From: Al Viro <viro@ZenIV.linux.org.uk>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mika Westerberg <mika.westerberg@linux.intel.com>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Miklos Szeredi <mszeredi@suse.cz>,
linux-fsdevel <linux-fsdevel@vger.kernel.org>
Subject: Re: fs/dcache.c - BUG: soft lockup - CPU#5 stuck for 22s! [systemd-udevd:1667]
Date: Thu, 29 May 2014 17:23:07 +0100 [thread overview]
Message-ID: <20140529162307.GL18016@ZenIV.linux.org.uk> (raw)
In-Reply-To: <20140529154454.GK18016@ZenIV.linux.org.uk>
On Thu, May 29, 2014 at 04:44:54PM +0100, Al Viro wrote:
> On Thu, May 29, 2014 at 08:10:57AM -0700, Linus Torvalds wrote:
> > If so, though, that brings up two questions:
> >
> > (a) do we really want to be that aggressive? Can we ever traverse
> > _past_ the point we're actually trying to shrink in
> > shrink_dcache_parent()?
>
> Caller of shrink_dcache_parent() would better hold a reference to the
> argument, or it might get freed right under us ;-) So no, we can't
> go past that point - the subtree root will stay busy.
>
> The reason we want to be aggressive there is to avoid excessive iterations -
> think what happens e.g. if we have a chain of N dentries, with nothing pinning
> them (i.e. the last one has refcount 0, the first - 2, everything else - 1).
> Simply doing dput() would result in O(N^2) vs. O(N)...
>
> > (b) why does the "dput()" (or rather, the dentry_kill()) locking
> > logic have to retain the old trylock case rather than share the parent
> > locking logic?
> >
> > I'm assuming the answer to (b) is that we can't afford to drop the
> > dentry lock in dentry_kill(), but I'd like that answer to the "Why" to
> > be documented somewhere.
>
> We actually might be able to do it that way (rechecking ->d_count after
> lock_parent()), but I would really prefer to leave that until after -final.
> I want to get profiling data from that first - dput() is a much hotter path
> than shrink_dcache_parent() and friends...
FWIW, I've just done more or less edible splitup of stuff past #for-linus -
see #experimental-dentry_kill for that. Again, I really want to get
profiling data to see if that hurts dput() - it takes ->d_lock on parent
before the trylock on ->i_lock and in case of ->d_lock on parent being
held by somebody else it bangs on rename_lock.lock cacheline. I'd expect
that to be non-issue on any loads, but we need something stronger than
my gut feelings...
BTW, lock_parent() might be better off if in contended case it would not
bother with rename_lock and did something like this:
again:
spin_unlock(&dentry->d_lock);
rcu_read_lock();
parent = ACCESS_ONCE(dentry->d_parent);
if (parent != dentry)
spin_lock(&parent->d_lock);
spin_lock(&dentry->d_lock);
if (likely(dentry->d_parent == parent)) {
rcu_read_unlock();
return parent;
}
if (parent)
spin_unlock(&parent->d_lock);
rcu_read_unlock();
goto again;
It's almost certainly not worth bothering with right now, but if dput()
starts using lock_parent(), it might be worth investigating...
next prev parent reply other threads:[~2014-05-29 16:23 UTC|newest]
Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20140526093741.GA1765@lahna.fi.intel.com>
2014-05-26 13:57 ` fs/dcache.c - BUG: soft lockup - CPU#5 stuck for 22s! [systemd-udevd:1667] Al Viro
2014-05-26 14:29 ` Mika Westerberg
2014-05-26 15:27 ` Al Viro
2014-05-26 16:42 ` Al Viro
2014-05-26 18:17 ` Linus Torvalds
2014-05-26 18:26 ` Al Viro
2014-05-26 20:24 ` Linus Torvalds
2014-05-27 1:40 ` Al Viro
2014-05-27 3:14 ` Al Viro
2014-05-27 4:00 ` Al Viro
2014-05-27 7:04 ` Mika Westerberg
2014-05-28 3:19 ` Al Viro
2014-05-28 7:37 ` Mika Westerberg
2014-05-28 11:57 ` Al Viro
2014-05-28 13:11 ` Mika Westerberg
2014-05-28 14:19 ` Al Viro
2014-05-28 18:39 ` Al Viro
2014-05-28 19:43 ` Linus Torvalds
2014-05-28 20:02 ` Linus Torvalds
2014-05-28 20:25 ` Al Viro
2014-05-29 10:42 ` Mika Westerberg
2014-05-28 20:14 ` Al Viro
2014-05-28 21:11 ` Linus Torvalds
2014-05-28 21:28 ` Al Viro
2014-05-29 3:11 ` Al Viro
2014-05-29 3:52 ` Al Viro
2014-05-29 5:34 ` Al Viro
2014-05-29 10:51 ` Mika Westerberg
2014-05-29 11:04 ` Mika Westerberg
2014-05-29 13:30 ` Al Viro
2014-05-29 14:56 ` Mika Westerberg
2014-05-29 15:10 ` Linus Torvalds
2014-05-29 15:44 ` Al Viro
2014-05-29 16:23 ` Al Viro [this message]
2014-05-29 16:29 ` Linus Torvalds
2014-05-29 16:53 ` Al Viro
2014-05-29 18:52 ` Al Viro
2014-05-29 19:14 ` Linus Torvalds
2014-05-30 4:50 ` Al Viro
2014-05-30 5:00 ` Linus Torvalds
2014-05-30 6:49 ` Al Viro
2014-05-30 8:12 ` Mika Westerberg
2014-05-30 15:21 ` Al Viro
2014-05-30 15:31 ` Linus Torvalds
2014-05-30 16:48 ` [git pull] " Al Viro
2014-05-30 17:14 ` Al Viro
2014-05-31 14:18 ` Josh Boyer
2014-05-31 14:48 ` Linus Torvalds
2014-05-31 14:58 ` Josh Boyer
2014-05-31 16:12 ` Josh Boyer
2014-05-30 17:15 ` Sedat Dilek
2014-05-29 4:21 ` Linus Torvalds
2014-05-29 5:16 ` Al Viro
2014-05-29 5:26 ` Al Viro
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=20140529162307.GL18016@ZenIV.linux.org.uk \
--to=viro@zeniv.linux.org.uk \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mika.westerberg@linux.intel.com \
--cc=mszeredi@suse.cz \
--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).