Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Thu, 4 Oct 2018 15:23:13 -0700
From:   Alexei Starovoitov <alexei.starovoitov@gmail.com>
To:     Andy Lutomirski <luto@amacapital.net>
Cc:     Roman Gushchin <guro@fb.com>, Alexei Starovoitov <ast@kernel.org>,
        "David S. Miller" <davem@davemloft.net>,
        Daniel Borkmann <daniel@iogearbox.net>,
        Al Viro <viro@zeniv.linux.org.uk>,
        Network Development <netdev@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>,
        kernel-team <Kernel-team@fb.com>
Subject: Re: [PATCH bpf-next 1/6] bpf: introduce BPF_PROG_TYPE_FILE_FILTER
Message-ID: <20181004222312.3vvyvibx6dc4vqwt@ast-mbp.dhcp.thefacebook.com>
References: <20181004025750.498303-1-ast@kernel.org>
 <20181004025750.498303-2-ast@kernel.org>
 <20181004194123.GA12697@castle.DHCP.thefacebook.com>
 <CALCETrUp69Y01v628KydCNxEqzGHSgs6dQQJxrzVAgs7DhPMxA@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CALCETrUp69Y01v628KydCNxEqzGHSgs6dQQJxrzVAgs7DhPMxA@mail.gmail.com>
User-Agent: NeoMutt/20180223
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Oct 04, 2018 at 12:51:00PM -0700, Andy Lutomirski wrote:
> On Thu, Oct 4, 2018 at 12:41 PM Roman Gushchin <guro@fb.com> wrote:
> >
> > On Wed, Oct 03, 2018 at 07:57:45PM -0700, Alexei Starovoitov wrote:
> > > Similar to networking sandboxing programs and cgroup-v2 based hooks
> > > (BPF_CGROUP_INET_[INGRESS|EGRESS,] BPF_CGROUP_INET[4|6]_[BIND|CONNECT], etc)
> > > introduce basic per-container sandboxing for file access via
> > > new BPF_PROG_TYPE_FILE_FILTER program type that attaches after
> > > security_file_open() LSM hook and works as additional file_open filter.
> > > The new cgroup bpf hook is called BPF_CGROUP_FILE_OPEN.
> > >
> > > Just like other cgroup-bpf programs new BPF_PROG_TYPE_FILE_FILTER type
> > > is only available to root.
> > >
> > > This program type has access to single argument 'struct bpf_file_info'
> > > that contains standard sys_stat fields:
> > > struct bpf_file_info {
> > >         __u64 inode;
> > >         __u32 dev_major;
> > >         __u32 dev_minor;
> > >         __u32 fs_magic;
> > >         __u32 mnt_id;
> > >         __u32 nlink;
> > >         __u32 mode;     /* file mode S_ISDIR, S_ISLNK, 0755, etc */
> > >         __u32 flags;    /* open flags O_RDWR, O_CREAT, etc */
> > > };
> >
> > It's probably nice to have file uid/gid as well.
> 
> And an indication of which mount namespace we're looking at.

Both certainly can be added in the future without breaking progs.
I didn't want to add too much all at once.
For file uid/gid I prototyped 
bpf_get_statx(struct bpf_file_info *file, int flags, int mask, struct statx *sx, int size);
helper that calls normal statx underneath.
But it's not fast, since sizeof(struct statx) == 256 and it has to be fully
inited by the helper or by the bpf prog (since bpf doesn't allow uninited memory anywhere).
Then I thought about going back to older sys_stat helper (without x), since structs
are smaller, but that didn't look as good either. So I've decied to table it for now
and get the basic support via 'struct bpf_file_info' first.
Then extend it later via new fields and new helpers.