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]
Date:   Mon, 30 Jul 2018 18:04:27 +0200
From:   Vincent Guittot <vincent.guittot@...aro.org>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     Ingo Molnar <mingo@...nel.org>,
        linux-kernel <linux-kernel@...r.kernel.org>,
        "Rafael J. Wysocki" <rjw@...ysocki.net>,
        viresh kumar <viresh.kumar@...aro.org>
Subject: Re: [PATCH] sched/fair: remove #ifdefs from scale_rt_capacity

Hi Peter,

On Fri, 20 Jul 2018 at 14:31, Peter Zijlstra <peterz@...radead.org> wrote:
>
> On Thu, Jul 19, 2018 at 02:00:06PM +0200, Vincent Guittot wrote:
> > But the compiler is not able to optimize the sequence (at least with
> > aarch64 GCC 7.2.1)
> >       free *= (max - irq);
> >       free /= max;
> > when irq is fixed to 0
>
> So much for compilers.... I though those things were supposed to be
> 'clever'.
>
> > +#if defined(SMP) \
> > + && (defined(CONFIG_IRQ_TIME_ACCOUNTING) || defined(CONFIG_PARAVIRT_TIME_ACCOUNTING))
>
> That's atrocious :-)
>
> Fixed it with the below.

Thanks for the fix


>
>
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -135,7 +135,7 @@ static void update_rq_clock_task(struct
>   * In theory, the compile should just see 0 here, and optimize out the call
>   * to sched_rt_avg_update. But I don't trust it...
>   */
> -#if defined(CONFIG_IRQ_TIME_ACCOUNTING) || defined(CONFIG_PARAVIRT_TIME_ACCOUNTING)
> +#ifdef HAVE_SCHED_AVG_IRQ
>         s64 steal = 0, irq_delta = 0;
>  #endif
>  #ifdef CONFIG_IRQ_TIME_ACCOUNTING
> @@ -177,7 +177,7 @@ static void update_rq_clock_task(struct
>
>         rq->clock_task += delta;
>
> -#if defined(CONFIG_IRQ_TIME_ACCOUNTING) || defined(CONFIG_PARAVIRT_TIME_ACCOUNTING)
> +#ifdef HAVE_SCHED_AVG_IRQ
>         if ((irq_delta + steal) && sched_feat(NONTASK_CAPACITY))
>                 update_irq_load_avg(rq, irq_delta + steal);
>  #endif
> --- a/kernel/sched/sched.h
> +++ b/kernel/sched/sched.h
> @@ -856,6 +856,7 @@ struct rq {
>         struct sched_avg        avg_rt;
>         struct sched_avg        avg_dl;
>  #if defined(CONFIG_IRQ_TIME_ACCOUNTING) || defined(CONFIG_PARAVIRT_TIME_ACCOUNTING)
> +#define HAVE_SCHED_AVG_IRQ
>         struct sched_avg        avg_irq;
>  #endif
>         u64                     idle_stamp;
> @@ -2212,8 +2213,7 @@ static inline unsigned long cpu_util_rt(
>  }
>  #endif
>
> -#if defined(SMP) \
> - && (defined(CONFIG_IRQ_TIME_ACCOUNTING) || defined(CONFIG_PARAVIRT_TIME_ACCOUNTING))
> +#ifdef HAVE_SCHED_AVG_IRQ
>  static inline unsigned long cpu_util_irq(struct rq *rq)
>  {
>         return rq->avg_irq.util_avg;

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ