linux-kernel - Re: [kernel-hardening] Re: [RFC v4 PATCH 00/13] HARDENED

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: <20161110213952.GB3117@twins.programming.kicks-ass.net>
Date:   Thu, 10 Nov 2016 22:39:52 +0100
From:   Peter Zijlstra <peterz@...radead.org>
To:     David Windsor <dave@...gbits.org>
Cc:     kernel-hardening@...ts.openwall.com,
        Will Deacon <will.deacon@....com>,
        Greg KH <gregkh@...uxfoundation.org>,
        Elena Reshetova <elena.reshetova@...el.com>,
        Arnd Bergmann <arnd@...db.de>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>,
        "H. Peter Anvin" <h.peter.anvin@...el.com>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [kernel-hardening] Re: [RFC v4 PATCH 00/13] HARDENED_ATOMIC

On Thu, Nov 10, 2016 at 04:23:08PM -0500, David Windsor wrote:
> Discussions have been occurring since KSPP has begun: do we need a

Note that I was not included in any of that. If you hide in a corner on
the intartubes don't be surprised people have no clue what you're on
about.

> specialized type for reference counters?  Oh, wait, we do: kref.
> Wait!  kref is implemented with atomic_t.
> 
> So, what?  We obviously need an atomicity for a reference counter
> type.  So, do we simply implement the HARDENED_ATOMIC protected
> version of atomic_t "inside" of kref and leave atomic_t alone?

But you could provide a small subset of the atomic_t API for that, under
a different type.

That way you don't get utter shite like atomic_cmpxchg_wrap() for
instance.

>From what I can see only all the add/sub variants have overflow checks,
but all the operations get _wrap() prefixes, even where it doesn't make
any bloody sense. _wrap() on bitops?, _wrap() on cmpxchg(). You must be
bloody joking right?