linux-kernel - Re: [v4 RFC PATCH 4/4] timers: logic to move non pinned timers

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 for Android: free password hash cracker in your pocket

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

Message-ID: <20090401114641.GF22478@linux.vnet.ibm.com>
Date:	Wed, 1 Apr 2009 17:16:41 +0530
From:	Arun R Bharadwaj <arun@...ux.vnet.ibm.com>
To:	linux-kernel@...r.kernel.org, linux-pm@...ts.linux-foundation.org
Cc:	a.p.zijlstra@...llo.nl, ego@...ibm.com, tglx@...utronix.de,
	mingo@...e.hu, andi@...stfloor.org, venkatesh.pallipadi@...el.com,
	vatsa@...ux.vnet.ibm.com, arjan@...radead.org,
	svaidy@...ux.vnet.ibm.com, Arun Bharadwaj <arun@...ux.vnet.ibm.com>
Subject: Re: [v4 RFC PATCH 4/4] timers: logic to move non pinned timers

* Arun R Bharadwaj <arun@...ux.vnet.ibm.com> [2009-04-01 17:07:38]:

--------snip--------------
> Index: linux.trees.git/include/linux/hrtimer.h
> ===================================================================
> --- linux.trees.git.orig/include/linux/hrtimer.h
> +++ linux.trees.git/include/linux/hrtimer.h
> @@ -22,7 +22,6 @@
>  #include <linux/wait.h>
>  #include <linux/percpu.h>
> 
> -
>  struct hrtimer_clock_base;
>  struct hrtimer_cpu_base;
> 
> @@ -376,6 +375,24 @@ extern ktime_t hrtimer_get_remaining(con
>  extern int hrtimer_get_res(const clockid_t which_clock, struct timespec *tp);
> 
>  extern ktime_t hrtimer_get_next_event(void);
> +extern ktime_t hrtimer_get_next_event_on(int cpu);
> +
> +#define TICK_PERIOD_IN_NS	NSEC_PER_SEC / HZ
> +/*
> + * Helper function to check before migrating a hrtimer if it expires
> + * before the next timer interrupt on the target cpu.
> + */
> +static inline int check_hrtimer_latency(struct hrtimer *timer, int cpu)
> +{
> +	unsigned long next_hrtimeout, next_jiffies;
> +	next_jiffies = get_next_timer_interrupt_on(jiffies, cpu);
> +	next_jiffies = next_jiffies * TICK_PERIOD_IN_NS;
> +	next_hrtimeout = hrtimer_get_expires_ns(timer);
> +
> +	if (next_hrtimeout > next_jiffies)
> +		return 1;
> +	return 0;
> +}
>

This function can be greatly simplified.
Here i am getting the next timer interrupt on target_cpu.
The target_cpu is the idle load balancer.
So we can check if the hrtimer expiry is before the next tick
on the ilb or not.
By doing this we can eliminate a lot of extra code, which would
otherwise be needed in the form of helper functions for
get_next_timer_interrupt_on().

After simplification the code would simply look like
static inline int check_hrtimer_latency(struct hrtimer *timer, int cpu)
{
        unsigned long next_hrtimeout, next_jiffies;
        next_jiffies = (jiffies + 1) * TICK_PERIOD_IN_NS;
        next_hrtimeout = hrtimer_get_expires_ns(timer);

        if(.....)
}

--arun

>  /*
>   * A timer is active, when it is enqueued into the rbtree or the callback
> Index: linux.trees.git/include/linux/timer.h
> ===================================================================
> --- linux.trees.git.orig/include/linux/timer.h
> +++ linux.trees.git/include/linux/timer.h
> @@ -184,7 +184,7 @@ extern unsigned long next_timer_interrup
>   * jiffie.
>   */
>  extern unsigned long get_next_timer_interrupt(unsigned long now);
> -
> +extern unsigned long get_next_timer_interrupt_on(unsigned long now, int cpu);
>  /*
>   * Timer-statistics info:
>   */
--
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/