netdev - Re: [PATCH net 0/6] net: better validate user provided tunnel names

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: <d904eaa7-4ea3-e55e-cfae-8060af021632@gmail.com>
Date:   Thu, 5 Apr 2018 12:40:40 -0700
From:   Eric Dumazet <eric.dumazet@...il.com>
To:     David Miller <davem@...emloft.net>, edumazet@...gle.com
Cc:     netdev@...r.kernel.org, steffen.klassert@...unet.com,
        eric.dumazet@...il.com
Subject: Re: [PATCH net 0/6] net: better validate user provided tunnel names



On 04/05/2018 12:21 PM, David Miller wrote:
> From: Eric Dumazet <edumazet@...gle.com>
> Date: Thu,  5 Apr 2018 06:39:25 -0700
> 
>> This series changes dev_valid_name() to not attempt reading
>> a possibly too long user-provided device name, then use
>> this helper in five different tunnel providers.
> 
> Series applied and queued up for -stable, thanks Eric.
> 
> Reading over this series makes me wonder if we generally have an
> off-by-one bug for device names which are exactly IFNAMSIZ.
> 
> We validate the size using the test:
> 
> 	if (strlen(name) >= IFNAMSIZ)
> 		return ERROR;
> 
> and thusly after Eric's changes:
> 
> 	if (strnlen(name, IFNAMSIZ) == IFNAMSIZ)
> 		return ERROR;
> 
> This value computed by str{,n}len() doesn't include the trailing null
> byte.
> 
> So we will accept a name that has exactly IFNAMSIZ bytes long not
> including the trailing null.

In this case strnlen(name, IFNAMSIZ) returns IFNAMSIZ.

So  (strnlen(name, IFNAMSIZ) == IFNAMSIZ) would definitely be true.

The only effect of the change is that strlen() would read 1000 bytes of a
malicious string before we reached the test on the length to reject such name.

While strnlen() is guaranteed to not read more than IFNAMSIZ bytes.