Age | Commit message (Collapse) | Author | Files | Lines |
|
Make sure that we notify the app of the data enqueued in the burst
before notifying of disconnect.
Change-Id: I7747a5cbb4c6bc9132007f849c24ce04b7841273
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Because the code is not optimized, newreno is still the default
congestion control algorithm.
Change-Id: I7061cc80c5a75fa8e8265901fae4ea2888e35173
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
- 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>
|
|
Instead of reusing buffers for acking, consume all buffers and program
output for (dup)ack generation. This implicitly fixes the drop counters
that were artificially inflated by both data and feedback traffic.
Moreover, the patch also significantly reduces the ack traffic as we now
only generate an ack per frame, unless duplicate acks need to be sent.
Because of the reduced feedback traffic, a sender's rx path and a
receiver's tx path are now significantly less loaded. In particular, a
sender can overwhelm a 40Gbps NIC and generate tx drop bursts for low
rtts. Consequently, tx pacing is now enforced by default.
Change-Id: I619c29a8945bf26c093f8f9e197e3c6d5d43868e
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Avoid dequeuing acked bytes more than once per burst for a connection.
Although the fifos do not use locks, size decrements are atomic, so they
rely on locked instructions.
Change-Id: Id65f4ea40b2c10057461402dfd0393034e6472d5
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Also reset pacer on tcp retransmit timeout
Change-Id: I5a9edee4c00d1d169248d79587a9b10437c2bd87
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Also propagate tcp worker context instead of retrieving it multiple
times.
Change-Id: I7b273b981826b37783566d0172a64cd6957f3b33
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Force pacing for fast retransmit to avoid bursts of retransmitted
packets.
Change-Id: I2ff42c328899b36322c4de557b1f7d853dba8fe2
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Adds tx pacing infrastructure for transport protocols that want to use
it. Particularly useful for connections with non-negligible rtt and
constrained network throughput as it avoids large tx bursts that lead to
local interface tx or network drops.
By default the pacer is disabled. To enabled it for tcp, add tx-pacing
to tcp's startup conf. We are still slightly inefficient in the handling
of incoming packets in established state so the pacer slightly affect
maximum throughput in low lacency scenarios.
Change-Id: Id445b2ffcd64cce015f75b773f7d722faa0f7ca9
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
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>
|
|
Change-Id: Idf83fce8ca176e57b323e3741034e3223f1d195a
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: Icff3d688506e7658330db004c58bcfcac273fcec
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
- cache and reuse tcp options and rcv_wnd for session layer tx bursts
- avoid reading/setting total_length_not_including_first_buffer. It's
part of a buffer's second cache line so it comes at a "cost".
Change-Id: Id18219c2f7e07cf4c63ee74f9cdd9e5918904036
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
- sprinkle statics for functions
- move some inlines from header files to corresponding .c files
- replace some always_inlines with statics where inlining is not
performance critical
Change-Id: I371dbf63431ce7e27e4ebbbdd844a9546a1f1849
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Because it avoids pool putting/getting the timer, this function is
somewhat faster than stopping and restarting a timer.
Change-Id: Id99ed9d356b0b1f7e12facfe8da193e1cd30b3ec
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: I28d3c31bdc4255a4ca223d80bcf44709fb39f4ed
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: Idf17a0633a1618b12c22b1119e40c2e9d3192df9
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
- Cleanup session state after last ack and avoid using a cleanup timer.
- Change session cleanup to free the session as opposed to waiting for
delete notify.
- When in close-wait, postpone sending the fin on close until all
outstanding data has been sent.
- Don't flush rx fifo unless in closed state
Change-Id: Ic2a4f0d5568b65c83f4b55b6c469a7b24b947f39
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
- 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>
|
|
Change-Id: I9ede6bc861350c7d9e78fa4d96cd584c2816d06f
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Should be more resilient to ack losses
Change-Id: Icec3b93c1d290dec437fcc4e6fe5171906c9ba8a
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
- compute session type out of transport and network protos
- make session, session lookup and session queue code network protocol
agnostic
This does not update the session layer to support non-ip network layer
protocols
Change-Id: Ifc2f92845e158b649d59462eb7d51c12af536691
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: I67a73e31bda9e497859297fcc1765e880572884a
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
- add debugging scaffolding for simulating buffer shortage
Change-Id: Ice519d74f9c4e4094c4586c548185135b7bb5f2d
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: I44d5c9df7c49b8d4d5677c6d319033b2da3e6b80
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
- add separate TIME_WAIT time constant
- fix output node for TIME_WAIT acks
- ensure snd_nxt is snd_una_max after retransmitting fin
- debugging improvements
Change-Id: Ic947153346979853f2526824b229126e47aead86
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Until now, if the stack didn't find a connection for a packet, it sent
back a reset. With the punt option enabled, packets are now enqueued to
error-punt where they can be handed off to the host os.
Change-Id: I12dea8694b8bd24c92b0d601412928aa7b8046cb
Signed-off-by: Florin Coras <fcoras@cisco.com>
Signed-off-by: Pierre Pfister <ppfister@cisco.com>
|
|
- 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>
|
|
- 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>
|
|
- Remove frame handoff support machinery. We haven't used it in a long
time.
- Configuration support for the local endpoints bihash table
- Drop lookup failure packets in tcp46_syn_sent
Change-Id: Icd51e6785f74661c741e76fac23d21c4cc998d17
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
- set session state as closed on session manager delete
- enable retransmit as opposed to persist timer after persist timer completes
- properly discard buffer chain bytes when new data overlaps ooo
segments
- don't use rxt bytes in snd space estimate used on tx path
Change-Id: Id9cab686e532e5fe70c775d5440260e8eb890a9f
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
- v6 support
- Non-default VRF ID collection
- Break up ip source address list into CLI + API-friendly functions
- Automate proxy arp / proxy nd configuration
- Automate local adjacency insertion
- Binary API support
Change-Id: Iede31184f65cc1ec8c414447d2d60a1334e3fe15
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
- Add SYN_RCVD timeout
- Fix FIN_WAIT_1 to CLOSING transition
Change-Id: I42ca7fc087f6fdfae15bd7a6175dd3226ed341c7
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: I58089d7a9867ede9d8a36b2aea62edef04cb5b81
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
- Fix tcp/udp sw checksum computation
- Fix allocation of multi buffer tcp segments for retransmits
- Send FIN only if/when tx fifo is empty
Change-Id: I2e43a14b87a72c9e547b4339b9a51811cf5732c4
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
- Make tcp output buffer allocation macro an inline function
- Use per ip version per thread tx frames for retransmits and timer
events
- Fix / parameterize tcp data structure preallocation
- Add a couple of gdb-callable show commands
- Fix local endpoint cleanup
Change-Id: I67b47b7570aa14cb4634b6fd93c57cd2eacbfa29
Signed-off-by: Florin Coras <fcoras@cisco.com>
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
- 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>
|
|
- 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>
|
|
- Fix rx sack option parsing
- Add session sack scoreboard tracing and replaying
- Add svm fifo tracing and replaying
- Scoreboard/svm fifo ooo segment reception fixes
- Improved overall debugging
Change-Id: Ieae07eba355e66f5935253232bb00f2dfb7ece00
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
- 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>
|
|
- multiarch on svm fifo
- avoid ip lookup on tx
Change-Id: Iab0d85204a710979417bca1d692cc47877131203
Signed-off-by: Florin Coras <fcoras@cisco.com>
Signed-off-by: Dave Barach <dbarach@cisco.com>
|
|
- 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>
|
|
- 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>
|
|
Also improves builtin client code.
Change-Id: I8bca1aa632028f95c373726efb0abf2ee0eff414
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: Iabeda0d0615b0f6fe20dd00611cb4c594d90b7eb
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Ensure that TCP data plus options does not exceed peer's advertised MSS.
Change-Id: I0de824cb3619346f0394dd694942fc1cf33a82b7
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Allows pure loss recovery retransmits only on timeout.
Change-Id: I563cdbf9e7b890a6569350bdbda4f746ace0544e
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: Ic5467df16e870b49c49678b1dbb40f4a2390b3c9
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: Ib91db6e531231bdc52b0104673a912bee024872f
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: If3c01e318bcb740ca5b240c63f712e2167082a80
Signed-off-by: Dave Barach <dave@barachs.net>
Signed-off-by: Florin Coras <fcoras@cisco.com>
|