diff mbox

[infiniband-diags,v3] ibportstate: Fixed switch peer port probing when using DR routing

Message ID 22b5963c36776d0b03eb0f6706323b4ba4bafac9.1477315879.git-series.knut.omang@oracle.com (mailing list archive)
State Superseded
Headers show

Commit Message

Knut Omang Oct. 24, 2016, 1:34 p.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 | 26 +++++++++++++++++---------
 1 file changed, 17 insertions(+), 9 deletions(-)


base-commit: 17e03b4738913365a3f947719c4897fcb92df32c

Comments

Hal Rosenstock Oct. 24, 2016, 1:50 p.m. UTC | #1
On 10/24/2016 9:34 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.
> 
> 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 | 26 +++++++++++++++++---------
>  1 file changed, 17 insertions(+), 9 deletions(-)
> 
> diff --git a/src/ibportstate.c b/src/ibportstate.c
> index cfb8be7..82cbcc2 100644
> --- a/src/ibportstate.c
> +++ b/src/ibportstate.c
> @@ -1,6 +1,7 @@
>  /*
>   * 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 +656,22 @@ 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,
> +						         &selfport, 0) < 0)
> +					IBEXIT("could not resolve self");
> +				peerportid.drpath.drslid = selfportid.lid;

Why was casting of selfportid.lid to (uint16_t) dropped ?

> +				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);
> 
> base-commit: 17e03b4738913365a3f947719c4897fcb92df32c
> 
--
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. 24, 2016, 2:37 p.m. UTC | #2
On 10/24/2016 03:50 PM, Hal Rosenstock wrote:
> On 10/24/2016 9:34 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.
>>
>> 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 | 26 +++++++++++++++++---------
>>   1 file changed, 17 insertions(+), 9 deletions(-)
>>
>> diff --git a/src/ibportstate.c b/src/ibportstate.c
>> index cfb8be7..82cbcc2 100644
>> --- a/src/ibportstate.c
>> +++ b/src/ibportstate.c
>> @@ -1,6 +1,7 @@
>>   /*
>>    * 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 +656,22 @@ 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,
>> +						         &selfport, 0) < 0)
>> +					IBEXIT("could not resolve self");
>> +				peerportid.drpath.drslid = selfportid.lid;
> Why was casting of selfportid.lid to (uint16_t) dropped ?
That was not intentional. Thanks for pointing it out.

Regards,
-Dag
>
>> +				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);
>>
>> base-commit: 17e03b4738913365a3f947719c4897fcb92df32c
>>

--
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 cfb8be7..82cbcc2 100644
--- a/src/ibportstate.c
+++ b/src/ibportstate.c
@@ -1,6 +1,7 @@ 
 /*
  * 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 +656,22 @@  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,
+						         &selfport, 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);