public inbox for linux-rdma@vger.kernel.org
 help / color / mirror / Atom feed
From: Leon Romanovsky <leon@kernel.org>
To: Long Li <longli@exchange.microsoft.com>
Cc: Steve French <sfrench@samba.org>,
	linux-cifs@vger.kernel.org, samba-technical@lists.samba.org,
	linux-kernel@vger.kernel.org, linux-rdma@vger.kernel.org,
	Christoph Hellwig <hch@infradead.org>,
	Tom Talpey <ttalpey@microsoft.com>,
	Matthew Wilcox <mawilcox@microsoft.com>,
	Long Li <longli@microsoft.com>
Subject: Re: [Patch v2 13/19] CIFS: SMBD: Use registered memory RDMA read for SMB write
Date: Wed, 23 Aug 2017 16:52:00 +0300	[thread overview]
Message-ID: <20170823135200.GP1724@mtr-leonro.local> (raw)
In-Reply-To: <1503255883-3041-14-git-send-email-longli@exchange.microsoft.com>

[-- Attachment #1: Type: text/plain, Size: 3111 bytes --]

On Sun, Aug 20, 2017 at 12:04:37PM -0700, Long Li wrote:
> From: Long Li <longli@microsoft.com>
>
> When sending I/O, if size is larger than rdma_readwrite_threshold we prepare to send SMB WRITE packet for a RDMA read via memory registration. The actual I/O is done out-of-the-band, so modify the relevant fields in the packet accordingly.
>
> Signed-off-by: Long Li <longli@microsoft.com>
> ---
>  fs/cifs/smb2pdu.c | 45 ++++++++++++++++++++++++++++++++++++++++++++-
>  1 file changed, 44 insertions(+), 1 deletion(-)
>
> diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c
> index 5cc5f6c..5581afd 100644
> --- a/fs/cifs/smb2pdu.c
> +++ b/fs/cifs/smb2pdu.c
> @@ -48,6 +48,7 @@
>  #include "smb2glob.h"
>  #include "cifspdu.h"
>  #include "cifs_spnego.h"
> +#include "smbdirect.h"
>
>  /*
>   *  The following table defines the expected "StructureSize" of SMB2 requests
> @@ -2716,6 +2717,41 @@ smb2_async_writev(struct cifs_writedata *wdata,
>  				offsetof(struct smb2_write_req, Buffer) - 4);
>  	req->RemainingBytes = 0;
>
> +	/*
> +	 * If we want to do a server RDMA read, fill in and append
> +	 * smbd_buffer_descriptor_v1 to the end of write request
> +	 */
> +	if (server->rdma && wdata->bytes >
> +		server->smbd_conn->rdma_readwrite_threshold) {
> +
> +		struct smbd_buffer_descriptor_v1 *v1;
> +		bool need_invalidate = server->dialect == SMB30_PROT_ID;
> +
> +		wdata->mr = smbd_register_mr(
> +				server->smbd_conn, wdata->pages,
> +				wdata->nr_pages, wdata->tailsz,
> +				false, need_invalidate);
> +		if (!wdata->mr) {
> +			rc = -ENOBUFS;
> +			goto async_writev_out;
> +		}
> +		req->Length = 0;
> +		req->DataOffset = 0;
> +		req->RemainingBytes =

Wow, we have CamelCase variables in linux kernel. It will help if you
start your patchset with small cleanup to convert those variables from
CamelCase to normal names.

Thanks

> +			(wdata->nr_pages-1)*PAGE_SIZE + wdata->tailsz;
> +		req->Channel = SMB2_CHANNEL_RDMA_V1_INVALIDATE;
> +		if (need_invalidate)
> +			req->Channel = SMB2_CHANNEL_RDMA_V1;
> +		req->WriteChannelInfoOffset =
> +			offsetof(struct smb2_write_req, Buffer) - 4;
> +		req->WriteChannelInfoLength =
> +			sizeof(struct smbd_buffer_descriptor_v1);
> +		v1 = (struct smbd_buffer_descriptor_v1 *) &req->Buffer[0];
> +		v1->offset = wdata->mr->mr->iova;
> +		v1->token = wdata->mr->mr->rkey;
> +		v1->length = wdata->mr->mr->length;
> +	}
> +
>  	/* 4 for rfc1002 length field and 1 for Buffer */
>  	iov[0].iov_len = 4;
>  	iov[0].iov_base = req;
> @@ -2729,10 +2765,17 @@ smb2_async_writev(struct cifs_writedata *wdata,
>  	rqst.rq_pagesz = wdata->pagesz;
>  	rqst.rq_tailsz = wdata->tailsz;
>
> +	if (wdata->mr) {
> +		iov[1].iov_len += sizeof(struct smbd_buffer_descriptor_v1);
> +		rqst.rq_npages = 0;
> +	}
> +
>  	cifs_dbg(FYI, "async write at %llu %u bytes\n",
>  		 wdata->offset, wdata->bytes);
>
> -	req->Length = cpu_to_le32(wdata->bytes);
> +	/* For RDMA read, I/O size is in RemainingBytes not in Length */
> +	if (!wdata->mr)
> +		req->Length = cpu_to_le32(wdata->bytes);
>
>  	inc_rfc1001_len(&req->hdr, wdata->bytes - 1 /* Buffer */);
>
> --
> 2.7.4
>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

  reply	other threads:[~2017-08-23 13:52 UTC|newest]

Thread overview: 38+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-08-20 19:04 [Patch v2 00/19] CIFS: Implement SMBDirect Long Li
     [not found] ` <1503255883-3041-1-git-send-email-longli-Lp/cVzEoVyZiJJESP9tAQJZ3qXmFLfmx@public.gmane.org>
2017-08-20 19:04   ` [Patch v2 01/19] CIFS: Add RDMA mount option Long Li
     [not found]     ` <1503255883-3041-2-git-send-email-longli-Lp/cVzEoVyZiJJESP9tAQJZ3qXmFLfmx@public.gmane.org>
2017-08-21  4:36       ` Leon Romanovsky
2017-08-21 18:18         ` Long Li
2017-08-20 19:04   ` [Patch v2 02/19] CIFS: SMBD: Add SMBDirect protocol and transport constants Long Li
2017-08-21 19:15   ` [Patch v2 00/19] CIFS: Implement SMBDirect Steve Wise
2017-08-21 19:50     ` Long Li
2017-08-21 19:56       ` Steve Wise
2017-08-21 20:23         ` Long Li
2017-08-29 18:10           ` Long Li
     [not found]             ` <MWHPR21MB0190050F3699CDF6F52B51A2CE9F0-saRRjQKJ25M/hL2NnenhuM1VXTxX1y3OvxpqHgZTriW3zl9H0oFU5g@public.gmane.org>
2017-09-05 16:30               ` Steve Wise
2017-09-05 17:42                 ` Long Li
2017-08-29 18:20   ` Roland Dreier
     [not found]     ` <CAL1RGDUGiOqjB8n0mJmF079jm7vXrdzE2rHdzoNOexvgHScjUQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-08-29 19:31       ` Long Li
     [not found]         ` <CY4PR21MB01821B704C2F64F72A8A92A9CE9F0-kUhI0YP1syo7ifcEnHlXec1VXTxX1y3OvxpqHgZTriW3zl9H0oFU5g@public.gmane.org>
2017-08-30  5:16           ` Stefan Metzmacher
2017-08-20 19:04 ` [Patch v2 03/19] CIFS: SMBD: " Long Li
2017-08-20 19:04 ` [Patch v2 04/19] CIFS: SMBD: Add SMBDirect transport to SMB connection and Makefile Long Li
2017-08-20 19:04 ` [Patch v2 05/19] CIFS: SMBD: Connect to SMBDirect session Long Li
2017-08-20 19:04 ` [Patch v2 06/19] CIFS: SMBD: Reconnect " Long Li
2017-08-20 19:04 ` [Patch v2 07/19] CIFS: SMBD: Destroy SMBDirect session on shutdown or umount Long Li
2017-08-20 19:04 ` [Patch v2 08/19] CIFS: SMBD: Set SMBDirect maximum read or write size for I/O Long Li
2017-08-20 19:04 ` [Patch v2 09/19] CIFS: SMBD: Read data from SMBDirect Long Li
2017-08-20 19:04 ` [Patch v2 10/19] CIFS: SMBD: Send data through SMBDirect Long Li
2017-08-20 19:04 ` [Patch v2 11/19] CIFS: SMBD: Define memory registration for I/O data Long Li
2017-08-20 19:04 ` [Patch v2 12/19] CIFS: SMBD: Fix the definition for SMB2_CHANNEL_RDMA_V1_INVALIDATE Long Li
2017-08-20 19:04 ` [Patch v2 13/19] CIFS: SMBD: Use registered memory RDMA read for SMB write Long Li
2017-08-23 13:52   ` Leon Romanovsky [this message]
     [not found]     ` <20170823135200.GP1724-U/DQcQFIOTAAJjI8aNfphQ@public.gmane.org>
2017-08-23 18:09       ` Long Li
     [not found]         ` <MWHPR21MB0190DBBDE3317D973FEF0E43CE850-saRRjQKJ25M/hL2NnenhuM1VXTxX1y3OvxpqHgZTriW3zl9H0oFU5g@public.gmane.org>
2017-08-23 19:02           ` Leon Romanovsky
     [not found]             ` <20170823190214.GY1724-U/DQcQFIOTAAJjI8aNfphQ@public.gmane.org>
2017-08-23 19:10               ` Long Li
     [not found]                 ` <MWHPR21MB0190DDCDB0B5F8FED3265A60CE850-saRRjQKJ25M/hL2NnenhuM1VXTxX1y3OvxpqHgZTriW3zl9H0oFU5g@public.gmane.org>
2017-08-23 19:23                   ` Leon Romanovsky
2017-08-23 19:39                 ` Steve French
2017-08-20 19:04 ` [Patch v2 14/19] CIFS: SMBD: Deregister memory when finishing " Long Li
2017-08-20 19:04 ` [Patch v2 15/19] CIFS: SMBD: Add parameter rdata to smb2_new_read_req Long Li
2017-08-20 19:04 ` [Patch v2 16/19] CIFS: SMBD: Read correct returned data length for RDMA write (SMB READ) I/O Long Li
2017-08-20 19:04 ` [Patch v2 17/19] CIFS: SMBD: Implement SMB READ via RDMA write through memory registration Long Li
2017-08-20 19:04 ` [Patch v2 18/19] CIFS: SMBD: Deregister memory when finishing SMB READ Long Li
2017-08-20 19:04 ` [Patch v2 19/19] CIFS: SMBD: Add SMBDirect debug counters Long Li

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20170823135200.GP1724@mtr-leonro.local \
    --to=leon@kernel.org \
    --cc=hch@infradead.org \
    --cc=linux-cifs@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-rdma@vger.kernel.org \
    --cc=longli@exchange.microsoft.com \
    --cc=longli@microsoft.com \
    --cc=mawilcox@microsoft.com \
    --cc=samba-technical@lists.samba.org \
    --cc=sfrench@samba.org \
    --cc=ttalpey@microsoft.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox