From mboxrd@z Thu Jan  1 00:00:00 1970
From: Linus Torvalds <torvalds-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org>
Subject: Re: [PATCH 2/2] mm: introduce MAP_VALIDATE, a mechanism for for
 safely defining new mmap flags
Date: Thu, 31 Aug 2017 18:27:31 -0700
Message-ID: <CA+55aFwsfUj1f41w8hqt9LN3-ajmJ=2AB1Nb6ZzwHgE1OKxGOw@mail.gmail.com>
References: <150413449482.5923.1348069619036923853.stgit@dwillia2-desk3.amr.corp.intel.com>
 <150413450616.5923.7069852068237042023.stgit@dwillia2-desk3.amr.corp.intel.com>
 <20170831100359.GD21443@lst.de> <CAPcyv4jvTB4Aiei1-fGybyJNopXQy9zADpnFcuRNdZCS4Mf1QQ@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Return-path: <linux-api-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org>
In-Reply-To: <CAPcyv4jvTB4Aiei1-fGybyJNopXQy9zADpnFcuRNdZCS4Mf1QQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
Sender: linux-api-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
To: Dan Williams <dan.j.williams-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Cc: Christoph Hellwig <hch-jcswGhMUV9g@public.gmane.org>, Linux MM <linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org>, Jan Kara <jack-AlSwsSmVLrQ@public.gmane.org>, Arnd Bergmann <arnd-r2nGTMty4D4@public.gmane.org>, "linux-nvdimm-hn68Rpc1hR1g9hUCZPvPmw@public.gmane.org" <linux-nvdimm-hn68Rpc1hR1g9hUCZPvPmw@public.gmane.org>, Linux API <linux-api-u79uwXL29TY76Z2rM5mHXA@public.gmane.org>, Andy Lutomirski <luto-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>, Andrew Morton <akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org>
List-Id: linux-api@vger.kernel.org

On Thu, Aug 31, 2017 at 6:01 PM, Dan Williams <dan.j.williams-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org> wrote:
>
> Ugh, nommu defeats the MAP_SHARED_VALIDATE proposal from Linus.
>
>         if ((flags & MAP_TYPE) != MAP_PRIVATE &&
>             (flags & MAP_TYPE) != MAP_SHARED)
>                 return -EINVAL;
>
> ...parisc strikes again.

Why? That's no different from the case statement for the mmu case,
just written differently.

You *want* existing kernels to fail, since they don't test the bits
you want to test.

So you just want to rewrite these all as

    switch (flags & MAP_TYPE) {
    case MAP_SHARED_VALIDATE:
        .. validate the other bits...
        /* fallhtough */
    case MAP_SHARED:
        .. do the shared case ..
    case MAP_PRIVATE:
        .. do the private case ..
    default:
        return -EINVAL;
    }

and you're all good.

I'm not seeing the problem.

Of course, I also suspect that for nommu you might as well just always
return -EINVAL anyway. The only people who would ever use
MAP_SHARED_VALIDATE are the kinds of people who do things that just
aren't likely relevant on nommu, but whatever..

               Linus