Re: [Qemu-devel] [PATCH] libqtest: fail if child coredumps

[Michael S. Tsirkin]

On Thu, May 24, 2018 at 05:04:56PM +0200, Marc-André Lureau wrote:
> Hi
> 
> On Thu, May 24, 2018 at 5:00 PM, Michael S. Tsirkin <address@hidden> wrote:
> > On Thu, May 24, 2018 at 04:45:31PM +0200, Thomas Huth wrote:
> >> On 24.05.2018 16:30, Michael S. Tsirkin wrote:
> >> > Right now tests report OK status if QEMU crashes during cleanup.
> >> > Let's catch that case and fail the test.
> >> >
> >> > Signed-off-by: Michael S. Tsirkin <address@hidden>
> >> > ---
> >> >  tests/libqtest.c | 9 ++++++++-
> >> >  1 file changed, 8 insertions(+), 1 deletion(-)
> >> >
> >> > diff --git a/tests/libqtest.c b/tests/libqtest.c
> >> > index 43fb97e..f869854 100644
> >> > --- a/tests/libqtest.c
> >> > +++ b/tests/libqtest.c
> >> > @@ -103,8 +103,15 @@ static int socket_accept(int sock)
> >> >  static void kill_qemu(QTestState *s)
> >> >  {
> >> >      if (s->qemu_pid != -1) {
> >> > +        int wstatus = 0;
> >> > +        pid_t pid;
> >> > +
> >> >          kill(s->qemu_pid, SIGTERM);
> >> > -        waitpid(s->qemu_pid, NULL, 0);
> >> > +        pid = waitpid(s->qemu_pid, &wstatus, 0);
> >> > +
> >> > +        if (pid == s->qemu_pid && WIFSIGNALED(wstatus)) {
> >> > +            assert(!WCOREDUMP(wstatus));
> >> > +        }
> >> >      }
> >> >  }
> >>
> >> That's basically a good idea ... but I've already seen yet another issue
> >> in the past already: QEMU sometimes simply hangs in an endless loop
> >> during clean up and never terminates. I think we should detect that
> >> situation, too. So instead of killing QEMU at the end of the testing, I 
> >> think we should
> >> rather try to terminate it with the QMP "quit" command. If QEMU does not
> >> terminate with an exit code of 0, then the test should be flagged a
> >> failed (and only if QEMU did not terminate at all, it should be killed
> >> with SIGKILL).
> >>
> >>  Thomas
> >
> > Fine but can we agree to do this as a patch on top? And do you have
> > the time to implement this?
> >
> > I'm seeing patches that cause crash on cleanup, it's not a theoretical
> > problem for me, so I'd like this one to go in first.
> 
> What is the difference between sending "quit" and SIGTERM (qemu has
> code to handle it in termsig_handler()). Is this documented somewhere?

SIGTERM might wake QEMU if it's blocked somewhere in kernel.
Sending quit also tests QMP isn't blocked.

> 
> -- 
> Marc-André Lureau