This allows all VMs to be used at the end of a test sequence by
assigning test cases to VMs based on which VM is available for a new
test case rather than splitting the full task at the beginning and
potentially getting stuck with the last VM running long test cases for
significantly longer than another VM that gets shorter duration tests
assigned to it.
Signed-off-by: Jouni Malinen <j@w1.fi>
Previously, it was possible for a kernel panic to be missed since the
only sign of it in stdout was reduced number of passed test cases.
Signed-off-by: Jouni Malinen <j@w1.fi>
This avoids possible mismatches in directory and log file timestamps if
the UNIX timestamp (seconds) changes during the startup sequence.
Signed-off-by: Jouni Malinen <j@w1.fi>
This was currently breaking parallel-run.*, as it was passing
--split num/num parameters (intended for rnu-tests.py)
to vm-run.sh which broke the --codecov and --timewrap options.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Update the code coverage documentation to also specify the
source base directory for the code coverage generation.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
When loading the hwsim module, disable support_p2p_device by default.
This will also become the default in the kernel, but until then it
makes sure it's not turned on by default.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
This is a more advanced version of the simple parallel-vm.sh script.
Status of each VM is printed out during the test and results are
provided in more convenient format in the end.
Signed-off-by: Jouni Malinen <j@w1.fi>
"parallel-vm.sh <number of VMs> [arguments..]" can now be used to run
multiple VMs in parallel to speed up full test cycle significantly. In
addition, the "--split srv/total" argument used in this design would
also make it possible to split this to multiple servers to speed up
testing.
Signed-off-by: Jouni Malinen <j@w1.fi>
This improves accuracy of the code coverage reports with hostapd-as-AS
and hlr_auc_gw getting analyzed separately.
Signed-off-by: Jouni Malinen <j@w1.fi>
To test the code under the influence of time jumps, add the option
(--timewarp) to the VM tests to reset the clock all the time, which
makes the wall clock time jump speed up 20x, causing gettimeofday()
to be unreliable for timeout calculations.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
The vm-config in the subdirectory is less useful as it
will get removed by "git clean" and similar, so read a
config file from ~/.wpas-vm-config in addition.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Use a more robust design for collecting the gcov logs from the case
where test cases are run within a virtual machine. This generates a
writable-from-vm build tree for each component separately so that the
lcov and gcov can easily find the matching source code and data files.
In addition, prepare the reports automatically at the end of the
vm-run.sh --codecov execution.
Signed-hostap: Jouni Malinen <j@w1.fi>
In order to handle regulatory domain requests, crda needs to be
installed on the host, but we also need to install a uevent helper in
the VM so that it gets executed (since we don't run udev).
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
If there's code coverage analysis data, copy it out of the VM
to be able to analyse it later. Also add a description to the
README file about how to use it.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
Add a CHANNELS configuration to the script running the VM
that can be added to the vm-config file to allow running
the tests with hwsim devices supporting more than a single
channel.
Eventually, with the (hopefully) upcoming dynamic work in
mac80211_hwsim, this might go away entirely, but for now
this allows testing more code paths.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
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>
Create a results.db in the output directory when running
the tests in a VM. To make that easier, create the tables
in the python script if they don't exist.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
Rather than just having KERNELDIR, allow setting KERNEL directly.
Also remove the -s option that prevents running multiple machines
at the same time, but add a KVMARGS= variable that can be used to
restore that if needed.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
Instead of running on the host, it can be useful to run in a
VM, particularly to test kernel rather than userspace changes,
so add a few scripts that allow doing so easily.
The basic idea is that the VM kernel is the same architecture
as the host kernel, so the host's root filesystem can be used
(in read-only mode) to run everything. Only a log filesystem
is mounted read-write and will get all the test output.
The kernel console output is collected to a special 'console'
file in the logs directory and kernel crashes are detected.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
