Files
deusdata--codebase-memory-mcp/tests/test_platform.c
Martin Vogel 0baddf7887 daemon: permanent lifecycle, daemon-backed CLI/hooks, real-Windows hardening, long-path launcher transactions
Daemon lifecycle and Windows correctness, verified on a real Windows 11
ARM64 VM through the maintained test-infrastructure/vm drivers, plus the
macOS and Linux arm64 suites and the container lint gate.

Daemon lifecycle:

- daemon start/stop/status subcommands. `daemon start` launches a
  PERMANENT daemon (spawn shape is byte-exact argv; survives idle
  periods and session ends) and reports an already-active daemon
  instead of failing. Permanence is honored at every stop latch:
  last-committed-client disconnect, host initial-client window,
  coordinator release, and application final-session close — a
  permanent daemon also keeps admitting new sessions after its last
  one closes.
- daemon stop refuses while sessions are active and lists the blocking
  peers (pid/role) that must finish first; an idle daemon drains
  through the activation-shutdown machinery with the ACK ordered after
  connection interrupts. A second stop is idempotent. The wire ops are
  no-cohort first-frame requests with peer fingerprint authentication,
  so stop/status never conflict with an exact-build admission gate.
- One-shot CLI commands now execute through the daemon (index workers
  keep their local supervised path). A cold CLI run that had to spawn a
  temporary daemon prints a hint that `daemon start` removes the
  per-command startup tax; a warm daemon is recycled silently.
- Hooks are connect-only fail-open: with no daemon present the hook
  emits a visible, rate-limited notice (Claude-dialect systemMessage
  plus stderr for other dialects) and always exits 0 — augmentation is
  never allowed to block the caller's tool use.
- Version skew: a newer-build client automatically drains an
  older-build permanent daemon (strict semantic-version triples only;
  dev builds never auto-drain) and the build-conflict message names
  `cbm daemon stop` as the manual escape hatch.

Windows IPC/runtime (real-VM verified):

- ipc(win): persistent pending overlapped ConnectNamedPipe. The accept
  path used to destroy its listening pipe instance on every 20 ms poll
  timeout; a client attaching in the teardown window was severed or left
  on an orphaned pipe object whose HELLO no server handle could ever
  read, absorbing the connect until the client's own timeout expired.
  The pending connect now survives poll timeouts and nothing is
  destroyed while a client could be attaching.
- ipc(win): drain-before-close for final responses. Closing a named-pipe
  server handle can discard a just-sent response before the peer reads
  it (POSIX stream sockets never lose buffered data on close). A bounded
  cbm_daemon_ipc_connection_drain (read-until-peer-EOF; no-op on POSIX,
  immediate on interrupted connections) now precedes close in
  runtime_worker_finish and runtime_reject_inline, so hello-conflict,
  capacity and disconnect acknowledgements reliably reach the peer.
- runtime: CLOSE_INTENT wire frame. A Windows named-pipe client has no
  transport half-close, so close_begin now announces departure with an
  explicit frame (ordered after APPLICATION_CANCEL, before the local
  interrupt); the server releases the client's admission on receipt
  instead of waiting for the handle to close. Admission-drop timing is
  now identical to POSIX shutdown() semantics on every platform.
- runtime(win): client close cancellation. close_begin serializes with
  request publication under the send lock, best-effort sends the active
  token's APPLICATION_CANCEL frame, then interrupts local I/O; the
  server cancels MCP/subprocess work promptly. Contract tests accept
  both correct outcomes (interrupted transport or decoded CANCELLED).
- runtime: activation acknowledgement ordering. The activation ACK is
  the requester's license to act on "snapshotted and draining", so every
  connection interrupt is now initiated before the ACK is sent; a
  session could previously get one more request serviced after the
  requester observed the ACK.
- service(win): deadline-bounded private-file prepare. The conflict-log
  prepare retry loop (100 x Sleep(2), which rounds up to the ~16 ms
  timer granularity) burned ~1.6 s against permanently obstructed paths,
  stalling hello rejections past the client's timeout. The retry budget
  is now a 250 ms deadline; transient share collisions still retry.
- subprocess(win): cmd.exe /C payload encoder quotes metacharacters
  correctly (root cause of the git-on-Windows failure cluster).
- watcher: SHA-256 buffer sizing (CBM_SZ_64 -> CBM_SZ_128) and a native
  Windows stop/unwatch cancellation test with exact-image verification.
- httpd: send_all writes in bounded 64 KiB slices. A single giant
  nonblocking send() on Windows is absorbed wholesale into AFD kernel
  buffering regardless of SO_SNDBUF, so send deadlines and interrupts
  could never engage against a slow peer (and the full payload was
  pinned in nonpaged pool). Slicing restores a deterministic
  backpressure point; a test hook pins SO_SNDBUF for the deadline and
  interrupt tests.
- ui/http: shutdown lifecycle — interrupt checks, response-wide send
  deadline, explicit connection states, refusal to free a server while
  a listener-owned connection is active.

Windows long-path support:

- Central path-aware wide conversion (canonicalize via GetFullPathNameW
  and prepend the extended-length prefix for absolute paths >=240) at
  the compat chokepoints (cbm_fopen/compat_fs/mkstemp/mkdtemp), sqlite
  store opens, and the daemon build-fingerprint/log paths. Deep managed
  installs (a 64-hex generation directory routinely exceeds MAX_PATH)
  now index, stage and activate correctly.
- activation transaction: its own file APIs and the component-walking
  ancestry validators now operate in the extended-length namespace;
  the launcher path is canonicalized (and prefixed when deep) once at
  entry so every downstream exact-string comparison stays
  form-consistent.
- Executable self-resolution uses the wide APIs (GetModuleFileNameW,
  GetFileAttributesW) so non-ASCII install paths survive argv[0]
  resolution.

Windows launcher install/uninstall transaction:

- FileRenameInfoEx names are NUL-terminated in an over-allocated
  buffer. FileNameLength governs per the contract, but filter drivers
  read FileName as NUL-terminated and appended adjacent heap bytes to
  created names — a flaky, garbage-suffixed rename target. Both the CLI
  and the launcher rename helpers are fixed.
- Uninstall retires state via rename-aside (.cbm ->
  .cbm-retired-v1-<tag>-<pid>) with the retired tag shortened to 16 hex
  chars so the bare rename target stays under the FileRenameInfoEx
  NT-conversion ceiling at guard depths; 64 bits still uniquely
  identify the generation.
- When the running launcher's mapped generation backings pin .cbm
  against rename, the backings are relocated to activation-<pid>-N
  .retired tombstones beside the install (a mapped image may be renamed,
  never deleted; the launcher's liveness-guarded sweep reclaims stale
  tombstones). Every relocation is recorded, and a FAILED uninstall
  reverses the moves after restoring .cbm — via MoveFileExW with
  extended-length paths on both arguments, since the deep generation
  target is beyond the handle-based rename's bare-path reach — so a
  restored install keeps its generation backings and stays runnable.
- After a committed uninstall the retired tree's backings are relocated
  out so the tree is shallow enough for the detached cleanup's rd, and
  the cleanup's working directory strips the extended-length prefix
  (CreateProcessW lpCurrentDirectory silently ignores prefixed paths).
- Files created under Administrators-default-owner directories
  (CopyFileW destinations, CREATE_NEW tombstones, probe directories)
  are explicitly owner-stamped so the exact-owner validators hold on
  runner images; guard fixtures stamp hand-built trees the same way.

Diagnostics, tests and infra:

- diagnostics: discovery is now an always-delivered JSON control record
  (new cbm_log_control) that survives CBM_LOG_LEVEL suppression and
  paths containing spaces; placement honors $TMPDIR with /tmp fallback
  via a diagnostics-local helper; the soak parser reads the JSON record;
  documented in docs/CONFIGURATION.md. Red-first coverage for suppressed
  log levels, TMPDIR-with-spaces, and native Windows output-contract
  assertions.
- tests(win): daemon_ipc/daemon_frontend fixtures now build endpoint
  parents with production-shaped ancestry (LocalAppData on Windows, via
  th_secure_runtime_parent_new) — the runtime ancestry validation
  correctly refuses temp roots whose ancestors grant mutation rights to
  Authenticated Users (C:/msys64/tmp, GitHub-runner work dirs) — and
  drive the documented startup-owner publication flow before reading
  generation-bound endpoint addresses. This turns the 26 Windows
  failures previously visible in CI's full-test job green without
  weakening any validation.
- tests(win): the launcher guard covers the full permanent-launcher
  contract including failed-uninstall restore and immediate reinstall
  after uninstall; new daemon lifecycle and reworked hook-augment
  guards run the start/recycle/stop flow end to end.
- tests: CBM_SKIP_PERF is now actually consumed by the test runner
  (it was set by CI but never read, so perf suites ran everywhere);
  four throughput/bench suites are classified as perf, the heavy
  store_arch suite moved to the slow-timeout tier, and two
  wall-clock-sensitive assertions were rewritten as invariant checks
  with coarse hang-detector backstops.
- build/test infra: build-dir safety contract, UI dev-proxy security
  contract, soak daemon-recovery contract, path-safety helper, the
  Windows VM worktree-sync contract wired into scripts/test.sh, and
  vm/win.sh guards building its clean embedded-UI product in an
  isolated BUILD_DIR so it cannot clobber the incremental test build.
  provision-windows.sh now installs Node.js for the guards UI build.

Signed-off-by: Martin Vogel <martin.vogel.tech@gmail.com>
2026-07-21 15:32:15 +02:00

651 lines
22 KiB
C

/*
* test_platform.c — RED phase tests for foundation/platform.
*/
#include "test_framework.h"
#include "../src/foundation/compat.h" /* cbm_setenv / cbm_unsetenv (Windows-portable) */
#include "../src/foundation/compat_fs.h"
#include "../src/foundation/constants.h"
#include "../src/foundation/compat_thread.h"
#include "../src/foundation/platform.h"
#include "../src/foundation/platform_internal.h"
#include "../src/foundation/system_info_internal.h"
#include <stdatomic.h>
#include <stdlib.h>
#include <unistd.h>
#ifdef __linux__
/* Linux-only cgroup tests need stdio for FILE*, stdlib for mkdtemp,
* string for strncpy/strchr, sys/stat for mkdir, dirent for the
* shell-free recursive teardown. */
#include <dirent.h>
#include <stdio.h>
#include <string.h>
#include <sys/stat.h>
#endif
enum { PLATFORM_TIME_THREADS = 8 };
#include <stdio.h>
#include <string.h>
/* Worker staging files land under CBM_CACHE_DIR, which users may place at
* non-ASCII paths. On Windows the templates must round-trip through the wide
* APIs; the ANSI CRT (_mktemp/_open) mangles UTF-8 bytes and fails. */
/* Store DB paths embed full repo-path-derived project names; deep repo or
* runner paths overflow the legacy 260-char limit. The compat file layer must
* carry absolute paths of any length (extended-length \\?\ form on
* Windows). */
TEST(platform_file_apis_survive_max_path_overflow) {
char base[CBM_SZ_512];
int written = snprintf(base, sizeof(base), "/tmp/cbm-longpath-XXXXXX");
ASSERT_TRUE(written > 0 && written < (int)sizeof(base));
ASSERT_NOT_NULL(cbm_mkdtemp(base));
enum { LONG_SEGMENTS = 5 };
static const char segment[] =
"segment-abcdefghijklmnopqrstuvwxyz0123456789-abcdefghijklmnop";
char deep[CBM_SZ_1K];
written = snprintf(deep, sizeof(deep), "%s", base);
ASSERT_TRUE(written > 0 && written < (int)sizeof(deep));
for (int index = 0; index < LONG_SEGMENTS; index++) {
written = snprintf(deep + strlen(deep), sizeof(deep) - strlen(deep), "/%s", segment);
ASSERT_TRUE(written > 0);
}
ASSERT_TRUE(strlen(deep) > 300);
ASSERT_TRUE(cbm_mkdir_p(deep, 0700));
char file_path[CBM_SZ_1K];
written = snprintf(file_path, sizeof(file_path), "%s/store.db", deep);
ASSERT_TRUE(written > 0 && written < (int)sizeof(file_path));
FILE *file = cbm_fopen(file_path, "wb");
ASSERT_NOT_NULL(file);
ASSERT_GTE(fputs("deep\n", file), 0);
ASSERT_EQ(fclose(file), 0);
ASSERT_TRUE(cbm_file_exists(file_path));
ASSERT_TRUE(cbm_is_dir(deep));
ASSERT_EQ(cbm_unlink(file_path), 0);
char cursor[CBM_SZ_1K];
(void)snprintf(cursor, sizeof(cursor), "%s", deep);
for (int index = 0; index < LONG_SEGMENTS; index++) {
ASSERT_EQ(cbm_rmdir(cursor), 0);
char *slash = strrchr(cursor, '/');
ASSERT_NOT_NULL(slash);
*slash = '\0';
}
(void)cbm_rmdir(base);
PASS();
}
TEST(platform_mkstemp_and_mkdtemp_survive_non_ascii_directory) {
char base[CBM_SZ_256];
int written = snprintf(base, sizeof(base), "/tmp/cbm-utf8-éè-XXXXXX");
ASSERT_TRUE(written > 0 && written < (int)sizeof(base));
ASSERT_NOT_NULL(cbm_mkdtemp(base));
char file_template[CBM_SZ_512];
written = snprintf(file_template, sizeof(file_template), "%s/.probe-XXXXXX", base);
ASSERT_TRUE(written > 0 && written < (int)sizeof(file_template));
int descriptor = cbm_mkstemp(file_template);
bool created = descriptor >= 0;
if (created) {
#ifdef _WIN32
_close(descriptor);
#else
close(descriptor);
#endif
(void)cbm_unlink(file_template);
}
(void)cbm_rmdir(base);
ASSERT_TRUE(created);
/* The returned path must keep the caller's UTF-8 directory intact. */
ASSERT_NOT_NULL(strstr(file_template, "éè"));
PASS();
}
TEST(platform_counter_scaling_avoids_intermediate_overflow) {
const uint64_t frequency = UINT64_C(10000000);
const uint64_t counter = UINT64_C(20000000000);
ASSERT(cbm_platform_scale_counter_ns(counter, frequency) == UINT64_C(2000000000000));
ASSERT(cbm_platform_scale_counter_ns(UINT64_MAX, UINT64_MAX) == UINT64_C(1000000000));
ASSERT(cbm_platform_scale_counter_ns(UINT64_MAX, 1) == UINT64_MAX);
PASS();
}
TEST(platform_counter_scaling_preserves_monotonic_deadlines) {
const uint64_t frequency = UINT64_C(10000000);
const uint64_t counter = UINT64_MAX / UINT64_C(1000000000);
const uint64_t now_ns = cbm_platform_scale_counter_ns(counter, frequency);
const uint64_t next_ns = cbm_platform_scale_counter_ns(counter + 1, frequency);
const uint64_t deadline_ns = cbm_platform_scale_counter_ns(counter + 5 * frequency, frequency);
ASSERT(next_ns >= now_ns);
ASSERT(deadline_ns > now_ns);
ASSERT(deadline_ns - now_ns == UINT64_C(5000000000));
PASS();
}
typedef struct {
atomic_int *ready;
atomic_bool *go;
uint64_t value;
} platform_time_worker_t;
static void *platform_time_first_call(void *opaque) {
platform_time_worker_t *worker = opaque;
(void)atomic_fetch_add_explicit(worker->ready, 1, memory_order_acq_rel);
while (!atomic_load_explicit(worker->go, memory_order_acquire)) {
atomic_signal_fence(memory_order_seq_cst);
}
worker->value = cbm_now_ns();
return NULL;
}
TEST(platform_now_ns_concurrent_first_call) {
cbm_thread_t threads[PLATFORM_TIME_THREADS];
platform_time_worker_t workers[PLATFORM_TIME_THREADS];
atomic_int ready;
atomic_bool go;
atomic_init(&ready, 0);
atomic_init(&go, false);
size_t created = 0;
for (; created < PLATFORM_TIME_THREADS; created++) {
workers[created] = (platform_time_worker_t){
.ready = &ready,
.go = &go,
};
if (cbm_thread_create(&threads[created], 0, platform_time_first_call, &workers[created]) !=
0) {
break;
}
}
while (created == PLATFORM_TIME_THREADS &&
atomic_load_explicit(&ready, memory_order_acquire) != PLATFORM_TIME_THREADS) {
atomic_signal_fence(memory_order_seq_cst);
}
atomic_store_explicit(&go, true, memory_order_release);
for (size_t index = 0; index < created; index++) {
(void)cbm_thread_join(&threads[index]);
}
ASSERT_EQ(created, PLATFORM_TIME_THREADS);
for (size_t index = 0; index < created; index++) {
ASSERT_GT(workers[index].value, 0);
}
PASS();
}
TEST(platform_now_ns) {
uint64_t t1 = cbm_now_ns();
ASSERT_GT(t1, 0);
/* Busy-wait a tiny bit */
for (volatile int i = 0; i < 100000; i++) {}
uint64_t t2 = cbm_now_ns();
ASSERT_GT(t2, t1);
PASS();
}
TEST(platform_now_ms) {
uint64_t t1 = cbm_now_ms();
ASSERT_GT(t1, 0);
PASS();
}
TEST(platform_nprocs) {
int n = cbm_nprocs();
ASSERT_GT(n, 0);
ASSERT_LT(n, 10000); /* sanity */
PASS();
}
TEST(platform_file_exists) {
/* This test file should exist */
ASSERT_TRUE(cbm_file_exists("tests/test_platform.c"));
ASSERT_FALSE(cbm_file_exists("nonexistent_file_xyz.txt"));
PASS();
}
TEST(platform_is_dir) {
ASSERT_TRUE(cbm_is_dir("tests"));
ASSERT_FALSE(cbm_is_dir("tests/test_platform.c"));
ASSERT_FALSE(cbm_is_dir("nonexistent_dir"));
PASS();
}
TEST(platform_file_size) {
int64_t sz = cbm_file_size("tests/test_platform.c");
ASSERT_GT(sz, 0);
ASSERT_EQ(cbm_file_size("nonexistent_file_xyz.txt"), -1);
PASS();
}
TEST(platform_mmap) {
/* mmap this test file and verify first bytes */
size_t sz = 0;
void *data = cbm_mmap_read("tests/test_platform.c", &sz);
ASSERT_NOT_NULL(data);
ASSERT_GT(sz, 0);
/* First line should be the comment */
ASSERT(memcmp(data, "/*", 2) == 0);
cbm_munmap(data, sz);
PASS();
}
TEST(platform_mmap_nonexistent) {
size_t sz = 0;
void *data = cbm_mmap_read("nonexistent_xyz.txt", &sz);
ASSERT_NULL(data);
PASS();
}
typedef struct {
const char *home;
const char *cache;
atomic_int *ready;
atomic_int *release;
} platform_path_thread_result_t;
static void *platform_capture_path_buffers(void *opaque) {
platform_path_thread_result_t *result = opaque;
result->home = cbm_get_home_dir();
result->cache = cbm_resolve_cache_dir();
atomic_fetch_add_explicit(result->ready, 1, memory_order_release);
while (atomic_load_explicit(result->release, memory_order_acquire) == 0) {
cbm_usleep(1000);
}
return NULL;
}
/* The daemon dispatches different sessions concurrently. Path helpers may
* retain their historical return-pointer API, but each live thread needs
* separate storage; process-global mutable buffers are a C data race. */
TEST(platform_path_helpers_use_per_thread_storage) {
const char *saved_home = getenv("HOME");
const char *saved_cache = getenv("CBM_CACHE_DIR");
char *saved_home_copy = saved_home ? strdup(saved_home) : NULL;
char *saved_cache_copy = saved_cache ? strdup(saved_cache) : NULL;
ASSERT_EQ(cbm_setenv("HOME", "/tmp/cbm-platform-thread-home", 1), 0);
ASSERT_EQ(cbm_setenv("CBM_CACHE_DIR", "/tmp/cbm-platform-thread-cache", 1), 0);
atomic_int ready;
atomic_int release;
atomic_init(&ready, 0);
atomic_init(&release, 0);
platform_path_thread_result_t results[2] = {
{.ready = &ready, .release = &release},
{.ready = &ready, .release = &release},
};
cbm_thread_t threads[2];
bool started0 =
cbm_thread_create(&threads[0], 0, platform_capture_path_buffers, &results[0]) == 0;
bool started1 =
cbm_thread_create(&threads[1], 0, platform_capture_path_buffers, &results[1]) == 0;
for (int spins = 0; started0 && started1 && spins < 5000 &&
atomic_load_explicit(&ready, memory_order_acquire) < 2;
spins++) {
cbm_usleep(1000);
}
bool both_ready = atomic_load_explicit(&ready, memory_order_acquire) == 2;
bool separate_home =
both_ready && results[0].home && results[1].home && results[0].home != results[1].home;
bool separate_cache =
both_ready && results[0].cache && results[1].cache && results[0].cache != results[1].cache;
atomic_store_explicit(&release, 1, memory_order_release);
if (started0) {
(void)cbm_thread_join(&threads[0]);
}
if (started1) {
(void)cbm_thread_join(&threads[1]);
}
if (saved_home_copy) {
(void)cbm_setenv("HOME", saved_home_copy, 1);
} else {
(void)cbm_unsetenv("HOME");
}
if (saved_cache_copy) {
(void)cbm_setenv("CBM_CACHE_DIR", saved_cache_copy, 1);
} else {
(void)cbm_unsetenv("CBM_CACHE_DIR");
}
free(saved_home_copy);
free(saved_cache_copy);
ASSERT_TRUE(started0);
ASSERT_TRUE(started1);
ASSERT_TRUE(both_ready);
ASSERT_TRUE(separate_home);
ASSERT_TRUE(separate_cache);
PASS();
}
/* A configured cache root is an identity boundary. Truncating an oversized
* value and silently using the prefix would admit/log one root while placing
* data somewhere the user never selected. Reject it instead of falling back. */
TEST(platform_cache_dir_rejects_truncated_override) {
const char *saved = getenv("CBM_CACHE_DIR");
char *saved_copy = saved ? strdup(saved) : NULL;
size_t length = 5000U;
char *value = malloc(length + 1U);
ASSERT_NOT_NULL(value);
memcpy(value, "/tmp/", 5U);
memset(value + 5U, 'a', length - 5U);
value[length] = '\0';
ASSERT_EQ(cbm_setenv("CBM_CACHE_DIR", value, 1), 0);
const char *resolved = cbm_resolve_cache_dir();
if (saved_copy) {
(void)cbm_setenv("CBM_CACHE_DIR", saved_copy, 1);
} else {
(void)cbm_unsetenv("CBM_CACHE_DIR");
}
free(saved_copy);
free(value);
ASSERT_NULL(resolved);
PASS();
}
#ifdef _WIN32
/* cbm_safe_getenv reads Windows' wide environment as UTF-8. Its matching
* setter must update that same wide environment; _putenv_s alone interprets
* UTF-8 path bytes through the active ANSI code page. */
TEST(platform_setenv_preserves_utf8_in_wide_environment) {
static const char utf8[] = "C:/cbm-cache-\xce\x94-\xe6\x97\xa5\xe6\x9c\xac";
static const wchar_t wide[] = L"C:/cbm-cache-\u0394-\u65e5\u672c";
ASSERT_EQ(cbm_setenv("CBM_CACHE_DIR", utf8, 1), 0);
wchar_t observed_wide[128];
ASSERT_EQ(GetEnvironmentVariableW(L"CBM_CACHE_DIR", observed_wide, 128), (DWORD)(wcslen(wide)));
ASSERT_EQ(wcscmp(observed_wide, wide), 0);
char observed_utf8[128];
ASSERT_NOT_NULL(cbm_safe_getenv("CBM_CACHE_DIR", observed_utf8, sizeof(observed_utf8), NULL));
ASSERT_STR_EQ(observed_utf8, utf8);
(void)cbm_unsetenv("CBM_CACHE_DIR");
PASS();
}
/* Empty and absent variables have different fallback semantics. In
* particular, an explicitly empty CBM_CACHE_DIR means "use the default"; it
* must not be misreported as a failed wide-environment read. Unset is also
* required to stay idempotent because test and process cleanup call it after
* partially initialized paths. */
TEST(platform_windows_empty_environment_is_read_and_unset_idempotently) {
ASSERT_EQ(cbm_setenv("CBM_CACHE_DIR", "", 1), 0);
char observed[9] = "sentinel";
ASSERT_NOT_NULL(cbm_safe_getenv("CBM_CACHE_DIR", observed, sizeof(observed), "fallback"));
ASSERT_STR_EQ(observed, "");
ASSERT_EQ(cbm_unsetenv("CBM_CACHE_DIR"), 0);
ASSERT_EQ(cbm_unsetenv("CBM_CACHE_DIR"), 0);
PASS();
}
#endif
/*
* CBM_WORKERS env override for cbm_default_worker_count.
*
* Containers running cbm on a host with more CPUs than the cgroup's
* effective quota currently see ~host_cpu workers spawned because
* sysconf(_SC_NPROCESSORS_ONLN) is not cgroup-aware (see GitHub
* issue for the cgroup-detection ask). CBM_WORKERS is the smaller,
* explicit-override path that ships independently.
*/
TEST(platform_default_workers_env_override) {
cbm_setenv("CBM_WORKERS", "4", 1);
int n = cbm_default_worker_count(true);
ASSERT_EQ(n, 4);
/* initial=false should also honor the explicit override. */
int m = cbm_default_worker_count(false);
ASSERT_EQ(m, 4);
cbm_unsetenv("CBM_WORKERS");
PASS();
}
TEST(platform_default_workers_env_invalid) {
/* Out-of-range values (< 1 or > 256) and non-numeric strings
* fall back to the sysconf-derived default. */
int baseline = cbm_default_worker_count(true);
ASSERT_GT(baseline, 0);
cbm_setenv("CBM_WORKERS", "0", 1);
ASSERT_EQ(cbm_default_worker_count(true), baseline);
cbm_setenv("CBM_WORKERS", "-1", 1);
ASSERT_EQ(cbm_default_worker_count(true), baseline);
cbm_setenv("CBM_WORKERS", "9999", 1);
ASSERT_EQ(cbm_default_worker_count(true), baseline);
cbm_setenv("CBM_WORKERS", "not-a-number", 1);
ASSERT_EQ(cbm_default_worker_count(true), baseline);
cbm_unsetenv("CBM_WORKERS");
PASS();
}
TEST(platform_default_workers_env_unset) {
/* When CBM_WORKERS is unset the result matches today's behaviour
* (info.total_cores for initial=true, perf_cores-1 for false). */
cbm_unsetenv("CBM_WORKERS");
cbm_system_info_t info = cbm_system_info();
ASSERT_EQ(cbm_default_worker_count(true), info.total_cores);
PASS();
}
TEST(platform_system_info) {
cbm_system_info_t info = cbm_system_info();
ASSERT_GT(info.total_cores, 0);
ASSERT_GT(info.total_ram, 0);
PASS();
}
/* ── cgroup-aware detection (Linux only) ─────────────────────────── */
#ifdef __linux__
/* Create a unique tmp directory the caller will own; returns 0 on success. */
static int cgroup_test_setup(char *root, size_t root_sz) {
strncpy(root, "/tmp/cbm_cgroup_test_XXXXXX", root_sz);
return mkdtemp(root) != NULL ? 0 : -1;
}
/* Write `content` to "<root>/<relpath>". Creates parent subdir if needed.
* Returns 0 on success, -1 on any failure. */
static int cgroup_test_write(const char *root, const char *relpath, const char *content) {
char path[1024];
const char *slash = strchr(relpath, '/');
if (slash != NULL) {
char subdir[1024];
size_t n = (size_t)(slash - relpath);
if (n >= sizeof(subdir)) {
return -1;
}
memcpy(subdir, relpath, n);
subdir[n] = '\0';
snprintf(path, sizeof(path), "%s/%s", root, subdir);
(void)mkdir(path, S_IRWXU);
}
snprintf(path, sizeof(path), "%s/%s", root, relpath);
FILE *fp = fopen(path, "we");
if (fp == NULL) {
return -1;
}
size_t n = strlen(content);
int rc = (fwrite(content, 1, n, fp) == n) ? 0 : -1;
fclose(fp);
return rc;
}
/* Recursively remove a tmp dir created by cgroup_test_setup. Best-effort.
* Uses opendir/unlink/rmdir rather than system("rm -rf ...") to avoid
* spawning a shell from the test binary. */
static void cgroup_test_teardown(const char *root) {
DIR *d = opendir(root);
if (d != NULL) {
struct dirent *ent;
while ((ent = readdir(d)) != NULL) {
if (strcmp(ent->d_name, ".") == 0 || strcmp(ent->d_name, "..") == 0) {
continue;
}
char child[1024];
snprintf(child, sizeof(child), "%s/%s", root, ent->d_name);
struct stat st;
if (stat(child, &st) == 0 && S_ISDIR(st.st_mode)) {
cgroup_test_teardown(child); /* recurse into subdir */
} else {
(void)unlink(child);
}
}
closedir(d);
}
(void)rmdir(root);
}
TEST(cgroup_v2_cpu_quota) {
char root[64];
ASSERT_EQ(cgroup_test_setup(root, sizeof(root)), 0);
/* 200ms quota in a 100ms period → 2 effective CPUs. */
ASSERT_EQ(cgroup_test_write(root, "cpu.max", "200000 100000\n"), 0);
ASSERT_EQ(cbm_detect_cgroup_cpus(root), 2);
cgroup_test_teardown(root);
PASS();
}
TEST(cgroup_v2_cpu_quota_rounds_up) {
char root[64];
ASSERT_EQ(cgroup_test_setup(root, sizeof(root)), 0);
/* 150ms quota / 100ms period = 1.5 → ceil = 2. */
ASSERT_EQ(cgroup_test_write(root, "cpu.max", "150000 100000\n"), 0);
ASSERT_EQ(cbm_detect_cgroup_cpus(root), 2);
cgroup_test_teardown(root);
PASS();
}
TEST(cgroup_v2_cpu_unlimited) {
char root[64];
ASSERT_EQ(cgroup_test_setup(root, sizeof(root)), 0);
ASSERT_EQ(cgroup_test_write(root, "cpu.max", "max 100000\n"), 0);
ASSERT_EQ(cbm_detect_cgroup_cpus(root), -1);
cgroup_test_teardown(root);
PASS();
}
TEST(cgroup_v1_cpu_quota) {
char root[64];
ASSERT_EQ(cgroup_test_setup(root, sizeof(root)), 0);
ASSERT_EQ(cgroup_test_write(root, "cpu/cpu.cfs_quota_us", "200000"), 0);
ASSERT_EQ(cgroup_test_write(root, "cpu/cpu.cfs_period_us", "100000"), 0);
ASSERT_EQ(cbm_detect_cgroup_cpus(root), 2);
cgroup_test_teardown(root);
PASS();
}
TEST(cgroup_v1_cpu_unlimited) {
char root[64];
ASSERT_EQ(cgroup_test_setup(root, sizeof(root)), 0);
/* quota=-1 is the cgroup-v1 sentinel for "no quota". */
ASSERT_EQ(cgroup_test_write(root, "cpu/cpu.cfs_quota_us", "-1"), 0);
ASSERT_EQ(cgroup_test_write(root, "cpu/cpu.cfs_period_us", "100000"), 0);
ASSERT_EQ(cbm_detect_cgroup_cpus(root), -1);
cgroup_test_teardown(root);
PASS();
}
TEST(cgroup_no_cpu_files) {
char root[64];
ASSERT_EQ(cgroup_test_setup(root, sizeof(root)), 0);
/* Empty tmp dir: no v2 file, no v1 file → fall through to sysconf. */
ASSERT_EQ(cbm_detect_cgroup_cpus(root), -1);
cgroup_test_teardown(root);
PASS();
}
TEST(cgroup_v2_mem) {
char root[64];
ASSERT_EQ(cgroup_test_setup(root, sizeof(root)), 0);
/* 2 GiB. */
ASSERT_EQ(cgroup_test_write(root, "memory.max", "2147483648\n"), 0);
ASSERT_EQ(cbm_detect_cgroup_mem(root), (size_t)2147483648UL);
cgroup_test_teardown(root);
PASS();
}
TEST(cgroup_v2_mem_unlimited) {
char root[64];
ASSERT_EQ(cgroup_test_setup(root, sizeof(root)), 0);
ASSERT_EQ(cgroup_test_write(root, "memory.max", "max\n"), 0);
ASSERT_EQ(cbm_detect_cgroup_mem(root), (size_t)0);
cgroup_test_teardown(root);
PASS();
}
TEST(cgroup_v1_mem) {
char root[64];
ASSERT_EQ(cgroup_test_setup(root, sizeof(root)), 0);
/* 1 GiB. */
ASSERT_EQ(cgroup_test_write(root, "memory/memory.limit_in_bytes", "1073741824"), 0);
ASSERT_EQ(cbm_detect_cgroup_mem(root), (size_t)1073741824UL);
cgroup_test_teardown(root);
PASS();
}
TEST(cgroup_v1_mem_unlimited_sentinel) {
char root[64];
ASSERT_EQ(cgroup_test_setup(root, sizeof(root)), 0);
/* cgroup v1 reports a huge near-ULLONG_MAX value when unlimited
* (PAGE_COUNTER_MAX). Our parser treats anything >= ULLONG_MAX/2
* as effectively unlimited. */
ASSERT_EQ(cgroup_test_write(root, "memory/memory.limit_in_bytes", "9223372036854775807"), 0);
ASSERT_EQ(cbm_detect_cgroup_mem(root), (size_t)0);
cgroup_test_teardown(root);
PASS();
}
TEST(cgroup_no_mem_files) {
char root[64];
ASSERT_EQ(cgroup_test_setup(root, sizeof(root)), 0);
ASSERT_EQ(cbm_detect_cgroup_mem(root), (size_t)0);
cgroup_test_teardown(root);
PASS();
}
#endif /* __linux__ */
SUITE(platform) {
RUN_TEST(platform_file_apis_survive_max_path_overflow);
RUN_TEST(platform_mkstemp_and_mkdtemp_survive_non_ascii_directory);
RUN_TEST(platform_counter_scaling_avoids_intermediate_overflow);
RUN_TEST(platform_counter_scaling_preserves_monotonic_deadlines);
RUN_TEST(platform_now_ns_concurrent_first_call);
RUN_TEST(platform_now_ns);
RUN_TEST(platform_now_ms);
RUN_TEST(platform_nprocs);
RUN_TEST(platform_file_exists);
RUN_TEST(platform_is_dir);
RUN_TEST(platform_file_size);
RUN_TEST(platform_mmap);
RUN_TEST(platform_mmap_nonexistent);
RUN_TEST(platform_path_helpers_use_per_thread_storage);
RUN_TEST(platform_cache_dir_rejects_truncated_override);
#ifdef _WIN32
RUN_TEST(platform_setenv_preserves_utf8_in_wide_environment);
RUN_TEST(platform_windows_empty_environment_is_read_and_unset_idempotently);
#endif
RUN_TEST(platform_default_workers_env_override);
RUN_TEST(platform_default_workers_env_invalid);
RUN_TEST(platform_default_workers_env_unset);
RUN_TEST(platform_system_info);
#ifdef __linux__
RUN_TEST(cgroup_v2_cpu_quota);
RUN_TEST(cgroup_v2_cpu_quota_rounds_up);
RUN_TEST(cgroup_v2_cpu_unlimited);
RUN_TEST(cgroup_v1_cpu_quota);
RUN_TEST(cgroup_v1_cpu_unlimited);
RUN_TEST(cgroup_no_cpu_files);
RUN_TEST(cgroup_v2_mem);
RUN_TEST(cgroup_v2_mem_unlimited);
RUN_TEST(cgroup_v1_mem);
RUN_TEST(cgroup_v1_mem_unlimited_sentinel);
RUN_TEST(cgroup_no_mem_files);
#endif
}