linux-kernel - Re: [PATCH] sched: cgroup: Move task_can

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: <20251003124300.4698dd18@gandalf.local.home>
Date: Fri, 3 Oct 2025 12:43:00 -0400
From: Steven Rostedt <rostedt@...dmis.org>
To: Andrea Righi <arighi@...dia.com>
Cc: LKML <linux-kernel@...r.kernel.org>, Peter Zijlstra
 <peterz@...radead.org>, Ingo Molnar <mingo@...nel.org>, Juri Lelli
 <juri.lelli@...hat.com>, Vincent Guittot <vincent.guittot@...aro.org>,
 Waiman Long <longman@...hat.com>, Tejun Heo <tj@...nel.org>, Johannes
 Weiner <hannes@...xchg.org>, Michal Koutný
 <mkoutny@...e.com>, Julia Lawall <Julia.Lawall@...ia.fr>, "Luis Claudio R.
 Goncalves" <lgoncalv@...hat.com>, Joseph Salisbury
 <joseph.salisbury@...cle.com>
Subject: Re: [PATCH] sched: cgroup: Move task_can_attach() to cpuset.c

On Fri, 3 Oct 2025 18:36:24 +0200
Andrea Righi <arighi@...dia.com> wrote:

> Hi Steven,
> 
> On Fri, Oct 03, 2025 at 12:14:21PM -0400, Steven Rostedt wrote:
> > From: Steven Rostedt <rostedt@...dmis.org>
> > 
> > At our monthly stable meeting, we were talking about documenting non
> > static functions and randomly picked a function to look at. That was
> > task_can_attach(). It was then noticed that it's only used by
> > cgroup/cpuset.c and nothing else. It's a simple function that doesn't
> > reference anything unique to sched/core.c, hence there's no reason that
> > function should be there.
> > 
> > Move it to cgroup/cpuset.c as that's the only place it is used. Also make
> > it a static inline as it is so small.
> > 
> > Signed-off-by: Steven Rostedt (Google) <rostedt@...dmis.org>  
> 
> Nice cleanup. :)

Thanks, since you also were the first to notice it is only used in one spot.

> 
> Apparently it became that small with commit 2ef269ef1ac00 ("cgroup/cpuset:
> Free DL BW in case can_attach() fails"), maybe we can mention that in the
> commit message?

Sure.

> 
> > ---
> >  include/linux/sched.h  |  1 -
> >  kernel/cgroup/cpuset.c | 19 +++++++++++++++++++
> >  kernel/sched/core.c    | 19 -------------------
> >  3 files changed, 19 insertions(+), 20 deletions(-)
> > 
> > diff --git a/include/linux/sched.h b/include/linux/sched.h
> > index e4ce0a76831e..4ee4fa973eda 100644
> > --- a/include/linux/sched.h
> > +++ b/include/linux/sched.h
> > @@ -1849,7 +1849,6 @@ current_restore_flags(unsigned long orig_flags, unsigned long flags)
> >  }
> >  
> >  extern int cpuset_cpumask_can_shrink(const struct cpumask *cur, const struct cpumask *trial);
> > -extern int task_can_attach(struct task_struct *p);
> >  extern int dl_bw_alloc(int cpu, u64 dl_bw);
> >  extern void dl_bw_free(int cpu, u64 dl_bw);
> >  
> > diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
> > index 27adb04df675..21fe872803e8 100644
> > --- a/kernel/cgroup/cpuset.c
> > +++ b/kernel/cgroup/cpuset.c
> > @@ -3009,6 +3009,25 @@ static void reset_migrate_dl_data(struct cpuset *cs)
> >  	cs->sum_migrate_dl_bw = 0;
> >  }
> >  
> > +static inline int task_can_attach(struct task_struct *p)
> > +{
> > +	int ret = 0;
> > +
> > +	/*
> > +	 * Kthreads which disallow setaffinity shouldn't be moved
> > +	 * to a new cpuset; we don't want to change their CPU
> > +	 * affinity and isolating such threads by their set of
> > +	 * allowed nodes is unnecessary.  Thus, cpusets are not
> > +	 * applicable for such threads.  This prevents checking for
> > +	 * success of set_cpus_allowed_ptr() on all attached tasks
> > +	 * before cpus_mask may be changed.
> > +	 */
> > +	if (p->flags & PF_NO_SETAFFINITY)
> > +		ret = -EINVAL;
> > +
> > +	return ret;  
> 
> As we're cleaning up, we could just return -EINVAL and 0 directly and get
> rid of that ret variable.

That should be a separate patch. Moves should really not do much else. I
even wondered about making it a static inline too, but figured that wasn't
touching the logic, and the function was going to become static anyway.

-- Steve


> 
> > +}
> > +
> >  /* Called by cgroups to determine if a cpuset is usable; cpuset_mutex held */
> >  static int cpuset_can_attach(struct cgroup_taskset *tset)
> >  {
> > diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> > index ccba6fc3c3fe..a195c4b25475 100644
> > --- a/kernel/sched/core.c
> > +++ b/kernel/sched/core.c
> > @@ -8070,25 +8070,6 @@ int cpuset_cpumask_can_shrink(const struct cpumask *cur,
> >  	return ret;
> >  }
> >  
> > -int task_can_attach(struct task_struct *p)
> > -{
> > -	int ret = 0;
> > -
> > -	/*
> > -	 * Kthreads which disallow setaffinity shouldn't be moved
> > -	 * to a new cpuset; we don't want to change their CPU
> > -	 * affinity and isolating such threads by their set of
> > -	 * allowed nodes is unnecessary.  Thus, cpusets are not
> > -	 * applicable for such threads.  This prevents checking for
> > -	 * success of set_cpus_allowed_ptr() on all attached tasks
> > -	 * before cpus_mask may be changed.
> > -	 */
> > -	if (p->flags & PF_NO_SETAFFINITY)
> > -		ret = -EINVAL;
> > -
> > -	return ret;
> > -}
> > -
> >  bool sched_smp_initialized __read_mostly;
> >  
> >  #ifdef CONFIG_NUMA_BALANCING
> > -- 
> > 2.50.1
> >   
> 
> Thanks,
> -Andrea