linux-kernel - Re: [PATCH v2 3/5] x86/mm: check exec permissions on fault

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: <5c06a3d7-bed8-4b02-6509-a7f6c138ee96@intel.com>
Date:   Mon, 25 Oct 2021 11:00:19 -0700
From:   Dave Hansen <dave.hansen@...el.com>
To:     Nadav Amit <nadav.amit@...il.com>
Cc:     Linux-MM <linux-mm@...ck.org>, LKML <linux-kernel@...r.kernel.org>,
        Andrea Arcangeli <aarcange@...hat.com>,
        Andrew Cooper <andrew.cooper3@...rix.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Andy Lutomirski <luto@...nel.org>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        Peter Xu <peterx@...hat.com>,
        Peter Zijlstra <peterz@...radead.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Will Deacon <will@...nel.org>, Yu Zhao <yuzhao@...gle.com>,
        Nick Piggin <npiggin@...il.com>, x86@...nel.org
Subject: Re: [PATCH v2 3/5] x86/mm: check exec permissions on fault

On 10/25/21 10:51 AM, Nadav Amit wrote:
>> On Oct 25, 2021, at 10:45 AM, Dave Hansen <dave.hansen@...el.com> wrote:
>> On 10/25/21 9:19 AM, Nadav Amit wrote:
>>> That was my first version, but I was concerned that perhaps there is
>>> some strange scenario in which both X86_PF_WRITE and X86_PF_INSN can
>>> be set. That is the reason that Peter asked you whether this is
>>> something that might happen.
>>>
>>> If you confirm they cannot be both set, I would the version you just
>>> mentioned.
>> I'm pretty sure they can't be set together on any sane hardware.  A
>> bonkers hypervisor or CPU could do it of course, but they'd be crazy.
>>
>> BTW, feel free to add a WARN_ON_ONCE() if WRITE and INSN are both set.
>> That would be a nice place to talk about the assumption.
>>
> I can do that. But be aware that if the assumption is broken, it might
> lead to the application getting stuck in an infinite loop of
> page-faults instead of receiving SIGSEGV.

If we have a bonkers hypervisor/CPU, I'm OK with a process that hangs
like that, especially if we can ^C it and see its stream of page faults
with tracing or whatever.

Couldn't we just also do:

	if ((code & (X86_PF_WRITE|X86_PF_INSN) ==
                    (X86_PF_WRITE|X86_PF_INSN)) {
		WARN_ON_ONCE(1);
		return 1;
	}

That should give you the WARN_ON_ONCE() and also return an affirmative
access_error(), resulting in a SIGSEGV.

(I'm not sure I like the indentation as I wrote it here... just do what
looks best in the code)