ip-sysctl.rst 89 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071727374757677787980818283848586878889909192939495969798991001011021031041051061071081091101111121131141151161171181191201211221231241251261271281291301311321331341351361371381391401411421431441451461471481491501511521531541551561571581591601611621631641651661671681691701711721731741751761771781791801811821831841851861871881891901911921931941951961971981992002012022032042052062072082092102112122132142152162172182192202212222232242252262272282292302312322332342352362372382392402412422432442452462472482492502512522532542552562572582592602612622632642652662672682692702712722732742752762772782792802812822832842852862872882892902912922932942952962972982993003013023033043053063073083093103113123133143153163173183193203213223233243253263273283293303313323333343353363373383393403413423433443453463473483493503513523533543553563573583593603613623633643653663673683693703713723733743753763773783793803813823833843853863873883893903913923933943953963973983994004014024034044054064074084094104114124134144154164174184194204214224234244254264274284294304314324334344354364374384394404414424434444454464474484494504514524534544554564574584594604614624634644654664674684694704714724734744754764774784794804814824834844854864874884894904914924934944954964974984995005015025035045055065075085095105115125135145155165175185195205215225235245255265275285295305315325335345355365375385395405415425435445455465475485495505515525535545555565575585595605615625635645655665675685695705715725735745755765775785795805815825835845855865875885895905915925935945955965975985996006016026036046056066076086096106116126136146156166176186196206216226236246256266276286296306316326336346356366376386396406416426436446456466476486496506516526536546556566576586596606616626636646656666676686696706716726736746756766776786796806816826836846856866876886896906916926936946956966976986997007017027037047057067077087097107117127137147157167177187197207217227237247257267277287297307317327337347357367377387397407417427437447457467477487497507517527537547557567577587597607617627637647657667677687697707717727737747757767777787797807817827837847857867877887897907917927937947957967977987998008018028038048058068078088098108118128138148158168178188198208218228238248258268278288298308318328338348358368378388398408418428438448458468478488498508518528538548558568578588598608618628638648658668678688698708718728738748758768778788798808818828838848858868878888898908918928938948958968978988999009019029039049059069079089099109119129139149159169179189199209219229239249259269279289299309319329339349359369379389399409419429439449459469479489499509519529539549559569579589599609619629639649659669679689699709719729739749759769779789799809819829839849859869879889899909919929939949959969979989991000100110021003100410051006100710081009101010111012101310141015101610171018101910201021102210231024102510261027102810291030103110321033103410351036103710381039104010411042104310441045104610471048104910501051105210531054105510561057105810591060106110621063106410651066106710681069107010711072107310741075107610771078107910801081108210831084108510861087108810891090109110921093109410951096109710981099110011011102110311041105110611071108110911101111111211131114111511161117111811191120112111221123112411251126112711281129113011311132113311341135113611371138113911401141114211431144114511461147114811491150115111521153115411551156115711581159116011611162116311641165116611671168116911701171117211731174117511761177117811791180118111821183118411851186118711881189119011911192119311941195119611971198119912001201120212031204120512061207120812091210121112121213121412151216121712181219122012211222122312241225122612271228122912301231123212331234123512361237123812391240124112421243124412451246124712481249125012511252125312541255125612571258125912601261126212631264126512661267126812691270127112721273127412751276127712781279128012811282128312841285128612871288128912901291129212931294129512961297129812991300130113021303130413051306130713081309131013111312131313141315131613171318131913201321132213231324132513261327132813291330133113321333133413351336133713381339134013411342134313441345134613471348134913501351135213531354135513561357135813591360136113621363136413651366136713681369137013711372137313741375137613771378137913801381138213831384138513861387138813891390139113921393139413951396139713981399140014011402140314041405140614071408140914101411141214131414141514161417141814191420142114221423142414251426142714281429143014311432143314341435143614371438143914401441144214431444144514461447144814491450145114521453145414551456145714581459146014611462146314641465146614671468146914701471147214731474147514761477147814791480148114821483148414851486148714881489149014911492149314941495149614971498149915001501150215031504150515061507150815091510151115121513151415151516151715181519152015211522152315241525152615271528152915301531153215331534153515361537153815391540154115421543154415451546154715481549155015511552155315541555155615571558155915601561156215631564156515661567156815691570157115721573157415751576157715781579158015811582158315841585158615871588158915901591159215931594159515961597159815991600160116021603160416051606160716081609161016111612161316141615161616171618161916201621162216231624162516261627162816291630163116321633163416351636163716381639164016411642164316441645164616471648164916501651165216531654165516561657165816591660166116621663166416651666166716681669167016711672167316741675167616771678167916801681168216831684168516861687168816891690169116921693169416951696169716981699170017011702170317041705170617071708170917101711171217131714171517161717171817191720172117221723172417251726172717281729173017311732173317341735173617371738173917401741174217431744174517461747174817491750175117521753175417551756175717581759176017611762176317641765176617671768176917701771177217731774177517761777177817791780178117821783178417851786178717881789179017911792179317941795179617971798179918001801180218031804180518061807180818091810181118121813181418151816181718181819182018211822182318241825182618271828182918301831183218331834183518361837183818391840184118421843184418451846184718481849185018511852185318541855185618571858185918601861186218631864186518661867186818691870187118721873187418751876187718781879188018811882188318841885188618871888188918901891189218931894189518961897189818991900190119021903190419051906190719081909191019111912191319141915191619171918191919201921192219231924192519261927192819291930193119321933193419351936193719381939194019411942194319441945194619471948194919501951195219531954195519561957195819591960196119621963196419651966196719681969197019711972197319741975197619771978197919801981198219831984198519861987198819891990199119921993199419951996199719981999200020012002200320042005200620072008200920102011201220132014201520162017201820192020202120222023202420252026202720282029203020312032203320342035203620372038203920402041204220432044204520462047204820492050205120522053205420552056205720582059206020612062206320642065206620672068206920702071207220732074207520762077207820792080208120822083208420852086208720882089209020912092209320942095209620972098209921002101210221032104210521062107210821092110211121122113211421152116211721182119212021212122212321242125212621272128212921302131213221332134213521362137213821392140214121422143214421452146214721482149215021512152215321542155215621572158215921602161216221632164216521662167216821692170217121722173217421752176217721782179218021812182218321842185218621872188218921902191219221932194219521962197219821992200220122022203220422052206220722082209221022112212221322142215221622172218221922202221222222232224222522262227222822292230223122322233223422352236223722382239224022412242224322442245224622472248224922502251225222532254225522562257225822592260226122622263226422652266226722682269227022712272227322742275227622772278227922802281228222832284228522862287228822892290229122922293229422952296229722982299230023012302230323042305230623072308230923102311231223132314231523162317231823192320232123222323232423252326232723282329233023312332233323342335233623372338233923402341234223432344234523462347234823492350235123522353235423552356235723582359236023612362236323642365236623672368236923702371237223732374237523762377237823792380238123822383238423852386238723882389239023912392239323942395239623972398239924002401240224032404240524062407240824092410241124122413241424152416241724182419242024212422242324242425242624272428242924302431243224332434243524362437243824392440244124422443244424452446244724482449245024512452245324542455245624572458245924602461246224632464246524662467246824692470247124722473247424752476247724782479248024812482248324842485248624872488248924902491249224932494249524962497249824992500250125022503250425052506250725082509251025112512251325142515251625172518251925202521252225232524252525262527252825292530253125322533253425352536253725382539254025412542254325442545254625472548254925502551255225532554255525562557255825592560256125622563256425652566256725682569257025712572257325742575257625772578257925802581258225832584258525862587258825892590259125922593259425952596259725982599260026012602260326042605260626072608260926102611261226132614261526162617261826192620262126222623262426252626262726282629263026312632263326342635263626372638263926402641264226432644264526462647264826492650265126522653265426552656265726582659266026612662266326642665266626672668266926702671
  1. .. SPDX-License-Identifier: GPL-2.0
  2. =========
  3. IP Sysctl
  4. =========
  5. /proc/sys/net/ipv4/* Variables
  6. ==============================
  7. ip_forward - BOOLEAN
  8. - 0 - disabled (default)
  9. - not 0 - enabled
  10. Forward Packets between interfaces.
  11. This variable is special, its change resets all configuration
  12. parameters to their default state (RFC1122 for hosts, RFC1812
  13. for routers)
  14. ip_default_ttl - INTEGER
  15. Default value of TTL field (Time To Live) for outgoing (but not
  16. forwarded) IP packets. Should be between 1 and 255 inclusive.
  17. Default: 64 (as recommended by RFC1700)
  18. ip_no_pmtu_disc - INTEGER
  19. Disable Path MTU Discovery. If enabled in mode 1 and a
  20. fragmentation-required ICMP is received, the PMTU to this
  21. destination will be set to min_pmtu (see below). You will need
  22. to raise min_pmtu to the smallest interface MTU on your system
  23. manually if you want to avoid locally generated fragments.
  24. In mode 2 incoming Path MTU Discovery messages will be
  25. discarded. Outgoing frames are handled the same as in mode 1,
  26. implicitly setting IP_PMTUDISC_DONT on every created socket.
  27. Mode 3 is a hardened pmtu discover mode. The kernel will only
  28. accept fragmentation-needed errors if the underlying protocol
  29. can verify them besides a plain socket lookup. Current
  30. protocols for which pmtu events will be honored are TCP, SCTP
  31. and DCCP as they verify e.g. the sequence number or the
  32. association. This mode should not be enabled globally but is
  33. only intended to secure e.g. name servers in namespaces where
  34. TCP path mtu must still work but path MTU information of other
  35. protocols should be discarded. If enabled globally this mode
  36. could break other protocols.
  37. Possible values: 0-3
  38. Default: FALSE
  39. min_pmtu - INTEGER
  40. default 552 - minimum discovered Path MTU
  41. ip_forward_use_pmtu - BOOLEAN
  42. By default we don't trust protocol path MTUs while forwarding
  43. because they could be easily forged and can lead to unwanted
  44. fragmentation by the router.
  45. You only need to enable this if you have user-space software
  46. which tries to discover path mtus by itself and depends on the
  47. kernel honoring this information. This is normally not the
  48. case.
  49. Default: 0 (disabled)
  50. Possible values:
  51. - 0 - disabled
  52. - 1 - enabled
  53. fwmark_reflect - BOOLEAN
  54. Controls the fwmark of kernel-generated IPv4 reply packets that are not
  55. associated with a socket for example, TCP RSTs or ICMP echo replies).
  56. If unset, these packets have a fwmark of zero. If set, they have the
  57. fwmark of the packet they are replying to.
  58. Default: 0
  59. fib_multipath_use_neigh - BOOLEAN
  60. Use status of existing neighbor entry when determining nexthop for
  61. multipath routes. If disabled, neighbor information is not used and
  62. packets could be directed to a failed nexthop. Only valid for kernels
  63. built with CONFIG_IP_ROUTE_MULTIPATH enabled.
  64. Default: 0 (disabled)
  65. Possible values:
  66. - 0 - disabled
  67. - 1 - enabled
  68. fib_multipath_hash_policy - INTEGER
  69. Controls which hash policy to use for multipath routes. Only valid
  70. for kernels built with CONFIG_IP_ROUTE_MULTIPATH enabled.
  71. Default: 0 (Layer 3)
  72. Possible values:
  73. - 0 - Layer 3
  74. - 1 - Layer 4
  75. - 2 - Layer 3 or inner Layer 3 if present
  76. fib_sync_mem - UNSIGNED INTEGER
  77. Amount of dirty memory from fib entries that can be backlogged before
  78. synchronize_rcu is forced.
  79. Default: 512kB Minimum: 64kB Maximum: 64MB
  80. ip_forward_update_priority - INTEGER
  81. Whether to update SKB priority from "TOS" field in IPv4 header after it
  82. is forwarded. The new SKB priority is mapped from TOS field value
  83. according to an rt_tos2priority table (see e.g. man tc-prio).
  84. Default: 1 (Update priority.)
  85. Possible values:
  86. - 0 - Do not update priority.
  87. - 1 - Update priority.
  88. route/max_size - INTEGER
  89. Maximum number of routes allowed in the kernel. Increase
  90. this when using large numbers of interfaces and/or routes.
  91. From linux kernel 3.6 onwards, this is deprecated for ipv4
  92. as route cache is no longer used.
  93. neigh/default/gc_thresh1 - INTEGER
  94. Minimum number of entries to keep. Garbage collector will not
  95. purge entries if there are fewer than this number.
  96. Default: 128
  97. neigh/default/gc_thresh2 - INTEGER
  98. Threshold when garbage collector becomes more aggressive about
  99. purging entries. Entries older than 5 seconds will be cleared
  100. when over this number.
  101. Default: 512
  102. neigh/default/gc_thresh3 - INTEGER
  103. Maximum number of non-PERMANENT neighbor entries allowed. Increase
  104. this when using large numbers of interfaces and when communicating
  105. with large numbers of directly-connected peers.
  106. Default: 1024
  107. neigh/default/unres_qlen_bytes - INTEGER
  108. The maximum number of bytes which may be used by packets
  109. queued for each unresolved address by other network layers.
  110. (added in linux 3.3)
  111. Setting negative value is meaningless and will return error.
  112. Default: SK_WMEM_MAX, (same as net.core.wmem_default).
  113. Exact value depends on architecture and kernel options,
  114. but should be enough to allow queuing 256 packets
  115. of medium size.
  116. neigh/default/unres_qlen - INTEGER
  117. The maximum number of packets which may be queued for each
  118. unresolved address by other network layers.
  119. (deprecated in linux 3.3) : use unres_qlen_bytes instead.
  120. Prior to linux 3.3, the default value is 3 which may cause
  121. unexpected packet loss. The current default value is calculated
  122. according to default value of unres_qlen_bytes and true size of
  123. packet.
  124. Default: 101
  125. mtu_expires - INTEGER
  126. Time, in seconds, that cached PMTU information is kept.
  127. min_adv_mss - INTEGER
  128. The advertised MSS depends on the first hop route MTU, but will
  129. never be lower than this setting.
  130. IP Fragmentation:
  131. ipfrag_high_thresh - LONG INTEGER
  132. Maximum memory used to reassemble IP fragments.
  133. ipfrag_low_thresh - LONG INTEGER
  134. (Obsolete since linux-4.17)
  135. Maximum memory used to reassemble IP fragments before the kernel
  136. begins to remove incomplete fragment queues to free up resources.
  137. The kernel still accepts new fragments for defragmentation.
  138. ipfrag_time - INTEGER
  139. Time in seconds to keep an IP fragment in memory.
  140. ipfrag_max_dist - INTEGER
  141. ipfrag_max_dist is a non-negative integer value which defines the
  142. maximum "disorder" which is allowed among fragments which share a
  143. common IP source address. Note that reordering of packets is
  144. not unusual, but if a large number of fragments arrive from a source
  145. IP address while a particular fragment queue remains incomplete, it
  146. probably indicates that one or more fragments belonging to that queue
  147. have been lost. When ipfrag_max_dist is positive, an additional check
  148. is done on fragments before they are added to a reassembly queue - if
  149. ipfrag_max_dist (or more) fragments have arrived from a particular IP
  150. address between additions to any IP fragment queue using that source
  151. address, it's presumed that one or more fragments in the queue are
  152. lost. The existing fragment queue will be dropped, and a new one
  153. started. An ipfrag_max_dist value of zero disables this check.
  154. Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
  155. result in unnecessarily dropping fragment queues when normal
  156. reordering of packets occurs, which could lead to poor application
  157. performance. Using a very large value, e.g. 50000, increases the
  158. likelihood of incorrectly reassembling IP fragments that originate
  159. from different IP datagrams, which could result in data corruption.
  160. Default: 64
  161. INET peer storage
  162. =================
  163. inet_peer_threshold - INTEGER
  164. The approximate size of the storage. Starting from this threshold
  165. entries will be thrown aggressively. This threshold also determines
  166. entries' time-to-live and time intervals between garbage collection
  167. passes. More entries, less time-to-live, less GC interval.
  168. inet_peer_minttl - INTEGER
  169. Minimum time-to-live of entries. Should be enough to cover fragment
  170. time-to-live on the reassembling side. This minimum time-to-live is
  171. guaranteed if the pool size is less than inet_peer_threshold.
  172. Measured in seconds.
  173. inet_peer_maxttl - INTEGER
  174. Maximum time-to-live of entries. Unused entries will expire after
  175. this period of time if there is no memory pressure on the pool (i.e.
  176. when the number of entries in the pool is very small).
  177. Measured in seconds.
  178. TCP variables
  179. =============
  180. somaxconn - INTEGER
  181. Limit of socket listen() backlog, known in userspace as SOMAXCONN.
  182. Defaults to 4096. (Was 128 before linux-5.4)
  183. See also tcp_max_syn_backlog for additional tuning for TCP sockets.
  184. tcp_abort_on_overflow - BOOLEAN
  185. If listening service is too slow to accept new connections,
  186. reset them. Default state is FALSE. It means that if overflow
  187. occurred due to a burst, connection will recover. Enable this
  188. option _only_ if you are really sure that listening daemon
  189. cannot be tuned to accept connections faster. Enabling this
  190. option can harm clients of your server.
  191. tcp_adv_win_scale - INTEGER
  192. Count buffering overhead as bytes/2^tcp_adv_win_scale
  193. (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
  194. if it is <= 0.
  195. Possible values are [-31, 31], inclusive.
  196. Default: 1
  197. tcp_allowed_congestion_control - STRING
  198. Show/set the congestion control choices available to non-privileged
  199. processes. The list is a subset of those listed in
  200. tcp_available_congestion_control.
  201. Default is "reno" and the default setting (tcp_congestion_control).
  202. tcp_app_win - INTEGER
  203. Reserve max(window/2^tcp_app_win, mss) of window for application
  204. buffer. Value 0 is special, it means that nothing is reserved.
  205. Default: 31
  206. tcp_autocorking - BOOLEAN
  207. Enable TCP auto corking :
  208. When applications do consecutive small write()/sendmsg() system calls,
  209. we try to coalesce these small writes as much as possible, to lower
  210. total amount of sent packets. This is done if at least one prior
  211. packet for the flow is waiting in Qdisc queues or device transmit
  212. queue. Applications can still use TCP_CORK for optimal behavior
  213. when they know how/when to uncork their sockets.
  214. Default : 1
  215. tcp_available_congestion_control - STRING
  216. Shows the available congestion control choices that are registered.
  217. More congestion control algorithms may be available as modules,
  218. but not loaded.
  219. tcp_base_mss - INTEGER
  220. The initial value of search_low to be used by the packetization layer
  221. Path MTU discovery (MTU probing). If MTU probing is enabled,
  222. this is the initial MSS used by the connection.
  223. tcp_mtu_probe_floor - INTEGER
  224. If MTU probing is enabled this caps the minimum MSS used for search_low
  225. for the connection.
  226. Default : 48
  227. tcp_min_snd_mss - INTEGER
  228. TCP SYN and SYNACK messages usually advertise an ADVMSS option,
  229. as described in RFC 1122 and RFC 6691.
  230. If this ADVMSS option is smaller than tcp_min_snd_mss,
  231. it is silently capped to tcp_min_snd_mss.
  232. Default : 48 (at least 8 bytes of payload per segment)
  233. tcp_congestion_control - STRING
  234. Set the congestion control algorithm to be used for new
  235. connections. The algorithm "reno" is always available, but
  236. additional choices may be available based on kernel configuration.
  237. Default is set as part of kernel configuration.
  238. For passive connections, the listener congestion control choice
  239. is inherited.
  240. [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
  241. tcp_dsack - BOOLEAN
  242. Allows TCP to send "duplicate" SACKs.
  243. tcp_early_retrans - INTEGER
  244. Tail loss probe (TLP) converts RTOs occurring due to tail
  245. losses into fast recovery (draft-ietf-tcpm-rack). Note that
  246. TLP requires RACK to function properly (see tcp_recovery below)
  247. Possible values:
  248. - 0 disables TLP
  249. - 3 or 4 enables TLP
  250. Default: 3
  251. tcp_ecn - INTEGER
  252. Control use of Explicit Congestion Notification (ECN) by TCP.
  253. ECN is used only when both ends of the TCP connection indicate
  254. support for it. This feature is useful in avoiding losses due
  255. to congestion by allowing supporting routers to signal
  256. congestion before having to drop packets.
  257. Possible values are:
  258. = =====================================================
  259. 0 Disable ECN. Neither initiate nor accept ECN.
  260. 1 Enable ECN when requested by incoming connections and
  261. also request ECN on outgoing connection attempts.
  262. 2 Enable ECN when requested by incoming connections
  263. but do not request ECN on outgoing connections.
  264. = =====================================================
  265. Default: 2
  266. tcp_ecn_fallback - BOOLEAN
  267. If the kernel detects that ECN connection misbehaves, enable fall
  268. back to non-ECN. Currently, this knob implements the fallback
  269. from RFC3168, section 6.1.1.1., but we reserve that in future,
  270. additional detection mechanisms could be implemented under this
  271. knob. The value is not used, if tcp_ecn or per route (or congestion
  272. control) ECN settings are disabled.
  273. Default: 1 (fallback enabled)
  274. tcp_fack - BOOLEAN
  275. This is a legacy option, it has no effect anymore.
  276. tcp_fin_timeout - INTEGER
  277. The length of time an orphaned (no longer referenced by any
  278. application) connection will remain in the FIN_WAIT_2 state
  279. before it is aborted at the local end. While a perfectly
  280. valid "receive only" state for an un-orphaned connection, an
  281. orphaned connection in FIN_WAIT_2 state could otherwise wait
  282. forever for the remote to close its end of the connection.
  283. Cf. tcp_max_orphans
  284. Default: 60 seconds
  285. tcp_frto - INTEGER
  286. Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
  287. F-RTO is an enhanced recovery algorithm for TCP retransmission
  288. timeouts. It is particularly beneficial in networks where the
  289. RTT fluctuates (e.g., wireless). F-RTO is sender-side only
  290. modification. It does not require any support from the peer.
  291. By default it's enabled with a non-zero value. 0 disables F-RTO.
  292. tcp_fwmark_accept - BOOLEAN
  293. If set, incoming connections to listening sockets that do not have a
  294. socket mark will set the mark of the accepting socket to the fwmark of
  295. the incoming SYN packet. This will cause all packets on that connection
  296. (starting from the first SYNACK) to be sent with that fwmark. The
  297. listening socket's mark is unchanged. Listening sockets that already
  298. have a fwmark set via setsockopt(SOL_SOCKET, SO_MARK, ...) are
  299. unaffected.
  300. Default: 0
  301. tcp_invalid_ratelimit - INTEGER
  302. Limit the maximal rate for sending duplicate acknowledgments
  303. in response to incoming TCP packets that are for an existing
  304. connection but that are invalid due to any of these reasons:
  305. (a) out-of-window sequence number,
  306. (b) out-of-window acknowledgment number, or
  307. (c) PAWS (Protection Against Wrapped Sequence numbers) check failure
  308. This can help mitigate simple "ack loop" DoS attacks, wherein
  309. a buggy or malicious middlebox or man-in-the-middle can
  310. rewrite TCP header fields in manner that causes each endpoint
  311. to think that the other is sending invalid TCP segments, thus
  312. causing each side to send an unterminating stream of duplicate
  313. acknowledgments for invalid segments.
  314. Using 0 disables rate-limiting of dupacks in response to
  315. invalid segments; otherwise this value specifies the minimal
  316. space between sending such dupacks, in milliseconds.
  317. Default: 500 (milliseconds).
  318. tcp_keepalive_time - INTEGER
  319. How often TCP sends out keepalive messages when keepalive is enabled.
  320. Default: 2hours.
  321. tcp_keepalive_probes - INTEGER
  322. How many keepalive probes TCP sends out, until it decides that the
  323. connection is broken. Default value: 9.
  324. tcp_keepalive_intvl - INTEGER
  325. How frequently the probes are send out. Multiplied by
  326. tcp_keepalive_probes it is time to kill not responding connection,
  327. after probes started. Default value: 75sec i.e. connection
  328. will be aborted after ~11 minutes of retries.
  329. tcp_l3mdev_accept - BOOLEAN
  330. Enables child sockets to inherit the L3 master device index.
  331. Enabling this option allows a "global" listen socket to work
  332. across L3 master domains (e.g., VRFs) with connected sockets
  333. derived from the listen socket to be bound to the L3 domain in
  334. which the packets originated. Only valid when the kernel was
  335. compiled with CONFIG_NET_L3_MASTER_DEV.
  336. Default: 0 (disabled)
  337. tcp_low_latency - BOOLEAN
  338. This is a legacy option, it has no effect anymore.
  339. tcp_max_orphans - INTEGER
  340. Maximal number of TCP sockets not attached to any user file handle,
  341. held by system. If this number is exceeded orphaned connections are
  342. reset immediately and warning is printed. This limit exists
  343. only to prevent simple DoS attacks, you _must_ not rely on this
  344. or lower the limit artificially, but rather increase it
  345. (probably, after increasing installed memory),
  346. if network conditions require more than default value,
  347. and tune network services to linger and kill such states
  348. more aggressively. Let me to remind again: each orphan eats
  349. up to ~64K of unswappable memory.
  350. tcp_max_syn_backlog - INTEGER
  351. Maximal number of remembered connection requests (SYN_RECV),
  352. which have not received an acknowledgment from connecting client.
  353. This is a per-listener limit.
  354. The minimal value is 128 for low memory machines, and it will
  355. increase in proportion to the memory of machine.
  356. If server suffers from overload, try increasing this number.
  357. Remember to also check /proc/sys/net/core/somaxconn
  358. A SYN_RECV request socket consumes about 304 bytes of memory.
  359. tcp_max_tw_buckets - INTEGER
  360. Maximal number of timewait sockets held by system simultaneously.
  361. If this number is exceeded time-wait socket is immediately destroyed
  362. and warning is printed. This limit exists only to prevent
  363. simple DoS attacks, you _must_ not lower the limit artificially,
  364. but rather increase it (probably, after increasing installed memory),
  365. if network conditions require more than default value.
  366. tcp_mem - vector of 3 INTEGERs: min, pressure, max
  367. min: below this number of pages TCP is not bothered about its
  368. memory appetite.
  369. pressure: when amount of memory allocated by TCP exceeds this number
  370. of pages, TCP moderates its memory consumption and enters memory
  371. pressure mode, which is exited when memory consumption falls
  372. under "min".
  373. max: number of pages allowed for queueing by all TCP sockets.
  374. Defaults are calculated at boot time from amount of available
  375. memory.
  376. tcp_min_rtt_wlen - INTEGER
  377. The window length of the windowed min filter to track the minimum RTT.
  378. A shorter window lets a flow more quickly pick up new (higher)
  379. minimum RTT when it is moved to a longer path (e.g., due to traffic
  380. engineering). A longer window makes the filter more resistant to RTT
  381. inflations such as transient congestion. The unit is seconds.
  382. Possible values: 0 - 86400 (1 day)
  383. Default: 300
  384. tcp_moderate_rcvbuf - BOOLEAN
  385. If set, TCP performs receive buffer auto-tuning, attempting to
  386. automatically size the buffer (no greater than tcp_rmem[2]) to
  387. match the size required by the path for full throughput. Enabled by
  388. default.
  389. tcp_mtu_probing - INTEGER
  390. Controls TCP Packetization-Layer Path MTU Discovery. Takes three
  391. values:
  392. - 0 - Disabled
  393. - 1 - Disabled by default, enabled when an ICMP black hole detected
  394. - 2 - Always enabled, use initial MSS of tcp_base_mss.
  395. tcp_probe_interval - UNSIGNED INTEGER
  396. Controls how often to start TCP Packetization-Layer Path MTU
  397. Discovery reprobe. The default is reprobing every 10 minutes as
  398. per RFC4821.
  399. tcp_probe_threshold - INTEGER
  400. Controls when TCP Packetization-Layer Path MTU Discovery probing
  401. will stop in respect to the width of search range in bytes. Default
  402. is 8 bytes.
  403. tcp_no_metrics_save - BOOLEAN
  404. By default, TCP saves various connection metrics in the route cache
  405. when the connection closes, so that connections established in the
  406. near future can use these to set initial conditions. Usually, this
  407. increases overall performance, but may sometimes cause performance
  408. degradation. If set, TCP will not cache metrics on closing
  409. connections.
  410. tcp_no_ssthresh_metrics_save - BOOLEAN
  411. Controls whether TCP saves ssthresh metrics in the route cache.
  412. Default is 1, which disables ssthresh metrics.
  413. tcp_orphan_retries - INTEGER
  414. This value influences the timeout of a locally closed TCP connection,
  415. when RTO retransmissions remain unacknowledged.
  416. See tcp_retries2 for more details.
  417. The default value is 8.
  418. If your machine is a loaded WEB server,
  419. you should think about lowering this value, such sockets
  420. may consume significant resources. Cf. tcp_max_orphans.
  421. tcp_recovery - INTEGER
  422. This value is a bitmap to enable various experimental loss recovery
  423. features.
  424. ========= =============================================================
  425. RACK: 0x1 enables the RACK loss detection for fast detection of lost
  426. retransmissions and tail drops. It also subsumes and disables
  427. RFC6675 recovery for SACK connections.
  428. RACK: 0x2 makes RACK's reordering window static (min_rtt/4).
  429. RACK: 0x4 disables RACK's DUPACK threshold heuristic
  430. ========= =============================================================
  431. Default: 0x1
  432. tcp_reordering - INTEGER
  433. Initial reordering level of packets in a TCP stream.
  434. TCP stack can then dynamically adjust flow reordering level
  435. between this initial value and tcp_max_reordering
  436. Default: 3
  437. tcp_max_reordering - INTEGER
  438. Maximal reordering level of packets in a TCP stream.
  439. 300 is a fairly conservative value, but you might increase it
  440. if paths are using per packet load balancing (like bonding rr mode)
  441. Default: 300
  442. tcp_retrans_collapse - BOOLEAN
  443. Bug-to-bug compatibility with some broken printers.
  444. On retransmit try to send bigger packets to work around bugs in
  445. certain TCP stacks.
  446. tcp_retries1 - INTEGER
  447. This value influences the time, after which TCP decides, that
  448. something is wrong due to unacknowledged RTO retransmissions,
  449. and reports this suspicion to the network layer.
  450. See tcp_retries2 for more details.
  451. RFC 1122 recommends at least 3 retransmissions, which is the
  452. default.
  453. tcp_retries2 - INTEGER
  454. This value influences the timeout of an alive TCP connection,
  455. when RTO retransmissions remain unacknowledged.
  456. Given a value of N, a hypothetical TCP connection following
  457. exponential backoff with an initial RTO of TCP_RTO_MIN would
  458. retransmit N times before killing the connection at the (N+1)th RTO.
  459. The default value of 15 yields a hypothetical timeout of 924.6
  460. seconds and is a lower bound for the effective timeout.
  461. TCP will effectively time out at the first RTO which exceeds the
  462. hypothetical timeout.
  463. RFC 1122 recommends at least 100 seconds for the timeout,
  464. which corresponds to a value of at least 8.
  465. tcp_rfc1337 - BOOLEAN
  466. If set, the TCP stack behaves conforming to RFC1337. If unset,
  467. we are not conforming to RFC, but prevent TCP TIME_WAIT
  468. assassination.
  469. Default: 0
  470. tcp_rmem - vector of 3 INTEGERs: min, default, max
  471. min: Minimal size of receive buffer used by TCP sockets.
  472. It is guaranteed to each TCP socket, even under moderate memory
  473. pressure.
  474. Default: 4K
  475. default: initial size of receive buffer used by TCP sockets.
  476. This value overrides net.core.rmem_default used by other protocols.
  477. Default: 131072 bytes.
  478. This value results in initial window of 65535.
  479. max: maximal size of receive buffer allowed for automatically
  480. selected receiver buffers for TCP socket. This value does not override
  481. net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
  482. automatic tuning of that socket's receive buffer size, in which
  483. case this value is ignored.
  484. Default: between 131072 and 6MB, depending on RAM size.
  485. tcp_sack - BOOLEAN
  486. Enable select acknowledgments (SACKS).
  487. tcp_comp_sack_delay_ns - LONG INTEGER
  488. TCP tries to reduce number of SACK sent, using a timer
  489. based on 5% of SRTT, capped by this sysctl, in nano seconds.
  490. The default is 1ms, based on TSO autosizing period.
  491. Default : 1,000,000 ns (1 ms)
  492. tcp_comp_sack_slack_ns - LONG INTEGER
  493. This sysctl control the slack used when arming the
  494. timer used by SACK compression. This gives extra time
  495. for small RTT flows, and reduces system overhead by allowing
  496. opportunistic reduction of timer interrupts.
  497. Default : 100,000 ns (100 us)
  498. tcp_comp_sack_nr - INTEGER
  499. Max number of SACK that can be compressed.
  500. Using 0 disables SACK compression.
  501. Default : 44
  502. tcp_slow_start_after_idle - BOOLEAN
  503. If set, provide RFC2861 behavior and time out the congestion
  504. window after an idle period. An idle period is defined at
  505. the current RTO. If unset, the congestion window will not
  506. be timed out after an idle period.
  507. Default: 1
  508. tcp_stdurg - BOOLEAN
  509. Use the Host requirements interpretation of the TCP urgent pointer field.
  510. Most hosts use the older BSD interpretation, so if you turn this on
  511. Linux might not communicate correctly with them.
  512. Default: FALSE
  513. tcp_synack_retries - INTEGER
  514. Number of times SYNACKs for a passive TCP connection attempt will
  515. be retransmitted. Should not be higher than 255. Default value
  516. is 5, which corresponds to 31seconds till the last retransmission
  517. with the current initial RTO of 1second. With this the final timeout
  518. for a passive TCP connection will happen after 63seconds.
  519. tcp_syncookies - INTEGER
  520. Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
  521. Send out syncookies when the syn backlog queue of a socket
  522. overflows. This is to prevent against the common 'SYN flood attack'
  523. Default: 1
  524. Note, that syncookies is fallback facility.
  525. It MUST NOT be used to help highly loaded servers to stand
  526. against legal connection rate. If you see SYN flood warnings
  527. in your logs, but investigation shows that they occur
  528. because of overload with legal connections, you should tune
  529. another parameters until this warning disappear.
  530. See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
  531. syncookies seriously violate TCP protocol, do not allow
  532. to use TCP extensions, can result in serious degradation
  533. of some services (f.e. SMTP relaying), visible not by you,
  534. but your clients and relays, contacting you. While you see
  535. SYN flood warnings in logs not being really flooded, your server
  536. is seriously misconfigured.
  537. If you want to test which effects syncookies have to your
  538. network connections you can set this knob to 2 to enable
  539. unconditionally generation of syncookies.
  540. tcp_fastopen - INTEGER
  541. Enable TCP Fast Open (RFC7413) to send and accept data in the opening
  542. SYN packet.
  543. The client support is enabled by flag 0x1 (on by default). The client
  544. then must use sendmsg() or sendto() with the MSG_FASTOPEN flag,
  545. rather than connect() to send data in SYN.
  546. The server support is enabled by flag 0x2 (off by default). Then
  547. either enable for all listeners with another flag (0x400) or
  548. enable individual listeners via TCP_FASTOPEN socket option with
  549. the option value being the length of the syn-data backlog.
  550. The values (bitmap) are
  551. ===== ======== ======================================================
  552. 0x1 (client) enables sending data in the opening SYN on the client.
  553. 0x2 (server) enables the server support, i.e., allowing data in
  554. a SYN packet to be accepted and passed to the
  555. application before 3-way handshake finishes.
  556. 0x4 (client) send data in the opening SYN regardless of cookie
  557. availability and without a cookie option.
  558. 0x200 (server) accept data-in-SYN w/o any cookie option present.
  559. 0x400 (server) enable all listeners to support Fast Open by
  560. default without explicit TCP_FASTOPEN socket option.
  561. ===== ======== ======================================================
  562. Default: 0x1
  563. Note that additional client or server features are only
  564. effective if the basic support (0x1 and 0x2) are enabled respectively.
  565. tcp_fastopen_blackhole_timeout_sec - INTEGER
  566. Initial time period in second to disable Fastopen on active TCP sockets
  567. when a TFO firewall blackhole issue happens.
  568. This time period will grow exponentially when more blackhole issues
  569. get detected right after Fastopen is re-enabled and will reset to
  570. initial value when the blackhole issue goes away.
  571. 0 to disable the blackhole detection.
  572. By default, it is set to 0 (feature is disabled).
  573. tcp_fastopen_key - list of comma separated 32-digit hexadecimal INTEGERs
  574. The list consists of a primary key and an optional backup key. The
  575. primary key is used for both creating and validating cookies, while the
  576. optional backup key is only used for validating cookies. The purpose of
  577. the backup key is to maximize TFO validation when keys are rotated.
  578. A randomly chosen primary key may be configured by the kernel if
  579. the tcp_fastopen sysctl is set to 0x400 (see above), or if the
  580. TCP_FASTOPEN setsockopt() optname is set and a key has not been
  581. previously configured via sysctl. If keys are configured via
  582. setsockopt() by using the TCP_FASTOPEN_KEY optname, then those
  583. per-socket keys will be used instead of any keys that are specified via
  584. sysctl.
  585. A key is specified as 4 8-digit hexadecimal integers which are separated
  586. by a '-' as: xxxxxxxx-xxxxxxxx-xxxxxxxx-xxxxxxxx. Leading zeros may be
  587. omitted. A primary and a backup key may be specified by separating them
  588. by a comma. If only one key is specified, it becomes the primary key and
  589. any previously configured backup keys are removed.
  590. tcp_syn_retries - INTEGER
  591. Number of times initial SYNs for an active TCP connection attempt
  592. will be retransmitted. Should not be higher than 127. Default value
  593. is 6, which corresponds to 63seconds till the last retransmission
  594. with the current initial RTO of 1second. With this the final timeout
  595. for an active TCP connection attempt will happen after 127seconds.
  596. tcp_timestamps - INTEGER
  597. Enable timestamps as defined in RFC1323.
  598. - 0: Disabled.
  599. - 1: Enable timestamps as defined in RFC1323 and use random offset for
  600. each connection rather than only using the current time.
  601. - 2: Like 1, but without random offsets.
  602. Default: 1
  603. tcp_min_tso_segs - INTEGER
  604. Minimal number of segments per TSO frame.
  605. Since linux-3.12, TCP does an automatic sizing of TSO frames,
  606. depending on flow rate, instead of filling 64Kbytes packets.
  607. For specific usages, it's possible to force TCP to build big
  608. TSO frames. Note that TCP stack might split too big TSO packets
  609. if available window is too small.
  610. Default: 2
  611. tcp_pacing_ss_ratio - INTEGER
  612. sk->sk_pacing_rate is set by TCP stack using a ratio applied
  613. to current rate. (current_rate = cwnd * mss / srtt)
  614. If TCP is in slow start, tcp_pacing_ss_ratio is applied
  615. to let TCP probe for bigger speeds, assuming cwnd can be
  616. doubled every other RTT.
  617. Default: 200
  618. tcp_pacing_ca_ratio - INTEGER
  619. sk->sk_pacing_rate is set by TCP stack using a ratio applied
  620. to current rate. (current_rate = cwnd * mss / srtt)
  621. If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio
  622. is applied to conservatively probe for bigger throughput.
  623. Default: 120
  624. tcp_tso_win_divisor - INTEGER
  625. This allows control over what percentage of the congestion window
  626. can be consumed by a single TSO frame.
  627. The setting of this parameter is a choice between burstiness and
  628. building larger TSO frames.
  629. Default: 3
  630. tcp_tw_reuse - INTEGER
  631. Enable reuse of TIME-WAIT sockets for new connections when it is
  632. safe from protocol viewpoint.
  633. - 0 - disable
  634. - 1 - global enable
  635. - 2 - enable for loopback traffic only
  636. It should not be changed without advice/request of technical
  637. experts.
  638. Default: 2
  639. tcp_window_scaling - BOOLEAN
  640. Enable window scaling as defined in RFC1323.
  641. tcp_wmem - vector of 3 INTEGERs: min, default, max
  642. min: Amount of memory reserved for send buffers for TCP sockets.
  643. Each TCP socket has rights to use it due to fact of its birth.
  644. Default: 4K
  645. default: initial size of send buffer used by TCP sockets. This
  646. value overrides net.core.wmem_default used by other protocols.
  647. It is usually lower than net.core.wmem_default.
  648. Default: 16K
  649. max: Maximal amount of memory allowed for automatically tuned
  650. send buffers for TCP sockets. This value does not override
  651. net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
  652. automatic tuning of that socket's send buffer size, in which case
  653. this value is ignored.
  654. Default: between 64K and 4MB, depending on RAM size.
  655. tcp_notsent_lowat - UNSIGNED INTEGER
  656. A TCP socket can control the amount of unsent bytes in its write queue,
  657. thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
  658. reports POLLOUT events if the amount of unsent bytes is below a per
  659. socket value, and if the write queue is not full. sendmsg() will
  660. also not add new buffers if the limit is hit.
  661. This global variable controls the amount of unsent data for
  662. sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
  663. to the global variable has immediate effect.
  664. Default: UINT_MAX (0xFFFFFFFF)
  665. tcp_workaround_signed_windows - BOOLEAN
  666. If set, assume no receipt of a window scaling option means the
  667. remote TCP is broken and treats the window as a signed quantity.
  668. If unset, assume the remote TCP is not broken even if we do
  669. not receive a window scaling option from them.
  670. Default: 0
  671. tcp_thin_linear_timeouts - BOOLEAN
  672. Enable dynamic triggering of linear timeouts for thin streams.
  673. If set, a check is performed upon retransmission by timeout to
  674. determine if the stream is thin (less than 4 packets in flight).
  675. As long as the stream is found to be thin, up to 6 linear
  676. timeouts may be performed before exponential backoff mode is
  677. initiated. This improves retransmission latency for
  678. non-aggressive thin streams, often found to be time-dependent.
  679. For more information on thin streams, see
  680. Documentation/networking/tcp-thin.rst
  681. Default: 0
  682. tcp_limit_output_bytes - INTEGER
  683. Controls TCP Small Queue limit per tcp socket.
  684. TCP bulk sender tends to increase packets in flight until it
  685. gets losses notifications. With SNDBUF autotuning, this can
  686. result in a large amount of packets queued on the local machine
  687. (e.g.: qdiscs, CPU backlog, or device) hurting latency of other
  688. flows, for typical pfifo_fast qdiscs. tcp_limit_output_bytes
  689. limits the number of bytes on qdisc or device to reduce artificial
  690. RTT/cwnd and reduce bufferbloat.
  691. Default: 1048576 (16 * 65536)
  692. tcp_challenge_ack_limit - INTEGER
  693. Limits number of Challenge ACK sent per second, as recommended
  694. in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
  695. Default: 1000
  696. tcp_rx_skb_cache - BOOLEAN
  697. Controls a per TCP socket cache of one skb, that might help
  698. performance of some workloads. This might be dangerous
  699. on systems with a lot of TCP sockets, since it increases
  700. memory usage.
  701. Default: 0 (disabled)
  702. UDP variables
  703. =============
  704. udp_l3mdev_accept - BOOLEAN
  705. Enabling this option allows a "global" bound socket to work
  706. across L3 master domains (e.g., VRFs) with packets capable of
  707. being received regardless of the L3 domain in which they
  708. originated. Only valid when the kernel was compiled with
  709. CONFIG_NET_L3_MASTER_DEV.
  710. Default: 0 (disabled)
  711. udp_mem - vector of 3 INTEGERs: min, pressure, max
  712. Number of pages allowed for queueing by all UDP sockets.
  713. min: Below this number of pages UDP is not bothered about its
  714. memory appetite. When amount of memory allocated by UDP exceeds
  715. this number, UDP starts to moderate memory usage.
  716. pressure: This value was introduced to follow format of tcp_mem.
  717. max: Number of pages allowed for queueing by all UDP sockets.
  718. Default is calculated at boot time from amount of available memory.
  719. udp_rmem_min - INTEGER
  720. Minimal size of receive buffer used by UDP sockets in moderation.
  721. Each UDP socket is able to use the size for receiving data, even if
  722. total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
  723. Default: 4K
  724. udp_wmem_min - INTEGER
  725. Minimal size of send buffer used by UDP sockets in moderation.
  726. Each UDP socket is able to use the size for sending data, even if
  727. total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
  728. Default: 4K
  729. RAW variables
  730. =============
  731. raw_l3mdev_accept - BOOLEAN
  732. Enabling this option allows a "global" bound socket to work
  733. across L3 master domains (e.g., VRFs) with packets capable of
  734. being received regardless of the L3 domain in which they
  735. originated. Only valid when the kernel was compiled with
  736. CONFIG_NET_L3_MASTER_DEV.
  737. Default: 1 (enabled)
  738. CIPSOv4 Variables
  739. =================
  740. cipso_cache_enable - BOOLEAN
  741. If set, enable additions to and lookups from the CIPSO label mapping
  742. cache. If unset, additions are ignored and lookups always result in a
  743. miss. However, regardless of the setting the cache is still
  744. invalidated when required when means you can safely toggle this on and
  745. off and the cache will always be "safe".
  746. Default: 1
  747. cipso_cache_bucket_size - INTEGER
  748. The CIPSO label cache consists of a fixed size hash table with each
  749. hash bucket containing a number of cache entries. This variable limits
  750. the number of entries in each hash bucket; the larger the value the
  751. more CIPSO label mappings that can be cached. When the number of
  752. entries in a given hash bucket reaches this limit adding new entries
  753. causes the oldest entry in the bucket to be removed to make room.
  754. Default: 10
  755. cipso_rbm_optfmt - BOOLEAN
  756. Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
  757. the CIPSO draft specification (see Documentation/netlabel for details).
  758. This means that when set the CIPSO tag will be padded with empty
  759. categories in order to make the packet data 32-bit aligned.
  760. Default: 0
  761. cipso_rbm_structvalid - BOOLEAN
  762. If set, do a very strict check of the CIPSO option when
  763. ip_options_compile() is called. If unset, relax the checks done during
  764. ip_options_compile(). Either way is "safe" as errors are caught else
  765. where in the CIPSO processing code but setting this to 0 (False) should
  766. result in less work (i.e. it should be faster) but could cause problems
  767. with other implementations that require strict checking.
  768. Default: 0
  769. IP Variables
  770. ============
  771. ip_local_port_range - 2 INTEGERS
  772. Defines the local port range that is used by TCP and UDP to
  773. choose the local port. The first number is the first, the
  774. second the last local port number.
  775. If possible, it is better these numbers have different parity
  776. (one even and one odd value).
  777. Must be greater than or equal to ip_unprivileged_port_start.
  778. The default values are 32768 and 60999 respectively.
  779. ip_local_reserved_ports - list of comma separated ranges
  780. Specify the ports which are reserved for known third-party
  781. applications. These ports will not be used by automatic port
  782. assignments (e.g. when calling connect() or bind() with port
  783. number 0). Explicit port allocation behavior is unchanged.
  784. The format used for both input and output is a comma separated
  785. list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
  786. 10). Writing to the file will clear all previously reserved
  787. ports and update the current list with the one given in the
  788. input.
  789. Note that ip_local_port_range and ip_local_reserved_ports
  790. settings are independent and both are considered by the kernel
  791. when determining which ports are available for automatic port
  792. assignments.
  793. You can reserve ports which are not in the current
  794. ip_local_port_range, e.g.::
  795. $ cat /proc/sys/net/ipv4/ip_local_port_range
  796. 32000 60999
  797. $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
  798. 8080,9148
  799. although this is redundant. However such a setting is useful
  800. if later the port range is changed to a value that will
  801. include the reserved ports.
  802. Default: Empty
  803. ip_local_unbindable_ports - list of comma separated ranges
  804. Specify the ports which are not directly bind()able.
  805. Usually you would use this to block the use of ports which
  806. are invalid due to something outside of the control of the
  807. kernel. For example a port stolen by the nic for serial
  808. console, remote power management or debugging.
  809. There's a relatively high chance you will also want to list
  810. these ports in 'ip_local_reserved_ports' to prevent autobinding.
  811. Default: Empty
  812. ip_unprivileged_port_start - INTEGER
  813. This is a per-namespace sysctl. It defines the first
  814. unprivileged port in the network namespace. Privileged ports
  815. require root or CAP_NET_BIND_SERVICE in order to bind to them.
  816. To disable all privileged ports, set this to 0. They must not
  817. overlap with the ip_local_port_range.
  818. Default: 1024
  819. ip_nonlocal_bind - BOOLEAN
  820. If set, allows processes to bind() to non-local IP addresses,
  821. which can be quite useful - but may break some applications.
  822. Default: 0
  823. ip_autobind_reuse - BOOLEAN
  824. By default, bind() does not select the ports automatically even if
  825. the new socket and all sockets bound to the port have SO_REUSEADDR.
  826. ip_autobind_reuse allows bind() to reuse the port and this is useful
  827. when you use bind()+connect(), but may break some applications.
  828. The preferred solution is to use IP_BIND_ADDRESS_NO_PORT and this
  829. option should only be set by experts.
  830. Default: 0
  831. ip_dynaddr - BOOLEAN
  832. If set non-zero, enables support for dynamic addresses.
  833. If set to a non-zero value larger than 1, a kernel log
  834. message will be printed when dynamic address rewriting
  835. occurs.
  836. Default: 0
  837. ip_early_demux - BOOLEAN
  838. Optimize input packet processing down to one demux for
  839. certain kinds of local sockets. Currently we only do this
  840. for established TCP and connected UDP sockets.
  841. It may add an additional cost for pure routing workloads that
  842. reduces overall throughput, in such case you should disable it.
  843. Default: 1
  844. ping_group_range - 2 INTEGERS
  845. Restrict ICMP_PROTO datagram sockets to users in the group range.
  846. The default is "1 0", meaning, that nobody (not even root) may
  847. create ping sockets. Setting it to "100 100" would grant permissions
  848. to the single group. "0 4294967295" would enable it for the world, "100
  849. 4294967295" would enable it for the users, but not daemons.
  850. tcp_early_demux - BOOLEAN
  851. Enable early demux for established TCP sockets.
  852. Default: 1
  853. udp_early_demux - BOOLEAN
  854. Enable early demux for connected UDP sockets. Disable this if
  855. your system could experience more unconnected load.
  856. Default: 1
  857. icmp_echo_ignore_all - BOOLEAN
  858. If set non-zero, then the kernel will ignore all ICMP ECHO
  859. requests sent to it.
  860. Default: 0
  861. icmp_echo_ignore_broadcasts - BOOLEAN
  862. If set non-zero, then the kernel will ignore all ICMP ECHO and
  863. TIMESTAMP requests sent to it via broadcast/multicast.
  864. Default: 1
  865. icmp_ratelimit - INTEGER
  866. Limit the maximal rates for sending ICMP packets whose type matches
  867. icmp_ratemask (see below) to specific targets.
  868. 0 to disable any limiting,
  869. otherwise the minimal space between responses in milliseconds.
  870. Note that another sysctl, icmp_msgs_per_sec limits the number
  871. of ICMP packets sent on all targets.
  872. Default: 1000
  873. icmp_msgs_per_sec - INTEGER
  874. Limit maximal number of ICMP packets sent per second from this host.
  875. Only messages whose type matches icmp_ratemask (see below) are
  876. controlled by this limit. For security reasons, the precise count
  877. of messages per second is randomized.
  878. Default: 1000
  879. icmp_msgs_burst - INTEGER
  880. icmp_msgs_per_sec controls number of ICMP packets sent per second,
  881. while icmp_msgs_burst controls the burst size of these packets.
  882. For security reasons, the precise burst size is randomized.
  883. Default: 50
  884. icmp_ratemask - INTEGER
  885. Mask made of ICMP types for which rates are being limited.
  886. Significant bits: IHGFEDCBA9876543210
  887. Default mask: 0000001100000011000 (6168)
  888. Bit definitions (see include/linux/icmp.h):
  889. = =========================
  890. 0 Echo Reply
  891. 3 Destination Unreachable [1]_
  892. 4 Source Quench [1]_
  893. 5 Redirect
  894. 8 Echo Request
  895. B Time Exceeded [1]_
  896. C Parameter Problem [1]_
  897. D Timestamp Request
  898. E Timestamp Reply
  899. F Info Request
  900. G Info Reply
  901. H Address Mask Request
  902. I Address Mask Reply
  903. = =========================
  904. .. [1] These are rate limited by default (see default mask above)
  905. icmp_ignore_bogus_error_responses - BOOLEAN
  906. Some routers violate RFC1122 by sending bogus responses to broadcast
  907. frames. Such violations are normally logged via a kernel warning.
  908. If this is set to TRUE, the kernel will not give such warnings, which
  909. will avoid log file clutter.
  910. Default: 1
  911. icmp_errors_use_inbound_ifaddr - BOOLEAN
  912. If zero, icmp error messages are sent with the primary address of
  913. the exiting interface.
  914. If non-zero, the message will be sent with the primary address of
  915. the interface that received the packet that caused the icmp error.
  916. This is the behaviour network many administrators will expect from
  917. a router. And it can make debugging complicated network layouts
  918. much easier.
  919. Note that if no primary address exists for the interface selected,
  920. then the primary address of the first non-loopback interface that
  921. has one will be used regardless of this setting.
  922. Default: 0
  923. igmp_max_memberships - INTEGER
  924. Change the maximum number of multicast groups we can subscribe to.
  925. Default: 20
  926. Theoretical maximum value is bounded by having to send a membership
  927. report in a single datagram (i.e. the report can't span multiple
  928. datagrams, or risk confusing the switch and leaving groups you don't
  929. intend to).
  930. The number of supported groups 'M' is bounded by the number of group
  931. report entries you can fit into a single datagram of 65535 bytes.
  932. M = 65536-sizeof (ip header)/(sizeof(Group record))
  933. Group records are variable length, with a minimum of 12 bytes.
  934. So net.ipv4.igmp_max_memberships should not be set higher than:
  935. (65536-24) / 12 = 5459
  936. The value 5459 assumes no IP header options, so in practice
  937. this number may be lower.
  938. igmp_max_msf - INTEGER
  939. Maximum number of addresses allowed in the source filter list for a
  940. multicast group.
  941. Default: 10
  942. igmp_qrv - INTEGER
  943. Controls the IGMP query robustness variable (see RFC2236 8.1).
  944. Default: 2 (as specified by RFC2236 8.1)
  945. Minimum: 1 (as specified by RFC6636 4.5)
  946. force_igmp_version - INTEGER
  947. - 0 - (default) No enforcement of a IGMP version, IGMPv1/v2 fallback
  948. allowed. Will back to IGMPv3 mode again if all IGMPv1/v2 Querier
  949. Present timer expires.
  950. - 1 - Enforce to use IGMP version 1. Will also reply IGMPv1 report if
  951. receive IGMPv2/v3 query.
  952. - 2 - Enforce to use IGMP version 2. Will fallback to IGMPv1 if receive
  953. IGMPv1 query message. Will reply report if receive IGMPv3 query.
  954. - 3 - Enforce to use IGMP version 3. The same react with default 0.
  955. .. note::
  956. this is not the same with force_mld_version because IGMPv3 RFC3376
  957. Security Considerations does not have clear description that we could
  958. ignore other version messages completely as MLDv2 RFC3810. So make
  959. this value as default 0 is recommended.
  960. ``conf/interface/*``
  961. changes special settings per interface (where
  962. interface" is the name of your network interface)
  963. ``conf/all/*``
  964. is special, changes the settings for all interfaces
  965. log_martians - BOOLEAN
  966. Log packets with impossible addresses to kernel log.
  967. log_martians for the interface will be enabled if at least one of
  968. conf/{all,interface}/log_martians is set to TRUE,
  969. it will be disabled otherwise
  970. accept_redirects - BOOLEAN
  971. Accept ICMP redirect messages.
  972. accept_redirects for the interface will be enabled if:
  973. - both conf/{all,interface}/accept_redirects are TRUE in the case
  974. forwarding for the interface is enabled
  975. or
  976. - at least one of conf/{all,interface}/accept_redirects is TRUE in the
  977. case forwarding for the interface is disabled
  978. accept_redirects for the interface will be disabled otherwise
  979. default:
  980. - TRUE (host)
  981. - FALSE (router)
  982. forwarding - BOOLEAN
  983. Enable IP forwarding on this interface. This controls whether packets
  984. received _on_ this interface can be forwarded.
  985. mc_forwarding - BOOLEAN
  986. Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
  987. and a multicast routing daemon is required.
  988. conf/all/mc_forwarding must also be set to TRUE to enable multicast
  989. routing for the interface
  990. medium_id - INTEGER
  991. Integer value used to differentiate the devices by the medium they
  992. are attached to. Two devices can have different id values when
  993. the broadcast packets are received only on one of them.
  994. The default value 0 means that the device is the only interface
  995. to its medium, value of -1 means that medium is not known.
  996. Currently, it is used to change the proxy_arp behavior:
  997. the proxy_arp feature is enabled for packets forwarded between
  998. two devices attached to different media.
  999. proxy_arp - BOOLEAN
  1000. Do proxy arp.
  1001. proxy_arp for the interface will be enabled if at least one of
  1002. conf/{all,interface}/proxy_arp is set to TRUE,
  1003. it will be disabled otherwise
  1004. proxy_arp_pvlan - BOOLEAN
  1005. Private VLAN proxy arp.
  1006. Basically allow proxy arp replies back to the same interface
  1007. (from which the ARP request/solicitation was received).
  1008. This is done to support (ethernet) switch features, like RFC
  1009. 3069, where the individual ports are NOT allowed to
  1010. communicate with each other, but they are allowed to talk to
  1011. the upstream router. As described in RFC 3069, it is possible
  1012. to allow these hosts to communicate through the upstream
  1013. router by proxy_arp'ing. Don't need to be used together with
  1014. proxy_arp.
  1015. This technology is known by different names:
  1016. In RFC 3069 it is called VLAN Aggregation.
  1017. Cisco and Allied Telesyn call it Private VLAN.
  1018. Hewlett-Packard call it Source-Port filtering or port-isolation.
  1019. Ericsson call it MAC-Forced Forwarding (RFC Draft).
  1020. shared_media - BOOLEAN
  1021. Send(router) or accept(host) RFC1620 shared media redirects.
  1022. Overrides secure_redirects.
  1023. shared_media for the interface will be enabled if at least one of
  1024. conf/{all,interface}/shared_media is set to TRUE,
  1025. it will be disabled otherwise
  1026. default TRUE
  1027. secure_redirects - BOOLEAN
  1028. Accept ICMP redirect messages only to gateways listed in the
  1029. interface's current gateway list. Even if disabled, RFC1122 redirect
  1030. rules still apply.
  1031. Overridden by shared_media.
  1032. secure_redirects for the interface will be enabled if at least one of
  1033. conf/{all,interface}/secure_redirects is set to TRUE,
  1034. it will be disabled otherwise
  1035. default TRUE
  1036. send_redirects - BOOLEAN
  1037. Send redirects, if router.
  1038. send_redirects for the interface will be enabled if at least one of
  1039. conf/{all,interface}/send_redirects is set to TRUE,
  1040. it will be disabled otherwise
  1041. Default: TRUE
  1042. bootp_relay - BOOLEAN
  1043. Accept packets with source address 0.b.c.d destined
  1044. not to this host as local ones. It is supposed, that
  1045. BOOTP relay daemon will catch and forward such packets.
  1046. conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
  1047. for the interface
  1048. default FALSE
  1049. Not Implemented Yet.
  1050. accept_source_route - BOOLEAN
  1051. Accept packets with SRR option.
  1052. conf/all/accept_source_route must also be set to TRUE to accept packets
  1053. with SRR option on the interface
  1054. default
  1055. - TRUE (router)
  1056. - FALSE (host)
  1057. accept_local - BOOLEAN
  1058. Accept packets with local source addresses. In combination with
  1059. suitable routing, this can be used to direct packets between two
  1060. local interfaces over the wire and have them accepted properly.
  1061. default FALSE
  1062. route_localnet - BOOLEAN
  1063. Do not consider loopback addresses as martian source or destination
  1064. while routing. This enables the use of 127/8 for local routing purposes.
  1065. default FALSE
  1066. rp_filter - INTEGER
  1067. - 0 - No source validation.
  1068. - 1 - Strict mode as defined in RFC3704 Strict Reverse Path
  1069. Each incoming packet is tested against the FIB and if the interface
  1070. is not the best reverse path the packet check will fail.
  1071. By default failed packets are discarded.
  1072. - 2 - Loose mode as defined in RFC3704 Loose Reverse Path
  1073. Each incoming packet's source address is also tested against the FIB
  1074. and if the source address is not reachable via any interface
  1075. the packet check will fail.
  1076. Current recommended practice in RFC3704 is to enable strict mode
  1077. to prevent IP spoofing from DDos attacks. If using asymmetric routing
  1078. or other complicated routing, then loose mode is recommended.
  1079. The max value from conf/{all,interface}/rp_filter is used
  1080. when doing source validation on the {interface}.
  1081. Default value is 0. Note that some distributions enable it
  1082. in startup scripts.
  1083. arp_filter - BOOLEAN
  1084. - 1 - Allows you to have multiple network interfaces on the same
  1085. subnet, and have the ARPs for each interface be answered
  1086. based on whether or not the kernel would route a packet from
  1087. the ARP'd IP out that interface (therefore you must use source
  1088. based routing for this to work). In other words it allows control
  1089. of which cards (usually 1) will respond to an arp request.
  1090. - 0 - (default) The kernel can respond to arp requests with addresses
  1091. from other interfaces. This may seem wrong but it usually makes
  1092. sense, because it increases the chance of successful communication.
  1093. IP addresses are owned by the complete host on Linux, not by
  1094. particular interfaces. Only for more complex setups like load-
  1095. balancing, does this behaviour cause problems.
  1096. arp_filter for the interface will be enabled if at least one of
  1097. conf/{all,interface}/arp_filter is set to TRUE,
  1098. it will be disabled otherwise
  1099. arp_announce - INTEGER
  1100. Define different restriction levels for announcing the local
  1101. source IP address from IP packets in ARP requests sent on
  1102. interface:
  1103. - 0 - (default) Use any local address, configured on any interface
  1104. - 1 - Try to avoid local addresses that are not in the target's
  1105. subnet for this interface. This mode is useful when target
  1106. hosts reachable via this interface require the source IP
  1107. address in ARP requests to be part of their logical network
  1108. configured on the receiving interface. When we generate the
  1109. request we will check all our subnets that include the
  1110. target IP and will preserve the source address if it is from
  1111. such subnet. If there is no such subnet we select source
  1112. address according to the rules for level 2.
  1113. - 2 - Always use the best local address for this target.
  1114. In this mode we ignore the source address in the IP packet
  1115. and try to select local address that we prefer for talks with
  1116. the target host. Such local address is selected by looking
  1117. for primary IP addresses on all our subnets on the outgoing
  1118. interface that include the target IP address. If no suitable
  1119. local address is found we select the first local address
  1120. we have on the outgoing interface or on all other interfaces,
  1121. with the hope we will receive reply for our request and
  1122. even sometimes no matter the source IP address we announce.
  1123. The max value from conf/{all,interface}/arp_announce is used.
  1124. Increasing the restriction level gives more chance for
  1125. receiving answer from the resolved target while decreasing
  1126. the level announces more valid sender's information.
  1127. arp_ignore - INTEGER
  1128. Define different modes for sending replies in response to
  1129. received ARP requests that resolve local target IP addresses:
  1130. - 0 - (default): reply for any local target IP address, configured
  1131. on any interface
  1132. - 1 - reply only if the target IP address is local address
  1133. configured on the incoming interface
  1134. - 2 - reply only if the target IP address is local address
  1135. configured on the incoming interface and both with the
  1136. sender's IP address are part from same subnet on this interface
  1137. - 3 - do not reply for local addresses configured with scope host,
  1138. only resolutions for global and link addresses are replied
  1139. - 4-7 - reserved
  1140. - 8 - do not reply for all local addresses
  1141. The max value from conf/{all,interface}/arp_ignore is used
  1142. when ARP request is received on the {interface}
  1143. arp_notify - BOOLEAN
  1144. Define mode for notification of address and device changes.
  1145. == ==========================================================
  1146. 0 (default): do nothing
  1147. 1 Generate gratuitous arp requests when device is brought up
  1148. or hardware address changes.
  1149. == ==========================================================
  1150. arp_accept - BOOLEAN
  1151. Define behavior for gratuitous ARP frames who's IP is not
  1152. already present in the ARP table:
  1153. - 0 - don't create new entries in the ARP table
  1154. - 1 - create new entries in the ARP table
  1155. Both replies and requests type gratuitous arp will trigger the
  1156. ARP table to be updated, if this setting is on.
  1157. If the ARP table already contains the IP address of the
  1158. gratuitous arp frame, the arp table will be updated regardless
  1159. if this setting is on or off.
  1160. mcast_solicit - INTEGER
  1161. The maximum number of multicast probes in INCOMPLETE state,
  1162. when the associated hardware address is unknown. Defaults
  1163. to 3.
  1164. ucast_solicit - INTEGER
  1165. The maximum number of unicast probes in PROBE state, when
  1166. the hardware address is being reconfirmed. Defaults to 3.
  1167. app_solicit - INTEGER
  1168. The maximum number of probes to send to the user space ARP daemon
  1169. via netlink before dropping back to multicast probes (see
  1170. mcast_resolicit). Defaults to 0.
  1171. mcast_resolicit - INTEGER
  1172. The maximum number of multicast probes after unicast and
  1173. app probes in PROBE state. Defaults to 0.
  1174. disable_policy - BOOLEAN
  1175. Disable IPSEC policy (SPD) for this interface
  1176. disable_xfrm - BOOLEAN
  1177. Disable IPSEC encryption on this interface, whatever the policy
  1178. igmpv2_unsolicited_report_interval - INTEGER
  1179. The interval in milliseconds in which the next unsolicited
  1180. IGMPv1 or IGMPv2 report retransmit will take place.
  1181. Default: 10000 (10 seconds)
  1182. igmpv3_unsolicited_report_interval - INTEGER
  1183. The interval in milliseconds in which the next unsolicited
  1184. IGMPv3 report retransmit will take place.
  1185. Default: 1000 (1 seconds)
  1186. promote_secondaries - BOOLEAN
  1187. When a primary IP address is removed from this interface
  1188. promote a corresponding secondary IP address instead of
  1189. removing all the corresponding secondary IP addresses.
  1190. drop_unicast_in_l2_multicast - BOOLEAN
  1191. Drop any unicast IP packets that are received in link-layer
  1192. multicast (or broadcast) frames.
  1193. This behavior (for multicast) is actually a SHOULD in RFC
  1194. 1122, but is disabled by default for compatibility reasons.
  1195. Default: off (0)
  1196. drop_gratuitous_arp - BOOLEAN
  1197. Drop all gratuitous ARP frames, for example if there's a known
  1198. good ARP proxy on the network and such frames need not be used
  1199. (or in the case of 802.11, must not be used to prevent attacks.)
  1200. Default: off (0)
  1201. tag - INTEGER
  1202. Allows you to write a number, which can be used as required.
  1203. Default value is 0.
  1204. xfrm4_gc_thresh - INTEGER
  1205. (Obsolete since linux-4.14)
  1206. The threshold at which we will start garbage collecting for IPv4
  1207. destination cache entries. At twice this value the system will
  1208. refuse new allocations.
  1209. igmp_link_local_mcast_reports - BOOLEAN
  1210. Enable IGMP reports for link local multicast groups in the
  1211. 224.0.0.X range.
  1212. Default TRUE
  1213. Alexey Kuznetsov.
  1214. kuznet@ms2.inr.ac.ru
  1215. Updated by:
  1216. - Andi Kleen
  1217. ak@muc.de
  1218. - Nicolas Delon
  1219. delon.nicolas@wanadoo.fr
  1220. /proc/sys/net/ipv6/* Variables
  1221. ==============================
  1222. IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
  1223. apply to IPv6 [XXX?].
  1224. bindv6only - BOOLEAN
  1225. Default value for IPV6_V6ONLY socket option,
  1226. which restricts use of the IPv6 socket to IPv6 communication
  1227. only.
  1228. - TRUE: disable IPv4-mapped address feature
  1229. - FALSE: enable IPv4-mapped address feature
  1230. Default: FALSE (as specified in RFC3493)
  1231. flowlabel_consistency - BOOLEAN
  1232. Protect the consistency (and unicity) of flow label.
  1233. You have to disable it to use IPV6_FL_F_REFLECT flag on the
  1234. flow label manager.
  1235. - TRUE: enabled
  1236. - FALSE: disabled
  1237. Default: TRUE
  1238. auto_flowlabels - INTEGER
  1239. Automatically generate flow labels based on a flow hash of the
  1240. packet. This allows intermediate devices, such as routers, to
  1241. identify packet flows for mechanisms like Equal Cost Multipath
  1242. Routing (see RFC 6438).
  1243. = ===========================================================
  1244. 0 automatic flow labels are completely disabled
  1245. 1 automatic flow labels are enabled by default, they can be
  1246. disabled on a per socket basis using the IPV6_AUTOFLOWLABEL
  1247. socket option
  1248. 2 automatic flow labels are allowed, they may be enabled on a
  1249. per socket basis using the IPV6_AUTOFLOWLABEL socket option
  1250. 3 automatic flow labels are enabled and enforced, they cannot
  1251. be disabled by the socket option
  1252. = ===========================================================
  1253. Default: 1
  1254. flowlabel_state_ranges - BOOLEAN
  1255. Split the flow label number space into two ranges. 0-0x7FFFF is
  1256. reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF
  1257. is reserved for stateless flow labels as described in RFC6437.
  1258. - TRUE: enabled
  1259. - FALSE: disabled
  1260. Default: true
  1261. flowlabel_reflect - INTEGER
  1262. Control flow label reflection. Needed for Path MTU
  1263. Discovery to work with Equal Cost Multipath Routing in anycast
  1264. environments. See RFC 7690 and:
  1265. https://tools.ietf.org/html/draft-wang-6man-flow-label-reflection-01
  1266. This is a bitmask.
  1267. - 1: enabled for established flows
  1268. Note that this prevents automatic flowlabel changes, as done
  1269. in "tcp: change IPv6 flow-label upon receiving spurious retransmission"
  1270. and "tcp: Change txhash on every SYN and RTO retransmit"
  1271. - 2: enabled for TCP RESET packets (no active listener)
  1272. If set, a RST packet sent in response to a SYN packet on a closed
  1273. port will reflect the incoming flow label.
  1274. - 4: enabled for ICMPv6 echo reply messages.
  1275. Default: 0
  1276. fib_multipath_hash_policy - INTEGER
  1277. Controls which hash policy to use for multipath routes.
  1278. Default: 0 (Layer 3)
  1279. Possible values:
  1280. - 0 - Layer 3 (source and destination addresses plus flow label)
  1281. - 1 - Layer 4 (standard 5-tuple)
  1282. - 2 - Layer 3 or inner Layer 3 if present
  1283. anycast_src_echo_reply - BOOLEAN
  1284. Controls the use of anycast addresses as source addresses for ICMPv6
  1285. echo reply
  1286. - TRUE: enabled
  1287. - FALSE: disabled
  1288. Default: FALSE
  1289. idgen_delay - INTEGER
  1290. Controls the delay in seconds after which time to retry
  1291. privacy stable address generation if a DAD conflict is
  1292. detected.
  1293. Default: 1 (as specified in RFC7217)
  1294. idgen_retries - INTEGER
  1295. Controls the number of retries to generate a stable privacy
  1296. address if a DAD conflict is detected.
  1297. Default: 3 (as specified in RFC7217)
  1298. mld_qrv - INTEGER
  1299. Controls the MLD query robustness variable (see RFC3810 9.1).
  1300. Default: 2 (as specified by RFC3810 9.1)
  1301. Minimum: 1 (as specified by RFC6636 4.5)
  1302. max_dst_opts_number - INTEGER
  1303. Maximum number of non-padding TLVs allowed in a Destination
  1304. options extension header. If this value is less than zero
  1305. then unknown options are disallowed and the number of known
  1306. TLVs allowed is the absolute value of this number.
  1307. Default: 8
  1308. max_hbh_opts_number - INTEGER
  1309. Maximum number of non-padding TLVs allowed in a Hop-by-Hop
  1310. options extension header. If this value is less than zero
  1311. then unknown options are disallowed and the number of known
  1312. TLVs allowed is the absolute value of this number.
  1313. Default: 8
  1314. max_dst_opts_length - INTEGER
  1315. Maximum length allowed for a Destination options extension
  1316. header.
  1317. Default: INT_MAX (unlimited)
  1318. max_hbh_length - INTEGER
  1319. Maximum length allowed for a Hop-by-Hop options extension
  1320. header.
  1321. Default: INT_MAX (unlimited)
  1322. skip_notify_on_dev_down - BOOLEAN
  1323. Controls whether an RTM_DELROUTE message is generated for routes
  1324. removed when a device is taken down or deleted. IPv4 does not
  1325. generate this message; IPv6 does by default. Setting this sysctl
  1326. to true skips the message, making IPv4 and IPv6 on par in relying
  1327. on userspace caches to track link events and evict routes.
  1328. Default: false (generate message)
  1329. nexthop_compat_mode - BOOLEAN
  1330. New nexthop API provides a means for managing nexthops independent of
  1331. prefixes. Backwards compatibilty with old route format is enabled by
  1332. default which means route dumps and notifications contain the new
  1333. nexthop attribute but also the full, expanded nexthop definition.
  1334. Further, updates or deletes of a nexthop configuration generate route
  1335. notifications for each fib entry using the nexthop. Once a system
  1336. understands the new API, this sysctl can be disabled to achieve full
  1337. performance benefits of the new API by disabling the nexthop expansion
  1338. and extraneous notifications.
  1339. Default: true (backward compat mode)
  1340. IPv6 Fragmentation:
  1341. ip6frag_high_thresh - INTEGER
  1342. Maximum memory used to reassemble IPv6 fragments. When
  1343. ip6frag_high_thresh bytes of memory is allocated for this purpose,
  1344. the fragment handler will toss packets until ip6frag_low_thresh
  1345. is reached.
  1346. ip6frag_low_thresh - INTEGER
  1347. See ip6frag_high_thresh
  1348. ip6frag_time - INTEGER
  1349. Time in seconds to keep an IPv6 fragment in memory.
  1350. IPv6 Segment Routing:
  1351. seg6_flowlabel - INTEGER
  1352. Controls the behaviour of computing the flowlabel of outer
  1353. IPv6 header in case of SR T.encaps
  1354. == =======================================================
  1355. -1 set flowlabel to zero.
  1356. 0 copy flowlabel from Inner packet in case of Inner IPv6
  1357. (Set flowlabel to 0 in case IPv4/L2)
  1358. 1 Compute the flowlabel using seg6_make_flowlabel()
  1359. == =======================================================
  1360. Default is 0.
  1361. ``conf/default/*``:
  1362. Change the interface-specific default settings.
  1363. ``conf/all/*``:
  1364. Change all the interface-specific settings.
  1365. [XXX: Other special features than forwarding?]
  1366. conf/all/forwarding - BOOLEAN
  1367. Enable global IPv6 forwarding between all interfaces.
  1368. IPv4 and IPv6 work differently here; e.g. netfilter must be used
  1369. to control which interfaces may forward packets and which not.
  1370. This also sets all interfaces' Host/Router setting
  1371. 'forwarding' to the specified value. See below for details.
  1372. This referred to as global forwarding.
  1373. proxy_ndp - BOOLEAN
  1374. Do proxy ndp.
  1375. fwmark_reflect - BOOLEAN
  1376. Controls the fwmark of kernel-generated IPv6 reply packets that are not
  1377. associated with a socket for example, TCP RSTs or ICMPv6 echo replies).
  1378. If unset, these packets have a fwmark of zero. If set, they have the
  1379. fwmark of the packet they are replying to.
  1380. Default: 0
  1381. ``conf/interface/*``:
  1382. Change special settings per interface.
  1383. The functional behaviour for certain settings is different
  1384. depending on whether local forwarding is enabled or not.
  1385. accept_ra - INTEGER
  1386. Accept Router Advertisements; autoconfigure using them.
  1387. It also determines whether or not to transmit Router
  1388. Solicitations. If and only if the functional setting is to
  1389. accept Router Advertisements, Router Solicitations will be
  1390. transmitted.
  1391. Possible values are:
  1392. == ===========================================================
  1393. 0 Do not accept Router Advertisements.
  1394. 1 Accept Router Advertisements if forwarding is disabled.
  1395. 2 Overrule forwarding behaviour. Accept Router Advertisements
  1396. even if forwarding is enabled.
  1397. == ===========================================================
  1398. Functional default:
  1399. - enabled if local forwarding is disabled.
  1400. - disabled if local forwarding is enabled.
  1401. accept_ra_defrtr - BOOLEAN
  1402. Learn default router in Router Advertisement.
  1403. Functional default:
  1404. - enabled if accept_ra is enabled.
  1405. - disabled if accept_ra is disabled.
  1406. accept_ra_from_local - BOOLEAN
  1407. Accept RA with source-address that is found on local machine
  1408. if the RA is otherwise proper and able to be accepted.
  1409. Default is to NOT accept these as it may be an un-intended
  1410. network loop.
  1411. Functional default:
  1412. - enabled if accept_ra_from_local is enabled
  1413. on a specific interface.
  1414. - disabled if accept_ra_from_local is disabled
  1415. on a specific interface.
  1416. accept_ra_min_hop_limit - INTEGER
  1417. Minimum hop limit Information in Router Advertisement.
  1418. Hop limit Information in Router Advertisement less than this
  1419. variable shall be ignored.
  1420. Default: 1
  1421. accept_ra_pinfo - BOOLEAN
  1422. Learn Prefix Information in Router Advertisement.
  1423. Functional default:
  1424. - enabled if accept_ra is enabled.
  1425. - disabled if accept_ra is disabled.
  1426. accept_ra_rt_info_min_plen - INTEGER
  1427. Minimum prefix length of Route Information in RA.
  1428. Route Information w/ prefix smaller than this variable shall
  1429. be ignored.
  1430. Functional default:
  1431. * 0 if accept_ra_rtr_pref is enabled.
  1432. * -1 if accept_ra_rtr_pref is disabled.
  1433. accept_ra_rt_info_max_plen - INTEGER
  1434. Maximum prefix length of Route Information in RA.
  1435. Route Information w/ prefix larger than this variable shall
  1436. be ignored.
  1437. Functional default:
  1438. * 0 if accept_ra_rtr_pref is enabled.
  1439. * -1 if accept_ra_rtr_pref is disabled.
  1440. accept_ra_rtr_pref - BOOLEAN
  1441. Accept Router Preference in RA.
  1442. Functional default:
  1443. - enabled if accept_ra is enabled.
  1444. - disabled if accept_ra is disabled.
  1445. accept_ra_mtu - BOOLEAN
  1446. Apply the MTU value specified in RA option 5 (RFC4861). If
  1447. disabled, the MTU specified in the RA will be ignored.
  1448. Functional default:
  1449. - enabled if accept_ra is enabled.
  1450. - disabled if accept_ra is disabled.
  1451. accept_redirects - BOOLEAN
  1452. Accept Redirects.
  1453. Functional default:
  1454. - enabled if local forwarding is disabled.
  1455. - disabled if local forwarding is enabled.
  1456. accept_source_route - INTEGER
  1457. Accept source routing (routing extension header).
  1458. - >= 0: Accept only routing header type 2.
  1459. - < 0: Do not accept routing header.
  1460. Default: 0
  1461. autoconf - BOOLEAN
  1462. Autoconfigure addresses using Prefix Information in Router
  1463. Advertisements.
  1464. Functional default:
  1465. - enabled if accept_ra_pinfo is enabled.
  1466. - disabled if accept_ra_pinfo is disabled.
  1467. dad_transmits - INTEGER
  1468. The amount of Duplicate Address Detection probes to send.
  1469. Default: 1
  1470. forwarding - INTEGER
  1471. Configure interface-specific Host/Router behaviour.
  1472. .. note::
  1473. It is recommended to have the same setting on all
  1474. interfaces; mixed router/host scenarios are rather uncommon.
  1475. Possible values are:
  1476. - 0 Forwarding disabled
  1477. - 1 Forwarding enabled
  1478. **FALSE (0)**:
  1479. By default, Host behaviour is assumed. This means:
  1480. 1. IsRouter flag is not set in Neighbour Advertisements.
  1481. 2. If accept_ra is TRUE (default), transmit Router
  1482. Solicitations.
  1483. 3. If accept_ra is TRUE (default), accept Router
  1484. Advertisements (and do autoconfiguration).
  1485. 4. If accept_redirects is TRUE (default), accept Redirects.
  1486. **TRUE (1)**:
  1487. If local forwarding is enabled, Router behaviour is assumed.
  1488. This means exactly the reverse from the above:
  1489. 1. IsRouter flag is set in Neighbour Advertisements.
  1490. 2. Router Solicitations are not sent unless accept_ra is 2.
  1491. 3. Router Advertisements are ignored unless accept_ra is 2.
  1492. 4. Redirects are ignored.
  1493. Default: 0 (disabled) if global forwarding is disabled (default),
  1494. otherwise 1 (enabled).
  1495. hop_limit - INTEGER
  1496. Default Hop Limit to set.
  1497. Default: 64
  1498. mtu - INTEGER
  1499. Default Maximum Transfer Unit
  1500. Default: 1280 (IPv6 required minimum)
  1501. ip_nonlocal_bind - BOOLEAN
  1502. If set, allows processes to bind() to non-local IPv6 addresses,
  1503. which can be quite useful - but may break some applications.
  1504. Default: 0
  1505. router_probe_interval - INTEGER
  1506. Minimum interval (in seconds) between Router Probing described
  1507. in RFC4191.
  1508. Default: 60
  1509. router_solicitation_delay - INTEGER
  1510. Number of seconds to wait after interface is brought up
  1511. before sending Router Solicitations.
  1512. Default: 1
  1513. router_solicitation_interval - INTEGER
  1514. Number of seconds to wait between Router Solicitations.
  1515. Default: 4
  1516. router_solicitations - INTEGER
  1517. Number of Router Solicitations to send until assuming no
  1518. routers are present.
  1519. Default: 3
  1520. use_oif_addrs_only - BOOLEAN
  1521. When enabled, the candidate source addresses for destinations
  1522. routed via this interface are restricted to the set of addresses
  1523. configured on this interface (vis. RFC 6724, section 4).
  1524. Default: false
  1525. use_tempaddr - INTEGER
  1526. Preference for Privacy Extensions (RFC3041).
  1527. * <= 0 : disable Privacy Extensions
  1528. * == 1 : enable Privacy Extensions, but prefer public
  1529. addresses over temporary addresses.
  1530. * > 1 : enable Privacy Extensions and prefer temporary
  1531. addresses over public addresses.
  1532. Default:
  1533. * 0 (for most devices)
  1534. * -1 (for point-to-point devices and loopback devices)
  1535. temp_valid_lft - INTEGER
  1536. valid lifetime (in seconds) for temporary addresses.
  1537. Default: 172800 (2 days)
  1538. temp_prefered_lft - INTEGER
  1539. Preferred lifetime (in seconds) for temporary addresses.
  1540. Default: 86400 (1 day)
  1541. keep_addr_on_down - INTEGER
  1542. Keep all IPv6 addresses on an interface down event. If set static
  1543. global addresses with no expiration time are not flushed.
  1544. * >0 : enabled
  1545. * 0 : system default
  1546. * <0 : disabled
  1547. Default: 0 (addresses are removed)
  1548. max_desync_factor - INTEGER
  1549. Maximum value for DESYNC_FACTOR, which is a random value
  1550. that ensures that clients don't synchronize with each
  1551. other and generate new addresses at exactly the same time.
  1552. value is in seconds.
  1553. Default: 600
  1554. regen_max_retry - INTEGER
  1555. Number of attempts before give up attempting to generate
  1556. valid temporary addresses.
  1557. Default: 5
  1558. max_addresses - INTEGER
  1559. Maximum number of autoconfigured addresses per interface. Setting
  1560. to zero disables the limitation. It is not recommended to set this
  1561. value too large (or to zero) because it would be an easy way to
  1562. crash the kernel by allowing too many addresses to be created.
  1563. Default: 16
  1564. disable_ipv6 - BOOLEAN
  1565. Disable IPv6 operation. If accept_dad is set to 2, this value
  1566. will be dynamically set to TRUE if DAD fails for the link-local
  1567. address.
  1568. Default: FALSE (enable IPv6 operation)
  1569. When this value is changed from 1 to 0 (IPv6 is being enabled),
  1570. it will dynamically create a link-local address on the given
  1571. interface and start Duplicate Address Detection, if necessary.
  1572. When this value is changed from 0 to 1 (IPv6 is being disabled),
  1573. it will dynamically delete all addresses and routes on the given
  1574. interface. From now on it will not possible to add addresses/routes
  1575. to the selected interface.
  1576. accept_dad - INTEGER
  1577. Whether to accept DAD (Duplicate Address Detection).
  1578. == ==============================================================
  1579. 0 Disable DAD
  1580. 1 Enable DAD (default)
  1581. 2 Enable DAD, and disable IPv6 operation if MAC-based duplicate
  1582. link-local address has been found.
  1583. == ==============================================================
  1584. DAD operation and mode on a given interface will be selected according
  1585. to the maximum value of conf/{all,interface}/accept_dad.
  1586. force_tllao - BOOLEAN
  1587. Enable sending the target link-layer address option even when
  1588. responding to a unicast neighbor solicitation.
  1589. Default: FALSE
  1590. Quoting from RFC 2461, section 4.4, Target link-layer address:
  1591. "The option MUST be included for multicast solicitations in order to
  1592. avoid infinite Neighbor Solicitation "recursion" when the peer node
  1593. does not have a cache entry to return a Neighbor Advertisements
  1594. message. When responding to unicast solicitations, the option can be
  1595. omitted since the sender of the solicitation has the correct link-
  1596. layer address; otherwise it would not have be able to send the unicast
  1597. solicitation in the first place. However, including the link-layer
  1598. address in this case adds little overhead and eliminates a potential
  1599. race condition where the sender deletes the cached link-layer address
  1600. prior to receiving a response to a previous solicitation."
  1601. ndisc_notify - BOOLEAN
  1602. Define mode for notification of address and device changes.
  1603. * 0 - (default): do nothing
  1604. * 1 - Generate unsolicited neighbour advertisements when device is brought
  1605. up or hardware address changes.
  1606. ndisc_tclass - INTEGER
  1607. The IPv6 Traffic Class to use by default when sending IPv6 Neighbor
  1608. Discovery (Router Solicitation, Router Advertisement, Neighbor
  1609. Solicitation, Neighbor Advertisement, Redirect) messages.
  1610. These 8 bits can be interpreted as 6 high order bits holding the DSCP
  1611. value and 2 low order bits representing ECN (which you probably want
  1612. to leave cleared).
  1613. * 0 - (default)
  1614. mldv1_unsolicited_report_interval - INTEGER
  1615. The interval in milliseconds in which the next unsolicited
  1616. MLDv1 report retransmit will take place.
  1617. Default: 10000 (10 seconds)
  1618. mldv2_unsolicited_report_interval - INTEGER
  1619. The interval in milliseconds in which the next unsolicited
  1620. MLDv2 report retransmit will take place.
  1621. Default: 1000 (1 second)
  1622. force_mld_version - INTEGER
  1623. * 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
  1624. * 1 - Enforce to use MLD version 1
  1625. * 2 - Enforce to use MLD version 2
  1626. suppress_frag_ndisc - INTEGER
  1627. Control RFC 6980 (Security Implications of IPv6 Fragmentation
  1628. with IPv6 Neighbor Discovery) behavior:
  1629. * 1 - (default) discard fragmented neighbor discovery packets
  1630. * 0 - allow fragmented neighbor discovery packets
  1631. optimistic_dad - BOOLEAN
  1632. Whether to perform Optimistic Duplicate Address Detection (RFC 4429).
  1633. * 0: disabled (default)
  1634. * 1: enabled
  1635. Optimistic Duplicate Address Detection for the interface will be enabled
  1636. if at least one of conf/{all,interface}/optimistic_dad is set to 1,
  1637. it will be disabled otherwise.
  1638. use_optimistic - BOOLEAN
  1639. If enabled, do not classify optimistic addresses as deprecated during
  1640. source address selection. Preferred addresses will still be chosen
  1641. before optimistic addresses, subject to other ranking in the source
  1642. address selection algorithm.
  1643. * 0: disabled (default)
  1644. * 1: enabled
  1645. This will be enabled if at least one of
  1646. conf/{all,interface}/use_optimistic is set to 1, disabled otherwise.
  1647. stable_secret - IPv6 address
  1648. This IPv6 address will be used as a secret to generate IPv6
  1649. addresses for link-local addresses and autoconfigured
  1650. ones. All addresses generated after setting this secret will
  1651. be stable privacy ones by default. This can be changed via the
  1652. addrgenmode ip-link. conf/default/stable_secret is used as the
  1653. secret for the namespace, the interface specific ones can
  1654. overwrite that. Writes to conf/all/stable_secret are refused.
  1655. It is recommended to generate this secret during installation
  1656. of a system and keep it stable after that.
  1657. By default the stable secret is unset.
  1658. addr_gen_mode - INTEGER
  1659. Defines how link-local and autoconf addresses are generated.
  1660. = =================================================================
  1661. 0 generate address based on EUI64 (default)
  1662. 1 do no generate a link-local address, use EUI64 for addresses
  1663. generated from autoconf
  1664. 2 generate stable privacy addresses, using the secret from
  1665. stable_secret (RFC7217)
  1666. 3 generate stable privacy addresses, using a random secret if unset
  1667. = =================================================================
  1668. drop_unicast_in_l2_multicast - BOOLEAN
  1669. Drop any unicast IPv6 packets that are received in link-layer
  1670. multicast (or broadcast) frames.
  1671. By default this is turned off.
  1672. drop_unsolicited_na - BOOLEAN
  1673. Drop all unsolicited neighbor advertisements, for example if there's
  1674. a known good NA proxy on the network and such frames need not be used
  1675. (or in the case of 802.11, must not be used to prevent attacks.)
  1676. By default this is turned off.
  1677. enhanced_dad - BOOLEAN
  1678. Include a nonce option in the IPv6 neighbor solicitation messages used for
  1679. duplicate address detection per RFC7527. A received DAD NS will only signal
  1680. a duplicate address if the nonce is different. This avoids any false
  1681. detection of duplicates due to loopback of the NS messages that we send.
  1682. The nonce option will be sent on an interface unless both of
  1683. conf/{all,interface}/enhanced_dad are set to FALSE.
  1684. Default: TRUE
  1685. ``icmp/*``:
  1686. ===========
  1687. ratelimit - INTEGER
  1688. Limit the maximal rates for sending ICMPv6 messages.
  1689. 0 to disable any limiting,
  1690. otherwise the minimal space between responses in milliseconds.
  1691. Default: 1000
  1692. ratemask - list of comma separated ranges
  1693. For ICMPv6 message types matching the ranges in the ratemask, limit
  1694. the sending of the message according to ratelimit parameter.
  1695. The format used for both input and output is a comma separated
  1696. list of ranges (e.g. "0-127,129" for ICMPv6 message type 0 to 127 and
  1697. 129). Writing to the file will clear all previous ranges of ICMPv6
  1698. message types and update the current list with the input.
  1699. Refer to: https://www.iana.org/assignments/icmpv6-parameters/icmpv6-parameters.xhtml
  1700. for numerical values of ICMPv6 message types, e.g. echo request is 128
  1701. and echo reply is 129.
  1702. Default: 0-1,3-127 (rate limit ICMPv6 errors except Packet Too Big)
  1703. echo_ignore_all - BOOLEAN
  1704. If set non-zero, then the kernel will ignore all ICMP ECHO
  1705. requests sent to it over the IPv6 protocol.
  1706. Default: 0
  1707. echo_ignore_multicast - BOOLEAN
  1708. If set non-zero, then the kernel will ignore all ICMP ECHO
  1709. requests sent to it over the IPv6 protocol via multicast.
  1710. Default: 0
  1711. echo_ignore_anycast - BOOLEAN
  1712. If set non-zero, then the kernel will ignore all ICMP ECHO
  1713. requests sent to it over the IPv6 protocol destined to anycast address.
  1714. Default: 0
  1715. xfrm6_gc_thresh - INTEGER
  1716. (Obsolete since linux-4.14)
  1717. The threshold at which we will start garbage collecting for IPv6
  1718. destination cache entries. At twice this value the system will
  1719. refuse new allocations.
  1720. IPv6 Update by:
  1721. Pekka Savola <pekkas@netcore.fi>
  1722. YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
  1723. /proc/sys/net/bridge/* Variables:
  1724. =================================
  1725. bridge-nf-call-arptables - BOOLEAN
  1726. - 1 : pass bridged ARP traffic to arptables' FORWARD chain.
  1727. - 0 : disable this.
  1728. Default: 1
  1729. bridge-nf-call-iptables - BOOLEAN
  1730. - 1 : pass bridged IPv4 traffic to iptables' chains.
  1731. - 0 : disable this.
  1732. Default: 1
  1733. bridge-nf-call-ip6tables - BOOLEAN
  1734. - 1 : pass bridged IPv6 traffic to ip6tables' chains.
  1735. - 0 : disable this.
  1736. Default: 1
  1737. bridge-nf-filter-vlan-tagged - BOOLEAN
  1738. - 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
  1739. - 0 : disable this.
  1740. Default: 0
  1741. bridge-nf-filter-pppoe-tagged - BOOLEAN
  1742. - 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
  1743. - 0 : disable this.
  1744. Default: 0
  1745. bridge-nf-pass-vlan-input-dev - BOOLEAN
  1746. - 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
  1747. interface on the bridge and set the netfilter input device to the
  1748. vlan. This allows use of e.g. "iptables -i br0.1" and makes the
  1749. REDIRECT target work with vlan-on-top-of-bridge interfaces. When no
  1750. matching vlan interface is found, or this switch is off, the input
  1751. device is set to the bridge interface.
  1752. - 0: disable bridge netfilter vlan interface lookup.
  1753. Default: 0
  1754. ``proc/sys/net/sctp/*`` Variables:
  1755. ==================================
  1756. addip_enable - BOOLEAN
  1757. Enable or disable extension of Dynamic Address Reconfiguration
  1758. (ADD-IP) functionality specified in RFC5061. This extension provides
  1759. the ability to dynamically add and remove new addresses for the SCTP
  1760. associations.
  1761. 1: Enable extension.
  1762. 0: Disable extension.
  1763. Default: 0
  1764. pf_enable - INTEGER
  1765. Enable or disable pf (pf is short for potentially failed) state. A value
  1766. of pf_retrans > path_max_retrans also disables pf state. That is, one of
  1767. both pf_enable and pf_retrans > path_max_retrans can disable pf state.
  1768. Since pf_retrans and path_max_retrans can be changed by userspace
  1769. application, sometimes user expects to disable pf state by the value of
  1770. pf_retrans > path_max_retrans, but occasionally the value of pf_retrans
  1771. or path_max_retrans is changed by the user application, this pf state is
  1772. enabled. As such, it is necessary to add this to dynamically enable
  1773. and disable pf state. See:
  1774. https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for
  1775. details.
  1776. 1: Enable pf.
  1777. 0: Disable pf.
  1778. Default: 1
  1779. pf_expose - INTEGER
  1780. Unset or enable/disable pf (pf is short for potentially failed) state
  1781. exposure. Applications can control the exposure of the PF path state
  1782. in the SCTP_PEER_ADDR_CHANGE event and the SCTP_GET_PEER_ADDR_INFO
  1783. sockopt. When it's unset, no SCTP_PEER_ADDR_CHANGE event with
  1784. SCTP_ADDR_PF state will be sent and a SCTP_PF-state transport info
  1785. can be got via SCTP_GET_PEER_ADDR_INFO sockopt; When it's enabled,
  1786. a SCTP_PEER_ADDR_CHANGE event will be sent for a transport becoming
  1787. SCTP_PF state and a SCTP_PF-state transport info can be got via
  1788. SCTP_GET_PEER_ADDR_INFO sockopt; When it's diabled, no
  1789. SCTP_PEER_ADDR_CHANGE event will be sent and it returns -EACCES when
  1790. trying to get a SCTP_PF-state transport info via SCTP_GET_PEER_ADDR_INFO
  1791. sockopt.
  1792. 0: Unset pf state exposure, Compatible with old applications.
  1793. 1: Disable pf state exposure.
  1794. 2: Enable pf state exposure.
  1795. Default: 0
  1796. addip_noauth_enable - BOOLEAN
  1797. Dynamic Address Reconfiguration (ADD-IP) requires the use of
  1798. authentication to protect the operations of adding or removing new
  1799. addresses. This requirement is mandated so that unauthorized hosts
  1800. would not be able to hijack associations. However, older
  1801. implementations may not have implemented this requirement while
  1802. allowing the ADD-IP extension. For reasons of interoperability,
  1803. we provide this variable to control the enforcement of the
  1804. authentication requirement.
  1805. == ===============================================================
  1806. 1 Allow ADD-IP extension to be used without authentication. This
  1807. should only be set in a closed environment for interoperability
  1808. with older implementations.
  1809. 0 Enforce the authentication requirement
  1810. == ===============================================================
  1811. Default: 0
  1812. auth_enable - BOOLEAN
  1813. Enable or disable Authenticated Chunks extension. This extension
  1814. provides the ability to send and receive authenticated chunks and is
  1815. required for secure operation of Dynamic Address Reconfiguration
  1816. (ADD-IP) extension.
  1817. - 1: Enable this extension.
  1818. - 0: Disable this extension.
  1819. Default: 0
  1820. prsctp_enable - BOOLEAN
  1821. Enable or disable the Partial Reliability extension (RFC3758) which
  1822. is used to notify peers that a given DATA should no longer be expected.
  1823. - 1: Enable extension
  1824. - 0: Disable
  1825. Default: 1
  1826. max_burst - INTEGER
  1827. The limit of the number of new packets that can be initially sent. It
  1828. controls how bursty the generated traffic can be.
  1829. Default: 4
  1830. association_max_retrans - INTEGER
  1831. Set the maximum number for retransmissions that an association can
  1832. attempt deciding that the remote end is unreachable. If this value
  1833. is exceeded, the association is terminated.
  1834. Default: 10
  1835. max_init_retransmits - INTEGER
  1836. The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
  1837. that an association will attempt before declaring the destination
  1838. unreachable and terminating.
  1839. Default: 8
  1840. path_max_retrans - INTEGER
  1841. The maximum number of retransmissions that will be attempted on a given
  1842. path. Once this threshold is exceeded, the path is considered
  1843. unreachable, and new traffic will use a different path when the
  1844. association is multihomed.
  1845. Default: 5
  1846. pf_retrans - INTEGER
  1847. The number of retransmissions that will be attempted on a given path
  1848. before traffic is redirected to an alternate transport (should one
  1849. exist). Note this is distinct from path_max_retrans, as a path that
  1850. passes the pf_retrans threshold can still be used. Its only
  1851. deprioritized when a transmission path is selected by the stack. This
  1852. setting is primarily used to enable fast failover mechanisms without
  1853. having to reduce path_max_retrans to a very low value. See:
  1854. http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
  1855. for details. Note also that a value of pf_retrans > path_max_retrans
  1856. disables this feature. Since both pf_retrans and path_max_retrans can
  1857. be changed by userspace application, a variable pf_enable is used to
  1858. disable pf state.
  1859. Default: 0
  1860. ps_retrans - INTEGER
  1861. Primary.Switchover.Max.Retrans (PSMR), it's a tunable parameter coming
  1862. from section-5 "Primary Path Switchover" in rfc7829. The primary path
  1863. will be changed to another active path when the path error counter on
  1864. the old primary path exceeds PSMR, so that "the SCTP sender is allowed
  1865. to continue data transmission on a new working path even when the old
  1866. primary destination address becomes active again". Note this feature
  1867. is disabled by initializing 'ps_retrans' per netns as 0xffff by default,
  1868. and its value can't be less than 'pf_retrans' when changing by sysctl.
  1869. Default: 0xffff
  1870. rto_initial - INTEGER
  1871. The initial round trip timeout value in milliseconds that will be used
  1872. in calculating round trip times. This is the initial time interval
  1873. for retransmissions.
  1874. Default: 3000
  1875. rto_max - INTEGER
  1876. The maximum value (in milliseconds) of the round trip timeout. This
  1877. is the largest time interval that can elapse between retransmissions.
  1878. Default: 60000
  1879. rto_min - INTEGER
  1880. The minimum value (in milliseconds) of the round trip timeout. This
  1881. is the smallest time interval the can elapse between retransmissions.
  1882. Default: 1000
  1883. hb_interval - INTEGER
  1884. The interval (in milliseconds) between HEARTBEAT chunks. These chunks
  1885. are sent at the specified interval on idle paths to probe the state of
  1886. a given path between 2 associations.
  1887. Default: 30000
  1888. sack_timeout - INTEGER
  1889. The amount of time (in milliseconds) that the implementation will wait
  1890. to send a SACK.
  1891. Default: 200
  1892. valid_cookie_life - INTEGER
  1893. The default lifetime of the SCTP cookie (in milliseconds). The cookie
  1894. is used during association establishment.
  1895. Default: 60000
  1896. cookie_preserve_enable - BOOLEAN
  1897. Enable or disable the ability to extend the lifetime of the SCTP cookie
  1898. that is used during the establishment phase of SCTP association
  1899. - 1: Enable cookie lifetime extension.
  1900. - 0: Disable
  1901. Default: 1
  1902. cookie_hmac_alg - STRING
  1903. Select the hmac algorithm used when generating the cookie value sent by
  1904. a listening sctp socket to a connecting client in the INIT-ACK chunk.
  1905. Valid values are:
  1906. * md5
  1907. * sha1
  1908. * none
  1909. Ability to assign md5 or sha1 as the selected alg is predicated on the
  1910. configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
  1911. CONFIG_CRYPTO_SHA1).
  1912. Default: Dependent on configuration. MD5 if available, else SHA1 if
  1913. available, else none.
  1914. rcvbuf_policy - INTEGER
  1915. Determines if the receive buffer is attributed to the socket or to
  1916. association. SCTP supports the capability to create multiple
  1917. associations on a single socket. When using this capability, it is
  1918. possible that a single stalled association that's buffering a lot
  1919. of data may block other associations from delivering their data by
  1920. consuming all of the receive buffer space. To work around this,
  1921. the rcvbuf_policy could be set to attribute the receiver buffer space
  1922. to each association instead of the socket. This prevents the described
  1923. blocking.
  1924. - 1: rcvbuf space is per association
  1925. - 0: rcvbuf space is per socket
  1926. Default: 0
  1927. sndbuf_policy - INTEGER
  1928. Similar to rcvbuf_policy above, this applies to send buffer space.
  1929. - 1: Send buffer is tracked per association
  1930. - 0: Send buffer is tracked per socket.
  1931. Default: 0
  1932. sctp_mem - vector of 3 INTEGERs: min, pressure, max
  1933. Number of pages allowed for queueing by all SCTP sockets.
  1934. min: Below this number of pages SCTP is not bothered about its
  1935. memory appetite. When amount of memory allocated by SCTP exceeds
  1936. this number, SCTP starts to moderate memory usage.
  1937. pressure: This value was introduced to follow format of tcp_mem.
  1938. max: Number of pages allowed for queueing by all SCTP sockets.
  1939. Default is calculated at boot time from amount of available memory.
  1940. sctp_rmem - vector of 3 INTEGERs: min, default, max
  1941. Only the first value ("min") is used, "default" and "max" are
  1942. ignored.
  1943. min: Minimal size of receive buffer used by SCTP socket.
  1944. It is guaranteed to each SCTP socket (but not association) even
  1945. under moderate memory pressure.
  1946. Default: 4K
  1947. sctp_wmem - vector of 3 INTEGERs: min, default, max
  1948. Currently this tunable has no effect.
  1949. addr_scope_policy - INTEGER
  1950. Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
  1951. - 0 - Disable IPv4 address scoping
  1952. - 1 - Enable IPv4 address scoping
  1953. - 2 - Follow draft but allow IPv4 private addresses
  1954. - 3 - Follow draft but allow IPv4 link local addresses
  1955. Default: 1
  1956. ``/proc/sys/net/core/*``
  1957. ========================
  1958. Please see: Documentation/admin-guide/sysctl/net.rst for descriptions of these entries.
  1959. ``/proc/sys/net/unix/*``
  1960. ========================
  1961. max_dgram_qlen - INTEGER
  1962. The maximum length of dgram socket receive queue
  1963. Default: 10