[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200622175343.GC13061@mtj.duckdns.org>
Date: Mon, 22 Jun 2020 13:53:43 -0400
From: Tejun Heo <tj@...nel.org>
To: Rick Lindsley <ricklind@...ux.vnet.ibm.com>
Cc: Ian Kent <raven@...maw.net>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Stephen Rothwell <sfr@...b.auug.org.au>,
Andrew Morton <akpm@...ux-foundation.org>,
Al Viro <viro@...iv.linux.org.uk>,
David Howells <dhowells@...hat.com>,
Miklos Szeredi <miklos@...redi.hu>,
linux-fsdevel <linux-fsdevel@...r.kernel.org>,
Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2 0/6] kernfs: proposed locking and concurrency
improvement
On Fri, Jun 19, 2020 at 07:44:29PM -0700, Rick Lindsley wrote:
> echo 0 > /sys/devices//system/memory/memory10374/online
>
> and boom - you've taken memory chunk 10374 offline.
>
> These changes are not just a whim. I used lockstat to measure contention
> during boot. The addition of 250,000 "devices" in parallel created
> tremendous contention on the kernfs_mutex and, it appears, on one of the
> directories within it where memory nodes are created. Using a mutex means
> that the details of that mutex must bounce around all the cpus ... did I
> mention 1500+ cpus? A whole lot of thrash ...
I don't know. The above highlights the absurdity of the approach itself to
me. You seem to be aware of it too in writing: 250,000 "devices".
Thanks.
--
tejun
Powered by blists - more mailing lists