On Mon, Sep 14, 2009 at 7:23 AM, Michael Goldish <mgoldish@xxxxxxxxxx> wrote:
> I think this is a very useful feature to have.
>
> Please see some very minor comments below.
>
> ----- "Lucas Meneghel Rodrigues" <lmr@xxxxxxxxxx> wrote:
>
>> This patch adds a system to watch user space segmentation
>> faults, writing core dumps and some degree of core dump
>> analysis report. We believe that such a system will be
>> beneficial for autotest as a whole, since the ability to
>> get core dumps and dump analysis for each app crashing
>> during an autotest execution can help test engineers with
>> richer debugging information.
>>
>> The system is comprised by 2 parts:
>>
>> * Modifications on test code that enable core dumps
>> generation, register a core handler script in the kernel
>> and check by generated core files at the end of each
>> test.
>>
>> * A core handler script that is going to write the
>> core on each test debug dir in a convenient way, with
>> a report that currently is comprised by the process that
>> died and a gdb stacktrace of the process. As the system
>> gets in shape, we could add more scripts that can do
>> fancier stuff (such as handlers that use frysk to get
>> more info such as memory maps, provided that we have
>> frysk installed in the machine).
>>
>> This is the proof of concept of the system. I am sending it
>> to the mailing list on this early stage so I can get
>> feedback on the feature. The system passes my basic
>> tests:
>>
>> * Run a simple long test, such as the kvm test, and
>> then crash an application while the test is running. I
>> get reports generated on test.debugdir
>>
>> * Run a slightly more complex control file, with 3 parallel
>> bonnie instances at once and crash an application while the
>> test is running. I get reports generated on all
>> test.debugdirs.
>>
>> 3rd try:
>> * Explicitely enable core dumps using the resource module
>> * Fixed a bug on the crash detection code, and factored
>> it into a utility function.
>>
>> I believe we are good to go now.
>>
>> Signed-off-by: Lucas Meneghel Rodrigues <lmr@xxxxxxxxxx>
>> ---
>> client/common_lib/test.py | 66 +++++++++++++-
>> client/tools/crash_handler.py | 202
>> +++++++++++++++++++++++++++++++++++++++++
>> 2 files changed, 266 insertions(+), 2 deletions(-)
>> create mode 100755 client/tools/crash_handler.py
>>
>> diff --git a/client/common_lib/test.py b/client/common_lib/test.py
>> index 362c960..65b78a3 100644
>> --- a/client/common_lib/test.py
>> +++ b/client/common_lib/test.py
>> @@ -17,7 +17,7 @@
>> # tmpdir eg. tmp/<tempname>_<testname.tag>
>>
>> import fcntl, os, re, sys, shutil, tarfile, tempfile, time,
>> traceback
>> -import warnings, logging
>> +import warnings, logging, glob, resource
>>
>> from autotest_lib.client.common_lib import error
>> from autotest_lib.client.bin import utils
>> @@ -31,7 +31,6 @@ class base_test:
>> self.job = job
>> self.pkgmgr = job.pkgmgr
>> self.autodir = job.autodir
>> -
>> self.outputdir = outputdir
>> self.tagged_testname = os.path.basename(self.outputdir)
>> self.resultsdir = os.path.join(self.outputdir, 'results')
>> @@ -40,6 +39,7 @@ class base_test:
>> os.mkdir(self.profdir)
>> self.debugdir = os.path.join(self.outputdir, 'debug')
>> os.mkdir(self.debugdir)
>> + self.configure_crash_handler()
>> self.bindir = bindir
>> if hasattr(job, 'libdir'):
>> self.libdir = job.libdir
>> @@ -54,6 +54,66 @@ class base_test:
>> self.after_iteration_hooks = []
>>
>>
>> + def configure_crash_handler(self):
>> + """
>> + Configure the crash handler by:
>> + * Setting up core size to unlimited
>> + * Putting an appropriate crash handler on
>> /proc/sys/kernel/core_pattern
>> + * Create files that the crash handler will use to figure
>> which tests
>> + are active at a given moment
>> +
>> + The crash handler will pick up the core file and write it to
>> + self.debugdir, and perform analysis on it to generate a
>> report. The
>> + program also outputs some results to syslog.
>> +
>> + If multiple tests are running, an attempt to verify if we
>> still have
>> + the old PID on the system process table to determine whether
>> it is a
>> + parent of the current test execution. If we can't determine
>> it, the
>> + core file and the report file will be copied to all test
>> debug dirs.
>> + """
>> + self.pattern_file = '/proc/sys/kernel/core_pattern'
>> + try:
>> + # Enable core dumps
>> + resource.setrlimit(resource.RLIMIT_CORE, (-1, -1))
>> + # Trying to backup core pattern and register our script
>> + self.core_pattern_backup = open(self.pattern_file,
>> 'r').read()
>> + pattern_file = open(self.pattern_file, 'w')
>> + tools_dir = os.path.join(self.autodir, 'tools')
>> + crash_handler_path = os.path.join(tools_dir,
>> 'crash_handler.py')
>> + pattern_file.write('|' + crash_handler_path + ' %p %t %u
>> %s %h %e')
>> + # Writing the files that the crash handler is going to
>> use
>> + self.debugdir_tmp_file = ('/tmp/autotest_results_dir.%s'
>> %
>> + os.getpid())
>> + utils.open_write_close(self.debugdir_tmp_file,
>> self.debugdir + "\n")
>> + except Exception, e:
>> + self.crash_handling_enabled = False
>> + logging.error('Crash handling system disabled: %s' % e)
>> + else:
>> + self.crash_handling_enabled = True
>> + logging.debug('Crash handling system enabled.')
>> +
>> +
>> + def crash_handler_report(self):
>> + """
>> + If core dumps are found on the debugdir after the execution
>> of the
>> + test, let the user know.
>> + """
>> + if self.crash_handling_enabled:
>> + core_dirs = glob.glob('%s/crash.*' % self.debugdir)
>> + if core_dirs:
>> + logging.warning('Programs crashed during test
>> execution:')
>> + for dir in core_dirs:
>> + logging.warning('Please verify %s for more info',
>> dir)
>> + # Remove the debugdir info file
>> + os.unlink(self.debugdir_tmp_file)
>> + # Restore the core pattern backup
>> + try:
>> + utils.open_write_close(self.pattern_file,
>> + self.core_pattern_backup)
>> + except EnvironmentError:
>> + pass
>> +
>> +
>> def assert_(self, expr, msg='Assertion failed.'):
>> if not expr:
>> raise error.TestError(msg)
>> @@ -377,6 +437,7 @@ class base_test:
>> traceback.print_exc()
>> print 'Now raising the earlier %s error' %
>> exc_info[0]
>> finally:
>> + self.crash_handler_report()
>> self.job.logging.restore()
>> try:
>> raise exc_info[0], exc_info[1], exc_info[2]
>> @@ -389,6 +450,7 @@ class base_test:
>> if run_cleanup:
>> _cherry_pick_call(self.cleanup, *args,
>> **dargs)
>> finally:
>> + self.crash_handler_report()
>> self.job.logging.restore()
>> except error.AutotestError:
>> if self.network_destabilizing:
>> diff --git a/client/tools/crash_handler.py
>> b/client/tools/crash_handler.py
>> new file mode 100755
>> index 0000000..e281eb5
>> --- /dev/null
>> +++ b/client/tools/crash_handler.py
>> @@ -0,0 +1,202 @@
>> +#!/usr/bin/python
>> +"""
>> +Simple crash handling application for autotest
>> +
>> +@copyright Red Hat Inc 2009
>> +@author Lucas Meneghel Rodrigues <lmr@xxxxxxxxxx>
>> +"""
>> +import sys, os, commands, glob, tempfile, shutil, syslog
>> +
>> +
>> +def get_parent_pid(pid):
>> + """
>> + Returns the parent PID for a given PID, converted to an integer.
>> +
>> + @param pid: Process ID.
>> + """
>> + try:
>> + stat_file_contents = open('/proc/%s/stat' % pid,
>> 'r').readline()
>> + ppid = int(stat_file_contents.split(" ")[3])
>
> I think there's no need for the " " in the split() call, because
> split() defaults to splitting around any kind of whitespace.
Right, fair enough!
> The two lines can be combined into a rather short one, BTW:
>
> ppid = int(open('/proc/%s/stat' % pid).read().split()[3])
>
> (open() defaults to 'r' mode.)
Done. I prefered the above, it was a bit more explicit, but in the end
I realized it was not harder to understand the condensed form.
>> + except:
>> + # It is not possible to determine the parent because the
>> process
>> + # already left the process table.
>> + ppid = 1
>> +
>> + return ppid
>> +
>> +
>> +def pid_descends_from(pid_a, pid_b):
>> + """
>> + Check whether pid_a descends from pid_b.
>> +
>> + @param pid_a: Process ID.
>> + @param pid_b: Process ID.
>> + """
>> + pid_a = int(pid_a)
>> + pid_b = int(pid_b)
>> + current_pid = pid_a
>> + while current_pid > 1:
>> + if current_pid == pid_b:
>> + syslog.syslog(syslog.LOG_INFO,
>> + "PID %s descends from PID %s!" % (pid_a,
>> pid_b))
>> + return True
>> + else:
>> + current_pid = get_parent_pid(current_pid)
>> + syslog.syslog(syslog.LOG_INFO,
>> + "PID %s does not descend from PID %s" % (pid_a,
>> pid_b))
>> + return False
>> +
>> +
>> +def write_to_file(file_path, contents):
>> + """
>> + Write contents to a given file path specified. If not specified,
>> the file
>> + will be created.
>> +
>> + @param file_path: Path to a given file.
>> + @param contents: File contents.
>> + """
>> + file_object = open(file_path, 'w')
>> + file_object.write(contents)
>> + file_object.close()
>> +
>> +
>> +def get_results_dir_list(pid, core_dir_basename):
>> + """
>> + Get all valid output directories for the core file and the
>> report. It works
>> + by inspecting files created by each test on /tmp and verifying if
>> the
>> + PID of the process that crashed is a child or grandchild of the
>> autotest
>> + test process. If it can't find any relationship (maybe a daemon
>> that died
>> + during a test execution), it will write the core file to the
>> debug dirs
>> + of all tests currently being executed. If there are no active
>> autotest
>> + tests at a particular moment, it will return a list with
>> ['/tmp'].
>> +
>> + @param pid: PID for the process that generated the core
>> + @param core_dir_basename: Basename for the directory that will
>> hold both
>> + the core dump and the crash report.
>> + """
>> + # Get all active test debugdir path files present
>> + debugdir_files = glob.glob("/tmp/autotest_results_dir.*")
>> + if debugdir_files:
>> + pid_dir_dict = {}
>> + for debugdir_file in debugdir_files:
>> + a_pid = debugdir_file.split('.')[-1]
>> + results_dir = open(debugdir_file, 'r').read().strip()
>> + pid_dir_dict[a_pid] = os.path.join(results_dir,
>> core_dir_basename)
>> +
>> + results_dir_list = []
>> + found_relation = False
>> + for a_pid, a_path in pid_dir_dict.iteritems():
>> + if pid_descends_from(pid, a_pid):
>> + results_dir_list.append(a_path)
>> + found_relation = True
>> +
>> + # If we could not find any relations between the pids in the
>> list with
>> + # the process that crashed, we can't tell for sure which test
>> spawned
>> + # the process (maybe it is a daemon and started even before
>> autotest
>> + # started), so we will have to output the core file to all
>> active test
>> + # directories.
>> + if not found_relation:
>> + return pid_dir_dict.values()
>> + else:
>> + return results_dir_list
>> +
>> + else:
>> + path_inactive_autotest = os.path.join('/tmp',
>> core_dir_basename)
>> + return [path_inactive_autotest]
>
> Here's a slightly shorter implementation of this function that doesn't
> need pid_descends_from():
>
> pid_dir_dict = {}
> for debugdir_file in glob.glob("/tmp/autotest_results_dir.*"):
> a_pid = os.path.splitext(debugdir_file)[1]
> results_dir = open(debugdir_file).read().strip()
> pid_dir_dict[a_pid] = os.path.join(results_dir, core_dir_basename)
>
> results_dir_list = []
> while pid > 1:
> if pid_dir_dict.has_key(pid):
> results_dir_list.append(pid_dir_dict[pid])
> pid = get_parent_pid(pid)
>
> return (results_dir_list or
> pid_dir_dict.values() or
> [os.path.join("/tmp", core_dir_basename)])
>
> It's not very different from your version -- I just find it a little
> simpler.
Ok, Implemented with the comments John made.
>> +def get_info_from_core(path):
>> + """
>> + Reads a core file and extracts a dictionary with useful core
>> information.
>> + Right now, the only information extracted is the full executable
>> name.
>> +
>> + @param path: Path to core file.
>> + """
>> + # Here we are getting the executable full path in a very
>> inelegant way :(
>> + # Since the 'right' solution for it is to make a library to get
>> information
>> + # from core dump files, properly written, I'll leave this as it
>> is for now.
>> + full_exe_path = commands.getoutput('strings %s | grep "_="' %
>> + path).strip("_=")
>
> If you think that's unelegant you can try using regular expressions,
> but then it might be slightly tricky to match only printable characters.
> You can also use filter(lambda x: x in string.printable, str), but that
> would return all printable strings concatenated, without newlines
> separating them (unless there are newlines in the core file itself).
>
>> + if full_exe_path.startswith("./"):
>
> I'm not sure, but it might be safer to use os.path.isabs().
> If an exe path is relative will it always be prefixed by "./" in the core
> file, or can the binary name appear without any prefix?
The session marked with _= contains the command line, I am not aware
of a case where there's nothing added to the prefix, so I left it the
way it was. If it turns out that this is not good enough, I will
change it. I was talking to Paul Muldoon, one of the frysk authors and
we might try to do a library for getting info from core files...
Thanks for your comments, Michael, I've sent an updated version for evaluation!
--
Lucas
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
