Re: [PATCH v4 2/2] livepatch: Convert timeouts to secs_to_jiffies()

[Christophe Leroy]

Le 18/12/2024 à 09:38, Petr Mladek a écrit :
> On Tue 2024-12-17 23:09:59, Easwar Hariharan wrote:
> > Commit b35108a51cf7 ("jiffies: Define secs_to_jiffies()") introduced
> > secs_to_jiffies(). As the value here is a multiple of 1000, use
> > secs_to_jiffies() instead of msecs_to_jiffies to avoid the multiplication.
> >
> > This is converted using scripts/coccinelle/misc/secs_to_jiffies.cocci with
> > the following Coccinelle rules:
> >
> > @@ constant C; @@
> >
> > - msecs_to_jiffies(C * 1000)
> > + secs_to_jiffies(C)
> >
> > @@ constant C; @@
> >
> > - msecs_to_jiffies(C * MSEC_PER_SEC)
> > + secs_to_jiffies(C)
> >
> > While here, replace the schedule_delayed_work() call with a 0 timeout
> > with an immediate schedule_work() call.
> >
> > --- a/samples/livepatch/livepatch-callbacks-busymod.c
> > +++ b/samples/livepatch/livepatch-callbacks-busymod.c
> > @@ -44,8 +44,7 @@ static void busymod_work_func(struct work_struct *work)
> >   static int livepatch_callbacks_mod_init(void)
> >   {
> >   	pr_info("%s\n", __func__);
> > -	schedule_delayed_work(&work,
> > -		msecs_to_jiffies(1000 * 0));
> > +	schedule_work(&work);
>
> Is it safe to use schedule_work() for struct delayed_work?

Should be, but you are right it should then be a standard work not a 
delayed work.

So probably the easiest is to keep

	schedule_delayed_work(&work, 0)

And eventually changing it to a not delayed work could be a follow-up patch.

> It might work in theory but I do not feel comfortable with it.
> Also I would expect a compiler warning.

__queue_delayed_work() does :

	if (!delay) {
		__queue_work(cpu, wq, &dwork->work);
		return;
	}


> If you really want to use schedule_work() then please
> also define the structure with DECLARE_WORK()
> and use cancel_work_sync() in livepatch_callbacks_mod_exit().
>
> Best Regards,
> Petr