zebra: Improve commenting for group requeue case

The commenting for why we would need to requeue a
group from the kernel to be later processed was not
sufficient. Add a better explanation for the flow
and state of the system.

Signed-off-by: Stephen Worley <sworley@cumulusnetworks.com>

diff --git a/zebra/zebra_nhg.c b/zebra/zebra_nhg.c
index 1b9440218b666011c6765b6c7f202180f02ae1ca..418c1102bdb0f9f2f47b6725e65e2ab6ca231f79 100644 (file)
--- a/zebra/zebra_nhg.c
+++ b/zebra/zebra_nhg.c
@@ -883,10 +883,17 @@ int nhg_ctx_process(struct nhg_ctx *ctx)
                ret = nhg_ctx_process_new(ctx);
                if (nhg_ctx_get_count(ctx) && ret == -ENOENT
                    && nhg_ctx_get_status(ctx) != NHG_CTX_REQUEUED) {
-                       /* Depends probably came before group, re-queue.
+                       /**
+                        * We have entered a situation where we are
+                        * processing a group from the kernel
+                        * that has a contained nexthop which
+                        * we have not yet processed.
                         *
-                        * Only going to retry once, hence just using status
-                        * flag rather than counter.
+                        * Re-enqueue this ctx to be handled exactly one
+                        * more time (indicated by the flag).
+                        *
+                        * By the time we get back to it, we
+                        * should have processed its depends.
                         */
                        nhg_ctx_set_status(ctx, NHG_CTX_NONE);
                        if (queue_add(ctx) == 0) {