From: Andi Kleen <ak@linux.intel.com>
To: Dave Hansen <dave.hansen@intel.com>
Cc: paulmck@linux.vnet.ibm.com, LKML <linux-kernel@vger.kernel.org>,
Josh Triplett <josh@joshtriplett.org>,
"Chen, Tim C" <tim.c.chen@intel.com>,
Christoph Lameter <cl@linux.com>
Subject: Re: [bisected] pre-3.16 regression on open() scalability
Date: Tue, 17 Jun 2014 17:15:17 -0700 [thread overview]
Message-ID: <20140618001517.GL8178@tassilo.jf.intel.com> (raw)
In-Reply-To: <53A0CAE5.9000702@intel.com>
> It also ends up eating a new cacheline in a bunch of pretty hot paths.
> It would be nice to be able to keep the fast path part of this as at
> least read-only.
>
> Could we do something (functionally) like the attached patch? Instead
> of counting cond_resched() calls, we could just specify some future time
> by which we want have a quiescent state. We could even push the time to
> be something _just_ before we would have declared a stall.
I still think it's totally the wrong place. cond_resched() is in so
many fast paths (every lock, every allocation). It just doesn't
make sense to add non essential things like this to it.
I would be rather to just revert the original patch.
-Andi
next prev parent reply other threads:[~2014-06-18 0:15 UTC|newest]
Thread overview: 49+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-06-13 20:04 [bisected] pre-3.16 regression on open() scalability Dave Hansen
2014-06-13 22:45 ` Paul E. McKenney
2014-06-13 23:35 ` Dave Hansen
2014-06-14 2:03 ` Paul E. McKenney
2014-06-17 23:10 ` Dave Hansen
2014-06-18 0:00 ` Josh Triplett
2014-06-18 0:15 ` Andi Kleen [this message]
2014-06-18 1:04 ` Paul E. McKenney
2014-06-18 2:27 ` Andi Kleen
2014-06-18 4:47 ` Paul E. McKenney
2014-06-18 12:40 ` Andi Kleen
2014-06-18 12:56 ` Paul E. McKenney
2014-06-18 14:29 ` Christoph Lameter
2014-06-18 0:18 ` Paul E. McKenney
2014-06-18 6:33 ` Dave Hansen
2014-06-18 12:58 ` Paul E. McKenney
2014-06-18 17:36 ` Dave Hansen
2014-06-18 20:30 ` Paul E. McKenney
2014-06-18 23:51 ` Paul E. McKenney
2014-06-19 1:42 ` Andi Kleen
2014-06-19 2:13 ` Paul E. McKenney
2014-06-19 2:29 ` Paul E. McKenney
2014-06-19 2:50 ` Mike Galbraith
2014-06-19 4:19 ` Paul E. McKenney
2014-06-19 3:38 ` Andi Kleen
2014-06-19 4:19 ` Paul E. McKenney
2014-06-19 5:24 ` Mike Galbraith
2014-06-19 18:14 ` Paul E. McKenney
2014-06-19 4:52 ` Eric Dumazet
2014-06-19 5:23 ` Paul E. McKenney
2014-06-19 14:42 ` Christoph Lameter
2014-06-19 18:09 ` Paul E. McKenney
2014-06-19 20:31 ` Christoph Lameter
2014-06-19 20:42 ` Paul E. McKenney
2014-06-19 20:50 ` Andi Kleen
2014-06-19 21:03 ` Paul E. McKenney
2014-06-19 21:13 ` Christoph Lameter
2014-06-19 21:16 ` Christoph Lameter
2014-06-19 21:32 ` josh
2014-06-19 23:07 ` Paul E. McKenney
2014-06-20 15:20 ` Christoph Lameter
2014-06-20 15:38 ` Paul E. McKenney
2014-06-20 16:07 ` Christoph Lameter
2014-06-20 16:30 ` Paul E. McKenney
2014-06-20 17:39 ` Dave Hansen
2014-06-20 18:15 ` Paul E. McKenney
2014-06-18 21:48 ` Paul E. McKenney
2014-06-18 22:03 ` Dave Hansen
2014-06-18 22:52 ` Paul E. McKenney
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=20140618001517.GL8178@tassilo.jf.intel.com \
--to=ak@linux.intel.com \
--cc=cl@linux.com \
--cc=dave.hansen@intel.com \
--cc=josh@joshtriplett.org \
--cc=linux-kernel@vger.kernel.org \
--cc=paulmck@linux.vnet.ibm.com \
--cc=tim.c.chen@intel.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 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.