myxce/src/scheduler/testsuite/test_core_isolation.c
Rongsong Shen a13068dca6
fix(scheduler): extend core-isolation nm-audit to core_events.c/core_session.c/core_clock.c (Req 5 AC 1/2)
test_core_isolation.c's build-time proof (nm(1) undefined-symbol scan of
sched_core_isolation_lib) previously covered only core.c -- core.c's own
800-line-cap siblings, holding scheduling logic just as boundary-sensitive
as core.c itself, were "isolation-clean by review" only, with nothing
mechanical catching a regression. Extends sched_core_isolation_lib to
compile all four files together, and the source-text scans (forbidden
headers/symbols, wall-clock calls) to also cover core_events.c/
core_session.c. core_clock.c joins only the nm(1) layer, not the
wall-clock text scan -- it's the one sanctioned exception carrying the
core family's single legitimate clock_gettime() call.
2026-08-19 11:00:41 +00:00

398 lines
15 KiB
C

/*
* test_core_isolation.c - regression guard for task 4.1's Observable
* (scheduler-core spec, Req 5.1/5.2): "the core structure compiles and
* links with zero symbols from the networking/API libraries; a build-time
* or link-time check confirms this" -- and the accompanying constraint
* "no wall-clock time call appears anywhere in the extracted core's
* source."
*
* Design review follow-up (2026-08-19, scheduler-core retrospective
* review): originally this guard mechanically covered ONLY core.c/core.h,
* leaving core_events.c/core_session.c/core_clock.c -- core.c's own
* 800-line-cap siblings, holding scheduling logic just as boundary-
* sensitive as core.c itself -- "isolation-clean by review" only (manual,
* not mechanical; scheduler/CLAUDE.md's own Architecture section used to
* say so explicitly). A future change to either sibling could silently
* reintroduce a networking/API dependency with nothing to catch it. Both
* layers below now cover core_events.c/core_session.c too. core_clock.c
* joins ONLY the nm(1) layer (2), not the wall-clock text scan: it is the
* one FILE-level sanctioned exception carrying the core family's single
* legitimate clock_gettime() call (Req 5 AC 3, core_clock.c's own header
* comment) -- scanning it for wall-clock text would trivially and
* incorrectly fail. It has no legitimate reason to reference any
* networking/API symbol, so it fully participates in layer 2.
*
* Two layers, same genre as test_picker_core_loop_boundary.c /
* test_picker_pendreason_factory_audit.c (source-text scans of the real,
* current on-disk files, not a frozen snapshot):
*
* 1. Source-text scan of core.h/core.c/core_events.c/core_session.c: no
* #include of client.h, netio.h, xceapi.h, delay.h, idgen.h, or
* xce/cluster.h (the "Allowed Dependencies" list in design.md names
* exactly these as forbidden); no call-site substring matching a
* symbol exported only by those libraries; no wall-clock call
* (time()/clock_gettime()/gettimeofday()) anywhere in core.c,
* core_events.c, or core_session.c (core_clock.c excepted -- see
* above).
*
* 2. The literal build-time/link-time proof the Observable asks for,
* not achievable by a text scan alone: core.c/core_events.c/
* core_session.c/core_clock.c are compiled together in isolation
* into one static library (meson.build's sched_core_isolation_lib,
* this test's XCE_CORE_LIB_PATH), and this test inspects that
* archive's UNDEFINED SYMBOL TABLE via nm(1) -- proving what this
* code's *compiled output actually calls*, independent of what any
* header happens to transitively declare (xce/types.h, used by
* literally every file in this directory, itself includes netio.h
* for unrelated reasons -- so a "no forbidden #include" check alone
* would either false-positive on that pre-existing, harmless
* transitive include, or have to special-case it; the nm check
* sidesteps that entirely by checking *usage*, not *visibility*).
*
* Why nm on a static_library() rather than the full xce_scheduler
* executable: the executable links the whole monolithic libxce_dep, so
* every symbol from netio/xceapi/client/delay/idgen would appear in ITS
* undefined-symbol table regardless of whether this code itself ever
* calls one -- that would prove nothing. Compiling this file set alone
* into its own archive (sched_core_isolation_lib, meson.build) and
* inspecting THAT archive's undefined symbols isolates exactly what these
* translation units reference.
*/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unity/unity.h>
#ifndef XCE_CORE_H_PATH
#error \
"XCE_CORE_H_PATH must be defined by meson.build to core.h's absolute path"
#endif
#ifndef XCE_CORE_C_PATH
#error \
"XCE_CORE_C_PATH must be defined by meson.build to core.c's absolute path"
#endif
#ifndef XCE_CORE_EVENTS_C_PATH
#error \
"XCE_CORE_EVENTS_C_PATH must be defined by meson.build to core_events.c's absolute path"
#endif
#ifndef XCE_CORE_SESSION_C_PATH
#error \
"XCE_CORE_SESSION_C_PATH must be defined by meson.build to core_session.c's absolute path"
#endif
#ifndef XCE_CORE_LIB_PATH
#error \
"XCE_CORE_LIB_PATH must be defined by meson.build to the absolute path of core.c's isolated static_library() archive"
#endif
/* Forbidden headers (design.md "Allowed Dependencies", Req 5 AC 3): the
* core must never directly include any of these. */
static const char* const FORBIDDEN_INCLUDES[] = {
"#include \"client.h\"", "#include \"netio.h\"",
"#include \"xceapi.h\"", "#include \"delay.h\"",
"#include \"idgen.h\"", "#include \"xce/cluster.h\"",
"#include <client.h>", "#include <netio.h>",
"#include <xceapi.h>", "#include <delay.h>",
"#include <idgen.h>", "#include <xce/cluster.h>",
};
#define NUM_FORBIDDEN_INCLUDES \
(sizeof(FORBIDDEN_INCLUDES) / sizeof(FORBIDDEN_INCLUDES[0]))
/* Forbidden call-site symbols: exported only by netio.c/xceapi.c/client.c/
* delay.c/idgen.c (verified against those headers' own extern lists) --
* none of these are legitimately callable from a core with zero controller
* I/O, network, or timer dependency. */
static const char* const FORBIDDEN_SYMBOLS[] = {
"xcecli_",
"endpoint_",
"iomgr_",
"idgen_",
"delaymgr_",
"delay_task",
"xce_connect_cluster",
"xce_reconnect_",
"xce_parse_peers",
"xce_cluster_config_validate",
"socket_connect",
"create_ssl_context",
"initialize_openssl",
"cleanup_openssl",
"io_fsm_start",
};
#define NUM_FORBIDDEN_SYMBOLS \
(sizeof(FORBIDDEN_SYMBOLS) / sizeof(FORBIDDEN_SYMBOLS[0]))
/* Wall-clock calls (Req 5 AC "no wall-clock time call appears anywhere in
* the extracted core's source"). */
static const char* const FORBIDDEN_WALLCLOCK[] = {
"time(",
"clock_gettime(",
"gettimeofday(",
};
#define NUM_FORBIDDEN_WALLCLOCK \
(sizeof(FORBIDDEN_WALLCLOCK) / sizeof(FORBIDDEN_WALLCLOCK[0]))
static char* read_whole_file(const char* path)
{
FILE* f = fopen(path, "rb");
if (f == NULL) {
return NULL;
}
if (fseek(f, 0, SEEK_END) != 0) {
fclose(f);
return NULL;
}
long len = ftell(f);
if (len < 0) {
fclose(f);
return NULL;
}
rewind(f);
char* buf = (char*)malloc((size_t)len + 1);
if (buf == NULL) {
fclose(f);
return NULL;
}
size_t nread = fread(buf, 1, (size_t)len, f);
fclose(f);
buf[nread] = '\0';
return buf;
}
void setUp(void) { return; }
void tearDown(void) { return; }
/* Sanity check: both source paths baked in by meson.build must actually be
* readable -- without this, a stale path would make every scan below
* silently pass on zero bytes of text (the exact false-confidence failure
* mode this whole test exists to avoid). */
static void test_core_source_paths_are_readable(void)
{
char* h = read_whole_file(XCE_CORE_H_PATH);
char* c = read_whole_file(XCE_CORE_C_PATH);
char* ce = read_whole_file(XCE_CORE_EVENTS_C_PATH);
char* cs = read_whole_file(XCE_CORE_SESSION_C_PATH);
TEST_ASSERT_NOT_NULL_MESSAGE(h, XCE_CORE_H_PATH);
TEST_ASSERT_NOT_NULL_MESSAGE(c, XCE_CORE_C_PATH);
TEST_ASSERT_NOT_NULL_MESSAGE(ce, XCE_CORE_EVENTS_C_PATH);
TEST_ASSERT_NOT_NULL_MESSAGE(cs, XCE_CORE_SESSION_C_PATH);
free(h);
free(c);
free(ce);
free(cs);
}
/* Scans one file's text for every FORBIDDEN_INCLUDES entry, tagging each
* failure message with `label` so a failure names the actual offending
* file -- shared by test_core_never_includes_forbidden_headers() across
* core.h/core.c/core_events.c/core_session.c. */
static void assert_no_forbidden_includes(const char* text, const char* label)
{
size_t i;
for (i = 0; i < NUM_FORBIDDEN_INCLUDES; i++) {
char msg[256];
snprintf(msg, sizeof(msg), "%s directly includes forbidden header '%s'",
label, FORBIDDEN_INCLUDES[i]);
TEST_ASSERT_NULL_MESSAGE(strstr(text, FORBIDDEN_INCLUDES[i]), msg);
}
}
static void test_core_never_includes_forbidden_headers(void)
{
char* h = read_whole_file(XCE_CORE_H_PATH);
char* c = read_whole_file(XCE_CORE_C_PATH);
char* ce = read_whole_file(XCE_CORE_EVENTS_C_PATH);
char* cs = read_whole_file(XCE_CORE_SESSION_C_PATH);
TEST_ASSERT_NOT_NULL(h);
TEST_ASSERT_NOT_NULL(c);
TEST_ASSERT_NOT_NULL(ce);
TEST_ASSERT_NOT_NULL(cs);
assert_no_forbidden_includes(h, "core.h");
assert_no_forbidden_includes(c, "core.c");
assert_no_forbidden_includes(ce, "core_events.c");
assert_no_forbidden_includes(cs, "core_session.c");
free(h);
free(c);
free(ce);
free(cs);
}
/* Shared by test_core_never_calls_forbidden_symbols() -- see
* assert_no_forbidden_includes() above for the same per-file-label
* rationale. */
static void assert_no_forbidden_symbols(const char* text, const char* label)
{
size_t i;
for (i = 0; i < NUM_FORBIDDEN_SYMBOLS; i++) {
char msg[256];
snprintf(msg, sizeof(msg),
"%s references forbidden networking/API symbol '%s'", label,
FORBIDDEN_SYMBOLS[i]);
TEST_ASSERT_NULL_MESSAGE(strstr(text, FORBIDDEN_SYMBOLS[i]), msg);
}
}
static void test_core_never_calls_forbidden_symbols(void)
{
char* h = read_whole_file(XCE_CORE_H_PATH);
char* c = read_whole_file(XCE_CORE_C_PATH);
char* ce = read_whole_file(XCE_CORE_EVENTS_C_PATH);
char* cs = read_whole_file(XCE_CORE_SESSION_C_PATH);
TEST_ASSERT_NOT_NULL(h);
TEST_ASSERT_NOT_NULL(c);
TEST_ASSERT_NOT_NULL(ce);
TEST_ASSERT_NOT_NULL(cs);
assert_no_forbidden_symbols(h, "core.h");
assert_no_forbidden_symbols(c, "core.c");
assert_no_forbidden_symbols(ce, "core_events.c");
assert_no_forbidden_symbols(cs, "core_session.c");
free(h);
free(c);
free(ce);
free(cs);
}
/* core_clock.c is deliberately NOT scanned here -- it is the one
* sanctioned exception carrying the core family's single legitimate
* clock_gettime() call (Req 5 AC 3); see this file's own header comment. */
static void test_core_never_calls_wallclock_functions(void)
{
char* c = read_whole_file(XCE_CORE_C_PATH);
char* ce = read_whole_file(XCE_CORE_EVENTS_C_PATH);
char* cs = read_whole_file(XCE_CORE_SESSION_C_PATH);
size_t i;
TEST_ASSERT_NOT_NULL(c);
TEST_ASSERT_NOT_NULL(ce);
TEST_ASSERT_NOT_NULL(cs);
for (i = 0; i < NUM_FORBIDDEN_WALLCLOCK; i++) {
char msg[256];
snprintf(msg, sizeof(msg),
"core.c calls wall-clock function '%s' -- the core must "
"access time exclusively through xce_core_now() "
"(scheduler-core spec, task 4.5)",
FORBIDDEN_WALLCLOCK[i]);
TEST_ASSERT_NULL_MESSAGE(strstr(c, FORBIDDEN_WALLCLOCK[i]), msg);
snprintf(msg, sizeof(msg),
"core_events.c calls wall-clock function '%s' -- the core "
"must access time exclusively through xce_core_now() "
"(scheduler-core spec, task 4.5)",
FORBIDDEN_WALLCLOCK[i]);
TEST_ASSERT_NULL_MESSAGE(strstr(ce, FORBIDDEN_WALLCLOCK[i]), msg);
snprintf(msg, sizeof(msg),
"core_session.c calls wall-clock function '%s' -- the core "
"must access time exclusively through xce_core_now() "
"(scheduler-core spec, task 4.5)",
FORBIDDEN_WALLCLOCK[i]);
TEST_ASSERT_NULL_MESSAGE(strstr(cs, FORBIDDEN_WALLCLOCK[i]), msg);
}
free(c);
free(ce);
free(cs);
}
/*
* test_core_object_has_no_forbidden_undefined_symbols - the literal
* build-time/link-time proof: core.c/core_events.c/core_session.c/
* core_clock.c, compiled together in isolation into
* sched_core_isolation_lib (meson.build), must not reference (as an
* undefined external symbol in the compiled archive) anything matching the
* forbidden prefixes above. This is independent of what any header
* transitively declares -- it reflects only what this file set's compiled
* code actually calls.
*/
static void test_core_object_has_no_forbidden_undefined_symbols(void)
{
char cmd[4096];
char line[1024];
FILE* p;
int any_undefined_seen = 0;
snprintf(cmd, sizeof(cmd), "nm --format=posix -u \"%s\" 2>&1",
XCE_CORE_LIB_PATH);
p = popen(cmd, "r");
TEST_ASSERT_NOT_NULL_MESSAGE(p, "failed to popen() nm(1)");
while (fgets(line, sizeof(line), p) != NULL) {
char symbol[512];
char type;
size_t i;
/* posix format for an undefined symbol: "<name> U" (no
* value/size columns). Skip anything that doesn't parse as
* exactly that -- e.g. a blank line or an archive member header
* some nm versions print. */
if (sscanf(line, "%511s %c", symbol, &type) != 2) {
continue;
}
if (type != 'U') {
continue;
}
any_undefined_seen = 1;
for (i = 0; i < NUM_FORBIDDEN_SYMBOLS; i++) {
if (strncmp(symbol, FORBIDDEN_SYMBOLS[i],
strlen(FORBIDDEN_SYMBOLS[i]))
== 0) {
char msg[768];
snprintf(msg, sizeof(msg),
"the core isolation archive (core.c/core_events.c/"
"core_session.c/core_clock.c) has an undefined "
"reference to '%s', matching forbidden "
"networking/API prefix '%s' -- the core must never "
"link networking/API library symbols (Req 5.1/"
"5.2)",
symbol, FORBIDDEN_SYMBOLS[i]);
pclose(p);
TEST_FAIL_MESSAGE(msg);
}
}
}
pclose(p);
/* If nm produced no undefined symbols at all, either the archive path
* is wrong or nm itself failed silently -- either way this test would
* be vacuously passing without having checked anything real. */
TEST_ASSERT_TRUE_MESSAGE(
any_undefined_seen,
"nm reported zero undefined symbols for XCE_CORE_LIB_PATH -- "
"the archive path is likely wrong (this test would otherwise pass "
"vacuously)");
}
int main(void)
{
UNITY_BEGIN();
RUN_TEST(test_core_source_paths_are_readable);
RUN_TEST(test_core_never_includes_forbidden_headers);
RUN_TEST(test_core_never_calls_forbidden_symbols);
RUN_TEST(test_core_never_calls_wallclock_functions);
RUN_TEST(test_core_object_has_no_forbidden_undefined_symbols);
return UNITY_END();
}