[PATCHv2 net-next 0/4] sctp: add subscribe per asoc and sockopt SCTP_EVENT

[PATCHv2 net-next 0/4] sctp: add subscribe per asoc and sockopt SCTP_EVENT

[Xin Long]

This patchset mainly adds the Event Subscription sockopt described in
rfc6525#section-6.2:

"Subscribing to events as described in [RFC6458] uses a setsockopt()
call with the SCTP_EVENT socket option.  This option takes the
following structure, which specifies the association, the event type
(using the same value found in the event type field), and an on/off
boolean.

  struct sctp_event {
    sctp_assoc_t se_assoc_id;
    uint16_t     se_type;
    uint8_t      se_on;
  };

The user fills in the se_type field with the same value found in the
strreset_type field, i.e., SCTP_STREAM_RESET_EVENT.  The user will
also fill in the se_assoc_id field with either the association to set
this event on (this field is ignored for one-to-one style sockets) or
one of the reserved constant values defined in [RFC6458].  Finally,
the se_on field is set with a 1 to enable the event or a 0 to disable
the event."

As for the old SCTP_EVENTS Option with struct sctp_event_subscribe,
it's being DEPRECATED.

Xin Long (4):
  sctp: define subscribe in sctp_sock as __u16
  sctp: add subscribe per asoc
  sctp: rename enum sctp_event to sctp_event_type
  sctp: add sockopt SCTP_EVENT

 include/net/sctp/constants.h |   2 +-
 include/net/sctp/sm.h        |   4 +-
 include/net/sctp/structs.h   |   4 +-
 include/net/sctp/ulpevent.h  |  39 ++++++++------
 include/uapi/linux/sctp.h    |  13 ++++-
 net/sctp/associola.c         |   2 +
 net/sctp/chunk.c             |   8 ++-
 net/sctp/primitive.c         |   2 +-
 net/sctp/sm_sideeffect.c     |  12 ++---
 net/sctp/sm_statetable.c     |   2 +-
 net/sctp/socket.c            | 126 ++++++++++++++++++++++++++++++++++++++++---
 net/sctp/stream_interleave.c |  12 +++--
 net/sctp/ulpqueue.c          |   8 +--
 13 files changed, 184 insertions(+), 50 deletions(-)

-- 
2.1.0

[PATCHv2 net-next 0/4] sctp: add subscribe per asoc and sockopt SCTP_EVENT

[Xin Long]

This patchset mainly adds the Event Subscription sockopt described in
rfc6525#section-6.2:

"Subscribing to events as described in [RFC6458] uses a setsockopt()
call with the SCTP_EVENT socket option.  This option takes the
following structure, which specifies the association, the event type
(using the same value found in the event type field), and an on/off
boolean.

  struct sctp_event {
    sctp_assoc_t se_assoc_id;
    uint16_t     se_type;
    uint8_t      se_on;
  };

The user fills in the se_type field with the same value found in the
strreset_type field, i.e., SCTP_STREAM_RESET_EVENT.  The user will
also fill in the se_assoc_id field with either the association to set
this event on (this field is ignored for one-to-one style sockets) or
one of the reserved constant values defined in [RFC6458].  Finally,
the se_on field is set with a 1 to enable the event or a 0 to disable
the event."

As for the old SCTP_EVENTS Option with struct sctp_event_subscribe,
it's being DEPRECATED.

Xin Long (4):
  sctp: define subscribe in sctp_sock as __u16
  sctp: add subscribe per asoc
  sctp: rename enum sctp_event to sctp_event_type
  sctp: add sockopt SCTP_EVENT

 include/net/sctp/constants.h |   2 +-
 include/net/sctp/sm.h        |   4 +-
 include/net/sctp/structs.h   |   4 +-
 include/net/sctp/ulpevent.h  |  39 ++++++++------
 include/uapi/linux/sctp.h    |  13 ++++-
 net/sctp/associola.c         |   2 +
 net/sctp/chunk.c             |   8 ++-
 net/sctp/primitive.c         |   2 +-
 net/sctp/sm_sideeffect.c     |  12 ++---
 net/sctp/sm_statetable.c     |   2 +-
 net/sctp/socket.c            | 126 ++++++++++++++++++++++++++++++++++++++++---
 net/sctp/stream_interleave.c |  12 +++--
 net/sctp/ulpqueue.c          |   8 +--
 13 files changed, 184 insertions(+), 50 deletions(-)

-- 
2.1.0

[PATCHv2 net-next 1/4] sctp: define subscribe in sctp_sock as __u16

[Xin Long]

The member subscribe in sctp_sock is used to indicate to which of
the events it is subscribed, more like a group of flags. So it's
better to be defined as __u16 (2 bytpes), instead of struct
sctp_event_subscribe (13 bytes).

Note that sctp_event_subscribe is an UAPI struct, used on sockopt
calls, and thus it will not be removed. This patch only changes
the internal storage of the flags.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
---
 include/net/sctp/structs.h   |  2 +-
 include/net/sctp/ulpevent.h  | 39 ++++++++++++++++++++++++---------------
 include/uapi/linux/sctp.h    |  6 +++++-
 net/sctp/chunk.c             |  4 ++--
 net/sctp/socket.c            | 35 ++++++++++++++++++++++++++---------
 net/sctp/stream_interleave.c | 11 ++++++-----
 net/sctp/ulpqueue.c          |  8 ++++----
 7 files changed, 68 insertions(+), 37 deletions(-)

diff --git a/include/net/sctp/structs.h b/include/net/sctp/structs.h
index af9d494..bc7808a 100644
--- a/include/net/sctp/structs.h
+++ b/include/net/sctp/structs.h
@@ -217,7 +217,7 @@ struct sctp_sock {
 	 * These two structures must be grouped together for the usercopy
 	 * whitelist region.
 	 */
-	struct sctp_event_subscribe subscribe;
+	__u16 subscribe;
 	struct sctp_initmsg initmsg;
 
 	int user_frag;
diff --git a/include/net/sctp/ulpevent.h b/include/net/sctp/ulpevent.h
index 51b4e06..bd922a0 100644
--- a/include/net/sctp/ulpevent.h
+++ b/include/net/sctp/ulpevent.h
@@ -164,30 +164,39 @@ void sctp_ulpevent_read_nxtinfo(const struct sctp_ulpevent *event,
 
 __u16 sctp_ulpevent_get_notification_type(const struct sctp_ulpevent *event);
 
+static inline void sctp_ulpevent_type_set(__u16 *subscribe,
+					  __u16 sn_type, __u8 on)
+{
+	if (sn_type > SCTP_SN_TYPE_MAX)
+		return;
+
+	if (on)
+		*subscribe |=  (1 << (sn_type - SCTP_SN_TYPE_BASE));
+	else
+		*subscribe &= ~(1 << (sn_type - SCTP_SN_TYPE_BASE));
+}
+
 /* Is this event type enabled? */
-static inline int sctp_ulpevent_type_enabled(__u16 sn_type,
-					     struct sctp_event_subscribe *mask)
+static inline bool sctp_ulpevent_type_enabled(__u16 subscribe, __u16 sn_type)
 {
-	int offset = sn_type - SCTP_SN_TYPE_BASE;
-	char *amask = (char *) mask;
+	if (sn_type > SCTP_SN_TYPE_MAX)
+		return false;
 
-	if (offset >= sizeof(struct sctp_event_subscribe))
-		return 0;
-	return amask[offset];
+	return subscribe & (1 << (sn_type - SCTP_SN_TYPE_BASE));
 }
 
 /* Given an event subscription, is this event enabled? */
-static inline int sctp_ulpevent_is_enabled(const struct sctp_ulpevent *event,
-					   struct sctp_event_subscribe *mask)
+static inline bool sctp_ulpevent_is_enabled(const struct sctp_ulpevent *event,
+					    __u16 subscribe)
 {
 	__u16 sn_type;
-	int enabled = 1;
 
-	if (sctp_ulpevent_is_notification(event)) {
-		sn_type = sctp_ulpevent_get_notification_type(event);
-		enabled = sctp_ulpevent_type_enabled(sn_type, mask);
-	}
-	return enabled;
+	if (!sctp_ulpevent_is_notification(event))
+		return true;
+
+	sn_type = sctp_ulpevent_get_notification_type(event);
+
+	return sctp_ulpevent_type_enabled(subscribe, sn_type);
 }
 
 #endif /* __sctp_ulpevent_h__ */
diff --git a/include/uapi/linux/sctp.h b/include/uapi/linux/sctp.h
index c81feb3..66afa5b 100644
--- a/include/uapi/linux/sctp.h
+++ b/include/uapi/linux/sctp.h
@@ -632,7 +632,9 @@ union sctp_notification {
  */
 
 enum sctp_sn_type {
-	SCTP_SN_TYPE_BASE     = (1<<15),
+	SCTP_SN_TYPE_BASE	= (1<<15),
+	SCTP_DATA_IO_EVENT	= SCTP_SN_TYPE_BASE,
+#define SCTP_DATA_IO_EVENT		SCTP_DATA_IO_EVENT
 	SCTP_ASSOC_CHANGE,
 #define SCTP_ASSOC_CHANGE		SCTP_ASSOC_CHANGE
 	SCTP_PEER_ADDR_CHANGE,
@@ -657,6 +659,8 @@ enum sctp_sn_type {
 #define SCTP_ASSOC_RESET_EVENT		SCTP_ASSOC_RESET_EVENT
 	SCTP_STREAM_CHANGE_EVENT,
 #define SCTP_STREAM_CHANGE_EVENT	SCTP_STREAM_CHANGE_EVENT
+	SCTP_SN_TYPE_MAX	= SCTP_STREAM_CHANGE_EVENT,
+#define SCTP_SN_TYPE_MAX		SCTP_SN_TYPE_MAX
 };
 
 /* Notification error codes used to fill up the error fields in some
diff --git a/net/sctp/chunk.c b/net/sctp/chunk.c
index ce80878..6c761af 100644
--- a/net/sctp/chunk.c
+++ b/net/sctp/chunk.c
@@ -109,8 +109,8 @@ static void sctp_datamsg_destroy(struct sctp_datamsg *msg)
 				error = asoc->outqueue.error;
 
 			sp = sctp_sk(asoc->base.sk);
-			notify = sctp_ulpevent_type_enabled(SCTP_SEND_FAILED,
-							    &sp->subscribe);
+			notify = sctp_ulpevent_type_enabled(sp->subscribe,
+							    SCTP_SEND_FAILED);
 		}
 
 		/* Generate a SEND FAILED event only if enabled. */
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index 5299add..48e0b45 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -2230,7 +2230,7 @@ static int sctp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
 	if (sp->recvrcvinfo)
 		sctp_ulpevent_read_rcvinfo(event, msg);
 	/* Check if we allow SCTP_SNDRCVINFO. */
-	if (sp->subscribe.sctp_data_io_event)
+	if (sctp_ulpevent_type_enabled(sp->subscribe, SCTP_DATA_IO_EVENT))
 		sctp_ulpevent_read_sndrcvinfo(event, msg);
 
 	err = copied;
@@ -2304,21 +2304,28 @@ static int sctp_setsockopt_disable_fragments(struct sock *sk,
 static int sctp_setsockopt_events(struct sock *sk, char __user *optval,
 				  unsigned int optlen)
 {
-	struct sctp_association *asoc;
-	struct sctp_ulpevent *event;
+	struct sctp_event_subscribe subscribe;
+	__u8 *sn_type = (__u8 *)&subscribe;
+	struct sctp_sock *sp = sctp_sk(sk);
+	int i;
 
 	if (optlen > sizeof(struct sctp_event_subscribe))
 		return -EINVAL;
-	if (copy_from_user(&sctp_sk(sk)->subscribe, optval, optlen))
+
+	if (copy_from_user(&subscribe, optval, optlen))
 		return -EFAULT;
 
+	for (i = 0; i < optlen; i++)
+		sctp_ulpevent_type_set(&sp->subscribe, SCTP_SN_TYPE_BASE + i,
+				       sn_type[i]);
+
 	/* At the time when a user app subscribes to SCTP_SENDER_DRY_EVENT,
 	 * if there is no data to be sent or retransmit, the stack will
 	 * immediately send up this notification.
 	 */
-	if (sctp_ulpevent_type_enabled(SCTP_SENDER_DRY_EVENT,
-				       &sctp_sk(sk)->subscribe)) {
-		asoc = sctp_id2assoc(sk, 0);
+	if (sctp_ulpevent_type_enabled(sp->subscribe, SCTP_SENDER_DRY_EVENT)) {
+		struct sctp_association *asoc = sctp_id2assoc(sk, 0);
+		struct sctp_ulpevent *event;
 
 		if (asoc && sctp_outq_is_empty(&asoc->outqueue)) {
 			event = sctp_ulpevent_make_sender_dry_event(asoc,
@@ -4722,7 +4729,7 @@ static int sctp_init_sock(struct sock *sk)
 	/* Initialize default event subscriptions. By default, all the
 	 * options are off.
 	 */
-	memset(&sp->subscribe, 0, sizeof(struct sctp_event_subscribe));
+	sp->subscribe = 0;
 
 	/* Default Peer Address Parameters.  These defaults can
 	 * be modified via SCTP_PEER_ADDR_PARAMS
@@ -5267,14 +5274,24 @@ static int sctp_getsockopt_disable_fragments(struct sock *sk, int len,
 static int sctp_getsockopt_events(struct sock *sk, int len, char __user *optval,
 				  int __user *optlen)
 {
+	struct sctp_event_subscribe subscribe;
+	__u8 *sn_type = (__u8 *)&subscribe;
+	int i;
+
 	if (len == 0)
 		return -EINVAL;
 	if (len > sizeof(struct sctp_event_subscribe))
 		len = sizeof(struct sctp_event_subscribe);
 	if (put_user(len, optlen))
 		return -EFAULT;
-	if (copy_to_user(optval, &sctp_sk(sk)->subscribe, len))
+
+	for (i = 0; i <= len; i++)
+		sn_type[i] = sctp_ulpevent_type_enabled(sctp_sk(sk)->subscribe,
+							SCTP_SN_TYPE_BASE + i);
+
+	if (copy_to_user(optval, &subscribe, len))
 		return -EFAULT;
+
 	return 0;
 }
 
diff --git a/net/sctp/stream_interleave.c b/net/sctp/stream_interleave.c
index 2b499a8..ceef5a3 100644
--- a/net/sctp/stream_interleave.c
+++ b/net/sctp/stream_interleave.c
@@ -503,7 +503,7 @@ static int sctp_enqueue_event(struct sctp_ulpq *ulpq,
 		sk_incoming_cpu_update(sk);
 	}
 
-	if (!sctp_ulpevent_is_enabled(event, &sp->subscribe))
+	if (!sctp_ulpevent_is_enabled(event, sp->subscribe))
 		goto out_free;
 
 	if (skb_list)
@@ -992,10 +992,11 @@ static void sctp_intl_stream_abort_pd(struct sctp_ulpq *ulpq, __u16 sid,
 				      __u32 mid, __u16 flags, gfp_t gfp)
 {
 	struct sock *sk = ulpq->asoc->base.sk;
+	struct sctp_sock *sp = sctp_sk(sk);
 	struct sctp_ulpevent *ev = NULL;
 
-	if (!sctp_ulpevent_type_enabled(SCTP_PARTIAL_DELIVERY_EVENT,
-					&sctp_sk(sk)->subscribe))
+	if (!sctp_ulpevent_type_enabled(sp->subscribe,
+					SCTP_PARTIAL_DELIVERY_EVENT))
 		return;
 
 	ev = sctp_ulpevent_make_pdapi(ulpq->asoc, SCTP_PARTIAL_DELIVERY_ABORTED,
@@ -1003,8 +1004,8 @@ static void sctp_intl_stream_abort_pd(struct sctp_ulpq *ulpq, __u16 sid,
 	if (ev) {
 		__skb_queue_tail(&sk->sk_receive_queue, sctp_event2skb(ev));
 
-		if (!sctp_sk(sk)->data_ready_signalled) {
-			sctp_sk(sk)->data_ready_signalled = 1;
+		if (!sp->data_ready_signalled) {
+			sp->data_ready_signalled = 1;
 			sk->sk_data_ready(sk);
 		}
 	}
diff --git a/net/sctp/ulpqueue.c b/net/sctp/ulpqueue.c
index 331cc73..b36dd90 100644
--- a/net/sctp/ulpqueue.c
+++ b/net/sctp/ulpqueue.c
@@ -219,7 +219,7 @@ int sctp_ulpq_tail_event(struct sctp_ulpq *ulpq, struct sctp_ulpevent *event)
 		sk_incoming_cpu_update(sk);
 	}
 	/* Check if the user wishes to receive this event.  */
-	if (!sctp_ulpevent_is_enabled(event, &sp->subscribe))
+	if (!sctp_ulpevent_is_enabled(event, sp->subscribe))
 		goto out_free;
 
 	/* If we are in partial delivery mode, post to the lobby until
@@ -1129,16 +1129,16 @@ void sctp_ulpq_renege(struct sctp_ulpq *ulpq, struct sctp_chunk *chunk,
 void sctp_ulpq_abort_pd(struct sctp_ulpq *ulpq, gfp_t gfp)
 {
 	struct sctp_ulpevent *ev = NULL;
-	struct sock *sk;
 	struct sctp_sock *sp;
+	struct sock *sk;
 
 	if (!ulpq->pd_mode)
 		return;
 
 	sk = ulpq->asoc->base.sk;
 	sp = sctp_sk(sk);
-	if (sctp_ulpevent_type_enabled(SCTP_PARTIAL_DELIVERY_EVENT,
-				       &sctp_sk(sk)->subscribe))
+	if (sctp_ulpevent_type_enabled(sp->subscribe,
+				       SCTP_PARTIAL_DELIVERY_EVENT))
 		ev = sctp_ulpevent_make_pdapi(ulpq->asoc,
 					      SCTP_PARTIAL_DELIVERY_ABORTED,
 					      0, 0, 0, gfp);
-- 
2.1.0

[PATCHv2 net-next 1/4] sctp: define subscribe in sctp_sock as __u16

[Xin Long]

The member subscribe in sctp_sock is used to indicate to which of
the events it is subscribed, more like a group of flags. So it's
better to be defined as __u16 (2 bytpes), instead of struct
sctp_event_subscribe (13 bytes).

Note that sctp_event_subscribe is an UAPI struct, used on sockopt
calls, and thus it will not be removed. This patch only changes
the internal storage of the flags.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
---
 include/net/sctp/structs.h   |  2 +-
 include/net/sctp/ulpevent.h  | 39 ++++++++++++++++++++++++---------------
 include/uapi/linux/sctp.h    |  6 +++++-
 net/sctp/chunk.c             |  4 ++--
 net/sctp/socket.c            | 35 ++++++++++++++++++++++++++---------
 net/sctp/stream_interleave.c | 11 ++++++-----
 net/sctp/ulpqueue.c          |  8 ++++----
 7 files changed, 68 insertions(+), 37 deletions(-)

diff --git a/include/net/sctp/structs.h b/include/net/sctp/structs.h
index af9d494..bc7808a 100644
--- a/include/net/sctp/structs.h
+++ b/include/net/sctp/structs.h
@@ -217,7 +217,7 @@ struct sctp_sock {
 	 * These two structures must be grouped together for the usercopy
 	 * whitelist region.
 	 */
-	struct sctp_event_subscribe subscribe;
+	__u16 subscribe;
 	struct sctp_initmsg initmsg;
 
 	int user_frag;
diff --git a/include/net/sctp/ulpevent.h b/include/net/sctp/ulpevent.h
index 51b4e06..bd922a0 100644
--- a/include/net/sctp/ulpevent.h
+++ b/include/net/sctp/ulpevent.h
@@ -164,30 +164,39 @@ void sctp_ulpevent_read_nxtinfo(const struct sctp_ulpevent *event,
 
 __u16 sctp_ulpevent_get_notification_type(const struct sctp_ulpevent *event);
 
+static inline void sctp_ulpevent_type_set(__u16 *subscribe,
+					  __u16 sn_type, __u8 on)
+{
+	if (sn_type > SCTP_SN_TYPE_MAX)
+		return;
+
+	if (on)
+		*subscribe |=  (1 << (sn_type - SCTP_SN_TYPE_BASE));
+	else
+		*subscribe &= ~(1 << (sn_type - SCTP_SN_TYPE_BASE));
+}
+
 /* Is this event type enabled? */
-static inline int sctp_ulpevent_type_enabled(__u16 sn_type,
-					     struct sctp_event_subscribe *mask)
+static inline bool sctp_ulpevent_type_enabled(__u16 subscribe, __u16 sn_type)
 {
-	int offset = sn_type - SCTP_SN_TYPE_BASE;
-	char *amask = (char *) mask;
+	if (sn_type > SCTP_SN_TYPE_MAX)
+		return false;
 
-	if (offset >= sizeof(struct sctp_event_subscribe))
-		return 0;
-	return amask[offset];
+	return subscribe & (1 << (sn_type - SCTP_SN_TYPE_BASE));
 }
 
 /* Given an event subscription, is this event enabled? */
-static inline int sctp_ulpevent_is_enabled(const struct sctp_ulpevent *event,
-					   struct sctp_event_subscribe *mask)
+static inline bool sctp_ulpevent_is_enabled(const struct sctp_ulpevent *event,
+					    __u16 subscribe)
 {
 	__u16 sn_type;
-	int enabled = 1;
 
-	if (sctp_ulpevent_is_notification(event)) {
-		sn_type = sctp_ulpevent_get_notification_type(event);
-		enabled = sctp_ulpevent_type_enabled(sn_type, mask);
-	}
-	return enabled;
+	if (!sctp_ulpevent_is_notification(event))
+		return true;
+
+	sn_type = sctp_ulpevent_get_notification_type(event);
+
+	return sctp_ulpevent_type_enabled(subscribe, sn_type);
 }
 
 #endif /* __sctp_ulpevent_h__ */
diff --git a/include/uapi/linux/sctp.h b/include/uapi/linux/sctp.h
index c81feb3..66afa5b 100644
--- a/include/uapi/linux/sctp.h
+++ b/include/uapi/linux/sctp.h
@@ -632,7 +632,9 @@ union sctp_notification {
  */
 
 enum sctp_sn_type {
-	SCTP_SN_TYPE_BASE     = (1<<15),
+	SCTP_SN_TYPE_BASE	= (1<<15),
+	SCTP_DATA_IO_EVENT	= SCTP_SN_TYPE_BASE,
+#define SCTP_DATA_IO_EVENT		SCTP_DATA_IO_EVENT
 	SCTP_ASSOC_CHANGE,
 #define SCTP_ASSOC_CHANGE		SCTP_ASSOC_CHANGE
 	SCTP_PEER_ADDR_CHANGE,
@@ -657,6 +659,8 @@ enum sctp_sn_type {
 #define SCTP_ASSOC_RESET_EVENT		SCTP_ASSOC_RESET_EVENT
 	SCTP_STREAM_CHANGE_EVENT,
 #define SCTP_STREAM_CHANGE_EVENT	SCTP_STREAM_CHANGE_EVENT
+	SCTP_SN_TYPE_MAX	= SCTP_STREAM_CHANGE_EVENT,
+#define SCTP_SN_TYPE_MAX		SCTP_SN_TYPE_MAX
 };
 
 /* Notification error codes used to fill up the error fields in some
diff --git a/net/sctp/chunk.c b/net/sctp/chunk.c
index ce80878..6c761af 100644
--- a/net/sctp/chunk.c
+++ b/net/sctp/chunk.c
@@ -109,8 +109,8 @@ static void sctp_datamsg_destroy(struct sctp_datamsg *msg)
 				error = asoc->outqueue.error;
 
 			sp = sctp_sk(asoc->base.sk);
-			notify = sctp_ulpevent_type_enabled(SCTP_SEND_FAILED,
-							    &sp->subscribe);
+			notify = sctp_ulpevent_type_enabled(sp->subscribe,
+							    SCTP_SEND_FAILED);
 		}
 
 		/* Generate a SEND FAILED event only if enabled. */
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index 5299add..48e0b45 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -2230,7 +2230,7 @@ static int sctp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
 	if (sp->recvrcvinfo)
 		sctp_ulpevent_read_rcvinfo(event, msg);
 	/* Check if we allow SCTP_SNDRCVINFO. */
-	if (sp->subscribe.sctp_data_io_event)
+	if (sctp_ulpevent_type_enabled(sp->subscribe, SCTP_DATA_IO_EVENT))
 		sctp_ulpevent_read_sndrcvinfo(event, msg);
 
 	err = copied;
@@ -2304,21 +2304,28 @@ static int sctp_setsockopt_disable_fragments(struct sock *sk,
 static int sctp_setsockopt_events(struct sock *sk, char __user *optval,
 				  unsigned int optlen)
 {
-	struct sctp_association *asoc;
-	struct sctp_ulpevent *event;
+	struct sctp_event_subscribe subscribe;
+	__u8 *sn_type = (__u8 *)&subscribe;
+	struct sctp_sock *sp = sctp_sk(sk);
+	int i;
 
 	if (optlen > sizeof(struct sctp_event_subscribe))
 		return -EINVAL;
-	if (copy_from_user(&sctp_sk(sk)->subscribe, optval, optlen))
+
+	if (copy_from_user(&subscribe, optval, optlen))
 		return -EFAULT;
 
+	for (i = 0; i < optlen; i++)
+		sctp_ulpevent_type_set(&sp->subscribe, SCTP_SN_TYPE_BASE + i,
+				       sn_type[i]);
+
 	/* At the time when a user app subscribes to SCTP_SENDER_DRY_EVENT,
 	 * if there is no data to be sent or retransmit, the stack will
 	 * immediately send up this notification.
 	 */
-	if (sctp_ulpevent_type_enabled(SCTP_SENDER_DRY_EVENT,
-				       &sctp_sk(sk)->subscribe)) {
-		asoc = sctp_id2assoc(sk, 0);
+	if (sctp_ulpevent_type_enabled(sp->subscribe, SCTP_SENDER_DRY_EVENT)) {
+		struct sctp_association *asoc = sctp_id2assoc(sk, 0);
+		struct sctp_ulpevent *event;
 
 		if (asoc && sctp_outq_is_empty(&asoc->outqueue)) {
 			event = sctp_ulpevent_make_sender_dry_event(asoc,
@@ -4722,7 +4729,7 @@ static int sctp_init_sock(struct sock *sk)
 	/* Initialize default event subscriptions. By default, all the
 	 * options are off.
 	 */
-	memset(&sp->subscribe, 0, sizeof(struct sctp_event_subscribe));
+	sp->subscribe = 0;
 
 	/* Default Peer Address Parameters.  These defaults can
 	 * be modified via SCTP_PEER_ADDR_PARAMS
@@ -5267,14 +5274,24 @@ static int sctp_getsockopt_disable_fragments(struct sock *sk, int len,
 static int sctp_getsockopt_events(struct sock *sk, int len, char __user *optval,
 				  int __user *optlen)
 {
+	struct sctp_event_subscribe subscribe;
+	__u8 *sn_type = (__u8 *)&subscribe;
+	int i;
+
 	if (len = 0)
 		return -EINVAL;
 	if (len > sizeof(struct sctp_event_subscribe))
 		len = sizeof(struct sctp_event_subscribe);
 	if (put_user(len, optlen))
 		return -EFAULT;
-	if (copy_to_user(optval, &sctp_sk(sk)->subscribe, len))
+
+	for (i = 0; i <= len; i++)
+		sn_type[i] = sctp_ulpevent_type_enabled(sctp_sk(sk)->subscribe,
+							SCTP_SN_TYPE_BASE + i);
+
+	if (copy_to_user(optval, &subscribe, len))
 		return -EFAULT;
+
 	return 0;
 }
 
diff --git a/net/sctp/stream_interleave.c b/net/sctp/stream_interleave.c
index 2b499a8..ceef5a3 100644
--- a/net/sctp/stream_interleave.c
+++ b/net/sctp/stream_interleave.c
@@ -503,7 +503,7 @@ static int sctp_enqueue_event(struct sctp_ulpq *ulpq,
 		sk_incoming_cpu_update(sk);
 	}
 
-	if (!sctp_ulpevent_is_enabled(event, &sp->subscribe))
+	if (!sctp_ulpevent_is_enabled(event, sp->subscribe))
 		goto out_free;
 
 	if (skb_list)
@@ -992,10 +992,11 @@ static void sctp_intl_stream_abort_pd(struct sctp_ulpq *ulpq, __u16 sid,
 				      __u32 mid, __u16 flags, gfp_t gfp)
 {
 	struct sock *sk = ulpq->asoc->base.sk;
+	struct sctp_sock *sp = sctp_sk(sk);
 	struct sctp_ulpevent *ev = NULL;
 
-	if (!sctp_ulpevent_type_enabled(SCTP_PARTIAL_DELIVERY_EVENT,
-					&sctp_sk(sk)->subscribe))
+	if (!sctp_ulpevent_type_enabled(sp->subscribe,
+					SCTP_PARTIAL_DELIVERY_EVENT))
 		return;
 
 	ev = sctp_ulpevent_make_pdapi(ulpq->asoc, SCTP_PARTIAL_DELIVERY_ABORTED,
@@ -1003,8 +1004,8 @@ static void sctp_intl_stream_abort_pd(struct sctp_ulpq *ulpq, __u16 sid,
 	if (ev) {
 		__skb_queue_tail(&sk->sk_receive_queue, sctp_event2skb(ev));
 
-		if (!sctp_sk(sk)->data_ready_signalled) {
-			sctp_sk(sk)->data_ready_signalled = 1;
+		if (!sp->data_ready_signalled) {
+			sp->data_ready_signalled = 1;
 			sk->sk_data_ready(sk);
 		}
 	}
diff --git a/net/sctp/ulpqueue.c b/net/sctp/ulpqueue.c
index 331cc73..b36dd90 100644
--- a/net/sctp/ulpqueue.c
+++ b/net/sctp/ulpqueue.c
@@ -219,7 +219,7 @@ int sctp_ulpq_tail_event(struct sctp_ulpq *ulpq, struct sctp_ulpevent *event)
 		sk_incoming_cpu_update(sk);
 	}
 	/* Check if the user wishes to receive this event.  */
-	if (!sctp_ulpevent_is_enabled(event, &sp->subscribe))
+	if (!sctp_ulpevent_is_enabled(event, sp->subscribe))
 		goto out_free;
 
 	/* If we are in partial delivery mode, post to the lobby until
@@ -1129,16 +1129,16 @@ void sctp_ulpq_renege(struct sctp_ulpq *ulpq, struct sctp_chunk *chunk,
 void sctp_ulpq_abort_pd(struct sctp_ulpq *ulpq, gfp_t gfp)
 {
 	struct sctp_ulpevent *ev = NULL;
-	struct sock *sk;
 	struct sctp_sock *sp;
+	struct sock *sk;
 
 	if (!ulpq->pd_mode)
 		return;
 
 	sk = ulpq->asoc->base.sk;
 	sp = sctp_sk(sk);
-	if (sctp_ulpevent_type_enabled(SCTP_PARTIAL_DELIVERY_EVENT,
-				       &sctp_sk(sk)->subscribe))
+	if (sctp_ulpevent_type_enabled(sp->subscribe,
+				       SCTP_PARTIAL_DELIVERY_EVENT))
 		ev = sctp_ulpevent_make_pdapi(ulpq->asoc,
 					      SCTP_PARTIAL_DELIVERY_ABORTED,
 					      0, 0, 0, gfp);
-- 
2.1.0

[PATCHv2 net-next 2/4] sctp: add subscribe per asoc

[Xin Long]

The member subscribe should be per asoc, so that sockopt SCTP_EVENT
in the next patch can subscribe a event from one asoc only.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
---
 include/net/sctp/structs.h   | 2 ++
 net/sctp/associola.c         | 2 ++
 net/sctp/chunk.c             | 6 ++----
 net/sctp/socket.c            | 6 +++++-
 net/sctp/stream_interleave.c | 7 ++++---
 net/sctp/ulpqueue.c          | 4 ++--
 6 files changed, 17 insertions(+), 10 deletions(-)

diff --git a/include/net/sctp/structs.h b/include/net/sctp/structs.h
index bc7808a..7eaa294 100644
--- a/include/net/sctp/structs.h
+++ b/include/net/sctp/structs.h
@@ -2077,6 +2077,8 @@ struct sctp_association {
 
 	int sent_cnt_removable;
 
+	__u16 subscribe;
+
 	__u64 abandoned_unsent[SCTP_PR_INDEX(MAX) + 1];
 	__u64 abandoned_sent[SCTP_PR_INDEX(MAX) + 1];
 };
diff --git a/net/sctp/associola.c b/net/sctp/associola.c
index 6a28b96..685c7ef 100644
--- a/net/sctp/associola.c
+++ b/net/sctp/associola.c
@@ -135,6 +135,8 @@ static struct sctp_association *sctp_association_init(
 	 */
 	asoc->max_burst = sp->max_burst;
 
+	asoc->subscribe = sp->subscribe;
+
 	/* initialize association timers */
 	asoc->timeouts[SCTP_EVENT_TIMEOUT_T1_COOKIE] = asoc->rto_initial;
 	asoc->timeouts[SCTP_EVENT_TIMEOUT_T1_INIT] = asoc->rto_initial;
diff --git a/net/sctp/chunk.c b/net/sctp/chunk.c
index 6c761af..0b203b8 100644
--- a/net/sctp/chunk.c
+++ b/net/sctp/chunk.c
@@ -86,11 +86,10 @@ void sctp_datamsg_free(struct sctp_datamsg *msg)
 /* Final destructruction of datamsg memory. */
 static void sctp_datamsg_destroy(struct sctp_datamsg *msg)
 {
+	struct sctp_association *asoc = NULL;
 	struct list_head *pos, *temp;
 	struct sctp_chunk *chunk;
-	struct sctp_sock *sp;
 	struct sctp_ulpevent *ev;
-	struct sctp_association *asoc = NULL;
 	int error = 0, notify;
 
 	/* If we failed, we may need to notify. */
@@ -108,8 +107,7 @@ static void sctp_datamsg_destroy(struct sctp_datamsg *msg)
 			else
 				error = asoc->outqueue.error;
 
-			sp = sctp_sk(asoc->base.sk);
-			notify = sctp_ulpevent_type_enabled(sp->subscribe,
+			notify = sctp_ulpevent_type_enabled(asoc->subscribe,
 							    SCTP_SEND_FAILED);
 		}
 
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index 48e0b45..789008d 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -2307,6 +2307,7 @@ static int sctp_setsockopt_events(struct sock *sk, char __user *optval,
 	struct sctp_event_subscribe subscribe;
 	__u8 *sn_type = (__u8 *)&subscribe;
 	struct sctp_sock *sp = sctp_sk(sk);
+	struct sctp_association *asoc;
 	int i;
 
 	if (optlen > sizeof(struct sctp_event_subscribe))
@@ -2319,14 +2320,17 @@ static int sctp_setsockopt_events(struct sock *sk, char __user *optval,
 		sctp_ulpevent_type_set(&sp->subscribe, SCTP_SN_TYPE_BASE + i,
 				       sn_type[i]);
 
+	list_for_each_entry(asoc, &sp->ep->asocs, asocs)
+		asoc->subscribe = sctp_sk(sk)->subscribe;
+
 	/* At the time when a user app subscribes to SCTP_SENDER_DRY_EVENT,
 	 * if there is no data to be sent or retransmit, the stack will
 	 * immediately send up this notification.
 	 */
 	if (sctp_ulpevent_type_enabled(sp->subscribe, SCTP_SENDER_DRY_EVENT)) {
-		struct sctp_association *asoc = sctp_id2assoc(sk, 0);
 		struct sctp_ulpevent *event;
 
+		asoc = sctp_id2assoc(sk, 0);
 		if (asoc && sctp_outq_is_empty(&asoc->outqueue)) {
 			event = sctp_ulpevent_make_sender_dry_event(asoc,
 					GFP_USER | __GFP_NOWARN);
diff --git a/net/sctp/stream_interleave.c b/net/sctp/stream_interleave.c
index ceef5a3..a6bf215 100644
--- a/net/sctp/stream_interleave.c
+++ b/net/sctp/stream_interleave.c
@@ -503,7 +503,7 @@ static int sctp_enqueue_event(struct sctp_ulpq *ulpq,
 		sk_incoming_cpu_update(sk);
 	}
 
-	if (!sctp_ulpevent_is_enabled(event, sp->subscribe))
+	if (!sctp_ulpevent_is_enabled(event, ulpq->asoc->subscribe))
 		goto out_free;
 
 	if (skb_list)
@@ -992,16 +992,17 @@ static void sctp_intl_stream_abort_pd(struct sctp_ulpq *ulpq, __u16 sid,
 				      __u32 mid, __u16 flags, gfp_t gfp)
 {
 	struct sock *sk = ulpq->asoc->base.sk;
-	struct sctp_sock *sp = sctp_sk(sk);
 	struct sctp_ulpevent *ev = NULL;
 
-	if (!sctp_ulpevent_type_enabled(sp->subscribe,
+	if (!sctp_ulpevent_type_enabled(ulpq->asoc->subscribe,
 					SCTP_PARTIAL_DELIVERY_EVENT))
 		return;
 
 	ev = sctp_ulpevent_make_pdapi(ulpq->asoc, SCTP_PARTIAL_DELIVERY_ABORTED,
 				      sid, mid, flags, gfp);
 	if (ev) {
+		struct sctp_sock *sp = sctp_sk(sk);
+
 		__skb_queue_tail(&sk->sk_receive_queue, sctp_event2skb(ev));
 
 		if (!sp->data_ready_signalled) {
diff --git a/net/sctp/ulpqueue.c b/net/sctp/ulpqueue.c
index b36dd90..5dde921 100644
--- a/net/sctp/ulpqueue.c
+++ b/net/sctp/ulpqueue.c
@@ -219,7 +219,7 @@ int sctp_ulpq_tail_event(struct sctp_ulpq *ulpq, struct sctp_ulpevent *event)
 		sk_incoming_cpu_update(sk);
 	}
 	/* Check if the user wishes to receive this event.  */
-	if (!sctp_ulpevent_is_enabled(event, sp->subscribe))
+	if (!sctp_ulpevent_is_enabled(event, ulpq->asoc->subscribe))
 		goto out_free;
 
 	/* If we are in partial delivery mode, post to the lobby until
@@ -1137,7 +1137,7 @@ void sctp_ulpq_abort_pd(struct sctp_ulpq *ulpq, gfp_t gfp)
 
 	sk = ulpq->asoc->base.sk;
 	sp = sctp_sk(sk);
-	if (sctp_ulpevent_type_enabled(sp->subscribe,
+	if (sctp_ulpevent_type_enabled(ulpq->asoc->subscribe,
 				       SCTP_PARTIAL_DELIVERY_EVENT))
 		ev = sctp_ulpevent_make_pdapi(ulpq->asoc,
 					      SCTP_PARTIAL_DELIVERY_ABORTED,
-- 
2.1.0

[PATCHv2 net-next 2/4] sctp: add subscribe per asoc

[Xin Long]

The member subscribe should be per asoc, so that sockopt SCTP_EVENT
in the next patch can subscribe a event from one asoc only.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
---
 include/net/sctp/structs.h   | 2 ++
 net/sctp/associola.c         | 2 ++
 net/sctp/chunk.c             | 6 ++----
 net/sctp/socket.c            | 6 +++++-
 net/sctp/stream_interleave.c | 7 ++++---
 net/sctp/ulpqueue.c          | 4 ++--
 6 files changed, 17 insertions(+), 10 deletions(-)

diff --git a/include/net/sctp/structs.h b/include/net/sctp/structs.h
index bc7808a..7eaa294 100644
--- a/include/net/sctp/structs.h
+++ b/include/net/sctp/structs.h
@@ -2077,6 +2077,8 @@ struct sctp_association {
 
 	int sent_cnt_removable;
 
+	__u16 subscribe;
+
 	__u64 abandoned_unsent[SCTP_PR_INDEX(MAX) + 1];
 	__u64 abandoned_sent[SCTP_PR_INDEX(MAX) + 1];
 };
diff --git a/net/sctp/associola.c b/net/sctp/associola.c
index 6a28b96..685c7ef 100644
--- a/net/sctp/associola.c
+++ b/net/sctp/associola.c
@@ -135,6 +135,8 @@ static struct sctp_association *sctp_association_init(
 	 */
 	asoc->max_burst = sp->max_burst;
 
+	asoc->subscribe = sp->subscribe;
+
 	/* initialize association timers */
 	asoc->timeouts[SCTP_EVENT_TIMEOUT_T1_COOKIE] = asoc->rto_initial;
 	asoc->timeouts[SCTP_EVENT_TIMEOUT_T1_INIT] = asoc->rto_initial;
diff --git a/net/sctp/chunk.c b/net/sctp/chunk.c
index 6c761af..0b203b8 100644
--- a/net/sctp/chunk.c
+++ b/net/sctp/chunk.c
@@ -86,11 +86,10 @@ void sctp_datamsg_free(struct sctp_datamsg *msg)
 /* Final destructruction of datamsg memory. */
 static void sctp_datamsg_destroy(struct sctp_datamsg *msg)
 {
+	struct sctp_association *asoc = NULL;
 	struct list_head *pos, *temp;
 	struct sctp_chunk *chunk;
-	struct sctp_sock *sp;
 	struct sctp_ulpevent *ev;
-	struct sctp_association *asoc = NULL;
 	int error = 0, notify;
 
 	/* If we failed, we may need to notify. */
@@ -108,8 +107,7 @@ static void sctp_datamsg_destroy(struct sctp_datamsg *msg)
 			else
 				error = asoc->outqueue.error;
 
-			sp = sctp_sk(asoc->base.sk);
-			notify = sctp_ulpevent_type_enabled(sp->subscribe,
+			notify = sctp_ulpevent_type_enabled(asoc->subscribe,
 							    SCTP_SEND_FAILED);
 		}
 
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index 48e0b45..789008d 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -2307,6 +2307,7 @@ static int sctp_setsockopt_events(struct sock *sk, char __user *optval,
 	struct sctp_event_subscribe subscribe;
 	__u8 *sn_type = (__u8 *)&subscribe;
 	struct sctp_sock *sp = sctp_sk(sk);
+	struct sctp_association *asoc;
 	int i;
 
 	if (optlen > sizeof(struct sctp_event_subscribe))
@@ -2319,14 +2320,17 @@ static int sctp_setsockopt_events(struct sock *sk, char __user *optval,
 		sctp_ulpevent_type_set(&sp->subscribe, SCTP_SN_TYPE_BASE + i,
 				       sn_type[i]);
 
+	list_for_each_entry(asoc, &sp->ep->asocs, asocs)
+		asoc->subscribe = sctp_sk(sk)->subscribe;
+
 	/* At the time when a user app subscribes to SCTP_SENDER_DRY_EVENT,
 	 * if there is no data to be sent or retransmit, the stack will
 	 * immediately send up this notification.
 	 */
 	if (sctp_ulpevent_type_enabled(sp->subscribe, SCTP_SENDER_DRY_EVENT)) {
-		struct sctp_association *asoc = sctp_id2assoc(sk, 0);
 		struct sctp_ulpevent *event;
 
+		asoc = sctp_id2assoc(sk, 0);
 		if (asoc && sctp_outq_is_empty(&asoc->outqueue)) {
 			event = sctp_ulpevent_make_sender_dry_event(asoc,
 					GFP_USER | __GFP_NOWARN);
diff --git a/net/sctp/stream_interleave.c b/net/sctp/stream_interleave.c
index ceef5a3..a6bf215 100644
--- a/net/sctp/stream_interleave.c
+++ b/net/sctp/stream_interleave.c
@@ -503,7 +503,7 @@ static int sctp_enqueue_event(struct sctp_ulpq *ulpq,
 		sk_incoming_cpu_update(sk);
 	}
 
-	if (!sctp_ulpevent_is_enabled(event, sp->subscribe))
+	if (!sctp_ulpevent_is_enabled(event, ulpq->asoc->subscribe))
 		goto out_free;
 
 	if (skb_list)
@@ -992,16 +992,17 @@ static void sctp_intl_stream_abort_pd(struct sctp_ulpq *ulpq, __u16 sid,
 				      __u32 mid, __u16 flags, gfp_t gfp)
 {
 	struct sock *sk = ulpq->asoc->base.sk;
-	struct sctp_sock *sp = sctp_sk(sk);
 	struct sctp_ulpevent *ev = NULL;
 
-	if (!sctp_ulpevent_type_enabled(sp->subscribe,
+	if (!sctp_ulpevent_type_enabled(ulpq->asoc->subscribe,
 					SCTP_PARTIAL_DELIVERY_EVENT))
 		return;
 
 	ev = sctp_ulpevent_make_pdapi(ulpq->asoc, SCTP_PARTIAL_DELIVERY_ABORTED,
 				      sid, mid, flags, gfp);
 	if (ev) {
+		struct sctp_sock *sp = sctp_sk(sk);
+
 		__skb_queue_tail(&sk->sk_receive_queue, sctp_event2skb(ev));
 
 		if (!sp->data_ready_signalled) {
diff --git a/net/sctp/ulpqueue.c b/net/sctp/ulpqueue.c
index b36dd90..5dde921 100644
--- a/net/sctp/ulpqueue.c
+++ b/net/sctp/ulpqueue.c
@@ -219,7 +219,7 @@ int sctp_ulpq_tail_event(struct sctp_ulpq *ulpq, struct sctp_ulpevent *event)
 		sk_incoming_cpu_update(sk);
 	}
 	/* Check if the user wishes to receive this event.  */
-	if (!sctp_ulpevent_is_enabled(event, sp->subscribe))
+	if (!sctp_ulpevent_is_enabled(event, ulpq->asoc->subscribe))
 		goto out_free;
 
 	/* If we are in partial delivery mode, post to the lobby until
@@ -1137,7 +1137,7 @@ void sctp_ulpq_abort_pd(struct sctp_ulpq *ulpq, gfp_t gfp)
 
 	sk = ulpq->asoc->base.sk;
 	sp = sctp_sk(sk);
-	if (sctp_ulpevent_type_enabled(sp->subscribe,
+	if (sctp_ulpevent_type_enabled(ulpq->asoc->subscribe,
 				       SCTP_PARTIAL_DELIVERY_EVENT))
 		ev = sctp_ulpevent_make_pdapi(ulpq->asoc,
 					      SCTP_PARTIAL_DELIVERY_ABORTED,
-- 
2.1.0

[PATCHv2 net-next 3/4] sctp: rename enum sctp_event to sctp_event_type

[Xin Long]

sctp_event is a structure name defined in RFC for sockopt
SCTP_EVENT. To avoid the conflict, rename it.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
---
 include/net/sctp/constants.h |  2 +-
 include/net/sctp/sm.h        |  4 ++--
 net/sctp/primitive.c         |  2 +-
 net/sctp/sm_sideeffect.c     | 12 ++++++------
 net/sctp/sm_statetable.c     |  2 +-
 5 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/include/net/sctp/constants.h b/include/net/sctp/constants.h
index 8dadc74..4588bdc 100644
--- a/include/net/sctp/constants.h
+++ b/include/net/sctp/constants.h
@@ -71,7 +71,7 @@ enum { SCTP_DEFAULT_INSTREAMS = SCTP_MAX_STREAM };
 					 SCTP_NUM_AUTH_CHUNK_TYPES)
 
 /* These are the different flavours of event.  */
-enum sctp_event {
+enum sctp_event_type {
 	SCTP_EVENT_T_CHUNK = 1,
 	SCTP_EVENT_T_TIMEOUT,
 	SCTP_EVENT_T_OTHER,
diff --git a/include/net/sctp/sm.h b/include/net/sctp/sm.h
index 9e3d327..24825a8 100644
--- a/include/net/sctp/sm.h
+++ b/include/net/sctp/sm.h
@@ -173,7 +173,7 @@ sctp_state_fn_t sctp_sf_autoclose_timer_expire;
 __u8 sctp_get_chunk_type(struct sctp_chunk *chunk);
 const struct sctp_sm_table_entry *sctp_sm_lookup_event(
 					struct net *net,
-					enum sctp_event event_type,
+					enum sctp_event_type event_type,
 					enum sctp_state state,
 					union sctp_subtype event_subtype);
 int sctp_chunk_iif(const struct sctp_chunk *);
@@ -313,7 +313,7 @@ struct sctp_chunk *sctp_process_strreset_resp(
 
 /* Prototypes for statetable processing. */
 
-int sctp_do_sm(struct net *net, enum sctp_event event_type,
+int sctp_do_sm(struct net *net, enum sctp_event_type event_type,
 	       union sctp_subtype subtype, enum sctp_state state,
 	       struct sctp_endpoint *ep, struct sctp_association *asoc,
 	       void *event_arg, gfp_t gfp);
diff --git a/net/sctp/primitive.c b/net/sctp/primitive.c
index c0817f7a..a8c4c33 100644
--- a/net/sctp/primitive.c
+++ b/net/sctp/primitive.c
@@ -53,7 +53,7 @@
 int sctp_primitive_ ## name(struct net *net, struct sctp_association *asoc, \
 			    void *arg) { \
 	int error = 0; \
-	enum sctp_event event_type; union sctp_subtype subtype; \
+	enum sctp_event_type event_type; union sctp_subtype subtype; \
 	enum sctp_state state; \
 	struct sctp_endpoint *ep; \
 	\
diff --git a/net/sctp/sm_sideeffect.c b/net/sctp/sm_sideeffect.c
index 85d3930..1d143bc 100644
--- a/net/sctp/sm_sideeffect.c
+++ b/net/sctp/sm_sideeffect.c
@@ -52,7 +52,7 @@
 #include <net/sctp/sm.h>
 #include <net/sctp/stream_sched.h>
 
-static int sctp_cmd_interpreter(enum sctp_event event_type,
+static int sctp_cmd_interpreter(enum sctp_event_type event_type,
 				union sctp_subtype subtype,
 				enum sctp_state state,
 				struct sctp_endpoint *ep,
@@ -61,7 +61,7 @@ static int sctp_cmd_interpreter(enum sctp_event event_type,
 				enum sctp_disposition status,
 				struct sctp_cmd_seq *commands,
 				gfp_t gfp);
-static int sctp_side_effects(enum sctp_event event_type,
+static int sctp_side_effects(enum sctp_event_type event_type,
 			     union sctp_subtype subtype,
 			     enum sctp_state state,
 			     struct sctp_endpoint *ep,
@@ -623,7 +623,7 @@ static void sctp_cmd_init_failed(struct sctp_cmd_seq *commands,
 /* Worker routine to handle SCTP_CMD_ASSOC_FAILED.  */
 static void sctp_cmd_assoc_failed(struct sctp_cmd_seq *commands,
 				  struct sctp_association *asoc,
-				  enum sctp_event event_type,
+				  enum sctp_event_type event_type,
 				  union sctp_subtype subtype,
 				  struct sctp_chunk *chunk,
 				  unsigned int error)
@@ -1162,7 +1162,7 @@ static void sctp_cmd_send_asconf(struct sctp_association *asoc)
  * If you want to understand all of lksctp, this is a
  * good place to start.
  */
-int sctp_do_sm(struct net *net, enum sctp_event event_type,
+int sctp_do_sm(struct net *net, enum sctp_event_type event_type,
 	       union sctp_subtype subtype, enum sctp_state state,
 	       struct sctp_endpoint *ep, struct sctp_association *asoc,
 	       void *event_arg, gfp_t gfp)
@@ -1199,7 +1199,7 @@ int sctp_do_sm(struct net *net, enum sctp_event event_type,
 /*****************************************************************
  * This the master state function side effect processing function.
  *****************************************************************/
-static int sctp_side_effects(enum sctp_event event_type,
+static int sctp_side_effects(enum sctp_event_type event_type,
 			     union sctp_subtype subtype,
 			     enum sctp_state state,
 			     struct sctp_endpoint *ep,
@@ -1285,7 +1285,7 @@ static int sctp_side_effects(enum sctp_event event_type,
  ********************************************************************/
 
 /* This is the side-effect interpreter.  */
-static int sctp_cmd_interpreter(enum sctp_event event_type,
+static int sctp_cmd_interpreter(enum sctp_event_type event_type,
 				union sctp_subtype subtype,
 				enum sctp_state state,
 				struct sctp_endpoint *ep,
diff --git a/net/sctp/sm_statetable.c b/net/sctp/sm_statetable.c
index 691d9dc..d239b94 100644
--- a/net/sctp/sm_statetable.c
+++ b/net/sctp/sm_statetable.c
@@ -79,7 +79,7 @@ static const struct sctp_sm_table_entry bug = {
 
 const struct sctp_sm_table_entry *sctp_sm_lookup_event(
 					struct net *net,
-					enum sctp_event event_type,
+					enum sctp_event_type event_type,
 					enum sctp_state state,
 					union sctp_subtype event_subtype)
 {
-- 
2.1.0

[PATCHv2 net-next 3/4] sctp: rename enum sctp_event to sctp_event_type

[Xin Long]

sctp_event is a structure name defined in RFC for sockopt
SCTP_EVENT. To avoid the conflict, rename it.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
---
 include/net/sctp/constants.h |  2 +-
 include/net/sctp/sm.h        |  4 ++--
 net/sctp/primitive.c         |  2 +-
 net/sctp/sm_sideeffect.c     | 12 ++++++------
 net/sctp/sm_statetable.c     |  2 +-
 5 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/include/net/sctp/constants.h b/include/net/sctp/constants.h
index 8dadc74..4588bdc 100644
--- a/include/net/sctp/constants.h
+++ b/include/net/sctp/constants.h
@@ -71,7 +71,7 @@ enum { SCTP_DEFAULT_INSTREAMS = SCTP_MAX_STREAM };
 					 SCTP_NUM_AUTH_CHUNK_TYPES)
 
 /* These are the different flavours of event.  */
-enum sctp_event {
+enum sctp_event_type {
 	SCTP_EVENT_T_CHUNK = 1,
 	SCTP_EVENT_T_TIMEOUT,
 	SCTP_EVENT_T_OTHER,
diff --git a/include/net/sctp/sm.h b/include/net/sctp/sm.h
index 9e3d327..24825a8 100644
--- a/include/net/sctp/sm.h
+++ b/include/net/sctp/sm.h
@@ -173,7 +173,7 @@ sctp_state_fn_t sctp_sf_autoclose_timer_expire;
 __u8 sctp_get_chunk_type(struct sctp_chunk *chunk);
 const struct sctp_sm_table_entry *sctp_sm_lookup_event(
 					struct net *net,
-					enum sctp_event event_type,
+					enum sctp_event_type event_type,
 					enum sctp_state state,
 					union sctp_subtype event_subtype);
 int sctp_chunk_iif(const struct sctp_chunk *);
@@ -313,7 +313,7 @@ struct sctp_chunk *sctp_process_strreset_resp(
 
 /* Prototypes for statetable processing. */
 
-int sctp_do_sm(struct net *net, enum sctp_event event_type,
+int sctp_do_sm(struct net *net, enum sctp_event_type event_type,
 	       union sctp_subtype subtype, enum sctp_state state,
 	       struct sctp_endpoint *ep, struct sctp_association *asoc,
 	       void *event_arg, gfp_t gfp);
diff --git a/net/sctp/primitive.c b/net/sctp/primitive.c
index c0817f7a..a8c4c33 100644
--- a/net/sctp/primitive.c
+++ b/net/sctp/primitive.c
@@ -53,7 +53,7 @@
 int sctp_primitive_ ## name(struct net *net, struct sctp_association *asoc, \
 			    void *arg) { \
 	int error = 0; \
-	enum sctp_event event_type; union sctp_subtype subtype; \
+	enum sctp_event_type event_type; union sctp_subtype subtype; \
 	enum sctp_state state; \
 	struct sctp_endpoint *ep; \
 	\
diff --git a/net/sctp/sm_sideeffect.c b/net/sctp/sm_sideeffect.c
index 85d3930..1d143bc 100644
--- a/net/sctp/sm_sideeffect.c
+++ b/net/sctp/sm_sideeffect.c
@@ -52,7 +52,7 @@
 #include <net/sctp/sm.h>
 #include <net/sctp/stream_sched.h>
 
-static int sctp_cmd_interpreter(enum sctp_event event_type,
+static int sctp_cmd_interpreter(enum sctp_event_type event_type,
 				union sctp_subtype subtype,
 				enum sctp_state state,
 				struct sctp_endpoint *ep,
@@ -61,7 +61,7 @@ static int sctp_cmd_interpreter(enum sctp_event event_type,
 				enum sctp_disposition status,
 				struct sctp_cmd_seq *commands,
 				gfp_t gfp);
-static int sctp_side_effects(enum sctp_event event_type,
+static int sctp_side_effects(enum sctp_event_type event_type,
 			     union sctp_subtype subtype,
 			     enum sctp_state state,
 			     struct sctp_endpoint *ep,
@@ -623,7 +623,7 @@ static void sctp_cmd_init_failed(struct sctp_cmd_seq *commands,
 /* Worker routine to handle SCTP_CMD_ASSOC_FAILED.  */
 static void sctp_cmd_assoc_failed(struct sctp_cmd_seq *commands,
 				  struct sctp_association *asoc,
-				  enum sctp_event event_type,
+				  enum sctp_event_type event_type,
 				  union sctp_subtype subtype,
 				  struct sctp_chunk *chunk,
 				  unsigned int error)
@@ -1162,7 +1162,7 @@ static void sctp_cmd_send_asconf(struct sctp_association *asoc)
  * If you want to understand all of lksctp, this is a
  * good place to start.
  */
-int sctp_do_sm(struct net *net, enum sctp_event event_type,
+int sctp_do_sm(struct net *net, enum sctp_event_type event_type,
 	       union sctp_subtype subtype, enum sctp_state state,
 	       struct sctp_endpoint *ep, struct sctp_association *asoc,
 	       void *event_arg, gfp_t gfp)
@@ -1199,7 +1199,7 @@ int sctp_do_sm(struct net *net, enum sctp_event event_type,
 /*****************************************************************
  * This the master state function side effect processing function.
  *****************************************************************/
-static int sctp_side_effects(enum sctp_event event_type,
+static int sctp_side_effects(enum sctp_event_type event_type,
 			     union sctp_subtype subtype,
 			     enum sctp_state state,
 			     struct sctp_endpoint *ep,
@@ -1285,7 +1285,7 @@ static int sctp_side_effects(enum sctp_event event_type,
  ********************************************************************/
 
 /* This is the side-effect interpreter.  */
-static int sctp_cmd_interpreter(enum sctp_event event_type,
+static int sctp_cmd_interpreter(enum sctp_event_type event_type,
 				union sctp_subtype subtype,
 				enum sctp_state state,
 				struct sctp_endpoint *ep,
diff --git a/net/sctp/sm_statetable.c b/net/sctp/sm_statetable.c
index 691d9dc..d239b94 100644
--- a/net/sctp/sm_statetable.c
+++ b/net/sctp/sm_statetable.c
@@ -79,7 +79,7 @@ static const struct sctp_sm_table_entry bug = {
 
 const struct sctp_sm_table_entry *sctp_sm_lookup_event(
 					struct net *net,
-					enum sctp_event event_type,
+					enum sctp_event_type event_type,
 					enum sctp_state state,
 					union sctp_subtype event_subtype)
 {
-- 
2.1.0

[PATCHv2 net-next 4/4] sctp: add sockopt SCTP_EVENT

[Xin Long]

This patch adds sockopt SCTP_EVENT described in rfc6525#section-6.2.
With this sockopt users can subscribe to an event from a specified
asoc.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
---
 include/uapi/linux/sctp.h |  7 ++++
 net/sctp/socket.c         | 89 +++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 96 insertions(+)

diff --git a/include/uapi/linux/sctp.h b/include/uapi/linux/sctp.h
index 66afa5b..d584073 100644
--- a/include/uapi/linux/sctp.h
+++ b/include/uapi/linux/sctp.h
@@ -129,6 +129,7 @@ typedef __s32 sctp_assoc_t;
 #define SCTP_STREAM_SCHEDULER_VALUE	124
 #define SCTP_INTERLEAVING_SUPPORTED	125
 #define SCTP_SENDMSG_CONNECT	126
+#define SCTP_EVENT	127
 
 /* PR-SCTP policies */
 #define SCTP_PR_SCTP_NONE	0x0000
@@ -1154,6 +1155,12 @@ struct sctp_add_streams {
 	uint16_t sas_outstrms;
 };
 
+struct sctp_event {
+	sctp_assoc_t se_assoc_id;
+	uint16_t se_type;
+	uint8_t se_on;
+};
+
 /* SCTP Stream schedulers */
 enum sctp_sched_type {
 	SCTP_SS_FCFS,
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index 789008d..1451211 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -4288,6 +4288,57 @@ static int sctp_setsockopt_reuse_port(struct sock *sk, char __user *optval,
 	return 0;
 }
 
+static int sctp_setsockopt_event(struct sock *sk, char __user *optval,
+				 unsigned int optlen)
+{
+	struct sctp_association *asoc;
+	struct sctp_ulpevent *event;
+	struct sctp_event param;
+	int retval = 0;
+
+	if (optlen < sizeof(param)) {
+		retval = -EINVAL;
+		goto out;
+	}
+
+	optlen = sizeof(param);
+	if (copy_from_user(&param, optval, optlen)) {
+		retval = -EFAULT;
+		goto out;
+	}
+
+	if (param.se_type < SCTP_SN_TYPE_BASE ||
+	    param.se_type > SCTP_SN_TYPE_MAX) {
+		retval = -EINVAL;
+		goto out;
+	}
+
+	asoc = sctp_id2assoc(sk, param.se_assoc_id);
+	if (!asoc) {
+		sctp_ulpevent_type_set(&sctp_sk(sk)->subscribe,
+				       param.se_type, param.se_on);
+		goto out;
+	}
+
+	sctp_ulpevent_type_set(&asoc->subscribe, param.se_type, param.se_on);
+
+	if (param.se_type == SCTP_SENDER_DRY_EVENT && param.se_on) {
+		if (sctp_outq_is_empty(&asoc->outqueue)) {
+			event = sctp_ulpevent_make_sender_dry_event(asoc,
+					GFP_USER | __GFP_NOWARN);
+			if (!event) {
+				retval = -ENOMEM;
+				goto out;
+			}
+
+			asoc->stream.si->enqueue_event(&asoc->ulpq, event);
+		}
+	}
+
+out:
+	return retval;
+}
+
 /* API 6.2 setsockopt(), getsockopt()
  *
  * Applications use setsockopt() and getsockopt() to set or retrieve
@@ -4485,6 +4536,9 @@ static int sctp_setsockopt(struct sock *sk, int level, int optname,
 	case SCTP_REUSE_PORT:
 		retval = sctp_setsockopt_reuse_port(sk, optval, optlen);
 		break;
+	case SCTP_EVENT:
+		retval = sctp_setsockopt_event(sk, optval, optlen);
+		break;
 	default:
 		retval = -ENOPROTOOPT;
 		break;
@@ -7430,6 +7484,38 @@ static int sctp_getsockopt_reuse_port(struct sock *sk, int len,
 	return 0;
 }
 
+static int sctp_getsockopt_event(struct sock *sk, int len, char __user *optval,
+				 int __user *optlen)
+{
+	struct sctp_association *asoc;
+	struct sctp_event param;
+	__u16 subscribe;
+
+	if (len < sizeof(param))
+		return -EINVAL;
+
+	len = sizeof(param);
+	if (copy_from_user(&param, optval, len))
+		return -EFAULT;
+
+	if (param.se_type < SCTP_SN_TYPE_BASE ||
+	    param.se_type > SCTP_SN_TYPE_MAX)
+		return -EINVAL;
+
+	asoc = sctp_id2assoc(sk, param.se_assoc_id);
+	subscribe = asoc ? asoc->subscribe : sctp_sk(sk)->subscribe;
+	param.se_on = sctp_ulpevent_type_enabled(subscribe, param.se_type);
+
+	if (put_user(len, optlen))
+		return -EFAULT;
+
+	if (copy_to_user(optval, &param, len))
+		return -EFAULT;
+
+	return 0;
+}
+
+
 static int sctp_getsockopt(struct sock *sk, int level, int optname,
 			   char __user *optval, int __user *optlen)
 {
@@ -7628,6 +7714,9 @@ static int sctp_getsockopt(struct sock *sk, int level, int optname,
 	case SCTP_REUSE_PORT:
 		retval = sctp_getsockopt_reuse_port(sk, len, optval, optlen);
 		break;
+	case SCTP_EVENT:
+		retval = sctp_getsockopt_event(sk, len, optval, optlen);
+		break;
 	default:
 		retval = -ENOPROTOOPT;
 		break;
-- 
2.1.0

[PATCHv2 net-next 4/4] sctp: add sockopt SCTP_EVENT

[Xin Long]

This patch adds sockopt SCTP_EVENT described in rfc6525#section-6.2.
With this sockopt users can subscribe to an event from a specified
asoc.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
---
 include/uapi/linux/sctp.h |  7 ++++
 net/sctp/socket.c         | 89 +++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 96 insertions(+)

diff --git a/include/uapi/linux/sctp.h b/include/uapi/linux/sctp.h
index 66afa5b..d584073 100644
--- a/include/uapi/linux/sctp.h
+++ b/include/uapi/linux/sctp.h
@@ -129,6 +129,7 @@ typedef __s32 sctp_assoc_t;
 #define SCTP_STREAM_SCHEDULER_VALUE	124
 #define SCTP_INTERLEAVING_SUPPORTED	125
 #define SCTP_SENDMSG_CONNECT	126
+#define SCTP_EVENT	127
 
 /* PR-SCTP policies */
 #define SCTP_PR_SCTP_NONE	0x0000
@@ -1154,6 +1155,12 @@ struct sctp_add_streams {
 	uint16_t sas_outstrms;
 };
 
+struct sctp_event {
+	sctp_assoc_t se_assoc_id;
+	uint16_t se_type;
+	uint8_t se_on;
+};
+
 /* SCTP Stream schedulers */
 enum sctp_sched_type {
 	SCTP_SS_FCFS,
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index 789008d..1451211 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -4288,6 +4288,57 @@ static int sctp_setsockopt_reuse_port(struct sock *sk, char __user *optval,
 	return 0;
 }
 
+static int sctp_setsockopt_event(struct sock *sk, char __user *optval,
+				 unsigned int optlen)
+{
+	struct sctp_association *asoc;
+	struct sctp_ulpevent *event;
+	struct sctp_event param;
+	int retval = 0;
+
+	if (optlen < sizeof(param)) {
+		retval = -EINVAL;
+		goto out;
+	}
+
+	optlen = sizeof(param);
+	if (copy_from_user(&param, optval, optlen)) {
+		retval = -EFAULT;
+		goto out;
+	}
+
+	if (param.se_type < SCTP_SN_TYPE_BASE ||
+	    param.se_type > SCTP_SN_TYPE_MAX) {
+		retval = -EINVAL;
+		goto out;
+	}
+
+	asoc = sctp_id2assoc(sk, param.se_assoc_id);
+	if (!asoc) {
+		sctp_ulpevent_type_set(&sctp_sk(sk)->subscribe,
+				       param.se_type, param.se_on);
+		goto out;
+	}
+
+	sctp_ulpevent_type_set(&asoc->subscribe, param.se_type, param.se_on);
+
+	if (param.se_type = SCTP_SENDER_DRY_EVENT && param.se_on) {
+		if (sctp_outq_is_empty(&asoc->outqueue)) {
+			event = sctp_ulpevent_make_sender_dry_event(asoc,
+					GFP_USER | __GFP_NOWARN);
+			if (!event) {
+				retval = -ENOMEM;
+				goto out;
+			}
+
+			asoc->stream.si->enqueue_event(&asoc->ulpq, event);
+		}
+	}
+
+out:
+	return retval;
+}
+
 /* API 6.2 setsockopt(), getsockopt()
  *
  * Applications use setsockopt() and getsockopt() to set or retrieve
@@ -4485,6 +4536,9 @@ static int sctp_setsockopt(struct sock *sk, int level, int optname,
 	case SCTP_REUSE_PORT:
 		retval = sctp_setsockopt_reuse_port(sk, optval, optlen);
 		break;
+	case SCTP_EVENT:
+		retval = sctp_setsockopt_event(sk, optval, optlen);
+		break;
 	default:
 		retval = -ENOPROTOOPT;
 		break;
@@ -7430,6 +7484,38 @@ static int sctp_getsockopt_reuse_port(struct sock *sk, int len,
 	return 0;
 }
 
+static int sctp_getsockopt_event(struct sock *sk, int len, char __user *optval,
+				 int __user *optlen)
+{
+	struct sctp_association *asoc;
+	struct sctp_event param;
+	__u16 subscribe;
+
+	if (len < sizeof(param))
+		return -EINVAL;
+
+	len = sizeof(param);
+	if (copy_from_user(&param, optval, len))
+		return -EFAULT;
+
+	if (param.se_type < SCTP_SN_TYPE_BASE ||
+	    param.se_type > SCTP_SN_TYPE_MAX)
+		return -EINVAL;
+
+	asoc = sctp_id2assoc(sk, param.se_assoc_id);
+	subscribe = asoc ? asoc->subscribe : sctp_sk(sk)->subscribe;
+	param.se_on = sctp_ulpevent_type_enabled(subscribe, param.se_type);
+
+	if (put_user(len, optlen))
+		return -EFAULT;
+
+	if (copy_to_user(optval, &param, len))
+		return -EFAULT;
+
+	return 0;
+}
+
+
 static int sctp_getsockopt(struct sock *sk, int level, int optname,
 			   char __user *optval, int __user *optlen)
 {
@@ -7628,6 +7714,9 @@ static int sctp_getsockopt(struct sock *sk, int level, int optname,
 	case SCTP_REUSE_PORT:
 		retval = sctp_getsockopt_reuse_port(sk, len, optval, optlen);
 		break;
+	case SCTP_EVENT:
+		retval = sctp_getsockopt_event(sk, len, optval, optlen);
+		break;
 	default:
 		retval = -ENOPROTOOPT;
 		break;
-- 
2.1.0

Re: [PATCHv2 net-next 1/4] sctp: define subscribe in sctp_sock as __u16

[Neil Horman]

On Tue, Nov 13, 2018 at 02:24:53PM +0800, Xin Long wrote:
>  
>  	/* Default Peer Address Parameters.  These defaults can
>  	 * be modified via SCTP_PEER_ADDR_PARAMS
> @@ -5267,14 +5274,24 @@ static int sctp_getsockopt_disable_fragments(struct sock *sk, int len,
>  static int sctp_getsockopt_events(struct sock *sk, int len, char __user *optval,
>  				  int __user *optlen)
>  {
> +	struct sctp_event_subscribe subscribe;
> +	__u8 *sn_type = (__u8 *)&subscribe;
> +	int i;
> +
>  	if (len == 0)
>  		return -EINVAL;
>  	if (len > sizeof(struct sctp_event_subscribe))
>  		len = sizeof(struct sctp_event_subscribe);
>  	if (put_user(len, optlen))
>  		return -EFAULT;
> -	if (copy_to_user(optval, &sctp_sk(sk)->subscribe, len))
> +
> +	for (i = 0; i <= len; i++)
> +		sn_type[i] = sctp_ulpevent_type_enabled(sctp_sk(sk)->subscribe,
> +							SCTP_SN_TYPE_BASE + i);
> +
This seems like an off by one error.  sctp_event_subscribe has N bytes in it (1
byte for each event), meaning that that events 0-(N-1) are subscribable.
Iterating this loop imples that you are going to check N events, overrunning the
sctp_event_subscribe struct.

Neil

> 

Re: [PATCHv2 net-next 1/4] sctp: define subscribe in sctp_sock as __u16

[Neil Horman]

On Tue, Nov 13, 2018 at 02:24:53PM +0800, Xin Long wrote:
>  
>  	/* Default Peer Address Parameters.  These defaults can
>  	 * be modified via SCTP_PEER_ADDR_PARAMS
> @@ -5267,14 +5274,24 @@ static int sctp_getsockopt_disable_fragments(struct sock *sk, int len,
>  static int sctp_getsockopt_events(struct sock *sk, int len, char __user *optval,
>  				  int __user *optlen)
>  {
> +	struct sctp_event_subscribe subscribe;
> +	__u8 *sn_type = (__u8 *)&subscribe;
> +	int i;
> +
>  	if (len = 0)
>  		return -EINVAL;
>  	if (len > sizeof(struct sctp_event_subscribe))
>  		len = sizeof(struct sctp_event_subscribe);
>  	if (put_user(len, optlen))
>  		return -EFAULT;
> -	if (copy_to_user(optval, &sctp_sk(sk)->subscribe, len))
> +
> +	for (i = 0; i <= len; i++)
> +		sn_type[i] = sctp_ulpevent_type_enabled(sctp_sk(sk)->subscribe,
> +							SCTP_SN_TYPE_BASE + i);
> +
This seems like an off by one error.  sctp_event_subscribe has N bytes in it (1
byte for each event), meaning that that events 0-(N-1) are subscribable.
Iterating this loop imples that you are going to check N events, overrunning the
sctp_event_subscribe struct.

Neil

> 

Re: [PATCHv2 net-next 1/4] sctp: define subscribe in sctp_sock as __u16

[Xin Long]

On Wed, Nov 14, 2018 at 2:16 AM Neil Horman <nhorman@tuxdriver.com> wrote:
>
> On Tue, Nov 13, 2018 at 02:24:53PM +0800, Xin Long wrote:
> >
> >       /* Default Peer Address Parameters.  These defaults can
> >        * be modified via SCTP_PEER_ADDR_PARAMS
> > @@ -5267,14 +5274,24 @@ static int sctp_getsockopt_disable_fragments(struct sock *sk, int len,
> >  static int sctp_getsockopt_events(struct sock *sk, int len, char __user *optval,
> >                                 int __user *optlen)
> >  {
> > +     struct sctp_event_subscribe subscribe;
> > +     __u8 *sn_type = (__u8 *)&subscribe;
> > +     int i;
> > +
> >       if (len == 0)
> >               return -EINVAL;
> >       if (len > sizeof(struct sctp_event_subscribe))
> >               len = sizeof(struct sctp_event_subscribe);
> >       if (put_user(len, optlen))
> >               return -EFAULT;
> > -     if (copy_to_user(optval, &sctp_sk(sk)->subscribe, len))
> > +
> > +     for (i = 0; i <= len; i++)
> > +             sn_type[i] = sctp_ulpevent_type_enabled(sctp_sk(sk)->subscribe,
> > +                                                     SCTP_SN_TYPE_BASE + i);
> > +
> This seems like an off by one error.  sctp_event_subscribe has N bytes in it (1
> byte for each event), meaning that that events 0-(N-1) are subscribable.
> Iterating this loop imples that you are going to check N events, overrunning the
> sctp_event_subscribe struct.
you're right, thanks.

>
> Neil
>
> >

Re: [PATCHv2 net-next 1/4] sctp: define subscribe in sctp_sock as __u16

[Xin Long]

On Wed, Nov 14, 2018 at 2:16 AM Neil Horman <nhorman@tuxdriver.com> wrote:
>
> On Tue, Nov 13, 2018 at 02:24:53PM +0800, Xin Long wrote:
> >
> >       /* Default Peer Address Parameters.  These defaults can
> >        * be modified via SCTP_PEER_ADDR_PARAMS
> > @@ -5267,14 +5274,24 @@ static int sctp_getsockopt_disable_fragments(struct sock *sk, int len,
> >  static int sctp_getsockopt_events(struct sock *sk, int len, char __user *optval,
> >                                 int __user *optlen)
> >  {
> > +     struct sctp_event_subscribe subscribe;
> > +     __u8 *sn_type = (__u8 *)&subscribe;
> > +     int i;
> > +
> >       if (len = 0)
> >               return -EINVAL;
> >       if (len > sizeof(struct sctp_event_subscribe))
> >               len = sizeof(struct sctp_event_subscribe);
> >       if (put_user(len, optlen))
> >               return -EFAULT;
> > -     if (copy_to_user(optval, &sctp_sk(sk)->subscribe, len))
> > +
> > +     for (i = 0; i <= len; i++)
> > +             sn_type[i] = sctp_ulpevent_type_enabled(sctp_sk(sk)->subscribe,
> > +                                                     SCTP_SN_TYPE_BASE + i);
> > +
> This seems like an off by one error.  sctp_event_subscribe has N bytes in it (1
> byte for each event), meaning that that events 0-(N-1) are subscribable.
> Iterating this loop imples that you are going to check N events, overrunning the
> sctp_event_subscribe struct.
you're right, thanks.

>
> Neil
>
> >