def run_module():
logfile = 'data.log'
os.system('rm -f ./data.log')
# Insert the module
print "Inserting module..."
stap_cmd = [ "staprun", "-o", "data.log", "probe_module.ko" ]
stap_pid = os.spawnvp(os.P_NOWAIT, stap_cmd[0], stap_cmd)
print "stap pid is", stap_pid
# Run the load generate commands (if any). Note we ignore the
# return values, since we don't really care if the commands
# succeed or not.
if config_opts.has_key('load_cmds'):
print "Running commands..."
# Redirect the output to 'run_module.log' by modifying
# what stdout/stderr points to.
sys.stdout.flush()
old_stdout = os.dup(sys.stdout.fileno())
sys.stderr.flush()
old_stderr = os.dup(sys.stderr.fileno())
fd = os.open('run_module.log', os.O_CREAT | os.O_WRONLY)
os.dup2(fd, sys.stdout.fileno())
os.dup2(fd, sys.stderr.fileno())
# Run the commands.
pid_list = list()
for cmd in config_opts['load_cmds']:
pid = os.spawnvp(os.P_NOWAIT, cmd[0], cmd)
pid_list.append(pid)
for pid in pid_list:
(childpid, status) = os.waitpid(pid, 0)
# Restore old value of stdout/stderr.
os.close(fd)
os.dup2(old_stdout, sys.stdout.fileno())
os.dup2(old_stderr, sys.stderr.fileno())
else:
time.sleep(10)
# Remove the module
print "Removing module..."
os.kill(stap_pid, signal.SIGINT)
(childpid, rc) = os.waitpid(stap_pid, 0)
#if os.WEXITSTATUS(rc) != 0:
# print >>sys.stderr, "Error: rmmod failed"
# return -1
# Now we have to wait until everything is flushed to the logfile
print "Looking for output..."
rc = 0
f = open(config_opts['probes_result'], 'w')
attempts = 0
start_pos = 0
while 1:
l = open(logfile, 'r')
# Find the ending size of the logfile
end_pos = os.path.getsize(logfile)
# Try to wait until data is available
while 1:
try:
input, output, exc = select.select([l.fileno()], [], [], 60)
break
except select.error, err:
if err[0] != EINTR:
raise
# Get the new stuff logged
data = l.read(end_pos - start_pos + 1)
if not data:
# ignore EOF
time.sleep(2)
attempts += 1
if attempts < 30:
l.close()
continue
else:
print >>sys.stderr, "Error: Couldn't find data"
rc = -1
break;
# Write results data
f.write(data)
# See if we can find 'probe_module unloaded' in the data we
# just read.
if data.find('probe_module unloaded') == -1:
start_pos = end_pos
else:
break
def run_module():
logfile = 'data.log'
os.system('rm -f ./data.log')
# Insert the module
print "Inserting module..."
stap_cmd = ["staprun", "-o", "data.log", "probe_module.ko"]
stap_pid = os.spawnvp(os.P_NOWAIT, stap_cmd[0], stap_cmd)
print "stap pid is", stap_pid
# Run the load generate commands (if any). Note we ignore the
# return values, since we don't really care if the commands
# succeed or not.
if config_opts.has_key('load_cmds'):
print "Running commands..."
# Redirect the output to 'run_module.log' by modifying
# what stdout/stderr points to.
sys.stdout.flush()
old_stdout = os.dup(sys.stdout.fileno())
sys.stderr.flush()
old_stderr = os.dup(sys.stderr.fileno())
fd = os.open('run_module.log', os.O_CREAT | os.O_WRONLY)
os.dup2(fd, sys.stdout.fileno())
os.dup2(fd, sys.stderr.fileno())
# Run the commands.
pid_list = list()
for cmd in config_opts['load_cmds']:
pid = os.spawnvp(os.P_NOWAIT, cmd[0], cmd)
pid_list.append(pid)
for pid in pid_list:
(childpid, status) = os.waitpid(pid, 0)
# Restore old value of stdout/stderr.
os.close(fd)
os.dup2(old_stdout, sys.stdout.fileno())
os.dup2(old_stderr, sys.stderr.fileno())
else:
time.sleep(10)
# Remove the module
print "Removing module..."
os.kill(stap_pid, signal.SIGINT)
(childpid, rc) = os.waitpid(stap_pid, 0)
#if os.WEXITSTATUS(rc) != 0:
# print >>sys.stderr, "Error: rmmod failed"
# return -1
# Now we have to wait until everything is flushed to the logfile
print "Looking for output..."
rc = 0
f = open(config_opts['probes_result'], 'w')
attempts = 0
start_pos = 0
while 1:
l = open(logfile, 'r')
# Find the ending size of the logfile
end_pos = os.path.getsize(logfile)
# Try to wait until data is available
while 1:
try:
input, output, exc = select.select([l.fileno()], [], [], 60)
break
except select.error, err:
if err[0] != EINTR:
raise
# Get the new stuff logged
data = l.read(end_pos - start_pos + 1)
if not data:
# ignore EOF
time.sleep(2)
attempts += 1
if attempts < 30:
l.close()
continue
else:
print >> sys.stderr, "Error: Couldn't find data"
rc = -1
break
# Write results data
f.write(data)
# See if we can find 'probe_module unloaded' in the data we
# just read.
if data.find('probe_module unloaded') == -1:
start_pos = end_pos
else:
break
def run_module():
# Find the current size of /var/log/messages
logfile = "/var/log/messages"
print >>sys.stderr, "Getting size of %s..." % logfile
start_pos = os.path.getsize(logfile)
l = open(logfile, "r")
l.seek(start_pos)
# Insert the module
print >>sys.stderr, "Inserting module..."
rc = os.system("/sbin/insmod %s" % os.path.join(os.getcwd(), "kprobe_module.ko"))
if os.WEXITSTATUS(rc) != 0:
# This is actually semi-OK, which is why there is no error
# message here, only a notice. This might mean that every
# probe tried cannot be registered.
print >>sys.stderr, "Notice: insmod failed"
else:
# Run the load generate commands (if any). Note we ignore the
# return values, since we don't really care if the commands
# succeed or not.
if config_opts.has_key("load_cmds"):
print >>sys.stderr, "Running commands..."
num_waits = 0
# Redirect the output to 'run_module.log' by modifying
# what stdout/stderr points to.
sys.stdout.flush()
old_stdout = os.dup(sys.stdout.fileno())
sys.stderr.flush()
old_stderr = os.dup(sys.stderr.fileno())
fd = os.open("run_module.log", os.O_CREAT | os.O_WRONLY)
os.dup2(fd, sys.stdout.fileno())
os.dup2(fd, sys.stderr.fileno())
# Run the commands.
pid_list = list()
for cmd in config_opts["load_cmds"]:
pid = os.spawnvp(os.P_NOWAIT, cmd[0], cmd)
pid_list.append(pid)
for pid in pid_list:
(childpid, status) = os.waitpid(pid, 0)
# Restore old value of stdout/stderr.
os.close(fd)
os.dup2(old_stdout, sys.stdout.fileno())
os.dup2(old_stderr, sys.stderr.fileno())
# Remove the module
print >>sys.stderr, "Removing module..."
rc = os.system("/sbin/rmmod kprobe_module")
if os.WEXITSTATUS(rc) != 0:
print >>sys.stderr, "Error: rmmod failed"
return -1
# Now we have to wait until everything is flushed to the logfile
print >>sys.stderr, "Looking for output..."
rc = 0
f = open(config_opts["probes_result"], "w")
attempts = 0
while 1:
# Find the ending size of /var/log/messages
end_pos = os.path.getsize(logfile)
if end_pos < start_pos:
# The log files have been rotated. Read any leftover data,
# then reopen file.
data = l.read()
if data:
f.write(data)
# See if we can find 'kprobe_module unloaded' in the
# data we just read.
if data.find("kprobe_module unloaded") != -1:
break
l.close()
l = open(logfile, "r")
start_pos = 0
continue
# Try to wait until data is available
while 1:
try:
input, output, exc = select.select([l.fileno()], [], [], 60)
break
except select.error, err:
if err[0] != EINTR:
raise
# Get the new stuff logged to /var/log/messages
data = l.read(end_pos - start_pos + 1)
if not data:
# ignore EOF
time.sleep(2)
attempts += 1
if attempts < 30:
continue
else:
print >>sys.stderr, "Error: Couldn't find data"
rc = -1
break
# Write results data
f.write(data)
# See if we can find 'kprobe_module unloaded' in the data we
# just read.
if data.find("kprobe_module unloaded") == -1:
start_pos = end_pos
else:
break
def run_module():
# Find the current size of /var/log/messages
logfile = '/var/log/messages'
print >>sys.stderr, "Getting size of %s..." % logfile
start_pos = os.path.getsize(logfile)
l = open(logfile, 'r')
l.seek(start_pos)
# Insert the module
print >>sys.stderr, "Inserting module..."
rc = os.system("/sbin/insmod %s" % os.path.join(os.getcwd(),
'kprobe_module.ko'))
if os.WEXITSTATUS(rc) != 0:
# This is actually semi-OK, which is why there is no error
# message here, only a notice. This might mean that every
# probe tried cannot be registered.
print >>sys.stderr, "Notice: insmod failed"
else:
# Run the load generate commands (if any). Note we ignore the
# return values, since we don't really care if the commands
# succeed or not.
if config_opts.has_key('load_cmds'):
print >>sys.stderr, "Running commands..."
num_waits = 0
# Redirect the output to 'run_module.log' by modifying
# what stdout/stderr points to.
sys.stdout.flush()
old_stdout = os.dup(sys.stdout.fileno())
sys.stderr.flush()
old_stderr = os.dup(sys.stderr.fileno())
fd = os.open('run_module.log', os.O_CREAT | os.O_WRONLY)
os.dup2(fd, sys.stdout.fileno())
os.dup2(fd, sys.stderr.fileno())
# Run the commands.
for cmd in config_opts['load_cmds']:
pid = os.spawnvp(os.P_NOWAIT, cmd[0], cmd)
num_waits += 1
while num_waits > 0:
(pid, status) = os.waitpid(-1, 0)
num_waits -= 1
# Restore old value of stdout/stderr.
os.close(fd)
os.dup2(old_stdout, sys.stdout.fileno())
os.dup2(old_stderr, sys.stderr.fileno())
# Remove the module
print >>sys.stderr, "Removing module..."
rc = os.system("/sbin/rmmod kprobe_module")
if os.WEXITSTATUS(rc) != 0:
print >>sys.stderr, "Error: rmmod failed"
return -1
# Now we have to wait until everything is flushed to the logfile
f = open(config_opts['probes_result'], 'w')
while 1:
# Find the ending size of /var/log/messages
end_pos = os.path.getsize(logfile)
if end_pos < start_pos:
# The log files have been rotated. Read any leftover data,
# then reopen file.
data = l.read()
if data:
f.write(data)
# See if we can find 'kprobe_module unloaded' in the
# data we just read.
if data.find('kprobe_module unloaded') != -1:
break
l.close()
l = open(logfile, 'r')
start_pos = 0
continue
# Try to wait until data is available
while 1:
try:
input, output, exc = select.select([l.fileno()], [], [], 60)
break
except select.error, err:
if err[0] != EINTR:
raise
# Get the new stuff logged to /var/log/messages
data = l.read(end_pos - start_pos + 1)
if not data:
# ignore EOF
time.sleep(2)
continue
# Write results data
f.write(data)
# See if we can find 'kprobe_module unloaded' in the data we
# just read.
if data.find('kprobe_module unloaded') == -1:
start_pos = end_pos
else:
break
