netdev - Re: [Linux-kernel-mentees][PATCH] net: usb: rtl8150: prevent set_ethernet

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]

Message-ID: <d7304f13-1dd8-0760-2c89-6d61c0f6ab7f@gmail.com>
Date:   Tue, 29 Sep 2020 14:16:39 +0530
From:   Anant Thazhemadam <anant.thazhemadam@...il.com>
To:     unlisted-recipients:; (no To-header on input)
Cc:     linux-kernel-mentees@...ts.linuxfoundation.org,
        syzbot+abbc768b560c84d92fd3@...kaller.appspotmail.com,
        Petko Manolov <petkan@...leusys.com>,
        "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>, linux-usb@...r.kernel.org,
        netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [Linux-kernel-mentees][PATCH] net: usb: rtl8150: prevent
 set_ethernet_addr from setting uninit address

A sample crash report can be found here.
    https://syzkaller.appspot.com/text?tag=CrashReport&x=17486911900000

The line where the bug seems to get triggered is,

if (!batadv_compare_eth(hard_iface->net_dev->dev_addr,
                    net_dev->dev_addr))
Looks like it goes through the list of ethernet interfaces, and
compares it with the address of the new device; which can
end up going uninitialized too.

The address should have been set by set_ethernet_addr:

    static inline void set_ethernet_addr(rtl8150_t * dev)
    {
        u8 node_id[6];

        get_registers(dev, IDR, sizeof(node_id), node_id);
        memcpy(dev->netdev->dev_addr, node_id, sizeof(node_id));
    }

However, when get_registers() fails (when ret <= 0 or ret > size),
no memory is copied back into node_id, which remains uninitialized.
The address is then set to be this uninitialized node_id value.

Checking for the return value of get_registers() in set_ethernet_addr()
and further checking the value of set_ethernet_addr() where ever it has
been invoked, and handling the condition wherein get_registers() fails
appropriately helps solve this issue.
Thank you for your time.

Thanks,
Anant