Add start_server() and accept() methods that can be used instead of
start_server_and_accept() to allow more fine-grained control over the
incoming connection process.
(Eagle-eyed reviewers will surely notice that it's a bit weird that
"CONNECTING" is a state that's shared between both the start_server()
and connect() states. That's absolutely true, and it's very true that
checking on the presence of _accepted as an indicator of state is a
hack. That's also very certainly true. But ... this keeps client code an
awful lot simpler, as it doesn't have to care exactly *how* the
connection is being made, just that it *is*. Is it worth disrupting that
simplicity in order to provide a better state guard on `accept()`? Hm.)
Signed-off-by: John Snow <jsnow@redhat.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20220225205948.3693480-9-jsnow@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
Refactor _do_accept() into _do_start_server() and _do_accept(). As of
this commit, the former calls the latter, but in subsequent commits
they'll be split apart.
(So please forgive the misnomer for _do_start_server(); it will live up
to its name shortly, and the docstring will be updated then too. I'm
just cutting down on some churn.)
Signed-off-by: John Snow <jsnow@redhat.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20220225205948.3693480-7-jsnow@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
As part of disentangling the monolithic nature of _do_accept(), split
out the incoming callback to prepare for factoring out the "wait for a
peer" step. Namely, this means using an event signal we can wait on from
outside of this method.
Signed-off-by: John Snow <jsnow@redhat.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20220225205948.3693480-5-jsnow@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
These two methods attempted to entirely envelop the logic of
establishing a connection to a peer start to finish. However, we need to
break apart the incoming connection step into more granular steps. We
will no longer be able to reasonably constrain the logic inside of these
helper functions.
So, remove them - with _session_guard(), they no longer serve a real
purpose.
Although the public API doesn't change, the internal API does. Now that
there are no intermediary methods between e.g. connect() and
_do_connect(), there's no hook where the runstate is set. As a result,
the test suite changes a little to cope with the new semantics of
_do_accept() and _do_connect().
Lastly, take some pieces of the now-deleted docstrings and move
them up to the public interface level. They were a little more detailed,
and it won't hurt to keep them.
Signed-off-by: John Snow <jsnow@redhat.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20220225205948.3693480-4-jsnow@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
Previously, I had a method named "accept()" that under-the-hood calls
bind(2), listen(2) *and* accept(2). I meant this as a simplification and
counterpart to the one-shot "connect()" method.
This is confusing to readers who expect accept() to mean *just*
accept(2). Since I need to split apart the "accept()" method into
multiple methods anyway (one of which strongly resembling accept(2)), it
feels pertinent to rename this method *now*.
Rename this all-in-one method "start_server_and_accept()" instead.
Signed-off-by: John Snow <jsnow@redhat.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20220225205948.3693480-3-jsnow@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
In _new_session, there's a fairly complex except clause that's used to
give semantic errors to callers of accept() and connect(). We need to
create a new two-step replacement for accept(), so factoring out this
piece of logic will be useful.
Bolster the comments and docstring here to try and demystify what's
going on in this fairly delicate piece of Python magic.
(If we were using Python 3.7+, this would be an @asynccontextmanager. We
don't have that very nice piece of magic, however, so this must take an
Awaitable to manage the Exception contexts properly. We pay the price
for platform compatibility.)
Signed-off-by: John Snow <jsnow@redhat.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20220225205948.3693480-2-jsnow@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
Setuptools v60 and later include a bundled version of distutils, a
deprecated standard library scheduled for removal in future versions of
Python. Setuptools v60 is only possible to install for Python 3.7 and later.
Python has a distutils.sysconfig.get_python_lib() function that returns
'/usr/lib/pythonX.Y' on posix systems. RPM-based systems actually use
'/usr/lib64/pythonX.Y' instead, so Fedora patches stdlib distutils for
Python 3.7 and Python 3.8 to return the correct value.
Python 3.9 and later introduce a sys.platlibdir property, which returns
the correct value on RPM-based systems.
The change to a distutils package not provided by Fedora on Python 3.7
and 3.8 causes a regression in distutils.sysconfig.get_python_lib() that
ultimately causes false positives to be emitted by pylint, because it
can no longer find the system source libraries.
Many Python tools are fairly aggressive about updating setuptools
packages, and so even though this package is a fair bit newer than
Python 3.7/3.8, it's not entirely unreasonable for a given user to have
such a modern package with a fairly old Python interpreter.
Updates to Python 3.7 and Python 3.8 are being produced for Fedora which
will fix the problem on up-to-date systems. Until then, we can force the
loading of platform-provided distutils when running the pylint
test. This is the least-invasive yet most comprehensive fix.
References:
https://github.com/pypa/setuptools/pull/2896https://github.com/PyCQA/pylint/issues/5704https://github.com/pypa/distutils/issues/110
Signed-off-by: John Snow <jsnow@redhat.com>
Message-id: 20220204221804.2047468-2-jsnow@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
When invoking setup.py directly, the default behavior for 'install' is
to run the bdist_egg installation hook, which is ... actually deprecated
by setuptools. It doesn't seem to work quite right anymore.
By contrast, 'pip install' will invoke the bdist_wheel hook
instead. This leads to differences in behavior for the two approaches. I
advocate using pip in the documentation in this directory, but the
'setup.py' which has been used for quite a long time in the Python world
may deceptively appear to work at first glance.
Add an error message that will save a bit of time and frustration
that points the user towards using the supported installation
invocation.
Reported-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Beraldo Leal <bleal@redhat.com>
Message-id: 20220207213039.2278569-1-jsnow@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
When running QMP commands with very large response payloads, it is often
not easy to spot the info you want. If we can save the response to a
file then tools like 'grep' or 'jq' can be used to extract information.
For convenience of processing, we merge the QMP command and response
dictionaries together:
{
"arguments": {},
"execute": "query-kvm",
"return": {
"enabled": false,
"present": true
}
}
Example usage
$ ./scripts/qmp/qmp-shell-wrap -l q.log -p -- ./build/qemu-system-x86_64 -display none
Welcome to the QMP low-level shell!
Connected
(QEMU) query-kvm
{
"return": {
"enabled": false,
"present": true
}
}
(QEMU) query-mice
{
"return": [
{
"absolute": false,
"current": true,
"index": 2,
"name": "QEMU PS/2 Mouse"
}
]
}
$ jq --slurp '. | to_entries[] | select(.value.execute == "query-kvm") |
.value.return.enabled' < q.log
false
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20220128161157.36261-3-berrange@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
With the current 'qmp-shell' tool developers must first spawn QEMU with
a suitable -qmp arg and then spawn qmp-shell in a separate terminal
pointing to the right socket.
With 'qmp-shell-wrap' developers can ignore QMP sockets entirely and
just pass the QEMU command and arguments they want. The program will
listen on a UNIX socket and tell QEMU to connect QMP to that.
For example, this:
# qmp-shell-wrap -- qemu-system-x86_64 -display none
Is roughly equivalent of running:
# qemu-system-x86_64 -display none -qmp qmp-shell-1234 &
# qmp-shell qmp-shell-1234
Except that 'qmp-shell-wrap' switches the socket peers around so that
it is the UNIX socket server and QEMU is the socket client. This makes
QEMU reliably go away when qmp-shell-wrap exits, closing the server
socket.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20220128161157.36261-2-berrange@redhat.com
[Edited for rebase. --js]
Signed-off-by: John Snow <jsnow@redhat.com>
The synchronous QMP library would bind to the server address during
__init__(). The new library delays this to the accept() call, because
binding occurs inside of the call to start_[unix_]server(), which is an
async method -- so it cannot happen during __init__ anymore.
Python 3.7+ adds the ability to create the server (and thus the bind()
call) and begin the active listening in separate steps, but we don't
have that functionality in 3.6, our current minimum.
Therefore ... Add a temporary workaround that allows the synchronous
version of the client to bind the socket in advance, guaranteeing that
there will be a UNIX socket in the filesystem ready for the QEMU client
to connect to without a race condition.
(Yes, it's a bit ugly. Fixing it more nicely will have to wait until our
minimum Python version is 3.7+.)
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-id: 20220201041134.1237016-5-jsnow@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
We need a slightly newer version of mypy in order to use some features
of the asyncio server functions in the next commit.
(Note: pipenv is not really suited to upgrading individual packages; I
need to replace this tool with something better for the task. For now,
the miscellaneous updates not related to the mypy upgrade are simply
beyond my control. It's on my list to take care of soon.)
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-id: 20220201041134.1237016-4-jsnow@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
This allows us to pack in some extra information about the failure,
which guarantees that if the caller did not *intentionally* cause a
failure (by capturing this Exception), some pretty good clues will be
printed at the bottom of the traceback information.
This will help make failures in the event of a non-negative return code
more obvious when they go unhandled; the current behavior in
_post_shutdown() is to print a warning message only in the event of
signal-based terminations (for negative return codes).
(Note: In Python, catching BaseException instead of Exception catches a
broader array of Exception events, including SystemExit and
KeyboardInterrupt. We do not want to "wrap" such exceptions as a
VMLaunchFailure, because that will 'downgrade' the exception from a
BaseException to a regular Exception. We do, however, want to perform
cleanup in either case, so catch on the broadest scope and
wrap-and-re-raise only in the more targeted scope.)
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-id: 20220201041134.1237016-3-jsnow@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
In order to upload a QMP package to PyPI, I want to remove any scripts
that I am not 100% confident I want to support upstream, beyond our
castle walls.
Move most of our QMP utilities into the utils package so we can split
them out from the PyPI upload.
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Beraldo Leal <bleal@redhat.com>
We have a replacement for async QMP, but it doesn't have feature parity
yet. For now, then, port the old tool onto the new backend.
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Async QMP always raises a "ConnectError" on any connection error which
houses the cause in a second exception. We can check if this root cause
was python's ConnectionError to determine a fairly similar condition to
the original error check here.
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Beraldo Leal <bleal@redhat.com>
This is in preparation for renaming qemu.aqmp to qemu.qmp. I should have
done this from this from the very beginning, but it's a convenient time
to make sure this churn is taken care of.
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
It's a commonly needed definition, it can be re-exported by the root.
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Beraldo Leal <bleal@redhat.com>
John Snow
Daniel P. BerrangéandJohn Snow