[PATCH] opensm/osm_mcast_mgr.c: fix osm_mcast_mgr_compute_max_hops for managed switch

[PATCH] opensm/osm_mcast_mgr.c: fix osm_mcast_mgr_compute_max_hops for managed switch

[Yevgeny Kliteynik]

If a certain mcast group has only one member, and this
member is a switch, OSM fails to select this switch
when trying to create spanning tree for this mcast group:

ERR 0A08: Unable to locate a suitable switch for group 0xC000
ERR 0A17: Unable to create spanning tree (IB_ERROR) for mlid 0xc000

Signed-off-by: Yevgeny Kliteynik <kliteyn-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
---
 opensm/opensm/osm_mcast_mgr.c |   16 ++++++++++------
 1 files changed, 10 insertions(+), 6 deletions(-)

diff --git a/opensm/opensm/osm_mcast_mgr.c b/opensm/opensm/osm_mcast_mgr.c
index 697fb58..16f578b 100644
--- a/opensm/opensm/osm_mcast_mgr.c
+++ b/opensm/opensm/osm_mcast_mgr.c
@@ -204,6 +204,12 @@ static float osm_mcast_mgr_compute_max_hops(osm_sm_t * sm, cl_qlist_t * l,

 	OSM_LOG_ENTER(sm->p_log);

+	if (cl_qlist_count(l) == 0) {
+		/*  We should be here if there aren't any ports in the group */
+		max_hops = 10001; /* see later - we use it to realize no hops */
+		goto Exit;
+	}
+
 	/*
 	   For each member of the multicast group, compute the
 	   number of hops to its base LID.
@@ -215,12 +221,10 @@ static float osm_mcast_mgr_compute_max_hops(osm_sm_t * sm, cl_qlist_t * l,
 			max_hops = hops;
 	}

-	if (max_hops == 0)
-		/*
-		   We should be here if there aren't any ports in the group.
-		 */
-		max_hops = 10001;	/* see later - we use it to realize no hops */
-
+	/* Note that at this point we might get (max_hops == 0),
+	   which means that there's only one member in the mcast
+	   group, and it's the current switch */
+Exit:
 	OSM_LOG_EXIT(sm->p_log);
 	return (float)max_hops;
 }
-- 
1.5.1.4

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] opensm/osm_mcast_mgr.c: fix osm_mcast_mgr_compute_max_hops for managed switch

[Sasha Khapyorsky]

Hi Yevgeny,

On 11:46 Mon 21 Dec     , Yevgeny Kliteynik wrote:
> If a certain mcast group has only one member, and this
> member is a switch, OSM fails to select this switch
> when trying to create spanning tree for this mcast group:
> 
> ERR 0A08: Unable to locate a suitable switch for group 0xC000
> ERR 0A17: Unable to create spanning tree (IB_ERROR) for mlid 0xc000
> 
> Signed-off-by: Yevgeny Kliteynik <kliteyn-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
> ---
>  opensm/opensm/osm_mcast_mgr.c |   16 ++++++++++------
>  1 files changed, 10 insertions(+), 6 deletions(-)
> 
> diff --git a/opensm/opensm/osm_mcast_mgr.c b/opensm/opensm/osm_mcast_mgr.c
> index 697fb58..16f578b 100644
> --- a/opensm/opensm/osm_mcast_mgr.c
> +++ b/opensm/opensm/osm_mcast_mgr.c
> @@ -204,6 +204,12 @@ static float osm_mcast_mgr_compute_max_hops(osm_sm_t * sm, cl_qlist_t * l,
> 
>  	OSM_LOG_ENTER(sm->p_log);
> 
> +	if (cl_qlist_count(l) == 0) {
> +		/*  We should be here if there aren't any ports in the group */
> +		max_hops = 10001; /* see later - we use it to realize no hops */
> +		goto Exit;
> +	}
> +

Could we reach this condition?

>From the caller code:

	num_ports = cl_qlist_count(&port_list);
	if (num_ports == 0) {
		OSM_LOG(sm->p_log, OSM_LOG_VERBOSE,
			"MLID 0x%X has no members - nothing to do\n",
			mbox->mlid);
		goto Exit;
	}

Sasha

>  	/*
>  	   For each member of the multicast group, compute the
>  	   number of hops to its base LID.
> @@ -215,12 +221,10 @@ static float osm_mcast_mgr_compute_max_hops(osm_sm_t * sm, cl_qlist_t * l,
>  			max_hops = hops;
>  	}
> 
> -	if (max_hops == 0)
> -		/*
> -		   We should be here if there aren't any ports in the group.
> -		 */
> -		max_hops = 10001;	/* see later - we use it to realize no hops */
> -
> +	/* Note that at this point we might get (max_hops == 0),
> +	   which means that there's only one member in the mcast
> +	   group, and it's the current switch */
> +Exit:
>  	OSM_LOG_EXIT(sm->p_log);
>  	return (float)max_hops;
>  }
> -- 
> 1.5.1.4
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] opensm/osm_mcast_mgr.c: fix osm_mcast_mgr_compute_max_hops for managed switch

[Yevgeny Kliteynik]

Hi Sasha,

On 22/Dec/09 23:23, Sasha Khapyorsky wrote:
> Hi Yevgeny,
>
> On 11:46 Mon 21 Dec     , Yevgeny Kliteynik wrote:
>> If a certain mcast group has only one member, and this
>> member is a switch, OSM fails to select this switch
>> when trying to create spanning tree for this mcast group:
>>
>> ERR 0A08: Unable to locate a suitable switch for group 0xC000
>> ERR 0A17: Unable to create spanning tree (IB_ERROR) for mlid 0xc000
>>
>> Signed-off-by: Yevgeny Kliteynik<kliteyn-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
>> ---
>>   opensm/opensm/osm_mcast_mgr.c |   16 ++++++++++------
>>   1 files changed, 10 insertions(+), 6 deletions(-)
>>
>> diff --git a/opensm/opensm/osm_mcast_mgr.c b/opensm/opensm/osm_mcast_mgr.c
>> index 697fb58..16f578b 100644
>> --- a/opensm/opensm/osm_mcast_mgr.c
>> +++ b/opensm/opensm/osm_mcast_mgr.c
>> @@ -204,6 +204,12 @@ static float osm_mcast_mgr_compute_max_hops(osm_sm_t * sm, cl_qlist_t * l,
>>
>>   	OSM_LOG_ENTER(sm->p_log);
>>
>> +	if (cl_qlist_count(l) == 0) {
>> +		/*  We should be here if there aren't any ports in the group */
>> +		max_hops = 10001; /* see later - we use it to realize no hops */
>> +		goto Exit;
>> +	}
>> +
>
> Could we reach this condition?
>
>> From the caller code:
>
> 	num_ports = cl_qlist_count(&port_list);
> 	if (num_ports == 0) {
> 		OSM_LOG(sm->p_log, OSM_LOG_VERBOSE,
> 			"MLID 0x%X has no members - nothing to do\n",
> 			mbox->mlid);
> 		goto Exit;
> 	}

You're fight, we can't reach this condition anyway.
So the fix should be just removing the special treatment
of the problematic case. Patch shortly.

-- Yevgeny
  
> Sasha
>
>>   	/*
>>   	   For each member of the multicast group, compute the
>>   	   number of hops to its base LID.
>> @@ -215,12 +221,10 @@ static float osm_mcast_mgr_compute_max_hops(osm_sm_t * sm, cl_qlist_t * l,
>>   			max_hops = hops;
>>   	}
>>
>> -	if (max_hops == 0)
>> -		/*
>> -		   We should be here if there aren't any ports in the group.
>> -		 */
>> -		max_hops = 10001;	/* see later - we use it to realize no hops */
>> -
>> +	/* Note that at this point we might get (max_hops == 0),
>> +	   which means that there's only one member in the mcast
>> +	   group, and it's the current switch */
>> +Exit:
>>   	OSM_LOG_EXIT(sm->p_log);
>>   	return (float)max_hops;
>>   }
>> --
>> 1.5.1.4
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
>> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html