Yevgeny Kliteynik wrote:
Nicolas Morey Chaisemartin wrote:
Yevgeny Kliteynik wrote:
Hi Nicolas,
/* foreach down-going port group (in indexing order) */
- i = p_sw->down_port_groups_idx;
+ i = (p_sw->down_port_groups_idx +
+ p_sw->down_port_groups_num) % p_sw->down_port_groups_num;
Perhaps it would be simpler just to init the down_port_groups_idx to
0 instead of -1?
Something like this:
diff --git a/opensm/opensm/osm_ucast_ftree.c
b/opensm/opensm/osm_ucast_ftree.c
index 4e65c87..eae1ed8 100644
--- a/opensm/opensm/osm_ucast_ftree.c
+++ b/opensm/opensm/osm_ucast_ftree.c
@@ -563,7 +563,7 @@ static ftree_sw_t *__osm_ftree_sw_create(IN
ftree_fabric_t * p_ftree,
/* initialize lft buffer */
memset(p_osm_sw->new_lft, OSM_NO_PATH, IB_LID_UCAST_END_HO + 1);
- p_sw->down_port_groups_idx = -1;
+ p_sw->down_port_groups_idx = 0;
return p_sw;
} /* __osm_ftree_sw_create() */
Sure. I wanted to ensure that whatever happens to the index it would
always be in the right interval but after checking I doubt anything
else than initialization could set it outside its normal interval.
Do you want me to make the patch and send it or will you just push
yours?
I'm ok with both options.
I can send a clean patch to Sasha tomorrow (I'm OOO today), or you can
do it today.
-- Yevgeny
Nicolas
Yours should be faster and I recheck and I see no reason to enforce a
"check" in the function so I prefer your solution.
I'll repost the patch today as it's breaking opensm/ftree.
Nicolas
_______________________________________________
general mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
