Re: [PATCH 1/1] Fix: trace sched switch start/stop racy updates

From: "Paul E. McKenney" <paulmck@linux.ibm.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	Valentin Schneider <valentin.schneider@arm.com>,
	Joel Fernandes <joel@joelfernandes.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Alan Stern <stern@rowland.harvard.edu>,
	Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
	rostedt <rostedt@goodmis.org>,
	linux-kernel <linux-kernel@vger.kernel.org>,
	Boqun Feng <boqun.feng@gmail.com>,
	Will Deacon <will.deacon@arm.com>,
	David Howells <dhowells@redhat.com>
Subject: Re: [PATCH 1/1] Fix: trace sched switch start/stop racy updates
Date: Tue, 20 Aug 2019 13:31:35 -0700	[thread overview]
Message-ID: <20190820203135.GX28441@linux.ibm.com> (raw)
In-Reply-To: <20190820140116.GT2332@hirez.programming.kicks-ass.net>

On Tue, Aug 20, 2019 at 04:01:16PM +0200, Peter Zijlstra wrote:
> On Fri, Aug 16, 2019 at 09:52:17PM -0700, Paul E. McKenney wrote:
> > On Fri, Aug 16, 2019 at 03:57:43PM -0700, Linus Torvalds wrote:
> > 
> > [ . . . ]
> > 
> > > We add READ_ONCE and WRITE_ONCE annotations when they make sense. Not
> > > because of some theoretical "compiler is free to do garbage"
> > > arguments. If such garbage happens, we need to fix the compiler, the
> > > same way we already do with
> > > 
> > >   -fno-strict-aliasing
> > 
> > Yeah, the compete-with-FORTRAN stuff.  :-/
> > 
> > There is some work going on in the C committee on this, where the
> > theorists would like to restrict strict-alias based optimizations to
> > enable better analysis tooling.  And no, although the theorists are
> > pushing in the direction we would like them to, as far as I can see
> > they are not pushing as far as we would like.  But it might be that
> > -fno-strict-aliasing needs some upgrades as well.  I expect to learn
> > more at the next meeting in a few months.
> > 
> > http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2364.pdf
> > http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2363.pdf
> > http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2362.pdf
> 
> We really should get the compiler folks to give us a
> -fno-pointer-provenance. Waiting on the standards committee to get their
> act together seems unlikely, esp. given that some people actually seem
> to _want_ this nonsense :/

The reason that they want it is to enable some significant optimizations
in numerical code on the one hand and in heavily templated C++ code on
the other.  Neither of which has much bearing on kernel code.

Interested in coming to the next C standards committee meeting in October
to help me push for this?  ;-)

							Thanx, Paul