List kern_ap_stub_ymm as dispatched in check_isa.py, so that a --cpu v2 build passes its instruction check again.

kernel.c's preemption stub has two bodies since 00f06e66: kern_ap_stub_ymm saves YMM0-15 with vmovdqa and is chosen only when ap_init's cpu_keeps_ymm() finds AVX, OSXSAVE and XCR0 bits 1 and 2, and kern_ap_stub_xmm serves every other processor. tools/check_isa.py's DISPATCHED listed only sha256.c's blocks_ni, so it took the YMM body for an undispatched AVX function and stopped every tools/build_game.py --cpu v2 build. The entry names that check. Players' builds (make_game.py) run no such check and were not affected.

Reproduced before the change: kernel.c built with the ucrt64 clang at -march=x86-64-v2 failed tools/check_isa.py with "kern_ap_stub_ymm ...: vmovdqa x32, vzeroupper x1", and passes now. Checks that passed: the new tests/check_isa_test.py (kernel.c and sha256.c built with -march=x86-64-v2 and without -march pass; an AVX function not in DISPATCHED still fails; it fails before the change) and tools/check_content.py --staged.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs/building.md

@@ -108,7 +108,8 @@ of them.
    `runtime/fma.c`, a correctly rounded software fma the FP lowering needs, and runs slower. Inline assembly,
    target attributes and intrinsics are not bound by `-march`, so a v2 build runs `tools/check_isa.py` over the runtime
    objects before it links: any AVX, BMI or other instruction above x86-64-v2 outside a function listed there as
-   dispatched by a runtime CPU check stops the build.
+   dispatched by a runtime CPU check stops the build. `python tests\check_isa_test.py` builds the files with such
+   functions both ways and checks that they pass.
 
    `gen.py` sizes its emit pool from the CPU count and the free memory; `--jobs N` sets it, and so does the environment variable `NX2NAVI_LIFT_JOBS=N`, which the installer's lift step honours too (it has no `--jobs`). When a worker runs out of memory, the pool restarts with half the workers (`emit: out of memory with N workers, retrying with M`), down to one; only an out-of-memory error with one worker fails the step.
 

tests/check_isa_test.py

@@ -0,0 +1,53 @@
+"""tools/check_isa.py on the runtime files that hold instructions above x86-64-v2 behind a runtime CPU check
+(runtime/kernel.c kern_ap_stub_ymm, runtime/sha256.c blocks_ni): built as tools/build_game.py --cpu v2 builds the
+runtime (-march=x86-64-v2) and as players' builds do (no -march), they pass, and a function with such an instruction
+and no entry in DISPATCHED still fails. Needs the MSYS2 UCRT64 clang and an objdump (skipped without them); no game
+files.
+
+Run: python tests/check_isa_test.py
+"""
+import shutil
+import subprocess
+import sys
+import tempfile
+import unittest
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parent.parent
+sys.path.insert(0, str(ROOT / 'tools'))
+
+import check_isa  # noqa: E402
+
+UCRT = Path(r'C:\msys64\ucrt64\bin')
+CC = next((str(c) for c in (UCRT / 'clang.exe', shutil.which('clang')) if c and Path(c).is_file()), None)
+OBJDUMP = check_isa.find_objdump(str(UCRT / 'objdump.exe') if (UCRT / 'objdump.exe').is_file() else None)
+FLAGS = ['-O2', '-g', '-fno-strict-aliasing', '-w']   # tools/build_game.py RT_FLAGS, without its warnings
+
+
+@unittest.skipUnless(CC and OBJDUMP, 'needs clang and objdump')
+class CheckIsaTest(unittest.TestCase):
+    def setUp(self):
+        td = tempfile.TemporaryDirectory()
+        self.addCleanup(td.cleanup)
+        self.tmp = Path(td.name)
+
+    def compile(self, src, march, name=None):
+        out = self.tmp / f'{name or Path(src).stem}{march or ""}.o'
+        subprocess.run([CC, *FLAGS, *([march] if march else []), '-I', str(ROOT / 'runtime'), '-c', str(src),
+                        '-o', str(out)], check=True, capture_output=True, timeout=600)
+        return out
+
+    def test_the_dispatched_functions_pass(self):
+        for march in ('-march=x86-64-v2', None):
+            objs = [self.compile(ROOT / 'runtime' / f, march) for f in ('kernel.c', 'sha256.c')]
+            self.assertEqual(check_isa.scan(objs, OBJDUMP), {}, march)
+
+    def test_an_undispatched_avx_function_fails(self):
+        src = self.tmp / 'wide.c'
+        src.write_text('void wide(void) { __asm__ volatile("vzeroupper"); }\n', encoding='utf-8')
+        hits = check_isa.scan([self.compile(src, '-march=x86-64-v2')], OBJDUMP)
+        self.assertEqual([f for _, f in hits], ['wide'])
+
+
+if __name__ == '__main__':
+    unittest.main(verbosity=2)

tools/check_isa.py

@@ -37,6 +37,7 @@ WIDE = re.compile(r'%[yz]mm\d|%k[0-7]\b')
 # Function -> the runtime check that keeps it off CPUs without the instructions.
 DISPATCHED = {
     'blocks_ni': 'runtime/sha256.c sha256_blocks: __builtin_cpu_supports("sha") && __builtin_cpu_supports("sse4.1")',
+    'kern_ap_stub_ymm': 'runtime/kernel.c ap_init: cpu_keeps_ymm() (CPUID AVX and OSXSAVE, XCR0 bits 1 and 2)',
 }