linux-kernel - Re: [PATCH mm 1/2] kfence: add option to use KFENCE without static keys

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: <20210111195323.GA842777@cork>
Date:   Mon, 11 Jan 2021 11:53:23 -0800
From:   Jörn Engel <joern@...estorage.com>
To:     Marco Elver <elver@...gle.com>
Cc:     akpm@...ux-foundation.org, glider@...gle.com, dvyukov@...gle.com,
        andreyknvl@...gle.com, jannh@...gle.com, mark.rutland@....com,
        linux-kernel@...r.kernel.org, linux-mm@...ck.org,
        kasan-dev@...glegroups.com
Subject: Re: [PATCH mm 1/2] kfence: add option to use KFENCE without static
 keys

On Mon, Jan 11, 2021 at 10:15:43AM +0100, Marco Elver wrote:
> For certain usecases, specifically where the sample interval is always
> set to a very low value such as 1ms, it can make sense to use a dynamic
> branch instead of static branches due to the overhead of toggling a
> static branch.

I ended up with 100µs and couldn't measure a performance problem in our
benchmarks.  My results don't have predictive value for anyone else, of
course.

> Therefore, add a new Kconfig option to remove the static branches and
> instead check kfence_allocation_gate if a KFENCE allocation should be
> set up.
> 
> Suggested-by: Jörn Engel <joern@...estorage.com>
> Signed-off-by: Marco Elver <elver@...gle.com>

Reviewed-by: Jörn Engel <joern@...estorage.com>

Jörn

--
One of the things I’ve discovered over the years, is that you can
create change or you can receive credit – not both.
-- Stephen Downes