linux-kernel - Re: [PATCH 3/3] cgroup: Implement cgroup2 basic CPU usage accounting

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:   Thu, 17 Aug 2017 06:13:59 -0700
From:   Tejun Heo <tj@...nel.org>
To:     Roman Gushchin <guro@...com>
Cc:     lizefan@...wei.com, hannes@...xchg.org, peterz@...radead.org,
        mingo@...hat.com, longman@...hat.com, cgroups@...r.kernel.org,
        linux-kernel@...r.kernel.org, kernel-team@...com, pjt@...gle.com,
        luto@...capital.net, efault@....de, torvalds@...ux-foundation.org
Subject: Re: [PATCH 3/3] cgroup: Implement cgroup2 basic CPU usage accounting

Hello,

On Thu, Aug 17, 2017 at 01:07:41PM +0100, Roman Gushchin wrote:
> On Fri, Aug 11, 2017 at 09:37:54AM -0700, Tejun Heo wrote:
> > In cgroup1, while cpuacct isn't actually controlling any resources, it
> > is a separate controller due to combinaton of two factors -
> 
> s/combinaton/combination

Fixed.

> > @@ -4466,6 +4470,8 @@ static void css_free_work_fn(struct work_struct *work)
> >  			 */
> >  			cgroup_put(cgroup_parent(cgrp));
> >  			kernfs_put(cgrp->kn);
> > +			if (cgroup_on_dfl(cgrp))
> > +				cgroup_stat_exit(cgrp);
> 
> It looks like this "if (cgroup_on_dfl(cgrp))" works here and further similar to
> "#ifdef CGROUP_V2". I wonder, if it's better to move this check into the
> calling function: cgroup_stat_exit() in this case.

I have a slight preference to keeping these topology-aware tests on
the core / management part of code because that makes it obvious that
these stats aren't available for all cgroups.  Also, during cgroup
creation, because @cgrp isn't linked to its parent yet, we'd have to
pass @parent to cgroup_stat_init/exit() too.

> > +void cgroup_stat_show_cputime(struct seq_file *seq, const char *prefix)
> > +{
> 
> What are any other possible prefix values except "cpu."?

Empty string when the stats are exposed through cpu.stat.

> > +void __init cgroup_stat_boot(void)
> > +{
> > +	int cpu;
> > +
> > +	for_each_possible_cpu(cpu)
> > +		raw_spin_lock_init(per_cpu_ptr(&cgroup_cpu_stat_lock, cpu));
> > +
> > +	WARN_ON(cgroup_stat_init(&cgrp_dfl_root.cgrp));
> 
> I'm not sure WARN_ON() is enough here: if cgroup_stat_init() returned -ENOMEM,
> the following OOPS is not avoidable, as you don't check cpu_stat pointer.
> But it's very unlikely, of course.

Sure, will switch to BUG_ON().

Thanks.

-- 
tejun