[PATCH net-next 01/24] sctp: introduce SCTP_FUTURE/CURRENT/ALL_ASSOC

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



This patch is to add 3 constants SCTP_FUTURE_ASSOC,
SCTP_CURRENT_ASSOC and SCTP_ALL_ASSOC for reserved
assoc_ids, as defined in rfc6458#section-7.2.

And add the process for them when doing lookup and
inserting in sctp_id2assoc and sctp_assoc_set_id.

Signed-off-by: Xin Long <lucien.xin@xxxxxxxxx>
---
 include/uapi/linux/sctp.h | 4 ++++
 net/sctp/associola.c      | 7 +++++--
 net/sctp/socket.c         | 2 +-
 3 files changed, 10 insertions(+), 3 deletions(-)

diff --git a/include/uapi/linux/sctp.h b/include/uapi/linux/sctp.h
index d584073..b8f2c4d 100644
--- a/include/uapi/linux/sctp.h
+++ b/include/uapi/linux/sctp.h
@@ -59,6 +59,10 @@
 
 typedef __s32 sctp_assoc_t;
 
+#define SCTP_FUTURE_ASSOC	0
+#define SCTP_CURRENT_ASSOC	1
+#define SCTP_ALL_ASSOC		2
+
 /* The following symbols come from the Sockets API Extensions for
  * SCTP <draft-ietf-tsvwg-sctpsocket-07.txt>.
  */
diff --git a/net/sctp/associola.c b/net/sctp/associola.c
index 201c888..b99f163 100644
--- a/net/sctp/associola.c
+++ b/net/sctp/associola.c
@@ -1651,8 +1651,11 @@ int sctp_assoc_set_id(struct sctp_association *asoc, gfp_t gfp)
 	if (preload)
 		idr_preload(gfp);
 	spin_lock_bh(&sctp_assocs_id_lock);
-	/* 0 is not a valid assoc_id, must be >= 1 */
-	ret = idr_alloc_cyclic(&sctp_assocs_id, asoc, 1, 0, GFP_NOWAIT);
+	/* 0, 1, 2 are used as SCTP_FUTURE_ASSOC, SCTP_CURRENT_ASSOC and
+	 * SCTP_ALL_ASSOC, so an available id must be > SCTP_ALL_ASSOC.
+	 */
+	ret = idr_alloc_cyclic(&sctp_assocs_id, asoc, SCTP_ALL_ASSOC + 1, 0,
+			       GFP_NOWAIT);
 	spin_unlock_bh(&sctp_assocs_id_lock);
 	if (preload)
 		idr_preload_end();
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index f93c3cf..a52d132 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -248,7 +248,7 @@ struct sctp_association *sctp_id2assoc(struct sock *sk, sctp_assoc_t id)
 	}
 
 	/* Otherwise this is a UDP-style socket. */
-	if (!id || (id == (sctp_assoc_t)-1))
+	if (id <= SCTP_ALL_ASSOC)
 		return NULL;
 
 	spin_lock_bh(&sctp_assocs_id_lock);
-- 
2.1.0




[Index of Archives]     [Linux Networking Development]     [Linux OMAP]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux