From: SeongJae Park <sjpark@amazon.com>
To: Eric Dumazet <eric.dumazet@gmail.com>
Cc: SeongJae Park <sjpark@amazon.com>,
Eric Dumazet <edumazet@google.com>,
David Miller <davem@davemloft.net>,
Al Viro <viro@zeniv.linux.org.uk>,
"Jakub Kicinski" <kuba@kernel.org>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
<sj38.park@gmail.com>, netdev <netdev@vger.kernel.org>,
LKML <linux-kernel@vger.kernel.org>,
SeongJae Park <sjpark@amazon.de>, <snu@amazon.com>,
<amit@kernel.org>, <stable@vger.kernel.org>
Subject: Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change
Date: Tue, 5 May 2020 17:46:44 +0200 [thread overview]
Message-ID: <20200505154644.18997-1-sjpark@amazon.com> (raw)
In-Reply-To: <a8510327-d4f0-1207-1342-d688e9d5b8c3@gmail.com> (raw)
On Tue, 5 May 2020 08:20:50 -0700 Eric Dumazet <eric.dumazet@gmail.com> wrote:
>
>
> On 5/5/20 8:07 AM, SeongJae Park wrote:
> > On Tue, 5 May 2020 07:53:39 -0700 Eric Dumazet <edumazet@google.com> wrote:
> >
>
> >> Why do we have 10,000,000 objects around ? Could this be because of
> >> some RCU problem ?
> >
> > Mainly because of a long RCU grace period, as you guess. I have no idea how
> > the grace period became so long in this case.
> >
> > As my test machine was a virtual machine instance, I guess RCU readers
> > preemption[1] like problem might affected this.
> >
> > [1] https://www.usenix.org/system/files/conference/atc17/atc17-prasad.pdf
> >
> >>
> >> Once Al patches reverted, do you have 10,000,000 sock_alloc around ?
> >
> > Yes, both the old kernel that prior to Al's patches and the recent kernel
> > reverting the Al's patches didn't reproduce the problem.
> >
>
> I repeat my question : Do you have 10,000,000 (smaller) objects kept in slab caches ?
>
> TCP sockets use the (very complex, error prone) SLAB_TYPESAFE_BY_RCU, but not the struct socket_wq
> object that was allocated in sock_alloc_inode() before Al patches.
>
> These objects should be visible in kmalloc-64 kmem cache.
Not exactly the 10,000,000, as it is only the possible highest number, but I
was able to observe clear exponential increase of the number of the objects
using slabtop. Before the start of the problematic workload, the number of
objects of 'kmalloc-64' was 5760, but I was able to observe the number increase
to 1,136,576.
OBJS ACTIVE USE OBJ SIZE SLABS OBJ/SLAB CACHE SIZE NAME
before: 5760 5088 88% 0.06K 90 64 360K kmalloc-64
after: 1136576 1136576 100% 0.06K 17759 64 71036K kmalloc-64
Thanks,
SeongJae Park
next prev parent reply other threads:[~2020-05-05 15:47 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-05 8:10 [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change SeongJae Park
2020-05-05 8:10 ` [PATCH net v2 1/2] Revert "coallocate socket_wq with socket itself" SeongJae Park
2020-05-06 4:55 ` kbuild test robot
2020-05-05 8:10 ` [PATCH net v2 2/2] Revert "sockfs: switch to ->free_inode()" SeongJae Park
2020-05-05 11:54 ` [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change SeongJae Park
2020-05-05 12:31 ` Nuernberger, Stefan
2020-05-05 14:53 ` Eric Dumazet
2020-05-05 15:07 ` SeongJae Park
2020-05-05 15:20 ` Eric Dumazet
2020-05-05 15:46 ` SeongJae Park [this message]
2020-05-05 16:00 ` Eric Dumazet
2020-05-05 16:13 ` SeongJae Park
2020-05-05 16:25 ` Eric Dumazet
2020-05-05 16:31 ` Eric Dumazet
2020-05-05 16:37 ` Eric Dumazet
2020-05-05 17:05 ` SeongJae Park
2020-05-05 17:30 ` Paul E. McKenney
2020-05-05 17:56 ` SeongJae Park
2020-05-05 18:17 ` Paul E. McKenney
2020-05-05 18:34 ` SeongJae Park
2020-05-05 18:49 ` Paul E. McKenney
2020-05-06 12:59 ` SeongJae Park
2020-05-06 14:33 ` Eric Dumazet
2020-05-06 14:41 ` Paul E. McKenney
2020-05-06 15:20 ` SeongJae Park
2020-05-05 17:28 ` Paul E. McKenney
2020-05-05 18:11 ` SeongJae Park
2020-05-05 17:23 ` Paul E. McKenney
2020-05-05 17:49 ` SeongJae Park
2020-05-05 18:27 ` Paul E. McKenney
2020-05-05 18:40 ` SeongJae Park
2020-05-05 18:48 ` Paul E. McKenney
2020-05-05 16:26 ` Al Viro
2020-05-05 18:48 ` David Miller
2020-05-05 19:00 ` David Miller
2020-05-06 6:24 ` SeongJae Park
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=20200505154644.18997-1-sjpark@amazon.com \
--to=sjpark@amazon.com \
--cc=amit@kernel.org \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=eric.dumazet@gmail.com \
--cc=gregkh@linuxfoundation.org \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=sj38.park@gmail.com \
--cc=sjpark@amazon.de \
--cc=snu@amazon.com \
--cc=stable@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;
as well as URLs for NNTP newsgroup(s).