Re: [PATCH] vsprintf: Fix memory barriers of ptr_key to have_filed_random_ptr_key

From: Linus Torvalds <torvalds@linux-foundation.org>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Kees Cook <keescook@chromium.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	tcharding <me@tobin.cc>
Subject: Re: [PATCH] vsprintf: Fix memory barriers of ptr_key to have_filed_random_ptr_key
Date: Tue, 15 May 2018 09:55:13 -0700	[thread overview]
Message-ID: <CA+55aFyBHRmXLo07f_7qt7szaDvppMStfX5Sq9Df6mZSy+qYvA@mail.gmail.com> (raw)
In-Reply-To: <20180515100558.21df515e@gandalf.local.home>

On Tue, May 15, 2018 at 7:06 AM Steven Rostedt <rostedt@goodmis.org> wrote:
> -       smp_mb();
> +       smp_wmb();
>          WRITE_ONCE(have_filled_random_ptr_key, true);

> +       /* Read ptr_key after reading have_filled_random_ptr_key */
> +       smp_rmb();
> +
>   #ifdef CONFIG_64BIT
>          hashval = (unsigned long)siphash_1u64((u64)ptr, &ptr_key);

Hmm. smp_wmb/rmb are basically free on x86, but on some architectures
smp_rmb() in particular can be pretty expensive.

So when you have a "handoff" situation like this, it's _probably_ better to
use use "smp_store_release()" and "smp_load_acquire()". To some degree that
might also be better for documentation purposes, because that's exactly the
"release-acquire" pattern.

That said, I'm not convinced this really matters all that much for a
boot-time flag like this. The race is pretty theoretical.

                    Linus