Re: [PATCH v4] /dev/mem: Revoke mappings when a driver claims the region

From: Dan Williams <dan.j.williams@intel.com>
To: Kees Cook <keescook@chromium.org>
Cc: Greg KH <gregkh@linuxfoundation.org>,
	Arnd Bergmann <arnd@arndb.de>,  Ingo Molnar <mingo@redhat.com>,
	Matthew Wilcox <willy@infradead.org>,
	 Russell King <linux@arm.linux.org.uk>,
	Andrew Morton <akpm@linux-foundation.org>,
	 Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Linux MM <linux-mm@kvack.org>
Subject: Re: [PATCH v4] /dev/mem: Revoke mappings when a driver claims the region
Date: Thu, 7 Apr 2022 11:47:40 -0700	[thread overview]
Message-ID: <CAPcyv4htK6Ur4-B=cfkQeOD__9s7ZfxMTo+uCCPVDLG=_WNy7A@mail.gmail.com> (raw)
In-Reply-To: <202204061243.FB134CA4B1@keescook>

On Wed, Apr 6, 2022 at 12:46 PM Kees Cook <keescook@chromium.org> wrote:
>
> *thread necromancy*

It's alive!

>
> Hi Dan,
>
> I'm doing a KSPP bug scrub and am reviewing
> https://github.com/KSPP/linux/issues/74 again.
>
> Do you have a chance to look at this? I'd love a way to make mmap()
> behave the same way as read() for the first meg of /dev/mem.

You want 0-reads or SIGBUS when attempting to access the first 1MB?

Because it sounds like what you want is instead of loudly failing with
-EPERM in drivers/char/mem.c::mmap_mem() you want it to silently
succeed but swap in the zero page, right? Otherwise if it's SIGBUS
then IO_STRICT_DEVMEM=y + marking that span as IORESOURCE_BUSY will
"Do the Right Thing (TM).".

>
> -Kees
>
> On Thu, May 21, 2020 at 08:01:53PM -0700, Kees Cook wrote:
> > On Thu, May 21, 2020 at 02:06:17PM -0700, Dan Williams wrote:
> > > The typical usage of unmap_mapping_range() is part of
> > > truncate_pagecache() to punch a hole in a file, but in this case the
> > > implementation is only doing the "first half" of a hole punch. Namely it
> > > is just evacuating current established mappings of the "hole", and it
> > > relies on the fact that /dev/mem establishes mappings in terms of
> > > absolute physical address offsets. Once existing mmap users are
> > > invalidated they can attempt to re-establish the mapping, or attempt to
> > > continue issuing read(2) / write(2) to the invalidated extent, but they
> > > will then be subject to the CONFIG_IO_STRICT_DEVMEM checking that can
> > > block those subsequent accesses.
> >
> > Nice!
> >
> > Reviewed-by: Kees Cook <keescook@chromium.org>
> >
> > And a thread hijack...   ;)
> >
> > I think this is very close to providing a way to solve another issue
> > I've had with /dev/mem, which is to zero the view of the first 1MB of
> > /dev/mem via mmap. I only fixed the read/write accesses:
> > a4866aa81251 ("mm: Tighten x86 /dev/mem with zeroing reads")
> > I.e. the low 1MB range should be considered allowed, but any reads will see
> > zeros.
> >
> > > +   unmap_mapping_range(inode->i_mapping, res->start, resource_size(res), 1);
> >
> > Is unmap_mapping_range() sufficient for this? Would it need to happen
> > once during open_port() or something more special during mmap_mem()?
> >
> > --
> > Kees Cook
>
> --
> Kees Cook