api: prevent early output retrieval

Change the API endpoint to prevent retrieving monitor-output from a
running instance. Instead, we require the caller to exit the API context
before querying the monitor-output. This guarantees that the api-thread
was synchronously taken down and scheduled any outstanding events.

This fixes an issue where a side-channel notifies us of a buildroot
exit, but the api-thread has not yet returned from epoll, and thus might
not have dispatched pending I/O events, yet. If we instead wait for the
thread to exit, we have a synchronous shutdown and know that all
*ordered* kernel events must have been handled.

In particular, imagine a build-root program running (like `echo` in the
test_monitor unittest) which writes data to the stdout-pipe and then
immediately exits. The syscall-order guarantees that the data is written
to the pipe before the SIGCHLD is sent (or wait(2) returns). However, we
retrieve the SIGCHLD from our main-thread usually (p.join() in our test,
and BuildRoot() in our main code), while the pipe-reading is done from
an API thread. Therefore, we might end up handling the SIGCHLD first
(just imagine a single-threaded CPU that schedules the main task before
the thread). To avoid this race, we can simply synchronize with the
api-thread. Since we already have this synchronization as part of the
api-thread takedown, it is as simple as stopping the api-thread before
continuing with operations.

Lastly, if a write operation to a pipe was issued, we are guaranteed
that a SIGCHLD synchronization across processes is ordered correctly.
Furthermore, the python event-loop also guarantees that stopping an
event-loop will necessarily dispatch all outstanding events. A read is
guaranteed to be outstanding in our race-scenario, so the read will be
dispatched. The only possible problem is `_output_ready()` only
dispatching a maximum of 4096 bytes. This might need to be fixed
separately. A comment is left in place.
This commit is contained in:
David Rheinsberg 2020-08-13 09:06:25 +02:00 committed by Christian Kellner
parent 818daef6cb
commit 803433fb62
3 changed files with 21 additions and 9 deletions

View file

@ -91,9 +91,13 @@ class BaseAPI(abc.ABC):
self.event_loop.run_forever()
self.event_loop.remove_reader(server)
@property
def running(self):
return self.event_loop is not None
def __enter__(self):
# We are not re-entrant, so complain if re-entered.
assert self.event_loop is None
assert not self.running
if not self.socket_address:
self._socketdir = self._make_socket_dir()
@ -141,6 +145,13 @@ class API(BaseAPI):
@property
def output(self):
# Only once the event-loop was stopped, you are guaranteed that the
# api-thread scheduled all outstanding events. Therefore, we disallow
# asking for the output-data from a running api context. If we happen
# to need live streaming access to the output in the future, we need
# to redesign the output-handling, anyway.
assert not self.running
return self._output_data.getvalue()
def _prepare_input(self):

View file

@ -73,8 +73,9 @@ class TestMonitor(unittest.TestCase):
path = os.path.join(tmpdir, "osbuild-api")
logfile = os.path.join(tmpdir, "log.txt")
with open(logfile, "w") as log, \
API(args, LogMonitor(log.fileno()), socket_address=path) as api:
with open(logfile, "w") as log:
api = API(args, LogMonitor(log.fileno()), socket_address=path)
with api as api:
p = mp.Process(target=echo, args=(path, ))
p.start()
p.join()