sctp_peeloff() returns a socket of wrong t ype: should be SOCK_STREAM but is SOCK_SEQPACKET

[marcellgg]

int sd, sock, type, len=sizeof(type);

sock = socket(AF_INET, SOCK_SEQPACKET, IPPROTO_SCTP);
sctp_connectx(sock, ...);
sd = sctp_peeloff(sock, id);
getsockopt(sd, SOL_SOCKET, SO_TYPE, &type, &len);

type is SOCK_SEQPACKET, not SOCK_STREAM as expected

Regards,
Yves.

[marcellgg]

RFC 6458 section 9.2 says : Note also that the new socket is a one-to-one style socket.

[lxin]

RFC 6458 section 9.2 says : Note also that the new socket is a one-to-one style socket.

This indeed cause differences in behavior on the peeled-off socket (newfd), for example in sendmsg() when a new destination address is supplied, in the handling of msg_flags such as SCTP_EOF, SCTP_ABORT, and SCTP_SENDALL, and in the behavior of certain socket options.

I also noticed that FreeBSD sets the socket type of the peeled-off socket to SOCK_SEQPACKET., @tuexen is this expected?

Thanks.

[tuexen]

On 6. Jan 2026, at 18:09, Xin Long ***@***.***> wrote: lxin left a comment (sctp/lksctp-tools#70) RFC 6458 section 9.2 says : Note also that the new socket is a one-to-one style socket. This indeed cause differences in behavior on the peeled-off socket (newfd), for example in sendmsg() when a new destination address is supplied, in the handling of msg_flags such as SCTP_EOF, SCTP_ABORT, and SCTP_SENDALL, and in the behavior of certain socket options. I also noticed that FreeBSD sets the socket type of the peeled-off socket to SOCK_SEQPACKET., @tuexen is this expected?

No. Let me check and get back to you. Best regards Michael

…

Thanks. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[marcellgg]

I confirm that FreeBSD (15.0, sctplib-1.0.32, libusrsctp-0.9.5.0) sets the socket type of a peeled-off socket to SOCK_SEQPACKET and not to SOCK_STREAM as expected. Regrads, Yves. Le 06/01/2026 à 21:12, Michael Tüxen a écrit :

…

*tuexen* left a comment (sctp/lksctp-tools#70) <#70 (comment)> > On 6. Jan 2026, at 18:09, Xin Long ***@***.***> wrote: > > lxin left a comment (sctp/lksctp-tools#70) > RFC 6458 section 9.2 says : Note also that the new socket is a one-to-one style socket. > This indeed cause differences in behavior on the peeled-off socket (newfd), for example in sendmsg() when a new destination address is supplied, in the handling of msg_flags such as SCTP_EOF, SCTP_ABORT, and SCTP_SENDALL, and in the behavior of certain socket options. > I also noticed that FreeBSD sets the socket type of the peeled-off socket to SOCK_SEQPACKET., @tuexen is this expected? No. Let me check and get back to you. Best regards Michael > Thanks. > — > Reply to this email directly, view it on GitHub, or unsubscribe. > You are receiving this because you were mentioned.Message ID: ***@***.***> — Reply to this email directly, view it on GitHub <#70 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BEPLWV732AIBKB65TRH5QOT4FQJJRAVCNFSM6AAAAACPSGDVDSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTOMJWGE2DQNZQGY>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[tuexen]

On 6. Jan 2026, at 18:09, Xin Long ***@***.***> wrote: lxin left a comment (sctp/lksctp-tools#70) RFC 6458 section 9.2 says : Note also that the new socket is a one-to-one style socket. This indeed cause differences in behavior on the peeled-off socket (newfd), for example in sendmsg() when a new destination address is supplied, in the handling of msg_flags such as SCTP_EOF, SCTP_ABORT, and SCTP_SENDALL, and in the behavior of certain socket options. I also noticed that FreeBSD sets the socket type of the peeled-off socket to SOCK_SEQPACKET., @tuexen is this expected?

This is fixed in https://cgit.FreeBSD.org/src/commit/?id=d195b3783fa4de5c1a95f6d95eb9444abce6778b <https://cgit.freebsd.org/src/commit/?id=d195b3783fa4de5c1a95f6d95eb9444abce6778b> Best regards Michael

…

Thanks. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>