Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20221025221755.3810809-1-glider@google.com> <Y1j+Tt9mnMDU0zO+@hirez.programming.kicks-ass.net>
 <CAG_fn=XDeghFBGXT37Mc-ky-8NaPaMmCLdo3Par=xh92Fk_CAQ@mail.gmail.com>
 <Y1o72704bVK0FgCr@hirez.programming.kicks-ass.net> <CAG_fn=XESk1PPqbAVDqMdGbRwyvLvLQrm2hybr2cXaaYjfZEKA@mail.gmail.com>
 <202210271155.33956B1@keescook>
In-Reply-To: <202210271155.33956B1@keescook>
From:   Alexander Potapenko <glider@google.com>
Date:   Thu, 27 Oct 2022 16:24:55 -0700
Message-ID: <CAG_fn=XX2exsGcD3ZR4LGw4Tqy_ietYe46WEOKt7a_nt9Gf=Cw@mail.gmail.com>
Subject: Re: [PATCH] x86/uaccess: instrument copy_from_user_nmi()
To:     Kees Cook <keescook@chromium.org>
Cc:     Peter Zijlstra <peterz@infradead.org>,
        Marco Elver <elver@google.com>, linux-kernel@vger.kernel.org,
        linux-mm@kvack.org, Andrew Morton <akpm@linux-foundation.org>,
        Dave Hansen <dave.hansen@linux.intel.com>, x86@kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, Oct 27, 2022 at 11:58 AM Kees Cook <keescook@chromium.org> wrote:
>
> On Thu, Oct 27, 2022 at 11:26:50AM -0700, Alexander Potapenko wrote:
> > On Thu, Oct 27, 2022 at 1:05 AM Peter Zijlstra <peterz@infradead.org> w=
rote:
> > >
> > > On Wed, Oct 26, 2022 at 11:38:53AM -0700, Alexander Potapenko wrote:
> > > > A bigger issue from the NMI perspective is probably
> > > > having __msan_poison_alloca() inserted in every non-noinstr kernel
> > > > function, because that hook may acquire the stackdepot lock.
> > >
> > > *urgghhh* that's broken, that must not be. There is a *TON* of NMI
> > > functions that are non-noinstr.
> >
> > __msan_poison_alloca() is guarded by kmsan_in_runtime(), which is
> > currently implemented as:
> >
> >   static __always_inline bool kmsan_in_runtime(void)
> >   {
> >           if ((hardirq_count() >> HARDIRQ_SHIFT) > 1)
> >                   return true;
> >           return kmsan_get_context()->kmsan_in_runtime;
> >   }
> >
> > I think the easiest way to fix the NMI situation would be adding "if
> > in_nmi() return true"?
>
> It might help to look through these threads:
>
> https://lore.kernel.org/lkml/20220916135953.1320601-1-keescook@chromium.o=
rg/
> https://lore.kernel.org/all/20220919201648.2250764-1-keescook@chromium.or=
g/

Sorry, I missed that letter, should have responded earlier.

> I wandered around attempting to deal with in_nmi(), etc. And in
> the end just drop the attempt to cover it. It's worth noting that
> copy_from_user_nmi() exists on 1 architecture and has exactly 1
> call-site...

It doesn't really matter for KASAN, because a missing addressability
check is a matter of missing some (possibly rare) bugs.
For KMSAN a missing initialization will result in false positives, and
we already started seeing them: show_opcodes() copies data to a local
and prints it, but without a call to kmsan_unpoison_memory() it will
result in error reports about opcodes[] being uninitialized.

So for this particular case I want to ensure kmsan_unpoison_memory()
can be called from NMI context (by removing the kmsan_in_runtime()
check from it), but to be on the safe side we'll also have to do
nothing in __msan_poison_alloca() under in_nmi().


> --
> Kees Cook


--
Alexander Potapenko
Software Engineer

Google Germany GmbH
Erika-Mann-Stra=C3=9Fe, 33
80636 M=C3=BCnchen

Gesch=C3=A4ftsf=C3=BChrer: Paul Manicle, Liana Sebastian
Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg