Re: [PATCH] Userfaultfd: Avoid double free of userfault_ctx and remove O_CLOEXEC

[Eric Biggers <ebiggers@kernel.org>]

On Tue, Aug 04, 2020 at 01:49:30PM -0700, Lokesh Gidra wrote:
> On Tue, Aug 4, 2020 at 1:45 PM Eric Biggers <ebiggers@kernel.org> wrote:
> >
> > On Tue, Aug 04, 2020 at 01:31:55PM -0700, Lokesh Gidra wrote:
> > > when get_unused_fd_flags returns error, ctx will be freed by
> > > userfaultfd's release function, which is indirectly called by fput().
> > > Also, if anon_inode_getfile_secure() returns an error, then
> > > userfaultfd_ctx_put() is called, which calls mmdrop() and frees ctx.
> > >
> > > Also, the O_CLOEXEC was inadvertently added to the call to
> > > get_unused_fd_flags() [1].
> > >
> > > Adding Al Viro's suggested-by, based on [2].
> > >
> > > [1] https://lore.kernel.org/lkml/1f69c0ab-5791-974f-8bc0-3997ab1d61ea@dancol.org/
> > > [2] https://lore.kernel.org/lkml/20200719165746.GJ2786714@ZenIV.linux.org.uk/
> > >
> > > Fixes: d08ac70b1e0d (Wire UFFD up to SELinux)
> > > Suggested-by: Al Viro <viro@zeniv.linux.org.uk>
> > > Reported-by: syzbot+75867c44841cb6373570@syzkaller.appspotmail.com
> > > Signed-off-by: Lokesh Gidra <lokeshgidra@google.com>
> >
> > What branch does this patch apply to?  Neither mainline nor linux-next works.
> >
> On James Morris' tree (secure_uffd_v5.9 branch).
> 

For those of us not "in the know", that apparently means branch secure_uffd_v5.9
of https://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security.git

Perhaps it would make more sense to resend your original patch series with this
fix folded in?

> diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c
> index ae859161908f..e15eb8fdc083 100644
> --- a/fs/userfaultfd.c
> +++ b/fs/userfaultfd.c
> @@ -2042,24 +2042,18 @@ SYSCALL_DEFINE1(userfaultfd, int, flags)
>  		O_RDWR | (flags & UFFD_SHARED_FCNTL_FLAGS),
>  		NULL);
>  	if (IS_ERR(file)) {
> -		fd = PTR_ERR(file);
> -		goto out;
> +		userfaultfd_ctx_put(ctx);
> +		return PTR_ERR(file);
>  	}
>  
> -	fd = get_unused_fd_flags(O_RDONLY | O_CLOEXEC);
> +	fd = get_unused_fd_flags(O_RDONLY);
>  	if (fd < 0) {
>  		fput(file);
> -		goto out;
> +		return fd;
>  	}
>  
>  	ctx->owner = file_inode(file);
>  	fd_install(fd, file);
> -
> -out:
> -	if (fd < 0) {
> -		mmdrop(ctx->mm);
> -		kmem_cache_free(userfaultfd_ctx_cachep, ctx);
> -	}
>  	return fd;

This introduces the opposite bug: now it's hardcoded to *not* use O_CLOEXEC,
instead of using the flag the user passed in the flags argument to the syscall.

- Eric