Received-SPF: pass (google.com: domain of linux-kernel+bounces-202729-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Date: Wed, 5 Jun 2024 09:40:20 -0400
From: Alan Stern <stern@rowland.harvard.edu>
To: Boqun Feng <boqun.feng@gmail.com>
Cc: Jonas Oberhauser <jonas.oberhauser@huaweicloud.com>, paulmck@kernel.org,
  parri.andrea@gmail.com, will@kernel.org, peterz@infradead.org,
  npiggin@gmail.com, dhowells@redhat.com, j.alglave@ucl.ac.uk,
  luc.maranget@inria.fr, akiyks@gmail.com, dlustig@nvidia.com,
  joel@joelfernandes.org, urezki@gmail.com, quic_neeraju@quicinc.com,
  frederic@kernel.org, linux-kernel@vger.kernel.org,
  Viktor Vafeiadis <viktor@mpi-sws.org>
Subject: Re: [PATCHv2 3/4] tools/memory-model: Define effect of Mb tags on
 RMWs in tools/...
Message-ID: <d8fb64c8-ec53-4087-bbbc-72c7e59d8dbf@rowland.harvard.edu>
References: <20240604152922.495908-1-jonas.oberhauser@huaweicloud.com>
 <20240604160440.498332-1-jonas.oberhauser@huaweicloud.com>
 <Zl_pehpjHVt2z95p@Boquns-Mac-mini.home>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Zl_pehpjHVt2z95p@Boquns-Mac-mini.home>

On Tue, Jun 04, 2024 at 09:28:42PM -0700, Boqun Feng wrote:
> On Tue, Jun 04, 2024 at 06:04:40PM +0200, Jonas Oberhauser wrote:
> > --- a/tools/memory-model/linux-kernel.cat
> > +++ b/tools/memory-model/linux-kernel.cat
> > @@ -34,6 +34,16 @@ let R4rmb = R \ Noreturn	(* Reads for which rmb works *)
> >  let rmb = [R4rmb] ; fencerel(Rmb) ; [R4rmb]
> >  let wmb = [W] ; fencerel(Wmb) ; [W]
> >  let mb = ([M] ; fencerel(Mb) ; [M]) |
> > +	(*
> > +	 * full-barrier RMWs (successful cmpxchg(), xchg(), etc.) act as
> > +	 * though there were enclosed by smp_mb().
> > +	 * The effect of these virtual smp_mb() is formalized by adding
> > +	 * Mb tags to the read and write of the operation, and providing
> > +	 * the same ordering as though there were additional po edges
> > +	 * between the Mb tag and the read resp. write.
> > +	 *)
> > +	([M] ; po ; [Mb & R]) |
> > +	([Mb & W] ; po ; [M]) |
> 
> I couldn't help suggestting:
> 
> 	([M] ; po ; [Mb & domain(rmw)]) |
> 	([Mb & range(rmw)] ; po ; [M]) |
> 
> , it's a bit more clear to me, but maybe the comment above is good
> enough?

We may want to use the patch's approach for other things besides RMW.  
For instance, it would be a good way to implement smp_store_mb() -- 
compare it to the existing implementation in the .def file.

Alan