Hello Johan,
On Tue, Jul 14, 2020 at 09:13:55AM +0200, Johan Hovold wrote:
> On Tue, Jul 07, 2020 at 06:59:58PM +0200, Uwe Kleine-König wrote:
> > + while (firstrun ||
> > + icount.rx != trigger_data->rx ||
> > + icount.tx != trigger_data->tx) {
> > +
> > + led_set_brightness(trigger_data->led_cdev, LED_ON);
> > +
> > + msleep(100);
> > +
> > + led_set_brightness(trigger_data->led_cdev, LED_OFF);
> > +
> > + trigger_data->rx = icount.rx;
> > + trigger_data->tx = icount.tx;
> > + firstrun = false;
> > +
> > + ret = tty_get_icount(trigger_data->tty, &icount);
> > + if (ret)
> > + return;
> > + }
>
> Haven't looked at the latest proposal in detail, but this looks broken
> as you can potentially loop indefinitely in a worker thread, and with no
> way to stop the trigger (delayed work).
I don't think that potentially looping indefinitely is a problem, but
indeed it should drop the lock during each iteration. Will think about
how to adapt.
Thanks
Uwe
--
Pengutronix e.K. | Uwe Kleine-König |
Industrial Linux Solutions | https://www.pengutronix.de/ |
Attachment:
signature.asc
Description: PGP signature
