linux-kernel - Re: [PATCH 3/4] time,acct: drop irq save & restore from __acct_update

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: <20160130162400.GD32581@lerouge>
Date:	Sat, 30 Jan 2016 17:24:01 +0100
From:	Frederic Weisbecker <fweisbec@...il.com>
To:	riel@...hat.com
Cc:	linux-kernel@...r.kernel.org, tglx@...utronix.de, mingo@...nel.org,
	luto@...capital.net, peterz@...radead.org, clark@...hat.com
Subject: Re: [PATCH 3/4] time,acct: drop irq save & restore from
 __acct_update_integrals

On Fri, Jan 29, 2016 at 10:36:04PM -0500, riel@...hat.com wrote:
> From: Rik van Riel <riel@...hat.com>
> 
> It looks like all the call paths that lead to __acct_update_integrals
> already have irqs disabled, and __acct_update_integrals does not need
> to disable irqs itself.
> 
> This is very convenient since about half the CPU time left in this
> function was spent in local_irq_save alone.
> 
> Performance of a microbenchmark that calls an invalid syscall
> ten million times in a row on a nohz_full CPU improves 21% vs.
> 4.5-rc1 with both the removal of divisions from __acct_update_integrals
> and this patch, with runtime dropping from 3.7 to 2.9 seconds.
> 
> With these patches applied, the highest remaining cpu user in
> the trace is native_sched_clock, which is addressed in the next
> patch.
> 
> Suggested-by: Peter Zijlstra <peterz@...radead.org>
> Signed-off-by: Rik van Riel <riel@...hat.com>
> ---
>  kernel/tsacct.c | 6 +-----
>  1 file changed, 1 insertion(+), 5 deletions(-)
> 
> diff --git a/kernel/tsacct.c b/kernel/tsacct.c
> index 8908f8b1d26e..b2663d699a72 100644
> --- a/kernel/tsacct.c
> +++ b/kernel/tsacct.c
> @@ -124,26 +124,22 @@ static void __acct_update_integrals(struct task_struct *tsk,
>  				    cputime_t utime, cputime_t stime)
>  {
>  	cputime_t time, dtime;
> -	unsigned long flags;
>  	u64 delta;
>  
>  	if (unlikely(!tsk->mm))
>  		return;
>  
> -	local_irq_save(flags);
>  	time = stime + utime;
>  	dtime = time - tsk->acct_timexpd;
>  	delta = cputime_to_nsecs(dtime);
>  
>  	if (delta < TICK_NSEC)
> -		goto out;
> +		return;
>  
>  	tsk->acct_timexpd = time;
>  	/* The final unit will be Mbyte-usecs, see xacct_add_tsk */
>  	tsk->acct_rss_mem1 += delta * get_mm_rss(tsk->mm) / 1024;
>  	tsk->acct_vm_mem1 += delta * tsk->mm->total_vm / 1024;
> -out:
> -	local_irq_restore(flags);
>  }

I think you need this as well, because do_exit() probably doesn't have
irqs disabled at that point:

diff --git a/kernel/tsacct.c b/kernel/tsacct.c
index 975cb49..12c6047 100644
--- a/kernel/tsacct.c
+++ b/kernel/tsacct.c
@@ -153,9 +153,12 @@ static void __acct_update_integrals(struct task_struct *tsk,
 void acct_update_integrals(struct task_struct *tsk)
 {
 	cputime_t utime, stime;
+	unsigned long flags;
 
 	task_cputime(tsk, &utime, &stime);
+	local_irq_save(flags);
 	__acct_update_integrals(tsk, utime, stime);
+	local_irq_restore(flags);
 }
 
 /**