public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: Jens Axboe <jens.axboe@oracle.com>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>,
	paulmck@linux.vnet.ibm.com, Ingo Molnar <mingo@elte.hu>,
	Vegard Nossum <vegard.nossum@gmail.com>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	Randy Dunlap <randy.dunlap@oracle.com>
Subject: Re: kmemcheck caught read from freed memory (cfq_free_io_context)
Date: Wed, 02 Apr 2008 14:13:42 +0200	[thread overview]
Message-ID: <1207138422.8514.799.camel@twins> (raw)
In-Reply-To: <20080402115337.GD12774@kernel.dk>

On Wed, 2008-04-02 at 13:53 +0200, Jens Axboe wrote:

> > Yeah, SLAB_DESTROY_BY_RCU should have a _HUGE_ comment explaining it,
> > I'm sure this is not the first (nor the last) time people get that
> > wrong.
> 
> It should, SLAB_DESTROY_BY_RCU is definitely useful, but it is expected
> to be an 'easier' way of doing the call_rcu() manually. So it definitely
> needs more documentation.
> 

Ok I gave it a go, how bad is this text?

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
---
diff --git a/include/linux/slab.h b/include/linux/slab.h
index f950a89..e049ddc 100644
--- a/include/linux/slab.h
+++ b/include/linux/slab.h
@@ -25,6 +25,32 @@
 #define SLAB_CACHE_DMA		0x00004000UL	/* Use GFP_DMA memory */
 #define SLAB_STORE_USER		0x00010000UL	/* DEBUG: Store the last owner for bug hunting */
 #define SLAB_PANIC		0x00040000UL	/* Panic if kmem_cache_create() fails */
+/*
+ * SLAB_DESTROY_BY_RCU - **WARNING** READ THIS!
+ *
+ * This delays freeing the SLAB page by a grace period, it does _NOT_
+ * delay object freeing. This means that if you do kmem_cache_free()
+ * that memory location is free to be reused at any time. Thus it may
+ * be possible to see another object there in the same RCU grace period.
+ *
+ * This feature only ensures the memory location backing the object
+ * stays valid, the trick to using this is relying on an independent
+ * object validation pass. Something like:
+ *
+ *  rcu_read_lock()
+ * again:
+ *  obj = lockless_lookup(key);
+ *  if (obj) {
+ *    if (!try_get_ref(obj)) // might fail for free objects
+ *      goto again;
+ *
+ *    if (obj->key != key) { // not the object we expected
+ *      put_ref(obj);
+ *      goto again;
+ *    }
+ *  }
+ *  rcu_read_unlock();
+ */
 #define SLAB_DESTROY_BY_RCU	0x00080000UL	/* Defer freeing slabs to RCU */
 #define SLAB_MEM_SPREAD		0x00100000UL	/* Spread some memory over cpuset */
 #define SLAB_TRACE		0x00200000UL	/* Trace allocations and frees */



  reply	other threads:[~2008-04-02 12:14 UTC|newest]

Thread overview: 69+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-04-01 21:08 kmemcheck caught read from freed memory (cfq_free_io_context) Vegard Nossum
2008-04-01 21:36 ` Peter Zijlstra
2008-04-01 22:51   ` Paul E. McKenney
2008-04-02  6:15     ` Peter Zijlstra
2008-04-02  7:19       ` Jens Axboe
2008-04-02 10:24       ` Paul E. McKenney
2008-04-02  7:17   ` Jens Axboe
2008-04-02  7:20     ` Pekka J Enberg
2008-04-02  7:24       ` Jens Axboe
2008-04-02  7:28         ` Ingo Molnar
2008-04-02  7:31           ` Jens Axboe
2008-04-02 10:55           ` Paul E. McKenney
2008-04-02 10:59             ` Peter Zijlstra
2008-04-02 11:33               ` Fabio Checconi
2008-04-02 11:43                 ` Jens Axboe
2008-04-02 12:36                   ` Jens Axboe
2008-04-02 12:36                     ` Jens Axboe
2008-04-02 12:55                       ` Fabio Checconi
2008-04-02 12:58                         ` Jens Axboe
2008-04-02 12:58                           ` Jens Axboe
2008-04-02 13:16                             ` Fabio Checconi
2008-04-02 16:14                               ` Paul E. McKenney
2008-04-02 13:37                           ` Paul E. McKenney
2008-04-02 13:41                             ` Jens Axboe
2008-04-02 15:33                               ` Paul E. McKenney
2008-04-02 16:31                                 ` Jens Axboe
2008-04-02 17:00                                   ` Paul E. McKenney
2008-04-02 13:32                 ` Paul E. McKenney
2008-04-02 13:40                   ` Jens Axboe
2008-04-02 16:15                     ` Paul E. McKenney
2008-04-02 11:01             ` Pekka Enberg
2008-04-02 11:07               ` Jens Axboe
2008-04-02 11:08                 ` Peter Zijlstra
2008-04-02 11:11                   ` Pekka Enberg
2008-04-02 11:14                     ` Peter Zijlstra
2008-04-02 11:18                       ` Pekka Enberg
2008-04-02 17:36                     ` Christoph Lameter
2008-04-02 11:14                   ` Jens Axboe
2008-04-02 11:20                     ` Peter Zijlstra
2008-04-02 11:25                       ` Peter Zijlstra
2008-04-02 11:32                       ` Jens Axboe
2008-04-02 11:37                         ` Peter Zijlstra
2008-04-02 11:42                           ` Jens Axboe
2008-04-02 11:47                             ` Peter Zijlstra
2008-04-02 11:53                               ` Jens Axboe
2008-04-02 12:13                                 ` Peter Zijlstra [this message]
2008-04-02 12:28                                   ` Jens Axboe
2008-04-02 13:26                                   ` Paul E. McKenney
2008-04-02 13:43                                   ` Andi Kleen
2008-04-02 12:26                                 ` Peter Zijlstra
2008-04-02 12:34                                   ` Jens Axboe
2008-04-02 16:08               ` Paul E. McKenney
2008-04-02 16:15                 ` Vegard Nossum
2008-04-02 16:32                   ` Pekka J Enberg
2008-04-02 18:23                     ` Paul E. McKenney
2008-04-02 19:53                       ` Pekka Enberg
2008-04-02 20:15                         ` Paul E. McKenney
2008-04-03 15:18                           ` Paul E. McKenney
2008-04-03 19:49                             ` Pekka J Enberg
2008-04-03 21:27                               ` Paul E. McKenney
2008-04-02 16:59                   ` Paul E. McKenney
2008-04-02 17:31                     ` Vegard Nossum
2008-04-02 10:40     ` Paul E. McKenney
2008-04-02 10:46       ` Pekka Enberg
2008-04-02 10:49         ` Peter Zijlstra
2008-04-02 10:54           ` Pekka J Enberg
2008-04-02 17:35           ` Christoph Lameter
2008-04-02 10:53       ` Peter Zijlstra
2008-04-02 11:13         ` Jens Axboe

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=1207138422.8514.799.camel@twins \
    --to=a.p.zijlstra@chello.nl \
    --cc=akpm@linux-foundation.org \
    --cc=jens.axboe@oracle.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=paulmck@linux.vnet.ibm.com \
    --cc=penberg@cs.helsinki.fi \
    --cc=randy.dunlap@oracle.com \
    --cc=vegard.nossum@gmail.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