From: ebiederm@xmission.com (Eric W. Biederman)
To: Christoph Lameter <cl@linux.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>,
Sasha Levin <levinsasha928@gmail.com>,
Dave Jones <davej@redhat.com>, davem <davem@davemloft.net>,
Pekka Enberg <penberg@kernel.org>, Matt Mackall <mpm@selenic.com>,
kaber@trash.net, pablo@netfilter.org,
linux-kernel <linux-kernel@vger.kernel.org>,
linux-mm <linux-mm@kvack.org>,
netfilter-devel@vger.kernel.org, netdev <netdev@vger.kernel.org>
Subject: Re: Hung task when calling clone() due to netfilter/slab
Date: Thu, 19 Jan 2012 14:15:01 -0800 [thread overview]
Message-ID: <m14nvr2vbu.fsf@fess.ebiederm.org> (raw)
In-Reply-To: <m1bopz2ws3.fsf@fess.ebiederm.org> (Eric W. Biederman's message of "Thu, 19 Jan 2012 13:43:40 -0800")
ebiederm@xmission.com (Eric W. Biederman) writes:
> Christoph Lameter <cl@linux.com> writes:
>
>> Another version that drops the slub lock for both invocations of sysfs
>> functions from kmem_cache_create. The invocation from slab_sysfs_init
>> is not a problem since user space is not active at that point.
>>
>>
>> Subject: slub: Do not take the slub lock while calling into sysfs
>>
>> This patch avoids holding the slub_lock during kmem_cache_create()
>> when calling sysfs. It is possible because kmem_cache_create()
>> allocates the kmem_cache object and therefore is the only one context
>> that can access the newly created object. It is therefore possible
>> to drop the slub_lock early. We defer the adding of the new kmem_cache
>> to the end of processing because the new kmem_cache structure would
>> be reachable otherwise via scans over slabs. This allows sysfs_slab_add()
>> to run without holding any locks.
>>
>> The case is different if we are creating an alias instead of a new
>> kmem_cache structure. In that case we can also drop the slub lock
>> early because we have taken a refcount on the kmem_cache structure.
>> It therefore cannot vanish from under us.
>> But if the sysfs_slab_alias() call fails we can no longer simply
>> decrement the refcount since the other references may have gone
>> away in the meantime. Call kmem_cache_destroy() to cause the
>> refcount to be decremented and the kmem_cache structure to be
>> freed if all references are gone.
>>
>> Signed-off-by: Christoph Lameter <cl@linux.com>
>
> I am dense. Is the deadlock here that you are fixing slub calling sysfs
> with the slub_lock held but sysfs then calling kmem_cache_zalloc?
>
> I don't see what sysfs is doing in the creation path that would cause
> a deadlock except for using slab.
Oh. I see. The problem is calling kobject_uevent (which happens to
live in slabs sysfs_slab_add) with a lock held. And kobject_uevent
makes a blocking call to userspace.
No locks held seems to be a good policy on that one.
Eric
next prev parent reply other threads:[~2012-01-19 22:12 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-01-14 16:30 Hung task when calling clone() due to netfilter/slab Sasha Levin
2012-01-14 17:10 ` Eric Dumazet
2012-01-14 17:18 ` Eric Dumazet
2012-01-15 12:59 ` Sasha Levin
2012-01-15 17:25 ` Eric Dumazet
2012-01-17 15:12 ` Christoph Lameter
2012-01-17 15:20 ` Eric Dumazet
2012-01-17 15:27 ` Christoph Lameter
2012-01-17 15:30 ` Eric Dumazet
2012-01-17 16:04 ` Christoph Lameter
2012-01-17 22:22 ` Christoph Lameter
2012-01-19 21:43 ` Eric W. Biederman
2012-01-19 22:15 ` Eric W. Biederman [this message]
2012-01-20 2:03 ` Christoph Lameter
2012-01-20 2:31 ` Eric W. Biederman
2012-01-20 14:49 ` Christoph Lameter
2012-01-20 20:40 ` Eric W. Biederman
2012-02-01 8:05 ` Pekka Enberg
2012-02-01 17:32 ` Eric W. Biederman
2012-02-01 8:07 ` Pekka Enberg
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=m14nvr2vbu.fsf@fess.ebiederm.org \
--to=ebiederm@xmission.com \
--cc=cl@linux.com \
--cc=davej@redhat.com \
--cc=davem@davemloft.net \
--cc=eric.dumazet@gmail.com \
--cc=kaber@trash.net \
--cc=levinsasha928@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mpm@selenic.com \
--cc=netdev@vger.kernel.org \
--cc=netfilter-devel@vger.kernel.org \
--cc=pablo@netfilter.org \
--cc=penberg@kernel.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).