linux-kernel - Re: [kernel-hardening] [PATCH] arm: kernel: implement fast refcount checking

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:   Wed, 3 Jan 2018 13:57:55 +0000
From:   Ard Biesheuvel <ard.biesheuvel@...aro.org>
To:     Jinbum Park <jinb.park7@...il.com>
Cc:     Dave Martin <Dave.Martin@....com>,
        linux-arm-kernel@...ts.infradead.org,
        LKML <linux-kernel@...r.kernel.org>,
        Kernel Hardening <kernel-hardening@...ts.openwall.com>,
        Russell King <linux@...linux.org.uk>,
        Will Deacon <will.deacon@....com>,
        Peter Zijlstra <peterz@...radead.org>, boqun.feng@...il.com,
        Ingo Molnar <mingo@...nel.org>,
        Mark Rutland <mark.rutland@....com>,
        Nicolas Pitre <nicolas.pitre@...aro.org>, mickael.guene@...com,
        Kees Cook <keescook@...omium.org>,
        Laura Abbott <labbott@...hat.com>
Subject: Re: [kernel-hardening] [PATCH] arm: kernel: implement fast refcount checking

On 3 January 2018 at 13:36, Jinbum Park <jinb.park7@...il.com> wrote:
>>> This is a nice result. However, without any insight into the presence
>>> of actual refcount hot spots, it is not obvious that we need this
>>> patch. This is the reason we ended up enabling CONFIG_REFCOUNT_FULL
>>> for arm64. I will let others comment on whether we want this patch in
>>> the first place,
>
> Dear Ard, Dave,
>
> I wanna hear some comment on above point.
> Is CONFIG_REFCOUNT_FULL much better for arm?
> If it is, I don't need to prepare v2 patch. (then, just needed to add
> "select REFCOUNT_FULL")
>

Well, we should probably turn that around. Please use REFCOUNT_FULL,
until you run into a use case where the slowdown is noticeable. If
nobody ever notices, we don't need to fix anything.