linux-kernel - [PATCH] Fix memory overwriting issue when copy an address to user space

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]

Message-ID: <CAEQHRfAE5pNCXko_HOamH-ffLe9fzSxvAh-yOnJe7OcPkFkV4A@mail.gmail.com>
Date:   Tue, 21 Jul 2020 04:23:03 +0000
From:   lebon zhou <lebon.zhou@...il.com>
To:     David Laight <David.Laight@...lab.com>
Cc:     "davem@...emloft.net" <davem@...emloft.net>,
        "kuba@...nel.org" <kuba@...nel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: [PATCH] Fix memory overwriting issue when copy an address to user space

On Mon, Jul 20, 2020 at 11:12 PM David Laight <David.Laight@...lab.com> wrote:
>
> From: lebon zhou
> > Sent: 20 July 2020 05:35
> > To: davem@...emloft.net; kuba@...nel.org
> > Cc: linux-kernel@...r.kernel.org; netdev@...r.kernel.org
> > Subject: [PATCH] Fix memory overwriting issue when copy an address to user space
> >
> > When application provided buffer size less than sockaddr_storage, then
> > kernel will overwrite some memory area which may cause memory corruption,
> > e.g.: in recvmsg case, let msg_name=malloc(8) and msg_namelen=8, then
> > usually application can call recvmsg successful but actually application
> > memory get corrupted.
>
> Where?
> The copy_to_user() uses the short length provided by the user.
> There is even a comment saying that if the address is truncated
> the length returned to the user is the full length.
>
> Maybe the application is reusing the msg without re-initialising
> it properly.

It is not related with copy_to_user(), it is about move_addr_to_user()
implementation itself,
there is comment /*After copying the data up to the limit the user
specifies...*/, but the fact
is when (ulen < klen), this function will copy more content to user
buffer over than user
specifies in @ulen, this will cause the user buffer to corrupt, this
patch fixes this issue.