From mboxrd@z Thu Jan 1 00:00:00 1970 From: Kay Sievers Date: Tue, 02 Jun 2009 11:39:06 +0000 Subject: Re: [GIT] Experimental threaded udev Message-Id: List-Id: References: <4A1EA138.10400@tuffmail.co.uk> In-Reply-To: <4A1EA138.10400@tuffmail.co.uk> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit To: linux-hotplug@vger.kernel.org On Tue, Jun 2, 2009 at 11:13, Alan Jenkins wrote: >> Version 3, which should clean up events with a worker that died. Also >> kills all workers if the config has changed. > > For strict correctness, I guess we need to do the same for "udevadm control > --log-priority" and "udevadm control --env". Yeah, makes sense, will add that. > Hmm, it says worker_kill_idle().  If there are running events at the time, > their stale workers will survive.  At the moment kill(worker->pid, SIGTERM) > allows the current event to finish before it terminates, so one solution > would be to just kill all the workers. The main daemon will wakeup with a timeout as long as workers exists. There meigt be several rounds of killing, until all workers are gone. Thanks, Kay