aboutsummaryrefslogtreecommitdiffstats
path: root/src/vnet/tcp/tcp_debug.h
AgeCommit message (Collapse)AuthorFilesLines
2019-08-14tcp: extend protocol configurationFlorin Coras1-2/+2
Type: feature Expose more configuration parameters and refactor some of the existing ones. Change-Id: If44c31ff77ce3d7e8da67d39a4ff61346bdf5ccc Signed-off-by: Florin Coras <fcoras@cisco.com>
2019-04-05tcp: do not delete session on establish popFlorin Coras1-14/+17
Also: - force reset if wait close pops in fin-wait-1 with unsent data - adds more event logging. Change-Id: I4ddada046214fa71e17514cdec57b3026f84a283 Signed-off-by: Florin Coras <fcoras@cisco.com>
2019-04-04tcp: shorten wait in fin-wait-1 with fin rcvdFlorin Coras1-4/+5
Change-Id: Ifddc32ab3da0e691ac3df74ff26e19f6fa00fef7 Signed-off-by: Florin Coras <fcoras@cisco.com>
2019-03-20tcp: disable cc elog debuggingFlorin Coras1-2/+2
Change-Id: Iffed748a15b9f01b985f9a6a9574a7bc42ab55aa Signed-off-by: Florin Coras <fcoras@cisco.com>
2019-02-22tcp: send enough dupacks to cover all sack holesFlorin Coras1-2/+2
Make sure we send enough dupacks to cover all the holes created in the last frame received. Also make sure we send all the blocks, not just the first. Change-Id: I9597a34ac14473d1cc3ad07d65bc37043e3d0582 Signed-off-by: Florin Coras <fcoras@cisco.com>
2019-01-20tcp: fix debugging without cc stats (VPP-1547)Florin Coras1-0/+1
Change-Id: I376856fcadce570bb555064435f5876eefb2befa Signed-off-by: Florin Coras <fcoras@cisco.com>
2018-12-26tcp: separate active and passive establish timersFlorin Coras1-41/+44
Change-Id: Ia2241e963cf45765d8d17c65eea781edbf74d4f9 Signed-off-by: Florin Coras <fcoras@cisco.com>
2018-12-24tcp: compute seq_end in tcp_inputFlorin Coras1-0/+2
syn/fin are no longer added to seq_end so they must be considered individually in each state. Change-Id: I5e3047194101c4fca2db9f9ad29a4a6468c397ab Signed-off-by: Florin Coras <fcoras@cisco.com>
2018-11-08tcp: pacer and mrtt estimation improvementsFlorin Coras1-4/+5
- update pacer once per burst - better estimate initial rtt - compute smoothed average for higher precision rtt estimate Change-Id: I06d41a98784cdf861bedfbee2e7d0afc0d0154ef Signed-off-by: Florin Coras <fcoras@cisco.com>
2018-11-01tcp: fast retransmit pacingFlorin Coras1-5/+13
Force pacing for fast retransmit to avoid bursts of retransmitted packets. Change-Id: I2ff42c328899b36322c4de557b1f7d853dba8fe2 Signed-off-by: Florin Coras <fcoras@cisco.com>
2018-10-23tcp: fast retransmit improvementsFlorin Coras1-12/+35
Patch is too large to be ported to 18.10 just days before release. - handle fast retransmits outside of established node and limit the retransmit burst size to avoid tx losses and worsening congestion. - in the absance of a tx pacer, use slow start after fast retransmit exists - add fast retransmit heuristic that re-retries sending the first segment if everything else fails - fine tuning Change-Id: I84a2ab8fbba8b97f1d2b26584dc11a1e2c33c8d2 Signed-off-by: Florin Coras <fcoras@cisco.com>
2018-09-29session: lock msg queues while generating control eventsFlorin Coras1-1/+4
Change-Id: I229ce9c306a97956c87ccb8f7a1c5c9dac94ab31 Signed-off-by: Florin Coras <fcoras@cisco.com>
2018-05-26tcp: loss recovery improvements/fixesFlorin Coras1-102/+172
- fix newreno cwnd computation - reset snd_una_max on entering recovery - accept acks beyond snd_nxt but less than snd_congestion when in recovery - avoid entering fast recovery multiple times when using sacks - avoid as much as possible sending small segments when doing fast retransmit - more event logging Change-Id: I19dd151d7704e39d4eae06de3a26f5e124875366 Signed-off-by: Florin Coras <fcoras@cisco.com>
2018-05-23tcp: cc improvements and fixesFlorin Coras1-1/+1
Change-Id: I6615bb612bcc3f795b5f822ea55209bb30ef35b5 Signed-off-by: Florin Coras <fcoras@cisco.com>
2018-04-30tcp/session: debug improvements/fixesFlorin Coras1-10/+9
Change-Id: I906e58b4f9827a79a6ab673f8fa2e03036c69820 Signed-off-by: Florin Coras <fcoras@cisco.com>
2018-04-20tcp: make newreno byte instead of acks dependentFlorin Coras1-1/+18
Should be more resilient to ack losses Change-Id: Icec3b93c1d290dec437fcc4e6fe5171906c9ba8a Signed-off-by: Florin Coras <fcoras@cisco.com>
2018-01-27Fix session/tcp coverity warningsFlorin Coras1-8/+11
Change-Id: I5c404eacb4a6c1e16485a6656168d9171ff49a8b Signed-off-by: Florin Coras <fcoras@cisco.com>
2018-01-27session: disconnect and echo client improvementsFlorin Coras1-1/+1
Change-Id: If421bad17b6cfe8e321257c93bb38931e37d7b64 Signed-off-by: Florin Coras <fcoras@cisco.com>
2018-01-25session: add support for memfd segmentsFlorin Coras1-1/+1
- update segment manager and session api to work with both flavors of ssvm segments - added generic ssvm slave/master init and del functions - cleanup/refactor tcp_echo - fixed uses of svm fifo pool as vector Change-Id: Ieee8b163faa407da6e77e657a2322de213a9d2a0 Signed-off-by: Florin Coras <fcoras@cisco.com>
2017-11-28tcp: fix retransmissions under buffer shortageFlorin Coras1-0/+34
- add debugging scaffolding for simulating buffer shortage Change-Id: Ice519d74f9c4e4094c4586c548185135b7bb5f2d Signed-off-by: Florin Coras <fcoras@cisco.com>
2017-09-25tcp: do not sample rtt for retransmitted segmentsFlorin Coras1-54/+54
Change-Id: I365c31607332a944ef498369881332b515894ed7 Signed-off-by: Florin Coras <fcoras@cisco.com>
2017-09-19session/tcp: improve preallocated segment handlingFlorin Coras1-7/+2
- add preallocated segment flag - don't remove pre-allocated segments except if application detaches - when preallocating fifos in multiple segments, completely fill a segment before moving to the next - detach server application from segment-managers when deleting app - batch syn/syn-ack/fin (re)transmissions - loosen up close-wait and time-wait times Change-Id: I412f53ce601cc83b3acc26aeffd7fa2d52d73b03 Signed-off-by: Florin Coras <fcoras@cisco.com>
2017-09-12tcp: horizontal scaling improvmentsFlorin Coras1-18/+31
- do not scale syn-ack window - fix the max number of outstanding syns in builtin client - fix syn-sent ack validation to use modulo arithmetic - improve retransmit timer handler - fix output buffer allocator leakeage - improved debugging Change-Id: Iac3bc0eadf7d0b494a93e22d210a3153b61b3273 Signed-off-by: Florin Coras <fcoras@cisco.com>
2017-07-30Make tcp active open data structures thread safeFlorin Coras1-16/+16
- Cleanup half-open connections and timers on the right thread - Ensure half-open connection and transport endpoint pools are thread safe - Enqueue TX events to the correct vpp thread in the builtin client - Use transport proto in transport connections instead of session type Change-Id: Id13239a206afbff6f34a38afa510fe014e4b2049 Signed-off-by: Florin Coras <fcoras@cisco.com> Signed-off-by: Dave Barach <dave@barachs.net>
2017-07-21Improvements to tcp rx path and debuggingFlorin Coras1-61/+185
- Increment rcv_nxt for fin packets - Call tcp_segment_rcv only if buffer has data - Parse rcv opts before deleting half-open connection - Fix initial rcv_wnd - Improved event logging Change-Id: I9b83c04f432c4cec832c480b03e534deff02c3b1 Signed-off-by: Florin Coras <fcoras@cisco.com>
2017-07-11Horizontal (nSessions) scaling draftDave Barach1-3/+10
- Data structure preallocation. - Input state machine fixes for mid-stream 3-way handshake retries. - Batch connections in the builtin_client - Multiple private fifo segment support - Fix elog simultaneous event type registration - Fix sacks when segment hole is added after highest sacked - Add "accepting" session state for sessions pending accept - Add ssvm non-recursive locking - Estimate RTT for syn-ack - Don't init fifo pointers. We're using relative offsets for ooo segments - CLI to dump individual session Change-Id: Ie0598563fd246537bafba4feed7985478ea1d415 Signed-off-by: Dave Barach <dbarach@cisco.com> Signed-off-by: Florin Coras <fcoras@cisco.com>
2017-06-19Overall tcp performance improvements (VPP-846)Florin Coras1-53/+133
- limit minimum rto per connection - cleanup sack scoreboard - switched svm fifo out-of-order data handling from absolute offsets to relative offsets. - improve cwnd handling when using sacks - add cc event debug stats - improved uri tcp test client/server: bugfixes and added half-duplex mode - expanded builtin client/server - updated uri socket client/server code to work in half-duplex - ensure session node unsets fifo event for empty fifo - fix session detach Change-Id: Ia446972340e32a65e0694ee2844355167d0c170d Signed-off-by: Florin Coras <fcoras@cisco.com>
2017-06-09Implement sack based tcp loss recovery (RFC 6675)Florin Coras1-8/+8
- refactor existing congestion control code (RFC 6582/5681). Handling of ack feedback now consists of: ack parsing, cc event detection, event handling, congestion control update - extend sack scoreboard to support sack based retransmissions - basic implementation of Eifel detection algorithm (RFC 3522) for detecting spurious retransmissions - actually initialize the per-thread frame freelist hash tables - increase worker stack size to 2mb - fix session queue node out-of-buffer handling - ensure that the local buffer cache vec_len matches reality - avoid 2x spurious event requeues when short of buffers - count out-of-buffer events - make the builtin server thread-safe - fix bihash template threading issue: need to paint -1 across uninitialized working_copy_length vector elements (via rebase from master) Change-Id: I646cb9f1add9a67d08f4a87badbcb117980ebfc4 Signed-off-by: Florin Coras <fcoras@cisco.com> Signed-off-by: Dave Barach <dbarach@cisco.com>
2017-05-01TCP ooo reception fixesFlorin Coras1-0/+15
- Improve svm fifo handling of out-of-order segments - Ensure tsval_recent is updated only if rcv_las falls withing the segments's sequence space - Avoid directly dropping old ACKs - Improve debugging Change-Id: I88dbe2394a0ad7eb389a4cc12d013a13733953aa Signed-off-by: Florin Coras <fcoras@cisco.com>
2017-04-27[VPP-710] tcp TCP_TIMER_ESTABLISH timeout_handler occur segment faultflyingeagle231-1/+2
Change-Id: Ie43b8f1c91600d44dff67e9348fb2fe83efc60c2 Signed-off-by: flyingeagle23 <wang.hui56@zte.com.cn>
2017-04-06Use thread local storage for thread indexDamjan Marion1-1/+1
This patch deprecates stack-based thread identification, Also removes requirement that thread stacks are adjacent. Finally, possibly annoying for some folks, it renames all occurences of cpu_index and cpu_number with thread index. Using word "cpu" is misleading here as thread can be migrated ti different CPU, and also it is not related to linux cpu index. Change-Id: I68cdaf661e701d2336fc953dcb9978d10a70f7c1 Signed-off-by: Damjan Marion <damarion@cisco.com>
2017-04-02TCP cc/window management fixes and debuggingFlorin Coras1-16/+50
- added persist timer - update rcv_las whenever sending an ack - moved fifo size to its own cache line - improved session and builtin client debugging Change-Id: Ia649cf942cf0c061a713e8b67f0eb6974a6cd55b Signed-off-by: Florin Coras <fcoras@cisco.com> Signed-off-by: Dave Barach <dave@barachs.net>
2017-03-27TCP/session improvementsFlorin Coras1-38/+214
- Added svm fifo flag for tracking fifo dequeue events (replaces event length). Updated all code to switch to the new scheme. - More session debugging - Fix peek index wrap - Add a trivial socket test client - Fast retransmit/cc fixes - tx and rx SACK fixes and unit testing - SRTT computation fix - remove dupack/ack burst filters - improve ack rx - improved segment rx - builtin client test code Change-Id: Ic4eb2d5ca446eb2260ccd3ccbcdaa73c64e7f4e1 Signed-off-by: Florin Coras <fcoras@cisco.com> Signed-off-by: Dave Barach <dbarach@cisco.com>
2017-03-13VPP-659 Improve tcp/session debugging and testingFlorin Coras1-0/+316
- event-logging support for tcp and session layer - improvements to uri test code - builtin_server on port 1234 - use the CLOSEWAIT timer when we rx FIN in FIN_WAIT_2 state Change-Id: Ibc445f164b2086b20323bf89c77cffd3059f570f Signed-off-by: Florin Coras <fcoras@cisco.com> Signed-off-by: Dave Barach <dbarach@cisco.com> Signed-off-by: Dave Barach <dave@barachs.net>
ors * sizeof (counters[0])); vec_validate (vm->error_elog_event_types, l - 1); /* Switch to the stats segment ... */ oldheap = vlib_stats_push_heap (0); /* Allocate a counter/elog type for each error. */ vec_validate (em->counters, l - 1); /* Zero counters for re-registrations of errors. */ if (n->error_heap_index + n_errors <= vec_len (em->counters_last_clear)) clib_memcpy (em->counters + n->error_heap_index, em->counters_last_clear + n->error_heap_index, n_errors * sizeof (em->counters[0])); else clib_memset (em->counters + n->error_heap_index, 0, n_errors * sizeof (em->counters[0])); /* Register counter indices in the stat segment directory */ { int i; u8 *error_name = 0; for (i = 0; i < n_errors; i++) { vec_reset_length (error_name); error_name = format (error_name, "/err/%v/%s%c", n->name, counters[i].name, 0); vlib_stats_register_error_index (oldheap, error_name, em->counters, n->error_heap_index + i); } vec_free (error_name); } /* (re)register the em->counters base address, switch back to main heap */ vlib_stats_pop_heap2 (em->counters, vm->thread_index, oldheap, 1); { elog_event_type_t t; uword i; clib_memset (&t, 0, sizeof (t)); if (n_errors > 0) vec_validate (nm->node_by_error, n->error_heap_index + n_errors - 1); for (i = 0; i < n_errors; i++) { t.format = (char *) format (0, "%v %s: %%d", n->name, counters[i].name); vm->error_elog_event_types[n->error_heap_index + i] = t; nm->node_by_error[n->error_heap_index + i] = n->index; } } } uword unformat_vlib_error (unformat_input_t *input, va_list *args) { vlib_main_t *vm = va_arg (*args, vlib_main_t *); const vlib_error_main_t *em = &vm->error_main; vlib_error_t *error_index = va_arg (*args, vlib_error_t *); const vlib_node_t *node; char *error_name; u32 node_index; vlib_error_t i; if (!unformat (input, "%U.%s", unformat_vlib_node, vm, &node_index, &error_name)) return 0; node = vlib_get_node (vm, node_index); for (i = 0; i < node->n_errors; i++) { vlib_error_t ei = node->error_heap_index + i; if (strcmp (em->counters_heap[ei].name, error_name) == 0) { *error_index = ei; vec_free (error_name); return 1; } } vec_free (error_name); return 0; } static char * sev2str (enum vl_counter_severity_e s) { switch (s) { case VL_COUNTER_SEVERITY_ERROR: return "error"; case VL_COUNTER_SEVERITY_WARN: return "warn"; case VL_COUNTER_SEVERITY_INFO: return "info"; default: return "unknown"; } } static clib_error_t * show_errors (vlib_main_t * vm, unformat_input_t * input, vlib_cli_command_t * cmd) { vlib_error_main_t *em = &vm->error_main; vlib_node_t *n; u32 code, i, ni; u64 c; int index = 0; int verbose = 0; u64 *sums = 0; if (unformat (input, "verbose %d", &verbose)) ; else if (unformat (input, "verbose")) verbose = 1; vec_validate (sums, vec_len (em->counters)); if (verbose) vlib_cli_output (vm, "%=10s%=35s%=35s%=10s%=6s", "Count", "Node", "Reason", "Severity", "Index"); else vlib_cli_output (vm, "%=10s%=35s%=35s%=10s", "Count", "Node", "Reason", "Severity"); foreach_vlib_main () { em = &this_vlib_main->error_main; if (verbose) vlib_cli_output (vm, "Thread %u (%v):", index, vlib_worker_threads[index].name); for (ni = 0; ni < vec_len (this_vlib_main->node_main.nodes); ni++) { n = vlib_get_node (this_vlib_main, ni); for (code = 0; code < n->n_errors; code++) { i = n->error_heap_index + code; c = em->counters[i]; if (i < vec_len (em->counters_last_clear)) c -= em->counters_last_clear[i]; sums[i] += c; if (c == 0 && verbose < 2) continue; if (verbose) vlib_cli_output (vm, "%10lu%=35v%=35s%=10s%=6d", c, n->name, em->counters_heap[i].desc, sev2str (em->counters_heap[i].severity), i); else vlib_cli_output (vm, "%10lu%=35v%=35s%=10s", c, n->name, em->counters_heap[i].desc, sev2str (em->counters_heap[i].severity)); } } index++; } if (verbose) vlib_cli_output (vm, "Total:"); for (ni = 0; ni < vec_len (vm->node_main.nodes); ni++) { n = vlib_get_node (vm, ni); for (code = 0; code < n->n_errors; code++) { i = n->error_heap_index + code; if (sums[i]) { if (verbose) vlib_cli_output (vm, "%10lu%=40v%=20s%=10d", sums[i], n->name, em->counters_heap[i].desc, i); } } } vec_free (sums); return 0; } /* *INDENT-OFF* */ VLIB_CLI_COMMAND (vlib_cli_show_errors) = { .path = "show errors", .short_help = "Show error counts", .function = show_errors, }; /* *INDENT-ON* */ /* *INDENT-OFF* */ VLIB_CLI_COMMAND (cli_show_node_counters, static) = { .path = "show node counters", .short_help = "Show node counters", .function = show_errors, }; /* *INDENT-ON* */ static clib_error_t * clear_error_counters (vlib_main_t * vm, unformat_input_t * input, vlib_cli_command_t * cmd) { vlib_error_main_t *em; u32 i; foreach_vlib_main () { em = &this_vlib_main->error_main; vec_validate (em->counters_last_clear, vec_len (em->counters) - 1); for (i = 0; i < vec_len (em->counters); i++) em->counters_last_clear[i] = em->counters[i]; } return 0; } /* *INDENT-OFF* */ VLIB_CLI_COMMAND (cli_clear_error_counters, static) = { .path = "clear errors", .short_help = "Clear error counters", .function = clear_error_counters, }; /* *INDENT-ON* */ /* *INDENT-OFF* */ VLIB_CLI_COMMAND (cli_clear_node_counters, static) = { .path = "clear node counters", .short_help = "Clear node counters", .function = clear_error_counters, }; /* *INDENT-ON* */ /* * fd.io coding-style-patch-verification: ON * * Local Variables: * eval: (c-set-style "gnu") * End: */