| #
32788 |
| 07-Dec-1987 |
bostic |
use Berkeley specific header
|
| #
32785 |
| 07-Dec-1987 |
karels |
synch with van, simplify negative-length (shrinking window) code
|
| #
32373 |
| 09-Oct-1987 |
karels |
don't send options if state is closed (from tcp_drop)
|
| #
32034 |
| 10-Aug-1987 |
karels |
Phil Karn's retransmit algorithm, using t_rxtcur for current rexmt value;
slow start from beginning if connection not local; get rtt estimate on SYN
|
| #
31758 |
| 03-Jul-1987 |
karels |
oops, don't time the acks
|
| #
31726 |
| 30-Jun-1987 |
karels |
redo retransmit calculations as srtt + 2*smoothed variance
(from Van Jacobson)
|
| #
31725 |
| 30-Jun-1987 |
karels |
silly-window: send full-sized packets only, or we start sending
funny sizes; don't decrement seq for FIN when retransmitting; rearrange code
to set rtseq properly on SYN (was one too high); don't cle
silly-window: send full-sized packets only, or we start sending
funny sizes; don't decrement seq for FIN when retransmitting; rearrange code
to set rtseq properly on SYN (was one too high); don't clear rxtshft
when persist tops out
show more ...
|
| #
31442 |
| 06-Jun-1987 |
karels |
slow-start changes from van
|
| #
30686 |
| 28-Mar-1987 |
karels |
branch to get fix from 7.4 applied to 7.2
|
| #
30657 |
| 20-Mar-1987 |
karels |
don't slide data back if resending data+FIN
|
| #
30526 |
| 19-Feb-1987 |
karels |
new stats
|
| #
29795 |
| 20-Aug-1986 |
karels |
don't send premature FIN
|
| #
29150 |
| 05-Jun-1986 |
mckusick |
4.3BSD release version
|
| #
27067 |
| 13-Apr-1986 |
karels |
more correct fix for TCP FIN seq problem: need to predecrement send_nxt
when resending FIN (for ack, etc.) but not retransmitting
|
| #
27048 |
| 12-Apr-1986 |
karels |
conservative fix for increasing FIN
|
| #
26834 |
| 12-Mar-1986 |
karels |
change fix to handle shrinking windows (previous was too far down);
reorder for convenience and put common case earliest; fix comments
(STILL CONTAINS DEBUG CODE logging persists at offset)
|
| #
26443 |
| 03-Mar-1986 |
karels |
one more time: don't clear persist on an ack-only packet
|
| #
26387 |
| 23-Feb-1986 |
karels |
lint
|
| #
26059 |
| 03-Feb-1986 |
karels |
consolidate route caching, manipulation in IP
|
| #
25940 |
| 22-Jan-1986 |
karels |
don't delay send if we have more than one segment to send,
but last is fractional; count the registers and reorder the tests;
more window anti-shrink code; sundry cleanups
|
| #
25261 |
| 23-Oct-1985 |
karels |
maintain estimates of peer's send and receive buffer sizes;
don't wait for bigger window if he hasn't got one
|
| #
24817 |
| 16-Sep-1985 |
karels |
ip options move to inpcb
|
| #
23191 |
| 08-Jun-1985 |
mckusick |
Add copyright
|
| #
21116 |
| 27-May-1985 |
karels |
make persist do sensible things if window is small, but nonzero
|
| #
17361 |
| 14-Nov-1984 |
karels |
add snd_cwndw to decrease outstanding data after a source quench;
fix retransmission change (1 seg when timer goes off,
then back to normal after an ack)
|