P2P: improve RX/TX flow control #2017

pull jgarzik wants to merge 2 commits into bitcoin:master from jgarzik:tx-full changing 2 files +40 −28

jgarzik commented at 11:27 PM on November 15, 2012: contributor
- "optimistic write": Push each message to kernel socket buffer immediately
- If there is write data at select time, that implies send() blocked during optimistic write. Drain write queue, before receiving any more messages.
This avoids needlessly queueing received data, if the remote peer is not themselves receiving data.

Result: write buffer (and thus memory usage) is kept small, DoS potential is slightly lower, and TCP flow control signalling is properly utilized.

The kernel will queue data into the socket buffer, then signal the remote peer to stop sending data, until we resume reading again.
P2P, cosmetic: break out buffer send(2) code into separate function fca5b98528

P2P: improve RX/TX flow control

1) "optimistic write": Push each message to kernel socket buffer immediately.

2) If there is write data at select time, that implies send() blocked
   during optimistic write.  Drain write queue, before receiving
   any more messages.

This avoids needlessly queueing received data, if the remote peer
is not themselves receiving data.

Result: write buffer (and thus memory usage) is kept small, DoS
potential is slightly lower, and TCP flow control signalling is
properly utilized.

The kernel will queue data into the socket buffer, then signal the
remote peer to stop sending data, until we resume reading again.

4e4e6d1fec

in src/net.cpp:None in 4e4e6d1fec
```
 789 |                  {
 790 |                      TRY_LOCK(pnode->cs_vSend, lockSend);
 791 | -                    if (lockSend && !pnode->vSend.empty())
 792 | -                        FD_SET(pnode->hSocket, &fdsetSend);
 793 | +                    if (lockSend) {
 794 | +                        // do not read, if draining write queue
```
laanwj commented at 12:30 PM on November 17, 2012:

I don't understand this part. Why make fdsetRecv depend on the state of the write queue?

jgarzik commented at 6:17 PM on November 17, 2012:

@laanwj That is explained in the commit message / pull request message.

mikehearn commented at 11:07 AM on November 21, 2012:

But future readers of the code won't see your commit message. They need comments to understand what's going on.
BitcoinPullTester commented at 6:42 PM on January 24, 2013: none

Automatic sanity-testing: PASSED, see http://jenkins.bluematt.me/pull-tester/4e4e6d1feca92c346389d4221d022db8f8cd3e78 for binaries and test log.
sipa commented at 2:12 PM on February 24, 2013: member

@jgarzik @gavinandresen I'd like to see this in 0.8.1.
jgarzik commented at 4:29 PM on March 24, 2013: contributor

Superceded by #2409
jgarzik closed this on Mar 24, 2013
jgarzik deleted the branch on Aug 24, 2014
KolbyML referenced this in commit 0fe1242ee6 on Dec 5, 2020
DrahtBot locked this on Sep 8, 2021

Contributors

Linked (view graph)

#2409 Network optimalizations