From: Tom Talpey <ttalpey@microsoft.com>
To: Long Li <longli@microsoft.com>, Steve French <sfrench@samba.org>,
        "linux-cifs@vger.kernel.org" <linux-cifs@vger.kernel.org>,
        "samba-technical@lists.samba.org" <samba-technical@lists.samba.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "linux-rdma@vger.kernel.org" <linux-rdma@vger.kernel.org>
Subject: RE: [[PATCH v1] 18/37] [CIFS] SMBD: Implement API for upper layer to
 send data
Thread-Topic: [[PATCH v1] 18/37] [CIFS] SMBD: Implement API for upper layer to
 send data
Thread-Index: AQHTC8xzKNs4k0d9aEewq2gj+iYcu6KEYFHw
Date: Mon, 14 Aug 2017 20:44:18 +0000
Message-ID: <CY4PR21MB01827F735F570CCAECD5AD8BA08C0@CY4PR21MB0182.namprd21.prod.outlook.com>
References: <1501704648-20159-1-git-send-email-longli@exchange.microsoft.com>
 <1501704648-20159-19-git-send-email-longli@exchange.microsoft.com>
In-Reply-To: <1501704648-20159-19-git-send-email-longli@exchange.microsoft.com>
Accept-Language: en-US
Content-Language: en-US
spamdiagnosticoutput: 1:99
spamdiagnosticmetadata: NSPM
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
X-MS-Exchange-CrossTenant-originalarrivaltime: 14 Aug 2017 20:44:18.6423
 (UTC)
X-MS-Exchange-CrossTenant-fromentityheader: Hosted
X-MS-Exchange-CrossTenant-id: 72f988bf-86f1-41af-91ab-2d7cd011db47
X-MS-Exchange-Transport-CrossTenantHeadersStamped: CY4PR21MB0629
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by nfs id v7EKihOn031810
Content-Length: 1958
Lines: 48

> -----Original Message-----
> From: linux-cifs-owner@vger.kernel.org [mailto:linux-cifs-
> owner@vger.kernel.org] On Behalf Of Long Li
> Sent: Wednesday, August 2, 2017 4:10 PM
> To: Steve French <sfrench@samba.org>; linux-cifs@vger.kernel.org; samba-
> technical@lists.samba.org; linux-kernel@vger.kernel.org
> Cc: Long Li <longli@microsoft.com>
> Subject: [[PATCH v1] 18/37] [CIFS] SMBD: Implement API for upper layer to
> send data
> 
> +/*
> + * Write data to transport
> + * Each rqst is transported as a SMBDirect payload
> + * rqst: the data to write
> + * return value: 0 if successfully write, otherwise error code
> + */
> +int cifs_rdma_write(struct cifs_rdma_info *info, struct smb_rqst *rqst)
> +{

!!!
This is a VERY confusing name. It is not sending an RDMA Write, which will
confuse any RDMA-enlightened reader. It's performing an RDMA Send, so
that name is perhaps one possibility. 

> +       if (info->transport_status != CIFS_RDMA_CONNECTED) {
> +               log_cifs_write("disconnected returning -EIO\n");
> +               return -EIO;
> +       }

Isn't this optimizing the error case? There's no guarantee it's still connected by
the time the following request construction occurs. Why not just proceed without
the check? 

> +       /* Strip the first 4 bytes MS-SMB2 section 2.1
> +        * they are used only for TCP transport */
> +       iov[0].iov_base = (char*)rqst->rq_iov[0].iov_base + 4;
> +       iov[0].iov_len = rqst->rq_iov[0].iov_len - 4;
> +       buflen += iov[0].iov_len;

Ok, that layering choice in the cifs.ko client code needs to be corrected. After all,
it will need to be RDMA-aware to build the SMB3 read/write channel structures.
And, the code in cifs_post_send_data() is allocating and building a structure that
could have been accounted for much earlier, avoiding the extra overhead.

That change could happen later, the hack is mostly ok for now. But something
needs to be said in a comment.
 
Tom.