netdev - Re: UBSAN: Undefined behaviour in net/ipv4/fib_trie.c:1397:22 and UBSAN: Undefined behaviour in ./arch/x86/include/asm/atomic.h:156:2

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] [day] [month] [year] [list]

Date:	Thu, 28 Jan 2016 09:15:09 -0800
From:	Alexander Duyck <alexander.duyck@...il.com>
To:	Eric Dumazet <eric.dumazet@...il.com>
Cc:	Toralf Förster <toralf.foerster@....de>,
	Alexander Duyck <alexander.duyck@...hat.com>,
	Linux Kernel <linux-kernel@...r.kernel.org>,
	netdev <netdev@...r.kernel.org>
Subject: Re: UBSAN: Undefined behaviour in net/ipv4/fib_trie.c:1397:22 and
 UBSAN: Undefined behaviour in ./arch/x86/include/asm/atomic.h:156:2

On Tue, Jan 26, 2016 at 1:36 PM, Eric Dumazet <eric.dumazet@...il.com> wrote:
> On Tue, 2016-01-26 at 20:45 +0100, Toralf Förster wrote:
>> at a 32 nbit KVM image of a Gentoo Linux runniogn v4.5-rc1 - attached is /var/log/messages
>>
>
> Jan 26 20:40:04 n22kvm-clone kernel: ================================================================================
> Jan 26 20:40:04 n22kvm-clone kernel: UBSAN: Undefined behaviour in net/ipv4/fib_trie.c:1397:22
> Jan 26 20:40:04 n22kvm-clone kernel: shift exponent 32 is too large for 32-bit type 'long unsigned int'
> Jan 26 20:40:04 n22kvm-clone kernel: CPU: 1 PID: 2050 Comm: dhcpcd Not tainted 4.5.0-rc1 #1
> Jan 26 20:40:04 n22kvm-clone kernel: Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.8.2-0-g33fbe13 by qemu-project.org 04/01/2014
>
> Thanks for the report.
>
> CC netdev and Alexander
>
> Commit 79e5ad2ceb006
> ("fib_trie: Remove leaf_info ")

I looked over the line in question, it is an issue where we are doing
the shift and then sanitizing for the undefined result after the fact.
I'll see about submitting a patch to address this.

- Alex