netdev - Re: [bug, bisected] pfifo_fast causes packet reordering

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 14 Mar 2018 11:09:02 +0100
From:   Jakob Unterwurzacher <jakob.unterwurzacher@...obroma-systems.com>
To:     John Fastabend <john.fastabend@...il.com>,
        Dave Taht <dave.taht@...il.com>
Cc:     netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
        "David S. Miller" <davem@...emloft.net>,
        "linux-can@...r.kernel.org" <linux-can@...r.kernel.org>,
        Martin Elshuber <martin.elshuber@...obroma-systems.com>
Subject: Re: [bug, bisected] pfifo_fast causes packet reordering

On 14.03.18 05:03, John Fastabend wrote:
>>> During stress-testing our "ucan" USB/CAN adapter SocketCAN driver on Linux
>>> v4.16-rc4-383-ged58d66f60b3 we observed that a small fraction of packets are
>>> delivered out-of-order.
>>>
> 
> Is the stress-testing tool available somewhere? What type of packets
> are being sent?

Not public, no, the problem is that you'd need a CAN adapter as well.

The test is simple, sending CAN frames with an increasing counter and a 
random length payload:

[  tx thread                      rx thread    ]
        |                              ^
        v                              |
[ interface 0 ] ---- cable ----> [ interface 1 ]

I'll see if I can come up with a UDP testcase that works with normal 
ethernet interfaces.

> Is this a single queue device or a multiqueue device? Running
> 'tc -s qdisc show dev foo' would help some.

Here you go:

root@...399-q7:~# tc -s qdisc show dev can0
qdisc pfifo_fast 0: root refcnt 2 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 
1 1 1 1 1
  Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
  backlog 0b 0p requeues 0

> If we introduced a OOO edge case somewhere that was not
> intended so I'll take a look into it. But, if you can provide
> a bit more details on how stress testing is done to cause the
> issue that would help.

Will do.

Thanks,
Jakob