Received-SPF: pass (google.com: domain of linux-nfs-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
From:   "NeilBrown" <neilb@suse.de>
To:     "Jeff Layton" <jlayton@kernel.org>
Cc:     "Chuck Lever" <chuck.lever@oracle.com>,
        "Olga Kornievskaia" <kolga@netapp.com>,
        "Dai Ngo" <Dai.Ngo@oracle.com>, "Tom Talpey" <tom@talpey.com>,
        linux-nfs@vger.kernel.org, linux-kernel@vger.kernel.org,
        "Jeff Layton" <jlayton@kernel.org>
Subject: Re: [PATCH RFC] nfsd: don't hand out write delegations on O_WRONLY opens
In-reply-to: <20230731-wdeleg-v1-1-f8fe1ce11b36@kernel.org>
References: <20230731-wdeleg-v1-1-f8fe1ce11b36@kernel.org>
Date:   Tue, 01 Aug 2023 08:11:18 +1000
Message-id: <169084147821.32308.9286837678268595107@noble.neil.brown.name>
Precedence: bulk

On Tue, 01 Aug 2023, Jeff Layton wrote:
> I noticed that xfstests generic/001 was failing against linux-next nfsd.
>=20
> The client would request a OPEN4_SHARE_ACCESS_WRITE open, and the server
> would hand out a write delegation. The client would then try to use that
> write delegation as the source stateid in a COPY or CLONE operation, and
> the server would respond with NFS4ERR_STALE.
>=20
> The problem is that the struct file associated with the delegation does
> not necessarily have read permissions. It's handing out a write
> delegation on what is effectively an O_WRONLY open. RFC 8881 states:
>=20
>  "An OPEN_DELEGATE_WRITE delegation allows the client to handle, on its
>   own, all opens."
>=20
> Given that the client didn't request any read permissions, and that nfsd
> didn't check for any, it seems wrong to give out a write delegation.
>=20
> Don't hand out a delegation if the client didn't request
> OPEN4_SHARE_ACCESS_BOTH.
>=20
> This fixes xfstest generic/001.
>=20
> Closes: https://bugzilla.linux-nfs.org/show_bug.cgi?id=3D412
> Signed-off-by: Jeff Layton <jlayton@kernel.org>
> ---
>  fs/nfsd/nfs4state.c | 2 ++
>  1 file changed, 2 insertions(+)
>=20
> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> index ef7118ebee00..9f1c90afed72 100644
> --- a/fs/nfsd/nfs4state.c
> +++ b/fs/nfsd/nfs4state.c
> @@ -5462,6 +5462,8 @@ nfs4_set_delegation(struct nfsd4_open *open, struct n=
fs4_ol_stateid *stp,
>  		return ERR_PTR(-EAGAIN);
> =20
>  	if (open->op_share_access & NFS4_SHARE_ACCESS_WRITE) {
> +		if (!(open->op_share_access & NFS4_SHARE_ACCESS_READ))
> +			return ERR_PTR(-EBADF);
<bikeshed>
The actual error code returned by nfs4_set_delegation() is ignored -
only the fact of an error is relevant.
Given that, how did you choose -EBADF.  nfsd doesn't use file
descriptors, and doesn't use EBADF anywhere else.
Given that you have just tested access, EACCES might be justifiable.
But I would prefer if nfs4_set_delegation() returns NULL if it could not
find or create a delegation, without bothering with giving a reason.
</bikeshed>

Reviewed-by: NeilBrown <neilb@suse.de>

NeilBrown

>  		nf =3D find_writeable_file(fp);
>  		dl_type =3D NFS4_OPEN_DELEGATE_WRITE;
>  	} else {
>=20
> ---
> base-commit: ec89391563792edd11d138a853901bce76d11f44
> change-id: 20230731-wdeleg-bbdb6b25a3c6
>=20
> Best regards,
> --=20
> Jeff Layton <jlayton@kernel.org>
>=20
>=20