Github user PSUdaemon commented on a diff in the pull request:
https://github.com/apache/trafficserver/pull/956#discussion_r77467242
--- Diff: iocore/eventsystem/UnixEventProcessor.cc ---
@@ -152,6 +155,59 @@ EventProcessor::start(int n_event_threads, size_t
stacksize)
#else
Debug("iocore_thread", "EThread: %d %s: %d", i, obj_name,
obj->logical_index);
#endif // HWLOC_API_VERSION
+ }
+#endif // TS_USE_HWLOC
+
+ snprintf(thr_name, MAX_THREAD_NAME_LENGTH, "[ET_NET %d]", i);
+#if TS_USE_HWLOC
+ if (obj_count > 0) {
+ hwloc_membind_policy_t mem_policy = HWLOC_MEMBIND_DEFAULT;
+ hwloc_nodeset_t nodeset = hwloc_bitmap_alloc();
+ int num_nodes = 0;
+
+ hwloc_cpuset_to_nodeset(ink_get_topology(), obj->cpuset, nodeset);
+ num_nodes =
hwloc_get_nbobjs_inside_cpuset_by_type(ink_get_topology(), obj->cpuset,
HWLOC_OBJ_NODE);
+
+ if (num_nodes == 1) {
+ mem_policy = HWLOC_MEMBIND_BIND;
+ } else if (num_nodes > 1) {
+ mem_policy = HWLOC_MEMBIND_INTERLEAVE;
+ }
+
+ if (mem_policy != HWLOC_MEMBIND_DEFAULT) {
+ hwloc_set_membind_nodeset(ink_get_topology(), nodeset, mem_policy,
HWLOC_MEMBIND_THREAD);
+ }
--- End diff --
One thing to clarify, that I think this may have been confusing on. We
don't set the memory binding for the threads that we are going to spawn for the
eventprocessor, we are just setting it for the current thread for the time we
need to allocate a new stack. We leave the newly spawned threads to have the
default memory binding, which is basically to allocate where they run.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
