From: Alan Stern <stern@rowland.harvard.edu>
To: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
linux-arch <linux-arch@vger.kernel.org>,
<andrea.parri@amarulasolutions.com>,
Will Deacon <will.deacon@arm.com>,
Peter Zijlstra <peterz@infradead.org>,
Boqun Feng <boqun.feng@gmail.com>,
Nick Piggin <npiggin@gmail.com>,
David Howells <dhowells@redhat.com>,
Jade Alglave <j.alglave@ucl.ac.uk>,
Luc Maranget <luc.maranget@inria.fr>,
Akira Yokosawa <akiyks@gmail.com>, Ingo Molnar <mingo@kernel.org>,
Roman Pen <roman.penyaev@profitbricks.com>
Subject: Re: LKMM litmus test for Roman Penyaev's rcu-rr
Date: Wed, 30 May 2018 16:28:56 -0400 (EDT) [thread overview]
Message-ID: <Pine.LNX.4.44L0.1805301620020.1502-100000@iolanthe.rowland.org> (raw)
In-Reply-To: <20180530194554.GM7063@linux.vnet.ibm.com>
On Wed, 30 May 2018, Paul E. McKenney wrote:
> > > My current guess is that we need to change the memory-model tool. If
> > > that is the case, here are some possible resolutions:
> > >
> > > 1. Make herd's C-language control dependencies work the same as its
> > > assembly language, so that they extend beyond the end of the
> > > "if" statement. I believe that this would make Roman's case
> > > work, but it could claim that other situations are safe that
> > > are actually problematic due to compiler optimizations.
> > >
> > > The fact that the model currently handles only READ_ONCE()
> > > and WRITE_ONCE() and not unmarked reads and writes make this
> > > option more attractive than it otherwise be, compilers not
> > > being allowed to reorder volatile accesses, but we are likely
> > > to introduce unmarked accesses sometime in the future.
> >
> > Preserving the order of volatile accesses isn't sufficient. The
> > compiler is still allowed to translate
> >
> > r1 = READ_ONCE(x);
> > if (r1) {
> > ...
> > }
> > WRITE_ONCE(y, r2);
> >
> > into something resembling
> >
> > r1 = READ_ONCE(x);
> > WRITE_ONCE(y, r2);
> > if (r1) {
> > ...
> > }
> >
> > (provided the "..." part doesn't contain any volatile accesses,
> > barriers, or anything affecting r2), which would destroy any run-time
> > control dependency. The CPU could then execute the write before the
> > read.
>
> True, but almost all current litmus tests do have at least one volatile
> access in their "if" statements. I am guessing that this has the same
> memory-model tooling issues as #2 below, but I am as usual happy to be
> proven wrong. ;-)
It shouldn't be all that bad. The dependencies are generated by herd,
meaning that the code would have to be changed to make control
dependencies extend beyond the ends of "if" statements. I don't think
the necessary changes would be tremendously big, especially since the
LISA front end already behaves this way.
> > > 2. Like #1 above, but only if something in one of the "if"'s
> > > branches would prevent the compiler from reordering
> > > (smp_mb(), synchronize_rcu(), value-returning non-relaxed
> > > RMW atomic, ...). Easy for me to say, but I am guessing
> > > that much violence would be needed to the tooling to make
> > > this work. ;-)
> >
> > This would be my preference. But I'm afraid it isn't practical at the
> > moment.
>
> I bet that some combination of scripting and smallish modifications to
> tooling could make it happen in reasonably short term. Might be more
> difficult to make something more future-proof, though, agreed.
I have no idea what sort of scripting/smallish modifications could do
the job. You could ask Luc, if you're not afraid of giving him an
aneurysm. :-)
> > > If I understand Alan correctly, there is not an obvious way to make
> > > this change within the confines of the memory model's .bell and .cat
> > > files.
> >
> > No way at all. It would require significant changes to herd's internal
> > workings and its external interface -- my original point.
>
> I was afraid of that. ;-)
>
> Though truth be told, I was expecting an issue like this to crop up
> sooner rather than later, so I was actually getting a bit nervous
> about the fact that it had not yet shown itself...
The fact is, herd was never meant to act like a compiler. Some
disagreements are inevitable.
Alan
next prev parent reply other threads:[~2018-05-30 20:28 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-05-28 22:08 LKMM litmus test for Roman Penyaev's rcu-rr Paul E. McKenney
2018-05-29 18:35 ` Alan Stern
2018-05-29 19:03 ` Paul E. McKenney
2018-05-29 20:49 ` Alan Stern
2018-05-29 21:10 ` Linus Torvalds
2018-05-29 22:53 ` Paul E. McKenney
2018-05-30 14:46 ` Alan Stern
2018-05-30 14:29 ` Alan Stern
2018-05-30 14:59 ` Linus Torvalds
2018-05-30 18:10 ` Alan Stern
2018-05-30 18:38 ` Paul E. McKenney
2018-05-30 19:08 ` Alan Stern
2018-05-30 19:45 ` Paul E. McKenney
2018-05-30 20:28 ` Alan Stern [this message]
2018-05-30 21:49 ` Paul E. McKenney
2018-05-30 22:01 ` Linus Torvalds
2018-05-30 23:14 ` Paul E. McKenney
2018-05-31 14:27 ` Alan Stern
2018-06-02 14:44 ` Paul E. McKenney
2018-06-04 14:17 ` Alan Stern
2018-06-04 16:01 ` Paul E. McKenney
2018-06-06 9:40 ` Roman Penyaev
2018-06-06 13:54 ` Alan Stern
2018-06-06 14:41 ` Roman Penyaev
2018-06-06 15:55 ` Alan Stern
2018-06-06 19:07 ` Paul E. McKenney
2018-06-06 19:23 ` Linus Torvalds
2018-06-07 9:43 ` Paul E. McKenney
2018-06-07 14:57 ` Alan Stern
2018-06-07 15:40 ` Linus Torvalds
2018-06-07 15:06 ` Linus Torvalds
2018-06-07 19:57 ` 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=Pine.LNX.4.44L0.1805301620020.1502-100000@iolanthe.rowland.org \
--to=stern@rowland.harvard.edu \
--cc=akiyks@gmail.com \
--cc=andrea.parri@amarulasolutions.com \
--cc=boqun.feng@gmail.com \
--cc=dhowells@redhat.com \
--cc=j.alglave@ucl.ac.uk \
--cc=linux-arch@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=luc.maranget@inria.fr \
--cc=mingo@kernel.org \
--cc=npiggin@gmail.com \
--cc=paulmck@linux.vnet.ibm.com \
--cc=peterz@infradead.org \
--cc=roman.penyaev@profitbricks.com \
--cc=torvalds@linux-foundation.org \
--cc=will.deacon@arm.com \
/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).