linux-kernel - Re: [PATCH 4/6] powerpc/64s: Enable barrier

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: <20180502134128.4a31b0fc@kitsune.suse.cz>
Date:   Wed, 2 May 2018 13:41:28 +0200
From:   Michal Suchánek <msuchanek@...e.de>
To:     Michael Ellerman <mpe@...erman.id.au>
Cc:     linuxppc-dev@...abs.org, linux-kernel@...r.kernel.org,
        npiggin@...il.com
Subject: Re: [PATCH 4/6] powerpc/64s: Enable barrier_nospec based on
 firmware settings

On Tue, 01 May 2018 21:11:06 +1000
Michael Ellerman <mpe@...erman.id.au> wrote:

> Michal Suchánek <msuchanek@...e.de> writes:
> > Hello,
> >
> > On Tue, 24 Apr 2018 14:15:57 +1000
> > Michael Ellerman <mpe@...erman.id.au> wrote:
> >  
> >> From: Michal Suchanek <msuchanek@...e.de>
> >> 
> >> Check what firmware told us and enable/disable the barrier_nospec
> >> as appropriate.
> >> 
> >> We err on the side of enabling the barrier, as it's no-op on older
> >> systems, see the comment for more detail.
> >> 
> >> Signed-off-by: Michael Ellerman <mpe@...erman.id.au>  
> ...
> >
> > I am missing the option for the barrier to be disabled by a kernel
> > commandline argument here.
> >
> > It does make sense to add a kernel parameter that is checked on
> > boot to be compatible with other platforms that implement one.  
> 
> No other platforms have an option to disable variant 1 mitigations, so
> there isn't an existing parameter we can use.

Right, I was looking at an older implementation which turned off both
v1 and v2 with same parameter. In current kernel the v1 mitigation is
not turned off at all.
> 
> Which is not to say we can't add one, but I wasn't sure if it was
> really worth it.

The current thinking is that most performance relevant cases are
covered with array_nospec which has little overhead. The less code we
have for this the better ;-)

Thanks

Michal