Re: [Qemu-devel] [PATCH v2 1/3] libqtest: Ignore QMP events when parsing the response for HMP commands

[Dr. David Alan Gilbert]

* John Snow (address@hidden) wrote:
> 
> 
> On 04/04/2017 03:31 AM, Thomas Huth wrote:
> > On 03.04.2017 21:09, John Snow wrote:
> >>
> >>
> >> On 03/30/2017 03:50 AM, Thomas Huth wrote:
> >>> When running certain HMP commands (like "device_del") via QMP, we
> >>> can sometimes get a QMP event in the response first, so that the
> >>> "g_assert(ret)" statement in qtest_hmp() triggers and the test
> >>> fails. Fix this by ignoring such QMP events while looking for the
> >>> real return value from QMP.
> >>>
> >>> Signed-off-by: Thomas Huth <address@hidden>
> >>> ---
> >>>  tests/libqtest.c | 6 ++++++
> >>>  1 file changed, 6 insertions(+)
> >>>
> >>> diff --git a/tests/libqtest.c b/tests/libqtest.c
> >>> index a5c3d2b..c9b2d76 100644
> >>> --- a/tests/libqtest.c
> >>> +++ b/tests/libqtest.c
> >>> @@ -580,6 +580,12 @@ char *qtest_hmpv(QTestState *s, const char *fmt, 
> >>> va_list ap)
> >>>                       " 'arguments': {'command-line': %s}}",
> >>>                       cmd);
> >>>      ret = g_strdup(qdict_get_try_str(resp, "return"));
> >>> +    while (ret == NULL && qdict_get_try_str(resp, "event")) {
> >>> +        /* Ignore asynchronous QMP events */
> >>> +        QDECREF(resp);
> >>> +        resp = qtest_qmp_receive(s);
> >>> +        ret = g_strdup(qdict_get_try_str(resp, "return"));
> >>> +    }
> >>>      g_assert(ret);
> >>>      QDECREF(resp);
> >>>      g_free(cmd);
> >>>
> >>
> >> You've probably been asked this, but can you just shove the QMP response
> >> you don't want into the event queue for consumption by other calls?
> > 
> > Well, this is the qtest_hmpv() function, so I assume that the caller
> > just wants to execute a HMP command and does not really care about QMP
> > events. If you care about QMP events, you should use the qmp functions
> > instead.
> > 
> >  Thomas
> > 
> 
> I don't think it's obvious that using HMP functions should cause the QMP
> stream to become faulty, though.
> 
> If someone uses an HMP function and then tries to wait on a QMP event to
> confirm that some key condition has occurred (pausing or resuming, for
> instance) it would not be immediately apparent from the user's POV that
> this function just eats replies because it was convenient to do so.
> 
> I guess the event queue only exists in python though, so it's not as
> trivial as I was thinking it would be...

I think it's OK to discard the QMP events - it feels rare to mix and match
in a test; if you care about QMP events you'll probably be basing the
test around QMP rather than HMP.

Dave

--
Dr. David Alan Gilbert / address@hidden / Manchester, UK