From: Tejun Heo <tj@kernel.org>
To: Jeff Layton <jlayton@primarydata.com>
Cc: linux-nfs@vger.kernel.org, linux-kernel@vger.kernel.org,
Al Viro <viro@zeniv.linux.org.uk>
Subject: Re: [RFC PATCH 00/14] nfsd/sunrpc: add support for a workqueue-based nfsd
Date: Tue, 2 Dec 2014 14:18:14 -0500 [thread overview]
Message-ID: <20141202191814.GK10918@htj.dyndns.org> (raw)
In-Reply-To: <1417544663-13299-1-git-send-email-jlayton@primarydata.com>
Hello, Jeff.
On Tue, Dec 02, 2014 at 01:24:09PM -0500, Jeff Layton wrote:
> 2) get some insight about the latency from those with a better
> understanding of the CMWQ code. Any thoughts as to why we might be
> seeing such high latency here? Any ideas of what we can do about it?
The latency is prolly from concurrency management. Work items which
participate in concurrency management (the ones on per-cpu workqueues
w/o WQ_CPU_INTENSIVE set) tend to get penalized on latency side quite
a bit as the "run" durations for all such work items end up being
serialized on the cpu. Setting WQ_CPU_INTENSIVE on the workqueue
disables concurrency management and so does making the workqueue
unbound. If strict cpu locality is likely to be beneficial and each
work item isn't likely to consume huge amount of cpu cycles,
WQ_CPU_INTENSIVE would fit better; otherwise, WQ_UNBOUND to let the
scheduler do its thing.
Thanks.
--
tejun
next prev parent reply other threads:[~2014-12-02 19:18 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-12-02 18:24 [RFC PATCH 00/14] nfsd/sunrpc: add support for a workqueue-based nfsd Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 01/14] sunrpc: add a new svc_serv_ops struct and move sv_shutdown into it Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 02/14] sunrpc: move sv_function into sv_ops Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 03/14] sunrpc: move sv_module parm " Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 04/14] sunrpc: turn enqueueing a svc_xprt into a svc_serv operation Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 05/14] sunrpc: abstract out svc_set_num_threads to sv_ops Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 06/14] sunrpc: move pool_mode definitions into svc.h Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 07/14] sunrpc: factor svc_rqst allocation and freeing from sv_nrthreads refcounting Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 08/14] sunrpc: set up workqueue function in svc_xprt Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 09/14] sunrpc: add basic support for workqueue-based services Jeff Layton
2014-12-08 20:47 ` J. Bruce Fields
2014-12-08 20:49 ` Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 10/14] nfsd: keep a reference to the fs_struct in svc_rqst Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 11/14] nfsd: add support for workqueue based service processing Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 12/14] sunrpc: keep a cache of svc_rqsts for each NUMA node Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 13/14] sunrpc: add more tracepoints around svc_xprt handling Jeff Layton
2014-12-02 18:24 ` [RFC PATCH 14/14] sunrpc: add tracepoints around svc_sock handling Jeff Layton
2014-12-02 19:18 ` Tejun Heo [this message]
2014-12-02 19:26 ` [RFC PATCH 00/14] nfsd/sunrpc: add support for a workqueue-based nfsd Jeff Layton
2014-12-02 19:29 ` Tejun Heo
2014-12-02 19:26 ` Tejun Heo
2014-12-02 19:46 ` Jeff Layton
2014-12-03 1:11 ` NeilBrown
2014-12-03 1:29 ` Jeff Layton
2014-12-03 15:56 ` Tejun Heo
2014-12-03 16:04 ` Jeff Layton
2014-12-03 19:02 ` Jeff Layton
2014-12-03 19:08 ` Trond Myklebust
2014-12-03 19:20 ` Jeff Layton
2014-12-03 19:59 ` Trond Myklebust
2014-12-03 20:21 ` Jeff Layton
2014-12-03 20:44 ` Trond Myklebust
2014-12-04 11:47 ` Jeff Layton
2014-12-04 17:17 ` Shirley Ma
2014-12-04 17:28 ` Jeff Layton
2014-12-04 17:44 ` Shirley Ma
2014-12-03 16:50 ` Chuck Lever
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=20141202191814.GK10918@htj.dyndns.org \
--to=tj@kernel.org \
--cc=jlayton@primarydata.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-nfs@vger.kernel.org \
--cc=viro@zeniv.linux.org.uk \
/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