diff options
author | Mark Johnston <markj@FreeBSD.org> | 2021-09-11 14:15:21 +0000 |
---|---|---|
committer | Michael Tuexen <tuexen@FreeBSD.org> | 2022-02-22 22:47:28 +0000 |
commit | aa16c2fcaf1d5e541fc91a66bbba352ba313429c (patch) | |
tree | c91f46d2c1df82b3c82433291496717dfa7de5c7 /sys/netinet/sctp_input.c | |
parent | 238e4a1179b0b03bf39bcc4e096f6771d2d82faa (diff) |
sctp: Tighten up locking around sctp_aloc_assoc()
All callers of sctp_aloc_assoc() mark the PCB as connected after a
successful call (for one-to-one-style sockets). In all cases this is
done without the PCB lock, so the PCB's flags can be corrupted. We also
do not atomically check whether a one-to-one-style socket is a listening
socket, which violates various assumptions in solisten_proto().
We need to hold the PCB lock across all of sctp_aloc_assoc() to fix
this. In order to do that without introducing lock order reversals, we
have to hold the global info lock as well.
So:
- Convert sctp_aloc_assoc() so that the inp and info locks are
consistently held. It returns with the association lock held, as
before.
- Fix an apparent bug where we failed to remove an association from a
global hash if sctp_add_remote_addr() fails.
- sctp_select_a_tag() is called when initializing an association, and it
acquires the global info lock. To avoid lock recursion, push locking
into its callers.
- Introduce sctp_aloc_assoc_connected(), which atomically checks for a
listening socket and sets SCTP_PCB_FLAGS_CONNECTED.
There is still one edge case in sctp_process_cookie_new() where we do
not update PCB/socket state correctly.
Reviewed by: tuexen
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D31908
(cherry picked from commit 2d5c48eccd9f29b9df8020bb1c3a8ffda38df37b)
Diffstat (limited to 'sys/netinet/sctp_input.c')
-rw-r--r-- | sys/netinet/sctp_input.c | 4 |
1 files changed, 4 insertions, 0 deletions
diff --git a/sys/netinet/sctp_input.c b/sys/netinet/sctp_input.c index b257a62e02f8..84d2a303303d 100644 --- a/sys/netinet/sctp_input.c +++ b/sys/netinet/sctp_input.c @@ -679,7 +679,9 @@ sctp_handle_nat_colliding_state(struct sctp_tcb *stcb) if ((SCTP_GET_STATE(stcb) == SCTP_STATE_COOKIE_WAIT) || (SCTP_GET_STATE(stcb) == SCTP_STATE_COOKIE_ECHOED)) { + SCTP_INP_INFO_RLOCK(); new_vtag = sctp_select_a_tag(stcb->sctp_ep, stcb->sctp_ep->sctp_lport, stcb->rport, 1); + SCTP_INP_INFO_RLOCK(); atomic_add_int(&stcb->asoc.refcnt, 1); SCTP_TCB_UNLOCK(stcb); SCTP_INP_INFO_WLOCK(); @@ -2183,6 +2185,8 @@ sctp_process_cookie_new(struct mbuf *m, int iphlen, int offset, * INIT/INIT-ACK/COOKIE arrived. But of course then it * should have went to the other code.. not here.. oh well.. * a bit of protection is worth having.. + * + * XXXMJ unlocked */ stcb->sctp_ep->sctp_flags |= SCTP_PCB_FLAGS_CONNECTED; soisconnected(stcb->sctp_socket); |