linux-kernel - Re: [PATCH v4 1/2] rcu: Eliminate rcu_state.nocb_is

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: <20220822163411.GF6159@paulmck-ThinkPad-P17-Gen-1>
Date:   Mon, 22 Aug 2022 09:34:11 -0700
From:   "Paul E. McKenney" <paulmck@...nel.org>
To:     Zhen Lei <thunder.leizhen@...wei.com>
Cc:     Frederic Weisbecker <frederic@...nel.org>,
        Neeraj Upadhyay <quic_neeraju@...cinc.com>,
        Josh Triplett <josh@...htriplett.org>,
        Steven Rostedt <rostedt@...dmis.org>,
        Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
        Lai Jiangshan <jiangshanlai@...il.com>,
        Joel Fernandes <joel@...lfernandes.org>, rcu@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v4 1/2] rcu: Eliminate rcu_state.nocb_is_setup

On Wed, Aug 17, 2022 at 09:42:52AM +0800, Zhen Lei wrote:
> 'rcu_state.nocb_is_setup' is initialized to true only if 'rcu_nocb_mask'
> successfully allocates memory. So it can be replaced by
> 'cpumask_available(rcu_nocb_mask)'. More importantly, the latter is more
> intuitive, and it has been used in several places.
> 
> Signed-off-by: Zhen Lei <thunder.leizhen@...wei.com>

One of the implementations of cpumask_available() does indeed check
for NULL.  But here is the other one:

static inline bool cpumask_available(cpumask_var_t mask)
{
	return true;
}

So I have to ask...  In a kernel built with CONFIG_CPUMASK_OFFSTACK=n,
will this change really work?

Another important question is "Do all of the existing uses of
cpumask_available() really work?"  Yes, I do believe that they do
work, but it would be good to get another set of eyes on that code.
"All software developers are blind!"  ;-)

							Thanx, Paul

> ---
>  kernel/rcu/tree.h      | 1 -
>  kernel/rcu/tree_nocb.h | 8 +++-----
>  2 files changed, 3 insertions(+), 6 deletions(-)
> 
> diff --git a/kernel/rcu/tree.h b/kernel/rcu/tree.h
> index d4a97e40ea9c3e2..06f659c63d2d192 100644
> --- a/kernel/rcu/tree.h
> +++ b/kernel/rcu/tree.h
> @@ -375,7 +375,6 @@ struct rcu_state {
>  	arch_spinlock_t ofl_lock ____cacheline_internodealigned_in_smp;
>  						/* Synchronize offline with */
>  						/*  GP pre-initialization. */
> -	int nocb_is_setup;			/* nocb is setup from boot */
>  };
>  
>  /* Values for rcu_state structure's gp_flags field. */
> diff --git a/kernel/rcu/tree_nocb.h b/kernel/rcu/tree_nocb.h
> index 0a5f0ef41484518..ff763e7dc53551f 100644
> --- a/kernel/rcu/tree_nocb.h
> +++ b/kernel/rcu/tree_nocb.h
> @@ -69,7 +69,6 @@ static int __init rcu_nocb_setup(char *str)
>  			cpumask_setall(rcu_nocb_mask);
>  		}
>  	}
> -	rcu_state.nocb_is_setup = true;
>  	return 1;
>  }
>  __setup("rcu_nocbs", rcu_nocb_setup);
> @@ -1215,7 +1214,7 @@ void __init rcu_init_nohz(void)
>  	struct rcu_data *rdp;
>  
>  #if defined(CONFIG_RCU_NOCB_CPU_DEFAULT_ALL)
> -	if (!rcu_state.nocb_is_setup) {
> +	if (!cpumask_available(rcu_nocb_mask)) {
>  		need_rcu_nocb_mask = true;
>  		offload_all = true;
>  	}
> @@ -1235,10 +1234,9 @@ void __init rcu_init_nohz(void)
>  				return;
>  			}
>  		}
> -		rcu_state.nocb_is_setup = true;
>  	}
>  
> -	if (!rcu_state.nocb_is_setup)
> +	if (!cpumask_available(rcu_nocb_mask))
>  		return;
>  
>  #if defined(CONFIG_NO_HZ_FULL)
> @@ -1299,7 +1297,7 @@ static void rcu_spawn_cpu_nocb_kthread(int cpu)
>  	struct task_struct *t;
>  	struct sched_param sp;
>  
> -	if (!rcu_scheduler_fully_active || !rcu_state.nocb_is_setup)
> +	if (!rcu_scheduler_fully_active || !cpumask_available(rcu_nocb_mask))
>  		return;
>  
>  	/* If there already is an rcuo kthread, then nothing to do. */
> -- 
> 2.25.1
>