From mboxrd@z Thu Jan  1 00:00:00 1970
From: Cyril Hrubis <chrubis@suse.cz>
Date: Thu, 11 Mar 2021 14:49:02 +0100
Subject: [LTP] [PATCH 6/7] tst_test.sh: Run cleanup also after test
 timeout
In-Reply-To: <YD4RairzgAuyBn7S@pevik>
References: <20210301220222.22705-1-pvorel@suse.cz>
 <20210301220222.22705-7-pvorel@suse.cz>
 <CAEemH2c2Bma-xXTcf+1M4CFVjM7ZQS36rV_6i9kJ6Es-19kghA@mail.gmail.com>
 <YD4RairzgAuyBn7S@pevik>
Message-ID: <YEofzk8nrVepc78j@yuki.lan>
List-Id: <ltp.lists.linux.it>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: ltp@lists.linux.it

Hi!
> > >  _tst_cleanup_timer()
> > >  {
> > > @@ -442,6 +442,14 @@ _tst_multiply_timeout()
> > >         return 0
> > >  }
> 
> > > +_tst_run_timer()
> 
> 
> > Hmm, this name is not good than before, or rename to _tst_kill_timer_pid(),
> > _tst_stop_timer()?
> 
> Good point. I slightly prefer _tst_stop_timer, but no hard feeling about it.

Or _tst_kill_test()?

> > > +{
> > > +       tst_res TBROK "test killed, timeout! If you are running on slow
> > > machine, try exporting LTP_TIMEOUT_MUL > 1"
> > > +       kill -INT -$pid
> > > +       sleep 5
> > > +       kill -KILL -$pid

Maybe we should change the messages to reflect what is happening and
maybe we should check if the test is still running before sending
SIGKILL with kill -0 $pid?

	tst_res TBROK "Test timeouted, sending SIGINT, ...."
	kill -INT -$pid

	sleep 5

	if kill -0 $pid 2>&1 > /dev/null; then
		tst_res TBROK "Test still running, sending SIGKILL"
		kill -KILL -$pid
	fi

We can also bussy loop wait for the process to terminate, e.g. loop 10
times with sleep 1 in the body and break the loop if kill -0 $pid
returns failure.

> > > +}
> > > +

-- 
Cyril Hrubis
chrubis@suse.cz