WIP: item 2.2 -- bound the VM registry (not complete, do not check off)

Partial work toward FABRIC.md punch list item 2.2. Adds the
STADIUM_MAX_VM_COUNT Kconfig symbol (default 4, per item 1.5) wired
through Makefile.starkernel, a new CAPSULE_RUN_ERR_FLEET_FULL result
code, and a vm_registry_live_count() helper in capsule_birth.c that
counts LIVE VMs only (distinct from the existing monotonic
vm_registry_count, which never decrements on death).

NOT YET DONE: nothing calls vm_registry_live_count() yet -- the actual
birth-refusal check is not wired into capsule_birth_baby(). Not built,
not boot-tested. FABRIC.md's item 2.2 checkbox is deliberately left
unchecked; this commit exists only to save in-progress work before a
pause, not to claim the item complete.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

src/starkernel/capsule/capsule_birth.c

@@ -180,6 +180,25 @@ uint32_t capsule_vm_registry_count(void) {
     return vm_registry_count;
 }
 
+/* Live population, distinct from vm_registry_count above: vm_registry_count
+ * is monotonic (incremented on every vm_registry_alloc(), never decremented
+ * on death), so it counts every VM ever born, not the outer Stadium's
+ * current occupancy. FABRIC.md item 1.5's bound is on LIVE VMs -- a dead or
+ * stillborn slot doesn't hold Stadium capacity, and gating on the monotonic
+ * total would mean the fleet could never regrow after any VM's death,
+ * which contradicts Hera's own kill-then-rebirth lifecycle (TRIPOD-TEST's
+ * "K soak" check, capsule_vm_physics.c). Not exposed in the public header:
+ * only capsule_birth_baby's bound check needs it today. */
+static uint32_t vm_registry_live_count(void) {
+    vm_node_t *node = vm_registry_head;
+    uint32_t   live = 0;
+    while (node) {
+        if (node->entry.state == VM_STATE_LIVE) live++;
+        node = node->next;
+    }
+    return live;
+}
+
 int capsule_vm_find_by_name(const char *name, VMRegistryEntry *out) {
     vm_node_t *node;
     if (!name || !out) return -1;