linux-kernel - Re: 32-bit PTI with THP = userspace corruption

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: <20180918140030.248afa21@alans-desktop>
Date:   Tue, 18 Sep 2018 14:00:30 +0100
From:   Alan Cox <gnomes@...rguk.ukuu.org.uk>
To:     Joerg Roedel <jroedel@...e.de>
Cc:     Meelis Roos <mroos@...ux.ee>, Thomas Gleixner <tglx@...utronix.de>,
        Linux Kernel list <linux-kernel@...r.kernel.org>,
        linux-mm@...ck.org, Andrea Arcangeli <aarcange@...hat.com>,
        Linus Torvalds <torvalds@...ux-foundation.org>
Subject: Re: 32-bit PTI with THP = userspace corruption

On Tue, 11 Sep 2018 14:12:22 +0200
Joerg Roedel <jroedel@...e.de> wrote:

> On Tue, Sep 11, 2018 at 02:58:10PM +0300, Meelis Roos wrote:
> > The machines where I have PAE off are the ones that have less memory. 
> > PAE is off just for performance reasons, not lack of PAE. PAE should be 
> > present on all of my affected machines anyway and current distributions 
> > seem to mostly assume 686 and PAE anyway for 32-bit systems.  
> 
> Right, most distributions don't even provide a non-PAE kernel for their
> users anymore.
> 
> How big is the performance impact of using PAE over legacy paging?

On what system. In the days of the original 36bit PAE Xeons it was around
10% when we measured it at Red Hat, but that was long ago and as you go
newer it really ought to be vanishingly small.

There are pretty much no machines that don't support PAE and are still
even vaguely able to boot a modern Linux kernel. The oddity is the
Pentium-M but most distros shipped a hack to use PAE on the Pentium M
anyway as it seems to work fine.

Alan