On Wed, 7 Jan 2026 10:24:43 -0300
Wander Lairson Costa <[email protected]> wrote:

> > In the kernel, this is handled via the READ_ONCE() macro. Perhaps rtla
> > should implement that too.
> >  
> 
> I considered that, but, in this use case, I saw no point because it
> didn't bring any advantage and volatile was simpler.
> Furthermore, as Crystal pointed out, using volatile for variables
> shared with signals is a pretty standard practice.

OK, I've just been broken in by Linus yelling at anyone defining any
variable as volatile ;-)  I now avoid it at all costs, and only have the
locations that need to be volatile defined that way.

-- Steve

Reply via email to