diff mbox

[infiniband-diags,v2] ibportstate: Fixed switch peer port probing when using DR routing

Message ID 8d3851995d389220a979543740862bd8884cacac.1476089300.git-series.knut.omang@oracle.com (mailing list archive)
State Superseded
Delegated to: Ira Weiny
Headers show

Commit Message

Knut Omang Oct. 10, 2016, 8:50 a.m. UTC
From: Dag Moxnes <dag.moxnes@oracle.com>

ibportstate queries to a remote peer port on a switch using direct
routing would result in timeouts. The reason for this is that the
DR path was not correctly constructed.

Signed-off-by: Dag Moxnes <dag.moxnes@oracle.com>
Reviewed-by: Line Holen <line.holen@oracle.com>
Signed-off-by: Knut Omang <knut.omang@oracle.com>
---
 src/ibportstate.c | 27 ++++++++++++++++++---------
 1 file changed, 18 insertions(+), 9 deletions(-)


base-commit: 2937cf99350a2e423b705e8b8dd10499796a7b41

Comments

Hal Rosenstock Oct. 10, 2016, 2:57 p.m. UTC | #1
On 10/10/2016 4:50 AM, Knut Omang wrote:
> From: Dag Moxnes <dag.moxnes@oracle.com>
> 
> ibportstate queries to a remote peer port on a switch using direct
> routing would result in timeouts. The reason for this is that the
> DR path was not correctly constructed.

Should the following be added here to make it clearer:
in case where local LID for switch port 0 is not yet configured.
?

> 
> Signed-off-by: Dag Moxnes <dag.moxnes@oracle.com>
> Reviewed-by: Line Holen <line.holen@oracle.com>
> Signed-off-by: Knut Omang <knut.omang@oracle.com>

Reviewed-by: Hal Rosenstock <hal@mellanox.com>
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Dag Moxnes Oct. 10, 2016, 8:29 p.m. UTC | #2
Den 10.10.2016 16.57, skrev Hal Rosenstock:
> On 10/10/2016 4:50 AM, Knut Omang wrote:
>> From: Dag Moxnes <dag.moxnes@oracle.com>
>>
>> ibportstate queries to a remote peer port on a switch using direct
>> routing would result in timeouts. The reason for this is that the
>> DR path was not correctly constructed.
> Should the following be added here to make it clearer:
> in case where local LID for switch port 0 is not yet configured.
> ?
I don't think so. The issue will happen every time the user is calling 
ibportstate
with direct routing (-D argument) irrespective of wether the local LID 
for switch
port 0 is configured or not.

-Dag
>
>> Signed-off-by: Dag Moxnes <dag.moxnes@oracle.com>
>> Reviewed-by: Line Holen <line.holen@oracle.com>
>> Signed-off-by: Knut Omang <knut.omang@oracle.com>
> Reviewed-by: Hal Rosenstock <hal@mellanox.com>

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Hal Rosenstock Oct. 10, 2016, 10:04 p.m. UTC | #3
On 10/10/2016 4:29 PM, Dag Moxnes wrote:
> 
> 
> Den 10.10.2016 16.57, skrev Hal Rosenstock:
>> On 10/10/2016 4:50 AM, Knut Omang wrote:
>>> From: Dag Moxnes <dag.moxnes@oracle.com>
>>>
>>> ibportstate queries to a remote peer port on a switch using direct
>>> routing would result in timeouts. The reason for this is that the
>>> DR path was not correctly constructed.
>> Should the following be added here to make it clearer:
>> in case where local LID for switch port 0 is not yet configured.
>> ?
> I don't think so. The issue will happen every time the user is calling
> ibportstate
> with direct routing (-D argument) irrespective of wether the local LID
> for switch
> port 0 is configured or not.

Oh right; portid.lid is not changed from 0 by ib_resolve_portid_str when
DR is used.

-- Hal

> -Dag
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/src/ibportstate.c b/src/ibportstate.c
index cb47aa9..7046f47 100644
--- a/src/ibportstate.c
+++ b/src/ibportstate.c
@@ -1,6 +1,8 @@ 
 /*
  * Copyright (c) 2004-2009 Voltaire Inc.  All rights reserved.
  * Copyright (c) 2010,2011 Mellanox Technologies LTD.  All rights reserved.
+ * Copyright (c) 2011, 2016, Oracle and/or its affiliates. All rights reserved.
+ *
  *
  * This software is available to you under a choice of one of two
  * licenses.  You may choose to be licensed under the terms of the GNU
@@ -655,15 +657,22 @@  int main(int argc, char **argv)
 
 			/* Setup portid for peer port */
 			memcpy(&peerportid, &portid, sizeof(peerportid));
-			peerportid.drpath.cnt = 1;
-			peerportid.drpath.p[1] = (uint8_t) portnum;
-
-			/* Set DrSLID to local lid */
-			if (resolve_self(ibd_ca, ibd_ca_port, &selfportid,
-						&selfport, 0) < 0)
-				IBEXIT("could not resolve self");
-			peerportid.drpath.drslid = (uint16_t) selfportid.lid;
-			peerportid.drpath.drdlid = 0xffff;
+			if (portid.lid == 0) {
+				peerportid.drpath.cnt++;
+				if (peerportid.drpath.cnt == IB_SUBNET_PATH_HOPS_MAX) {
+					IBEXIT("Too many hops");
+				}
+			} else {
+				peerportid.drpath.cnt = 1;
+
+				/* Set DrSLID to local lid */
+				if (resolve_self(ibd_ca, ibd_ca_port, &selfportid, argv[0],
+						         &selfportid, 0) < 0)
+					IBEXIT("could not resolve self");
+				peerportid.drpath.drslid = selfportid.lid;
+				peerportid.drpath.drdlid = 0xffff;
+			}
+			peerportid.drpath.p[peerportid.drpath.cnt] = (uint8_t) portnum;
 
 			/* Get peer port NodeInfo to obtain peer port number */
 			is_peer_switch = get_node_info(&peerportid, data);