It is possible to execute the tests directly inside the a VM without
necessarily running the repository wrapper script to do so. In these
cases, passing the VM flag using an environment variable allows
correctly flagging this fact to the tests.
Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>
git has started rejecting repositories owned by other users and refusing
to run the "git rev-parse HEAD" command in this type of cases. That
resulted in issues with the VM testing model where the VM is practically
running everything as root while the host is a normal development
environment and likely a non-root user owned files.
Fix this by fetching the commitid on the host and pass it to the VM so
that no git operations need to be run within the VM itself.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
Do not add the --commit argument if the current git commitid cannot be
determined. This prevents complete failure to run the tests if the git
command cannot be used for some reason (like a recent change that
stopped allowing root user within the VM from running the git operation
for the case where the host system uses non-root account).
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
There is no point in building this tarball in /tmp that is on the
ramdisk of the VM since it will go away when the VM exits.
Signed-off-by: Jouni Malinen <j@w1.fi>
The hwsim's start.sh script spawns hostapd process using "sudo".
Since sudo forks a child process, $! holds the pid of sudo itself.
Fix that by storing the PID of the child process instead.
Since in VM "sudo" is replaced with a dummy script, pass an additional
argument to run-all.sh and start.sh scripts to indicate that they are
running inside a VM.
This is needed to fix ap_config_reload and ap_config_reload_file test
cases on some platforms where sudo is apparently not relaying the
signals properly.
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
The number of channels was not properly passed from the
run-all.sh script to the start.sh script. Fix it.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Add an option --build to run-all.sh to build before starting to run all
the tests. In addition, add an option --codecov to extract the code
coverage data at the end of the run.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
This will be needed to be able to control dynamic mac80211_hwsim
operations like adding and removing a phy. In the past, it has been
possible to start the main programs as root and then use non-root
account for run-tests.py. However, there is already a large number of
cases within the test scripts where sudo is needed. Moving that
requirement to execution of run-tests.py allows those to be simplified
as well.
Signed-off-by: Jouni Malinen <j@w1.fi>
Add an option to run-all.sh and start.sh to get as an argument the
number of concurrent channels that mac80211_hwsim will be loaded with.
To start mac80211_hwsim with more than one channel, the following
parameter should be added to command line: channels=<num_channels>
The default is one channel (no concurrent channels).
The driver should be loaded with multi channel support in order to run
some tests.
Signed-off-by: Haim Dreyfuss <haim.dreyfuss@intel.com>
These were not really used in practice and better quality test cases for
concurrent P2P operations are now available in tests_p2p_concurrency.py
using the standard test framework design. Remove the special concurrent
option for start.sh and the run-all.sh concurrent and
concurrent-valgrind operations to clean up scripts.
Signed-hostap: Jouni Malinen <j@w1.fi>
Remove the -l command like option from run-tests.py and always enable
writing of debug level logs to files. The stdout debug verbosity is
controlled independently of the debug log files.
Signed-hostap: Jouni Malinen <j@w1.fi>
In some cases, e.g., with the VM tests if the VM crashes, it
can be useful to know which tests should have run but didn't
(or didn't finish). In order to catch these more easily, add
an option to prefill the database with all tests at the very
beginning of the testing (in a new NOTRUN state) and use the
option in the VM tests.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
The run-tests.py -l argument does not take an argument value anymore.
Instead, debug output is directed to a separate file <test>.log for each
test case.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
This is unnecessary extra complexity for user, so use the 'test_' prefix
only internally within the python scripts and file names.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
This file was used for adding debug info into the buildbot logs. The new
sqlite database-based design will replace that, so there is no need to
create last-debug and getting rid of it will make it easier to split the
run log into per-test case files.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
In addition to tracing, allow collecting dmesg. There's no
provision for actually looking at it and finding problems
in it yet though.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
Instead of passing the log directory for each option
(-l, -r, -e, and -T) pass it once and make the other
options just take the filename (optionally, even).
This will also make it easier to extend later.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
Instead of logging many files called "<timestamp>-*", log into
a new "<timestamp>/*" in the logs/ subdirectory and only put
the last-debug file into logs/. If a LOGDIR is specified in the
environment, instead just put everything into that directory
(so the caller should make sure to give it a timestamp or so)
and skip the creation of last-debug entirely.
Also clean up a bit and pass the LOGDIR from run-all.sh to
start.sh rather than having start.sh create the timestamp and
run-all.sh detect the latest one when having run start.sh.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
In order to get tracing per test, allow run-tests.py to start
and stop tracing per test case. This is implemented using a
python 'with' context so it starts/stops automatically at the
right spots.
Instead of starting global tracing, also use it from run-all.sh
and put the trace files into the log dir.
Note that this only works right if you use a separate log dir
for all test runs as the trace files aren't timestamped.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
Allow run-all to pass on extra arguments from the command
line that aren't built by the script itself.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
Reuse the code rather than duplicating the implementation
of starting the tests. To make that easier, allow passing
multiple modules with -f to run-tests.py.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
Instead of re-implementing a command-line parser, use the
argparse module.
The only real change (I hope) is that the test module must
now be given to the -f option without the .py suffix.
Also, --help now works, and if a test module/test name is
given that doesn't exist, the valid list is printed.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
Instead of hardcoding logs/, allow putting LOGDIR into
the environment. This helps run tests in a VM where the
rootfs might be mounted read-only.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
There is no point trying to go through the tests if hostapd and
wpa_supplicant processes cannot be started properly.
Signed-hostap: Jouni Malinen <j@w1.fi>
run-all.sh and start.sh scripts can now take 'trace' command line
argument to request Linux tracing information from mac80211, cfg80211,
and wpa_supplicant to be recorded.
Signed-hostap: Jouni Malinen <j@w1.fi>
The previous design of having to provide valgrind parameter to
stop-wifi.sh did not work since the new test setup may no match the
previous one. Instead of trying to figure out what needs to be stopped
based on previous run, stop all wpa_supplicant/hostapd processes running
under valgrind unconditionally to avoid cases where new set of processes
cannot be started due to existing processed.
Signed-hostap: Jouni Malinen <j@w1.fi>
This makes it easier to figure out what failed and allows builbot to
integrate multiple logs and state information about the test cases.
Signed-hostap: Jouni Malinen <j@w1.fi>
./run-all.sh can now take an optional argument to select whether to run
all test cases as before (default), to run these under valgrind
("valgrind"), to run P2P test cases with concurrent station interface
("concurrent"), or the concurrent P2P test cases under valgrind
("concurrent-valgrind").
valgrind cases report errors if a test case fails or valgrind reports an
error.
Signed-hostap: Jouni Malinen <j@w1.fi>
This removes the unnecessary separation of P2P (no hostapd) and AP
tests. The same scripts can be used to prepare for these tests and to
execute the tests.
Signed-hostap: Jouni Malinen <j@w1.fi>
buildbot seems to be running with more limited umask by default and that
prevented the hostapd per-interface socket from being accessed.
Signed-hostap: Jouni Malinen <j@w1.fi>
