linux-kernel - Re: [PATCH 2/2] sched: Pull resched loop to _

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CA+55aFwhEkBc899yMY3KugwF0QGF8r7iM_6+DoYa0RgtApv5aw@mail.gmail.com>
Date:	Sun, 14 Dec 2014 18:46:27 -0800
From:	Linus Torvalds <torvalds@...ux-foundation.org>
To:	Frederic Weisbecker <fweisbec@...il.com>
Cc:	Ingo Molnar <mingo@...nel.org>,
	LKML <linux-kernel@...r.kernel.org>,
	Peter Zijlstra <peterz@...radead.org>
Subject: Re: [PATCH 2/2] sched: Pull resched loop to __schedule() callers

On Sun, Dec 14, 2014 at 6:24 PM, Frederic Weisbecker <fweisbec@...il.com> wrote:
> -need_resched:
>         preempt_disable();
>         cpu = smp_processor_id();
>         rq = cpu_rq(cpu);
> @@ -2821,8 +2824,6 @@ need_resched:
>         post_schedule(rq);
>
>         sched_preempt_enable_no_resched();
> -       if (need_resched())
> -               goto need_resched;

So my suggestion was to move the
"preempt_disable()/enable_no_resched()" into the callers too.

Everybody already does that - except for "schedule()" itself. So that
would involve adding it here too:

> @@ -2842,7 +2843,9 @@ asmlinkage __visible void __sched schedule(void)
>         struct task_struct *tsk = current;
>
>         sched_submit_work(tsk);
> -       __schedule();
> +       do {
preempt_disable();
> +               __schedule();
sched_preempt_enable_no_resched();
> +       } while (need_resched());
>  }

Hmm?

That would mean that we have one less preempt update in the
__sched_preempt() cases. If somebody cares about the preempt counter
value, we'd have to increemnt the preempt count add values too, ie do

    __preempt_count_add(PREEMPT_ACTIVE+1);

there, but I'm not convicned anybody cares about the exact count.

As it is, you end up doing the preempt count things twice in the
__sched_preempt() case: first the PREEMPT_ACTIVE count, and then the
count inside __schedule().

Hmm?

                     Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/