tls_sendmsg should send partial data for TCP

[djwatson]

Currently tls_sendmsg does a size check:

if (size > KTLS_MAX_PAYLOAD_SIZE) { ret = -E2BIG; goto send_end; }

For tcp, we should be sending KTLS_MAX_PAYLOAD_SIZE bytes, and returning the number of bytes sent, or providing a socket buffer somehow.

I think this is probably OK for UDP/DTLS though.

The man page says it should be EMSGSIZE:

If the message is too long to pass atomically through the underlying protocol, the error EMSGSIZE is returned, and the message is not transmitted.

[fridex]

For tcp, we should be sending KTLS_MAX_PAYLOAD_SIZE bytes, and returning the number of bytes sent

Sounds reasonable to transparently assemble records from kvec (simply by iterating over scatter/gather array passed) so user space does not have to care about it. I don't like the buffer approach, is it really needed here?

Nevertheless, I would consider to send payload based on MTU set by setsockopt(2) (which is KTLS_MAX_PAYLOAD_SIZE by default) for both, TLS and DTLS (for DTLS limit what can be sent).

The man page says it should be EMSGSIZE:

Return values have to be reviewed. To be honest, I did not give any special care to return values at all (https://github.com/fridex/af_ktls/issues/3).

[lancerchao]

OpenSSL[1] segments data into record buffers if size overflows.

[1] https://www.ibm.com/support/knowledgecenter/ssw_i5_54/apis/sslwrite.htm

[djwatson]

fixed by #62

[lancerchao]

Why is this closed? #62 fixes recv partial data. It hasn't changed send.

[fridex]

Reopening...

[djwatson]

implemented in #84