netdev - OOM triggered by SCTP

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-next>] [day] [month] [year] [list]

Message-ID: <CAJPywTL5aKYB40FsAFYFEuhErhgQpYZP5Q_ipMG9pDxqipcEDg@mail.gmail.com>
Date:   Tue, 16 Jul 2019 23:47:40 +0200
From:   Marek Majkowski <marek@...udflare.com>
To:     vyasevich@...il.com, nhorman@...driver.com,
        marcelo.leitner@...il.com, linux-sctp@...r.kernel.org
Cc:     netdev@...r.kernel.org, kernel-team <kernel-team@...udflare.com>
Subject: OOM triggered by SCTP

Morning,

My poor man's fuzzer found something interesting in SCTP. It seems
like creating large number of SCTP sockets + some magic dance, upsets
a memory subsystem related to SCTP. The sequence:

 - create SCTP socket
 - call setsockopts (SCTP_EVENTS)
 - call bind(::1, port)
 - call sendmsg(long buffer, MSG_CONFIRM, ::1, port)
 - close SCTP socket
 - repeat couple thousand times

Full code:
https://gist.github.com/majek/bd083dae769804d39134ce01f4f802bb#file-test_sctp-c

I'm running it on virtme the simplest way:
$ virtme-run --show-boot-console --rw --pwd --kimg bzImage --memory
512M --script-sh ./test_sctp

Originally I was running it inside net namespace, and just having a
localhost interface is sufficient to trigger the problem.

Kernel is 5.2.1 (with KASAN and such, but that shouldn't be a factor).
In some tests I saw a message that might indicate something funny
hitting neighbor table:

neighbour: ndisc_cache: neighbor table overflow!

I'm not addr-decoding the stack trace, since it seems unrelated to the
root cause.

Cheers,
    Marek