Re: [PATCH 1/2] kernel/sofirq: consolidate common code in __tasklet_schedule() + _hi_

From: Steven Rostedt
Date: Fri Feb 16 2018 - 13:02:45 EST


On Fri, 16 Feb 2018 18:55:09 +0100
Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx> wrote:


> > Should we add something like:
> >
> > WARN_ON_ONCE(!in_atomic());
> >
> > ?
>
> Doubt it. this_cpu_ptr() screams already with CONFIG_DEBUG_PREEMPT.

If that's the case then, yeah I agree. I couldn't remember if
this_cpu_ptr() did that or not. I remember having an argument with
Christoph Lameter about whether or not this_cpu_* functions would
complain with preemption off, as some of the use cases were for being
used with preemption enabled. I remember there was some kind of
compromise but didn't remember exactly what that was.

-- Steve