Documentation/networking/ip-sysctl.rst

44 	protocols for which pmtu events will be honored are TCP and
48 	TCP path mtu must still work but path MTU information of other
78 	associated with a socket for example, TCP RSTs or ICMP echo replies).
330 TCP variables
336 	See also tcp_max_syn_backlog for additional tuning for TCP sockets.
372 	Enable TCP auto corking :
404 	TCP SYN and SYNACK messages usually advertise an ADVMSS option,
423 	Allows TCP to send "duplicate" SACKs.
445 	Control use of Explicit Congestion Notification (ECN) by TCP.
446 	ECN is used only when both ends of the TCP connection indicate support
450 	feeds back ECN at the TCP layer. The highest variant of ECN feedback
484 	2 Send AccECN option on every packet whenever it fits into TCP
486 	3 Send AccECN option on every packet whenever it fits into TCP
530 	F-RTO is an enhanced recovery algorithm for TCP retransmission
555 	in response to incoming TCP packets that are for an existing
564 	rewrite TCP header fields in manner that causes each endpoint
565 	to think that the other is sending invalid TCP segments, thus
576 	How often TCP sends out keepalive messages when keepalive is enabled.
580 	How many keepalive probes TCP sends out, until it decides that the
608 	Maximal number of TCP sockets not attached to any user file handle,
642 	min: below this number of pages TCP is not bothered about its
645 	pressure: when amount of memory allocated by TCP exceeds this number
646 	of pages, TCP moderates its memory consumption and enters memory
650 	max: number of pages allowed for queueing by all TCP sockets.
667 	If enabled, TCP performs receive buffer auto-tuning, attempting to
689 	Controls TCP Packetization-Layer Path MTU Discovery.  Takes three
697 	Controls how often to start TCP Packetization-Layer Path MTU
702 	Controls when TCP Packetization-Layer Path MTU Discovery probing
707 	By default, TCP saves various connection metrics in the route cache
711 	degradation.  If enabled, TCP will not cache metrics on closing
722 	Controls whether TCP saves ssthresh metrics in the route cache.
733 	This value influences the timeout of a locally closed TCP connection,
763 	for outgoing packets. This allows to have both directions of a TCP
777 	Initial reordering level of packets in a TCP stream.
778 	TCP stack can then dynamically adjust flow reordering level
784 	Maximal reordering level of packets in a TCP stream.
793 	certain TCP stacks.
803 	This value influences the time, after which TCP decides, that
812 	This value influences the timeout of an alive TCP connection,
814 	Given a value of N, a hypothetical TCP connection following
820 	TCP will effectively time out at the first RTO which exceeds the
829 	If enabled, the TCP stack behaves conforming to RFC1337. If unset,
830 	we are not conforming to RFC, but prevent TCP TIME_WAIT
841 	min: Minimal size of receive buffer used by TCP sockets.
842 	It is guaranteed to each TCP socket, even under moderate memory
847 	default: initial size of receive buffer used by TCP sockets.
853 	selected receiver buffers for TCP socket.
878 	TCP tries to reduce number of SACK sent, using a timer based
903 	long latencies at end of a TCP socket syscall.
926 	Use the Host requirements interpretation of the TCP urgent pointer field.
938 	Number of times SYNACKs for a passive TCP connection attempt will
942 	for a passive TCP connection will happen after 63seconds.
958 	syncookies seriously violate TCP protocol, do not allow
959 	to use TCP extensions, can result in serious degradation
1000 	Enable TCP Fast Open (RFC7413) to send and accept data in the opening
1032 	Initial time period in second to disable Fastopen on active TCP sockets
1062 	Number of times initial SYNs for an active TCP connection attempt
1066 	With this the final timeout for an active TCP connection attempt
1082 	Since linux-3.12, TCP does an automatic sizing of TSO frames,
1084 	For specific usages, it's possible to force TCP to build big
1085 	TSO frames. Note that TCP stack might split too big TSO packets
1093 	Starting from linux-5.18, TCP autosizing can be tweaked
1114 	sk->sk_pacing_rate is set by TCP stack using a ratio applied
1116 	If TCP is in slow start, tcp_pacing_ss_ratio is applied
1117 	to let TCP probe for bigger speeds, assuming cwnd can be
1123 	sk->sk_pacing_rate is set by TCP stack using a ratio applied
1125 	If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio
1131 	The number of times for an active TCP connection to retransmit SYNs with
1133 	timeout. This has no effect on SYNACK at the passive TCP side.
1168         This setting contains an assumption about the other TCP timestamp clock
1186 	This changes how the TCP receive window is calculated.
1189 	window can be offered, and that TCP implementations MUST ensure
1203 	min: Amount of memory reserved for send buffers for TCP sockets.
1204 	Each TCP socket has rights to use it due to fact of its birth.
1208 	default: initial size of send buffer used by TCP sockets.  This
1216 	send buffers for TCP sockets. This value does not override
1224 	A TCP socket can control the amount of unsent bytes in its write queue,
1238 	remote TCP is broken and treats the window as a signed quantity.
1239 	If disabled, assume the remote TCP is not broken even if we do
1268 	Controls TCP Small Queue limit per tcp socket.
1269 	TCP bulk sender tends to increase packets in flight until it
1281 	in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
1284 	TCP stack implements per TCP socket limits anyway.
1288 	Show the number of hash buckets for TCP sockets in the current
1295 	Control the number of hash buckets for TCP sockets in the child
1318 	and enables PLB feature, TCP PLB (Protective Load Balancing) is
1321 	upon sensing sustained congestion, TCP triggers a change in
1328 	to apply PLB for IPv4 with other network header fields (e.g. TCP
1368 	having PLB repath onto a connectivity "black hole", after an RTO a TCP
1371 	of multiple TCP connections. This should be set corresponding to the
1398 	requests that must happen before TCP considers that a connection is a
1410 	Minimal TCP retransmission timeout (in microseconds). Note that the
1423 	Maximal TCP retransmission timeout (in ms).
1574 	Defines the local port range that is used by TCP and UDP to
1661 	for established TCP and connected UDP sockets.
1681 	Enable early demux for established TCP sockets.
2398 	- 2: enabled for TCP RESET packets (no active listener)
2643 	associated with a socket for example, TCP RSTs or ICMPv6 echo replies).
3747         Like in TCP, ECN is used only when both ends of the SCTP connection