github/verilator
1
0
Fork 1
mirror of https://github.com/verilator/verilator.git synced 2025-01-01 04:07:34 +00:00
Code Issues Packages Projects Releases Wiki Activity
685ea0bc65
verilator/test_regress/driver.py
Geza Lore 59b07529b1
driver.py: fix aslr_off (#5477)
2024-09-21 16:37:51 +01:00

2827 lines
116 KiB
Python
Executable File
Raw Blame History

#!/usr/bin/env python3
# pylint: disable=C0103,C0114,C0115,C0116,C0209,C0302,R0902,R0903,R0904,R0912,R0913,R0914,R0915,R0916,W0212,W0511,W0603,W1201
######################################################################
import argparse
import collections
import glob
import hashlib
import json
import logging
import multiprocessing
import os
import pickle
import platform
import re
import runpy
import shutil
import signal
import subprocess
import sys
import time
from functools import lru_cache # Eventually use python 3.9's cache
from pprint import pformat, pprint
from packaging import version
if False: # pylint: disable=using-constant-test
pprint(pformat("Ignored")) # Prevent unused warning
# Map of all scenarios, with the names used to enable them
All_Scenarios = {
# yapf: disable
'dist': ['dist'],
'atsim': ['simulator', 'simulator_st', 'atsim'],
'ghdl': ['linter', 'simulator', 'simulator_st', 'ghdl'],
'iv': ['simulator', 'simulator_st', 'iv'],
'ms': ['linter', 'simulator', 'simulator_st', 'ms'],
'nc': ['linter', 'simulator', 'simulator_st', 'nc'],
'vcs': ['linter', 'simulator', 'simulator_st', 'vcs'],
'xrun': ['linter', 'simulator', 'simulator_st', 'xrun'],
'xsim': ['linter', 'simulator', 'simulator_st', 'xsim'],
'vlt': ['linter', 'simulator', 'simulator_st', 'vlt_all', 'vlt'],
'vltmt': ['simulator', 'vlt_all', 'vltmt'],
# yapf: enable
}
# Globals
test = None
Arg_Tests = []
Quitting = False
Vltmt_Threads = 3
# So an 'import vltest_bootstrap' inside test files will do nothing
sys.modules['vltest_bootstrap'] = {}
#######################################################################
#######################################################################
# Decorators
class staticproperty(property):
def __get__(self, owner_self, owner_cls):
return self.fget()
#######################################################################
#######################################################################
# VtOs -- OS extensions
class VtOs:
@staticmethod
def delenv(var: str) -> None:
"""Delete environment variable, if exists"""
if var in os.environ:
del os.environ[var]
@staticmethod
def getenv_def(var: str, default=None) -> str:
"""Return environment variable, returning default if does not exist"""
if var in os.environ:
return os.environ[var]
return default
@staticmethod
def mkdir_ok(path: str) -> None:
"""Make directory, no error if exists"""
try:
os.mkdir(path)
except FileExistsError:
pass
@staticmethod
def run_capture(command: str, check=True) -> str:
"""Run a command and return results"""
proc = subprocess.run([command], capture_output=True, text=True, shell=True, check=False)
if check and proc.returncode:
sys.exit("%Error: command failed '" + command + "':\n" + proc.stderr + proc.stdout)
return str(proc.stdout)
@staticmethod
def unlink_ok(filename: str) -> None:
"""Unlink a file, no error if fails"""
try:
os.unlink(filename)
except OSError:
pass
#######################################################################
#######################################################################
# Capabilities -- What OS/Verilator is built to support
class Capabilities:
# @lru_cache(maxsize=1024) broken with @staticmethod on older pythons we use
_cached_cmake_version = None
_cached_cxx_version = None
_cached_have_coroutines = None
_cached_have_gdb = None
_cached_have_sc = None
_cached_have_solver = None
_cached_make_version = None
@staticproperty
def cmake_version() -> str: # pylint: disable=no-method-argument
if Capabilities._cached_cmake_version is None:
out = VtOs.run_capture('cmake --version', check=False)
match = re.search(r'cmake version (\d+)\.(\d+)', out, re.IGNORECASE)
if match:
Capabilities._cached_cmake_version = match.group(1) + "." + match.group(2)
else:
Capabilities._cached_cmake_version = 0
return Capabilities._cached_cmake_version
@staticproperty
def cxx_version() -> str: # pylint: disable=no-method-argument
if Capabilities._cached_cxx_version is None:
Capabilities._cached_cxx_version = VtOs.run_capture(os.environ['MAKE'] + " -C " +
os.environ['TEST_REGRESS'] +
" -f Makefile print-cxx-version",
check=False)
return Capabilities._cached_cxx_version
@staticproperty
def have_coroutines() -> bool: # pylint: disable=no-method-argument
if Capabilities._cached_have_coroutines is None:
Capabilities._cached_have_coroutines = bool(
Capabilities._verilator_get_supported('COROUTINES'))
return Capabilities._cached_have_coroutines
@staticproperty
def have_gdb() -> bool: # pylint: disable=no-method-argument
if Capabilities._cached_have_gdb is None:
out = VtOs.run_capture('gdb --version 2>/dev/null', check=False)
Capabilities._cached_have_gdb = bool('Copyright' in out)
return Capabilities._cached_have_gdb
@staticproperty
def have_sc() -> bool: # pylint: disable=no-method-argument
if Capabilities._cached_have_sc is None:
if 'SYSTEMC' in os.environ:
Capabilities._cached_have_sc = True
elif 'SYSTEMC_INCLUDE' in os.environ:
Capabilities._cached_have_sc = True
elif 'CFG_HAVE_SYSTEMC' in os.environ:
Capabilities._cached_have_sc = True
else:
Capabilities._cached_have_sc = bool(
Capabilities._verilator_get_supported('SYSTEMC'))
return Capabilities._cached_have_sc
@staticproperty
def have_solver() -> bool: # pylint: disable=no-method-argument
if Capabilities._cached_have_solver is None:
out = VtOs.run_capture('(z3 --help || cvc5 --help || cvc4 --help) 2>/dev/null',
check=False)
Capabilities._cached_have_solver = bool('Usage' in out)
return Capabilities._cached_have_solver
@staticproperty
@lru_cache(maxsize=1024)
def make_version() -> str: # pylint: disable=no-method-argument
if Capabilities._cached_make_version is None:
out = VtOs.run_capture(os.environ['MAKE'] + ' --version', check=False)
match = re.search(r'make ([0-9]+\.[0-9]+)', out, flags=re.IGNORECASE)
if match:
Capabilities._cached_make_version = match.group(1)
else:
Capabilities._cached_make_version = -1
return Capabilities._cached_make_version
# Fetch
@staticmethod
def warmup_cache() -> None:
_ignore = Capabilities.have_coroutines
_ignore = Capabilities.have_gdb
_ignore = Capabilities.have_sc
_ignore = Capabilities.have_solver
# Internals
@staticmethod
def _verilator_get_supported(feature) -> str:
# Returns if given feature is supported
cmd = "perl " + os.environ['VERILATOR_ROOT'] + "/bin/verilator -get-supported " + feature
out = VtOs.run_capture(cmd, check=False).strip()
if out == '1':
return True
if out in ('', '0'):
return False
sys.exit("%Error: couldn't verilator_get_supported: " + cmd + "\n" + out)
#######################################################################
#######################################################################
# Forker class - run multprocess pool of processes
# Similar interface to Perl's Parallel::Forker.
class Forker:
class Job:
def __init__(self, _id, name, scenario, args, quiet, rerun_skipping, run_pre_start,
run_on_start, run_on_finish):
self.fail_max_skip = False
self.id = _id
self.name = name
self.scenario = scenario
self.args = args
self.quiet = quiet
self.rerun_skipping = rerun_skipping
self.run_pre_start = run_pre_start
self.run_on_start = run_on_start
self.run_on_finish = run_on_finish
self.mprocess = None # Set once call run()
@property
def pid(self) -> int:
return self.mprocess.pid
@property
def exitcode(self) -> int:
return self.mprocess.exitcode
def __init__(self, max_processes):
self._max_processes = max_processes
self._id_next = 0
self._left = collections.deque() # deque of Job
self._running = {} # key of pid, value of Job
def is_any_left(self) -> bool:
return self.num_running() > 0 or (len(self._left) > 0 and not Quitting)
def max_proc(self, n: int) -> None:
self._max_processes = n
def poll(self) -> bool:
"""Run threads, returning if more work to do (if False, sleep)"""
# We don't use SIGCHLD as conflicted with other handler, instead just poll
completed = [] # Need two passes to avoid changing the list we are iterating
nrunning = 0
more_now = False
for process in self._running.values():
if process.exitcode is not None:
completed.append(process)
more_now = True
else:
nrunning += 1
# Start new work now, so running in background while we then collect completions
while len(self._left) and nrunning < self._max_processes and not Quitting:
process = self._left.popleft()
self._run(process)
nrunning += 1
more_now = True
for process in completed:
self._finished(process)
return more_now
def running(self) -> list:
return self._running.values()
def num_running(self) -> int:
return len(self._running)
def schedule(self, name, scenario, args, quiet, rerun_skipping, run_pre_start, run_on_start,
run_on_finish):
# print("-Forker::schedule: [" + name + "]")
process = Forker.Job(self._id_next,
name=name,
scenario=scenario,
args=args,
quiet=quiet,
rerun_skipping=rerun_skipping,
run_pre_start=run_pre_start,
run_on_start=run_on_start,
run_on_finish=run_on_finish)
self._id_next += 1
self._left.append(process)
def kill_tree_all(self) -> None:
# print("-Forker: kill_tree_all")
for process in self._running.values():
process.mprocess.kill()
def _run(self, process) -> None:
# print("-Forker: [" + process.name + "] run_pre_start")
process.run_pre_start(process)
ctx = multiprocessing.get_context('forkserver')
process.mprocess = ctx.Process( #
target=forker.child_start, # pylint: disable=used-before-assignment
args=(process, ))
process.mprocess.start()
# print("-Forker: [" + process.name + "] RUNNING pid=" + str(process.pid))
self._running[process.pid] = process
@staticmethod
def child_start(process) -> None:
# Runs in context of child
# print("-Forker: [" + process.name + "] run_on_start")
process.run_on_start(process)
# print("-Forker: [" + process.name + "] FINISHED run_on_start")
sys.exit(0)
# Don't close anything
def _finished(self, process) -> None:
del self._running[process.pid]
# print("-Forker: [" + process.name + "] run_on_finish exitcode=" + str(process.exitcode))
process.run_on_finish(process)
process.mprocess.close()
#######################################################################
#######################################################################
# Runner class
class Runner:
def __init__(self, driver_log_filename, quiet, ok_cnt=0, fail1_cnt=0, skip_cnt=0):
self.driver_log_filename = driver_log_filename
self.quiet = quiet
# Counts
self.all_cnt = 0
self.left_cnt = 0
self.ok_cnt = ok_cnt
self.fail1_cnt = fail1_cnt
self.fail_cnt = 0
self.skip_cnt = skip_cnt
self.skip_msgs = []
self.fail_msgs = []
self.fail_tests = []
self._last_proc_finish_time = 0
self._last_summary_time = 0
self._last_summary_left = 0
self._running_ids = {}
self._msg_fail_max_skip = False
Runner.runner = self
def one_test(self, py_filename: str, scenario: str, rerun_skipping=False) -> None:
self.all_cnt += 1
self.left_cnt += 1
forker.schedule(name=py_filename,
scenario=scenario,
args=Args,
quiet=self.quiet,
rerun_skipping=rerun_skipping,
run_pre_start=self._run_pre_start_static,
run_on_start=self._run_on_start_static,
run_on_finish=self._run_on_finish_static)
@staticmethod
def _run_pre_start_static(process) -> None:
Runner.runner._run_pre_start(process) # pylint: disable=protected-access
def _run_pre_start(self, process) -> None:
# Running in context of parent, before run_on_start
# Make an identifier that is unique across all current running jobs
i = 1
while i in self._running_ids:
i += 1
process.running_id = i
self._running_ids[process.running_id] = 1
if Args.fail_max and Args.fail_max <= self.fail_cnt:
if not self._msg_fail_max_skip:
self._msg_fail_max_skip = True
print("== Too many test failures; exceeded --fail-max\n", file=sys.stderr)
process.fail_max_skip = 1
@staticmethod
def _run_on_start_static(process) -> None:
# Running in context of child, so can't pass data to parent directly
if process.quiet:
sys.stdout = open(os.devnull, 'w') # pylint: disable=R1732,unspecified-encoding
sys.stderr = open(os.devnull, 'w') # pylint: disable=R1732,unspecified-encoding
print("=" * 70)
global Args
Args = process.args
global test
test = VlTest(py_filename=process.name,
scenario=process.scenario,
running_id=process.running_id)
test.oprint("=" * 50)
test._prep()
if process.rerun_skipping:
print(" ---------- Earlier logfiles below; test was rerunnable = False\n")
os.system("cat " + test.obj_dir + "/*.log")
print(" ---------- Earlier logfiles above; test was rerunnable = False\n")
elif process.fail_max_skip:
test.skip("Too many test failures; exceeded --fail-max")
else:
VtOs.unlink_ok(test._status_filename)
test._read()
# Don't put anything other than _exit after _read,
# as may call _exit via another path
test._exit()
@staticmethod
def _run_on_finish_static(process) -> None:
Runner.runner._run_on_finish(process) # pylint: disable=protected-access
def _run_on_finish(self, process) -> None:
# Running in context of parent
global test
test = VlTest(py_filename=process.name,
scenario=process.scenario,
running_id=process.running_id)
test._quit = Quitting
test._read_status()
if test.ok:
self.ok_cnt += 1
elif test._quit:
pass
elif test._scenario_off and not test.errors:
pass
elif test._skips and not test.errors and not test.errors_keep_going:
self.skip_msgs.append("\t#" + test.soprint("-Skip: " + test._skips))
self.skip_cnt += 1
else:
error_msg = test.errors if test.errors else test.errors_keep_going
test.oprint("FAILED: " + error_msg)
makecmd = VtOs.getenv_def('VERILATOR_MAKE', os.environ['MAKE']) + " &&"
upperdir = 'test_regress/' if re.search(r'test_regress', os.getcwd()) else ''
self.fail_msgs.append("\t#" + test.soprint("%Error: " + error_msg) + "\t\t" + makecmd +
" " + upperdir + test.py_filename + ' ' +
' '.join(self._manual_args()) + " --" + test.scenario + "\n")
self.fail_tests.append(test)
self.fail_cnt += 1
self.report(self.driver_log_filename)
other = ""
for proc in forker.running():
other += " " + proc.name
if other != "" and not Args.quiet:
test.oprint("Simultaneous running tests:" + other)
if Args.stop:
sys.exit("%Error: --stop and errors found\n")
self.left_cnt -= 1
self._last_proc_finish_time = time.time()
if process.running_id:
del self._running_ids[process.running_id]
def wait_and_report(self) -> None:
self.print_summary(force=True)
# Wait for all children to finish
while forker.is_any_left():
more_now = forker.poll()
if not Args.interactive_debugger:
self.print_summary(force=False)
if not more_now:
time.sleep(0.1)
self.report(None)
self.report(self.driver_log_filename)
def report(self, filename: str) -> None:
if filename:
with open(filename, "w", encoding="utf8") as fh:
self._report_fh(fh)
else:
self._report_fh(sys.stdout)
def _report_fh(self, fh) -> None:
fh.write("\n")
fh.write('=' * 70 + "\n")
for f in sorted(self.fail_msgs):
fh.write(f.strip() + "\n")
for f in sorted(self.skip_msgs):
fh.write(f.strip() + "\n")
if self.fail_cnt:
sumtxt = 'FAILED'
elif self.skip_cnt:
sumtxt = 'PASSED w/SKIPS'
else:
sumtxt = 'PASSED'
fh.write("==TESTS DONE, " + sumtxt + ": " + self.sprint_summary() + "\n")
def print_summary(self, force=False):
change = self._last_summary_left != self.left_cnt
if (force or ((time.time() - self._last_summary_time) >= 15)
or (not self.quiet and change)):
self._last_summary_left = self.left_cnt
self._last_summary_time = time.time()
print("==SUMMARY: " + self.sprint_summary(), file=sys.stderr)
if (self._last_proc_finish_time != 0
and ((time.time() - self._last_proc_finish_time) > 15)):
self._last_proc_finish_time = time.time()
other = ""
for proc in forker.running():
other += " " + proc.name
print("==STILL RUNNING:" + other, file=sys.stderr)
@staticmethod
def _py_filename_adjust(py_filename: str,
tdir_def: str) -> list: # Return (py_filename, t_dir)
for tdir in Args.test_dirs: # pylint: disable=redefined-outer-name
# t_dir used both absolutely and under obj_dir
try_py_filename = tdir + "/" + os.path.basename(py_filename)
if os.path.exists(try_py_filename):
# Note most tests require error messages of the form t/x.v
# Therefore py_filename must be t/ for local tests
# t_dir must be absolute - used under t or under obj_dir
tdir_abs = os.path.abspath(tdir)
return (try_py_filename, tdir_abs)
return (py_filename, os.path.abspath(tdir_def))
def sprint_summary(self) -> str:
delta = time.time() - Start # pylint: disable=used-before-assignment
# Fudge of 120% works out about right so ETA correctly predicts completion time
eta = 1.2 * ((self.all_cnt * (delta / ((self.all_cnt - self.left_cnt) + 0.001))) - delta)
if delta < 10:
eta = 0
out = ""
if self.left_cnt:
out += "Left " + str(self.left_cnt) + " "
out += "Passed " + str(self.ok_cnt)
# Ordered below most severe to least severe
out += " Failed " + str(self.fail_cnt)
if self.fail1_cnt:
out += " Failed-First " + str(self.fail1_cnt)
if self.skip_cnt:
out += " Skipped " + str(self.skip_cnt)
if forker.num_running():
out += " Running " + str(forker.num_running())
if self.left_cnt > 10 and eta > 10:
out += " Eta %d:%02d" % (int(eta / 60), eta % 60)
out += " Time %d:%02d" % (int(delta / 60), delta % 60)
return out
def _manual_args(self) -> str:
# Return command line with scenarios stripped
out = []
for oarg in Args.orig_argv_sw:
showarg = True
for val in All_Scenarios.values():
for allscarg in val:
if oarg == "--" + allscarg:
showarg = False
# Also strip certain flags that per-test debugging won't want
if showarg and oarg != '--rerun' and oarg != '--quiet':
out.append(oarg)
return out
#######################################################################
#######################################################################
# Test exceptions
class VtSkipException(Exception):
pass
class VtErrorException(Exception):
pass
#######################################################################
#######################################################################
# Test class
class VlTest:
_file_contents_cache = {}
# @lru_cache(maxsize=1024) broken with @staticmethod on older pythons we use
_cached_aslr_off = None
_cached_cfg_with_ccache = None
def __init__(self, py_filename, scenario, running_id):
self.py_filename = py_filename # Name of .py file to get setup from
self.running_id = running_id
self.scenario = scenario
self._have_solver_called = False
self._inputs = {}
self._ok = False
self._quit = False
self._scenario_off = False # scenarios() didn't match running scenario
self._skips = None
match = re.match(r'^(.*/)?([^/]*)\.py', self.py_filename)
self.name = match.group(2) # Name of this test
self.benchmark = Args.benchmark
self.benchmarksim = False
self.clean_command = None
self.context_threads = 0 # Number of threads to allocate in the context
self.errors = None
self.errors_keep_going = None
self.main_time_multiplier = 1
self.make_main = 1 # Make __main.cpp
self.make_pli = 0 # need to compile pli
self.make_top_shell = 1 # Make a default __top.v file
self.rerunnable = True # Rerun if fails
self.sc_time_resolution = "SC_PS" # Keep - PS is SystemC default
self.sim_time = 1100 # simulation time units for main wrapper
self.threads = -1 # --threads (negative means auto based on scenario)
self.verbose = Args.verbose
self.verilated_randReset = ""
self.vm_prefix = "V" + self.name
# Make e.g. self.vlt, self.vltmt etc
self.vlt = False # Set below also
self.vltmt = False # Set below also
self.xsim = False # Set below also
for ascenario in All_Scenarios:
self.__dict__[ascenario] = False
self.__dict__[scenario] = True
self.vlt_all = self.vlt or self.vltmt # Any Verilator scenario
(self.py_filename, self.t_dir) = Runner._py_filename_adjust(self.py_filename, ".")
for tdir in Args.test_dirs: # pylint: disable=redefined-outer-name
# t_dir used both absolutely and under obj_dir
self.t_dir = None
if os.path.exists(tdir + "/" + self.name + ".py"):
# Note most tests require error messages of the form t/x.v
# Therefore py_filename must be t/ for local tests
self.py_filename = os.path.relpath(tdir + "/" + self.name + ".py")
# t_dir must be absolute - used under t or under obj_dir
self.t_dir = os.path.abspath(tdir)
break
if not self.t_dir:
sys.exit("%Error: Can't locate dir for " + self.name)
scen_dir = os.path.relpath(self.t_dir + "/../obj_" + self.scenario)
# Simplify filenames on local runs
scen_dir = re.sub(r'^t/\.\./', '', scen_dir)
# Not mkpath so error if try to build somewhere odd
VtOs.mkdir_ok(scen_dir)
self.obj_dir = scen_dir + "/" + self.name
define_opt = self._define_opt_calc()
# All compilers
self.v_flags = []
if self.xsim:
self.v_flags += ["-f input.xsim.vc"]
elif os.path.exists('input.vc'):
self.v_flags += ["-f input.vc"]
if not re.search(r'/test_regress', self.t_dir): # Don't include standard dir, only site's
self.v_flags += ["+incdir+" + self.t_dir + " -y " + self.t_dir]
self.v_flags += [define_opt + "TEST_OBJ_DIR=" + self.obj_dir]
if Args.verbose:
self.v_flags += [define_opt + "TEST_VERBOSE=1"]
if Args.benchmark:
self.v_flags += [define_opt + "TEST_BENCHMARK=Args.benchmark"]
if Args.trace:
self.v_flags += [define_opt + "WAVES=1"]
self.v_flags2 = [] # Overridden in some sim files
self.v_other_filenames = []
self.all_run_flags = []
self.pli_flags = [
"-I" + os.environ['VERILATOR_ROOT'] + "/include/vltstd", "-fPIC", "-shared"
]
if platform.system() == 'Darwin':
self.pli_flags += ["-Wl,-undefined,dynamic_lookup"]
else:
self.pli_flags += ["-rdynamic"]
if Args.verbose:
self.pli_flags += ["-DTEST_VERBOSE=1"]
self.pli_flags += ["-o", self.obj_dir + "/libvpi.so"]
self.tool_c_flags = []
# ATSIM
self.atsim_define = 'ATSIM'
self.atsim_flags = [
"-c", "+sv", "+define+ATSIM", ("+sv_dir+" + self.obj_dir + "/.athdl_compile")
]
self.atsim_flags2 = [] # Overridden in some sim files
self.atsim_run_flags = []
# GHDL
self.ghdl_define = 'GHDL'
self.ghdl_work_dir = self.obj_dir + "/ghdl_compile"
self.ghdl_flags = [("-v" if Args.debug else ""),
("--workdir=" + self.obj_dir + "/ghdl_compile")]
self.ghdl_flags2 = [] # Overridden in some sim files
self.ghdl_run_flags = []
# IV
self.iv_define = 'IVERILOG'
self.iv_flags = ["+define+IVERILOG", "-g2012", ("-o" + self.obj_dir + "/simiv")]
self.iv_flags2 = [] # Overridden in some sim files
self.iv_run_flags = []
# VCS
self.vcs_define = 'VCS'
self.vcs_flags = [
"+vcs+lic+wait", "+cli", "-debug_access", "+define+VCS+1", "-q", "-sverilog",
"-CFLAGS", "'-DVCS'"
]
self.vcs_flags2 = [] # Overridden in some sim files
self.vcs_run_flags = ["+vcs+lic_wait"]
# NC
self.nc_define = 'NC'
self.nc_flags = [
"+licqueue", "+nowarn+LIBNOU", "+define+NC=1", "-q", "+assert", "+sv", "-c",
("+access+r" if Args.trace else "")
]
self.nc_flags2 = [] # Overridden in some sim files
self.nc_run_flags = ["+licqueue", "-q", "+assert", "+sv", "-R"]
# ModelSim
self.ms_define = 'MS'
self.ms_flags = [
"-sv", "-work", (self.obj_dir + "/work"), "+define+MS=1", "-ccflags", '\"-DMS=1\"'
]
self.ms_flags2 = [] # Overridden in some sim files
self.ms_run_flags = ["-lib", self.obj_dir + "/work", "-c", "-do", "'run -all;quit'"]
# XSim
self.xsim_define = 'XSIM'
self.xsim_flags = [
"--nolog", "--sv", "--define", "XSIM", "--work",
(self.name + "=" + self.obj_dir + "/xsim")
]
self.xsim_flags2 = [] # Overridden in some sim files
self.xsim_run_flags = [
"--nolog", "--runall", "--lib", (self.name + "=" + self.obj_dir + "/xsim"),
(" --debug all" if Args.trace else "")
]
self.xsim_run_flags2 = [] # Overridden in some sim files
# Xcelium (xrun)
self.xrun = 0
self.xrun_define = 'XRUN'
self.xrun_flags = [] # Doesn't really have a compile step
self.xrun_flags2 = [] # Overridden in some sim files
self.xrun_run_flags = [
"-64", "-access", "+rwc", "-newsv", "-sv", "-xmlibdirname", self.obj_dir + "/work",
"-l", self.obj_dir + "/history", "-quiet", "-plinowarn"
]
# Verilator
self.verilator_define = 'VERILATOR'
self.verilator_flags = [
"-cc",
"-Mdir",
self.obj_dir,
"--fdedup", # As currently disabled unless -O3
"--debug-check",
"--comp-limit-members",
"10"
]
self.verilator_flags2 = []
self.verilator_flags3 = ["--clk clk"]
self.verilator_make_gmake = True
self.verilator_make_cmake = False
self.verilated_debug = Args.verilated_debug
self._status_filename = self.obj_dir + "/V" + self.name + ".status"
self.coverage_filename = self.obj_dir + "/coverage.dat"
self.golden_filename = re.sub(r'\.py$', '.out', self.py_filename)
self.main_filename = self.obj_dir + "/" + self.vm_prefix + "__main.cpp"
self.compile_log_filename = self.obj_dir + "/vlt_compile.log"
self.run_log_filename = self.obj_dir + "/vlt_sim.log"
self.stats = self.obj_dir + "/V" + self.name + "__stats.txt"
self.top_filename = re.sub(r'\.py$', '', self.py_filename) + '.' + self.v_suffix
self.pli_filename = re.sub(r'\.py$', '', self.py_filename) + '.cpp'
self.top_shell_filename = self.obj_dir + "/" + self.vm_prefix + "__top.v"
def _define_opt_calc(self) -> str:
return "--define " if self.xsim else "+define+"
def init_benchmarksim(self) -> None:
# Simulations with benchmarksim enabled append to the same file between runs.
# Test files must ensure a clean benchmark data file before executing tests.
filename = self.benchmarksim_filename
with open(filename, 'w', encoding="utf8") as fh:
fh.write("# Verilator simulation benchmark data\n")
fh.write("# Test name: " + self.name + "\n")
fh.write("# Top file: " + self.top_filename + "\n")
fh.write("evals, time[s]\n")
def soprint(self, message: str) -> str:
message = message.rstrip() + "\n"
message = self.scenario + "/" + self.name + ": " + message
return message
def oprint(self, message: str) -> None:
message = message.rstrip() + "\n"
print(self.soprint(message), end="")
def error(self, message: str) -> None:
"""Called from tests as: error("Reason message")
Newline is optional. Only first line is passed to summaries
Throws a VtErrorException, so rest of testing is not executed"""
message = message.rstrip() + "\n"
print("%Warning: " + self.scenario + "/" + self.name + ": " + message,
file=sys.stderr,
end="")
if not self.errors:
message = re.sub(r'\n.*', '\n', message)
self.errors = message
raise VtErrorException
def error_keep_going(self, message: str) -> None:
"""Called from tests as: error_keep_going("Reason message")
Newline is optional. Only first line is passed to summaries"""
if self._quit:
return
message = message.rstrip() + "\n"
print("%Warning: " + self.scenario + "/" + self.name + ": " + message,
file=sys.stderr,
end="")
if not self.errors_keep_going:
message = re.sub(r'\n.*', '\n', message)
self.errors_keep_going = message
def skip(self, message: str) -> None:
"""Called from tests as: skip("Reason message"[, ...])
Newline is optional. Only first line is passed to summaries.
Throws a VtSkipException, so rest of testing is not executed"""
message = message.rstrip() + "\n"
print("-Skip: " + self.scenario + "/" + self.name + ": " + message,
file=sys.stderr,
end="")
if not self._skips:
message = re.sub(r'\n.*', '\n', message)
self._skips = message
raise VtSkipException
def scenarios(self, *scenario_list) -> None:
"""Called from tests as: scenarios(<list_of_scenarios>) to
specify which scenarios this test runs under. Where ... is
one cases listed in All_Scenarios.
All scenarios must be on one line; this is parsed outside Python."""
enabled_scenarios = {}
for param in scenario_list:
hit = False
for allsc, allscs in All_Scenarios.items():
for allscarg in allscs:
if param == allscarg:
hit = True
enabled_scenarios[allsc] = True
if not hit:
self.error("scenarios(...) has unknown scenario '" + param + "'")
if not enabled_scenarios.get(self.scenario, None):
self._scenario_off = True
self.skip("scenario '" + self.scenario + "' not enabled for test")
# self._exit() implied by skip's exception
@staticmethod
def _prefilter_scenario(py_filename: str, scenario: str) -> bool:
"""Read a python file to see if scenarios require it to be run.
Much faster than parsing the file for a runtime check."""
(py_filename, _) = Runner._py_filename_adjust(py_filename, ".")
with open(py_filename, 'r', encoding="utf-8") as fh:
for line in fh:
m = re.search(r'^\s*test.scenarios\((.*?)\)', line)
if m:
for param in re.findall(r"""["']([^,]*)["']""", m.group(1)):
for allscarg in All_Scenarios[scenario]:
if param == allscarg:
return True
return False
def _prep(self) -> None:
VtOs.mkdir_ok(self.obj_dir) # Ok if already exists
def _read(self) -> None:
if not os.path.exists(self.py_filename):
self.error("Can't open " + self.py_filename)
return
global test
test = self
sys.path.append(self.t_dir) # To find vltest_bootstrap.py
# print("_read/exec py_filename=" + self.py_filename)
# print("_read/exec dir=", ' '.join(dir()))
# print("_read/exec vars=", ' '.join(vars().keys()))
# print("_read/exec globals=", ' '.join(globals().keys()))
# print("_read/exec locals=", ' '.join(locals().keys()))
try:
runpy.run_path(self.py_filename, globals())
except (VtErrorException, VtSkipException):
pass
def _exit(self):
if self.ok:
self.oprint("Self PASSED")
elif self._skips and not self.errors:
self.oprint("-Skip: " + self._skips)
else:
# If no worse errors, promote errors_keep_going to normal errors
if not self.errors and self.errors_keep_going:
self.errors = self.errors_keep_going
if not self.errors:
self.error("Missing ok")
self.oprint("%Error: " + self.errors)
self._write_status()
sys.exit(0)
def _write_status(self) -> None:
with open(self._status_filename, "wb") as fh:
pass_to_driver = {
'_ok': self._ok,
'_scenario_off': self._scenario_off,
'_skips': self._skips,
'errors': self.errors,
}
pickle.dump(pass_to_driver, fh)
def _read_status(self) -> None:
filename = self._status_filename
if not os.path.isfile(filename):
self.error_keep_going("Child test did not return status (test has Python error?): " +
self.py_filename)
return
with open(filename, "rb") as fh:
dic = pickle.load(fh)
for k in dic.keys():
# print("_read_status " + filename + ".test['" + k + "]=" + pformat(dic[k]))
setattr(self, k, dic[k])
#----------------------------------------------------------------------
# Methods invoked by tests
def clean(self) -> None:
"""Called on a rerun to cleanup files."""
if self.clean_command:
os.system(self.clean_command)
# Prevents false-failures when switching compilers
# Remove old results to force hard rebuild
os.system('/bin/rm -rf ' + self.obj_dir + '__fail1')
os.system('/bin/mv ' + self.obj_dir + ' ' + self.obj_dir + '__fail1')
def clean_objs(self) -> None:
os.system("/bin/rm -rf " + ' '.join(glob.glob(self.obj_dir + "/*")))
def _checkflags(self, param):
checkflags = (
' ' + ' '.join(param['v_flags']) + #
' ' + ' '.join(param['v_flags2']) + #
' ' + ' '.join(param['verilator_flags']) + #
' ' + ' '.join(param['verilator_flags2']) + #
' ' + ' '.join(param['verilator_flags3']) + ' ')
return checkflags
def compile_vlt_cmd(self, **kwargs) -> list:
"""Return default command list to run verilator"""
param = {'stdout_filename': None}
param.update(vars(self))
param.update(kwargs)
vlt_cmd = [
"perl", os.environ["VERILATOR_ROOT"] + "/bin/verilator",
*self._compile_vlt_flags(**param), param['top_filename'], *param['v_other_filenames']
]
if param['stdout_filename']:
vlt_cmd += ["> " + param['stdout_filename']]
return vlt_cmd
def _compile_vlt_flags(self, **param) -> list:
checkflags = self._checkflags(param)
d_verilator_flags = ' ' + ' '.join(self.driver_verilator_flags) + ' '
self.pins_sc_uint_bool = ( # pylint: disable=attribute-defined-outside-init
bool(
re.search(r'-pins-sc-uint-bool\b', checkflags)
or re.search(r'-pins-sc-uint-bool\b', d_verilator_flags)))
self.savable = ( # pylint: disable=attribute-defined-outside-init
bool(re.search(r'-savable\b', checkflags)))
self.coverage = ( # pylint: disable=attribute-defined-outside-init
bool(re.search(r'-coverage\b', checkflags)))
self.sanitize = param.get('sanitize', None) # pylint: disable=attribute-defined-outside-init
self.sc = ( # pylint: disable=attribute-defined-outside-init
bool(re.search(r'-sc\b', checkflags)))
self.timing = ( # pylint: disable=attribute-defined-outside-init
bool(re.search(r'( -?-timing\b| -?-binary\b)', checkflags)))
self.trace = ( # pylint: disable=attribute-defined-outside-init
bool(Args.trace or re.search(r'-trace\b|-trace-fst\b', checkflags)))
if re.search(r'-trace-fst', checkflags):
if self.sc:
self.trace_format = 'fst-sc' # pylint: disable=attribute-defined-outside-init
else:
self.trace_format = 'fst-c' # pylint: disable=attribute-defined-outside-init
elif self.sc:
self.trace_format = 'vcd-sc' # pylint: disable=attribute-defined-outside-init
else:
self.trace_format = 'vcd-c' # pylint: disable=attribute-defined-outside-init
if param.get('benchmarksim', None):
self.benchmarksim = True # pylint: disable=attribute-defined-outside-init
verilator_flags = [*param.get('verilator_flags', "")]
if Args.gdb:
verilator_flags += ["--gdb"]
if Args.gdbbt:
verilator_flags += ["--gdbbt"]
if Args.rr:
verilator_flags += ["--rr"]
if Args.trace:
verilator_flags += ["--trace"]
if Args.gdbsim or Args.rrsim:
verilator_flags += ["-CFLAGS -ggdb -LDFLAGS -ggdb"]
verilator_flags += ["--x-assign unique"] # More likely to be buggy
if param['vltmt']:
verilator_flags += ["--debug-partition"]
if param['threads'] >= 0:
verilator_flags += ["--threads", str(param['threads'])]
if param['vltmt'] and re.search(r'-trace-fst ', checkflags):
verilator_flags += ["--trace-threads 2"]
if self.sanitize:
verilator_flags += [
"-CFLAGS -fsanitize=address,undefined -LDFLAGS -fsanitize=address,undefined"
]
if param['verilator_make_cmake']:
verilator_flags += ["--make cmake"]
if param['verilator_make_gmake']:
verilator_flags += ["--make gmake"]
if param['make_main'] and param['verilator_make_gmake']:
verilator_flags += ["--exe"]
if param['make_main'] and param['verilator_make_gmake']:
verilator_flags += ["../" + self.main_filename]
cmdargs = [
"--prefix",
param['vm_prefix'],
*verilator_flags,
*param['verilator_flags2'],
*param['verilator_flags3'],
*param['v_flags'],
*param['v_flags2'],
# Flags from driver cmdline override default flags and
# flags from the test itself
*self.driver_verilator_flags,
]
return cmdargs
def lint(self, **kwargs) -> None:
"""Run a linter. Arguments similar to run(); default arguments are from self"""
param = {}
param.update(vars(self))
param.update({ # Lint-specific default overrides
'make_main': False,
'make_top_shell': False,
'verilator_flags2': ["--lint-only"],
'verilator_make_gmake': False
})
param.update(kwargs)
self.compile(**param)
def compile(self, **kwargs) -> None:
"""Run simulation compiler. Arguments similar to run(); default arguments are from self"""
param = {
'expect_filename': None,
'fails': False,
'make_flags': [],
'tee': True,
'timing_loop': False,
}
param.update(vars(self))
param.update(kwargs)
if self.verbose:
self.oprint("Compile")
if param['vlt'] and param['threads'] > 1:
self.error("'threads =' argument must be <= 1 for vlt scenario")
# Compute automatic parameter values
checkflags = self._checkflags(param)
if re.search(r'(^|\s)-?-threads\s', checkflags):
self.error("Specify threads via 'threads=' argument, not as a command line option")
if param['threads'] < 0 and param['vltmt']:
param['threads'] = calc_threads(Vltmt_Threads)
if not param['context_threads']:
param['context_threads'] = param['threads'] if (param['threads'] >= 1) else 1
if re.search(r'( -?-main\b| -?-binary\b)', checkflags):
param['make_main'] = False
if re.search(r'( -?-build\b| -?-binary\b)', checkflags):
param['verilator_make_cmake'] = False
param['verilator_make_gmake'] = False
self.threads = param['threads']
self.context_threads = param['context_threads']
self.compile_vlt_cmd(**param)
if not re.search(r'TEST_DUMPFILE', ' '.join(self.v_flags)):
self.v_flags += [self._define_opt_calc() + "TEST_DUMPFILE=" + self.trace_filename]
if not param['make_top_shell']:
self.top_shell_filename = ""
else:
self.top_shell_filename = self.obj_dir + "/" + self.vm_prefix + "__top." + self.v_suffix
param['top_shell_filename'] = self.top_shell_filename
if param['atsim']:
tool_define = param['atsim_define']
self._make_top(param['make_top_shell'])
self.run(logfile=self.obj_dir + "/atsim_compile.log",
fails=param['fails'],
cmd=[
VtOs.getenv_def('VERILATOR_ATSIM', "atsim"),
' '.join(param['atsim_flags']),
' '.join(param['atsim_flags2']),
' '.join(param['v_flags']),
' '.join(param['v_flags2']),
param['top_filename'],
param['top_shell_filename'],
' '.join(param['v_other_filenames']),
])
elif param['ghdl']:
tool_define = param['ghdl_define']
VtOs.mkdir_ok(self.ghdl_work_dir)
self._make_top(param['make_top_shell'])
tool_exe = VtOs.getenv_def('VERILATOR_GHDL', "ghdl")
self.run(
logfile=self.obj_dir + "/ghdl_compile.log",
fails=param['fails'],
cmd=[
tool_exe,
# Add -c here, as having -c twice freaks it out
("" if re.search(r' -c\b', tool_exe) else "-c"),
' '.join(param['ghdl_flags']),
' '.join(param['ghdl_flags2']),
#' '.join(param['v_flags']), # Not supported
#' '.join(param['v_flags2']), # Not supported
param['top_filename'],
param['top_shell_filename'],
' '.join(param['v_other_filenames']),
"-e t",
])
elif param['vcs']:
tool_define = param['vcs_define']
self._make_top(param['make_top_shell'])
self.run(logfile=self.obj_dir + "/vcs_compile.log",
fails=param['fails'],
cmd=[
VtOs.getenv_def('VERILATOR_VCS', "vcs"),
' '.join(param['vcs_flags']),
' '.join(param['vcs_flags2']),
("-CFLAGS -DTEST_VERBOSE=1" if Args.verbose else ""),
' '.join(param['v_flags']),
' '.join(param['v_flags2']),
param['top_filename'],
param['top_shell_filename'],
' '.join(param['v_other_filenames']),
])
elif param['nc']:
tool_define = param['nc_define']
self._make_top(param['make_top_shell'])
self.run(logfile=self.obj_dir + "/nc_compile.log",
fails=param['fails'],
cmd=[
VtOs.getenv_def('VERILATOR_NCVERILOG', "ncverilog"),
' '.join(param['nc_flags']),
' '.join(param['nc_flags2']),
' '.join(param['v_flags']),
' '.join(param['v_flags2']),
param['top_filename'],
param['top_shell_filename'],
' '.join(param['v_other_filenames']),
])
elif param['ms']:
tool_define = param['ms_define']
self._make_top(param['make_top_shell'])
self.run(logfile=self.obj_dir + "/ms_compile.log",
fails=param['fails'],
cmd=[
("vlib " + self.obj_dir + "/work && "),
VtOs.getenv_def('VERILATOR_MODELSIM', "vlog"),
' '.join(param['ms_flags']),
' '.join(param['ms_flags2']),
' '.join(param['v_flags']),
' '.join(param['v_flags2']),
param['top_filename'],
param['top_shell_filename'],
])
elif param['iv']:
tool_define = param['iv_define']
self._make_top(param['make_top_shell'])
cmd = (VtOs.getenv_def('VERILATOR_IVERILOG', "iverilog"), ' '.join(param['iv_flags']),
' '.join(param['iv_flags2']), ' '.join(param['v_flags']),
' '.join(param['v_flags2']), param['top_filename'], param['top_shell_filename'],
' '.join(param['v_other_filenames']))
cmd = list(map(lambda str: re.sub(r'\+define\+', '-D ', str), cmd))
self.run(logfile=self.obj_dir + "/iv_compile.log", fails=param['fails'], cmd=cmd)
elif param['xrun']:
tool_define = param['xrun_define']
self._make_top(param['make_top_shell'])
elif param['xsim']:
tool_define = param['xsim_define']
self._make_top(param['make_top_shell'])
self.run(logfile=self.obj_dir + "/xsim_compile.log",
fails=param['fails'],
cmd=[
VtOs.getenv_def('VERILATOR_XVLOG', "xvlog"),
' '.join(param['xsim_flags']),
' '.join(param['xsim_flags2']),
' '.join(param['v_flags']),
' '.join(param['v_flags2']),
param['top_filename'],
param['top_shell_filename'],
])
elif param['vlt_all']:
tool_define = param['verilator_define']
if self.sc and not self.have_sc:
self.skip("Test requires SystemC; ignore error since not installed\n")
return
if self.timing and not self.have_coroutines:
self.skip("Test requires Coroutines; ignore error since not available\n")
return
if param['verilator_make_cmake'] and not self.have_cmake:
self.skip(
"Test requires CMake; ignore error since not available or version too old\n")
return
if not param['fails'] and param['make_main']:
self._make_main(param['timing_loop'])
if (param['verilator_make_gmake']
or (not param['verilator_make_gmake'] and not param['verilator_make_cmake'])):
vlt_cmd = self.compile_vlt_cmd(**param)
if self.verbose:
self.oprint("Running Verilator (gmake)")
if Args.verilation:
self.run(logfile=self.obj_dir + "/vlt_compile.log",
fails=param['fails'],
tee=param['tee'],
expect_filename=param['expect_filename'],
verilator_run=True,
cmd=vlt_cmd)
if param['verilator_make_cmake']:
vlt_args = self._compile_vlt_flags(**param)
if self.verbose:
self.oprint("Running cmake")
VtOs.mkdir_ok(self.obj_dir)
csources = []
if param['make_main']:
csources.append(self.main_filename)
self.run(
logfile=self.obj_dir + "/vlt_cmake.log",
fails=param['fails'],
tee=param['tee'],
expect_filename=param['expect_filename'],
verilator_run=True,
cmd=[
"cd \"" + self.obj_dir + "\" && cmake",
"\"" + self.t_dir + "/..\"",
"-DTEST_VERILATOR_ROOT=" + os.environ['VERILATOR_ROOT'],
"-DTEST_NAME=" + self.name,
"-DTEST_CSOURCES=\"" + ' '.join(csources) + "\"",
"-DTEST_VERILATOR_ARGS=\"" + ' '.join(vlt_args) + "\"",
"-DTEST_VERILATOR_SOURCES=\"" + param['top_filename'] + ' ' +
' '.join(param['v_other_filenames']) + "\"",
"-DTEST_VERBOSE=\"" + ("1" if self.verbose else "0") + "\"",
"-DTEST_SYSTEMC=\"" + ("1" if self.sc else "0") + "\"",
"-DCMAKE_PREFIX_PATH=\"" +
(VtOs.getenv_def('SYSTEMC_INCLUDE', VtOs.getenv_def('SYSTEMC', '')) +
"/..\""),
"-DTEST_OPT_FAST=\"" + ("-Os" if param['benchmark'] else "-O0") + "\"",
"-DTEST_OPT_GLOBAL=\"" + ("-Os" if param['benchmark'] else "-O0") + "\"",
"-DTEST_VERILATION=\"" + ("1" if Args.verilation else "0") + "\"",
])
if not param['fails'] and param['verilator_make_gmake']:
if self.verbose:
self.oprint("Running make (gmake)")
self.run(
logfile=self.obj_dir + "/vlt_gcc.log",
entering=self.obj_dir,
cmd=[
os.environ['MAKE'],
"-C " + self.obj_dir,
"-f " + os.path.abspath(os.path.dirname(__file__)) + "/Makefile_obj",
("" if self.verbose else "--no-print-directory"),
"VM_PREFIX=" + self.vm_prefix,
"TEST_OBJ_DIR=" + self.obj_dir,
"CPPFLAGS_DRIVER=-D" + self.name.upper(),
("CPPFLAGS_DRIVER2=-DTEST_VERBOSE=1" if self.verbose else ""),
("" if param['benchmark'] else "OPT_FAST=-O0"),
("" if param['benchmark'] else "OPT_GLOBAL=-O0"),
self.vm_prefix, # bypass default rule, as we don't need archive
*param['make_flags'],
])
if not param['fails'] and param['verilator_make_cmake']:
if self.verbose:
self.oprint("Running cmake --build")
self.run(logfile=self.obj_dir + "/vlt_cmake_build.log",
cmd=[
"cmake",
"--build",
self.obj_dir,
("--verbose" if self.verbose else ""),
])
else:
self.error("No compile step defined for '%s' scenario" % self.scenario())
if param['make_pli']:
if self.verbose:
self.oprint("Compile vpi")
cmd = [
os.environ['CXX'], *param['pli_flags'], "-D" + tool_define, "-DIS_VPI",
VtOs.getenv_def('CFLAGS', ''), self.pli_filename
]
self.run(logfile=self.obj_dir + "/pli_compile.log", fails=param['fails'], cmd=cmd)
def execute(self, **kwargs) -> None:
"""Run simulation executable.
Arguments similar to run(); default arguments are from self"""
# Default arguments are from self
# params may be expect or {tool}_expect
param = {
'aslr_off': False,
'entering': False,
'check_finished': False,
'executable': None,
'expect_filename': None,
'fails': False,
'run_env': '',
'tee': True,
}
param.update(vars(self))
param.update(kwargs)
if self.verbose:
self.oprint("Run")
if not self.verbose:
os.environ['SYSTEMC_DISABLE_COPYRIGHT_MESSAGE'] = 'DISABLE'
else:
VtOs.delenv('SYSTEMC_DISABLE_COPYRIGHT_MESSAGE')
if not self._have_solver_called:
os.environ['VERILATOR_SOLVER'] = "test.py-file-needs-have_solver()-call"
if param['check_finished'] is None and not param['fails']:
param['check_finished'] = 1
run_env = param['run_env']
if run_env:
run_env = run_env + ' '
if param['atsim']:
self.run(
logfile=self.obj_dir + "/atsim_sim.log",
fails=param['fails'],
cmd=[
"echo q | " + run_env + self.obj_dir + "/athdl_sv",
' '.join(param['atsim_run_flags']),
' '.join(param['all_run_flags']),
],
*param,
expect_filename=param['atsim_run_expect_filename'],
)
elif param['ghdl']:
self.run(
logfile=self.obj_dir + "/ghdl_sim.log",
fails=param['fails'],
cmd=[
run_env + self.obj_dir + "/simghdl",
' '.join(param['ghdl_run_flags']),
' '.join(param['all_run_flags']),
],
*param,
expect_filename=param['ghdl_run_expect_filename'],
)
elif param['iv']:
cmd = [
run_env + self.obj_dir + "/simiv",
' '.join(param['iv_run_flags']),
' '.join(param['all_run_flags']),
]
if param['use_libvpi']:
# Don't enter command line on $stop, include vpi
cmd += ["vvp -n -m " + self.obj_dir + "/libvpi.so"]
self.run(
logfile=self.obj_dir + "/iv_sim.log",
fails=param['fails'],
cmd=cmd,
*param,
expect_filename=param['iv_run_expect_filename'],
)
elif param['ms']:
pli_opt = ""
if param['use_libvpi']:
pli_opt = "-pli " + self.obj_dir + "/libvpi.so"
self.run(
logfile=self.obj_dir + "/ms_sim.log",
fails=param['fails'],
cmd=[
"echo q | " + run_env + VtOs.getenv_def('VERILATOR_MODELSIM', "vsim"),
' '.join(param['ms_run_flags']), ' '.join(param['all_run_flags']), pli_opt,
(" top")
],
*param,
expect_filename=param['ms_expect_filename'],
)
elif param['nc']:
self.run(
logfile=self.obj_dir + "/nc_sim.log",
fails=param['fails'],
cmd=[
"echo q | " + run_env + VtOs.getenv_def('VERILATOR_NCVERILOG', "ncverilog"),
' '.join(param['nc_run_flags']),
' '.join(param['all_run_flags']),
],
*param,
expect_filename=param['nc_run_expect_filename'],
)
elif param['vcs']:
# my $fh = IO::File->new(">simv.key") or die "%Error: $! simv.key,"
# fh.print("quit\n"); fh.close()
self.run(
logfile=self.obj_dir + "/vcs_sim.log",
cmd=[
"echo q | " + run_env + "./simv",
' '.join(param['vcs_run_flags']),
' '.join(param['all_run_flags']),
],
*param,
expect_filename=param['vcs_run_expect_filename'],
)
elif param['xrun']:
pli_opt = ""
if param['use_libvpi']:
pli_opt = "-loadvpi " + self.obj_dir + "/libvpi.so:vpi_compat_bootstrap"
self.run(
logfile=self.obj_dir + "/xrun_sim.log",
fails=param['fails'],
cmd=[
"echo q | " + run_env + VtOs.getenv_def('VERILATOR_XRUN', "xrun"),
' '.join(param['xrun_run_flags']),
' '.join(param['xrun_flags2']),
' '.join(param['all_run_flags']),
pli_opt,
param['top_filename'],
],
*param,
expect_filename=param['xrun_run_expect_filename'],
)
elif param['xsim']:
self.run(
logfile=self.obj_dir + "/xsim_sim.log",
fails=param['fails'],
cmd=[
run_env + VtOs.getenv_def('VERILATOR_XELAB', "xelab"),
' '.join(param['xsim_run_flags']), ' '.join(param['xsim_run_flags2']),
' '.join(param['all_run_flags']), (" " + self.name + ".top")
],
*param,
expect_filename=param['xsim_expect_filename'],
)
elif param['vlt_all']:
if not param['executable']:
param['executable'] = self.obj_dir + "/" + param['vm_prefix']
debugger = ""
if Args.gdbsim:
debugger = VtOs.getenv_def('VERILATOR_GDB', "gdb") + " "
elif Args.rrsim:
debugger = "rr record "
self.run(
cmd=[
(run_env + debugger + param['executable'] +
(" -ex 'run " if Args.gdbsim else "")),
*param['all_run_flags'],
("'" if Args.gdbsim else ""),
],
aslr_off=param['aslr_off'], # Disable address space layour randomization
check_finished=param['check_finished'], # Check for All Finished
entering=param['entering'], # Print entering directory information
expect_filename=param['expect_filename'],
fails=param['fails'],
logfile=param.get('logfile', self.obj_dir + "/vlt_sim.log"),
tee=param['tee'],
verilator_run=True,
)
else:
self.error("No execute step for this simulator")
#---------------------------------------------------------------
# Accessors
@property
def aslr_off(self) -> str:
if VlTest._cached_aslr_off is None:
out = VtOs.run_capture('setarch --addr-no-randomize echo OK 2>/dev/null', check=False)
if re.search(r'OK', out):
VlTest._cached_aslr_off = "setarch --addr-no-randomize "
else:
VlTest._cached_aslr_off = ""
return VlTest._cached_aslr_off
@property
def benchmarksim_filename(self) -> str:
return self.obj_dir + "/" + self.name + "_benchmarksim.csv"
@property
def driver_verilator_flags(self) -> list:
return Args.passdown_verilator_flags
@property
def get_default_vltmt_threads(self) -> int:
return Vltmt_Threads
@property
def ok(self) -> bool:
if self.errors or self.errors_keep_going or self._skips:
self._ok = False
return self._ok
def passes(self, is_ok=True):
if not self.errors:
self._ok = is_ok
@property
def too_few_cores(self) -> bool:
return calc_threads(Vltmt_Threads) < Vltmt_Threads
@property
def trace_filename(self) -> str:
if re.match(r'^fst', self.trace_format):
return self.obj_dir + "/simx.fst"
return self.obj_dir + "/simx.vcd"
def skip_if_too_few_cores(self) -> None:
if self.too_few_cores:
self.skip("Skipping due to too few cores")
@property
def v_suffix(self) -> str:
return "v"
@property
def wno_unopthreads_for_few_cores(self) -> str:
if self.too_few_cores:
print("Too few cores, using -Wno-UNOPTTHREADS")
return "-Wno-UNOPTTHREADS"
return ""
#---------------------------------------------------------------
# Capabilities
@property
def cmake_version(self) -> str:
return Capabilities.cmake_version
@property
def cxx_version(self) -> str:
return Capabilities.cxx_version
@property
def have_cmake(self) -> bool:
ver = Capabilities.cmake_version
return ver and version.parse(ver) >= version.parse("3.8")
@property
def have_coroutines(self) -> bool:
return Capabilities.have_coroutines
@property
def have_gdb(self) -> bool:
return Capabilities.have_gdb
@property
def have_sc(self) -> bool:
return Capabilities.have_sc
@property
def have_solver(self) -> bool:
self._have_solver_called = True
return Capabilities.have_solver
@property
def make_version(self) -> str:
return Capabilities.make_version
#---------------------------------------------------------------
# OS functions
def getenv_def(self, var: str, default=None) -> str:
"""Return environment variable, returning default if does not exist"""
return VtOs.getenv_def(var, default)
def mkdir_ok(self, filename) -> None:
"""Make directory, no error if exists"""
if test.verbose:
print("\tmkdir " + filename)
VtOs.mkdir_ok(filename)
def run_capture(self, cmd: str, check=True) -> str:
"""Run a command and return results"""
if test.verbose:
print("\t" + cmd)
return VtOs.run_capture(cmd, check=check)
def setenv(self, var: str, val: str) -> None:
"""Set enviornment variable"""
print("\texport %s='%s'" % (var, val))
os.environ[var] = val
def unlink_ok(self, filename) -> None:
"""Unlink a file, no error if fails"""
if test.verbose:
print("\trm " + filename)
VtOs.unlink_ok(filename)
#----------------------------------------------------------------------
def run(
self, #
cmd: list,
aslr_off=False, # Disable address space layour randomization
check_finished=False, # Check for All Finished
entering=None, # Print entering directory information
expect_filename=None, # Filename that should match logfile
fails=False, # Command should fail
logfile=None, # Filename to write putput to
tee=True,
verilator_run=False) -> str: # Move gcov data to parallel area
try:
command = ' '.join(cmd)
except TypeError:
print('run(cmd=' + pformat(cmd))
command = ' '.join(cmd)
if aslr_off and aslr_off != "":
prefix = self.aslr_off
if prefix:
command = prefix + " " + command
if Args.benchmark and re.match(r'^cd ', command):
command = "time " + command
if verilator_run:
# Gcov fails when parallel jobs write same data file,
# so we make sure .gcda output dir is unique across all running jobs.
# We can't just put each one in an unique obj_dir as it uses too much disk.
# Must use absolute path as some execute()s have different PWD
self.setenv('GCOV_PREFIX_STRIP', '99')
self.setenv('GCOV_PREFIX',
os.path.abspath(__file__ + "/../obj_dist/gcov_" + str(self.running_id)))
os.makedirs(os.environ['GCOV_PREFIX'], exist_ok=True)
else:
VtOs.delenv('GCOV_PREFIX_STRIP')
VtOs.delenv('GCOV_PREFIX')
print("\t" + command + ((" > " + logfile) if logfile else ""))
if entering:
print("driver: Entering directory '" + os.path.abspath(entering) + "'")
# Execute command redirecting output, keeping order between stderr and stdout.
# Must do low-level IO so GCC interaction works (can't be line-based)
status = None
if True: # process_caller_block # pylint: disable=using-constant-test
logfh = None
if logfile:
logfh = open(logfile, 'wb') # pylint: disable=consider-using-with
with subprocess.Popen(command,
shell=True,
bufsize=0,
stdout=subprocess.PIPE,
stderr=subprocess.STDOUT) as proc:
rawbuf = bytearray(2048)
while True:
finished = proc.poll()
# Need to check readinto once, even after poll "completes"
got = proc.stdout.readinto(rawbuf)
if got:
data = rawbuf[0:got]
if tee:
sys.stdout.write(data.decode('latin-1'))
if Args.interactive_debugger:
sys.stdout.flush()
if logfh:
logfh.write(data)
elif finished is not None:
break
if logfh:
logfh.close()
rc = proc.returncode # Negative if killed by signal
if (rc in (
-4, # SIGILL
-8, # SIGFPA
-11)): # SIGSEGV
self.error("Exec failed with core dump")
status = 10
elif rc:
status = 10
else:
status = 0
sys.stdout.flush()
sys.stderr.flush()
if entering:
print("driver: Leaving directory '" + os.path.abspath(entering) + "'")
if not fails and status:
firstline = ""
if logfile:
with open(logfile, 'r', encoding="utf8") as fh:
for line in fh:
line = line.rstrip()
if re.match(r'^- ', line): # Debug message
continue
firstline = line
break
self.error("Exec of " + cmd[0] + " failed: " + firstline)
if fails and status:
print("(Exec expected to fail, and did.)")
if fails and not status:
self.error("Exec of " + cmd[0] + " ok, but expected to fail")
if self.errors or self._skips:
return False
# Read the log file a couple of times to allow for NFS delays
if check_finished:
for tryn in range(Args.log_retries - 1, -1, -1):
if tryn != Args.log_retries - 1:
time.sleep(1)
moretry = tryn != 0
if not self._run_log_try(logfile, check_finished, moretry):
break
if expect_filename:
self.files_identical(logfile, expect_filename, is_logfile=True)
return False
return True
def _run_log_try(self, logfile: str, check_finished: bool, moretry: bool) -> bool:
# If moretry, then return true to try again
with open(logfile, 'r', encoding='latin-1', newline='\n') as fh:
if not fh and moretry:
return True
wholefile = str(fh.read())
# Finished?
if check_finished and not re.search(r'\*\-\* All Finished \*\-\*', wholefile):
if moretry:
return True
self.error("Missing '*-* All Finished *-*'")
return False
#######################################################################
# Little utilities
@staticmethod
def _try_regex(text: str, regex) -> None:
# Try to eval a regexp
# Returns:
# 1 if $text ~= /$regex/ms
# 0 if no match
# -1 if $regex is invalid, doesn't compile
try:
m = re.search(regex, text)
return 1 if m else 0
except re.error:
return -1
def _make_main(self, timing_loop: bool) -> None:
if timing_loop and self.sc:
self.error("Cannot use timing loop and SystemC together!")
self._read_inputs_v()
filename = self.main_filename
with open(filename, "w", encoding="utf8") as fh:
fh.write("// Test defines\n")
fh.write("#define MAIN_TIME_MULTIPLIER " +
str(int(round(self.main_time_multiplier, 0))) + "\n")
fh.write("#include <memory>\n")
if self.benchmarksim:
fh.write("#include <fstream>\n")
fh.write("#include <chrono>\n")
fh.write("#include <iomanip>\n")
fh.write("// OS header\n")
fh.write('#include "verilatedos.h"' + "\n")
fh.write("// Generated header\n")
fh.write('#include "' + self.vm_prefix + '.h"' + "\n")
fh.write("// General headers\n")
fh.write('#include "verilated.h"' + "\n")
if self.sc:
fh.write('#include "systemc.h"' + "\n")
if self.trace and self.trace_format == 'fst-c':
fh.write("#include \"verilated_fst_c.h\"\n")
if self.trace and self.trace_format == 'fst-sc':
fh.write("#include \"verilated_fst_sc.h\"\n")
if self.trace and self.trace_format == 'vcd-c':
fh.write("#include \"verilated_vcd_c.h\"\n")
if self.trace and self.trace_format == 'vcd-sc':
fh.write("#include \"verilated_vcd_sc.h\"\n")
if self.savable:
fh.write("#include \"verilated_save.h\"\n")
fh.write("std::unique_ptr<" + self.vm_prefix + "> topp;\n")
if self.savable:
fh.write("\n")
fh.write("void save_model(const char* filenamep) {\n")
fh.write(" VL_PRINTF(\"Saving model to '%s'\\n\", filenamep);\n")
fh.write(" VerilatedSave os;\n")
fh.write(" os.open(filenamep);\n")
fh.write(" os << *topp;\n")
fh.write(" os.close();\n")
fh.write("}\n")
fh.write("\n")
fh.write("void restore_model(const char* filenamep) {\n")
fh.write(" VL_PRINTF(\"Restoring model from '%s'\\n\", filenamep);\n")
fh.write(" VerilatedRestore os;\n")
fh.write(" os.open(filenamep);\n")
fh.write(" os >> *topp;\n")
fh.write(" os.close();\n")
fh.write("}\n")
#### Main
if self.sc:
fh.write("extern int sc_main(int argc, char** argv);\n")
fh.write("int sc_main(int argc, char** argv) {\n")
if 'fastclk' in self._inputs:
if self.pins_sc_uint_bool:
fh.write(" sc_signal<sc_dt::sc_uint<1>> fastclk;\n")
else:
fh.write(" sc_signal<bool> fastclk;\n")
if 'clk' in self._inputs:
if self.pins_sc_uint_bool:
fh.write(" sc_signal<sc_dt::sc_uint<1>> clk;\n")
else:
fh.write(" sc_signal<bool> clk;\n")
fh.write(" sc_set_time_resolution(1, " + self.sc_time_resolution + ");\n")
fh.write(" sc_time sim_time(" + str(self.sim_time) + ", " +
self.sc_time_resolution + ");\n")
else:
fh.write("int main(int argc, char** argv) {\n")
fh.write(" uint64_t sim_time = " + str(self.sim_time) + ";\n")
fh.write(
" const std::unique_ptr<VerilatedContext> contextp{new VerilatedContext};\n")
fh.write(" contextp->threads(" + str(self.context_threads) + ");\n")
fh.write(" contextp->commandArgs(argc, argv);\n")
fh.write(" contextp->debug(" + ('1' if self.verilated_debug else '0') + ");\n")
fh.write(" srand48(5);\n") # Ensure determinism
if self.verilated_randReset is not None and self.verilated_randReset != "":
fh.write(" contextp->randReset(" + str(self.verilated_randReset) + ");\n")
fh.write(" topp.reset(new " + self.vm_prefix + "{\"top\"});\n")
if self.verilated_debug:
fh.write(" contextp->internalsDump()\n;")
if self.sc:
if 'fastclk' in self._inputs:
fh.write(" topp->fastclk(fastclk);\n")
if 'clk' in self._inputs:
fh.write(" topp->clk(clk);\n")
setp = ""
else:
fh.write(" topp->eval();\n")
setp = "topp->"
if self.benchmarksim:
fh.write(" std::chrono::time_point<std::chrono::steady_clock> starttime;\n")
fh.write(" bool warm = false;\n")
fh.write(" uint64_t n_evals = 0;\n")
if self.trace:
fh.write("\n")
fh.write("#if VM_TRACE\n")
fh.write(" contextp->traceEverOn(true);\n")
if self.trace_format == 'fst-c':
fh.write(" std::unique_ptr<VerilatedFstC> tfp{new VerilatedFstC};\n")
if self.trace_format == 'fst-sc':
fh.write(" std::unique_ptr<VerilatedFstSc> tfp{new VerilatedFstSc};\n")
if self.trace_format == 'vcd-c':
fh.write(" std::unique_ptr<VerilatedVcdC> tfp{new VerilatedVcdC};\n")
if self.trace_format == 'vcd-sc':
fh.write(" std::unique_ptr<VerilatedVcdSc> tfp{new VerilatedVcdSc};\n")
if self.sc:
fh.write(" sc_core::sc_start(sc_core::SC_ZERO_TIME);" +
" // Finish elaboration before trace and open\n")
fh.write(" topp->trace(tfp.get(), 99);\n")
fh.write(" tfp->open(\"" + self.trace_filename + "\");\n")
if self.trace and not self.sc:
fh.write(" if (tfp) tfp->dump(contextp->time());\n")
fh.write("#endif\n")
if self.savable:
fh.write(" const char* save_time_strp"
" = contextp->commandArgsPlusMatch(\"save_time=\");\n")
fh.write(" unsigned int save_time = !save_time_strp[0]"
" ? 0 : std::atoi(save_time_strp + std::strlen(\"+save_time=\"));\n")
fh.write(" const char* save_restore_strp"
" = contextp->commandArgsPlusMatch(\"save_restore=\");\n")
fh.write(" unsigned int save_restore = !save_restore_strp[0] ? 0 : 1;\n")
if self.savable:
fh.write(" if (save_restore) {\n")
fh.write(" restore_model(\"" + self.obj_dir + "/saved.vltsv\");\n")
fh.write(" } else {\n")
else:
fh.write(" {\n")
if 'fastclk' in self._inputs:
fh.write(" " + setp + "fastclk = false;\n")
if 'clk' in self._inputs:
fh.write(" " + setp + "clk = false;\n")
if not timing_loop:
self._print_advance_time(fh, 10, None)
fh.write(" }\n")
timestamp = "sc_time_stamp()" if self.sc else "contextp->time()"
fh.write(" while (")
if not timing_loop or 'clk' in self._inputs:
fh.write("(" + timestamp + " < sim_time * MAIN_TIME_MULTIPLIER) && ")
fh.write("!contextp->gotFinish()) {\n")
if timing_loop:
fh.write(" topp->eval();\n")
if self.trace:
fh.write("#if VM_TRACE\n")
fh.write(" if (tfp) tfp->dump(contextp->time());\n")
fh.write("#endif // VM_TRACE\n")
if 'clk' in self._inputs:
fh.write(" const uint64_t cycles"
" = contextp->time() / MAIN_TIME_MULTIPLIER;\n")
fh.write(" uint64_t new_time = (cycles + 1) * MAIN_TIME_MULTIPLIER;\n")
fh.write(" if (topp->eventsPending() &&\n")
fh.write(" topp->nextTimeSlot()"
" / MAIN_TIME_MULTIPLIER <= cycles) {\n")
fh.write(" new_time = topp->nextTimeSlot();\n")
fh.write(" } else {\n")
if self.pins_sc_uint_bool:
fh.write(" " + setp + "clk.write(!" + setp + "clk.read());\n")
else:
fh.write(" " + setp + "clk = !" + setp + "clk;\n")
fh.write(" }\n")
fh.write(" contextp->time(new_time);\n")
else:
fh.write(" if (!topp->eventsPending()) break;\n")
fh.write(" contextp->time(topp->nextTimeSlot());\n")
else:
for i in range(5):
action = False
if 'fastclk' in self._inputs:
if self.pins_sc_uint_bool:
fh.write(" " + setp + "fastclk.write(!" + setp +
"fastclk.read());\n")
else:
fh.write(" " + setp + "fastclk = !" + setp + "fastclk;\n")
action = True
if i == 0 and 'clk' in self._inputs:
if self.pins_sc_uint_bool:
fh.write(" " + setp + "clk.write(!" + setp + "clk.read());\n")
else:
fh.write(" " + setp + "clk = !" + setp + "clk;\n")
action = True
if self.savable:
fh.write(" if (save_time && " + timestamp + " == save_time) {\n")
fh.write(" save_model(\"" + self.obj_dir + "/saved.vltsv\");\n")
fh.write(" printf(\"Exiting after save_model\\n\");\n")
fh.write(" topp.reset(nullptr);\n")
fh.write(" return 0;\n")
fh.write(" }\n")
self._print_advance_time(fh, 1, action)
if self.benchmarksim:
fh.write(" if (VL_UNLIKELY(!warm)) {\n")
fh.write(" starttime = std::chrono::steady_clock::now();\n")
fh.write(" warm = true;\n")
fh.write(" } else {\n")
fh.write(" ++n_evals;\n")
fh.write(" }\n")
fh.write(" }\n")
if self.benchmarksim:
fh.write(" {\n")
fh.write(" const std::chrono::duration<double> exec_s"
" = std::chrono::steady_clock::now() - starttime;\n")
fh.write(" std::ofstream benchfile(\"" + self.benchmarksim_filename +
"\", std::ofstream::out | std::ofstream::app);\n")
fh.write(" benchfile << std::fixed << std::setprecision(9)"
" << n_evals << \",\" << exec_s.count() << std::endl;\n")
fh.write(" benchfile.close();\n")
fh.write(" }\n")
fh.write(" if (!contextp->gotFinish()) {\n")
fh.write(' vl_fatal(__FILE__, __LINE__, "main",' +
' "%Error: Timeout; never got a $finish");' + "\n")
fh.write(" }\n")
fh.write(" topp->final();\n")
fh.write("\n")
if self.coverage:
fh.write("#if VM_COVERAGE\n")
fh.write(" contextp->coveragep()->write(\"" + self.coverage_filename + "\");\n")
fh.write("#endif // VM_COVERAGE\n")
if self.trace:
fh.write("#if VM_TRACE\n")
fh.write(" if (tfp) tfp->close();\n")
fh.write(" tfp.reset();\n")
fh.write("#endif // VM_TRACE\n")
fh.write(" topp.reset();\n")
fh.write(" return 0;\n")
fh.write("}\n")
def _print_advance_time(self, fh, timeinc: str, action: bool) -> None:
setp = "" if self.sc else "topp->"
if self.sc:
fh.write(" sc_start(" + str(timeinc) + " * MAIN_TIME_MULTIPLIER, " +
self.sc_time_resolution + ");\n")
else:
if action:
fh.write(" " + setp + "eval();\n")
if self.trace and not self.sc:
fh.write("#if VM_TRACE\n")
fh.write(" if (tfp) tfp->dump(contextp->time());\n")
fh.write("#endif // VM_TRACE\n")
fh.write(" contextp->timeInc(" + str(timeinc) + " * MAIN_TIME_MULTIPLIER);\n")
#######################################################################
def _make_top(self, needed=True) -> None:
if not needed:
return
self._make_top_v()
def _make_top_v(self) -> None:
self._read_inputs_v()
with open(self.top_shell_filename(), 'w', encoding="utf8") as fh:
fh.write("module top;\n")
for inp in sorted(self._inputs.keys()):
fh.write(" reg " + inp + ";\n")
# Inst
fh.write(" t t (\n")
comma = ""
for inp in sorted(self._inputs.keys()):
fh.write(" " + comma + "." + inp + " (" + inp + ")\n")
comma = ","
fh.write(" );\n")
# Waves
fh.write("\n")
fh.write("`ifdef WAVES\n")
fh.write(" initial begin\n")
fh.write(" $display(\"-Tracing Waves to Dumpfile: " + self.trace_filename +
"\");\n")
fh.write(" $dumpfile(\"" + self.trace_filename + "\");\n")
fh.write(" $dumpvars(0, top);\n")
fh.write(" end\n")
fh.write("`endif\n")
# Test
fh.write("\n")
fh.write(" initial begin\n")
if 'fastclk' in self._inputs:
fh.write(" fastclk = 0;\n")
if 'clk' in self._inputs:
fh.write(" clk = 0;\n")
fh.write(" #10;\n")
if 'fastclk' in self._inputs:
fh.write(" fastclk = 1;\n")
if 'clk' in self._inputs:
fh.write(" clk = 1;\n")
fh.write(" while (" + time + " < " + self.sim_time + ") begin\n")
for i in range(6):
fh.write(" #1;\n")
if 'fastclk' in self._inputs:
fh.write(" fastclk = !fastclk;\n")
if i == 4 and 'clk' in self._inputs:
fh.write(" clk = !clk;\n")
fh.write(" end\n")
fh.write(" end\n")
fh.write("endmodule\n")
#######################################################################
def _read_inputs_v(self) -> None:
filename = self.top_filename
if not os.path.exists(filename):
filename = self.t_dir + '/' + filename
with open(filename, 'r', encoding="utf8") as fh:
get_sigs = True
inputs = {}
for line in fh:
if get_sigs:
m = re.match(r'^\s*input\s*(\S+)\s*(\/[^\/]+\/|)\s*;', line)
if m:
inputs[m.group(1)] = m.group(1)
if re.match(r'^\s*(function|task|endmodule)', line):
get_sigs = False
# Ignore any earlier inputs; Module 't' has precedence
if re.match(r'^\s*module\s+t\b', line):
inputs = {}
get_sigs = True
for sig, val in inputs.items():
self._inputs[sig] = val
#######################################################################
# File utilities
def files_identical(self, fn1: str, fn2: str, is_logfile=False) -> None:
"""Test if two files have identical contents"""
for tryn in range(Args.log_retries, -1, -1):
if tryn != Args.log_retries - 1:
time.sleep(1)
moretry = tryn != 0
if not self._files_identical_try(
fn1=fn1, fn2=fn2, is_logfile=is_logfile, moretry=moretry):
break
def _files_identical_try(self, fn1: str, fn2: str, is_logfile: bool, moretry: bool) -> bool:
# If moretry, then return true to try again
try:
f1 = open( # pylint: disable=consider-using-with
fn1, 'r', encoding='latin-1', newline='\n')
except FileNotFoundError:
f1 = None
if not moretry:
self.error("Files_identical file does not exist: " + fn1)
return True # Retry
try:
f2 = open( # pylint: disable=consider-using-with
fn2, 'r', encoding='latin-1', newline='\n')
except FileNotFoundError:
f2 = None
if not moretry:
self.copy_if_golden(fn1, fn2)
self.error("Files_identical file does not exist: " + fn2)
return True # Retry
again = self._files_identical_reader(f1,
f2,
fn1=fn1,
fn2=fn2,
is_logfile=is_logfile,
moretry=moretry)
if f1:
f1.close()
if f2:
f2.close()
return again
def _files_identical_reader(self, f1, f2, fn1: str, fn2: str, is_logfile: bool,
moretry: bool) -> None:
# If moretry, then return true to try again
l1s = f1.readlines()
l2s = f2.readlines() if f2 else []
# print(" rawGOT="+pformat(l1s)+"\n rawEXP="+pformat(l2s))
if is_logfile:
l1o = []
for line in l1s:
if (re.match(r'^- [^\n]+\n', line) #
or re.match(r'^- [a-z.0-9]+:\d+:[^\n]+\n', line)
or re.match(r'^-node:', line) #
or re.match(r'^dot [^\n]+\n', line) #
or re.match(r'^Aborted', line) #
or re.match(r'^In file: .*\/sc_.*:\d+', line) #
or re.match(r'^libgcov.*', line) #
or re.match(r'--- \/tmp\/', line) # t_difftree.py
or re.match(r'\+\+\+ \/tmp\/', line) # t_difftree.py
or re.match(r'^==[0-9]+== ?[^\n]*\n', line)): # valgrind
continue
# Don't put control chars or unstable lines into source repository
while True:
(line, didn) = re.subn(r'(Internal Error: [^\n]+?\.(cpp|h)):[0-9]+', r'\1:#',
line)
if not didn:
break
# --vlt vs --vltmt run differences
line = re.sub(r'^-V\{t[0-9]+,[0-9]+\}', '-V{t#,#}', line)
line = re.sub(r'\r', '<#013>', line)
line = re.sub(r'Command Failed[^\n]+', 'Command Failed', line)
line = re.sub(r'Version: Verilator[^\n]+', 'Version: Verilator ###', line)
line = re.sub(r'CPU Time: +[0-9.]+ seconds[^\n]+', 'CPU Time: ###', line)
line = re.sub(r'\?v=[0-9.]+', '?v=latest', line) # warning URL
line = re.sub(r'_h[0-9a-f]{8}_', '_h########_', line)
line = re.sub(r'%Error: /[^: ]+/([^/:])', r'%Error: .../\1',
line) # Avoid absolute paths
line = re.sub(r' \/[^ ]+\/verilated_std.sv', ' verilated_std.sv', line)
#
(line, n) = re.subn(r'Exiting due to.*', r"Exiting due to", line)
if n:
l1o.append(line)
break # Trunc rest
l1o.append(line)
#
l1s = l1o
for lineno_m1 in range(0, max(len(l1s), len(l2s))):
l1 = l1s[lineno_m1] if lineno_m1 < len(l1s) else "*EOF*\n"
l2 = l2s[lineno_m1] if lineno_m1 < len(l2s) else "*EOF*\n"
if l1 != l2:
# print(" clnGOT="+pformat(l1s)+"\n clnEXP="+pformat(l2s))
if moretry:
return True # Retry
self.error_keep_going("Line " + str(lineno_m1) + " miscompares; " + fn1 + " != " +
fn2)
for c in range(min(len(l1), len(l2))):
if ord(l1[c]) != ord(l2[c]):
print("Miscompare starts at column " + str(c) +
(" w/ F1(got)=0x%02x F2(exp)=0x%02x" % (ord(l1[c]), ord(l2[c]))),
file=sys.stderr)
break
print("F1(got): " + l1 + "F2(exp): " + l2, file=sys.stderr)
if 'HARNESS_UPDATE_GOLDEN' in os.environ: # Update golden files with current
print("%Warning: HARNESS_UPDATE_GOLDEN set: cp " + fn1 + " " + fn2,
file=sys.stderr)
with open(fn2, 'w', encoding="utf8") as fhw:
fhw.write(''.join(l1s))
else:
print("To update reference: HARNESS_UPDATE_GOLDEN=1 {command} or --golden",
file=sys.stderr)
return False # No retry - bad
return False # No retry - good
def files_identical_sorted(self, fn1: str, fn2: str, is_logfile=False) -> None:
"""Test if two files, after sorting both, have identical contents"""
# Set LC_ALL as suggested in the sort manpage to avoid sort order
# changes from the locale.
os.environ['LC_ALL'] = 'C'
fn1sort = fn1 + '.sort'
self.run(cmd=['sort', fn1, "> " + fn1sort])
self.files_identical(fn1sort, fn2, is_logfile)
def copy_if_golden(self, fn1: str, fn2: str) -> None:
"""Copy a file if updating golden .out files"""
if 'HARNESS_UPDATE_GOLDEN' in os.environ: # Update golden files with current
print("%Warning: HARNESS_UPDATE_GOLDEN set: cp " + fn1 + " " + fn2, file=sys.stderr)
shutil.copy(fn1, fn2)
def vcd_identical(self, fn1: str, fn2: str) -> None:
"""Test if two VCD files have logically-identical contents"""
# vcddiff to check transitions, if installed
cmd = "vcddiff --help"
out = test.run_capture(cmd, check=True)
cmd = 'vcddiff ' + fn1 + ' ' + fn2
out = test.run_capture(cmd, check=True)
if out != "":
cmd = 'vcddiff ' + fn2 + " " + fn1 # Reversed arguments
out = VtOs.run_capture(cmd, check=False)
if out != "":
print(out)
self.copy_if_golden(fn1, fn2)
self.error("VCD miscompares " + fn2 + " " + fn1)
# vcddiff doesn't check module and variable scope, so check that
# Also provides backup if vcddiff not installed
h1 = self._vcd_read(fn1)
h2 = self._vcd_read(fn2)
a = json.dumps(h1, sort_keys=True, indent=1)
b = json.dumps(h2, sort_keys=True, indent=1)
if a != b:
self.copy_if_golden(fn1, fn2)
self.error("VCD hier miscompares " + fn1 + " " + fn2 + "\nGOT=" + a + "\nEXP=" + b +
"\n")
def fst2vcd(self, fn1: str, fn2: str) -> None:
cmd = "fst2vcd -h"
out = VtOs.run_capture(cmd, check=False)
if out == "" or not re.search(r'Usage:', out):
self.skip("No fst2vcd installed")
return
cmd = 'fst2vcd -e -f "' + fn1 + '" -o "' + fn2 + '"'
print("\t " + cmd + "\n") # Always print to help debug race cases
out = VtOs.run_capture(cmd, check=False)
print(out)
def fst_identical(self, fn1: str, fn2: str) -> None:
"""Test if two FST files have logically-identical contents"""
tmp = fn1 + ".vcd"
self.fst2vcd(fn1, tmp)
self.vcd_identical(tmp, fn2)
def _vcd_read(self, filename: str) -> str:
data = {}
with open(filename, 'r', encoding='latin-1') as fh:
hier_stack = ["TOP"]
var = []
for line in fh:
match1 = re.search(r'\$scope (module|struct|interface)\s+(\S+)', line)
match2 = re.search(r'(\$var (\S+)\s+\d+\s+)\S+\s+(\S+)', line)
match3 = re.search(r'(\$attrbegin .* \$end)', line)
line = line.rstrip()
# print("VR"+ ' '*len(hier_stack) +" L " + line)
if match1: # $scope
name = match1.group(2)
# print("VR"+ ' '*len(hier_stack) +" scope " + line)
hier_stack += [name]
scope = '.'.join(hier_stack)
data[scope] = match1.group(1) + " " + name
elif match2: # $var
# print("VR"+ ' '*len(hier_stack) +" var " + line)
scope = '.'.join(hier_stack)
var = match2.group(2)
data[scope + "." + var] = match2.group(1) + match2.group(3)
elif match3: # $attrbegin
# print("VR"+ ' '*len(hier_stack) +" attr " + line)
if var:
scope = '.'.join(hier_stack)
data[scope + "." + var + "#"] = match3.group(1)
elif re.search(r'\$enddefinitions', line):
break
n = len(re.findall(r'\$upscope', line))
if n:
for i in range(0, n): # pylint: disable=unused-variable
# print("VR"+ ' '*len(hier_stack) +" upscope " + line)
hier_stack.pop()
return data
def inline_checks(self) -> None:
covfn = self.coverage_filename
contents = self.file_contents(covfn)
if self.verbose:
self.oprint("Extract checks")
with open(self.top_filename, 'r', encoding="utf8") as fh:
flineno = 0
for line in fh:
flineno += 1
if re.search(r'CHECK', line):
match1 = re.search(
r'CHECK_COVER *\( *((-|[0-9])+) *,'
r'*"([^"]+)" *, *("([^"]+)" *,|) *(\d+) *\)', line)
match2 = re.search(r'CHECK_COVER_MISSING *\( *((-|[0-9])+) *\)', line)
if match1:
lineno = flineno + int(match1.group(1))
hier = match1.group(3)
comment = match1.group(5)
count = match1.group(6)
regexp = "\001l\002" + str(lineno)
if comment:
regexp += ".*\001o\002" + re.escape(comment)
if hier:
regexp += ".*\001h\002" + re.escape(hier)
regexp += ".*' " + str(count)
if not re.search(regexp, contents):
self.error("CHECK_COVER: " + covfn + ":" + str(flineno) +
": Regexp not found: " + regexp + "\n" + "From " +
self.top_filename + ": " + line)
elif match2:
lineno = flineno + int(match2.group(1))
regexp = "\001l\002" + str(lineno)
if re.search(regexp, contents):
self.error("CHECK_COVER_MISSING: " + covfn + ":" + str(flineno) +
": Regexp found: " + regexp + "\n" + "From " +
self.top_filename + ": " + line)
else:
self.error(self.top_filename + ":" + str(flineno) +
": Unknown CHECK request: " + line)
@staticmethod
def cfg_with_ccache() -> bool:
if VlTest._cached_cfg_with_ccache is None:
mkf = VlTest._file_contents_static(os.environ['VERILATOR_ROOT'] +
"/include/verilated.mk")
VlTest._cached_cfg_with_ccache = bool(re.match(r'OBJCACHE \?= ccache', mkf))
return VlTest._cached_cfg_with_ccache
def glob_some(self, pattern: str) -> list:
"""Return list of filenames matching a glob, with at least one match required."""
files = glob.glob(pattern)
# print("glob_some('" + pattern + "') files =\n " + pformat(files))
if not files:
self.error("glob_one: pattern '" + pattern + "' does not match any files")
return ['No_file_found']
return sorted(files)
def glob_one(self, pattern: str) -> str:
"""Return a filename matching a glob, with exactly one match required."""
files = self.glob_some(pattern)
if files and len(files) > 1:
msg = "glob_one: pattern '" + pattern + "' matches multiple files:\n"
for file in files:
msg += file + "\n"
self.error(msg)
return 'Multiple_files_found'
return files[0]
def file_grep_not(self, filename: str, regexp) -> None:
contents = self.file_contents(filename)
if contents == "_Already_Errored_":
return
if re.search(regexp, contents, re.MULTILINE):
self.error("File_grep_not: " + filename + ": Regexp found: '" + regexp + "'")
def file_grep(self, filename: str, regexp, expvalue=None) -> list:
contents = self.file_contents(filename)
if contents == "_Already_Errored_":
return None
match = re.search(regexp, contents, re.MULTILINE)
if not match:
self.error("File_grep: " + filename + ": Regexp not found: " + regexp)
return None
if expvalue and str(expvalue) != match.group(1):
self.error("File_grep: " + filename + ": Got='" + match.group(1) + "' Expected='" +
str(expvalue) + "' in regexp: '" + regexp + "'")
return None
return [match.groups()]
def file_grep_count(self, filename: str, regexp, expcount: int) -> None:
contents = self.file_contents(filename)
if contents == "_Already_Errored_":
return
count = len(re.findall(regexp, contents))
if expcount != count:
self.error("File_grep_count: " + filename + ": Got='" + count + "' Expected='" +
expcount + "' in regexp: '" + regexp + "'")
def file_grep_any(self, filenames: list, regexp, expvalue=None) -> None:
for filename in filenames:
contents = self.file_contents(filename)
if contents == "_Already_Errored_":
return
match = re.search(regexp, contents)
if match:
if expvalue and str(expvalue) != match.group(1):
self.error("file_grep: " + filename + ": Got='" + match.group(1) +
"' Expected='" + str(expvalue) + "' in regexp: " + regexp)
return
msg = "file_grep_any: Regexp '" + regexp + "' not found in any of the following files:\n"
for filename in filenames:
msg += filename + "\n"
self.error(msg)
def file_contents(self, filename: str) -> str:
if filename not in self._file_contents_cache:
if not os.path.exists(filename):
self._file_contents_cache[filename] = "_Already_Errored_"
self.error("File_contents file not found: " + filename)
else:
with open(filename, 'r', encoding='latin-1') as fh:
self._file_contents_cache[filename] = str(fh.read())
return self._file_contents_cache[filename]
@staticmethod
def _file_contents_static(filename: str) -> str:
if filename not in VlTest._file_contents_cache:
if not os.path.exists(filename):
VlTest._file_contents_cache[filename] = "_Already_Errored_"
sys.exit("_file_contents_static file not found: " + filename)
else:
with open(filename, 'r', encoding='latin-1') as fh:
VlTest._file_contents_cache[filename] = str(fh.read())
return VlTest._file_contents_cache[filename]
def write_wholefile(self, filename: str, contents: str) -> None:
with open(filename, 'wb') as fh:
fh.write(contents.encode('latin-1'))
if filename in self._file_contents_cache:
del self._file_contents_cache[filename]
def file_sed(self, in_filename: str, out_filename, edit_lambda) -> None:
contents = self.file_contents(in_filename)
contents = edit_lambda(contents)
self.write_wholefile(out_filename, contents)
def extract(
self,
in_filename: str,
out_filename: str,
regexp=r'.*',
lineno_adjust=-9999, #
lines=None) -> None: #'#, #-#'
temp_fn = out_filename
temp_fn = re.sub(r'.*/', '', temp_fn)
temp_fn = self.obj_dir + "/" + temp_fn
out = []
emph = ""
lineno = 0
lineno_out = 0
with open(in_filename, 'r', encoding="latin-1") as fh:
for line in fh:
lineno += 1
if re.search(regexp, line) and self._lineno_match(lineno, lines):
match = re.search(r't/[A-Za-z0-9_]+.v:(\d+):(\d+):', line)
if match:
mlineno = int(match.group(1)) + lineno_adjust
col = int(match.group(2))
mlineno = max(1, mlineno)
line = re.sub(r't/[A-Za-z0-9_]+.v:(\d+):(\d+):',
'example.v:' + str(mlineno) + ':' + str(col), line)
out.append(" " + line)
lineno_out += 1
if '<--' in line:
if emph:
emph += ","
emph += str(lineno_out)
with open(temp_fn, 'w', encoding="latin-1") as fhw:
lang = " sv" if re.search(r'\.s?vh?$', in_filename) else ""
fhw.write(".. comment: generated by " + self.name + "\n")
fhw.write(".. code-block::" + lang + "\n")
if lang != "" and len(out) > 1:
fhw.write(" :linenos:\n")
if emph:
fhw.write(" :emphasize-lines: " + emph + "\n")
fhw.write("\n")
for line in out:
fhw.write(line)
self.files_identical(temp_fn, out_filename)
@staticmethod
def _lineno_match(lineno: int, lines: str) -> bool:
if not lines:
return True
for lc in lines.split(','):
match1 = re.match(r'^(\d+)$', lc)
match2 = re.match(r'^(\d+)-(\d+)$', lc)
if match1 and int(match1.group(1)) == lineno:
return True
if match2 and int(match2.group(1)) <= lineno <= int(match2.group(2)):
return True
return False
######################################################################
######################################################################
# Global Functions
def calc_jobs() -> int:
ok_threads = max_procs()
print("driver.py: Found %d cores, using -j %d" % (ok_threads, ok_threads))
return ok_threads
def calc_threads(default_threads) -> int:
ok_threads = max_procs()
return ok_threads if (ok_threads < default_threads) else default_threads
def _calc_hashset() -> list:
match = re.match(r'^(\d+)/(\d+)$', Args.hashset)
if not match:
sys.exit("%Error: Need number/number format for --hashset: " + Args.hashset)
setn = int(match.group(1))
nsets = int(match.group(2))
new = []
global Arg_Tests
for t in Arg_Tests:
checksum = int(hashlib.sha256(t.encode("utf-8")).hexdigest()[0:4], 16)
if (setn % nsets) == (checksum % nsets):
new.append(t)
Arg_Tests = new
#######################################################################
#######################################################################
# Verilator utilities
@lru_cache(maxsize=1)
def max_procs() -> int:
procs = multiprocessing.cpu_count()
if procs < 2:
print("driver.py: Python didn't find at least two CPUs")
return procs
def _parameter(param: str) -> None:
global _Parameter_Next_Level
if _Parameter_Next_Level:
if not re.match(r'^(\d+)$', param):
sys.exit("%Error: Expected number following " + _Parameter_Next_Level + ": " + param)
Args.passdown_verilator_flags.append(param)
_Parameter_Next_Level = None
elif re.search(r'\.py', param):
Arg_Tests.append(param)
elif re.match(r'^-?(-debugi|-dumpi)', param):
Args.passdown_verilator_flags.append(param)
_Parameter_Next_Level = param
elif re.match(r'^-?(-W||-debug-check)', param):
Args.passdown_verilator_flags.append(param)
else:
sys.exit("%Error: Unknown parameter: " + param)
def run_them() -> None:
VtOs.mkdir_ok("obj_dist")
timestart = time.strftime("%Y%m%d_%H%M%S")
runner = Runner(driver_log_filename="obj_dist/driver_" + timestart + ".log", quiet=Args.quiet)
for test_py in Arg_Tests:
for scenario in sorted(set(Args.scenarios)):
if VlTest._prefilter_scenario(test_py, scenario):
runner.one_test(py_filename=test_py, scenario=scenario)
runner.wait_and_report()
if Args.rerun and runner.fail_cnt and not Quitting:
print('=' * 70)
print('=' * 70)
print("RERUN ==\n")
# Avoid parallel run to ensure that isn't causing problems
# If > 10 failures something more wrong and get results quickly
if runner.fail_cnt < 10:
forker.max_proc(1)
orig_runner = runner
runner = Runner(driver_log_filename="obj_dist/driver_" + timestart + "_rerun.log",
quiet=False,
fail1_cnt=orig_runner.fail_cnt,
ok_cnt=orig_runner.ok_cnt,
skip_cnt=orig_runner.skip_cnt)
for ftest in orig_runner.fail_tests:
# Reschedule test
if ftest.rerunnable:
ftest.clean()
runner.one_test(py_filename=ftest.py_filename,
scenario=ftest.scenario,
rerun_skipping=not ftest.rerunnable)
runner.wait_and_report()
if runner.fail_cnt:
sys.exit(10)
######################################################################
######################################################################
# Main
if __name__ == '__main__':
os.environ['PYTHONUNBUFFERED'] = "1"
if ('VERILATOR_ROOT' not in os.environ) and os.path.isfile('../bin/verilator'):
os.environ['VERILATOR_ROOT'] = os.getcwd() + "/.."
if 'MAKE' not in os.environ:
os.environ['MAKE'] = "make"
if 'CXX' not in os.environ:
os.environ['CXX'] = "c++"
if 'TEST_REGRESS' in os.environ:
sys.exit("%Error: TEST_REGRESS environment variable is already set")
os.environ['TEST_REGRESS'] = os.getcwd()
forker = None
Start = time.time()
_Parameter_Next_Level = None
def sig_int(signum, env) -> None: # pylint: disable=unused-argument
global Quitting
if Quitting:
sys.exit("\nQuitting (immediately)...")
print("\nQuitting... (send another interrupt signal for immediate quit)")
Quitting = True
if forker:
forker.kill_tree_all()
signal.signal(signal.SIGINT, sig_int)
#---------------------------------------------------------------------
parser = argparse.ArgumentParser(
allow_abbrev=False,
formatter_class=argparse.RawDescriptionHelpFormatter,
description="""Run Verilator regression tests""",
epilog="""driver.py invokes Verilator or another simulator on each test file.
See docs/internals.rst in the distribution for more information.
Copyright 2024-2024 by Wilson Snyder. This program is free software; you
can redistribute it and/or modify it under the terms of either the GNU
Lesser General Public License Version 3 or the Perl Artistic License
Version 2.0.
SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0""")
parser.add_argument('--benchmark', action='store', help='enable benchmarking')
parser.add_argument('--debug', action='store_const', const=9, help='enable debug')
# --debugi: see _parameter()
parser.add_argument('--fail-max',
action='store',
default=None,
help='run Verilator executable with gdb')
parser.add_argument('--gdb', action='store_true', help='run Verilator executable with gdb')
parser.add_argument('--gdbbt',
action='store_true',
help='run Verilated executable with gdb and backtrace')
parser.add_argument('--gdbsim', action='store_true', help='run Verilated executable with gdb')
parser.add_argument('--golden', '--gold', action='store_true', help='update golden .out files')
parser.add_argument('--hashset', action='store', help='split tests based on <set>/<numsets>')
parser.add_argument('--jobs',
'-j',
action='store',
default=0,
type=int,
help='parallel job count (0=cpu count)')
parser.add_argument('--quiet',
action='store_true',
help='suppress output except failures and progress')
parser.add_argument('--rerun', action='store_true', help='rerun all tests that fail')
parser.add_argument('--rr', action='store_true', help='run Verilator executable with rr')
parser.add_argument('--rrsim', action='store_true', help='run Verilated executable with rr')
parser.add_argument('--sanitize', action='store_true', help='run address sanitizer')
parser.add_argument('--site',
action='store_true',
help='include VERILATOR_TEST_SITE test list')
parser.add_argument('--stop', action='store_true', help='stop on the first error')
parser.add_argument('--trace', action='store_true', help='enable simulator waveform tracing')
parser.add_argument('--verbose',
action='store_true',
help='compile and run test in verbose mode')
parser.add_argument(
'--verilation', # -no-verilation undocumented debugging
action='store_true',
default=True,
help="don't run verilator compile() phase")
parser.add_argument('--verilated-debug',
action='store_true',
help='enable Verilated executable debug')
## Scenarios
for scen, v in All_Scenarios.items():
parser.add_argument('--' + scen,
dest='scenarios',
action='append_const',
const=scen,
help='scenario-enable ' + scen)
(Args, rest) = parser.parse_known_intermixed_args()
Args.passdown_verilator_flags = []
for arg in rest:
_parameter(arg)
if Args.debug:
Args.passdown_verilator_flags.append("--debug --no-skip-identical")
logging.basicConfig(level=logging.DEBUG)
logging.info("In driver.py, ARGV=" + ' '.join(sys.argv))
# Use Args for some global information, so gets passed to forked process
Args.interactive_debugger = Args.gdb or Args.gdbsim or Args.rr or Args.rrsim
if Args.jobs > 1 and Args.interactive_debugger:
sys.exit("%Error: Unable to use -j > 1 with --gdb* and --rr* options")
if Args.golden:
os.environ['HARNESS_UPDATE_GOLDEN'] = '1'
if Args.jobs == 0:
Args.jobs = 1 if Args.interactive_debugger else calc_jobs()
if not Args.scenarios:
Args.scenarios = []
Args.scenarios.append('dist')
Args.scenarios.append('vlt')
Args.orig_argv_sw = []
for arg in sys.argv:
if re.match(r'^-', arg) and not re.match(r'^-j$', arg):
Args.orig_argv_sw.append(arg)
Args.test_dirs = ["t"]
if 'VERILATOR_TESTS_SITE' in os.environ:
if Args.site or len(Arg_Tests) >= 1:
for tdir in os.environ['VERILATOR_TESTS_SITE'].split(':'):
Args.test_dirs.append(tdir)
if not Arg_Tests: # Run everything
uniq = {}
for tdir in Args.test_dirs:
# Uniquify by inode, so different paths to same place get combined
stats = os.stat(tdir)
if stats.st_ino not in uniq:
uniq[stats.st_ino] = 1
Arg_Tests += sorted(glob.glob(tdir + "/t_*.py"))
if Args.hashset:
_calc_hashset()
# Number of retries when reading logfiles, generally only need many
# retries when system is busy running a lot of tests
Args.log_retries = 10 if (len(Arg_Tests) > 3) else 2
forker = Forker(Args.jobs)
if len(Arg_Tests) >= 2 and Args.jobs >= 2:
# Read supported into master process, so don't call every subprocess
Capabilities.warmup_cache()
# Without this tests such as t_debug_sigsegv_bt_bad.py will occasionally
# block on input and cause a SIGSTOP, then a "fg" was needed to resume testing.
print("== Many jobs; redirecting STDIN", file=sys.stderr)
#
sys.stdin = open("/dev/null", 'r', encoding="utf8") # pylint: disable=consider-using-with
run_them()
Reference in New Issue View Git Blame Copy Permalink