[ofa-general] [OpenSM] Traps 64 and 65 sent too early.

[Sasha Khapyorsky]

Hi Terry,

On 00:59 Tue 04 Mar     , Terry Greeniaus wrote:
> 
> I think this could be fixed by moving the
> __osm_state_mgr_report_new_ports() call in osm_state_mgr.c to follow the
> call that transitions the subnet to Armed or Active, however I am not
> very familiar with the OpenSM code so am appealing to the list to help
> me out.

This seems correct to me. Could you try the patch (against master)?

>From 1dfc192236c47edfa24d70967a3027af01aaa28d Mon Sep 17 00:00:00 2001
From: Sasha Khapyorsky <sashak at voltaire.com>
Date: Wed, 5 Mar 2008 13:28:37 +0200
Subject: [PATCH] opensm: send trap 64 only after new ports are in ACTIVE state.

Send trap 64 only when new ports were moved to ACTIVE state and routing
is configured.

Signed-off-by: Sasha Khapyorsky <sashak at voltaire.com>
---
 opensm/opensm/osm_state_mgr.c |    8 +++++---
 1 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/opensm/opensm/osm_state_mgr.c b/opensm/opensm/osm_state_mgr.c
index 38b2c4e..9b03314 100644
--- a/opensm/opensm/osm_state_mgr.c
+++ b/opensm/opensm/osm_state_mgr.c
@@ -1169,10 +1169,7 @@ _repeat_discovery:
 
 	/*
 	 * Proceed with unicast forwarding table configuration.
-	 * First - send trap 64 on newly discovered endports
 	 */
-	__osm_state_mgr_report_new_ports(sm);
-
 	osm_ucast_mgr_process(&sm->ucast_mgr);
 	if (wait_for_pending_transactions(&sm->p_subn->p_osm->stats))
 		return;
@@ -1223,6 +1220,11 @@ _repeat_discovery:
 	 * The sweep completed!
 	 */
 
+	/*
+	 * Send trap 64 on newly discovered endports
+	 */
+	__osm_state_mgr_report_new_ports(sm);
+
 	/* in any case we zero this flag */
 	sm->p_subn->coming_out_of_standby = FALSE;
 
-- 
1.5.4.1.122.gaa8d


> Also, perhaps there is a bug tracker somewhere against which I should 
> file a bug report?

Sure, it is https://bugs.openfabrics.org .

Sasha