Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Uladzislau Rezki <urezki@gmail.com>
Date:   Wed, 22 Apr 2020 12:35:36 +0200
To:     "Joel Fernandes (Google)" <joel@joelfernandes.org>
Cc:     linux-kernel@vger.kernel.org,
        Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
        Uladzislau Rezki <urezki@gmail.com>,
        Josh Triplett <josh@joshtriplett.org>,
        Lai Jiangshan <jiangshanlai@gmail.com>,
        Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
        "Paul E. McKenney" <paulmck@kernel.org>, rcu@vger.kernel.org,
        Steven Rostedt <rostedt@goodmis.org>
Subject: Re: [PATCH rcu/dev -fixes 2/4] rcu/tree: Skip entry into the page
 allocator for PREEMPT_RT
Message-ID: <20200422103536.GA16263@pc636>
References: <20200420153837.194532-1-joel@joelfernandes.org>
 <20200420153837.194532-3-joel@joelfernandes.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20200420153837.194532-3-joel@joelfernandes.org>
User-Agent: Mutt/1.10.1 (2018-07-13)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Apr 20, 2020 at 11:38:35AM -0400, Joel Fernandes (Google) wrote:
> To keep kfree_rcu() path working on raw non-preemptible sections,
> prevent the optional entry into the allocator as it uses sleeping locks.
> In fact, even if the caller of kfree_rcu() is preemptible, this path
> still is not, as krcp->lock is a raw spinlock as done in previous
> patches. With additional page pre-allocation in the works, hitting this
> return is going to be much less likely soon so just prevent it for now
> so that PREEMPT_RT does not break. Note that page allocation here is an
> optimization and skipping it still makes kfree_rcu() work.
> 
> Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> Co-developed-by: Uladzislau Rezki <urezki@gmail.com>
> Signed-off-by: Uladzislau Rezki <urezki@gmail.com>
> Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
> ---
>  kernel/rcu/tree.c | 12 ++++++++++++
>  1 file changed, 12 insertions(+)
> 
> diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> index cf68d3d9f5b81..cd61649e1b001 100644
> --- a/kernel/rcu/tree.c
> +++ b/kernel/rcu/tree.c
> @@ -3092,6 +3092,18 @@ kfree_call_rcu_add_ptr_to_bulk(struct kfree_rcu_cpu *krcp,
>  		if (!bnode) {
>  			WARN_ON_ONCE(sizeof(struct kfree_rcu_bulk_data) > PAGE_SIZE);
>  
> +			/*
> +			 * To keep this path working on raw non-preemptible
> +			 * sections, prevent the optional entry into the
> +			 * allocator as it uses sleeping locks. In fact, even
> +			 * if the caller of kfree_rcu() is preemptible, this
> +			 * path still is not, as krcp->lock is a raw spinlock.
> +			 * With additional page pre-allocation in the works,
> +			 * hitting this return is going to be much less likely.
> +			 */
> +			if (IS_ENABLED(CONFIG_PREEMPT_RT))
> +				return false;
> +
>  			bnode = (struct kfree_rcu_bulk_data *)
>  				__get_free_page(GFP_NOWAIT | __GFP_NOWARN);
>  		}
This will not be correctly working by just reverting everything to the
"rcu_head path" for CONFIG_PREEMPT_RT kernel. We need to preallocate at
least once. I can add caching on top of this series to address it.

--
Vlad Rezki