linux-arch.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Boqun Feng <boqun.feng@gmail.com>
To: Alan Stern <stern@rowland.harvard.edu>
Cc: Peter Zijlstra <peterz@infradead.org>,
	"Paul E. McKenney" <paulmck@kernel.org>,
	parri.andrea@gmail.com, will@kernel.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,
	linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org
Subject: Re: "Verifying and Optimizing Compact NUMA-Aware Locks on Weak Memory Models"
Date: Sat, 27 Aug 2022 09:44:17 -0700	[thread overview]
Message-ID: <YwpJ4ZPVbuCnnFKS@boqun-archlinux> (raw)
In-Reply-To: <YwpAzTwSRCK5kdLN@rowland.harvard.edu>

On Sat, Aug 27, 2022 at 12:05:33PM -0400, Alan Stern wrote:
> On Sat, Aug 27, 2022 at 01:47:48AM +0200, Peter Zijlstra wrote:
> > On Fri, Aug 26, 2022 at 01:10:39PM -0400, Alan Stern wrote:
> > 
> > > >  - some babbling about a missing propagation -- ISTR Linux if stuffed
> > > >    full of them, specifically we require stores to auto propagate
> > > >    without help from barriers
> > > 
> > > Not a missing propagation; a late one.
> > > 
> > > Don't understand what you mean by "auto propagate without help from 
> > > barriers".
> > 
> > Linux hard relies on:
> > 
> > 	CPU0				CPU1
> > 
> > 	WRITE_ONCE(foo, 1);		while (!READ_ONCE(foo));
> > 
> > making forward progress.
> 
> Indeed yes.  As far as I can tell, this requirement is not explicitly 
> mentioned in the LKMM, although it certainly is implicit.  I can't even 
> think of a way to express it in a form Herd could verify.
> 

FWIW, C++ defines this as (in https://eel.is/c++draft/atomics#order-11):

	Implementations should make atomic stores visible to atomic
	loads within a reasonable amount of time.

in other words:

if one thread does an atomic store, then all other threads must see that
store eventually.

(from: https://rust-lang.zulipchat.com/#narrow/stream/136281-t-lang.2Fwg-unsafe-code-guidelines/topic/Rust.20forward.20progress.20guarantees/near/294702950)

Should we add something somewhere in our model, maybe in the
explanation.txt?

Plus, I think we cannot express this in Herd because Herd uses
graph-based model (axiomatic model) instead of an operational model to
describe the model: axiomatic model cannot describe "something will
eventually happen". There was also some discussion in the zulip steam
of Rust unsafe-code-guidelines.

Regards,
Boqun

> > There were a few 'funny' uarchs that were broken, see for example commit
> > a30718868915f.
> 
> Ha!  That commit should be a lesson in something, although I'm not sure 
> what.  :-)
> 
> Alan

  reply	other threads:[~2022-08-27 16:45 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-08-26 12:48 "Verifying and Optimizing Compact NUMA-Aware Locks on Weak Memory Models" Paul E. McKenney
2022-08-26 16:21 ` Boqun Feng
2022-08-26 16:23 ` Peter Zijlstra
2022-08-26 17:10   ` Alan Stern
2022-08-26 20:42     ` Paul E. McKenney
2022-08-27 16:00       ` Alan Stern
2022-08-27 17:04         ` Paul E. McKenney
2022-09-13 11:24       ` Will Deacon
2022-09-13 12:21         ` Dan Lustig
2022-09-16  8:18           ` Will Deacon
2022-08-26 23:47     ` Peter Zijlstra
2022-08-27 16:05       ` Alan Stern
2022-08-27 16:44         ` Boqun Feng [this message]
2022-08-29  2:15           ` Andrea Parri
2022-09-09 11:45           ` Jonas Oberhauser
2022-09-10 12:11             ` Hernan Luis Ponce de Leon
2022-09-10 15:03               ` Alan Stern
2022-09-10 20:41                 ` Hernan Luis Ponce de Leon
2022-09-11 10:20                   ` Joel Fernandes
2022-09-12 10:13                     ` Jonas Oberhauser
2022-09-12 11:10                       ` Hernan Luis Ponce de Leon
2022-09-14 14:41                         ` Alan Stern
2022-09-12 12:01                       ` Alan Stern
2022-09-11 14:53                   ` Andrea Parri
2022-09-12 10:46                 ` Jonas Oberhauser
2022-09-12 12:02                   ` Alan Stern
2022-08-29  2:33 ` Andrea Parri
2022-08-29 12:25   ` Paul E. McKenney

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=YwpJ4ZPVbuCnnFKS@boqun-archlinux \
    --to=boqun.feng@gmail.com \
    --cc=akiyks@gmail.com \
    --cc=dhowells@redhat.com \
    --cc=dlustig@nvidia.com \
    --cc=j.alglave@ucl.ac.uk \
    --cc=joel@joelfernandes.org \
    --cc=linux-arch@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=luc.maranget@inria.fr \
    --cc=npiggin@gmail.com \
    --cc=parri.andrea@gmail.com \
    --cc=paulmck@kernel.org \
    --cc=peterz@infradead.org \
    --cc=stern@rowland.harvard.edu \
    --cc=will@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).