linux-kernel - Re: [PATCH 01/13] mm: Update ptep_get

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: <CAHk-=wgiDa9Xy5rLMR_gvC-WVvdkxABgXXADbKhwswYof7ptdg@mail.gmail.com>
Date:   Sat, 29 Oct 2022 12:14:21 -0700
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     Nadav Amit <nadav.amit@...il.com>
Cc:     Peter Zijlstra <peterz@...radead.org>,
        Jann Horn <jannh@...gle.com>,
        John Hubbard <jhubbard@...dia.com>, X86 ML <x86@...nel.org>,
        Matthew Wilcox <willy@...radead.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        kernel list <linux-kernel@...r.kernel.org>,
        Linux-MM <linux-mm@...ck.org>,
        Andrea Arcangeli <aarcange@...hat.com>,
        "Kirill A . Shutemov" <kirill.shutemov@...ux.intel.com>,
        jroedel@...e.de, ubizjak@...il.com,
        Alistair Popple <apopple@...dia.com>
Subject: Re: [PATCH 01/13] mm: Update ptep_get_lockless()s comment

On Sat, Oct 29, 2022 at 11:58 AM Linus Torvalds
<torvalds@...ux-foundation.org> wrote:
>
> Still not tested in any way, shape, or form. I decided I wanted to
> send this one before booting into this and possibly blowing up ;^)

Well, it boots, and I see no difference with your PoC code.

It didn't fail for me before, it doesn't fail for me with those patches.

Again, the "it doesn't fail for me" is probably because I'm running it
incorrectly, although for all I know there can also be hardware
differences.

I'm testing on an older AMD threadripper, and as I'm sure you are very
aware, some AMD cores used to have special support for keeping the TLB
coherent with the actual page table contents in order to then avoid
TLB flushes entirely.

Those things ended up being buggy and disabled, but my point is that
hardware differences can obviously actively hide this issue by making
the TLB contents track page table changes.

So even if I were to run it the same way you do, I might not see the
failure due to just running it on different hardware with different
TLB and timing.

Anyway, the patches don't seem to cause any *obvious* problems. That's
not to say that they are correct, or that they fix anything, but it's
certainly a fairly simple and straightforward patch, and it "feels
right" to me.

Sadly, reality doesn't always agree with my feelings. Damn.

                Linus