Re: [PATCH v2 3/3] mm: perform the mapping_map_writable() check after call_mmap()

From: Lorenzo Stoakes <lstoakes@gmail.com>
To: Andy Lutomirski <luto@kernel.org>
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Andrew Morton <akpm@linux-foundation.org>,
	Matthew Wilcox <willy@infradead.org>,
	Mike Kravetz <mike.kravetz@oracle.com>,
	Muchun Song <muchun.song@linux.dev>,
	Alexander Viro <viro@zeniv.linux.org.uk>,
	Christian Brauner <brauner@kernel.org>,
	linux-fsdevel@vger.kernel.org, Jan Kara <jack@suse.cz>,
	Hugh Dickins <hughd@google.com>
Subject: Re: [PATCH v2 3/3] mm: perform the mapping_map_writable() check after call_mmap()
Date: Tue, 2 May 2023 08:57:07 +0100	[thread overview]
Message-ID: <7565426e-1080-4521-afdd-4dfbfbc63c9b@lucifer.local> (raw)
In-Reply-To: <CALCETrV1QWSjZR_PQgQdyS8rrg4hhrs1u+FyJh43H-gA7CzkFg@mail.gmail.com>

On Mon, May 01, 2023 at 12:02:00PM -0700, Andy Lutomirski wrote:
> On Sun, Apr 30, 2023 at 3:26 PM Lorenzo Stoakes <lstoakes@gmail.com> wrote:
> >
> > In order for a F_SEAL_WRITE sealed memfd mapping to have an opportunity to
> > clear VM_MAYWRITE, we must be able to invoke the appropriate vm_ops->mmap()
> > handler to do so. We would otherwise fail the mapping_map_writable() check
> > before we had the opportunity to avoid it.
>
> Is there any reason this can't go before patch 3?

I don't quite understand what you mean by this? I mean sure, we could, but
intent was to build to this point and leave the most controversial change
for last :)

>
> If I'm understanding correctly, a comment like the following might
> make this a lot more comprehensible:
>
> >
> > This patch moves this check after the call_mmap() invocation. Only memfd
> > actively denies write access causing a potential failure here (in
> > memfd_add_seals()), so there should be no impact on non-memfd cases.
> >
> > This patch makes the userland-visible change that MAP_SHARED, PROT_READ
> > mappings of an F_SEAL_WRITE sealed memfd mapping will now succeed.
> >
> > Link: https://bugzilla.kernel.org/show_bug.cgi?id=217238
> > Signed-off-by: Lorenzo Stoakes <lstoakes@gmail.com>
> > ---
> >  mm/mmap.c | 12 ++++++------
> >  1 file changed, 6 insertions(+), 6 deletions(-)
> >
> > diff --git a/mm/mmap.c b/mm/mmap.c
> > index 646e34e95a37..1608d7f5a293 100644
> > --- a/mm/mmap.c
> > +++ b/mm/mmap.c
> > @@ -2642,17 +2642,17 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
> >         vma->vm_pgoff = pgoff;
> >
> >         if (file) {
> > -               if (is_shared_maywrite(vm_flags)) {
> > -                       error = mapping_map_writable(file->f_mapping);
> > -                       if (error)
> > -                               goto free_vma;
> > -               }
> > -
> >                 vma->vm_file = get_file(file);
> >                 error = call_mmap(file, vma);
> >                 if (error)
> >                         goto unmap_and_free_vma;
> >
>
> /* vm_ops->mmap() may have changed vma->flags.  Check for writability now. */
>

Ack, will add on next spin.

> > +               if (vma_is_shared_maywrite(vma)) {
> > +                       error = mapping_map_writable(file->f_mapping);
> > +                       if (error)
> > +                               goto close_and_free_vma;
> > +               }
> > +
>
> Alternatively, if anyone is nervous about the change in ordering here,
> there could be a whole new vm_op like adjust_vma_flags() that happens
> before any of this.

Agreed, clearly this change is the most controversial thing here. I did
look around and couldn't find any instance where this could cause an issue,
since it is purely the mapping_map_writable() that gets run at a different
point, but this is certainly an alterative.

I have a feeling people might find adding a new op there possibly _more_
nerve-inducing :) but it's an option.

>
> --Andy