linux-kernel - Re: Wrong module .text address in 4.16.0

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]

Date:   Mon, 16 Apr 2018 08:13:16 -0700
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     Jessica Yu <jeyu@...nel.org>
Cc:     Christian Borntraeger <borntraeger@...ibm.com>,
        Thomas-Mich Richter <tmricht@...ux.ibm.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Martin Schwidefsky <schwidefsky@...ibm.com>,
        Hendrik Brueckner <brueckner@...ux.vnet.ibm.com>,
        Heiko Carstens <heiko.carstens@...ibm.com>,
        Peter Zijlstra <peterz@...radead.org>,
        Arnaldo Carvalho de Melo <acme@...nel.org>,
        "Tobin C . Harding" <me@...in.cc>,
        Kees Cook <keescook@...omium.org>
Subject: Re: Wrong module .text address in 4.16.0

On Mon, Apr 16, 2018 at 6:43 AM, Jessica Yu <jeyu@...nel.org> wrote:
>
> So for users of /sys/module/*/sections, we will need to work around
> this and possibly use %px for the real address. But perhaps we should
> base the usage of %px on kptr_restrict?

Maybe. I was hoping we would be able to get rid of it eventually.

The real problem is that those darn module_attribute things don't have
proper IO routines. They *only* have the show routine, and that
doesn't even get the 'struct file' pointer passed to it, just the
buffer to fill in (not even a _size_ of a buffer - we're talking the
bad bad old days of nasty /proc interfaces).

Why is that a problem? Without a 'struct file' we can't even do
permission checking right. %pK worked by doing disgusting wrong
things.

Now, in this case, at least the files are root-owned, and legible only
to root, so I guess we can say that permissions have been properly
checked at open time (not really true: the CAP_SYSLOG bit wasn't!, but
I doubt anybody really cares), and so we could just check
kptr_restrict.

Oh well.

Something like the attached, perhaps? Completely untested, and I don't
even want credit for this if it is used.

              Linus

View attachment "patch.diff" of type "text/x-patch" (675 bytes)