Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <alpine.LRH.2.21.1808301639570.15669@math.ut.ee>
 <20180830205527.dmemjwxfbwvkdzk2@suse.de> <alpine.LRH.2.21.1808310711380.17865@math.ut.ee>
 <20180831070722.wnulbbmillxkw7ke@suse.de> <alpine.DEB.2.21.1809081223450.1402@nanos.tec.linutronix.de>
 <20180911114927.gikd3uf3otxn2ekq@suse.de>
In-Reply-To: <20180911114927.gikd3uf3otxn2ekq@suse.de>
From:   Linus Torvalds <torvalds@linux-foundation.org>
Date:   Tue, 11 Sep 2018 08:20:33 -1000
Message-ID: <CA+55aFzo3b2aChbJ2aOSvKbguYKMG8wv02NS8qzp6w2T5z8WTg@mail.gmail.com>
Subject: Re: 32-bit PTI with THP = userspace corruption
To:     Joerg Roedel <jroedel@suse.de>
Cc:     Thomas Gleixner <tglx@linutronix.de>, Meelis Roos <mroos@linux.ee>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        linux-mm <linux-mm@kvack.org>,
        Andrea Arcangeli <aarcange@redhat.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Sep 11, 2018 at 1:49 AM Joerg Roedel <jroedel@suse.de> wrote:
>
> I had a look into the THP and the HugeTLBfs code, and that is not
> really easy to fix there. As I can see it now, there are a few options
> to fix that, but most of them are ugly:

Just do (4): disable PTI with PAE.

Then we can try to make people perhaps not use !PAE very much, and
warn if you have PAE disabled on a machine that supports it.

As you say, there shouldn't be much of a performance impact from PAE.
There is a more noticeable performance impact from HIGHMEM, not from
HIGHMEM_64G, iirc.

                Linus