From: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
To: netdev@vger.kernel.org
Cc: Eric Dumazet <eric.dumazet@gmail.com>,
Neal Cardwell <ncardwell@google.com>
Subject: Hitting WARN_ON_ONCE on skb_try_coalesce
Date: Wed, 3 Aug 2016 20:36:05 -0300 [thread overview]
Message-ID: <20160803233605.GF2967@localhost.localdomain> (raw)
Hi,
I have two namespaces linked with a veth and netem in one of them adding 20ms
latency, and doing netperf from one to another. I'm on commit
7cf210dc267861360cf6968b69bf512475aca985, net updated today, and I'm hitting
this:
[ 1043.024555] WARNING: CPU: 3 PID: 19902 at ...../linux/net/core/skbuff.c:4289 skb_try_coalesce+0x414/0x430
[ 1043.024556] Kernel panic - not syncing: panic_on_warn set ...
[ 1043.024601] CPU: 3 PID: 19902 Comm: netserver Not tainted 4.7.0-standard+ #20
[ 1043.024631] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.9.1-1.fc24 04/01/2014
[ 1043.024667] 0000000000000086 00000000db033d46 ffff880035733ad8 ffffffff813d630d
[ 1043.024705] 0000000000000000 ffffffff81a35655 ffff880035733b60 ffffffff811b5e3d
[ 1043.024741] ffff880000000008 ffff880035733b70 ffff880035733b08 00000000db033d46
[ 1043.024777] Call Trace:
[ 1043.024794] [<ffffffff813d630d>] dump_stack+0x63/0x86
[ 1043.024818] [<ffffffff811b5e3d>] panic+0xe4/0x226
[ 1043.024842] [<ffffffff816ae244>] ? skb_try_coalesce+0x414/0x430
[ 1043.024870] [<ffffffff810a5f13>] __warn+0xe3/0xf0
[ 1043.024892] [<ffffffff810a602d>] warn_slowpath_null+0x1d/0x20
[ 1043.024919] [<ffffffff816ae244>] skb_try_coalesce+0x414/0x430
[ 1043.024945] [<ffffffff817165f3>] tcp_try_coalesce+0x63/0xd0
[ 1043.024977] [<ffffffff81718497>] tcp_queue_rcv+0x57/0x140
[ 1043.025003] [<ffffffff8171e132>] tcp_rcv_established+0x3f2/0x6f0
[ 1043.025029] [<ffffffff81728815>] tcp_v4_do_rcv+0x145/0x200
[ 1043.025054] [<ffffffff816a4387>] __release_sock+0x87/0xf0
[ 1043.025078] [<ffffffff816a4420>] release_sock+0x30/0xa0
[ 1043.025102] [<ffffffff81711d8c>] tcp_recvmsg+0x5fc/0xb70
[ 1043.025126] [<ffffffff8174051e>] inet_recvmsg+0x7e/0xb0
[ 1043.025151] [<ffffffff8169f61d>] sock_recvmsg+0x3d/0x50
[ 1043.025175] [<ffffffff8169f85a>] SYSC_recvfrom+0xda/0x150
[ 1043.025201] [<ffffffff811157df>] ? do_setitimer+0x1bf/0x220
[ 1043.025226] [<ffffffff81115891>] ? alarm_setitimer+0x51/0x90
[ 1043.025252] [<ffffffff816a0bbe>] SyS_recvfrom+0xe/0x10
[ 1043.025277] [<ffffffff817df132>] entry_SYSCALL_64_fastpath+0x1a/0xa4
That's:
WARN_ON_ONCE(delta < len);
Netperf command line is:
netperf -cC -l 10 -t TCP_STREAM \
-H 192.168.10.1 -L 192.168.10.2
netem is added on netperf/client side.
It triggers it right at the beginning of the test.
I can reproduce it 100% of the times with above setup.
- Removing netem made it stop reproducing.
- Reducing netem latency to 10ms, still warns
- Reducing netem latency to 1ms, still warns
- Reducing netem latency to 0ms, it stops reproducing it
This system is not generating a kdump, I'm looking into it.
Thanks,
Marcelo
next reply other threads:[~2016-08-03 23:36 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-08-03 23:36 Marcelo Ricardo Leitner [this message]
2016-08-04 5:55 ` Hitting WARN_ON_ONCE on skb_try_coalesce Eric Dumazet
2016-08-04 14:07 ` Marcelo Ricardo Leitner
2016-08-04 16:48 ` Marcelo Ricardo Leitner
2016-08-04 16:56 ` Eric Dumazet
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=20160803233605.GF2967@localhost.localdomain \
--to=marcelo.leitner@gmail.com \
--cc=eric.dumazet@gmail.com \
--cc=ncardwell@google.com \
--cc=netdev@vger.kernel.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).