llvm/utils/collect_and_build_with_pgo.py

   1 #!/usr/bin/env python3
   2 """
   3 This script:
   4 - Builds clang with user-defined flags
   5 - Uses that clang to build an instrumented clang, which can be used to collect
   6   PGO samples
   7 - Builds a user-defined set of sources (default: clang) to act as a
   8   "benchmark" to generate a PGO profile
   9 - Builds clang once more with the PGO profile generated above
  10
  11 This is a total of four clean builds of clang (by default). This may take a
  12 while. :)
  13
  14 This scripts duplicates https://llvm.org/docs/AdvancedBuilds.html#multi-stage-pgo
  15 Eventually, it will be updated to instead call the cmake cache mentioned there.
  16 """
  17
  18 import argparse
  19 import collections
  20 import multiprocessing
  21 import os
  22 import shlex
  23 import shutil
  24 import subprocess
  25 import sys
  26
  27 ### User configuration
  28
  29
  30 # If you want to use a different 'benchmark' than building clang, make this
  31 # function do what you want. out_dir is the build directory for clang, so all
  32 # of the clang binaries will live under "${out_dir}/bin/". Using clang in
  33 # ${out_dir} will magically have the profiles go to the right place.
  34 #
  35 # You may assume that out_dir is a freshly-built directory that you can reach
  36 # in to build more things, if you'd like.
  37 def _run_benchmark(env, out_dir, include_debug_info):
  38     """The 'benchmark' we run to generate profile data."""
  39     target_dir = env.output_subdir("instrumentation_run")
  40
  41     # `check-llvm` and `check-clang` are cheap ways to increase coverage. The
  42     # former lets us touch on the non-x86 backends a bit if configured, and the
  43     # latter gives us more C to chew on (and will send us through diagnostic
  44     # paths a fair amount, though the `if (stuff_is_broken) { diag() ... }`
  45     # branches should still heavily be weighted in the not-taken direction,
  46     # since we built all of LLVM/etc).
  47     _build_things_in(env, out_dir, what=["check-llvm", "check-clang"])
  48
  49     # Building tblgen gets us coverage; don't skip it. (out_dir may also not
  50     # have them anyway, but that's less of an issue)
  51     cmake = _get_cmake_invocation_for_bootstrap_from(env, out_dir, skip_tablegens=False)
  52
  53     if include_debug_info:
  54         cmake.add_flag("CMAKE_BUILD_TYPE", "RelWithDebInfo")
  55
  56     _run_fresh_cmake(env, cmake, target_dir)
  57
  58     # Just build all the things. The more data we have, the better.
  59     _build_things_in(env, target_dir, what=["all"])
  60
  61
  62 ### Script
  63
  64
  65 class CmakeInvocation:
  66     _cflags = ["CMAKE_C_FLAGS", "CMAKE_CXX_FLAGS"]
  67     _ldflags = [
  68         "CMAKE_EXE_LINKER_FLAGS",
  69         "CMAKE_MODULE_LINKER_FLAGS",
  70         "CMAKE_SHARED_LINKER_FLAGS",
  71     ]
  72
  73     def __init__(self, cmake, maker, cmake_dir):
  74         self._prefix = [cmake, "-G", maker, cmake_dir]
  75
  76         # Map of str -> (list|str).
  77         self._flags = {}
  78         for flag in CmakeInvocation._cflags + CmakeInvocation._ldflags:
  79             self._flags[flag] = []
  80
  81     def add_new_flag(self, key, value):
  82         self.add_flag(key, value, allow_overwrites=False)
  83
  84     def add_flag(self, key, value, allow_overwrites=True):
  85         if key not in self._flags:
  86             self._flags[key] = value
  87             return
  88
  89         existing_value = self._flags[key]
  90         if isinstance(existing_value, list):
  91             existing_value.append(value)
  92             return
  93
  94         if not allow_overwrites:
  95             raise ValueError("Invalid overwrite of %s requested" % key)
  96
  97         self._flags[key] = value
  98
  99     def add_cflags(self, flags):
 100         # No, I didn't intend to append ['-', 'O', '2'] to my flags, thanks :)
 101         assert not isinstance(flags, str)
 102         for f in CmakeInvocation._cflags:
 103             self._flags[f].extend(flags)
 104
 105     def add_ldflags(self, flags):
 106         assert not isinstance(flags, str)
 107         for f in CmakeInvocation._ldflags:
 108             self._flags[f].extend(flags)
 109
 110     def to_args(self):
 111         args = self._prefix.copy()
 112         for key, value in sorted(self._flags.items()):
 113             if isinstance(value, list):
 114                 # We preload all of the list-y values (cflags, ...). If we've
 115                 # nothing to add, don't.
 116                 if not value:
 117                     continue
 118                 value = " ".join(value)
 119
 120             arg = "-D" + key
 121             if value != "":
 122                 arg += "=" + value
 123             args.append(arg)
 124         return args
 125
 126
 127 class Env:
 128     def __init__(self, llvm_dir, use_make, output_dir, default_cmake_args, dry_run):
 129         self.llvm_dir = llvm_dir
 130         self.use_make = use_make
 131         self.output_dir = output_dir
 132         self.default_cmake_args = default_cmake_args.copy()
 133         self.dry_run = dry_run
 134
 135     def get_default_cmake_args_kv(self):
 136         return self.default_cmake_args.items()
 137
 138     def get_cmake_maker(self):
 139         return "Ninja" if not self.use_make else "Unix Makefiles"
 140
 141     def get_make_command(self):
 142         if self.use_make:
 143             return ["make", "-j{}".format(multiprocessing.cpu_count())]
 144         return ["ninja"]
 145
 146     def output_subdir(self, name):
 147         return os.path.join(self.output_dir, name)
 148
 149     def has_llvm_subproject(self, name):
 150         if name == "compiler-rt":
 151             subdir = "../compiler-rt"
 152         elif name == "clang":
 153             subdir = "../clang"
 154         else:
 155             raise ValueError("Unknown subproject: %s" % name)
 156
 157         return os.path.isdir(os.path.join(self.llvm_dir, subdir))
 158
 159     # Note that we don't allow capturing stdout/stderr. This works quite nicely
 160     # with dry_run.
 161     def run_command(self, cmd, cwd=None, check=False, silent_unless_error=False):
 162         print("Running `%s` in %s" % (cmd, shlex.quote(cwd or os.getcwd())))
 163
 164         if self.dry_run:
 165             return
 166
 167         if silent_unless_error:
 168             stdout, stderr = subprocess.PIPE, subprocess.STDOUT
 169         else:
 170             stdout, stderr = None, None
 171
 172         # Don't use subprocess.run because it's >= py3.5 only, and it's not too
 173         # much extra effort to get what it gives us anyway.
 174         popen = subprocess.Popen(
 175             cmd, stdin=subprocess.DEVNULL, stdout=stdout, stderr=stderr, cwd=cwd
 176         )
 177         stdout, _ = popen.communicate()
 178         return_code = popen.wait(timeout=0)
 179
 180         if not return_code:
 181             return
 182
 183         if silent_unless_error:
 184             print(stdout.decode("utf-8", "ignore"))
 185
 186         if check:
 187             raise subprocess.CalledProcessError(
 188                 returncode=return_code, cmd=cmd, output=stdout, stderr=None
 189             )
 190
 191
 192 def _get_default_cmake_invocation(env):
 193     inv = CmakeInvocation(
 194         cmake="cmake", maker=env.get_cmake_maker(), cmake_dir=env.llvm_dir
 195     )
 196     for key, value in env.get_default_cmake_args_kv():
 197         inv.add_new_flag(key, value)
 198     return inv
 199
 200
 201 def _get_cmake_invocation_for_bootstrap_from(env, out_dir, skip_tablegens=True):
 202     clang = os.path.join(out_dir, "bin", "clang")
 203     cmake = _get_default_cmake_invocation(env)
 204     cmake.add_new_flag("CMAKE_C_COMPILER", clang)
 205     cmake.add_new_flag("CMAKE_CXX_COMPILER", clang + "++")
 206
 207     # We often get no value out of building new tblgens; the previous build
 208     # should have them. It's still correct to build them, just slower.
 209     def add_tablegen(key, binary):
 210         path = os.path.join(out_dir, "bin", binary)
 211
 212         # Check that this exists, since the user's allowed to specify their own
 213         # stage1 directory (which is generally where we'll source everything
 214         # from). Dry runs should hope for the best from our user, as well.
 215         if env.dry_run or os.path.exists(path):
 216             cmake.add_new_flag(key, path)
 217
 218     if skip_tablegens:
 219         add_tablegen("LLVM_TABLEGEN", "llvm-tblgen")
 220         add_tablegen("CLANG_TABLEGEN", "clang-tblgen")
 221
 222     return cmake
 223
 224
 225 def _build_things_in(env, target_dir, what):
 226     cmd = env.get_make_command() + what
 227     env.run_command(cmd, cwd=target_dir, check=True)
 228
 229
 230 def _run_fresh_cmake(env, cmake, target_dir):
 231     if not env.dry_run:
 232         try:
 233             shutil.rmtree(target_dir)
 234         except FileNotFoundError:
 235             pass
 236
 237         os.makedirs(target_dir, mode=0o755)
 238
 239     cmake_args = cmake.to_args()
 240     env.run_command(cmake_args, cwd=target_dir, check=True, silent_unless_error=True)
 241
 242
 243 def _build_stage1_clang(env):
 244     target_dir = env.output_subdir("stage1")
 245     cmake = _get_default_cmake_invocation(env)
 246     _run_fresh_cmake(env, cmake, target_dir)
 247     _build_things_in(env, target_dir, what=["clang", "llvm-profdata", "profile"])
 248     return target_dir
 249
 250
 251 def _generate_instrumented_clang_profile(env, stage1_dir, profile_dir, output_file):
 252     llvm_profdata = os.path.join(stage1_dir, "bin", "llvm-profdata")
 253     if env.dry_run:
 254         profiles = [os.path.join(profile_dir, "*.profraw")]
 255     else:
 256         profiles = [
 257             os.path.join(profile_dir, f)
 258             for f in os.listdir(profile_dir)
 259             if f.endswith(".profraw")
 260         ]
 261     cmd = [llvm_profdata, "merge", "-output=" + output_file] + profiles
 262     env.run_command(cmd, check=True)
 263
 264
 265 def _build_instrumented_clang(env, stage1_dir):
 266     assert os.path.isabs(stage1_dir)
 267
 268     target_dir = os.path.join(env.output_dir, "instrumented")
 269     cmake = _get_cmake_invocation_for_bootstrap_from(env, stage1_dir)
 270     cmake.add_new_flag("LLVM_BUILD_INSTRUMENTED", "IR")
 271
 272     # libcxx's configure step messes with our link order: we'll link
 273     # libclang_rt.profile after libgcc, and the former requires atexit from the
 274     # latter. So, configure checks fail.
 275     #
 276     # Since we don't need libcxx or compiler-rt anyway, just disable them.
 277     cmake.add_new_flag("LLVM_BUILD_RUNTIME", "No")
 278
 279     _run_fresh_cmake(env, cmake, target_dir)
 280     _build_things_in(env, target_dir, what=["clang", "lld"])
 281
 282     profiles_dir = os.path.join(target_dir, "profiles")
 283     return target_dir, profiles_dir
 284
 285
 286 def _build_optimized_clang(env, stage1_dir, profdata_file):
 287     if not env.dry_run and not os.path.exists(profdata_file):
 288         raise ValueError(
 289             "Looks like the profdata file at %s doesn't exist" % profdata_file
 290         )
 291
 292     target_dir = os.path.join(env.output_dir, "optimized")
 293     cmake = _get_cmake_invocation_for_bootstrap_from(env, stage1_dir)
 294     cmake.add_new_flag("LLVM_PROFDATA_FILE", os.path.abspath(profdata_file))
 295
 296     # We'll get complaints about hash mismatches in `main` in tools/etc. Ignore
 297     # it.
 298     cmake.add_cflags(["-Wno-backend-plugin"])
 299     _run_fresh_cmake(env, cmake, target_dir)
 300     _build_things_in(env, target_dir, what=["clang"])
 301     return target_dir
 302
 303
 304 Args = collections.namedtuple(
 305     "Args",
 306     [
 307         "do_optimized_build",
 308         "include_debug_info",
 309         "profile_location",
 310         "stage1_dir",
 311     ],
 312 )
 313
 314
 315 def _parse_args():
 316     parser = argparse.ArgumentParser(
 317         description="Builds LLVM and Clang with instrumentation, collects "
 318         "instrumentation profiles for them, and (optionally) builds things "
 319         "with these PGO profiles. By default, it's assumed that you're "
 320         "running this from your LLVM root, and all build artifacts will be "
 321         "saved to $PWD/out."
 322     )
 323     parser.add_argument(
 324         "--cmake-extra-arg",
 325         action="append",
 326         default=[],
 327         help="an extra arg to pass to all cmake invocations. Note that this "
 328         "is interpreted as a -D argument, e.g. --cmake-extra-arg FOO=BAR will "
 329         "be passed as -DFOO=BAR. This may be specified multiple times.",
 330     )
 331     parser.add_argument(
 332         "--dry-run", action="store_true", help="print commands instead of running them"
 333     )
 334     parser.add_argument(
 335         "--llvm-dir",
 336         default=".",
 337         help="directory containing an LLVM checkout (default: $PWD)",
 338     )
 339     parser.add_argument(
 340         "--no-optimized-build",
 341         action="store_true",
 342         help="disable the final, PGO-optimized build",
 343     )
 344     parser.add_argument(
 345         "--out-dir", help="directory to write artifacts to (default: $llvm_dir/out)"
 346     )
 347     parser.add_argument(
 348         "--profile-output",
 349         help="where to output the profile (default is $out/pgo_profile.prof)",
 350     )
 351     parser.add_argument(
 352         "--stage1-dir",
 353         help="instead of having an initial build of everything, use the given "
 354         "directory. It is expected that this directory will have clang, "
 355         "llvm-profdata, and the appropriate libclang_rt.profile already built",
 356     )
 357     parser.add_argument(
 358         "--use-debug-info-in-benchmark",
 359         action="store_true",
 360         help="use a regular build instead of RelWithDebInfo in the benchmark. "
 361         "This increases benchmark execution time and disk space requirements, "
 362         "but gives more coverage over debuginfo bits in LLVM and clang.",
 363     )
 364     parser.add_argument(
 365         "--use-make",
 366         action="store_true",
 367         default=shutil.which("ninja") is None,
 368         help="use Makefiles instead of ninja",
 369     )
 370
 371     args = parser.parse_args()
 372
 373     llvm_dir = os.path.abspath(args.llvm_dir)
 374     if args.out_dir is None:
 375         output_dir = os.path.join(llvm_dir, "out")
 376     else:
 377         output_dir = os.path.abspath(args.out_dir)
 378
 379     extra_args = {
 380         "CMAKE_BUILD_TYPE": "Release",
 381         "LLVM_ENABLE_PROJECTS": "clang;compiler-rt;lld",
 382     }
 383     for arg in args.cmake_extra_arg:
 384         if arg.startswith("-D"):
 385             arg = arg[2:]
 386         elif arg.startswith("-"):
 387             raise ValueError(
 388                 "Unknown not- -D arg encountered; you may need "
 389                 "to tweak the source..."
 390             )
 391         split = arg.split("=", 1)
 392         if len(split) == 1:
 393             key, val = split[0], ""
 394         else:
 395             key, val = split
 396         extra_args[key] = val
 397
 398     env = Env(
 399         default_cmake_args=extra_args,
 400         dry_run=args.dry_run,
 401         llvm_dir=llvm_dir,
 402         output_dir=output_dir,
 403         use_make=args.use_make,
 404     )
 405
 406     if args.profile_output is not None:
 407         profile_location = args.profile_output
 408     else:
 409         profile_location = os.path.join(env.output_dir, "pgo_profile.prof")
 410
 411     result_args = Args(
 412         do_optimized_build=not args.no_optimized_build,
 413         include_debug_info=args.use_debug_info_in_benchmark,
 414         profile_location=profile_location,
 415         stage1_dir=args.stage1_dir,
 416     )
 417
 418     return env, result_args
 419
 420
 421 def _looks_like_llvm_dir(directory):
 422     """Arbitrary set of heuristics to determine if `directory` is an llvm dir.
 423
 424     Errs on the side of false-positives."""
 425
 426     contents = set(os.listdir(directory))
 427     expected_contents = [
 428         "CODE_OWNERS.TXT",
 429         "cmake",
 430         "docs",
 431         "include",
 432         "utils",
 433     ]
 434
 435     if not all(c in contents for c in expected_contents):
 436         return False
 437
 438     try:
 439         include_listing = os.listdir(os.path.join(directory, "include"))
 440     except NotADirectoryError:
 441         return False
 442
 443     return "llvm" in include_listing
 444
 445
 446 def _die(*args, **kwargs):
 447     kwargs["file"] = sys.stderr
 448     print(*args, **kwargs)
 449     sys.exit(1)
 450
 451
 452 def _main():
 453     env, args = _parse_args()
 454
 455     if not _looks_like_llvm_dir(env.llvm_dir):
 456         _die("Looks like %s isn't an LLVM directory; please see --help" % env.llvm_dir)
 457     if not env.has_llvm_subproject("clang"):
 458         _die("Need a clang checkout at tools/clang")
 459     if not env.has_llvm_subproject("compiler-rt"):
 460         _die("Need a compiler-rt checkout at projects/compiler-rt")
 461
 462     def status(*args):
 463         print(*args, file=sys.stderr)
 464
 465     if args.stage1_dir is None:
 466         status("*** Building stage1 clang...")
 467         stage1_out = _build_stage1_clang(env)
 468     else:
 469         stage1_out = args.stage1_dir
 470
 471     status("*** Building instrumented clang...")
 472     instrumented_out, profile_dir = _build_instrumented_clang(env, stage1_out)
 473     status("*** Running profdata benchmarks...")
 474     _run_benchmark(env, instrumented_out, args.include_debug_info)
 475     status("*** Generating profile...")
 476     _generate_instrumented_clang_profile(
 477         env, stage1_out, profile_dir, args.profile_location
 478     )
 479
 480     print("Final profile:", args.profile_location)
 481     if args.do_optimized_build:
 482         status("*** Building PGO-optimized binaries...")
 483         optimized_out = _build_optimized_clang(env, stage1_out, args.profile_location)
 484         print("Final build directory:", optimized_out)
 485
 486
 487 if __name__ == "__main__":
 488     _main()