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;
Message-ID: <f9a48b3f0a6497014dc35ceb7835b5c543f48f59.camel@kernel.org>
Subject: Re: [PATCH v4 6/8] NFSD: Make nfsd4_setattr() wait before returning
 NFS4ERR_DELAY
From:   Jeff Layton <jlayton@kernel.org>
To:     Chuck Lever <chuck.lever@oracle.com>, linux-nfs@vger.kernel.org
Date:   Mon, 12 Sep 2022 12:20:37 -0400
In-Reply-To: <166267525333.1842.10708885965116635193.stgit@manet.1015granger.net>
References: <166267495153.1842.14474564029477470642.stgit@manet.1015granger.net>
         <166267525333.1842.10708885965116635193.stgit@manet.1015granger.net>
Content-Type: text/plain; charset="ISO-8859-15"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.44.4 (3.44.4-1.fc36) 
MIME-Version: 1.0
Precedence: bulk

On Thu, 2022-09-08 at 18:14 -0400, Chuck Lever wrote:
> nfsd_setattr() can kick off a CB_RECALL (via
> notify_change() -> break_lease()) if a delegation is present. Before
> returning NFS4ERR_DELAY, give the client holding that delegation a
> chance to return it and then retry the nfsd_setattr() again, once.
>=20
> Link: https://bugzilla.linux-nfs.org/show_bug.cgi?id=3D354
> Tested-by: Igor Mammedov <imammedo@redhat.com>
> Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
> ---
>  fs/nfsd/vfs.c |    9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
>=20
> diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
> index 02f31d8c727a..03a826ccc165 100644
> --- a/fs/nfsd/vfs.c
> +++ b/fs/nfsd/vfs.c
> @@ -394,6 +394,7 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *f=
hp,
>  	int		host_err;
>  	bool		get_write_count;
>  	bool		size_change =3D (iap->ia_valid & ATTR_SIZE);
> +	int		retries;
> =20
>  	if (iap->ia_valid & ATTR_SIZE) {
>  		accmode |=3D NFSD_MAY_WRITE|NFSD_MAY_OWNER_OVERRIDE;
> @@ -455,7 +456,13 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *=
fhp,
>  	}
> =20
>  	inode_lock(inode);
> -	host_err =3D __nfsd_setattr(dentry, iap);
> +	for (retries =3D 1;;) {
> +		host_err =3D __nfsd_setattr(dentry, iap);
> +		if (host_err !=3D -EAGAIN || !retries--)
> +			break;

Can __nfsd_setattr return -EAGAIN for entirely different reasons? I
try_break_deleg will, but could an underlying filesystem's ->setattr
operation return -EAGAIN for an unrelated reason?

Then again, you're only retrying once, so maybe it's no big deal.

> +		if (!nfsd_wait_for_delegreturn(rqstp, inode))
> +			break;
> +	}
>  	if (attr->na_seclabel && attr->na_seclabel->len)
>  		attr->na_labelerr =3D security_inode_setsecctx(dentry,
>  			attr->na_seclabel->data, attr->na_seclabel->len);
>=20
>=20

--=20
Jeff Layton <jlayton@kernel.org>