gh-105481: generate op IDs from bytecode.c instead of hard coding them in opcode.py (#107971)

This commit is contained in:
Irit Katriel 2023-08-16 23:25:18 +01:00 committed by GitHub
parent e88eb3775e
commit 665a4391e1
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
21 changed files with 1593 additions and 1521 deletions

View file

@ -27,27 +27,6 @@ opcode_h_footer = """
#endif /* !Py_OPCODE_H */
"""
opcode_ids_h_header = f"""
// Auto-generated by {SCRIPT_NAME} from {PYTHON_OPCODE}
#ifndef Py_OPCODE_IDS_H
#define Py_OPCODE_IDS_H
#ifdef __cplusplus
extern "C" {{
#endif
/* Instruction opcodes for compiled code */
""".lstrip()
opcode_ids_h_footer = """
#ifdef __cplusplus
}
#endif
#endif /* !Py_OPCODE_IDS_H */
"""
internal_header = f"""
// Auto-generated by {SCRIPT_NAME} from {PYTHON_OPCODE}
@ -83,52 +62,10 @@ def get_python_module_dict(filename):
return mod
def main(opcode_py,
_opcode_metadata_py='Lib/_opcode_metadata.py',
opcode_ids_h='Include/opcode_ids.h',
opcode_h='Include/opcode.h',
opcode_targets_h='Python/opcode_targets.h',
internal_opcode_h='Include/internal/pycore_opcode.h'):
_opcode_metadata = get_python_module_dict(_opcode_metadata_py)
opcode = get_python_module_dict(opcode_py)
opmap = opcode['opmap']
opname = opcode['opname']
MIN_INSTRUMENTED_OPCODE = opcode["MIN_INSTRUMENTED_OPCODE"]
NUM_OPCODES = len(opname)
used = [ False ] * len(opname)
next_op = 1
for name, op in opmap.items():
used[op] = True
specialized_opmap = {}
opname_including_specialized = opname.copy()
for name in _opcode_metadata['_specialized_instructions']:
while used[next_op]:
next_op += 1
specialized_opmap[name] = next_op
opname_including_specialized[next_op] = name
used[next_op] = True
with open(opcode_ids_h, 'w') as fobj:
fobj.write(opcode_ids_h_header)
for name in opname:
if name in opmap:
op = opmap[name]
if op == MIN_INSTRUMENTED_OPCODE:
fobj.write(DEFINE.format("MIN_INSTRUMENTED_OPCODE", MIN_INSTRUMENTED_OPCODE))
fobj.write(DEFINE.format(name, op))
for name, op in specialized_opmap.items():
fobj.write(DEFINE.format(name, op))
fobj.write(opcode_ids_h_footer)
with open(opcode_h, 'w') as fobj:
fobj.write(opcode_h_header)
@ -143,7 +80,6 @@ def main(opcode_py,
iobj.write(internal_header)
iobj.write("\nextern const uint8_t _PyOpcode_Caches[256];\n")
iobj.write("\nextern const uint8_t _PyOpcode_Deopt[256];\n")
iobj.write("\n#ifdef NEED_OPCODE_TABLES\n")
iobj.write("\nconst uint8_t _PyOpcode_Caches[256] = {\n")
@ -151,52 +87,12 @@ def main(opcode_py,
iobj.write(f" [{name}] = {entries},\n")
iobj.write("};\n")
deoptcodes = {}
for basic, op in opmap.items():
if op < 256:
deoptcodes[basic] = basic
for basic, family in _opcode_metadata["_specializations"].items():
for specialized in family:
deoptcodes[specialized] = basic
iobj.write("\nconst uint8_t _PyOpcode_Deopt[256] = {\n")
for opt, deopt in sorted(deoptcodes.items()):
iobj.write(f" [{opt}] = {deopt},\n")
iobj.write("};\n")
iobj.write("#endif // NEED_OPCODE_TABLES\n")
iobj.write("\n")
iobj.write(f"\nextern const char *const _PyOpcode_OpName[{NUM_OPCODES}];\n")
iobj.write("\n#ifdef NEED_OPCODE_TABLES\n")
iobj.write(f"const char *const _PyOpcode_OpName[{NUM_OPCODES}] = {{\n")
for op, name in enumerate(opname_including_specialized):
if name[0] != "<":
op = name
iobj.write(f''' [{op}] = "{name}",\n''')
iobj.write("};\n")
iobj.write("#endif // NEED_OPCODE_TABLES\n")
iobj.write("\n")
iobj.write("#define EXTRA_CASES \\\n")
for i, flag in enumerate(used):
if not flag:
iobj.write(f" case {i}: \\\n")
iobj.write(" ;\n")
iobj.write(internal_footer)
with open(opcode_targets_h, "w") as f:
targets = ["_unknown_opcode"] * 256
for op, name in enumerate(opname_including_specialized):
if op < 256 and not name.startswith("<"):
targets[op] = f"TARGET_{name}"
f.write("static void *opcode_targets[256] = {\n")
f.write(",\n".join([f" &&{s}" for s in targets]))
f.write("\n};\n")
print(f"{opcode_h} regenerated from {opcode_py}")
if __name__ == '__main__':
main(sys.argv[1], sys.argv[2], sys.argv[3], sys.argv[4],
sys.argv[5], sys.argv[6])
main(sys.argv[1], sys.argv[2], sys.argv[3])