diff options
| author | Andrei Vagin <avagin@google.com> | 2019-03-28 11:02:23 -0700 |
|---|---|---|
| committer | Shentubot <shentubot@google.com> | 2019-03-28 11:03:41 -0700 |
| commit | f4105ac21a9f11f5231681239ca92ac814b5149d (patch) | |
| tree | 2ecd11f283e674bce8cd29b514ae0745fdd0fa83 /pkg/tcpip/transport/tcp/accept.go | |
| parent | 9c188978870051f0b42ceb1a3f16320286936976 (diff) | |
netstack/fdbased: add generic segmentation offload (GSO) support
The linux packet socket can handle GSO packets, so we can segment packets to
64K instead of the MTU which is usually 1500.
Here are numbers for the nginx-1m test:
runsc: 579330.01 [Kbytes/sec] received
runsc-gso: 1794121.66 [Kbytes/sec] received
runc: 2122139.06 [Kbytes/sec] received
and for tcp_benchmark:
$ tcp_benchmark --duration 15 --ideal
[ 4] 0.0-15.0 sec 86647 MBytes 48456 Mbits/sec
$ tcp_benchmark --client --duration 15 --ideal
[ 4] 0.0-15.0 sec 2173 MBytes 1214 Mbits/sec
$ tcp_benchmark --client --duration 15 --ideal --gso 65536
[ 4] 0.0-15.0 sec 19357 MBytes 10825 Mbits/sec
PiperOrigin-RevId: 240809103
Change-Id: I2637f104db28b5d4c64e1e766c610162a195775a
Diffstat (limited to 'pkg/tcpip/transport/tcp/accept.go')
| -rw-r--r-- | pkg/tcpip/transport/tcp/accept.go | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/pkg/tcpip/transport/tcp/accept.go b/pkg/tcpip/transport/tcp/accept.go index 7a19737c7..a3894ed8f 100644 --- a/pkg/tcpip/transport/tcp/accept.go +++ b/pkg/tcpip/transport/tcp/accept.go @@ -214,6 +214,8 @@ func (l *listenContext) createConnectedEndpoint(s *segment, iss seqnum.Value, ir n.maybeEnableTimestamp(rcvdSynOpts) n.maybeEnableSACKPermitted(rcvdSynOpts) + n.initGSO() + // Register new endpoint so that packets are routed to it. if err := n.stack.RegisterTransportEndpoint(n.boundNICID, n.effectiveNetProtos, ProtocolNumber, n.id, n, n.reusePort); err != nil { n.Close() |