Re: [PATCH -tip] x86: kprobes: Cleanup preempt disabling and enabling

[Masami Hiramatsu <mhiramat@kernel.org>]

On Sat, 3 Mar 2018 10:58:23 +0100
Ingo Molnar <mingo@kernel.org> wrote:

> 
> * Masami Hiramatsu <mhiramat@kernel.org> wrote:
> 
> > +/*
> > + * Interrupts are disabled on entry as trap3 is an interrupt gate and they
> > + * remain disabled throughout this function.
> > + */
> > +int kprobe_int3_handler(struct pt_regs *regs)
> > +{
> > +	struct kprobe_ctlblk *kcb;
> > +	int ret;
> > +
> > +	if (user_mode(regs))
> > +		return 0;
> > +
> > +	/*
> > +	 * We don't want to be preempted for the entire
> > +	 * duration of kprobe processing.
> > +	 */
> > +	preempt_disable();
> > +
> > +	kcb = get_kprobe_ctlblk();
> > +	ret = kprobe_int3_dispatcher(regs, kcb);
> > +
> > +	if (!kprobe_ready_for_singlestep(regs))
> > +		preempt_enable_no_resched();
> > +
> > +	return ret;
> 
> What's the point of disabling preemption, if IRQs are disabled already?
> 
> There's no preemption when IRQs are off...

Ahh, right! Whole the kprobe singlestepping, IRQs are off (kprobes drops
IF from regs->flags for single stepping) so we don't need to care about
preempt count anymore...

Thank you!


-- 
Masami Hiramatsu <mhiramat@kernel.org>