an over loaded server with
1612 syncookies enabled.
1614 tcp_window_scaling
1615 ------------------
1617 Enable window scaling as defined in RFC1323.
1619 tcp_fin_timeout
1620 ---------------
1622 The length of time in seconds it takes to receive a final FIN before the
1623 socket is always closed. This is strictly a violation of the TCP
1624 specification, but required to prevent denial-of-service attacks.
1626 tcp_max_ka_probes
1627 -----------------
1629 Indicates how many keep alive probes are sent per slow timer run. Should not
1630 be set too high to prevent bursts.
1632 tcp_max_syn_backlog
1633 -------------------
1635 Length of the per socket backlog queue. Since Linux 2.2 the backlog specified
1636 in listen(2) only specifies the length of the backlog queue of already
1637 established sockets. When more connection requests arrive Linux starts to drop
1638 packets. When syncookies are enabled the packets are still answered and the
1639 maximum queue is effectively ignored.
1641 tcp_retries1
1642 ------------
1644 Defines how often an answer to a TCP connection request is retransmitted
1645 before giving up.
1647 tcp_retries2
1648 ------------
1650 Defines how often a TCP packet is retransmitted before giving up.
1652 Interface specific settings
1653 ---------------------------
1655 In the directory /proc/sys/net/ipv4/conf you'll find one subdirectory for each
1656 interface the system knows about and one directory calls all. Changes in the
1657 all subdirectory affect all interfaces, whereas changes in the other
1658 subdirectories affect only one interface. All directories have the same
1659 entries:
1661 accept_redirects
1662 ----------------
1664 This switch decides if the kernel accepts ICMP redirect messages or not. The
1665 default is 'yes' if the kernel is configured for a regular host and 'no' for a
1666 router configuration.
1668 accept_source_route
1669 -------------------
1671 Should source routed packages be accepted or declined. The default is
1672 dependent on the kernel configuration. It's 'yes' for routers and 'no' for
1673 hosts.
1675 bootp_relay
1676 ~~~~~~~~~~~
1678 Accept packets with source address 0.b.c.d with destinations not to this host
1679 as local ones. It is supposed that a BOOTP relay daemon will catch and forward
1680 such packets.
1682 The default is 0, since this feature is not implemented yet (kernel version
1683 2.2.12).
1685 forwarding
1686 ----------
1688 Enable or disable IP forwarding on this interface.
1690 log_martians
1691 ------------
1693 Log packets with source addresses with no known route to kernel log.
1695 mc_forwarding
1696 -------------
1698 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE and a
1699 multicast routing daemon is required.
1701 proxy_arp
1702 ---------
1704 Does (1) or does not (0) perform proxy ARP.
1706 rp_filter
1707 ---------
1709 Integer value determines if a source validation should be made. 1 means yes, 0
1710 means no. Disabled by default, but local/broadcast address spoofing is always
1711 on.
1713 If you set this to 1 on a router that is the only connection for a network to
1714 the net, it will prevent spoofing attacks against your internal networks
1715 (external addresses can still be spoofed), without the need for additional
1716 firewall rules.
1718 secure_redirects
1719 ----------------
1721 Accept ICMP redirect messages only for gateways, listed in default gateway
1722 list. Enabled by default.
1724 shared_media
1725 ------------
1727 If it is not set the kernel does not assume that different subnets on this
1728 device can communicate directly. Default setting is 'yes'.
1730 send_redirects
1731 --------------
1733 Determines whether to send ICMP redirects to other hosts.
1735 Routing settings
1736 ----------------
1738 The directory /proc/sys/net/ipv4/route contains several file to control
1739 routing issues.
1741 error_burst and error_cost
1742 --------------------------
1744 These parameters are used to limit how many ICMP destination unreachable to
1745 send from the host in question. ICMP destination unreachable messages are
1746 sent when we cannot reach the next hop while trying to transmit a packet.
1747 It will also print some error messages to kernel logs if someone is ignoring
1748 our ICMP redirects. The higher the error_cost factor is, the fewer
1749 destination unreachable and error messages will be let through. Error_burst
1750 controls when destination unreachable messages and error messages will be
1751 dropped. The default settings limit warning messages to five every second.
1753 flush
1754 -----
1756 Writing to this file results in a flush of the routing cache.
1758 gc_elasticity, gc_interval, gc_min_interval_ms, gc_timeout, gc_thresh
1759 ---------------------------------------------------------------------
1761 Values to control the frequency and behavior of the garbage collection
1762 algorithm for the routing cache. gc_min_interval is deprecated and replaced
1763 by gc_min_interval_ms.
1766 max_size
1767 --------
1769 Maximum size of the routing cache. Old entries will be purged once the cache
1770 reached has this size.
1772 max_delay, min_delay
1773 --------------------
1775 Delays for flushing the routing cache.
1777 redirect_load, redirect_number
1778 ------------------------------
1780 Factors which determine if more ICPM redirects should be sent to a specific
1781 host. No redirects will be sent once the load limit or the maximum number of
1782 redirects has been reached.
1784 redirect_silence
1785 ----------------
1787 Timeout for redirects. After this period redirects will be sent again, even if
1788 this has been stopped, because the load or number limit has been reached.
1790 Network Neighbor handling
1791 -------------------------
1793 Settings about how to handle connections with direct neighbors (nodes attached
1794 to the same link) can be found in the directory /proc/sys/net/ipv4/neigh.
1796 As we saw it in the conf directory, there is a default subdirectory which
1797 holds the default values, and one directory for each interface. The contents
1798 of the directories are identical, with the single exception that the default
1799 settings contain additional options to set garbage collection parameters.
1801 In the interface directories you'll find the following entries:
1803 base_reachable_time, base_reachable_time_ms
1804 -------------------------------------------
1806 A base value used for computing the random reachable time value as specified
1807 in RFC2461.
1809 Expression of base_reachable_time, which is deprecated, is in seconds.
1810 Expression of base_reachable_time_ms is in milliseconds.
1812 retrans_time, retrans_time_ms
1813 -----------------------------
1815 The time between retransmitted Neighbor Solicitation messages.
1816 Used for address resolution and to determine if a neighbor is
1817 unreachable.
1819 Expression of retrans_time, which is deprecated, is in 1/100 seconds (for
1820 IPv4) or in jiffies (for IPv6).
1821 Expression of retrans_time_ms is in milliseconds.
1823 unres_qlen
1824 ----------
1826 Maximum queue length for a pending arp request - the number of packets which
1827 are accepted from other layers while the ARP address is still resolved.
1829 anycast_delay
1830 -------------
1832 Maximum for random delay of answers to neighbor solicitation messages in
1833 jiffies (1/100 sec). Not yet implemented (Linux does not have anycast support
1834 yet).
1836 ucast_solicit
1837 -------------
1839 Maximum number of retries for unicast solicitation.
1841 mcast_solicit
1842 -------------
1844 Maximum number of retries for multicast solicitation.
1846 delay_first_probe_time
1847 ----------------------
1849 Delay for the first time probe if the neighbor is reachable. (see
1850 gc_stale_time)
1852 locktime
1853 --------
1855 An ARP/neighbor entry is only replaced with a new one if the old is at least
1856 locktime old. This prevents ARP cache thrashing.
1858 proxy_delay
1859 -----------
1861 Maximum time (real time is random [0..proxytime]) before answering to an ARP
1862 request for which we have an proxy ARP entry. In some cases, this is used to
1863 prevent network flooding.
1865 proxy_qlen
1866 ----------
1868 Maximum queue length of the delayed proxy arp timer. (see proxy_delay).
1870 app_solicit
1871 ----------
1873 Determines the number of requests to send to the user level ARP daemon. Use 0
1874 to turn off.
1876 gc_stale_time
1877 -------------
1879 Determines how often to check for stale ARP entries. After an ARP entry is
1880 stale it will be resolved again (which is useful when an IP address migrates
1881 to another machine). When ucast_solicit is greater than 0 it first tries to
1882 send an ARP packet directly to the known host When that fails and
1883 mcast_solicit is greater than 0, an ARP request is broadcasted.
1885 2.9 Appletalk
1886 -------------
1888 The /proc/sys/net/appletalk directory holds the Appletalk configuration data
1889 when Appletalk is loaded. The configurable parameters are:
1891 aarp-expiry-time
1892 ----------------
1894 The amount of time we keep an ARP entry before expiring it. Used to age out
1895 old hosts.
1897 aarp-resolve-time
1898 -----------------
1900 The amount of time we will spend trying to resolve an Appletalk address.
1902 aarp-retransmit-limit
1903 ---------------------
1905 The number of times we will retransmit a query before giving up.
1907 aarp-tick-time
1908 --------------
1910 Controls the rate at which expires are checked.
1912 The directory /proc/net/appletalk holds the list of active Appletalk sockets
1913 on a machine.
1915 The fields indicate the DDP type, the local address (in network:node format)
1916 the remote address, the size of the transmit pending queue, the size of the
1917 received queue (bytes waiting for applications to read) the state and the uid
1918 owning the socket.
1920 /proc/net/atalk_iface lists all the interfaces configured for appletalk.It
1921 shows the name of the interface, its Appletalk address, the network range on
1922 that address (or network number for phase 1 networks), and the status of the
1923 interface.
1925 /proc/net/atalk_route lists each known network route. It lists the target
1926 (network) that the route leads to, the router (may be directly connected), the
1927 route flags, and the device the route is using.
1929 2.10 IPX
1930 --------
1932 The IPX protocol has no tunable values in proc/sys/net.
1934 The IPX protocol does, however, provide proc/net/ipx. This lists each IPX
1935 socket giving the local and remote addresses in Novell format (that is
1936 network:node:port). In accordance with the strange Novell tradition,
1937 everything but the port is in hex. Not_Connected is displayed for sockets that
1938 are not tied to a specific remote address. The Tx and Rx queue sizes indicate
1939 the number of bytes pending for transmission and reception. The state
1940 indicates the state the socket is in and the uid is the owning uid of the
1941 socket.
1943 The /proc/net/ipx_interface file lists all IPX interfaces. For each interface
1944 it gives the network number, the node number, and indicates if the network is
1945 the primary network. It also indicates which device it is bound to (or
1946 Internal for internal networks) and the Frame Type if appropriate. Linux
1947 supports 802.3, 802.2, 802.2 SNAP and DIX (Blue Book) ethernet framing for
1948 IPX.
1950 The /proc/net/ipx_route table holds a list of IPX routes. For each route it
1951 gives the destination network, the router node (or Directly) and the network
1952 address of the router (or Connected) for internal networks.
1954 2.11 /proc/sys/fs/mqueue - POSIX message queues filesystem
1955 ----------------------------------------------------------
1957 The "mqueue" filesystem provides the necessary kernel features to enable the
1958 creation of a user space library that implements the POSIX message queues
1959 API (as noted by the MSG tag in the POSIX 1003.1-2001 version of the System
1960 Interfaces specification.)
1962 The "mqueue" filesystem contains values for determining/setting the amount of
1963 resources used by the file system.
1965 /proc/sys/fs/mqueue/queues_max is a read/write file for setting/getting the
1966 maximum number of message queues allowed on the system.
1968 /proc/sys/fs/mqueue/msg_max is a read/write file for setting/getting the
1969 maximum number of messages in a queue value. In fact it is the limiting value
1970 for another (user) limit which is set in mq_open invocation. This attribute of
1971 a queue must be less or equal then msg_max.
1973 /proc/sys/fs/mqueue/msgsize_max is a read/write file for setting/getting the
1974 maximum message size value (it is every message queue's attribute set during
1975 its creation).
1977 2.12 /proc/<pid>/oom_adj - Adjust the oom-killer score
1978 ------------------------------------------------------
1980 This file can be used to adjust the score used to select which processes
1981 should be killed in an out-of-memory situation. Giving it a high score will
1982 increase the likelihood of this process being killed by the oom-killer. Valid
1983 values are in the range -16 to +15, plus the special value -17, which disables
1984 oom-killing altogether for this process.
1986 2.13 /proc/<pid>/oom_score - Display current oom-killer score
1987 -------------------------------------------------------------
1989 ------------------------------------------------------------------------------
1990 This file can be used to check the current score used by the oom-killer is for
1991 any given <pid>. Use it together with /proc/<pid>/oom_adj to tune which
1992 process should be killed in an out-of-memory situation.
1994 ------------------------------------------------------------------------------
1995 Summary
1996 ------------------------------------------------------------------------------
1997 Certain aspects of kernel behavior can be modified at runtime, without the
1998 need to recompile the kernel, or even to reboot the system. The files in the
1999 /proc/sys tree can not only be read, but also modified. You can use the echo
2000 command to write value into these files, thereby changing the default settings
2001 of the kernel.
2002 ------------------------------------------------------------------------------
2004 2.14 /proc/<pid>/io - Display the IO accounting fields
2005 -------------------------------------------------------
2007 This file contains IO statistics for each running process
2009 Example
2010 -------
2012 test:/tmp # dd if=/dev/zero of=/tmp/test.dat &
2013 [1] 3828
2015 test:/tmp # cat /proc/3828/io
2016 rchar: 323934931
2017 wchar: 323929600
2018 syscr: 632687
2019 syscw: 632675
2020 read_bytes: 0
2021 write_bytes: 323932160
2022 cancelled_write_bytes: 0
2025 Description
2026 -----------
2028 rchar
2029 -----
2031 I/O counter: chars read
2032 The number of bytes which this task has caused to be read from storage. This
2033 is simply the sum of bytes which this process passed to read() and pread().
2034 It includes things like tty IO and it is unaffected by whether or not actual
2035 physical disk IO was required (the read might have been satisfied from
2036 pagecache)
2039 wchar
2040 -----
2042 I/O counter: chars written
2043 The number of bytes which this task has caused, or shall cause to be written
2044 to disk. Similar caveats apply here as with rchar.
2047 syscr
2048 -----
2050 I/O counter: read syscalls
2051 Attempt to count the number of read I/O operations, i.e. syscalls like read()
2052 and pread().
2055 syscw
2056 -----
2058 I/O counter: write syscalls
2059 Attempt to count the number of write I/O operations, i.e. syscalls like
2060 write() and pwrite().
2063 read_bytes
2064 ----------
2066 I/O counter: bytes read
2067 Attempt to count the number of bytes which this process really did cause to
2068 be fetched from the storage layer. Done at the submit_bio() level, so it is
2069 accurate for block-backed filesystems. <please add status regarding NFS and
2070 CIFS at a later time>
2073 write_bytes
2074 -----------
2076 I/O counter: bytes written
2077 Attempt to count the number of bytes which this process caused to be sent to
2078 the storage layer. This is done at page-dirtying time.
2081 cancelled_write_bytes
2082 ---------------------
2084 The big inaccuracy here is truncate. If a process writes 1MB to a file and
2085 then deletes the file, it will in fact perform no writeout. But it will have
2086 been accounted as having caused 1MB of write.
2087 In other words: The number of bytes which this process caused to not happen,
2088 by truncating pagecache. A task can cause "negative" IO too. If this task
2089 truncates some dirty pagecache, some IO which another task has been accounted
2090 for (in it's write_bytes) will not be happening. We _could_ just subtract that
2091 from the truncating task's write_bytes, but there is information loss in doing
2092 that.
2095 Note
2096 ----
2098 At its current implementation state, this is a bit racy on 32-bit machines: if
2099 process A reads process B's /proc/pid/io while process B is updating one of
2100 those 64-bit counters, process A could see an intermediate result.
2103 More information about this can be found within the taskstats documentation in
2104 Documentation/accounting.
2106 ------------------------------------------------------------------------------