|
|
01c892df8c
|
tweak finishing timer
|
2024-11-06 10:49:24 +01:00 |
|
|
|
2a0350a564
|
minor tweaks and fixes
especially preventing a stall on some packetloss scenarios
|
2024-10-31 11:39:16 +01:00 |
|
|
|
6f2fa60394
|
handle init2 in ft1 (hacky)
|
2024-10-30 11:25:33 +01:00 |
|
|
|
c7485c4577
|
use sr
|
2024-10-24 14:00:16 +02:00 |
|
|
|
51396314d1
|
big fix for not using the found free slot index!!
minor logging and tweaking changes
|
2024-10-24 11:50:03 +02:00 |
|
|
|
5e884fd3ee
|
dont delay recv_done and check on init info
|
2024-10-23 13:46:00 +02:00 |
|
|
|
c3c2d0f133
|
fixes for 32bit
|
2024-10-20 16:35:55 +02:00 |
|
|
|
1efae931d1
|
better receiving transfer cleanup (reduces log spam)
|
2024-07-15 14:56:52 +02:00 |
|
|
|
6e681aa3fd
|
light cca refator and expose some cca values to the outside
|
2024-07-12 13:14:24 +02:00 |
|
|
|
92373d34f7
|
work around missing contact events (better now)
fix missing ft event on reset (oops)
hard assert sending transfers can not time out higher level
|
2024-07-09 11:00:59 +02:00 |
|
|
|
fab3d42ee9
|
transfer time temporality buffer
|
2024-07-07 15:27:30 +02:00 |
|
|
|
eaaf798661
|
clear receiving transfers
TODO: actually keep around for 2*delay, so missing packets can still be retransmitted
but this fixes perf issues
|
2024-07-07 11:07:31 +02:00 |
|
|
|
3827733f08
|
and remove the old code
|
2024-06-23 12:31:01 +02:00 |
|
|
|
b27107af4c
|
start moving pkg sending to ngcext
wip, but working as far as its implemented
|
2024-06-23 10:14:03 +02:00 |
|
|
|
0eb30246a8
|
small refactor and print in flight packages when timing out
|
2024-05-31 17:03:22 +02:00 |
|
|
|
c52ac19285
|
print window on done
|
2024-05-31 15:36:18 +02:00 |
|
|
|
61b667a4aa
|
reserve memory to reduce number of allocations in hotspots
especially on the sender side
|
2024-03-16 11:30:55 +01:00 |
|
|
|
5fd1f2ab84
|
fix missing virtual destructor and scale tranfer timeout with concurency
|
2024-03-05 16:48:58 +01:00 |
|
|
|
ccf66fb80c
|
update hex conv
|
2024-01-13 22:34:42 +01:00 |
|
|
|
0df0760c06
|
failing to send is now also a congestion event (hacky and only the first time we send data)
|
2024-01-11 00:48:57 +01:00 |
|
|
|
f02b03da7c
|
update to plugin 7 and refactor (should improve speed)
|
2024-01-07 17:23:06 +01:00 |
|
|
|
103f36f2d2
|
update to new ngc_events
|
2023-12-26 21:16:35 +01:00 |
|
|
|
ad918a3253
|
add random cap (1020-1220) and tighten cubic rate limit more
|
2023-12-15 15:31:32 +01:00 |
|
|
|
70cea0d219
|
small fixes
|
2023-12-13 19:38:55 +01:00 |
|
|
|
77f21f01e9
|
extend the protocol to support larger data packets and set it to the new tox constants numbers
|
2023-10-11 03:00:03 +02:00 |
|
|
|
070585ab3d
|
remeber the first sending transfer that could not send any packets and start there next iterate
|
2023-09-01 23:20:03 +02:00 |
|
|
|
ba8befbb2d
|
more fixes
|
2023-09-01 17:34:05 +02:00 |
|
|
|
a1a9bf886a
|
make cubic and flow more resilient
|
2023-09-01 15:51:28 +02:00 |
|
|
|
4ee5dd6ca5
|
cubic almost working, general fixes
|
2023-08-30 13:45:09 +02:00 |
|
|
|
0d49752c3e
|
cubic mostly working (simple), flow rtt seems funky ???
|
2023-08-30 03:03:43 +02:00 |
|
|
|
d957f9496a
|
add and switch to flow only
|
2023-08-29 18:21:12 +02:00 |
|
|
|
89d8d728ff
|
after 2 weeks of porting over the ngc_ft1 code to solanaceae and rewriting the highlevel logic
(29 commits predate this)
|
2023-08-19 22:37:55 +02:00 |
|
