# # Wireshark tests # By Gerald Combs # # Ported from a set of Bash scripts which were copyright 2005 Ulf Lamping # # SPDX-License-Identifier: GPL-2.0-or-later # '''Command line option tests''' import json import sys import os.path import subprocess import subprocesstest from subprocesstest import ExitCodes, grep_output, count_output import shutil import pytest #glossaries = ('fields', 'protocols', 'values', 'decodes', 'defaultprefs', 'currentprefs') glossaries = ('decodes', 'values') testout_pcap = 'testout.pcap' class TestDumpcapOptions: # XXX Should we generate individual test functions instead of looping? def test_dumpcap_invalid_chars(self, cmd_dumpcap, base_env): '''Invalid dumpcap parameters''' for char_arg in 'CEFGHJKNORTUVWXYejloxz': process = subprocesstest.run((cmd_dumpcap, '-' + char_arg), env=base_env) assert process.returncode == ExitCodes.COMMAND_LINE # XXX Should we generate individual test functions instead of looping? def test_dumpcap_valid_chars(self, cmd_dumpcap, base_env): for char_arg in 'hv': process = subprocesstest.run((cmd_dumpcap, '-' + char_arg), env=base_env) assert process.returncode == ExitCodes.OK # XXX Should we generate individual test functions instead of looping? def test_dumpcap_interface_chars(self, cmd_dumpcap, base_env): '''Valid dumpcap parameters requiring capture permissions''' valid_returns = [ExitCodes.OK, ExitCodes.INVALID_INTERFACE] for char_arg in 'DL': process = subprocesstest.run((cmd_dumpcap, '-' + char_arg), env=base_env) assert process.returncode in valid_returns class TestDumpcapClopts: def test_dumpcap_invalid_capfilter(self, cmd_dumpcap, capture_interface, result_file, base_env): '''Invalid capture filter''' invalid_filter = '__invalid_protocol' # $DUMPCAP -f 'jkghg' -w './testout.pcap' > ./testout.txt 2>&1 testout_file = result_file(testout_pcap) process = subprocesstest.run((cmd_dumpcap, '-f', invalid_filter, '-w', testout_file), capture_output=True, env=base_env) assert grep_output(process.stderr, 'Invalid capture filter "' + invalid_filter + '" for interface') def test_dumpcap_invalid_interface_name(self, cmd_dumpcap, capture_interface, result_file, base_env): '''Invalid capture interface name''' invalid_interface = '__invalid_interface' # $DUMPCAP -i invalid_interface -w './testout.pcap' > ./testout.txt 2>&1 testout_file = result_file(testout_pcap) process = subprocesstest.run((cmd_dumpcap, '-i', invalid_interface, '-w', testout_file), capture_output=True, env=base_env) assert grep_output(process.stderr, 'There is no device named "__invalid_interface"') or \ grep_output(process.stderr, 'The capture session could not be initiated on capture device "__invalid_interface"') def test_dumpcap_invalid_interface_index(self, cmd_dumpcap, capture_interface, result_file, base_env): '''Invalid capture interface index''' invalid_index = '0' # $DUMPCAP -i 0 -w './testout.pcap' > ./testout.txt 2>&1 testout_file = result_file(testout_pcap) process = subprocesstest.run((cmd_dumpcap, '-i', invalid_index, '-w', testout_file), capture_output=True, env=base_env) assert grep_output(process.stderr, 'There is no interface with that adapter index') class TestBasicClopts: def test_existing_file(self, cmd_tshark, capture_file, test_env): # $TSHARK -r "${CAPTURE_DIR}dhcp.pcap" > ./testout.txt 2>&1 process = subprocesstest.run((cmd_tshark, '-r', capture_file('dhcp.pcap')), env=test_env) assert process.returncode == ExitCodes.OK def test_existing_file_longopt(self, cmd_tshark, capture_file, test_env): # $TSHARK -r "${CAPTURE_DIR}dhcp.pcap" > ./testout.txt 2>&1 process = subprocesstest.run((cmd_tshark, '--read-file', capture_file('dhcp.pcap'), '--display-filter', 'dhcp'), env=test_env) assert process.returncode == ExitCodes.OK def test_nonexistent_file(self, cmd_tshark, capture_file, test_env): # $TSHARK - r ThisFileDontExist.pcap > ./testout.txt 2 > &1 process = subprocesstest.run((cmd_tshark, '-r', capture_file('__ceci_nest_pas_une.pcap')), env=test_env) assert process.returncode == ExitCodes.INVALID_FILE_ERROR class TestTsharkOptions: # XXX Should we generate individual test functions instead of looping? def test_tshark_invalid_chars(self, cmd_tshark, test_env): '''Invalid tshark parameters''' for char_arg in 'ABCEFHJKMNORTUWXYZabcdefijkmorstuwyz': process = subprocesstest.run((cmd_tshark, '-' + char_arg), env=test_env) assert process.returncode == ExitCodes.COMMAND_LINE # XXX Should we generate individual test functions instead of looping? def test_tshark_valid_chars(self, cmd_tshark, test_env): for char_arg in 'Ghv': process = subprocesstest.run((cmd_tshark, '-' + char_arg), env=test_env) assert process.returncode == ExitCodes.OK # XXX Should we generate individual test functions instead of looping? def test_tshark_interface_chars(self, cmd_tshark, cmd_dumpcap, test_env): '''Valid tshark parameters requiring capture permissions''' # These options require dumpcap, but may fail with a pacp error # if WinPcap or Npcap are not present valid_returns = [ExitCodes.OK, ExitCodes.PCAP_ERROR, ExitCodes.INVALID_CAPABILITY, ExitCodes.INVALID_INTERFACE] for char_arg in 'DL': process = subprocesstest.run((cmd_tshark, '-' + char_arg), env=test_env) assert process.returncode in valid_returns def test_tshark_disable_protos(self, cmd_tshark, capture_file, test_env): '''--disable-protocol/--enable-protocol from !16923''' process = subprocesstest.run((cmd_tshark, "-r", capture_file("http.pcap"), "--disable-protocol", "ALL", "--enable-protocol", "eth,ip", "-Tjson", "-eeth.type", "-eip.proto", "-ehttp.host", ), capture_output=True, env=test_env) assert process.returncode == ExitCodes.OK obj = json.loads(process.stdout)[0]['_source']['layers'] assert obj.get('eth.type', 'NOT FOUND') == ['0x0800'] assert obj.get('ip.proto', 'NOT FOUND') == ['6'] assert obj.get('http.host', 'NOT FOUND') == 'NOT FOUND' class TestTsharkCaptureClopts: def test_tshark_invalid_capfilter(self, cmd_tshark, capture_interface, result_file, test_env): '''Invalid capture filter''' invalid_filter = '__invalid_protocol' # $TSHARK -f 'jkghg' -w './testout.pcap' > ./testout.txt 2>&1 testout_file = result_file(testout_pcap) process = subprocesstest.run((cmd_tshark, '-f', invalid_filter, '-w', testout_file ), capture_output=True, env=test_env) assert grep_output(process.stderr, 'Invalid capture filter "' + invalid_filter + '" for interface') def test_tshark_invalid_interface_name(self, cmd_tshark, capture_interface, result_file, test_env): '''Invalid capture interface name''' invalid_interface = '__invalid_interface' # $TSHARK -i invalid_interface -w './testout.pcap' > ./testout.txt 2>&1 testout_file = result_file(testout_pcap) process = subprocesstest.run((cmd_tshark, '-i', invalid_interface, '-w', testout_file), capture_output=True, env=test_env) assert grep_output(process.stderr, 'There is no device named "__invalid_interface"') or \ grep_output(process.stderr, 'The capture session could not be initiated on capture device "__invalid_interface"') def test_tshark_invalid_interface_index(self, cmd_tshark, capture_interface, result_file, test_env): '''Invalid capture interface index''' invalid_index = '0' # $TSHARK -i 0 -w './testout.pcap' > ./testout.txt 2>&1 testout_file = result_file(testout_pcap) process = subprocesstest.run((cmd_tshark, '-i', invalid_index, '-w', testout_file), capture_output=True, env=test_env) assert grep_output(process.stderr, 'There is no interface with that adapter index') class TestTsharkNameResolutionClopts: def test_tshark_valid_name_resolution(self, cmd_tshark, capture_file, test_env): # $TSHARK -N mnNtdv -a duration:1 > ./testout.txt 2>&1 process = subprocesstest.run((cmd_tshark, '-r', capture_file('empty.pcap'), '-N', 'mnNtdv', ), env=test_env) assert process.returncode == 0 # XXX Add invalid name resolution. class TestTsharkUnicodeClopts: def test_tshark_unicode_display_filter(self, cmd_tshark, capture_file, test_env): '''Unicode (UTF-8) display filter''' process = subprocesstest.run((cmd_tshark, '-r', capture_file('http.pcap'), '-Y', 'tcp.flags.str == "·······AP···"'), capture_output=True, env=test_env) assert grep_output(process.stdout, 'HEAD.*/v4/iuident.cab') class TestTsharkDumpGlossaries: def test_tshark_dump_glossary(self, cmd_tshark, base_env): for glossary in glossaries: process = subprocesstest.run((cmd_tshark, '-G', glossary), capture_output=True, env=base_env) assert not process.stderr, 'Found error output while printing glossary ' + glossary def test_tshark_glossary_valid_utf8(self, cmd_tshark, base_env): for glossary in glossaries: env = base_env env['LANG'] = 'en_US.UTF-8' # subprocess.run() returns bytes here. proc = subprocess.run((cmd_tshark, '-G', glossary), capture_output=True, env=env) assert proc.returncode == 0 proc.stdout.decode('UTF-8') def test_tshark_glossary_plugin_count(self, cmd_tshark, base_env, features): if not features.have_plugins: pytest.skip('Test requires binary plugin support.') process = subprocesstest.run((cmd_tshark, '-G', 'plugins'), capture_output=True, env=base_env) assert count_output(process.stdout, 'dissector') >= 10, 'Fewer than 10 dissector plugins found' def test_tshark_elastic_mapping(self, cmd_tshark, dirs, base_env): def get_ip_props(obj): return obj['mappings']['properties']['layers']['properties']['ip']['properties'] baseline_file = os.path.join(dirs.baseline_dir, 'elastic-mapping-ip-subset.json') with open(baseline_file) as f: expected_obj = json.load(f) keys_to_check = get_ip_props(expected_obj).keys() proc = subprocesstest.run((cmd_tshark, '-G', 'elastic-mapping', '--elastic-mapping-filter', 'ip'), capture_output=True, env=base_env) actual_obj = json.loads(proc.stdout) ip_props = get_ip_props(actual_obj) for key in list(ip_props.keys()): if key not in keys_to_check: del ip_props[key] assert actual_obj == expected_obj def test_tshark_unicode_folders(self, cmd_tshark, unicode_env, features): '''Folders output with unicode''' if not features.have_lua: pytest.skip('Test requires Lua scripting support.') if sys.platform == 'win32' and not features.have_lua_unicode: pytest.skip('Test requires a patched Lua build with UTF-8 support.') proc = subprocesstest.run((cmd_tshark, '-G', 'folders'), capture_output=True, env=unicode_env.env) out = proc.stdout pluginsdir = [x.split('\t', 1)[1] for x in out.splitlines() if x.startswith('Personal Lua Plugins:')] assert [unicode_env.pluginsdir] == pluginsdir class TestTsharkZExpert: def test_tshark_z_expert_all(self, cmd_tshark, capture_file, test_env): proc = subprocesstest.run((cmd_tshark, '-q', '-z', 'expert', '-o', 'tcp.check_checksum:TRUE', '-r', capture_file('http-ooo-fuzzed.pcapng')), capture_output=True, env=test_env) # http2-data-reassembly.pcap has Errors, Warnings, Notes, and Chats # when TCP checksum are verified. assert grep_output(proc.stdout, 'Errors') assert grep_output(proc.stdout, 'Warns') assert grep_output(proc.stdout, 'Notes') assert grep_output(proc.stdout, 'Chats') def test_tshark_z_expert_error(self, cmd_tshark, capture_file, test_env): proc = subprocesstest.run((cmd_tshark, '-q', '-z', 'expert,error', '-o', 'tcp.check_checksum:TRUE', '-r', capture_file('http-ooo-fuzzed.pcapng')), capture_output=True, env=test_env) assert grep_output(proc.stdout, 'Errors') assert not grep_output(proc.stdout, 'Warns') assert not grep_output(proc.stdout, 'Notes') assert not grep_output(proc.stdout, 'Chats') def test_tshark_z_expert_warn(self, cmd_tshark, capture_file, test_env): proc = subprocesstest.run((cmd_tshark, '-q', '-z', 'expert,warn', '-o', 'tcp.check_checksum:TRUE', '-r', capture_file('http-ooo-fuzzed.pcapng')), capture_output=True, env=test_env) assert grep_output(proc.stdout, 'Errors') assert grep_output(proc.stdout, 'Warns') assert not grep_output(proc.stdout, 'Notes') assert not grep_output(proc.stdout, 'Chats') def test_tshark_z_expert_note(self, cmd_tshark, capture_file, test_env): proc = subprocesstest.run((cmd_tshark, '-q', '-z', 'expert,note', '-o', 'tcp.check_checksum:TRUE', '-r', capture_file('http-ooo-fuzzed.pcapng')), capture_output=True, env=test_env) assert grep_output(proc.stdout, 'Errors') assert grep_output(proc.stdout, 'Warns') assert grep_output(proc.stdout, 'Notes') assert not grep_output(proc.stdout, 'Chats') def test_tshark_z_expert_chat(self, cmd_tshark, capture_file, test_env): proc = subprocesstest.run((cmd_tshark, '-q', '-z', 'expert,chat', '-o', 'tcp.check_checksum:TRUE', '-r', capture_file('http-ooo-fuzzed.pcapng')), capture_output=True, env=test_env) assert grep_output(proc.stdout, 'Errors') assert grep_output(proc.stdout, 'Warns') assert grep_output(proc.stdout, 'Notes') assert grep_output(proc.stdout, 'Chats') def test_tshark_z_expert_comment(self, cmd_tshark, capture_file, test_env): proc = subprocesstest.run((cmd_tshark, '-q', '-z', 'expert,comment', '-r', capture_file('sip.pcapng')), capture_output=True, env=test_env) assert grep_output(proc.stdout, 'Notes') assert grep_output(proc.stdout, 'Comments') def test_tshark_z_expert_invalid_filter(self, cmd_tshark, capture_file, test_env): invalid_filter = '__invalid_protocol' proc = subprocesstest.run((cmd_tshark, '-q', '-z', 'expert,' + invalid_filter, '-r', capture_file('http-ooo.pcap')), capture_output=True, env=test_env) assert proc.returncode == ExitCodes.COMMAND_LINE assert grep_output(proc.stdout, 'Filter "' + invalid_filter + '" is invalid') def test_tshark_z_expert_error_invalid_filter(self, cmd_tshark, capture_file, test_env): invalid_filter = '__invalid_protocol' proc = subprocesstest.run((cmd_tshark, '-q', '-z', 'expert,error,' + invalid_filter, '-r', capture_file('http-ooo.pcap')), capture_output=True, env=test_env) assert proc.returncode == ExitCodes.COMMAND_LINE assert grep_output(proc.stdout, 'Filter "' + invalid_filter + '" is invalid') def test_tshark_z_expert_filter(self, cmd_tshark, capture_file, test_env): proc = subprocesstest.run((cmd_tshark, '-q', '-z', 'expert,udp', '-o', 'tcp.check_checksum:TRUE', '-r', capture_file('http-ooo-fuzzed.pcapng')), capture_output=True, env=test_env) # Filtering for UDP should produce no expert infos. assert not grep_output(proc.stdout, 'Errors') assert not grep_output(proc.stdout, 'Warns') assert not grep_output(proc.stdout, 'Notes') assert not grep_output(proc.stdout, 'Chats') def test_tshark_z_expert_error_filter(self, cmd_tshark, capture_file, test_env): proc = subprocesstest.run((cmd_tshark, '-q', '-z', 'expert,note,http', # tls is a filter '-o', 'tcp.check_checksum:TRUE', '-r', capture_file('http-ooo-fuzzed.pcapng')), capture_output=True, env=test_env) # Filtering for HTTP and Note level expert info should produce only # Error and Warning level expert infos with checksumming turned on. # The Note warnings on are packets with TCP but not HTTP, and we're # filtering out the Chat level. assert grep_output(proc.stdout, 'Errors') assert grep_output(proc.stdout, 'Warns') assert not grep_output(proc.stdout, 'Notes') assert not grep_output(proc.stdout, 'Chats') class TestTsharkExtcap: # dumpcap dependency has been added to run this test only with capture support def test_tshark_extcap_interfaces(self, cmd_tshark, cmd_dumpcap, test_env, home_path): # Script extcaps don't work with the current code on windows. # https://www.wireshark.org/docs/wsdg_html_chunked/ChCaptureExtcap.html # TODO: skip this test until it will get fixed. if sys.platform == 'win32': pytest.skip('FIXME extcap .py scripts needs special treatment on Windows') extcap_dir_path = os.path.join(home_path, 'extcap') os.makedirs(extcap_dir_path) test_env['WIRESHARK_EXTCAP_DIR'] = extcap_dir_path source_file = os.path.join(os.path.dirname(__file__), 'sampleif.py') # We run our tests in a bare, reproducible home environment. This can result in an # invalid or missing Python interpreter if our main environment has a wonky Python # path, as is the case in the GitLab SaaS macOS runners which use `asdf`. Force # sampleif.py to use our current Python executable. with open(source_file, 'r') as sf: sampleif_py = sf.read() sampleif_py = sampleif_py.replace('/usr/bin/env python3', sys.executable) sys.stderr.write(sampleif_py) extcap_file = os.path.join(extcap_dir_path, 'sampleif.py') with open(extcap_file, 'w') as ef: ef.write(sampleif_py) os.fchmod(ef.fileno(), os.fstat(sf.fileno()).st_mode) # Ensure the test extcap_tool is properly loaded proc = subprocesstest.run((cmd_tshark, '-D'), capture_output=True, env=test_env) assert count_output(proc.stdout, 'sampleif') == 1 # Ensure tshark lists 2 interfaces in the preferences proc = subprocesstest.run((cmd_tshark, '-G', 'currentprefs'), capture_output=True, env=test_env) assert count_output(proc.stdout, 'extcap.sampleif.test') == 2