venugopal iyer wrote:
Some comments inline. Also, I looked at the rest of the SCTP files and the
changes look fine to me, except for a couple of nits.
usr/src/uts/common/inet/sctp/sctp_cookie.c
L814: change printf() before putback?
Either it will become an ASSERT, or it will be removed.
Which one depends on the TX review.
usr/src/uts/common/inet/sctp/sctp_error.c
L289,444: change printf() before putback?
Ditto.
usr/src/uts/common/inet/sctp/sctp_input.c
L357: True, it is not required.
Will fix.
usr/src/uts/common/inet/sctp/sctp_opt_data.c
L600-601: SCTP instead of TCP
OK
L648,653: in theory the max len would be the max number of
logical addresses that can be/is plumbed * struct sockaddr_in6
(for both v4 and v6). In practice, you can get the nladdrs
and * struct sockaddr_in6.
I see the code actually does it gracefully by taking
addr_cnt = buflen / addr_size;
and then only copying out that many addresses.
L1061: Is SCTP in secs now after this?
Yes.
I'll rephrase the comment to say that SCTP, just like TCP, has l_linger
in seconds.
Erik
_______________________________________________
networking-discuss mailing list
[email protected]
