Add support for TCP receive buffer auto tuning.

The implementation is similar to linux where we track the number of bytes
consumed by the application to grow the receive buffer of a given TCP endpoint.

This ensures that the advertised window grows at a reasonable rate to accomodate
for the sender's rate and prevents large amounts of data being held in stack
buffers if the application is not actively reading or not reading fast enough.

The original paper that was used to implement the linux receive buffer auto-
tuning is available @ https://public.lanl.gov/radiant/pubs/drs/lacsi2001.pdf

NOTE: Linux does not implement DRS as defined in that paper, it's just a good
reference to understand the solution space.

Updates #230

PiperOrigin-RevId: 253168283

2 files changed, 46 insertions, 0 deletions

diff --git a/pkg/tcpip/stack/stack.go b/pkg/tcpip/stack/stack.go
index 8ecc51a58..2d7f56ca9 100644
--- a/pkg/tcpip/stack/stack.go
+++ b/pkg/tcpip/stack/stack.go
@@ -225,6 +225,45 @@ type TCPSACKInfo struct {
 	MaxSACKED seqnum.Value
 }
 
+// RcvBufAutoTuneParams holds state related to TCP receive buffer auto-tuning.
+type RcvBufAutoTuneParams struct {
+	// MeasureTime is the time at which the current measurement
+	// was started.
+	MeasureTime time.Time
+
+	// CopiedBytes is the number of bytes copied to user space since
+	// this measure began.
+	CopiedBytes int
+
+	// PrevCopiedBytes is the number of bytes copied to user space in
+	// the previous RTT period.
+	PrevCopiedBytes int
+
+	// RcvBufSize is the auto tuned receive buffer size.
+	RcvBufSize int
+
+	// RTT is the smoothed RTT as measured by observing the time between
+	// when a byte is first acknowledged and the receipt of data that is at
+	// least one window beyond the sequence number that was acknowledged.
+	RTT time.Duration
+
+	// RTTVar is the "round-trip time variation" as defined in section 2
+	// of RFC6298.
+	RTTVar time.Duration
+
+	// RTTMeasureSeqNumber is the highest acceptable sequence number at the
+	// time this RTT measurement period began.
+	RTTMeasureSeqNumber seqnum.Value
+
+	// RTTMeasureTime is the absolute time at which the current RTT
+	// measurement period began.
+	RTTMeasureTime time.Time
+
+	// Disabled is true if an explicit receive buffer is set for the
+	// endpoint.
+	Disabled bool
+}
+
 // TCPEndpointState is a copy of the internal state of a TCP endpoint.
 type TCPEndpointState struct {
 	// ID is a copy of the TransportEndpointID for the endpoint.
@@ -240,6 +279,10 @@ type TCPEndpointState struct {
 	// buffer for the endpoint.
 	RcvBufUsed int
 
+	// RcvBufAutoTuneParams is used to hold state variables to compute
+	// the auto tuned receive buffer size.
+	RcvAutoParams RcvBufAutoTuneParams
+
 	// RcvClosed if true, indicates the endpoint has been closed for reading.
 	RcvClosed bool
 
diff --git a/pkg/tcpip/stack/transport_test.go b/pkg/tcpip/stack/transport_test.go
index 10fc8f195..788ffcc8c 100644
--- a/pkg/tcpip/stack/transport_test.go
+++ b/pkg/tcpip/stack/transport_test.go
@@ -192,6 +192,9 @@ func (f *fakeTransportEndpoint) State() uint32 {
 	return 0
 }
 
+func (f *fakeTransportEndpoint) ModerateRecvBuf(copied int) {
+}
+
 type fakeTransportGoodOption bool
 
 type fakeTransportBadOption bool