All of lore.kernel.org
 help / color / mirror / Atom feed
From: Eric Dumazet <eric.dumazet@gmail.com>
To: Roman Gushchin <klamm@yandex-team.ru>
Cc: paulmck@linux.vnet.ibm.com, Dipankar Sarma <dipankar@in.ibm.com>,
	zhmurov@yandex-team.ru, linux-kernel@vger.kernel.org,
	netdev@vger.kernel.org, "David S. Miller" <davem@davemloft.net>,
	Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>,
	James Morris <jmorris@namei.org>,
	Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>,
	Patrick McHardy <kaber@trash.net>
Subject: Re: [PATCH v2] rcu: fix a race in hlist_nulls_for_each_entry_rcu macro
Date: Mon, 27 May 2013 17:12:10 -0700	[thread overview]
Message-ID: <1369699930.3301.494.camel@edumazet-glaptop> (raw)
In-Reply-To: <51A39E11.5020405@yandex-team.ru>

On Mon, 2013-05-27 at 21:55 +0400, Roman Gushchin wrote:
> Hi, Paul!
> 
> > On 25.05.2013 15:37, Paul E. McKenney wrote:
> >> Again, I believe that your retry logic needs to extend back into the
> >> calling function for your some_func() example above.
> 
> And what do you think about the following approach (diff below)?
> 
> It seems to me, it's enough clear (especially with good accompanying comments)
> and produces a good binary code (without significant overhead).
> Also, we will remove a hidden reef in using rcu-protected (h)list traverses with restarts.
> 

> diff --git a/include/linux/rculist_nulls.h b/include/linux/rculist_nulls.h
> index 2ae1371..4af5ee5 100644
> --- a/include/linux/rculist_nulls.h
> +++ b/include/linux/rculist_nulls.h
> @@ -107,7 +107,8 @@ static inline void hlist_nulls_add_head_rcu(struct hlist_nulls_node *n,
>    *
>    */
>   #define hlist_nulls_for_each_entry_rcu(tpos, pos, head, member)                        \
> -       for (pos = rcu_dereference_raw(hlist_nulls_first_rcu(head));            \
> +       for (ACCESS_ONCE(*(head)),                                              \
> +               pos = rcu_dereference_raw(hlist_nulls_first_rcu(head));         \
>                  (!is_a_nulls(pos)) &&                                           \
>                  ({ tpos = hlist_nulls_entry(pos, typeof(*tpos), member); 1; }); \
>                  pos = rcu_dereference_raw(hlist_nulls_next_rcu(pos)))

It looks like this still relies on gcc being friendly here.

I repeat again : @head here is a constant.

Macro already uses ACCESS_ONCE(), we only have to instruct gcc that
caching the value is forbidden if we restart the loop 
(aka "goto begin;" see Documentation/RCU/rculist_nulls.txt line 146)

Adding a barrier() is probably what we want.

I cooked followed patch and it fixes the problem.

diff --git a/include/linux/rculist_nulls.h b/include/linux/rculist_nulls.h
index 2ae1371..4dc51b2 100644
--- a/include/linux/rculist_nulls.h
+++ b/include/linux/rculist_nulls.h
@@ -105,8 +105,12 @@ static inline void hlist_nulls_add_head_rcu(struct hlist_nulls_node *n,
  * @head:	the head for your list.
  * @member:	the name of the hlist_nulls_node within the struct.
  *
+ * The barrier() is needed to make sure compiler doesn't cache first element,
+ * as this loop can be restarted.
+ * (cf Documentation/RCU/rculist_nulls.txt around line 146)
  */
 #define hlist_nulls_for_each_entry_rcu(tpos, pos, head, member)			\
+	barrier();								\
 	for (pos = rcu_dereference_raw(hlist_nulls_first_rcu(head));		\
 		(!is_a_nulls(pos)) &&						\
 		({ tpos = hlist_nulls_entry(pos, typeof(*tpos), member); 1; }); \



  reply	other threads:[~2013-05-28  0:12 UTC|newest]

Thread overview: 57+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-05-21  9:05 [PATCH] rcu: fix a race in hlist_nulls_for_each_entry_rcu macro Roman Gushchin
2013-05-21 10:40 ` David Laight
2013-05-21 11:55   ` Roman Gushchin
2013-05-21 13:42     ` David Laight
2013-05-21 12:09 ` Paul E. McKenney
2013-05-21 12:46   ` Roman Gushchin
2013-05-21 12:58     ` Paul E. McKenney
2013-05-21 13:37   ` Eric Dumazet
2013-05-21 13:44   ` Eric Dumazet
2013-05-21 14:47     ` Roman Gushchin
2013-05-21 15:16       ` Eric Dumazet
2013-05-21 15:51         ` Roman Gushchin
2013-05-21 15:38       ` Eric Dumazet
2013-05-21 15:51         ` Roman Gushchin
2013-05-21 18:12         ` [PATCH v2] " Roman Gushchin
2013-05-22  2:01           ` Eric Dumazet
2013-05-22  5:49             ` Eric Dumazet
2013-05-22 11:58               ` Roman Gushchin
2013-05-22 12:30                 ` Eric Dumazet
2013-05-22 13:07                   ` Roman Gushchin
2013-05-22 17:45                     ` Paul E. McKenney
2013-05-22 19:17                       ` Roman Gushchin
2013-05-25 11:37                         ` Paul E. McKenney
2013-05-27 11:34                           ` Roman Gushchin
2013-05-27 17:55                           ` Roman Gushchin
2013-05-28  0:12                             ` Eric Dumazet [this message]
2013-05-28  9:10                               ` Roman Gushchin
2013-05-29  0:34                                 ` Eric Dumazet
2013-05-29  1:31                                   ` Paul E. McKenney
2013-05-29  5:08                                     ` Eric Dumazet
2013-05-29 10:09                                       ` Roman Gushchin
2013-05-29 19:06                                         ` Eric Dumazet
2013-05-30  8:25                                           ` Roman Gushchin
2013-06-02 23:31                                             ` Eric Dumazet
2013-06-03  2:58                                               ` David Miller
2013-06-03  3:12                                                 ` Eric Dumazet
2013-06-03  3:27                                                   ` David Miller
2013-06-03  3:42                                                     ` Paul E. McKenney
2013-06-03  3:47                                                       ` Eric Dumazet
2013-06-03  3:49                                                       ` David Miller
2013-06-03  6:05                                                         ` Paul E. McKenney
2013-06-10 18:29                                                         ` Boris B. Zhmurov
2013-06-10 18:51                                                           ` Eric Dumazet
2013-06-03  3:48                                                   ` Paul E. McKenney
2013-06-03  3:42                                                 ` Paul E. McKenney
2013-05-29  9:17                                   ` Roman Gushchin
2013-05-29  1:19                                 ` Paul E. McKenney
2013-05-22 13:27                   ` David Laight
2013-05-22 13:27                     ` David Laight
2013-05-22 13:36                     ` Eric Dumazet
2013-05-22 14:23                       ` David Laight
2013-05-22 14:23                         ` David Laight
2013-05-22 13:55                     ` Roman Gushchin
2013-05-22  9:58             ` Paul E. McKenney
2013-05-22 12:28               ` Eric Dumazet
2013-05-22 13:00                 ` Paul E. McKenney
2013-05-22 14:16                   ` Eric Dumazet

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=1369699930.3301.494.camel@edumazet-glaptop \
    --to=eric.dumazet@gmail.com \
    --cc=davem@davemloft.net \
    --cc=dipankar@in.ibm.com \
    --cc=jmorris@namei.org \
    --cc=kaber@trash.net \
    --cc=klamm@yandex-team.ru \
    --cc=kuznet@ms2.inr.ac.ru \
    --cc=linux-kernel@vger.kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=paulmck@linux.vnet.ibm.com \
    --cc=yoshfuji@linux-ipv6.org \
    --cc=zhmurov@yandex-team.ru \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.