netdev - Re: [PATCH net-next] ethtool: copy reglen to userspace

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: <20190530064848.GA27401@unicorn.suse.cz>
Date:   Thu, 30 May 2019 08:48:48 +0200
From:   Michal Kubecek <mkubecek@...e.cz>
To:     netdev@...r.kernel.org
Cc:     David Miller <davem@...emloft.net>, vivien.didelot@...il.com,
        linux-kernel@...r.kernel.org, kernel@...oirfairelinux.com,
        linville@...hat.com, f.fainelli@...il.com
Subject: Re: [PATCH net-next] ethtool: copy reglen to userspace

On Wed, May 29, 2019 at 10:17:44PM -0700, David Miller wrote:
> From: Vivien Didelot <vivien.didelot@...il.com>
> Date: Tue, 28 May 2019 16:58:48 -0400
> 
> > ethtool_get_regs() allocates a buffer of size reglen obtained from
> > ops->get_regs_len(), thus only this value must be used when copying
> > the buffer back to userspace. Also no need to check regbuf twice.
> > 
> > Signed-off-by: Vivien Didelot <vivien.didelot@...il.com>
> 
> Hmmm, can't regs.len be modified by the driver potentially?

The driver certainly shouldn't raise it as that could result in kernel
writing past the buffer provided by userspace. (I'll check some drivers
to see if they truncate the dump or return an error if regs.len from
userspace is insufficient.) And lowering it would be also wrong as that
would mean dump would be shorter than what ops->get_regs_len() returned.

Michal Kubecek