| /linux-6.14.4/net/ipv4/ |
| D | Kconfig | 470 bool "TCP: advanced congestion control"
472 Support for selection of various TCP congestion control
483 tristate "Binary Increase Congestion (BIC) control"
490 congestion window is large, additive increase with a large
492 scalability. Under small congestion windows, binary search
509 protocol stack that optimizes the performance of TCP congestion
511 congestion window and slow start threshold after a congestion
513 slow start threshold and a congestion window which takes into
514 account the bandwidth used at the time congestion is experienced.
524 congestion control for high speed network links. It uses a
[all …]
|
| D | tcp_nv.c | 3 * TCP NV: TCP with Congestion Avoidance
7 * Like TCP-Vegas, TCP-NV supports true congestion avoidance,
8 * the ability to detect congestion before packet losses occur.
9 * When congestion (queue buildup) starts to occur, TCP-NV
17 * detect congestion (congestion control) and those that use queue
18 * buildup to detect congestion (congestion avoidance).
25 * 1) Add mechanism to deal with reverse congestion.
39 * nv_cong_dec_mult Decrease cwnd by X% (30%) of congestion when detected
40 * nv_ssthresh_factor On congestion set ssthresh to this * <desired cwnd> / 8
46 * slow-start due to congestion
[all …]
|
| D | tcp_cong.c | 3 * Pluggable TCP congestion control support and newReno
4 * congestion control.
89 /* Attach new congestion control algorithm to the list
118 * Remove congestion control algorithm, called from
159 pr_notice("invalid old congestion control algorithm to replace\n"); in tcp_update_congestion_control()
214 /* Assign choice of congestion control. */
279 /* Used by sysctl to change default congestion control */
317 /* Build string with list of available congestion control values */
335 /* Get current default congestion control */
346 /* Built list of non-restricted congestion control values */
[all …]
|
| D | tcp_plb.c | 4 * switches. PLB is a host-based optimization; it leverages congestion
6 * connection experiencing sustained congestion. PLB prefers to repath
16 * "PLB: Congestion Signals are Simple and Effective for
43 /* Check whether recent congestion has been persistent enough to warrant
105 * congestion characteristics. in tcp_plb_update_state_upon_rto()
|
| D | tcp_bic.c | 3 * Binary Increase Congestion control for TCP
8 * "Binary Increase Congestion Control for Fast, Long Distance
13 * Unless BIC is enabled and congestion window is large
41 MODULE_PARM_DESC(low_window, "lower bound on congestion window (for TCP friendliness)");
81 * Compute congestion window to use.
159 * then increase congestion window slowly
|
| D | tcp_vegas.c | 3 * TCP Vegas congestion control
5 * This is based on the congestion detection/avoidance scheme described in
7 * "TCP Vegas: End to end congestion avoidance on a global internet."
65 * stale info -- both the saved cwnd and congestion feedback are
209 * of delayed ACKs, at the cost of noticing congestion in tcp_vegas_cong_avoid()
231 * and switch to congestion avoidance. in tcp_vegas_cong_avoid()
249 /* Congestion avoidance. */ in tcp_vegas_cong_avoid()
|
| D | tcp_lp.c | 9 * As of 2.6.13, Linux supports pluggable congestion control algorithms.
18 * o Handle reaction for Early Congestion Indication (ECI) within
92 * @sk: socket to initialize congestion control algorithm for
263 * @sk: socket requiring congestion avoidance calculations
266 * Deal with active drop under Early Congestion Indication.
|
| /linux-6.14.4/drivers/crypto/caam/ |
| D | dpseci.h | 36 * Enable the Congestion Group support
257 * enum dpseci_congestion_unit - DPSECI congestion units
268 * congestion state (see 'threshold_entry')
274 * congestion state (see 'threshold_exit')
286 * DPIO/DPCON's WQ channel once entering a congestion state
293 * DPIO/DPCON's WQ channel once exiting a congestion state
306 * struct dpseci_congestion_notification_cfg - congestion notification
309 * @threshold_entry: above this threshold we enter a congestion state.
311 * @threshold_exit: below this threshold we exit the congestion state.
|
| D | debugfs.c | 27 * This is a counter for the number of times the congestion group (where all
28 * the request and response queueus are) reached congestion. Incremented
29 * each time the congestion callback is called with congested == true.
|
| /linux-6.14.4/Documentation/networking/ |
| D | rds.rst | 90 mark the port as "congested" and send a congestion update to
118 a congestion update is sent to all peers.
122 notification when a congestion update arrived, or when a RDMA
144 this - by using congestion notifications, and by checking for
178 Congestion Notifications
207 CONG_BITMAP this is a congestion update bitmap
248 Congestion
271 communication of per-port congestion. The sender does a very
282 The general layer handles the socket API, congestion handling,
318 wraps the raw congestion bitmap, contains rbnode, waitq, etc.
[all …]
|
| /linux-6.14.4/drivers/infiniband/hw/qib/ |
| D | qib_mad.h | 97 * Congestion control class attributes
139 * Congestion Control class portinfo capability mask bits
166 /* Congestion control traps */
181 /* Congestion info flags */
222 /* Sixteen congestion entries */
257 /* 64 Congestion Control table entries in a single MAD */
|
| /linux-6.14.4/net/dccp/ |
| D | Kconfig | 6 Datagram Congestion Control Protocol (RFC 4340)
10 The Datagram Congestion Control Protocol (DCCP) is a transport
12 congestion-controlled, unreliable datagrams. It should be suitable
|
| /linux-6.14.4/net/rds/ |
| D | cong.c | 42 * This file implements the receive side of the unconventional congestion
64 * finer-grained communication of per-port congestion. The sender does a very
76 * We maintain a global counter that is incremented each time a congestion map
84 * Congestion monitoring
304 rdsdebug("setting congestion for %pI4:%u in map %p\n", in rds_cong_set_bit()
318 rdsdebug("clearing congestion for %pI4:%u in map %p\n", in rds_cong_clear_bit()
357 /* update congestion map for now-closed port */ in rds_cong_remove_socket()
383 /* Test again - a congestion update may have arrived in in rds_cong_wait()
416 * Allocate a RDS message containing a congestion update.
|
| /linux-6.14.4/samples/bpf/ |
| D | hbm_out_kern.c | 31 * to reduce the congestion window. The current implementation uses a linear
122 // Set flags (drop, congestion, cwr) in _hbm_out_cg()
123 // Dropping => we are congested, so ignore congestion flag in _hbm_out_cg()
133 // Congested, set congestion flag in _hbm_out_cg()
153 // Do congestion control in _hbm_out_cg()
|
| /linux-6.14.4/include/uapi/linux/ |
| D | rds.h | 295 * Congestion monitoring.
296 * Congestion control in RDS happens at the host connection
299 * up when the congestion map is updated.
304 * When a congestion update arrives, RDS checks the set of
309 * To install the congestion monitor bitmask, use RDS_CONG_MONITOR
311 * Congestion updates are received via RDS_CMSG_CONG_UPDATE
|
| /linux-6.14.4/drivers/net/ethernet/freescale/dpaa/ |
| D | dpaa_eth.h | 88 u64 cg_tdrop; /* Congestion group taildrop */
89 u64 wred; /* WRED congestion */
153 * (and the same) congestion group.
161 * entered congestion state
|
| /linux-6.14.4/drivers/net/ethernet/freescale/dpaa2/ |
| D | dpni.h | 62 * DPNI_OPT_SHARED_CONGESTION - Congestion can be managed in several ways,
64 * use congestion groups for sets of queues. If set, it configures a single
65 * congestion groups across all TCs. If reset, a congestion group is allocated
440 * @page_4: Page_4 statistics structure: congestion points
441 * @page_4.cgr_reject_frames: number of rejected frames due to congestion point
442 * @page_4.cgr_reject_bytes: number of rejected bytes due to congestion point
853 * enum dpni_congestion_unit - DPNI congestion units
863 * enum dpni_congestion_point - Structure representing congestion point
890 /* DPNI congestion options */
893 * DPNI_CONG_OPT_FLOW_CONTROL - This congestion will trigger flow control or
[all …]
|
| /linux-6.14.4/include/soc/fsl/ |
| D | qman.h | 47 #define QM_PIRQ_CSCI 0x00100000 /* Congestion State Change */
83 __be32 cfg; /* format, offset, length / congestion */
106 * interpreted purely for congestion calculations, ie. a "congestion
137 * If 'format' is _compound, 29b "congestion weight".
502 #define QM_FQCTRL_CGE 0x0400 /* Congestion Group Enable */
521 #define QM_OAC_CG 0x1 /* Accounting for Congestion Groups */
538 * how we represent it here. The congestion state threshold is calculated from
661 * This object type represents a QMan congestion group, it is defined further
776 * This callback type is used when handling congestion group entry/exit.
777 * 'congested' is non-zero on congestion-entry, and zero on congestion-exit.
[all …]
|
| /linux-6.14.4/tools/include/uapi/linux/ |
| D | tcp.h | 104 #define TCP_CONGESTION 13 /* Congestion control algorithm */
116 #define TCP_CC_INFO 26 /* Get Congestion Control (optional) info */
175 * Sender's congestion state indicating normal or abnormal situations
195 * The sender enters Congestion Window Reduction (CWR) state when it
196 * has received ACKs with ECN-ECE marks, or has experienced congestion
299 TCP_NLA_SND_CWND, /* Sending congestion window */
|
| /linux-6.14.4/Documentation/networking/devlink/ |
| D | hns3.rst | 28 - Used to represent the Soft Congestion Control (SSC) firmware version.
29 SCC is a firmware component which provides multiple RDMA congestion
|
| /linux-6.14.4/drivers/soc/fsl/qbman/ |
| D | qman_priv.h | 58 /* "Query Congestion Group State" */
91 /* Congestion Groups */
94 * This wrapper represents a bit-array for the state of the 256 QMan congestion
95 * groups. Is also used as a *mask* for congestion groups, eg. so we ignore
97 * already used in the management command to query congestion groups.
|
| /linux-6.14.4/net/sctp/ |
| D | transport.c | 579 * cwnd only if the current congestion window is being fully in sctp_transport_raise_cwnd()
594 * if, and only if the congestion window is being fully in sctp_transport_raise_cwnd()
641 pr_debug("%s: congestion avoidance: transport:%p, " in sctp_transport_raise_cwnd()
652 /* This routine is used to lower the transport's cwnd when congestion is
705 * then the sender knows that congestion was encountered in the in sctp_transport_lower_cwnd()
707 * indication of congestion should be treated just as a in sctp_transport_lower_cwnd()
708 * congestion loss in non-ECN Capable TCP. That is, the TCP in sctp_transport_lower_cwnd()
709 * source halves the congestion window "cwnd" and reduces the in sctp_transport_lower_cwnd()
712 * congestion indications more than once every window of in sctp_transport_lower_cwnd()
747 /* Apply Max.Burst limit to the congestion window:
[all …]
|
| /linux-6.14.4/tools/testing/selftests/net/netfilter/packetdrill/ |
| D | common.sh | 16 # TCP congestion control.
22 # TCP Explicit Congestion Notification (ECN)
|
| /linux-6.14.4/net/dccp/ccids/ |
| D | ccid2.c | 98 * need to check that the ack ratio is still less than the congestion in ccid2_check_l_ack_ratio()
99 * window. Otherwise, we will send an entire congestion window of in ccid2_check_l_ack_ratio()
103 * the congestion window (or 1 if that's zero) instead of to the in ccid2_check_l_ack_ratio()
104 * congestion window. This prevents problems if one ack is lost. in ccid2_check_l_ack_ratio()
176 * Congestion window validation (RFC 2861).
180 MODULE_PARM_DESC(ccid2_do_cwv, "Perform RFC2861 Congestion Window Validation");
390 * (see P. Sarolahti, A. Kuznetsov,"Congestion Control in ccid2_rtt_estimator()
523 /* check reverse path congestion */ in ccid2_hc_tx_packet_recv()
548 * FIXME: Ack Congestion Control is broken; in in ccid2_hc_tx_packet_recv()
560 /* check forward path congestion */ in ccid2_hc_tx_packet_recv()
[all …]
|
| D | Kconfig | 19 rate-controlled congestion control mechanism. TFRC is designed to
31 The TFRC congestion control algorithms were initially described in
|