Hi Jim,
On 20/Nov/09 21:15, Jim Schutt wrote:
For a fabric that requires routing with an engine with special properties,
say avoiding credit loops via making use of SLs in routing, it might
be preferable to not fall back to minhop if the configured routing engine
fails.
E.g. the torus-2QoS routing engine uses both SL2VL maps and path SL values
to provide routing free of credit loops, but cannot route fabrics for
some patterns of failed switches. Should a switch fail that creates such
a pattern, it may be preferable to keep the previous routing information
loaded in the switches until a switch can be replaced that restores
torus-2QoS's ability to route the fabric.
The alternative, having some other engine route the fabric, will immediately
introduce credit loops.
This is a great idea.
Regarding the implementation: I would prefer seeing this
as a purely OpenSM option and not as a new routing engine
keyword.
I think it would be cleaner to leave the list of routing
engines w/o special keys, and have a general option
that would prevent SM from falling back. Actually, the
fall-back itself is not bad, as it is defined by the list
of routing engines, and SM should try them one by one.
The problem is with using default routing that is not
specified in the routing engines list.
Here's the patch that implements OSM option
"use_default_routing", and a command line parameter
"no_default_routing" to control this option.
I'll write the patch that adds this option to the
OSM trunk and send it to Sasha shortly.
Signed-off-by: Yevgeny Kliteynik <klit...@dev.mellanox.co.il>
---
opensm/include/opensm/osm_subnet.h | 2 +-
opensm/opensm/main.c | 9 +++++++++
opensm/opensm/osm_opensm.c | 10 ++++------
opensm/opensm/osm_subnet.c | 8 ++++++++
opensm/opensm/osm_ucast_mgr.c | 7 +++++--
5 files changed, 27 insertions(+), 9 deletions(-)
diff --git a/opensm/include/opensm/osm_subnet.h
b/opensm/include/opensm/osm_subnet.h
index a4133a0..905f64d 100644
--- a/opensm/include/opensm/osm_subnet.h
+++ b/opensm/include/opensm/osm_subnet.h
@@ -190,6 +190,7 @@ typedef struct osm_subn_opt {
boolean_t sweep_on_trap;
char *routing_engine_names;
boolean_t use_ucast_cache;
+ boolean_t use_default_routing;
boolean_t connect_roots;
char *lid_matrix_dump_file;
char *lfts_file;
@@ -215,7 +216,6 @@ typedef struct osm_subn_opt {
osm_qos_options_t qos_rtr_options;
boolean_t enable_quirks;
boolean_t no_clients_rereg;
- boolean_t no_fallback_routing_engine;
#ifdef ENABLE_OSM_PERF_MGR
boolean_t perfmgr;
boolean_t perfmgr_redir;
diff --git a/opensm/opensm/main.c b/opensm/opensm/main.c
index 096bf5f..47075a2 100644
--- a/opensm/opensm/main.c
+++ b/opensm/opensm/main.c
@@ -175,6 +175,10 @@ static void show_usage(void)
" separated by commas so that specific ordering of
routing\n"
" algorithms will be tried if earlier routing engines
fail.\n"
" Supported engines: updn, file, ftree, lash, dor,
torus-2QoS\n\n");
+ printf("--no_default_routing\n"
+ " This option prevents OpenSM from falling back to
default\n"
+ " routing if none of the provided engines was able to\n"
+ " configure the subnet.\n\n");
printf("--do_mesh_analysis\n"
" This option enables additional analysis for the
lash\n"
" routing engine to precondition switch port
assignments\n"
@@ -612,6 +616,7 @@ int main(int argc, char *argv[])
{"sm_sl", 1, NULL, 7},
{"retries", 1, NULL, 8},
{"torus_config", 1, NULL, 9},
+ {"no_default_routing", 0, NULL, 10},
{NULL, 0, NULL, 0} /* Required at the end of the array */
};
@@ -993,6 +998,10 @@ int main(int argc, char *argv[])
case 9:
SET_STR_OPT(opt.torus_conf_file, optarg);
break;
+ case 10:
+ opt.use_default_routing = FALSE;
+ printf(" No fall back to default routing\n");
+ break;
case 'h':
case '?':
case ':':
diff --git a/opensm/opensm/osm_opensm.c b/opensm/opensm/osm_opensm.c
index e7ef55c..d153be5 100644
--- a/opensm/opensm/osm_opensm.c
+++ b/opensm/opensm/osm_opensm.c
@@ -159,11 +159,6 @@ static struct osm_routing_engine
*setup_routing_engine(osm_opensm_t *osm,
struct osm_routing_engine *re;
const struct routing_engine_module *m;
- if (!strcmp(name, "no_fallback")) {
- osm->subn.opt.no_fallback_routing_engine = TRUE;
- return NULL;
- }
-
for (m = routing_modules; m->name && *m->name; m++) {
if (!strcmp(m->name, name)) {
re = malloc(sizeof(struct osm_routing_engine));
@@ -212,7 +207,10 @@ static void setup_routing_engines(osm_opensm_t *osm, const
char *engine_names)
}
free(str);
}
- if (!osm->default_routing_engine) {
+
+ if (!engine_names || !*engine_names ||
+ (!osm->default_routing_engine &&
+ osm->subn.opt.use_default_routing)) {
re = setup_routing_engine(osm, "minhop");
if (!osm->routing_engine_list && re)
append_routing_engine(osm, re);
diff --git a/opensm/opensm/osm_subnet.c b/opensm/opensm/osm_subnet.c
index 03d9538..274e807 100644
--- a/opensm/opensm/osm_subnet.c
+++ b/opensm/opensm/osm_subnet.c
@@ -327,6 +327,7 @@ static const opt_rec_t opt_tbl[] = {
{ "port_profile_switch_nodes", OPT_OFFSET(port_profile_switch_nodes),
opts_parse_boolean, NULL, 1 },
{ "sweep_on_trap", OPT_OFFSET(sweep_on_trap), opts_parse_boolean, NULL,
1 },
{ "routing_engine", OPT_OFFSET(routing_engine_names), opts_parse_charp,
NULL, 0 },
+ { "use_default_routing", OPT_OFFSET(use_default_routing),
opts_parse_boolean, NULL, 1 },
{ "connect_roots", OPT_OFFSET(connect_roots), opts_parse_boolean, NULL,
1 },
{ "use_ucast_cache", OPT_OFFSET(use_ucast_cache), opts_parse_boolean,
NULL, 1 },
{ "log_file", OPT_OFFSET(log_file), opts_parse_charp, NULL, 0 },
@@ -743,6 +744,7 @@ void osm_subn_set_default_opt(IN osm_subn_opt_t * p_opt)
p_opt->port_profile_switch_nodes = FALSE;
p_opt->sweep_on_trap = TRUE;
p_opt->use_ucast_cache = FALSE;
+ p_opt->use_default_routing = TRUE;
p_opt->routing_engine_names = NULL;
p_opt->connect_roots = FALSE;
p_opt->lid_matrix_dump_file = NULL;
@@ -1392,6 +1394,12 @@ int osm_subn_output_conf(FILE *out, IN osm_subn_opt_t *
p_opts)
p_opts->routing_engine_names : null_str);
fprintf(out,
+ "# Fall back to default routing engine if the provided\n"
+ "# routing engine(s) failed to configure the subnet\n"
+ "use_default_routing %s\n\n",
+ p_opts->use_default_routing ? "TRUE" : "FALSE");
+
+ fprintf(out,
"# Connect roots (use FALSE if unsure)\n"
"connect_roots %s\n\n",
p_opts->connect_roots ? "TRUE" : "FALSE");
diff --git a/opensm/opensm/osm_ucast_mgr.c b/opensm/opensm/osm_ucast_mgr.c
index fbc9244..9264753 100644
--- a/opensm/opensm/osm_ucast_mgr.c
+++ b/opensm/opensm/osm_ucast_mgr.c
@@ -979,8 +979,11 @@ int osm_ucast_mgr_process(IN osm_ucast_mgr_t * p_mgr)
}
if (!p_osm->routing_engine_used &&
- p_osm->subn.opt.no_fallback_routing_engine != TRUE) {
- /* If configured routing algorithm failed, use default MinHop */
+ p_osm->default_routing_engine) {
+ /*
+ * If configured routing algorithms failed,
+ * and default routing has been set, use it.
+ */
struct osm_routing_engine *r = p_osm->default_routing_engine;
r->build_lid_matrices(r->context);
--
1.5.1.4
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
