KVM test: kvm_subprocess: use read_nonblocking(0) instead of read_nonblocking (0.1)
PATCH 1/6 KVM test: kvm_subprocess: use read_nonblocking(0) instead of read_nonblocking (0.1)
In get_command_status_output() and is_responsive() use read_nonblocking(0) to read the unread output before sending input (e.g. a command).
The timeout is currently 0.1 because theoretically it should help if the guest still produces output when the function is called, but in practice there's no guarantee that a value of 0.1 will suffice. Therefore, it is be the user's responsibility to make sure the guest stopped producing output before get_command_status_output() is called. This can be guaranteed (in most cases) by using get_command_status_output() and friends instead of sendline() to send commands (because the former waits for the prompt to return, whereas the latter returns immediately).
--- client/tests/kvm/kvm_subprocess.py | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-)
client/tests/kvm/kvm_subprocess.py +2 -2
@@ -1028,7 +1028,7 @@ class kvm_shell_session(kvm_expect): """ # Read all output that's waiting to be read, to make sure the output # we read next is in response to the newline sent - self.read_nonblocking(timeout=0.1) + self.read_nonblocking(timeout=0) # Send a newline self.sendline() # Wait up to timeout seconds for some output from the child
@@ -1095,7 +1095,7 @@ class kvm_shell_session(kvm_expect): logging.debug("Sending command: %s" % command) # Read everything that's waiting to be read - self.read_nonblocking(0.1) + self.read_nonblocking(timeout=0) # Send the command and get its output self.sendline(command)
Signature · 1 line
1.5.4.1
PATCH 2/6 KVM test: AutoIt and Autotest wrappers: use get_command_output() instead of sendline()
get_command_output() is safer because it waits for the prompt to return. sendline() returns immediately, and the output generated in response to the command can appear later and interfere with the test.
For example, the prompt can appear while the Autotest wrapper waits for an Autotest test to complete, and this will make the wrapper exit as if the test has completed.
--- client/tests/kvm/tests/autoit.py | 7 +++---- client/tests/kvm/tests/autotest.py | 9 ++++----- 2 files changed, 7 insertions(+), 9 deletions(-)
client/tests/kvm/tests/autoit.py +3 -4
@@ -30,18 +30,17 @@ def run_autoit(test, params, env): # Send AutoIt script to guest (this code will be replaced once we # support sending files to Windows guests) - session.sendline("del script.au3") + session.get_command_output("del script.au3", internal_timeout=0) file = open(kvm_utils.get_path(test.bindir, script)) for line in file.readlines(): # Insert a '^' before each character line = "".join("^" + c for c in line.rstrip()) if line: # Append line to the file - session.sendline("echo %s>>script.au3" % line) + session.get_command_output("echo %s>>script.au3" % line, + internal_timeout=0) file.close() - session.read_up_to_prompt() - command = "cmd /c %s script.au3 %s" % (binary, script_params) logging.info("---------------- Script output ----------------")
client/tests/kvm/tests/autotest.py +4 -5
@@ -85,7 +85,7 @@ def run_autotest(test, params, env): extract(vm, "autotest.tar.bz2") # mkdir autotest/tests - session.sendline("mkdir autotest/tests") + session.get_command_output("mkdir autotest/tests") # Extract <test_name>.tar.bz2 into autotest/tests extract(vm, test_name + ".tar.bz2", "autotest/tests")
@@ -99,10 +99,9 @@ def run_autotest(test, params, env): # Run the test logging.info("Running test '%s'..." % test_name) - session.sendline("cd autotest") - session.sendline("rm -f control.state") - session.sendline("rm -rf results/*") - session.read_up_to_prompt() + session.get_command_output("cd autotest") + session.get_command_output("rm -f control.state") + session.get_command_output("rm -rf results/*") logging.info("---------------- Test output ----------------") status = session.get_command_status("bin/autotest control", timeout=test_timeout,
Signature · 1 line
1.5.4.1
PATCH 3/6 KVM test: kvm_vm.py: change qemu-img timeout to 120 seconds
The timeout of qemu-img commands is currently 30 seconds.
This may not suffice under heavy load (e.g. when multiple tests run in parallel and use the same physical disk).
--- client/tests/kvm/kvm_vm.py | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
client/tests/kvm/kvm_vm.py +1 -1
@@ -55,7 +55,7 @@ def create_image(params, root_dir): logging.debug("Running qemu-img command:\n%s" % qemu_img_cmd) (status, output) = kvm_subprocess.run_fg(qemu_img_cmd, logging.debug, - "(qemu-img) ", timeout=30) + "(qemu-img) ", timeout=120) if status is None: logging.error("Timeout elapsed while waiting for qemu-img command "
Signature · 1 line
1.5.4.1
PATCH 4/6 KVM test: step file tests: add parameter timeout_multiplier
This parameter multiplies the timeout values of all the barriers in a step file test.
It is useful for slower hosts, under load (e.g. when executing multiple tests in parallel) and for testing QEMU without KVM. In any of these cases, the multiplier should be greater than 1 in order to give the test more time to complete.
In addition to modifying tests/steps.py, this patch adds a usage example to control and control.parallel.
--- client/tests/kvm/control | 1 + client/tests/kvm/control.parallel | 1 + client/tests/kvm/tests/steps.py | 22 +++++++++++++++------- 3 files changed, 17 insertions(+), 7 deletions(-)
client/tests/kvm/control +1 -0
@@ -140,6 +140,7 @@ filename = os.path.join(pwd, "kvm_tests.cfg") cfg = kvm_config.config(filename) # If desirable, make changes to the test configuration here. For example: +# cfg.parse_string("install|setup: timeout_multiplier = 2") # cfg.parse_string("only fc8_quick") # cfg.parse_string("display = sdl")
client/tests/kvm/control.parallel +1 -0
@@ -136,6 +136,7 @@ filename = os.path.join(pwd, "kvm_tests.cfg") cfg = kvm_config.config(filename) # If desirable, make changes to the test configuration here. For example: +# cfg.parse_string("install|setup: timeout_multiplier = 2") # cfg.parse_string("only fc8_quick") # cfg.parse_string("display = sdl")
client/tests/kvm/tests/steps.py +15 -7
@@ -33,6 +33,21 @@ def barrier_2(vm, words, params, debug_dir, data_scrdump_filename, cmd, dx, dy, x1, y1, md5sum, timeout = words[:7] dx, dy, x1, y1, timeout = map(int, [dx, dy, x1, y1, timeout]) + scrdump_filename = os.path.join(debug_dir, "scrdump.ppm") + cropped_scrdump_filename = os.path.join(debug_dir, "cropped_scrdump.ppm") + expected_scrdump_filename = os.path.join(debug_dir, "scrdump_expected.ppm") + expected_cropped_scrdump_filename = os.path.join(debug_dir, + "cropped_scrdump_expected.ppm") + comparison_filename = os.path.join(debug_dir, "comparison.ppm") + + # Multiply timeout by the timeout multiplier + timeout_multiplier = params.get("timeout_multiplier") + if timeout_multiplier: + timeout_multiplier = float(timeout_multiplier) + else: + timeout_multiplier = 1.0 + timeout *= timeout_multiplier + # Timeout/5 is the time it took stepmaker to complete this step. # Divide that number by 10 to poll 10 times, just in case # current machine is stronger then the "stepmaker machine".
@@ -41,13 +56,6 @@ def barrier_2(vm, words, params, debug_dir, data_scrdump_filename, if sleep_duration < 1.0: sleep_duration = 1.0 if sleep_duration > 10.0: sleep_duration = 10.0 - scrdump_filename = os.path.join(debug_dir, "scrdump.ppm") - cropped_scrdump_filename = os.path.join(debug_dir, "cropped_scrdump.ppm") - expected_scrdump_filename = os.path.join(debug_dir, "scrdump_expected.ppm") - expected_cropped_scrdump_filename = os.path.join(debug_dir, - "cropped_scrdump_expected.ppm") - comparison_filename = os.path.join(debug_dir, "comparison.ppm") - fail_if_stuck_for = params.get("fail_if_stuck_for") if fail_if_stuck_for: fail_if_stuck_for = float(fail_if_stuck_for)
Signature · 1 line
1.5.4.1
PATCH 5/6 KVM test: tests/steps.py: simplify barrier_2()
Add a few comments, group similar lines and replace some blocks with one-liners.
Note: the weird formtimeout_multiplier = float(params.get("timeout_multiplier") or 1) allows the user to fall back to the default value by setting timeout_multiplier to "".
The more common formtimeout_multiplier = float(params.get("timeout_multiplier", 1)) will raise an exception if timeout_multiplier is "".
--- client/tests/kvm/tests/steps.py | 32 ++++++++++---------------------- 1 files changed, 10 insertions(+), 22 deletions(-)
client/tests/kvm/tests/steps.py +10 -22
@@ -30,22 +30,27 @@ def barrier_2(vm, words, params, debug_dir, data_scrdump_filename, logging.error("Bad barrier_2 command line") return False + # Parse barrier command line cmd, dx, dy, x1, y1, md5sum, timeout = words[:7] dx, dy, x1, y1, timeout = map(int, [dx, dy, x1, y1, timeout]) + # Define some paths scrdump_filename = os.path.join(debug_dir, "scrdump.ppm") cropped_scrdump_filename = os.path.join(debug_dir, "cropped_scrdump.ppm") expected_scrdump_filename = os.path.join(debug_dir, "scrdump_expected.ppm") expected_cropped_scrdump_filename = os.path.join(debug_dir, "cropped_scrdump_expected.ppm") comparison_filename = os.path.join(debug_dir, "comparison.ppm") + history_dir = os.path.join(debug_dir, "barrier_history") + + # Collect a few parameters + timeout_multiplier = float(params.get("timeout_multiplier") or 1) + fail_if_stuck_for = float(params.get("fail_if_stuck_for") or 1e308) + stuck_detection_history = int(params.get("stuck_detection_history") or 2) + keep_screendump_history = params.get("keep_screendump_history") == "yes" + keep_all_history = params.get("keep_all_history") == "yes" # Multiply timeout by the timeout multiplier - timeout_multiplier = params.get("timeout_multiplier") - if timeout_multiplier: - timeout_multiplier = float(timeout_multiplier) - else: - timeout_multiplier = 1.0 timeout *= timeout_multiplier # Timeout/5 is the time it took stepmaker to complete this step.
@@ -56,23 +61,6 @@ def barrier_2(vm, words, params, debug_dir, data_scrdump_filename, if sleep_duration < 1.0: sleep_duration = 1.0 if sleep_duration > 10.0: sleep_duration = 10.0 - fail_if_stuck_for = params.get("fail_if_stuck_for") - if fail_if_stuck_for: - fail_if_stuck_for = float(fail_if_stuck_for) - else: - fail_if_stuck_for = 1e308 - - stuck_detection_history = params.get("stuck_detection_history") - if stuck_detection_history: - stuck_detection_history = int(stuck_detection_history) - else: - stuck_detection_history = 2 - - keep_screendump_history = params.get("keep_screendump_history") == "yes" - if keep_screendump_history: - keep_all_history = params.get("keep_all_history") == "yes" - history_dir = os.path.join(debug_dir, "barrier_history") - end_time = time.time() + timeout end_time_stuck = time.time() + fail_if_stuck_for start_time = time.time()
Signature · 1 line
1.5.4.1
PATCH 6/6 KVM test: kvm_subprocess: use select() in read_until_output_matches()
Currently read_nonblocking() is called repeatedly until a match is found. This is fine as long as internal_timeout, the timeout parameter passed to read_nonblocking(), is greater than zero. If it equals zero the loop will keep the CPU busy and stress the host.
To avoid this, use select() to wait until there's output to read from the child process.
--- client/tests/kvm/kvm_subprocess.py | 9 +++++++-- 1 files changed, 7 insertions(+), 2 deletions(-)
client/tests/kvm/kvm_subprocess.py +7 -2
@@ -848,8 +848,12 @@ class kvm_expect(kvm_tail): match = None data = "" + fd = self._get_fd("expect") end_time = time.time() + timeout - while time.time() < end_time: + while True: + r, w, x = select.select([fd], [], [], + max(0, end_time - time.time())) + if fd not in r: break # Read data from child newdata = self.read_nonblocking(internal_timeout) # Print it if necessary
@@ -868,7 +872,8 @@ class kvm_expect(kvm_tail): done = True # Check if child has died if not self.is_alive(): - logging.debug("Process terminated with status %s" % self.get_status()) + logging.debug("Process terminated with status %s" % + self.get_status()) done = True # Are we done? if done: break
Signature · 1 line
1.5.4.1
Patchset applied!
Quoted 46 lines · Michael Goldish
In
get_command_status_output()andis_responsive()useread_nonblocking(0)to read the unread output before sending input (e.g. a command).
The timeout is currently 0.1 because theoretically it should help if the guest still produces output when the function is called, but in practice there's no guarantee that a value of 0.1 will suffice. Therefore, it is be the user's responsibility to make sure the guest stopped producing output beforeget_command_status_output()is called. This can be guaranteed (in most cases) by usingget_command_status_output()and friends instead ofsendline()to send commands (because the former waits for the prompt to return, whereas the latter returns immediately).Signed-off-by Michael Goldish <[email protected]>---
client/tests/kvm/kvm_subprocess.py | 4 ++--1 files changed, 2 insertions(+), 2 deletions(-)client/tests/kvm/kvm_subprocess.py +0
-0@@ -1028,7 +1028,7 @@ class kvm_shell_session(kvm_expect):"""
# Read all output that's waiting to be read, to make sure the output# we read next is in response to the newline sent - self.read_nonblocking(timeout=0.1) + self.read_nonblocking(timeout=0) # Send a newline self.sendline() # Wait up to timeout seconds for some output from the childclient/tests/kvm/kvm_subprocess.py +1
-2@@ -1095,7 +1095,7 @@ class kvm_shell_session(kvm_expect): logging.debug("Sending command: %s" % command) # Read everything that's waiting to be read - self.read_nonblocking(0.1) + self.read_nonblocking(timeout=0) # Send the command and get its output self.sendline(command) --1.5.4.1
_______________________________________________
Autotest mailing list
[email protected]
http://test.kernel.org/cgi-bin/mailman/listinfo/autotest
Signature · 1 line
Lucas