netdev - Linux kernel and RFC 7217 support

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

Message-ID: <fae45803-d031-1897-073a-5edccf197426@strotmann.de>
Date:   Sun, 3 Feb 2019 22:31:03 +0100
From:   Carsten Strotmann <cas@...otmann.de>
To:     davem@...emloft.net, kuznet@....inr.ac.ru, yoshfuji@...ux-ipv6.org,
        netdev@...r.kernel.org, fgont@...networks.com
Subject: Linux kernel and RFC 7217 support

Hi,

I'm testing IPv6 stable-secret identifiers SII under Linux RFC 7217.
<https://tools.ietf.org/html/rfc7217>

One important feature of RFC 7217 is to deliver stable IPv6 addresses
for server addressing use that will not change in case of new network
hardware (changed hardware address). Besides other benefits of SII, this
is an important feature compared with EUI64 IPv6 addresses.

However during my testing, the SII addresses were always changing when
changing the network hardware address. After investigating into
NetworkManager (Red Hat), wicked (SUSE) and disabling all these, working
with the kernel alone and sysctl values, the issue was still showing.

I'm not a C programmer, but I took a look into the source code and in

linux-4.20.6/net/ipv6/addrconf.c

I found:

[...]
        sha_init(digest);

        memset(&data, 0, sizeof(data));

        memset(workspace, 0, sizeof(workspace));

--->    memcpy(data.hwaddr, idev->dev->perm_addr, idev->dev->addr_len);

        data.prefix[0] = address->s6_addr32[0];

        data.prefix[1] = address->s6_addr32[1];

        data.secret = secret;

        data.dad_count = dad_count;

        sha_transform(digest, data.__data, workspace);

        temp = *address;

        temp.s6_addr32[2] = (__force __be32)digest[0];

        temp.s6_addr32[3] = (__force __be32)digest[1];

[...]

The marked line looks to me as if the interface hardware address is
mixed into the SHA hash that will generate the SII address.

Is this an correct observation?

If yes, what is the rationale for this?

If my suspected behavior of the Linux kernel is the true, it does really
reduce the usefulness of IPv6 SII addresses in datacenter/enterprise
environments.

Best regards

Carsten Strotmann