On Tue, Apr 17, 2018 at 10:25 PM, Samudrala, Sridhar
[off-list ref] wrote:

On 4/17/2018 2:07 PM, Willem de Bruijn wrote:

quoted

On Tue, Apr 17, 2018 at 4:48 PM, Sowmini Varadhan
[off-list ref] wrote:

quoted

On (04/17/18 16:23), Willem de Bruijn wrote:

quoted

Assuming IPv4 with an MTU of 1500 and the maximum segment
size of 1472, the receiver will see three datagrams with MSS of
1472B, 528B and 512B.

so the recvmsg will also pass up 1472, 526, 512, right?

That's right.

quoted

If yes, how will the recvmsg differentiate between the case
(2000 byte message followed by 512 byte message) and
(1472 byte message, 526 byte message, then 512 byte message),
in other words, how are UDP message boundary semantics preserved?

They aren't. This is purely an optimization to amortize the cost of
repeated tx stack traversal. Unlike UFO, which would preserve the
boundaries of the original larger than MTU datagram.

Doesn't this break UDP applications that expect message boundary
preservation semantics? Is it possible to negotiate this feature?

A process has to explicitly request the feature with socket option
or cmsg UDP_SEGMENT. By setting that to gso size is signals
its intent to send multiple datagrams in one call.

Or were you responding to the hypothetical GRO example below?
Yes, that clearly would have to be limited to negotiated flows, not
unlike how foo-over-udp tunneling is detected. It is also not a serious
suggestion at this point.

quoted

A prime use case is bulk transfer of data. Think video streaming
with QUIC. It must send MTU sized or smaller packets, but has
no application-layer requirement to reconstruct large packets on
the peer.

That said, for negotiated flows an inverse GRO feature could
conceivably be implemented to reduce rx stack traversal, too.
Though due to interleaving of packets on the wire, it aggregation
would be best effort, similar to TCP TSO and GRO using the
PSH bit as packetization signal.