sctp: a bunch of fixes for stream reconfig

When doing asoc reset, if the sender of the response has already sent some chunk and increased asoc->next_tsn before the duplicate request comes, the response will use the old result with an incorrect sender next_tsn.

Better than asoc->next_tsn, asoc->ctsn_ack_point can't be changed after the sender of the response has performed the asoc reset and before the peer has confirmed it, and it's value is still asoc->next_tsn original value minus 1.

This patch sets sender next_tsn for the old result with ctsn_ack_point plus 1 when processing the duplicate request, to make sure the sender next_tsn value peer gets will be always right.

Fixes 692787cef651 ("sctp: implement receiver-side procedures for the SSN/TSN Reset Request Parameter")

Signed-off-by Xin Long <[email protected]>

---
 net/sctp/stream.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

net/sctp/stream.c +1 -1

index 9dd5bfe..a20145b 100644
@@ -725,7 +725,7 @@ struct sctp_chunk *sctp_process_strreset_tsnreq(
 		i = asoc->strreset_inseq - request_seq - 1;
 		result = asoc->strreset_result[i];
 		if (result = SCTP_STRRESET_PERFORMED) {
-			next_tsn = asoc->next_tsn;
+			next_tsn = asoc->ctsn_ack_point + 1;
 			init_tsn >  				sctp_tsnmap_get_ctsn(&asoc->peer.tsn_map) + 1;
 		}

Signature · 6 lines

2.1.0

--
To unsubscribe from this list: send the line "unsubscribe linux-sctp" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Now when doing asoc reset, it cleans up sacked and abandoned queues by calling sctp_outq_free where it also cleans up unsent, retransmit and transmitted queues.

It's safe for the sender of response, as these 3 queues are empty at that time. But when the receiver of response is doing the reset, the users may already enqueue some chunks into unsent during the time waiting the response, and these chunks should not be flushed.

To void the chunks in it would be removed, it moves the queue into a temp list, then gets it back after sctp_outq_free is done.

The patch also fixes some incorrect comments in
sctp_process_strreset_tsnreq.

Signed-off-by Xin Long <[email protected]>

---
 net/sctp/stream.c | 21 ++++++++++++++-------
 1 file changed, 14 insertions(+), 7 deletions(-)

net/sctp/stream.c +14 -7

index f3b7d27..9dd5bfe 100644
@@ -747,9 +747,10 @@ struct sctp_chunk *sctp_process_strreset_tsnreq(
 		goto out;
 	}
 
-	/* G3: The same processing as though a SACK chunk with no gap report
-	 *     and a cumulative TSN ACK of the Sender's Next TSN minus 1 were
-	 *     received MUST be performed.
+	/* G4: The same processing as though a FWD-TSN chunk (as defined in
+	 *     [RFC3758]) with all streams affected and a new cumulative TSN
+	 *     ACK of the Receiver's Next TSN minus 1 were received MUST be
+	 *     performed.
 	 */
 	max_tsn_seen = sctp_tsnmap_get_max_tsn_seen(&asoc->peer.tsn_map);
 	sctp_ulpq_reasm_flushtsn(&asoc->ulpq, max_tsn_seen);

@@ -764,10 +765,9 @@ struct sctp_chunk *sctp_process_strreset_tsnreq(
 	sctp_tsnmap_init(&asoc->peer.tsn_map, SCTP_TSN_MAP_INITIAL,
 			 init_tsn, GFP_ATOMIC);
 
-	/* G4: The same processing as though a FWD-TSN chunk (as defined in
-	 *     [RFC3758]) with all streams affected and a new cumulative TSN
-	 *     ACK of the Receiver's Next TSN minus 1 were received MUST be
-	 *     performed.
+	/* G3: The same processing as though a SACK chunk with no gap report
+	 *     and a cumulative TSN ACK of the Sender's Next TSN minus 1 were
+	 *     received MUST be performed.
 	 */
 	sctp_outq_free(&asoc->outqueue);
 

@@ -1021,6 +1021,7 @@ struct sctp_chunk *sctp_process_strreset_resp(
 		if (result = SCTP_STRRESET_PERFORMED) {
 			__u32 mtsn = sctp_tsnmap_get_max_tsn_seen(
 						&asoc->peer.tsn_map);
+			LIST_HEAD(temp);
 
 			sctp_ulpq_reasm_flushtsn(&asoc->ulpq, mtsn);
 			sctp_ulpq_abort_pd(&asoc->ulpq, GFP_ATOMIC);

@@ -1029,7 +1030,13 @@ struct sctp_chunk *sctp_process_strreset_resp(
 					 SCTP_TSN_MAP_INITIAL,
 					 stsn, GFP_ATOMIC);
 
+			/* Clean up sacked and abandoned queues only. As the
+			 * out_chunk_list may not be empty, splice it to temp,
+			 * then get it back after sctp_outq_free is done.
+			 */
+			list_splice_init(&asoc->outqueue.out_chunk_list, &temp);
 			sctp_outq_free(&asoc->outqueue);
+			list_splice_init(&temp, &asoc->outqueue.out_chunk_list);
 
 			asoc->next_tsn = rtsn;
 			asoc->ctsn_ack_point = asoc->next_tsn - 1;

Signature · 6 lines

2.1.0

--
To unsubscribe from this list: send the line "unsubscribe linux-sctp" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Sent: 25 November 2017 13:06

As it says in rfc6525#section5.1.4, before sending the request,

C2: The sender has either no outstanding TSNs or considers all
outstanding TSNs abandoned.

Prior to this patch, it tried to consider all outstanding TSNs abandoned by dropping all chunks in all outqs with sctp_outq_free (even including sacked, retransmit and transmitted queues) when doing this reset, which is too aggressive.

To make it work gently, this patch will only allow the asoc reset when the sender has no outstanding TSNs by checking if unsent, transmitted and retransmit are all empty with sctp_outq_is_empty before sending and processing the request.

From: Xin Long

Quoted 12 lines

Sent: 25 November 2017 13:06

As it says in rfc6525#section5.1.4, before sending the request,

C2: The sender has either no outstanding TSNs or considers all
outstanding TSNs abandoned.

Prior to this patch, it tried to consider all outstanding TSNs abandoned by dropping all chunks in all outqs with sctp_outq_free (even including sacked, retransmit and transmitted queues) when doing this reset, which is too aggressive.

Quoted 4 lines

To make it work gently, this patch will only allow the asoc reset when the sender has no outstanding TSNs by checking if unsent, transmitted and retransmit are all empty with sctp_outq_is_empty before sending and processing the request.

Doesn't that rather defeat the point of a reset?

As it says in rfc6525#section5.1.4, before sending the request,

C2: The sender has either no outstanding TSNs or considers all
outstanding TSNs abandoned.

Prior to this patch, it tried to consider all outstanding TSNs abandoned by dropping all chunks in all outqs with sctp_outq_free (even including sacked, retransmit and transmitted queues) when doing this reset, which is too aggressive.

To make it work gently, this patch will only allow the asoc reset when the sender has no outstanding TSNs by checking if unsent, transmitted and retransmit are all empty with sctp_outq_is_empty before sending and processing the request.

Fixes 692787cef651 ("sctp: implement receiver-side procedures for the SSN/TSN Reset Request Parameter")

Signed-off-by Xin Long <[email protected]>

---
 net/sctp/stream.c | 9 +++++++++
 1 file changed, 9 insertions(+)

net/sctp/stream.c +9 -0

index b209037..f3b7d27 100644
@@ -406,6 +406,9 @@ int sctp_send_reset_assoc(struct sctp_association *asoc)
 	if (asoc->strreset_outstanding)
 		return -EINPROGRESS;
 
+	if (!sctp_outq_is_empty(&asoc->outqueue))
+		return -EAGAIN;
+
 	chunk = sctp_make_strreset_tsnreq(asoc);
 	if (!chunk)
 		return -ENOMEM;

@@ -728,6 +731,12 @@ struct sctp_chunk *sctp_process_strreset_tsnreq(
 		}
 		goto err;
 	}
+
+	if (!sctp_outq_is_empty(&asoc->outqueue)) {
+		result = SCTP_STRRESET_IN_PROGRESS;
+		goto err;
+	}
+
 	asoc->strreset_inseq++;
 
 	if (!(asoc->strreset_enable & SCTP_ENABLE_RESET_ASSOC_REQ))

Signature · 6 lines

2.1.0

--
To unsubscribe from this list: send the line "unsubscribe linux-sctp" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Now the out stream reset in sctp stream reconf could be done even if the stream outq is not empty. It means that users can not be sure since which msg the new ssn will be used.

To make this more synchronous, it shouldn't allow to do out stream reset until these chunks in unsent outq all are sent out.

This patch checks the corresponding stream outqs when sending and processing the request . If any of them has unsent chunks in outq, it will return -EAGAIN instead or send SCTP_STRRESET_IN_PROGRESS back to the sender.

Fixes 7f9d68ac944e ("sctp: implement sender-side procedures for SSN Reset Request Parameter")

Suggested-by Marcelo Ricardo Leitner <[email protected]>

Signed-off-by Xin Long <[email protected]>

---
 net/sctp/stream.c | 35 +++++++++++++++++++++++++++++++++++
 1 file changed, 35 insertions(+)

net/sctp/stream.c +35 -0

index 09c797a..b209037 100644
@@ -254,6 +254,30 @@ static int sctp_send_reconf(struct sctp_association *asoc,
 	return retval;
 }
 
+static bool sctp_stream_outq_is_empty(struct sctp_stream *stream,
+				      __u16 str_nums, __be16 *str_list)
+{
+	struct sctp_association *asoc;
+	__u16 i;
+
+	asoc = container_of(stream, struct sctp_association, stream);
+	if (!asoc->outqueue.out_qlen)
+		return true;
+
+	if (!str_nums)
+		return false;
+
+	for (i = 0; i < str_nums; i++) {
+		__u16 sid = ntohs(str_list[i]);
+
+		if (stream->out[sid].ext &&
+		    !list_empty(&stream->out[sid].ext->outq))
+			return false;
+	}
+
+	return true;
+}
+
 int sctp_send_reset_streams(struct sctp_association *asoc,
 			    struct sctp_reset_streams *params)
 {

@@ -317,6 +341,11 @@ int sctp_send_reset_streams(struct sctp_association *asoc,
 	for (i = 0; i < str_nums; i++)
 		nstr_list[i] = htons(str_list[i]);
 
+	if (out && !sctp_stream_outq_is_empty(stream, str_nums, nstr_list)) {
+		retval = -EAGAIN;
+		goto out;
+	}
+
 	chunk = sctp_make_strreset_req(asoc, str_nums, nstr_list, out, in);
 
 	kfree(nstr_list);

@@ -636,6 +665,12 @@ struct sctp_chunk *sctp_process_strreset_inreq(
 		}
 	}
 
+	if (!sctp_stream_outq_is_empty(stream, nums, str_p)) {
+		result = SCTP_STRRESET_IN_PROGRESS;
+		asoc->strreset_inseq--;
+		goto err;
+	}
+
 	chunk = sctp_make_strreset_req(asoc, nums, str_p, 1, 0);
 	if (!chunk)
 		goto out;

Signature · 6 lines

2.1.0

--
To unsubscribe from this list: send the line "unsubscribe linux-sctp" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Now in stream reconf part there are still some places using magic number 2 for each stream number length. To make it more readable, this patch is to replace them with sizeof(__u16).

Reported-by Marcelo Ricardo Leitner <[email protected]>

Signed-off-by Xin Long <[email protected]>

---
 net/sctp/stream.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

net/sctp/stream.c +6 -4

index a11db21..09c797a 100644
@@ -563,7 +563,7 @@ struct sctp_chunk *sctp_process_strreset_outreq(
 		flags = SCTP_STREAM_RESET_INCOMING_SSN;
 	}
 
-	nums = (ntohs(param.p->length) - sizeof(*outreq)) / 2;
+	nums = (ntohs(param.p->length) - sizeof(*outreq)) / sizeof(__u16);
 	if (nums) {
 		str_p = outreq->list_of_streams;
 		for (i = 0; i < nums; i++) {

@@ -627,7 +627,7 @@ struct sctp_chunk *sctp_process_strreset_inreq(
 		goto out;
 	}
 
-	nums = (ntohs(param.p->length) - sizeof(*inreq)) / 2;
+	nums = (ntohs(param.p->length) - sizeof(*inreq)) / sizeof(__u16);
 	str_p = inreq->list_of_streams;
 	for (i = 0; i < nums; i++) {
 		if (ntohs(str_p[i]) >= stream->outcnt) {

@@ -927,7 +927,8 @@ struct sctp_chunk *sctp_process_strreset_resp(
 
 		outreq = (struct sctp_strreset_outreq *)req;
 		str_p = outreq->list_of_streams;
-		nums = (ntohs(outreq->param_hdr.length) - sizeof(*outreq)) / 2;
+		nums = (ntohs(outreq->param_hdr.length) - sizeof(*outreq)) /
+		       sizeof(__u16);
 
 		if (result = SCTP_STRRESET_PERFORMED) {
 			if (nums) {

@@ -956,7 +957,8 @@ struct sctp_chunk *sctp_process_strreset_resp(
 
 		inreq = (struct sctp_strreset_inreq *)req;
 		str_p = inreq->list_of_streams;
-		nums = (ntohs(inreq->param_hdr.length) - sizeof(*inreq)) / 2;
+		nums = (ntohs(inreq->param_hdr.length) - sizeof(*inreq)) /
+		       sizeof(__u16);
 
 		*evp = sctp_ulpevent_make_stream_reset_event(asoc, flags,
 			nums, str_p, GFP_ATOMIC);

Signature · 6 lines

2.1.0

--
To unsubscribe from this list: send the line "unsubscribe linux-sctp" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

This patchset is to make stream reset and asoc reset work more correctly for stream reconfig.

Thank to Marcelo making them very clear.

Xin Long (5):
sctp: use sizeof(__u16) for each stream number length instead of magic
number
sctp: only allow the out stream reset when the stream outq is empty
sctp: only allow the asoc reset when the asoc outq is empty
sctp: avoid flushing unsent queue when doing asoc reset
sctp: set sender next_tsn for the old result with ctsn_ack_point plus

    1

 net/sctp/stream.c | 77 ++++++++++++++++++++++++++++++++++++++++++++++---------
 1 file changed, 65 insertions(+), 12 deletions(-)

Signature · 1 line

2.1.0

This patchset is to make stream reset and asoc reset work more correctly for stream reconfig.

Thank to Marcelo making them very clear.