Skip to contentSkip to conversation

Remove dependency to implementation-specific a_rwnd behavior

PATCH Applied Matija Glavinic Pecotic 2 participants 4 messages Feb 9, 2014 – Feb 18, 2014 SCTP protocol

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Remove dependency to implementation-specific a_rwnd behavior

test_timetolive and test_sctp_sendrecvmsq take advantage of the following lksctp characteristic:

- SO_RECVBUF being set prior association is established will control initially advertised rwnd
- Modifying SO_RECVBUF later will keep maximum rwnd at the initial value which is defined at the association startup

Test cases use this fact in order to initially advertise small rwnd, increase buffer later, but keeping the initial rwnd as maximum one. However, we may at later time increase buffer, and implementation may advertise this to the partner. RFC4960 only forbids decrease of the buffer space:

Advertised Receiver Window Credit (a_rwnd): 32 bits (unsigned integer)

This value represents the dedicated buffer space, in number of bytes, the sender of the INIT has reserved in association with this window. During the life of the association, this buffer space SHOULD NOT be lessened (i.e., dedicated buffers taken away from this association); however, an endpoint MAY change the value of a_rwnd it sends in SACK chunks.

Having in mind that rwnd reflects free space in receiver's buffer, implementation may advertise new value, and therefore test case should not rely on above defined behavior. There is no impact on functional behavior of test cases due to this change.

This change was done in course of '[PATCH v2] net: sctp: Fix a_rwnd/rwnd management to reflect real state of the receiver's buffer', as with this patch behavior is changed from described to rwnd reflecting real buffer state.

src/func_tests/test_timetolive.c +1 -11
@@ -172,10 +172,8 @@ int main(int argc, char *argv[])
 	 * This code sets the associations RWND very small so we can
 	 * fill it.  It does this by manipulating the rcvbuf as follows:
 	 * 1) Reduce the rcvbuf size on the socket
-	 * 2) create an association so that we advertize rcvbuf/2 as
+	 * 2) create an association so that we advertise rcvbuf/2 as
 	 *    our initial rwnd
-	 * 3) raise the rcvbuf value so that we don't drop data wile 
-	 *    receiving later data
 	 */
 	len = SMALL_RCVBUF;
 	error = setsockopt(sk2, SOL_SOCKET, SO_RCVBUF, &len,
@@ -239,14 +237,6 @@ int main(int argc, char *argv[])
 	sac = (struct sctp_assoc_change *)iov.iov_base;
 	associd1 = sac->sac_assoc_id;
 
-	/* restore the rcvbuffer size for the receiving socket */
-	error = setsockopt(sk2, SOL_SOCKET, SO_RCVBUF, &orig_len,
-			   sizeof(orig_len));
-
-	if (error)
-		tst_brkm(TBROK, tst_exit, "setsockopt(SO_RCVBUF): %s",
-			strerror(errno));
-
         /* Get the first data message which was sent.  */
         inmessage.msg_controllen = sizeof(incmsg);
         error = test_recvmsg(sk2, &inmessage, MSG_WAITALL);
src/func_tests/test_sctp_sendrecvmsg.c +0 -9
@@ -142,7 +142,6 @@ int main(int argc, char *argv[])
 
 	/*
 	 * Set the RWND small so we can fill it up easily.
-	 * then reset RCVBUF to avoid frame droppage
 	 */
 	len = sizeof(int);
 	error = getsockopt(sk2, SOL_SOCKET, SO_RCVBUF, &oldlen, &len);
@@ -186,14 +185,6 @@ int main(int argc, char *argv[])
 				    SCTP_ASSOC_CHANGE, SCTP_COMM_UP);
 
 
-	/* restore the rcvbuffer size for the receiving socket */
-	error = setsockopt(sk2, SOL_SOCKET, SO_RCVBUF, &oldlen,
-			   sizeof(oldlen));
-
-	if (error)
-		tst_brkm(TBROK, tst_exit, "setsockopt(SO_RCVBUF): %s",
-			 strerror(errno));
-
 	/* Get the communication up message on sk1.  */
 	buflen = REALLY_BIG;
 	msgname_len = sizeof(msgname);
#1 lore raw
Quoted 23 lines · Matija Glavinic Pecotic

test_timetolive and test_sctp_sendrecvmsq take advantage of the following lksctp characteristic:

- SO_RECVBUF being set prior association is established will control initially advertised rwnd
- Modifying SO_RECVBUF later will keep maximum rwnd at the initial value which is defined at the association startup

Test cases use this fact in order to initially advertise small rwnd, increase buffer later, but keeping the initial rwnd as maximum one. However, we may at later time increase buffer, and implementation may advertise this to the partner. RFC4960 only forbids decrease of the buffer space:

Advertised Receiver Window Credit (a_rwnd): 32 bits (unsigned integer)

This value represents the dedicated buffer space, in number of bytes, the sender of the INIT has reserved in association with this window. During the life of the association, this buffer space SHOULD NOT be lessened (i.e., dedicated buffers taken away from this association); however, an endpoint MAY change the value of a_rwnd it sends in SACK chunks.

Quoted 8 lines

Having in mind that rwnd reflects free space in receiver's buffer, implementation may advertise new value, and therefore test case should not rely on above defined behavior. There is no impact on functional behavior of test cases due to this change.

This change was done in course of '[PATCH v2] net: sctp: Fix a_rwnd/rwnd management to reflect real state of the receiver's buffer', as with this patch behavior is changed from described to rwnd reflecting real buffer state.

Before applying this, does that break user space? I had a test run of test_timetolive of the latest net tree and without this patch here, test_timetolive is stuck forever; with this patch, it works again. Before your kernel change, the test suite passed.

So you're saying that the test case is simply implemented wrong by "violating" the RFC in the fact that it keeps/resets the old small init rwnd along. Hence, these hunks should be removed as it "SHOULD NOT" be lessened.

Quoted 1 line
#2 lore raw
In reply to Matija Glavinic Pecotic

Hello,

Quoted 29 lines · ext Daniel Borkmann
Quoted 28 lines · Matija Glavinic Pecotic

test_timetolive and test_sctp_sendrecvmsq take advantage of the following lksctp characteristic:

- SO_RECVBUF being set prior association is established will control initially advertised rwnd
- Modifying SO_RECVBUF later will keep maximum rwnd at the initial value which is defined at the association startup

Test cases use this fact in order to initially advertise small rwnd, increase buffer later, but keeping the initial rwnd as maximum one. However, we may at later time increase buffer, and implementation may advertise this to the partner. RFC4960 only forbids decrease of the buffer space:

Advertised Receiver Window Credit (a_rwnd): 32 bits (unsigned integer)

This value represents the dedicated buffer space, in number of bytes, the sender of the INIT has reserved in association with this window. During the life of the association, this buffer space SHOULD NOT be lessened (i.e., dedicated buffers taken away from this association); however, an endpoint MAY change the value of a_rwnd it sends in SACK chunks.

Having in mind that rwnd reflects free space in receiver's buffer, implementation may advertise new value, and therefore test case should not rely on above defined behavior. There is no impact on functional behavior of test cases due to this change.

Quoted 8 lines
Quoted 3 lines

This change was done in course of '[PATCH v2] net: sctp: Fix a_rwnd/rwnd management to reflect real state of the receiver's buffer', as with this patch behavior is changed from described to rwnd reflecting real buffer state.

Before applying this, does that break user space? I had a test run of test_timetolive of the latest net tree and without this patch here, test_timetolive is stuck forever; with this patch, it works again. Before your kernel change, the test suite passed.

this is correct observation

Quoted 4 lines

So you're saying that the test case is simply implemented wrong by "violating" the RFC in the fact that it keeps/resets the old small init rwnd along. Hence, these hunks should be removed as it "SHOULD NOT" be lessened.

Here is discussed "it may be changed", and in this particular case, it may grow. What test case relies is that initially set rwnd will never change, regardless of the socket buffer change. However, with proposed fair advertisement, if SO_RECVBUF is changed, new buffer size will be reported to the sender in form of increased/decreased maximum rwnd. This makes tc hang, fillmessage being created with size of increased buffer (much more then REALLY_BIG, at least on my machine where rmem_defualt is 160k) and not the original (SMALL_RCVBUF). This will deplete receiver, hang sender on next message, and since everything is done in same thread, receiver never gets opportunity to read.

maximum (and minimum?) rwnd might be clamped to deal with this particular scenario, but I would still propose to go for completely fair advertisement.

Quoted 5 lines
Quoted 1 line
Signature · 3 lines
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
#3 lore raw
In reply to Matija Glavinic Pecotic
Quoted 3 lines · Matija Glavinic Pecotic

test_timetolive and test_sctp_sendrecvmsq take advantage of the following lksctp characteristic:

...

Quoted 5 lines

This change was done in course of '[PATCH v2] net: sctp: Fix a_rwnd/rwnd management to reflect real state of the receiver's buffer', as with this patch behavior is changed from described to rwnd reflecting real buffer state.

Applied.