TheRock/compiler/amd-llvm/llvm/utils/update_llc_test_checks.py

#!/usr/bin/env python3

"""A test case update script.

This script is a utility to update LLVM 'llc' based test cases with new
FileCheck patterns. It can either update all of the tests in the file or
a single test function.
"""

from __future__ import print_function

from sys import stderr
from traceback import print_exc
import argparse
import os  # Used to advertise this file's name ("autogenerated_note").
import sys

from UpdateTestChecks import common, mir

# llc is the only llc-like in the LLVM tree but downstream forks can add
# additional ones here if they have them.
LLC_LIKE_TOOLS = [
    "llc",
]


def update_test(ti: common.TestInfo):
    triple_in_ir = None
    for l in ti.input_lines:
        m = common.TRIPLE_IR_RE.match(l)
        if m:
            triple_in_ir = m.groups()[0]
            break

    run_list = []
    mir_run_list = []
    for l in ti.run_lines:
        if "|" not in l:
            common.warn("Skipping unparsable RUN line: " + l)
            continue

        commands = [cmd.strip() for cmd in l.split("|")]
        assert len(commands) >= 2
        preprocess_cmd = None
        if len(commands) > 2:
            preprocess_cmd = " | ".join(commands[:-2])
        llc_cmd = commands[-2]
        filecheck_cmd = commands[-1]
        llc_tool = llc_cmd.split(" ")[0]

        triple_in_cmd = None
        m = common.TRIPLE_ARG_RE.search(llc_cmd)
        if m:
            triple_in_cmd = m.groups()[0]

        march_in_cmd = ti.args.default_march
        m = common.MARCH_ARG_RE.search(llc_cmd)
        if m:
            march_in_cmd = m.groups()[0]

        target_list = run_list
        m = common.DEBUG_ONLY_ARG_RE.search(llc_cmd)
        if m and m.groups()[0] == "isel":
            from UpdateTestChecks import isel as output_type
        elif not m and common.STOP_PASS_RE.search(llc_cmd):
            # MIR output mode. If -debug-only is present assume
            # the debug output is the main point of interest.
            target_list = mir_run_list
        else:
            from UpdateTestChecks import asm as output_type

        common.verify_filecheck_prefixes(filecheck_cmd)

        llc_like_tools = LLC_LIKE_TOOLS[:]
        if ti.args.tool:
            llc_like_tools.append(ti.args.tool)
        if llc_tool not in llc_like_tools:
            common.warn("Skipping non-llc RUN line: " + l)
            continue

        if not filecheck_cmd.startswith("FileCheck "):
            common.warn("Skipping non-FileChecked RUN line: " + l)
            continue

        llc_cmd_args = llc_cmd[len(llc_tool) :].strip()
        llc_cmd_args = llc_cmd_args.replace("< %s", "").replace("%s", "").strip()
        if ti.path.endswith(".mir"):
            llc_cmd_args += " -x mir"
        check_prefixes = common.get_check_prefixes(filecheck_cmd)

        # FIXME: We should use multiple check prefixes to common check lines. For
        # now, we just ignore all but the last.
        target_list.append(
            (
                check_prefixes,
                llc_tool,
                llc_cmd_args,
                preprocess_cmd,
                triple_in_cmd,
                march_in_cmd,
            )
        )

    if ti.path.endswith(".mir"):
        check_indent = "  "
    else:
        check_indent = ""

    ginfo = common.make_asm_generalizer(version=1)
    builder = common.FunctionTestBuilder(
        run_list=run_list,
        flags=type(
            "",
            (object,),
            {
                "verbose": ti.args.verbose,
                "filters": ti.args.filters,
                "function_signature": False,
                "check_attributes": False,
                "replace_value_regex": [],
            },
        ),
        scrubber_args=[ti.args],
        path=ti.path,
        ginfo=ginfo,
    )

    # Dictionary to store MIR function bodies separately
    mir_func_dict = {}
    for run_tuple, is_mir in [(run, False) for run in run_list] + [
        (run, True) for run in mir_run_list
    ]:
        (
            prefixes,
            llc_tool,
            llc_args,
            preprocess_cmd,
            triple_in_cmd,
            march_in_cmd,
        ) = run_tuple

        common.debug("Extracted LLC cmd:", llc_tool, llc_args)
        common.debug("Extracted FileCheck prefixes:", str(prefixes))

        raw_tool_output = common.invoke_tool(
            ti.args.llc_binary or llc_tool,
            llc_args,
            ti.path,
            preprocess_cmd,
            verbose=ti.args.verbose,
        )
        triple = triple_in_cmd or triple_in_ir
        if not triple:
            triple = common.get_triple_from_march(march_in_cmd)

        if is_mir:
            # MIR output mode
            common.debug("Detected MIR output mode for prefixes:", str(prefixes))
            for prefix in prefixes:
                if prefix not in mir_func_dict:
                    mir_func_dict[prefix] = {}

            mir.build_function_info_dictionary(
                ti.path,
                raw_tool_output,
                triple,
                prefixes,
                mir_func_dict,
                ti.args.verbose,
            )
        else:
            # ASM output mode
            scrubber, function_re = output_type.get_run_handler(triple)
            if 0 == builder.process_run_line(
                function_re, scrubber, raw_tool_output, prefixes
            ):
                common.warn(
                    "Couldn't match any function. Possibly the wrong target triple has been provided"
                )
            builder.processed_prefixes(prefixes)

    func_dict = builder.finish_and_get_func_dict()

    # Check for conflicts: same prefix used for both ASM and MIR
    conflicting_prefixes = set(func_dict.keys()) & set(mir_func_dict.keys())
    if conflicting_prefixes:
        common.warn(
            "The following prefixes are used for both ASM and MIR output, which will cause FileCheck failures: {}".format(
                ", ".join(sorted(conflicting_prefixes))
            ),
            test_file=ti.path,
        )
        for prefix in conflicting_prefixes:
            mir_func_dict[prefix] = {}
            func_dict[prefix] = {}

    global_vars_seen_dict = {}

    is_in_function = False
    is_in_function_start = False
    func_name = None
    prefix_set = set([prefix for p in run_list for prefix in p[0]])
    prefix_set.update([prefix for p in mir_run_list for prefix in p[0]])
    common.debug("Rewriting FileCheck prefixes:", str(prefix_set))
    output_lines = []

    include_generated_funcs = common.find_arg_in_test(
        ti,
        lambda args: ti.args.include_generated_funcs,
        "--include-generated-funcs",
        True,
    )

    generated_prefixes = []
    if include_generated_funcs:
        # Generate the appropriate checks for each function.  We need to emit
        # these in the order according to the generated output so that CHECK-LABEL
        # works properly.  func_order provides that.

        # We can't predict where various passes might insert functions so we can't
        # be sure the input function order is maintained.  Therefore, first spit
        # out all the source lines.
        common.dump_input_lines(output_lines, ti, prefix_set, ";")

        # Now generate all the checks.
        generated_prefixes = common.add_checks_at_end(
            output_lines,
            run_list,
            builder.func_order(),
            check_indent + ";",
            lambda my_output_lines, prefixes, func: output_type.add_checks(
                my_output_lines,
                check_indent + ";",
                prefixes,
                func_dict,
                func,
                ginfo,
                global_vars_seen_dict,
                is_filtered=builder.is_filtered(),
            ),
        )
    else:
        for input_info in ti.iterlines(output_lines):
            input_line = input_info.line
            args = input_info.args
            if is_in_function_start:
                if input_line == "":
                    continue
                if input_line.lstrip().startswith(";"):
                    m = common.CHECK_RE.match(input_line)
                    if not m or m.group(1) not in prefix_set:
                        output_lines.append(input_line)
                        continue

                # Print out the various check lines here.
                generated_prefixes.extend(
                    output_type.add_checks(
                        output_lines,
                        check_indent + ";",
                        run_list,
                        func_dict,
                        func_name,
                        ginfo,
                        global_vars_seen_dict,
                        is_filtered=builder.is_filtered(),
                    )
                )

                # Also add MIR checks if we have them for this function
                if mir_run_list and func_name:
                    mir.add_mir_checks_for_function(
                        ti.path,
                        output_lines,
                        mir_run_list,
                        mir_func_dict,
                        func_name,
                        single_bb=False,  # Don't skip basic block labels.
                        print_fixed_stack=False,  # Don't print fixed stack (ASM tests don't need it).
                        first_check_is_next=False,  # First check is LABEL, not NEXT.
                        at_the_function_name=False,  # Use "name:" not "@name".
                        check_indent="",  # No indentation for IR files (not MIR files).
                    )

                is_in_function_start = False

            if is_in_function:
                if common.should_add_line_to_output(input_line, prefix_set):
                    # This input line of the function body will go as-is into the output.
                    output_lines.append(input_line)
                else:
                    continue
                if input_line.strip() == "}":
                    is_in_function = False
                continue

            # If it's outside a function, it just gets copied to the output.
            output_lines.append(input_line)

            m = common.IR_FUNCTION_RE.match(input_line)
            if not m:
                continue
            func_name = m.group(1)
            if args.function is not None and func_name != args.function:
                # When filtering on a specific function, skip all others.
                continue
            is_in_function = is_in_function_start = True

    if ti.args.gen_unused_prefix_body:
        output_lines.extend(
            ti.get_checks_for_unused_prefixes(run_list, generated_prefixes)
        )

    common.debug("Writing %d lines to %s..." % (len(output_lines), ti.path))
    with open(ti.path, "wb") as f:
        f.writelines(["{}\n".format(l).encode("utf-8") for l in output_lines])


def main():
    parser = argparse.ArgumentParser(description=__doc__)
    parser.add_argument(
        "--llc-binary",
        default=None,
        help='The "llc" binary to use to generate the test case',
    )
    parser.add_argument("--function", help="The function in the test file to update")
    parser.add_argument(
        "--extra_scrub",
        action="store_true",
        help="Always use additional regex to further reduce diffs between various subtargets",
    )
    parser.add_argument(
        "--x86_scrub_sp",
        action="store_true",
        default=True,
        help="Use regex for x86 sp matching to reduce diffs between various subtargets",
    )
    parser.add_argument("--no_x86_scrub_sp", action="store_false", dest="x86_scrub_sp")
    parser.add_argument(
        "--x86_scrub_rip",
        action="store_true",
        default=False,
        help="Use more regex for x86 rip matching to reduce diffs between various subtargets",
    )
    parser.add_argument(
        "--no_x86_scrub_rip", action="store_false", dest="x86_scrub_rip"
    )
    parser.add_argument(
        "--no_x86_scrub_mem_shuffle",
        action="store_true",
        default=False,
        help="Reduce scrubbing shuffles with memory operands",
    )
    parser.add_argument(
        "--tool",
        default=None,
        help="Treat the given tool name as an llc-like tool for which check lines should be generated",
    )
    parser.add_argument(
        "--default-march",
        default=None,
        help="Set a default -march for when neither triple nor arch are found in a RUN line",
    )
    parser.add_argument("tests", nargs="+")
    initial_args = common.parse_commandline_args(parser)

    script_name = os.path.basename(__file__)

    returncode = 0
    for ti in common.itertests(
        initial_args.tests, parser, script_name="utils/" + script_name
    ):
        try:
            update_test(ti)
        except Exception as e:
            stderr.write(f"Error: Failed to update test {ti.path}\n")
            print_exc()
            returncode = 1
    return returncode


if __name__ == "__main__":
    sys.exit(main())