linux-kernel - Re: [PATCH 2/6] sched/vtime: Bring up complete kcpustat accessor

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 Dec 2019 02:08:40 +0100
From:   Frederic Weisbecker <frederic@...nel.org>
To:     Chris Wilson <chris@...is-wilson.co.uk>
Cc:     Ingo Molnar <mingo@...nel.org>,
        Peter Zijlstra <peterz@...radead.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Jacek Anaszewski <jacek.anaszewski@...il.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        "Rafael J . Wysocki" <rjw@...ysocki.net>,
        Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        Rik van Riel <riel@...riel.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Yauheni Kaliuta <yauheni.kaliuta@...hat.com>,
        Viresh Kumar <viresh.kumar@...aro.org>,
        Pavel Machek <pavel@....cz>
Subject: Re: [PATCH 2/6] sched/vtime: Bring up complete kcpustat accessor

On Sat, Dec 28, 2019 at 08:56:19PM +0000, Chris Wilson wrote:
> I'm randomly hitting this WARN on a non-virtualised system reading
> /proc/stat.
> 
> vtime->state is updated under the write_seqcount, so the access here is
> deliberately racey, and the change in vtime->state would be picked up
> the seqcount_retry.
> 
> Quick suggestion would be something along the lines of
> 
>  static int vtime_state_check(struct vtime *vtime, int cpu)
>  {
> +	int state = READ_ONCE(vtime->state);
> +
>  	/*
>  	 * We raced against a context switch, fetch the
>  	 * kcpustat task again.
> @@ -930,10 +932,10 @@ static int vtime_state_check(struct vtime *vtime, int cpu)
>  	 *
>  	 * Case 1) is ok but 2) is not. So wait for a safe VTIME state.
>  	 */
> -	if (vtime->state == VTIME_INACTIVE)
> +	if (state == VTIME_INACTIVE)
>  		return -EAGAIN;
> 
> -	return 0;
> +	return state;
>  }
> 
>  static u64 kcpustat_user_vtime(struct vtime *vtime)
> @@ -1055,7 +1057,7 @@ static int kcpustat_cpu_fetch_vtime(struct kernel_cpustat *dst,
>  		cpustat = dst->cpustat;
> 
>  		/* Task is sleeping, dead or idle, nothing to add */
> -		if (vtime->state < VTIME_SYS)
> +		if (err < VTIME_SYS)
>  			continue;
> 
>  		delta = vtime_delta(vtime);
> @@ -1064,15 +1066,15 @@ static int kcpustat_cpu_fetch_vtime(struct kernel_cpustat *dst,
>  		 * Task runs either in user (including guest) or kernel space,
>  		 * add pending nohz time to the right place.
>  		 */
> -		if (vtime->state == VTIME_SYS) {
> +		if (err == VTIME_SYS) {
>  			cpustat[CPUTIME_SYSTEM] += vtime->stime + delta;
> -		} else if (vtime->state == VTIME_USER) {
> +		} else if (err == VTIME_USER) {
>  			if (task_nice(tsk) > 0)
>  				cpustat[CPUTIME_NICE] += vtime->utime + delta;
>  			else
>  				cpustat[CPUTIME_USER] += vtime->utime + delta;
>  		} else {
> -			WARN_ON_ONCE(vtime->state != VTIME_GUEST);
> +			WARN_ON_ONCE(err != VTIME_GUEST);
>  			if (task_nice(tsk) > 0) {
>  				cpustat[CPUTIME_GUEST_NICE] += vtime->gtime + delta;
>  				cpustat[CPUTIME_NICE] += vtime->gtime + delta;
> 
> Or drop the warn.

Good catch, can I use your Signed-off-by ?

Thanks.