linux-kernel - Re: [PATCH v5] powerpc: Do not make the entire heap executable

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]

Date:   Tue, 27 Sep 2016 20:55:44 -0600
From:   Jason Gunthorpe <jgunthorpe@...idianresearch.com>
To:     Michael Ellerman <mpe@...erman.id.au>
Cc:     Al Viro <viro@...iv.linux.org.uk>, linuxppc-dev@...ts.ozlabs.org,
        Andrew Morton <akpm@...ux-foundation.org>,
        Denys Vlasenko <dvlasenk@...hat.com>,
        Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        Paul Mackerras <paulus@...ba.org>,
        "Aneesh Kumar K.V" <aneesh.kumar@...ux.vnet.ibm.com>,
        Kees Cook <keescook@...omium.org>,
        Oleg Nesterov <oleg@...hat.com>,
        Florian Weimer <fweimer@...hat.com>, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v5] powerpc: Do not make the entire heap executable

On Wed, Sep 28, 2016 at 11:42:11AM +1000, Michael Ellerman wrote:

> But this is not really a powerpc patch, and I'm not an ELF expert. So
> I'm not comfortable merging it via the powerpc tree. It doesn't look
> like we really have a maintainer for binfmt_elf.c, so I'm not sure who
> should be acking that part.

Thanks a bunch for looking at this Michael.

> I've added Al Viro to Cc, he maintains fs/ and might be interested.

> I've also added Andrew Morton who might be happy to put this in his
> tree, and see if anyone complains?

For those added to the CC, I would re-state my original commit message
more clearly.

My research showed that the ELF loader bug fixed in this patch is the
root cause bug fix required to implement this hunk:

> > -#define VM_DATA_DEFAULT_FLAGS32	(VM_READ | VM_WRITE | VM_EXEC | \
> > +#define VM_DATA_DEFAULT_FLAGS32 \
> > +	(((current->personality & READ_IMPLIES_EXEC) ? VM_EXEC : 0) | \
> > +				 VM_READ | VM_WRITE | \
> >  				 VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)

Eg that 32 bit powerpc currently unconditionally injects writable,
executable pages into a user space process.

This critically undermines all the W^X security work that has been
done in the tool chain and user space by the PPC community.

I would encourage people to view this as an important security patch
for 32 bit powerpc environments.

Regards,
Jason