As we expand this test for more virtio devices we will need to support
different feature sets. Add a mandatory op field to fetch the list of
features needed for the test itself.
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220802095010.3330793-22-alex.bennee@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
No device driver (which is what the qvirtio_ access functions
represent) should be setting UNUSED(30) in the feature space. Although
existing libqos users mask it out lets ensure nothing sneaks through.
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220802095010.3330793-20-alex.bennee@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The vhost-user tests respawn qos-test as a standalone process. As a
result the gtester framework squashes all messages coming out of it
which make it hard to debug. As the test does not care about asserting
certain messages just convert the tests to use the direct qos_printf.
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220802095010.3330793-17-alex.bennee@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
These public functions are not used anywhere, thus can be dropped.
Also, since this is the final job API that doesn't use AioContext
lock and replaces it with job_lock, adjust all remaining function
documentation to clearly specify if the job lock is taken or not.
Also document the locking requirements for a few functions
where the second version is not removed.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20220926093214.506243-22-eesposit@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Change the job_{lock/unlock} and macros to use job_mutex.
Now that they are not nop anymore, remove the aiocontext
to avoid deadlocks.
Therefore:
- when possible, remove completely the aiocontext lock/unlock pair
- if it is used by some other function too, reduce the locking
section as much as possible, leaving the job API outside.
- change AIO_WAIT_WHILE in AIO_WAIT_WHILE_UNLOCKED, since we
are not using the aiocontext lock anymore
The only functions that still need the aiocontext lock are:
- the JobDriver callbacks, already documented in job.h
- job_cancel_sync() in replication.c is called with aio_context_lock
taken, but now job is using AIO_WAIT_WHILE_UNLOCKED so we need to
release the lock.
Reduce the locking section to only cover the callback invocation
and document the functions that take the AioContext lock,
to avoid taking it twice.
Also remove real_job_{lock/unlock}, as they are replaced by the
public functions.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Message-Id: <20220926093214.506243-19-eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Add missing job synchronization in the unit tests, with
explicit locks.
We are deliberately using _locked functions wrapped by a guard
instead of a normal call because the normal call will be removed
in future, as the only usage is limited to the tests.
In other words, if a function like job_pause() is/will be only used
in tests to avoid:
WITH_JOB_LOCK_GUARD(){
job_pause_locked();
}
then it is not worth keeping job_pause(), and just use the guard.
Note: at this stage, job_{lock/unlock} and job lock guard macros
are *nop*.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20220926093214.506243-10-eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Almost all invocations of run-test have either "$* on $(TARGET_NAME)"
or "$< on $(TARGET_NAME)" as the last argument. So provide a default
test name, while allowing an escape hatch for custom names.
As an additional simplification, remove the need to do shell quoting.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220929114231.583801-24-alex.bennee@linaro.org>
Make tests/tcg/ppc64le include tests/tcg/ppc64 instead of duplicating
the rules. Because the ppc64le vpath includes tests/tcg/ppc64 but
not vice versa, the tests have to be moved from tests/tcg/ppc64le/
to tests/tcg/ppc64.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220929114231.583801-23-alex.bennee@linaro.org>
Instead of linking tests/tcg/Makefile.target into the build tree, name
the symbolic link "Makefile" and create it in every target subdirectory.
This makes it possible to just invoke "make" in tests/tcg subdirectories.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220929114231.583801-21-alex.bennee@linaro.org>
Remove the DOCKER_SCRIPT and TARGET variable from the Makefile invocation
for tests/tcg. For DOCKER_SCRIPT, resolve the path to docker.py in configure;
for TARGET, move it to config-$(TARGET).mak and use a symbolic link to break
the cycle.
The symbolic link is still needed because tests/tcg includes dummy config files
for targets that are not buildable. Once that is cleaned up, the symbolic link
will go away too.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220929114231.583801-20-alex.bennee@linaro.org>
Tracking alpine-edge like debian-sid is a moving target. Usually such
rolling releases are marked as "allow_failure: true" in our CI.
However as alpine presents a musl based distro and provides useful
extra coverage lets track a release branch instead to avoid random
breakages.
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220929114231.583801-11-alex.bennee@linaro.org>
Note, the glib2-native mapping exists separately from the normal glib2
mapping. The latter uses a `foreign` cross-policy-default, and
libvirt-ci is not able to support package mappings for multiple
cross-compilation policies.
This will probably change in the future.
Signed-off-by: Alessandro Di Federico <ale@rev.ng>
Signed-off-by: Paolo Montesel <babush@rev.ng>
Signed-off-by: Anton Johansson <anjo@rev.ng>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220804115548.13024-9-anjo@rev.ng>
Message-Id: <20220929114231.583801-7-alex.bennee@linaro.org>
Change 'tmp_path' into an array of two members to accommodate another
disk image of size TEST_IMAGE_SIZE. This facilitates testing ATA
protocol aspects peculiar to secondary devices on the same controller.
Signed-off-by: Lev Kujawski <lkujaw@member.fsf.org>
Message-Id: <20220707031140.158958-2-lkujaw@member.fsf.org>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Add a new test to see what happens when you migrate a VM with a backing
chain that has json:{} backing file strings, which, when opened, will be
resolved to plain filenames.
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20220803144446.20723-4-hreitz@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Commit 5f76a7aac1 is looking harmless from
the first glance, but it has changed things a lot. 'libvirt' uses it to
detect that it should follow new initialization way and this changes
things considerably. With this procedure followed, blockdev_init() is
not called anymore and thus block_acct_setup() helper is not called.
This means in particular that defaults for block accounting statistics
are changed and account_invalid/account_failed are actually initialized
as false instead of true originally.
This commit changes things to match original world. There are the following
constraints:
* new default value in block_acct_init() is set to true
* block_acct_setup() inside blockdev_init() is called before
blkconf_apply_backend_options()
* thus newly created option in block device properties has precedence if
specified
Signed-off-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
CC: Peter Krempa <pkrempa@redhat.com>
CC: Markus Armbruster <armbru@redhat.com>
CC: John Snow <jsnow@redhat.com>
CC: Kevin Wolf <kwolf@redhat.com>
CC: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20220824095044.166009-3-den@openvz.org>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
While the source directory is always included in the include path,
the corresponding directory in the build tree is not. Therefore,
custom_targets (e.g. ui/dbus-display1.h) must be referred to using
the full path.
This avoids a build failure when ui/dbus-chardev.c is not built as
a module:
In file included from ../ui/dbus-chardev.c:32:
../ui/dbus.h:34:10: fatal error: dbus-display1.h: No such file or directory
34 | #include "dbus-display1.h"
| ^~~~~~~~~~~~~~~~~
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This qtest executable created a serial chardev file to be passed to
the QEMU executable. The serial file was created by g_file_open_tmp(),
which internally opens the file with FILE_SHARE_WRITE security attribute
on Windows. Based on [1], there is only one case that allows the first
call to CreateFile() with GENERIC_READ & FILE_SHARE_WRITE, and second
call to CreateFile() with GENERIC_WRITE & FILE_SHARE_READ. All other
combinations require FILE_SHARE_WRITE in the second call. But there is
no way for the second call (in this case the QEMU executable) to know
what combination was passed to the first call, unless FILE_SHARE_WRITE
is passed to the second call.
Two processes shouldn't share the same file for writing with a chardev.
Let's close the serial file before starting QEMU.
[1] https://docs.microsoft.com/en-us/windows/win32/fileio/creating-and-opening-files
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Message-Id: <20220927110632.1973965-40-bmeng.cn@gmail.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Some migration test cases use TLS to communicate, but they fail on
Windows with the following error messages:
qemu-system-x86_64: TLS handshake failed: Insufficient credentials for that request.
qemu-system-x86_64: TLS handshake failed: Error in the pull function.
query-migrate shows failed migration: TLS handshake failed: Error in the pull function.
Disable them temporarily.
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Message-Id: <20220925113032.1949844-51-bmeng.cn@gmail.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
close() is a *nix function. It works on any file descriptor, and
sockets in *nix are an example of a file descriptor.
closesocket() is a Windows-specific function, which works only
specifically with sockets. Sockets on Windows do not use *nix-style
file descriptors, and socket() returns a handle to a kernel object
instead, so it must be closed with closesocket().
In QEMU there is already a logic to handle such platform difference
in os-posix.h and os-win32.h, that:
* closesocket maps to close on POSIX
* closesocket maps to a wrapper that calls the real closesocket()
on Windows
Replace the call to close a socket with closesocket() instead.
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20220925113032.1949844-46-bmeng.cn@gmail.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
By default Windows opens file in text mode, while a POSIX compliant
implementation treats text files and binary files the same.
The fopen() 'mode' string can include the letter 'b' to indicate
binary mode shall be used. POSIX spec says the character 'b' shall
have no effect, but is allowed for ISO C standard conformance.
Let's add the letter 'b' which works on both POSIX and Windows.
Signed-off-by: Xuzhou Cheng <xuzhou.cheng@windriver.com>
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20220925113032.1949844-41-bmeng.cn@gmail.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Single quotes in the arguments (oem_id='CRASH ') are not removed in
the Windows environment before it is passed to the QEMU executable.
The space in the argument causes the "-acpitable" option parser to
think that all of its parameters are done, hence it complains:
'-acpitable' requires one of 'data' or 'file'
Change to use double quotes which works fine on all platforms.
Also /dev/null does not work on win32, and nul should be used.
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20220925113032.1949844-39-bmeng.cn@gmail.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
libqmp.c::qmp_fd_vsend_fds() is not available on Windows, hence any
APIs in libqtest that call libqmp.c::qmp_fd_vsend_fds() should be
excluded for win32 too. This includes the following:
* qtest_qmp_vsend_fds()
* qtest_vqmp_fds()
* qtest_qmp_fds()
* qtest_qmp_add_client()
Note qtest_qmp_vsend() was wrongly written to call qmp_fd_vsend_fds()
previously, but it should call the non fds version API qmp_fd_vsend().
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20220925113032.1949844-35-bmeng.cn@gmail.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
