myxce/src/lib/testsuite/test_ctrlstatus.c

/* test_ctrlstatus.c — Unity tests for the controller-status (ctrlstatus)
 *                     libxce client reply path (controller-status task 3.1).
 *
 * Proves the reply path that xcecli_ctrlstatus relies on:
 *   1. A true ENCODE→DECODE round-trip of xce_ctrlstatus_t through the
 *      generated Cap'n Proto codec (encode_CtrlStatus/write_CtrlStatus →
 *      bytes → read_CtrlStatus/decode_CtrlStatus): every status field,
 *      including the peers list and the configChecksum hex Text, survives
 *      the wire (Req 1.1 — the client decodes every field a handler emits).
 *   2. xcecli_ctrlstatus(NULL, ...) returns a typed negative rather than a
 *      partial/empty success (Req 7.3 — no reachable controller ⇒ typed
 *      failure).
 *
 * A full live round-trip (client → running controller → handler reply)
 * needs a controller and is the E2E task; here we prove the encode/decode
 * round-trip + the no-context typed failure only.
 */
#include <errno.h>
#include <stdlib.h>
#include <string.h>
#include <unity/unity.h>

#include "client.h"
#include "config.h" /* XCE_MAXHOSTNAME_LEN */
#include "request.h"
#include "xce/types.h"
#include "xcectx.h" /* xce_ctx_t, xce_handle_reply (bug-1077 regression) */

/* generated Cap'n Proto codec for CtrlStatus / CtrlPeer */
#include "controller.capnp.h"

/* ── encode→decode round-trip of xce_ctrlstatus_t ───────────────────────── */

static void test_ctrlstatus_roundtrip_all_fields(void)
{
    /* Arrange: a fully-populated status with two configured peers */
    xce_ctrlpeer_t p0 = {
        .name = "ctrl-a", .host = "10.0.0.1", .port = 9991
    };
    xce_ctrlpeer_t p1 = {
        .name = "ctrl-b", .host = "10.0.0.2", .port = 9992
    };
    xce_ctrlpeer_t* peerlist[2] = { &p0, &p1 };

    xce_ctrlstatus_t in;
    memset(&in, 0, sizeof(in));
    in.clusterName            = "lab-cluster";
    in.primaryName            = "ctrl-a";
    in.responderName          = "ctrl-b";
    in.responderRole          = 3;        /* STANDBY */
    in.term                   = 42;
    in.configVersion          = 0x1122334455667788ULL;
    in.configChecksum         = "1122334455667788aabbccddeeff0011"
                                "2233445566778899aabbccddeeff0011";
    in.committedConfigVersion = 0x1122334455667788ULL;
    in.committedLsn           = 5000;
    in.appliedLsn             = 4999;
    in.fenced                 = 1;
    in.peers.count            = 2;
    in.peers.list             = peerlist;
    in.peerCount              = 2;

    /* Act: encode into a capnp segment, serialize to bytes, then decode
     * a fresh struct back out — mirrors xce_encode/xce_decode. */
    struct capn          c;
    struct capn_segment* cs;
    struct CtrlStatus    cap_in;
    CtrlStatus_ptr       sp;

    capn_init_malloc(&c);
    cs = capn_root(&c).seg;

    memset(&cap_in, 0, sizeof(cap_in));
    encode_CtrlStatus(cs, &cap_in, &in);

    sp = new_CtrlStatus(cs);
    write_CtrlStatus(&cap_in, sp);
    capn_setp(capn_root(&c), 0, sp.p);

    size_t   sz  = capn_size(&c);
    uint8_t* buf = (uint8_t*)malloc(sz);
    TEST_ASSERT_NOT_NULL(buf);
    TEST_ASSERT_EQUAL_size_t(sz, capn_write_mem(&c, buf, sz, 0));
    capn_free(&c);

    /* decode side */
    struct capn       dc;
    struct CtrlStatus cap_out;
    CtrlStatus_ptr    dp;
    xce_ctrlstatus_t  out;

    capn_init_mem(&dc, buf, sz, 0);
    dp.p = capn_getp(capn_root(&dc), 0, 1);
    read_CtrlStatus(&cap_out, dp);
    memset(&out, 0, sizeof(out));
    decode_CtrlStatus(&out, &cap_out);

    /* Assert: every scalar/text field survived */
    TEST_ASSERT_NOT_NULL(out.clusterName);
    TEST_ASSERT_EQUAL_STRING("lab-cluster", out.clusterName);
    TEST_ASSERT_NOT_NULL(out.primaryName);
    TEST_ASSERT_EQUAL_STRING("ctrl-a", out.primaryName);
    TEST_ASSERT_NOT_NULL(out.responderName);
    TEST_ASSERT_EQUAL_STRING("ctrl-b", out.responderName);
    TEST_ASSERT_EQUAL_UINT8(3, out.responderRole);
    TEST_ASSERT_EQUAL_UINT64(42, out.term);
    TEST_ASSERT_EQUAL_UINT64(0x1122334455667788ULL, out.configVersion);
    TEST_ASSERT_NOT_NULL(out.configChecksum);
    TEST_ASSERT_EQUAL_STRING(in.configChecksum, out.configChecksum);
    TEST_ASSERT_EQUAL_UINT64(0x1122334455667788ULL, out.committedConfigVersion);
    TEST_ASSERT_EQUAL_UINT64(5000, out.committedLsn);
    TEST_ASSERT_EQUAL_UINT64(4999, out.appliedLsn);
    TEST_ASSERT_EQUAL_UINT32(1, out.fenced);
    TEST_ASSERT_EQUAL_UINT32(2, out.peerCount);

    /* Assert: the peers list survived with both entries */
    TEST_ASSERT_EQUAL_INT(2, out.peers.count);
    TEST_ASSERT_NOT_NULL(out.peers.list);
    TEST_ASSERT_NOT_NULL(out.peers.list[0]);
    TEST_ASSERT_EQUAL_STRING("ctrl-a", out.peers.list[0]->name);
    TEST_ASSERT_EQUAL_STRING("10.0.0.1", out.peers.list[0]->host);
    TEST_ASSERT_EQUAL_UINT32(9991, out.peers.list[0]->port);
    TEST_ASSERT_NOT_NULL(out.peers.list[1]);
    TEST_ASSERT_EQUAL_STRING("ctrl-b", out.peers.list[1]->name);
    TEST_ASSERT_EQUAL_STRING("10.0.0.2", out.peers.list[1]->host);
    TEST_ASSERT_EQUAL_UINT32(9992, out.peers.list[1]->port);

    free_CtrlStatus(&out);
    capn_free(&dc);
    free(buf);

    printf("test_ctrlstatus_roundtrip_all_fields: PASS\n");
}

/* An "unknown" primary (empty primaryName) must round-trip as empty/NULL,
 * never garbage — the client renders it as "unknown" (Req 1.3, 2.2). */
static void test_ctrlstatus_roundtrip_unknown_primary(void)
{
    xce_ctrlstatus_t in;
    memset(&in, 0, sizeof(in));
    in.clusterName   = "lab-cluster";
    in.primaryName   = "";          /* unknown */
    in.responderName = "ctrl-a";
    in.responderRole = 1;           /* ELECTING */
    in.peers.count   = 0;
    in.peers.list    = NULL;
    in.peerCount     = 0;

    struct capn          c;
    struct capn_segment* cs;
    struct CtrlStatus    cap_in;
    CtrlStatus_ptr       sp;

    capn_init_malloc(&c);
    cs = capn_root(&c).seg;
    memset(&cap_in, 0, sizeof(cap_in));
    encode_CtrlStatus(cs, &cap_in, &in);
    sp = new_CtrlStatus(cs);
    write_CtrlStatus(&cap_in, sp);
    capn_setp(capn_root(&c), 0, sp.p);

    size_t   sz  = capn_size(&c);
    uint8_t* buf = (uint8_t*)malloc(sz);
    TEST_ASSERT_NOT_NULL(buf);
    capn_write_mem(&c, buf, sz, 0);
    capn_free(&c);

    struct capn       dc;
    struct CtrlStatus cap_out;
    CtrlStatus_ptr    dp;
    xce_ctrlstatus_t  out;

    capn_init_mem(&dc, buf, sz, 0);
    dp.p = capn_getp(capn_root(&dc), 0, 1);
    read_CtrlStatus(&cap_out, dp);
    memset(&out, 0, sizeof(out));
    decode_CtrlStatus(&out, &cap_out);

    TEST_ASSERT_EQUAL_STRING("ctrl-a", out.responderName);
    TEST_ASSERT_EQUAL_UINT8(1, out.responderRole);
    /* empty primaryName decodes to NULL or "" — not a stale/guessed value */
    if (out.primaryName != NULL) {
        TEST_ASSERT_EQUAL_STRING("", out.primaryName);
    }
    TEST_ASSERT_EQUAL_INT(0, out.peers.count);
    TEST_ASSERT_EQUAL_UINT32(0, out.peerCount);

    free_CtrlStatus(&out);
    capn_free(&dc);
    free(buf);

    printf("test_ctrlstatus_roundtrip_unknown_primary: PASS\n");
}

/* ── no reachable controller ⇒ typed failure (Req 7.3) ──────────────────── */

static void test_xcecli_ctrlstatus_null_ctx_typed_failure(void)
{
    xce_ctrlstatus_t* status = (xce_ctrlstatus_t*)0xdeadbeef;
    int               rc     = xcecli_ctrlstatus(NULL, &status);
    /* typed negative, not a partial/empty success */
    TEST_ASSERT_EQUAL_INT(-EINVAL, rc);
    printf("test_xcecli_ctrlstatus_null_ctx_typed_failure: PASS\n");
}

/* ── authoritative (primary-only) read: task 3.2 ─────────────────────────── */

/* The authoritative call must reject a NULL/unconnected ctx with the same
 * typed negative as the any-node read — it must NEVER hand back a hint reply
 * as the authoritative result (Req 9.2). The `status` out-param is left
 * untouched on failure (it still holds the poisoned sentinel below). */
static void test_xcecli_ctrlstatus_primary_null_ctx_typed_failure(void)
{
    xce_ctrlstatus_t* status = (xce_ctrlstatus_t*)0xdeadbeef;
    int               rc     = xcecli_ctrlstatus_primary(NULL, &status);
    TEST_ASSERT_EQUAL_INT(-EINVAL, rc);
    /* never returned a hint reply as the authoritative result */
    TEST_ASSERT_EQUAL_PTR((xce_ctrlstatus_t*)0xdeadbeef, status);
    printf("test_xcecli_ctrlstatus_primary_null_ctx_typed_failure: PASS\n");
}

/* NULL status out-param is also a typed failure (boundary validation). */
static void test_xcecli_ctrlstatus_primary_null_out_typed_failure(void)
{
    int rc = xcecli_ctrlstatus_primary(NULL, NULL);
    TEST_ASSERT_EQUAL_INT(-EINVAL, rc);
    printf("test_xcecli_ctrlstatus_primary_null_out_typed_failure: PASS\n");
}

/* The pure authoritative-decision helper: a reply is authoritative iff the
 * responder's role is PRIMARY (ctrl_role_t PRIMARY == 0). A standby (1),
 * init (2), or electing (3) responder is a non-authoritative HINT — the call
 * must keep following the redirect (Req 9.1, 2.3, 3.2). */
static void test_ctrlstatus_is_authoritative_decision(void)
{
    xce_ctrlstatus_t reply;
    memset(&reply, 0, sizeof(reply));

    reply.responderRole = 0; /* PRIMARY */
    TEST_ASSERT_TRUE(xcecli_ctrlstatus_is_authoritative(&reply));

    reply.responderRole = 1; /* STANDBY */
    TEST_ASSERT_FALSE(xcecli_ctrlstatus_is_authoritative(&reply));

    reply.responderRole = 2; /* INIT */
    TEST_ASSERT_FALSE(xcecli_ctrlstatus_is_authoritative(&reply));

    reply.responderRole = 3; /* ELECTING */
    TEST_ASSERT_FALSE(xcecli_ctrlstatus_is_authoritative(&reply));

    /* NULL reply is never authoritative (defensive). */
    TEST_ASSERT_FALSE(xcecli_ctrlstatus_is_authoritative(NULL));

    printf("test_ctrlstatus_is_authoritative_decision: PASS\n");
}

/* The pure next-target helper: given a standby's hint reply, map the reported
 * primaryName to its host:port via the reply's peers list so the next query
 * can target the primary directly (the redirect step of Req 9.1). */
static void test_ctrlstatus_primary_addr_from_peers(void)
{
    xce_ctrlpeer_t  p0 = { .name = "ctrl-a", .host = "10.0.0.1", .port = 9991 };
    xce_ctrlpeer_t  p1 = { .name = "ctrl-b", .host = "10.0.0.2", .port = 9992 };
    xce_ctrlpeer_t* peerlist[2] = { &p0, &p1 };

    xce_ctrlstatus_t hint;
    memset(&hint, 0, sizeof(hint));
    hint.responderName = "ctrl-a";
    hint.responderRole = 1;        /* STANDBY */
    hint.primaryName   = "ctrl-b"; /* the standby knows ctrl-b is primary */
    hint.peers.count   = 2;
    hint.peers.list    = peerlist;
    hint.peerCount     = 2;

    char     host[XCE_MAXHOSTNAME_LEN];
    uint16_t port = 0;
    int rc = xcecli_ctrlstatus_primary_addr(&hint, host, sizeof(host), &port);
    TEST_ASSERT_EQUAL_INT(0, rc);
    TEST_ASSERT_EQUAL_STRING("10.0.0.2", host);
    TEST_ASSERT_EQUAL_UINT16(9992, port);

    printf("test_ctrlstatus_primary_addr_from_peers: PASS\n");
}

/* When the standby reports no known primary (empty primaryName) or the named
 * primary is not in the peers list, the address lookup fails (-ENOENT) so the
 * caller falls back to plain master rotation rather than targeting a bogus
 * address (Req 2.3 unknown-primary, redirect best-effort). */
static void test_ctrlstatus_primary_addr_unknown_or_missing(void)
{
    xce_ctrlpeer_t  p0 = { .name = "ctrl-a", .host = "10.0.0.1", .port = 9991 };
    xce_ctrlpeer_t* peerlist[1] = { &p0 };

    xce_ctrlstatus_t hint;
    memset(&hint, 0, sizeof(hint));
    hint.responderName = "ctrl-a";
    hint.responderRole = 3; /* ELECTING */
    hint.peers.count   = 1;
    hint.peers.list    = peerlist;
    hint.peerCount     = 1;

    char     host[XCE_MAXHOSTNAME_LEN];
    uint16_t port = 7;

    /* unknown primary (empty) */
    hint.primaryName = "";
    TEST_ASSERT_EQUAL_INT(-ENOENT, xcecli_ctrlstatus_primary_addr(
                                       &hint, host, sizeof(host), &port));

    /* NULL primary */
    hint.primaryName = NULL;
    TEST_ASSERT_EQUAL_INT(-ENOENT, xcecli_ctrlstatus_primary_addr(
                                       &hint, host, sizeof(host), &port));

    /* named primary not present in peers */
    hint.primaryName = "ctrl-z";
    TEST_ASSERT_EQUAL_INT(-ENOENT, xcecli_ctrlstatus_primary_addr(
                                       &hint, host, sizeof(host), &port));

    /* NULL/zero args are rejected */
    hint.primaryName = "ctrl-a";
    TEST_ASSERT_EQUAL_INT(-EINVAL, xcecli_ctrlstatus_primary_addr(
                                       NULL, host, sizeof(host), &port));
    TEST_ASSERT_EQUAL_INT(-EINVAL, xcecli_ctrlstatus_primary_addr(
                                       &hint, NULL, sizeof(host), &port));
    TEST_ASSERT_EQUAL_INT(
        -EINVAL, xcecli_ctrlstatus_primary_addr(&hint, host, 0, &port));
    TEST_ASSERT_EQUAL_INT(
        -EINVAL,
        xcecli_ctrlstatus_primary_addr(&hint, host, sizeof(host), NULL));

    printf("test_ctrlstatus_primary_addr_unknown_or_missing: PASS\n");
}

/* ── reply-union type-confusion regression (bug-1077) ───────────────────────
 *
 * ctx->reply.u is a single untagged union shared by every opcode (debugmsg /
 * get / jobid / clusterconf / ctrlstatus all alias the same 8 bytes). A
 * failover-aborted reslist/joblist leaves ctx->reply.u.get pointing at the
 * caller's (now-dead) stack `getreply`. When the next ctrlstatus reply lands on
 * a worker thread, handle_reply must NOT treat that foreign, non-NULL slot as a
 * CtrlStatus* and free it — doing so dereferenced a dead stack address inside
 * free_CtrlStatus and crashed (ASan stack-use-after-return at
 * controller.capnp.c:11638).
 *
 * This drives xce_handle_reply directly with a poisoned union slot and proves
 * the foreign occupant is left untouched (no free, no deref). Pre-fix, the
 * ctrlstatus branch frees the sentinel and the explicit free() below is a
 * double-free that ASan aborts on. */
static void test_handle_reply_ctrlstatus_does_not_free_foreign_slot(void)
{
    xce_ctx_t ctx;
    memset(&ctx, 0, sizeof(ctx));

    /* A prior get/reslist leaked a non-CtrlStatus pointer into the shared slot.
     * Sized as a zeroed xce_ctrlstatus_t so that, even on the buggy path,
     * free_CtrlStatus(sentinel) only walks all-NULL fields (no wild deref) and
     * the regression surfaces as a clean double-free rather than a hard crash. */
    xce_ctrlstatus_t* sentinel = (xce_ctrlstatus_t*)calloc(1, sizeof(*sentinel));
    TEST_ASSERT_NOT_NULL(sentinel);
    ctx.reply.u.get = (dataset_t*)sentinel; /* same 8 bytes as u.ctrlstatus */

    /* A genuine ctrlstatus reply arrives for this ctx. */
    xce_ctrlstatus_t* cs = (xce_ctrlstatus_t*)calloc(1, sizeof(*cs));
    TEST_ASSERT_NOT_NULL(cs);
    cs->clusterName = strdup("eda1");

    xce_reply_t reply;
    memset(&reply, 0, sizeof(reply));
    reply.opcode     = OPCode_ctrlstatus;
    reply.status     = 0;
    reply.data_which = RequestReply_data_ctrlstatus;
    reply.data.ctrlstatus = cs;

    int rc = xce_handle_reply(&ctx, &reply);

    /* The genuine reply is now owned by the ctx (MOVE_PTR nulls the source) … */
    TEST_ASSERT_EQUAL_INT(1, rc);
    TEST_ASSERT_EQUAL_PTR(cs, ctx.reply.u.ctrlstatus);
    TEST_ASSERT_NULL(reply.data.ctrlstatus);

    /* … and the foreign sentinel was NOT freed by the reply handler. Freeing it
     * here is the sole owner's free; on the buggy code it is the second free. */
    free(sentinel);

    free_CtrlStatus(cs);
    free(cs);

    printf("test_handle_reply_ctrlstatus_does_not_free_foreign_slot: PASS\n");
}

/* ── Suite entry point ──────────────────────────────────────────────────── */

void test_ctrlstatus(void)
{
    RUN_TEST(test_handle_reply_ctrlstatus_does_not_free_foreign_slot);
    RUN_TEST(test_ctrlstatus_roundtrip_all_fields);
    RUN_TEST(test_ctrlstatus_roundtrip_unknown_primary);
    RUN_TEST(test_xcecli_ctrlstatus_null_ctx_typed_failure);
    RUN_TEST(test_xcecli_ctrlstatus_primary_null_ctx_typed_failure);
    RUN_TEST(test_xcecli_ctrlstatus_primary_null_out_typed_failure);
    RUN_TEST(test_ctrlstatus_is_authoritative_decision);
    RUN_TEST(test_ctrlstatus_primary_addr_from_peers);
    RUN_TEST(test_ctrlstatus_primary_addr_unknown_or_missing);
}