From: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
To: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
"Rafael J. Wysocki" <rjw@sisk.pl>,
LKML <linux-kernel@vger.kernel.org>, Ingo Molnar <mingo@elte.hu>,
Andrew Morton <akpm@linux-foundation.org>,
linux-ext4@vger.kernel.org
Subject: Re: 2.6.25-git2: BUG: unable to handle kernel paging request at ffffffffffffffff
Date: Sun, 20 Apr 2008 19:08:07 -0700 [thread overview]
Message-ID: <20080421020806.GL20138@linux.vnet.ibm.com> (raw)
In-Reply-To: <20080421011855.GA6243@gondor.apana.org.au>
On Mon, Apr 21, 2008 at 09:18:55AM +0800, Herbert Xu wrote:
> Hi Linus:
>
> On Sun, Apr 20, 2008 at 02:31:48PM -0700, Linus Torvalds wrote:
> >
> > Talking about RCU I also think that whoever did those "rcu_dereference()"
> > macros in <linux/list.h> was insane. It's totally pointless to do
> > "rcu_dereference()" on a local variable. It simply *cannot* make sense.
> > Herbert, Paul, you guys should look at it.
>
> Since I made the macros look this way I'm obliged to defend it :)
>
> > #define list_for_each_rcu(pos, head) \
> > - for (pos = (head)->next; \
> > - prefetch(rcu_dereference(pos)->next), pos != (head); \
> > - pos = pos->next)
> > + for (pos = rcu_dereference((head)->next); \
> > + prefetch(pos->next), pos != (head); \
> > + pos = rcu_dereference(pos->next))
>
> Semantically there should be no difference between the two versions.
> The purpose of rcu_dereference is really similar to smp_rmb, i.e.,
> it adds a (conditional) read barrier between what has been read so
> far (including its argument), and what will be read subsequently.
>
> So if we expand out the current code it would look like
>
> fetch (head)->next
> store into pos
> again:
> smp_read_barrier_depends()
> prefetch(pos->next)
> pos != (head)
>
> ...loop body...
>
> fetch pos->next
> store into pos
> goto again
>
> Yours looks like
>
> fetch (head)->next
> smp_read_barrier_depends()
> store into pos
> again:
> prefetch(pos->next)
> pos != (head)
>
> ...loop body...
>
> fetch pos->next
> smp_read_barrier_depends()
> store into pos
> goto again
>
> As the objective here is to insert a barrier before dereferencing
> pos (e.g., reading pos->next or using it in the loop body), these
> two should be identical.
>
> But I do concede that your version looks clearer, and has the
> benefit that should prefetch ever be optimised out with no side-
> effects, yours would still be correct while the current one will
> lose the barrier completely.
Agreed as well -- compilers would also be within their right to bypass
the rcu_dereference() around the test/prefetch, which would allow
them to refetch. For example, with __list_for_each_rcu(), the original
implementation allows the compiler to treat a use of "pos" within the body
of the loop as if it was a use of (head)->next, refetching if convenient.
Not so good.
So good catch, Linus!!!
Could we also eliminate the (both unused in 2.6.25 and useless as
well) list_for_each_safe_rcu()? After all, if you use list_del_rcu()
and call_rcu(), all the RCU list-traversal primitives are "safe" in
this sense. Patch attached (testing in progress), based on Linus's
earlier patch.
Signed_off_by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
---
list.h | 47 +++++++++++++++--------------------------------
1 file changed, 15 insertions(+), 32 deletions(-)
diff -urpNa linux-2.6.25/include/linux/list.h linux-2.6.25-rcu-list/include/linux/list.h
--- linux-2.6.25/include/linux/list.h 2008-04-16 19:49:44.000000000 -0700
+++ linux-2.6.25-rcu-list/include/linux/list.h 2008-04-20 18:44:55.000000000 -0700
@@ -631,31 +631,14 @@ static inline void list_splice_init_rcu(
* as long as the traversal is guarded by rcu_read_lock().
*/
#define list_for_each_rcu(pos, head) \
- for (pos = (head)->next; \
- prefetch(rcu_dereference(pos)->next), pos != (head); \
- pos = pos->next)
+ for (pos = rcu_dereference((head)->next); \
+ prefetch(pos->next), pos != (head); \
+ pos = rcu_dereference(pos->next))
#define __list_for_each_rcu(pos, head) \
- for (pos = (head)->next; \
- rcu_dereference(pos) != (head); \
- pos = pos->next)
-
-/**
- * list_for_each_safe_rcu
- * @pos: the &struct list_head to use as a loop cursor.
- * @n: another &struct list_head to use as temporary storage
- * @head: the head for your list.
- *
- * Iterate over an rcu-protected list, safe against removal of list entry.
- *
- * This list-traversal primitive may safely run concurrently with
- * the _rcu list-mutation primitives such as list_add_rcu()
- * as long as the traversal is guarded by rcu_read_lock().
- */
-#define list_for_each_safe_rcu(pos, n, head) \
- for (pos = (head)->next; \
- n = rcu_dereference(pos)->next, pos != (head); \
- pos = n)
+ for (pos = rcu_dereference((head)->next); \
+ pos != (head); \
+ pos = rcu_dereference(pos->next))
/**
* list_for_each_entry_rcu - iterate over rcu list of given type
@@ -668,10 +651,10 @@ static inline void list_splice_init_rcu(
* as long as the traversal is guarded by rcu_read_lock().
*/
#define list_for_each_entry_rcu(pos, head, member) \
- for (pos = list_entry((head)->next, typeof(*pos), member); \
- prefetch(rcu_dereference(pos)->member.next), \
+ for (pos = list_entry(rcu_dereference((head)->next), typeof(*pos), member); \
+ prefetch(pos->member.next), \
&pos->member != (head); \
- pos = list_entry(pos->member.next, typeof(*pos), member))
+ pos = list_entry(rcu_dereference(pos->member.next), typeof(*pos), member))
/**
@@ -686,9 +669,9 @@ static inline void list_splice_init_rcu(
* as long as the traversal is guarded by rcu_read_lock().
*/
#define list_for_each_continue_rcu(pos, head) \
- for ((pos) = (pos)->next; \
- prefetch(rcu_dereference((pos))->next), (pos) != (head); \
- (pos) = (pos)->next)
+ for ((pos) = rcu_dereference((pos)->next); \
+ prefetch((pos)->next), (pos) != (head); \
+ (pos) = rcu_dereference((pos)->next))
/*
* Double linked lists with a single pointer list head.
@@ -986,10 +969,10 @@ static inline void hlist_add_after_rcu(s
* as long as the traversal is guarded by rcu_read_lock().
*/
#define hlist_for_each_entry_rcu(tpos, pos, head, member) \
- for (pos = (head)->first; \
- rcu_dereference(pos) && ({ prefetch(pos->next); 1;}) && \
+ for (pos = rcu_dereference((head)->first); \
+ ({ prefetch(pos->next); 1;}) && \
({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \
- pos = pos->next)
+ pos = rcu_dereference(pos->next))
#else
#warning "don't include kernel headers in userspace"
next prev parent reply other threads:[~2008-04-21 2:08 UTC|newest]
Thread overview: 173+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <200804191522.54334.rjw@sisk.pl>
2008-04-20 19:04 ` 2.6.25-git2: BUG: unable to handle kernel paging request at ffffffffffffffff Rafael J. Wysocki
2008-04-20 19:14 ` Rafael J. Wysocki
2008-04-20 21:31 ` Linus Torvalds
2008-04-21 1:18 ` Herbert Xu
2008-04-21 2:08 ` Paul E. McKenney [this message]
2008-04-21 4:59 ` Paul E. McKenney
2008-04-21 5:47 ` Paul E. McKenney
2008-04-21 13:00 ` Ingo Molnar
2008-04-21 16:06 ` Linus Torvalds
2008-04-21 16:24 ` Rafael J. Wysocki
2008-04-21 15:49 ` Linus Torvalds
2008-04-21 17:05 ` Paul E. McKenney
2008-04-21 17:30 ` Linus Torvalds
2008-04-21 17:43 ` Paul E. McKenney
2008-04-22 1:03 ` Herbert Xu
2008-04-22 13:36 ` Paul E. McKenney
2008-04-21 16:12 ` Rafael J. Wysocki
2008-04-21 16:54 ` Linus Torvalds
2008-04-21 17:06 ` Jiri Slaby
2008-04-21 17:19 ` Rafael J. Wysocki
2008-04-21 17:48 ` Linus Torvalds
2008-04-21 18:22 ` Rafael J. Wysocki
2008-04-21 19:38 ` Jiri Slaby
2008-04-21 20:39 ` David Miller
2008-04-21 21:18 ` Jiri Slaby
2008-04-21 21:58 ` Jiri Slaby
2008-04-21 22:26 ` Jiri Slaby
2008-04-21 22:54 ` Paul E. McKenney
2008-04-21 23:02 ` Jiri Slaby
2008-04-21 23:11 ` Zdenek Kabelac
2008-04-21 23:17 ` Jiri Slaby
2008-04-22 0:54 ` Rafael J. Wysocki
2008-04-22 1:14 ` Linus Torvalds
2008-04-22 1:30 ` Rafael J. Wysocki
2008-04-22 9:49 ` Jiri Slaby
2008-04-22 9:53 ` Ingo Molnar
2008-04-22 18:35 ` Zdenek Kabelac
2008-04-22 18:48 ` Linus Torvalds
2008-04-23 8:50 ` Zdenek Kabelac
2008-04-23 15:53 ` Linus Torvalds
2008-04-23 16:58 ` Pekka Enberg
2008-04-23 17:28 ` Zdenek Kabelac
2008-04-23 17:40 ` Ingo Molnar
2008-04-23 18:52 ` Pekka Enberg
2008-04-23 19:05 ` Christoph Lameter
2008-04-23 19:19 ` Pekka J Enberg
2008-04-23 19:28 ` Christoph Lameter
2008-04-23 20:27 ` Zdenek Kabelac
2008-04-24 22:26 ` Jiri Slaby
2008-04-24 22:41 ` Linus Torvalds
2008-04-25 0:57 ` Jiri Slaby
2008-04-24 23:45 ` Linus Torvalds
2008-04-25 7:36 ` Jiri Slaby
2008-04-25 14:09 ` Pavel Machek
2008-04-25 15:30 ` Rafael J. Wysocki
2008-04-25 17:10 ` Jiri Slaby
2008-04-25 9:13 ` David Miller
2008-04-25 12:15 ` Zdenek Kabelac
2008-04-25 12:27 ` Zdenek Kabelac
2008-04-25 15:12 ` [PATCH 1/1] x86: fix text_poke Jiri Slaby
2008-04-25 15:03 ` Linus Torvalds
2008-04-25 15:17 ` Andi Kleen
2008-04-25 19:36 ` Christoph Lameter
2008-04-26 9:59 ` Andi Kleen
2008-04-26 11:16 ` Jiri Slaby
2008-04-26 11:34 ` Andi Kleen
2008-04-28 20:24 ` VIRTUAL_BUG_ON() Christoph Lameter
2008-05-01 19:22 ` [RFC 1/1] mm: add virt to phys debug Jiri Slaby
2008-05-01 20:18 ` Christoph Lameter
2008-05-06 21:54 ` Jiri Slaby
2008-05-07 17:30 ` Christoph Lameter
2008-05-13 14:38 ` Jiri Slaby
2008-04-25 15:19 ` [PATCH 1/1] x86: fix text_poke Ingo Molnar
2008-04-25 15:26 ` Ingo Molnar
2008-04-25 15:32 ` Ingo Molnar
2008-04-25 15:33 ` Linus Torvalds
2008-04-25 15:48 ` Andi Kleen
2008-04-25 16:06 ` Linus Torvalds
2008-04-25 16:19 ` Andi Kleen
2008-04-25 16:24 ` Linus Torvalds
2008-04-25 16:33 ` Ingo Molnar
2008-04-25 18:13 ` Jeremy Fitzhardinge
2008-05-05 2:36 ` Nick Piggin
2008-04-25 16:30 ` Mathieu Desnoyers
2008-04-25 16:42 ` H. Peter Anvin
2008-04-25 17:09 ` Mathieu Desnoyers
2008-04-25 18:37 ` Mathieu Desnoyers
2008-04-25 18:47 ` H. Peter Anvin
2008-04-25 19:19 ` H. Peter Anvin
2008-04-25 20:04 ` Mathieu Desnoyers
2008-04-25 20:09 ` H. Peter Anvin
2008-04-25 20:18 ` H. Peter Anvin
2008-04-25 20:37 ` Mathieu Desnoyers
2008-04-25 20:41 ` H. Peter Anvin
2008-04-25 20:51 ` Linus Torvalds
2008-04-25 21:12 ` Mathieu Desnoyers
2008-04-25 21:15 ` H. Peter Anvin
2008-04-25 21:47 ` Mathieu Desnoyers
2008-04-25 22:07 ` H. Peter Anvin
2008-04-25 22:30 ` Mathieu Desnoyers
2008-04-25 22:36 ` Linus Torvalds
2008-04-28 20:21 ` Ingo Molnar
2008-04-28 20:55 ` Jeremy Fitzhardinge
2008-04-28 21:01 ` H. Peter Anvin
2008-04-28 22:42 ` Mathieu Desnoyers
2008-04-28 20:43 ` Mathieu Desnoyers
2008-04-28 21:02 ` Jeremy Fitzhardinge
2008-05-04 15:03 ` Mathieu Desnoyers
2008-05-04 16:18 ` H. Peter Anvin
2008-04-25 22:38 ` H. Peter Anvin
2008-04-25 22:04 ` Linus Torvalds
2008-04-25 23:00 ` Mathieu Desnoyers
2008-04-25 23:13 ` Jeremy Fitzhardinge
2008-04-25 23:34 ` Masami Hiramatsu
2008-04-26 6:21 ` Jeremy Fitzhardinge
2008-04-26 11:56 ` Arnaldo Carvalho de Melo
2008-04-26 23:38 ` Jeremy Fitzhardinge
2008-04-27 1:00 ` Arnaldo Carvalho de Melo
2008-04-26 2:12 ` Frank Ch. Eigler
2008-06-05 17:44 ` Frank Ch. Eigler
2008-04-26 6:50 ` Jeremy Fitzhardinge
2008-04-28 0:49 ` Masami Hiramatsu
2008-04-25 21:02 ` David Miller
2008-04-25 21:11 ` H. Peter Anvin
2008-04-25 16:22 ` Ingo Molnar
2008-04-25 16:37 ` Linus Torvalds
2008-04-25 16:43 ` Ingo Molnar
2008-04-25 16:45 ` Ingo Molnar
2008-04-25 16:51 ` Linus Torvalds
2008-04-25 17:02 ` Ingo Molnar
2008-04-25 17:13 ` Linus Torvalds
2008-04-25 17:26 ` Andi Kleen
2008-04-25 17:29 ` Linus Torvalds
2008-04-25 17:53 ` Ingo Molnar
2008-04-25 18:04 ` Ingo Molnar
2008-04-25 18:09 ` Linus Torvalds
2008-04-25 18:19 ` Ingo Molnar
2008-04-25 18:56 ` Ingo Molnar
2008-04-25 18:13 ` Ingo Molnar
2008-04-25 16:52 ` Ingo Molnar
2008-04-25 16:56 ` Andi Kleen
2008-04-25 15:50 ` Ingo Molnar
2008-04-25 15:57 ` H. Peter Anvin
2008-04-25 18:53 ` Pavel Machek
2008-04-25 16:11 ` Linus Torvalds
2008-04-25 15:54 ` Mathieu Desnoyers
2008-04-25 15:59 ` Ingo Molnar
2008-04-25 16:11 ` Mathieu Desnoyers
2008-04-25 15:27 ` Andi Kleen
2008-04-25 20:18 ` David Miller
2008-04-25 15:23 ` Jiri Slaby
2008-04-25 1:35 ` 2.6.25-git2: BUG: unable to handle kernel paging request at ffffffffffffffff David Miller
2008-04-25 1:48 ` Linus Torvalds
2008-04-25 1:57 ` David Miller
2008-04-25 7:41 ` Jiri Slaby
2008-04-25 7:45 ` David Miller
2008-04-25 8:02 ` Jiri Slaby
2008-04-25 10:53 ` Craig Schlenter
2008-04-25 7:42 ` Jiri Slaby
2008-04-25 7:49 ` David Miller
2008-04-25 7:56 ` Jiri Slaby
2008-04-25 7:58 ` David Miller
2008-04-25 8:00 ` Jiri Slaby
2008-04-25 15:47 ` Randy Dunlap
2008-04-22 21:46 ` Rafael J. Wysocki
2008-04-22 19:09 ` Ingo Molnar
2008-04-22 1:15 ` Rafael J. Wysocki
2008-04-22 1:25 ` [ProbableSpam]Re: " Paul E. McKenney
2008-04-21 21:19 ` Linus Torvalds
2008-04-21 21:54 ` David Miller
2008-04-21 13:17 ` Ingo Molnar
2008-04-21 13:35 ` Rafael J. Wysocki
2008-04-21 18:56 ` Ingo Molnar
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=20080421020806.GL20138@linux.vnet.ibm.com \
--to=paulmck@linux.vnet.ibm.com \
--cc=akpm@linux-foundation.org \
--cc=herbert@gondor.apana.org.au \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=rjw@sisk.pl \
--cc=torvalds@linux-foundation.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).