Files
graphene/Scripts/regression.py
Michał Kowalczyk d53729b201 [Pal] Rework manifest loading
This is a major refactor of the way manifests are loaded and handled,
which will be followed by a complete rework of the loader code (which
will include e.g. centralized config).

Changes/fixes:
- Huge part of manifest handling was refactored and untangled.
- Starting without a manifest is now disallowed. This was actually
  accidentally broken for some time and no one complained. It also makes
  little sense in practice and in Graphene's overall design, e.g. it
  conflicts with protected argv.
- Now we only allow starting by giving the executable, not manifest (the
  magic resolution logic was removed).
- Now manifests are sent over pipes between parent and children, instead
  of children finding and loading them on their own. This is a
  preparation for the upcoming centralized manifests change.
- Previously manifests were parsed 2 times on Linux and 3 times on
  Linux-SGX (by untrusted PAL, trusted PAL and LibOS). This is now
  fixed.
- The common `pal_main()` now requires that the backend-specific PAL
  loader loads the manifest before calling it. SGX code already has to
  do it (for proper initialization), so let's unify this interface for
  all PALs.
- Fix for a PAL crash when manifest size was divisible by page size
  (sic!). NULL termination was missing, but most of the time the padding
  to page size saved Graphene from crashing.
2020-12-05 01:46:03 +01:00

137 lines
4.8 KiB
Python

import contextlib
import os
import pathlib
import signal
import subprocess
import sys
import unittest
# pylint: disable=subprocess-popen-preexec-fn,subprocess-run-check
HAS_SGX = os.environ.get('SGX') == '1'
ON_X86 = os.uname().machine in ['x86_64']
def expectedFailureIf(predicate):
if predicate:
return unittest.expectedFailure
return lambda func: func
class RegressionTestCase(unittest.TestCase):
LOADER_ENV = 'PAL_LOADER'
LIBPAL_PATH_ENV = 'LIBPAL_PATH'
HOST_PAL_PATH_ENV = 'HOST_PAL_PATH'
DEFAULT_TIMEOUT = (20 if HAS_SGX else 10)
def get_env(self, name):
try:
return os.environ[name]
except KeyError:
self.fail('environment variable {} not set'.format(name))
def has_debug(self):
libpal = self.get_env(self.LIBPAL_PATH_ENV)
p = subprocess.run(['objdump', '-x', libpal], check=True, stdout=subprocess.PIPE)
dump = p.stdout.decode()
return '.debug_info' in dump
def run_gdb(self, args, gdb_script, **kwds):
host_pal_path = self.get_env(self.HOST_PAL_PATH_ENV)
# See also pal_loader.
prefix = ['gdb', '-q']
env = os.environ.copy()
if HAS_SGX:
prefix += ['-x', os.path.join(host_pal_path, 'gdb_integration/graphene_sgx_gdb.py')]
sgx_gdb = os.path.join(host_pal_path, 'gdb_integration/sgx_gdb.so')
env['LD_PRELOAD'] = sgx_gdb + ':' + env.get('LD_PRELOAD', '')
else:
prefix += ['-x', os.path.join(host_pal_path, 'gdb_integration/graphene_gdb.py')]
# Override TTY, as apparently os.setpgrp() confuses GDB and causes it to hang.
prefix += ['-x', gdb_script, '-batch', '-tty=/dev/null']
prefix += ['--args']
return self.run_binary(args, prefix=prefix, env=env, **kwds)
def run_binary(self, args, *, timeout=None, prefix=None, **kwds):
timeout = (max(self.DEFAULT_TIMEOUT, timeout) if timeout is not None
else self.DEFAULT_TIMEOUT)
loader = self.get_env(self.LOADER_ENV)
if not pathlib.Path(loader).exists():
self.skipTest('loader ({}) not found'.format(loader))
libpal = self.get_env(self.LIBPAL_PATH_ENV)
if not pathlib.Path(libpal).exists():
self.skipTest('libpal ({}) not found'.format(libpal))
if prefix is None:
prefix = []
with subprocess.Popen([*prefix, loader, libpal, 'init', *args],
stdout=subprocess.PIPE, stderr=subprocess.PIPE,
preexec_fn=os.setpgrp,
**kwds) as process:
try:
stdout, stderr = process.communicate(timeout=timeout)
except subprocess.TimeoutExpired:
os.killpg(process.pid, signal.SIGKILL)
self.fail('timeout ({} s) expired'.format(timeout))
self.print_output(stdout, stderr)
if process.returncode:
raise subprocess.CalledProcessError(
process.returncode, args, stdout, stderr)
return stdout.decode(), stderr.decode()
@classmethod
def run_native_binary(cls, args, timeout=None, libpath=None, **kwds):
timeout = (max(cls.DEFAULT_TIMEOUT, timeout) if timeout is not None
else cls.DEFAULT_TIMEOUT)
my_env = os.environ.copy()
if not libpath is None:
my_env["LD_LIBRARY_PATH"] = libpath
with subprocess.Popen(args,
stdout=subprocess.PIPE, stderr=subprocess.PIPE,
env=my_env,
preexec_fn=os.setpgrp,
**kwds) as process:
try:
stdout, stderr = process.communicate(timeout=timeout)
except subprocess.TimeoutExpired:
os.killpg(process.pid, signal.SIGKILL)
raise AssertionError('timeout ({} s) expired'.format(timeout))
cls.print_output(stdout, stderr)
if process.returncode:
raise subprocess.CalledProcessError(
process.returncode, args, stdout, stderr)
return stdout.decode(), stderr.decode()
@staticmethod
def print_output(stdout: bytes, stderr: bytes):
'''
Print command output (stdout, stderr) so that pytest can capture it.
'''
sys.stdout.write(stdout.decode(errors='surrogateescape'))
sys.stderr.write(stderr.decode(errors='surrogateescape'))
@contextlib.contextmanager
def expect_returncode(self, returncode):
if returncode == 0:
raise ValueError('expected returncode should be nonzero')
try:
yield
self.fail('did not fail (expected {})'.format(returncode))
except subprocess.CalledProcessError as e:
self.assertEqual(e.returncode, returncode,
'failed with returncode {} (expected {})'.format(
e.returncode, returncode))