Re: [PATCH] sched/cputime: correct account of irqtime

[jun qian]

Pingfan Liu <kernelfans@xxxxxxxxx> 于2020年10月12日周一 下午9:54写道：
>
> __do_softirq() may be interrupted by hardware interrupts. In this case,
> irqtime_account_irq() will account the time slice as CPUTIME_SOFTIRQ by
> mistake.
>
> By passing irqtime_account_irq() an extra param about either hardirq or
> softirq, irqtime_account_irq() can handle the above case.
>
> Signed-off-by: Pingfan Liu <kernelfans@xxxxxxxxx>
> Cc: Ingo Molnar <mingo@xxxxxxxxxx>
> Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> Cc: Juri Lelli <juri.lelli@xxxxxxxxxx>
> Cc: Vincent Guittot <vincent.guittot@xxxxxxxxxx>
> Cc: Dietmar Eggemann <dietmar.eggemann@xxxxxxx>
> Cc: Steven Rostedt <rostedt@xxxxxxxxxxx>
> Cc: Ben Segall <bsegall@xxxxxxxxxx>
> Cc: Mel Gorman <mgorman@xxxxxxx>
> Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> Cc: Andy Lutomirski <luto@xxxxxxxxxx>
> Cc: Will Deacon <will@xxxxxxxxxx>
> Cc: "Paul E. McKenney" <paulmck@xxxxxxxxxx>
> Cc: Frederic Weisbecker <frederic@xxxxxxxxxx>
> Cc: Allen Pais <allen.lkml@xxxxxxxxx>
> Cc: Romain Perier <romain.perier@xxxxxxxxx>
> To: linux-kernel@xxxxxxxxxxxxxxx
> ---
>  include/linux/hardirq.h |  4 ++--
>  include/linux/vtime.h   | 12 ++++++------
>  kernel/sched/cputime.c  |  4 ++--
>  kernel/softirq.c        |  6 +++---
>  4 files changed, 13 insertions(+), 13 deletions(-)
>
> diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h
> index 754f67a..56e7bb5 100644
> --- a/include/linux/hardirq.h
> +++ b/include/linux/hardirq.h
> @@ -32,7 +32,7 @@ static __always_inline void rcu_irq_enter_check_tick(void)
>   */
>  #define __irq_enter()                                  \
>         do {                                            \
> -               account_irq_enter_time(current);        \
> +               account_irq_enter_time(current, true);  \
>                 preempt_count_add(HARDIRQ_OFFSET);      \
>                 lockdep_hardirq_enter();                \
>         } while (0)
> @@ -63,7 +63,7 @@ void irq_enter_rcu(void);
>  #define __irq_exit()                                   \
>         do {                                            \
>                 lockdep_hardirq_exit();                 \
> -               account_irq_exit_time(current);         \
> +               account_irq_exit_time(current, true);   \
>                 preempt_count_sub(HARDIRQ_OFFSET);      \
>         } while (0)
>
> diff --git a/include/linux/vtime.h b/include/linux/vtime.h
> index 2cdeca0..294188ae1 100644
> --- a/include/linux/vtime.h
> +++ b/include/linux/vtime.h
> @@ -98,21 +98,21 @@ static inline void vtime_flush(struct task_struct *tsk) { }
>
>
>  #ifdef CONFIG_IRQ_TIME_ACCOUNTING
> -extern void irqtime_account_irq(struct task_struct *tsk);
> +extern void irqtime_account_irq(struct task_struct *tsk, bool hardirq);
>  #else
> -static inline void irqtime_account_irq(struct task_struct *tsk) { }
> +static inline void irqtime_account_irq(struct task_struct *tsk, bool hardirq) { }
>  #endif
>
> -static inline void account_irq_enter_time(struct task_struct *tsk)
> +static inline void account_irq_enter_time(struct task_struct *tsk, bool hardirq)
>  {
>         vtime_account_irq_enter(tsk);
> -       irqtime_account_irq(tsk);
> +       irqtime_account_irq(tsk, hardirq);
>  }
>
> -static inline void account_irq_exit_time(struct task_struct *tsk)
> +static inline void account_irq_exit_time(struct task_struct *tsk, bool hardirq)
>  {
>         vtime_account_irq_exit(tsk);
> -       irqtime_account_irq(tsk);
> +       irqtime_account_irq(tsk, hardirq);
>  }
>
>  #endif /* _LINUX_KERNEL_VTIME_H */
> diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c
> index 5a55d23..166f1d7 100644
> --- a/kernel/sched/cputime.c
> +++ b/kernel/sched/cputime.c
> @@ -47,7 +47,7 @@ static void irqtime_account_delta(struct irqtime *irqtime, u64 delta,
>   * Called before incrementing preempt_count on {soft,}irq_enter
>   * and before decrementing preempt_count on {soft,}irq_exit.
>   */
> -void irqtime_account_irq(struct task_struct *curr)
> +void irqtime_account_irq(struct task_struct *curr, bool hardirq)
>  {
>         struct irqtime *irqtime = this_cpu_ptr(&cpu_irqtime);
>         s64 delta;
> @@ -68,7 +68,7 @@ void irqtime_account_irq(struct task_struct *curr)
>          */
>         if (hardirq_count())
>                 irqtime_account_delta(irqtime, delta, CPUTIME_IRQ);
> -       else if (in_serving_softirq() && curr != this_cpu_ksoftirqd())
> +       else if (in_serving_softirq() && curr != this_cpu_ksoftirqd() && !hardirq)
>                 irqtime_account_delta(irqtime, delta, CPUTIME_SOFTIRQ);
>  }

In my opinion, we don't need to use the hardirq flag, the code: if
(hardirq_count())
already tell us that where the delt time is from.

Thanks

>  EXPORT_SYMBOL_GPL(irqtime_account_irq);
> diff --git a/kernel/softirq.c b/kernel/softirq.c
> index bf88d7f6..da59ea39 100644
> --- a/kernel/softirq.c
> +++ b/kernel/softirq.c
> @@ -270,7 +270,7 @@ asmlinkage __visible void __softirq_entry __do_softirq(void)
>         current->flags &= ~PF_MEMALLOC;
>
>         pending = local_softirq_pending();
> -       account_irq_enter_time(current);
> +       account_irq_enter_time(current, false);
>
>         __local_bh_disable_ip(_RET_IP_, SOFTIRQ_OFFSET);
>         in_hardirq = lockdep_softirq_start();
> @@ -321,7 +321,7 @@ asmlinkage __visible void __softirq_entry __do_softirq(void)
>         }
>
>         lockdep_softirq_end(in_hardirq);
> -       account_irq_exit_time(current);
> +       account_irq_exit_time(current, false);
>         __local_bh_enable(SOFTIRQ_OFFSET);
>         WARN_ON_ONCE(in_interrupt());
>         current_restore_flags(old_flags, PF_MEMALLOC);
> @@ -417,7 +417,7 @@ static inline void __irq_exit_rcu(void)
>  #else
>         lockdep_assert_irqs_disabled();
>  #endif
> -       account_irq_exit_time(current);
> +       account_irq_exit_time(current, true);
>         preempt_count_sub(HARDIRQ_OFFSET);
>         if (!in_interrupt() && local_softirq_pending())
>                 invoke_softirq();
> --
> 2.7.5
>