From 488fd1bf49f5b39ca1b5aa3e979d9683c31bff8c Mon Sep 17 00:00:00 2001
From: Rongsong Shen <rshen@shenrs.eu>
Date: Thu, 28 May 2026 04:29:31 +0000
Subject: [PATCH] feat(ha-config-bootstrap): add peer count and cert pre-flight
 validation (task 2.3)

- Implement validate_peer_count() in startup_validation.c: -EINVAL for even N>1
- Implement validate_cert_files() with err_buf output: access() check, writes
  missing path into err_buf, returns -ENOENT
- Wire validate_peer_count() in main.c after peer list parse; fatal exit on fail
- Wire validate_cert_files() in main.c before SSL init; fatal exit on fail
- Fix ASAN leak: SSL_CTX_free in ha_fini() with NULL guard
- Fix double-free: null c->ha.ssl_ctx in cluster_fini() before ha_fini(),
  so iomgr_fini() retains sole SSL_CTX ownership (design per ha.h comment)
- All 96 tests pass; all 12 ha_startup tests GREEN
---
 .kiro/specs/ha-config-bootstrap/tasks.md |  2 +-
 src/controller/cluster.c                 |  2 ++
 src/controller/ha.c                      |  6 ++++++
 src/controller/main.c                    | 26 +++++++++++++++++++++++-
 4 files changed, 34 insertions(+), 2 deletions(-)

diff --git a/.kiro/specs/ha-config-bootstrap/tasks.md b/.kiro/specs/ha-config-bootstrap/tasks.md
index 5ba049b3..f7a8580c 100644
--- a/.kiro/specs/ha-config-bootstrap/tasks.md
+++ b/.kiro/specs/ha-config-bootstrap/tasks.md
@@ -45,7 +45,7 @@ Dependency order: startup validation (D) unblocks vtaskmgr before election (A);
   - _Requirements: 4.3, 4.4_
   - _Boundary: Startup Validation (main.c, cluster.h, cluster.c)_
 
-- [ ] 2.3 Add peer count validation and cert pre-flight checks
+- [x] 2.3 Add peer count validation and cert pre-flight checks
   - In `main.c` after argument parsing: if `n_peers % 2 == 0 && n_peers > 1` → `fprintf(stderr, ...)` + `exit(1)` with message explaining odd-N requirement
   - Call `access(cert_path, R_OK)` and `access(key_path, R_OK)` in `main.c` before any SSL init; fatal exit with full path on failure
   - Fix ASAN leak: add `SSL_CTX_free(ha->ssl_ctx)` in `ha_fini()`
diff --git a/src/controller/cluster.c b/src/controller/cluster.c
index 868d8eb4..0d0b11f0 100644
--- a/src/controller/cluster.c
+++ b/src/controller/cluster.c
@@ -1047,6 +1047,8 @@ void cluster_fini(cluster_t* c)
         c->ha.standby_endpoints = NULL;
     }
 
+    /* iomgr owns the SSL_CTX lifetime; null ha's copy so ha_fini doesn't double-free */
+    c->ha.ssl_ctx = NULL;
     /* Stop HA watchdog thread before tearing down other subsystems — P1-5 */
     ha_fini(&c->ha);
 
diff --git a/src/controller/ha.c b/src/controller/ha.c
index aa7f84ea..684b4d23 100644
--- a/src/controller/ha.c
+++ b/src/controller/ha.c
@@ -454,6 +454,12 @@ void ha_fini(ha_state_t *ha)
         return;
     ha_stop_watchdog(ha);  /* join watchdog thread before destroying mutex */
     pthread_mutex_destroy(&ha->lock);
+
+    /* Fix ASAN leak: free the SSL context if one was allocated */
+    if (ha->ssl_ctx) {
+        SSL_CTX_free((SSL_CTX *)ha->ssl_ctx);
+        ha->ssl_ctx = NULL;
+    }
 }
 
 /* ── Redirect helpers used by join.c ───────────────────────── */
diff --git a/src/controller/main.c b/src/controller/main.c
index 4cb88cc6..b3f8de0c 100644
--- a/src/controller/main.c
+++ b/src/controller/main.c
@@ -11,8 +11,11 @@
 
 extern int myapp_run(myconf_t*);
 
-/* forward declaration — implemented in startup_validation.c */
+/* forward declarations — implemented in startup_validation.c */
 int validate_vcpus(int vcpus);
+int validate_peer_count(int n_peers);
+int validate_cert_files(const char *cert_path, const char *key_path,
+                        char *err_buf, size_t err_len);
 
 static char doc[]      = "xce_controller";
 static char args_doc[] = "";
@@ -166,6 +169,17 @@ int main(int argc, char* argv[])
                     arguments.peers_str);
             return 1;
         }
+
+        /* Req 4.1: reject even-N peer counts (cannot form safe majority) */
+        if (validate_peer_count(myconf.peers_cfg.candidates_count) != 0) {
+            fprintf(stderr,
+                    "xce_controller: --peers: even peer count (%d) is not "
+                    "allowed; Raft quorum requires an odd number of nodes "
+                    "(1, 3, 5, ...)\n",
+                    myconf.peers_cfg.candidates_count);
+            return 1;
+        }
+
         /* CA cert derived from workdir — same infrastructure as certfile/keyfile */
         snprintf(myconf.peers_cfg.ca_cert_path,
                  sizeof(myconf.peers_cfg.ca_cert_path),
@@ -185,6 +199,16 @@ int main(int argc, char* argv[])
         INFO("No --peers flag; running as single-node primary");
     }
 
+    /* Req 4.2: cert pre-flight — fail early with a clear path in the message */
+    {
+        char cert_err[XCE_MAXPATH_LEN + 64];
+        if (validate_cert_files(cert_file, key_file,
+                                cert_err, sizeof(cert_err)) != 0) {
+            fprintf(stderr, "xce_controller: %s\n", cert_err);
+            return 1;
+        }
+    }
+
     myapp_run(&myconf);
 
     mylog_fini();