diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-07 18:49:45 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-07 18:49:45 +0000 |
commit | 2c3c1048746a4622d8c89a29670120dc8fab93c4 (patch) | |
tree | 848558de17fb3008cdf4d861b01ac7781903ce39 /tools/testing/selftests/tc-testing/plugin-lib/valgrindPlugin.py | |
parent | Initial commit. (diff) | |
download | linux-2c3c1048746a4622d8c89a29670120dc8fab93c4.tar.xz linux-2c3c1048746a4622d8c89a29670120dc8fab93c4.zip |
Adding upstream version 6.1.76.upstream/6.1.76upstream
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'tools/testing/selftests/tc-testing/plugin-lib/valgrindPlugin.py')
-rw-r--r-- | tools/testing/selftests/tc-testing/plugin-lib/valgrindPlugin.py | 160 |
1 files changed, 160 insertions, 0 deletions
diff --git a/tools/testing/selftests/tc-testing/plugin-lib/valgrindPlugin.py b/tools/testing/selftests/tc-testing/plugin-lib/valgrindPlugin.py new file mode 100644 index 000000000..4bb866575 --- /dev/null +++ b/tools/testing/selftests/tc-testing/plugin-lib/valgrindPlugin.py @@ -0,0 +1,160 @@ +''' +run the command under test, under valgrind and collect memory leak info +as a separate test. +''' + + +import os +import re +import signal +from string import Template +import subprocess +import time +from TdcPlugin import TdcPlugin +from TdcResults import * + +from tdc_config import * + +def vp_extract_num_from_string(num_as_string_maybe_with_commas): + return int(num_as_string_maybe_with_commas.replace(',','')) + +class SubPlugin(TdcPlugin): + def __init__(self): + self.sub_class = 'valgrind/SubPlugin' + self.tap = '' + self._tsr = TestSuiteReport() + super().__init__() + + def pre_suite(self, testcount, testidlist): + '''run commands before test_runner goes into a test loop''' + super().pre_suite(testcount, testidlist) + if self.args.verbose > 1: + print('{}.pre_suite'.format(self.sub_class)) + if self.args.valgrind: + self._add_to_tap('1..{}\n'.format(self.testcount)) + + def post_suite(self, index): + '''run commands after test_runner goes into a test loop''' + super().post_suite(index) + if self.args.verbose > 1: + print('{}.post_suite'.format(self.sub_class)) + #print('{}'.format(self.tap)) + for xx in range(index - 1, self.testcount): + res = TestResult('{}-mem'.format(self.testidlist[xx]), 'Test skipped') + res.set_result(ResultState.skip) + res.set_errormsg('Skipped because of prior setup/teardown failure') + self._add_results(res) + if self.args.verbose < 4: + subprocess.check_output('rm -f vgnd-*.log', shell=True) + + def add_args(self, parser): + super().add_args(parser) + self.argparser_group = self.argparser.add_argument_group( + 'valgrind', + 'options for valgrindPlugin (run command under test under Valgrind)') + + self.argparser_group.add_argument( + '-V', '--valgrind', action='store_true', + help='Run commands under valgrind') + + return self.argparser + + def adjust_command(self, stage, command): + super().adjust_command(stage, command) + cmdform = 'list' + cmdlist = list() + + if not self.args.valgrind: + return command + + if self.args.verbose > 1: + print('{}.adjust_command'.format(self.sub_class)) + + if not isinstance(command, list): + cmdform = 'str' + cmdlist = command.split() + else: + cmdlist = command + + if stage == 'execute': + if self.args.verbose > 1: + print('adjust_command: stage is {}; inserting valgrind stuff in command [{}] list [{}]'. + format(stage, command, cmdlist)) + cmdlist.insert(0, '--track-origins=yes') + cmdlist.insert(0, '--show-leak-kinds=definite,indirect') + cmdlist.insert(0, '--leak-check=full') + cmdlist.insert(0, '--log-file=vgnd-{}.log'.format(self.args.testid)) + cmdlist.insert(0, '-v') # ask for summary of non-leak errors + cmdlist.insert(0, ENVIR['VALGRIND_BIN']) + else: + pass + + if cmdform == 'str': + command = ' '.join(cmdlist) + else: + command = cmdlist + + if self.args.verbose > 1: + print('adjust_command: return command [{}]'.format(command)) + return command + + def post_execute(self): + if not self.args.valgrind: + return + + res = TestResult('{}-mem'.format(self.args.testid), + '{} memory leak check'.format(self.args.test_name)) + if self.args.test_skip: + res.set_result(ResultState.skip) + res.set_errormsg('Test case designated as skipped.') + self._add_results(res) + return + + self.definitely_lost_re = re.compile( + r'definitely lost:\s+([,0-9]+)\s+bytes in\s+([,0-9]+)\sblocks', re.MULTILINE | re.DOTALL) + self.indirectly_lost_re = re.compile( + r'indirectly lost:\s+([,0-9]+)\s+bytes in\s+([,0-9]+)\s+blocks', re.MULTILINE | re.DOTALL) + self.possibly_lost_re = re.compile( + r'possibly lost:\s+([,0-9]+)bytes in\s+([,0-9]+)\s+blocks', re.MULTILINE | re.DOTALL) + self.non_leak_error_re = re.compile( + r'ERROR SUMMARY:\s+([,0-9]+) errors from\s+([,0-9]+)\s+contexts', re.MULTILINE | re.DOTALL) + + def_num = 0 + ind_num = 0 + pos_num = 0 + nle_num = 0 + + # what about concurrent test runs? Maybe force them to be in different directories? + with open('vgnd-{}.log'.format(self.args.testid)) as vfd: + content = vfd.read() + def_mo = self.definitely_lost_re.search(content) + ind_mo = self.indirectly_lost_re.search(content) + pos_mo = self.possibly_lost_re.search(content) + nle_mo = self.non_leak_error_re.search(content) + + if def_mo: + def_num = int(def_mo.group(2)) + if ind_mo: + ind_num = int(ind_mo.group(2)) + if pos_mo: + pos_num = int(pos_mo.group(2)) + if nle_mo: + nle_num = int(nle_mo.group(1)) + + mem_results = '' + if (def_num > 0) or (ind_num > 0) or (pos_num > 0) or (nle_num > 0): + mem_results += 'not ' + res.set_result(ResultState.fail) + res.set_failmsg('Memory leak detected') + res.append_failmsg(content) + else: + res.set_result(ResultState.success) + + self._add_results(res) + + + def _add_results(self, res): + self._tsr.add_resultdata(res) + + def _add_to_tap(self, more_tap_output): + self.tap += more_tap_output |