From: Andrea Parri <andrea.parri@amarulasolutions.com>
To: Daniel Jordan <daniel.m.jordan@oracle.com>
Cc: Steffen Klassert <steffen.klassert@secunet.com>,
Herbert Xu <herbert@gondor.apana.org.au>,
boqun.feng@gmail.com, paulmck@linux.ibm.com,
peterz@infradead.org, linux-arch@vger.kernel.org,
linux-crypto@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] padata: use smp_mb in padata_reorder to avoid orphaned padata jobs
Date: Tue, 16 Jul 2019 14:53:09 +0200 [thread overview]
Message-ID: <20190716125309.GA10672@andrea> (raw)
In-Reply-To: <20190712160737.iniaaxlsnhs6azg5@ca-dmjordan1.us.oracle.com>
Hi Daniel,
My two cents (summarizing some findings we discussed privately):
> I think adding the full barrier guarantees the following ordering, and the
> memory model people can correct me if I'm wrong:
>
> CPU21 CPU22
> ------------------------ --------------------------
> UNLOCK pd->lock
> smp_mb()
> LOAD reorder_objects
> INC reorder_objects
> spin_unlock(&pqueue->reorder.lock) // release barrier
> TRYLOCK pd->lock
>
> So if CPU22 has incremented reorder_objects but CPU21 reads 0 for it, CPU21
> should also have unlocked pd->lock so CPU22 can get it and serialize any
> remaining jobs.
This information inspired me to write down the following litmus test:
(AFAICT, you independently wrote a very similar test, which is indeed
quite reassuring! ;D)
C daniel-padata
{ }
P0(atomic_t *reorder_objects, spinlock_t *pd_lock)
{
int r0;
spin_lock(pd_lock);
spin_unlock(pd_lock);
smp_mb();
r0 = atomic_read(reorder_objects);
}
P1(atomic_t *reorder_objects, spinlock_t *pd_lock, spinlock_t *reorder_lock)
{
int r1;
spin_lock(reorder_lock);
atomic_inc(reorder_objects);
spin_unlock(reorder_lock);
//smp_mb();
r1 = spin_trylock(pd_lock);
}
exists (0:r0=0 /\ 1:r1=0)
It seems worth noticing that this test's "exists" clause is satisfiable
according to the (current) memory consistency model. (Informally, this
can be explained by noticing that the RELEASE from the spin_unlock() in
P1 does not provide any order between the atomic increment and the read
part of the spin_trylock() operation.) FWIW, uncommenting the smp_mb()
in P1 would suffice to prevent this clause from being satisfiable; I am
not sure, however, whether this approach is feasible or ideal... (sorry,
I'm definitely not too familiar with this code... ;/)
Thanks,
Andrea
next prev parent reply other threads:[~2019-07-16 12:53 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-07-11 22:12 [PATCH] padata: use smp_mb in padata_reorder to avoid orphaned padata jobs Daniel Jordan
2019-07-12 10:06 ` Herbert Xu
2019-07-12 10:10 ` Steffen Klassert
2019-07-12 16:07 ` Daniel Jordan
2019-07-13 5:03 ` Herbert Xu
2019-07-15 16:10 ` Daniel Jordan
2019-07-16 10:04 ` Herbert Xu
2019-07-16 11:14 ` Steffen Klassert
2019-07-16 12:57 ` [PATCH] padata: Use RCU when fetching pd from do_serial Herbert Xu
2019-07-16 13:09 ` Herbert Xu
2019-07-16 13:23 ` [v2 PATCH] " Herbert Xu
2019-07-17 8:36 ` Steffen Klassert
2019-07-17 8:50 ` Herbert Xu
2019-07-17 8:28 ` [PATCH] " Steffen Klassert
2019-07-17 8:47 ` Herbert Xu
2019-07-17 8:53 ` Steffen Klassert
2019-07-16 13:15 ` Peter Zijlstra
2019-07-16 13:18 ` Herbert Xu
2019-07-16 13:50 ` Peter Zijlstra
2019-07-16 13:52 ` Herbert Xu
2019-07-16 12:53 ` Andrea Parri [this message]
2019-07-16 13:13 ` [PATCH] padata: use smp_mb in padata_reorder to avoid orphaned padata jobs Peter Zijlstra
2019-07-16 15:01 ` Herbert Xu
2019-07-16 15:44 ` Daniel Jordan
2019-07-16 16:32 ` [PATCH v2] " Daniel Jordan
2019-07-17 11:11 ` [PATCH] padata: Replace delayed timer with immediate workqueue in padata_reorder Herbert Xu
2019-07-17 18:32 ` Daniel Jordan
2019-07-18 3:31 ` Herbert Xu
2019-07-18 14:27 ` Daniel Jordan
2019-07-18 14:56 ` Herbert Xu
2019-07-18 15:01 ` [v2 PATCH] " Herbert Xu
2019-07-19 14:37 ` Daniel Jordan
2019-07-19 14:55 ` Herbert Xu
2019-07-19 19:04 ` [PATCH] padata: purge get_cpu and reorder_via_wq from padata_do_serial Daniel Jordan
2019-07-26 12:36 ` Herbert Xu
2019-07-19 14:27 ` [PATCH] padata: Replace delayed timer with immediate workqueue in padata_reorder Daniel Jordan
2019-07-17 23:21 ` Daniel Jordan
2019-07-18 3:30 ` Herbert Xu
2019-07-18 14:25 ` Daniel Jordan
2019-07-18 14:49 ` Herbert Xu
2019-07-19 14:21 ` Daniel Jordan
2019-07-18 5:42 ` [PATCH v2] padata: use smp_mb in padata_reorder to avoid orphaned padata jobs Herbert Xu
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=20190716125309.GA10672@andrea \
--to=andrea.parri@amarulasolutions.com \
--cc=boqun.feng@gmail.com \
--cc=daniel.m.jordan@oracle.com \
--cc=herbert@gondor.apana.org.au \
--cc=linux-arch@vger.kernel.org \
--cc=linux-crypto@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=paulmck@linux.ibm.com \
--cc=peterz@infradead.org \
--cc=steffen.klassert@secunet.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).