Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Thu, 13 Feb 2020 17:40:31 +0100
From:   Peter Zijlstra <peterz@infradead.org>
To:     "Paul E. McKenney" <paulmck@kernel.org>
Cc:     Joel Fernandes <joel@joelfernandes.org>,
        linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org,
        rostedt@goodmis.org, mingo@kernel.org, gregkh@linuxfoundation.org,
        gustavo@embeddedor.com, tglx@linutronix.de, josh@joshtriplett.org,
        mathieu.desnoyers@efficios.com, jiangshanlai@gmail.com
Subject: Re: [PATCH v2 3/9] rcu,tracing: Create trace_rcu_{enter,exit}()
Message-ID: <20200213164031.GH14914@hirez.programming.kicks-ass.net>
References: <20200212210139.382424693@infradead.org>
 <20200212210749.971717428@infradead.org>
 <20200212232005.GC115917@google.com>
 <20200213082716.GI14897@hirez.programming.kicks-ass.net>
 <20200213135138.GB2935@paulmck-ThinkPad-P72>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20200213135138.GB2935@paulmck-ThinkPad-P72>
User-Agent: Mutt/1.10.1 (2018-07-13)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Feb 13, 2020 at 05:51:38AM -0800, Paul E. McKenney wrote:

> The reason for the irq argument is to avoid invoking
> rcu_prepare_for_idle() and rcu_dynticks_task_enter() from NMI context
> from rcu_nmi_exit_common().  Similarly, we need to avoid invoking
> rcu_dynticks_task_exit() and rcu_cleanup_after_idle() from NMI context
> from rcu_nmi_enter_common().

Aaah, I see. I didn't grep hard enough earlier today (I only found
stubs). Yes, those take locks, we mustn't call them from NMI context.

> It might well be that I could make these functions be NMI-safe, but
> rcu_prepare_for_idle() in particular would be a bit ugly at best.
> So, before looking into that, I have a question.  Given these proposed
> changes, will rcu_nmi_exit_common() and rcu_nmi_enter_common() be able
> to just use in_nmi()?

That _should_ already be the case today. That is, if we end up in a
tracer and in_nmi() is unreliable we're already screwed anyway.