From patchwork Wed Oct 18 23:09:16 2017 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Long Li X-Patchwork-Id: 827849 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=none (mailfrom) smtp.mailfrom=vger.kernel.org (client-ip=209.132.180.67; helo=vger.kernel.org; envelope-from=linux-cifs-owner@vger.kernel.org; receiver=) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by ozlabs.org (Postfix) with ESMTP id 3yHSWl6bJxz9t44 for ; Thu, 19 Oct 2017 10:12:35 +1100 (AEDT) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751858AbdJRXMX (ORCPT ); Wed, 18 Oct 2017 19:12:23 -0400 Received: from a2nlsmtp01-04.prod.iad2.secureserver.net ([198.71.225.38]:49964 "EHLO a2nlsmtp01-04.prod.iad2.secureserver.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751572AbdJRXK6 (ORCPT ); Wed, 18 Oct 2017 19:10:58 -0400 Received: from linuxonhyperv.com ([107.180.71.197]) by : HOSTING RELAY : with SMTP id 4xT9evRY6hvnb4xT9e49ci; Wed, 18 Oct 2017 16:09:56 -0700 x-originating-ip: 107.180.71.197 Received: from longli by linuxonhyperv.com with local (Exim 4.89) (envelope-from ) id 1e4xT9-0005Vx-Bz; Wed, 18 Oct 2017 16:09:47 -0700 From: Long Li To: Steve French , linux-cifs@vger.kernel.org, samba-technical@lists.samba.org, linux-kernel@vger.kernel.org, linux-rdma@vger.kernel.org, Tom Talpey , Matthew Wilcox , Stephen Hemminger Cc: Long Li Subject: [Patch v5 17/21] CIFS: SMBD: Upper layer performs SMB write via RDMA read through memory registration Date: Wed, 18 Oct 2017 16:09:16 -0700 Message-Id: <20171018230920.21042-18-longli@exchange.microsoft.com> X-Mailer: git-send-email 2.14.1 In-Reply-To: <20171018230920.21042-1-longli@exchange.microsoft.com> References: <20171018230920.21042-1-longli@exchange.microsoft.com> X-CMAE-Envelope: MS4wfLQ2gA5yFHHfArk85NsyHLQ6jYIT2aVKDnXXwwCG68a8cmh8fdyfg+tSQt6QpbOOT598fQPuPPTxHB99dyI0xdekeMNvV+GVxJz4iJS14sw/cpX/M0Hc lVcWfLteMhugw7eowfy7sB2sNLwkpGtPkrBnP6Slys1ACBwsVrPzbweAqwjSrilYrypuX0VGpY9WcgSqzZeK939HIoMx+cjfUvEika8YvYcfYvlasAubGZO0 OKUlZMKFL5BV5VnlBD3ndNspMlYZC0UMbjvOJDZK3AZwIWvknOnu3vbC84tsIO0eY4vy78KzIwoxJmr59ZHvrU67lKC8CnV5oSznA+KtL14XcwvGYfmFt0rm FNMOdNnynH5Y4MBT7+rBOulj5vHyZo9Kqi7lA1RDvpN6l7FbvegstNa7LuM03JPhnqWeHSV8uxkSRPkAQuyt/3ivD1vngBXyWiu/NPP9mDih8Sq9ymEV9BiE aagguOylPw2Tz+/I Sender: linux-cifs-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-cifs@vger.kernel.org From: Long Li 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 by remote peer through local RDMA hardware. Modify the relevant fields in the packet accordingly, and append a smbd_buffer_descriptor_v1 to the end of the SMB write packet. On write I/O finish, deregister the memory region if this was for a RDMA read. If remote invalidation is not used, the call to smbd_deregister_mr will do local invalidation and possibly wait. Memory region is normally deregistered in MID callback as soon as it's used. There are situations where the MID may not be created on I/O failure, under which memory region is deregistered when write data context is released. Signed-off-by: Long Li --- fs/cifs/cifsglob.h | 3 +++ fs/cifs/cifssmb.c | 9 ++++++++ fs/cifs/smb2pdu.c | 68 ++++++++++++++++++++++++++++++++++++++++++++++++++---- 3 files changed, 76 insertions(+), 4 deletions(-) diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h index 09f9a71..2ae7d02 100644 --- a/fs/cifs/cifsglob.h +++ b/fs/cifs/cifsglob.h @@ -1170,6 +1170,9 @@ struct cifs_writedata { pid_t pid; unsigned int bytes; int result; +#ifdef CONFIG_CIFS_SMB_DIRECT + struct smbd_mr *mr; +#endif unsigned int pagesz; unsigned int tailsz; unsigned int credits; diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c index 5857009..08ff56a 100644 --- a/fs/cifs/cifssmb.c +++ b/fs/cifs/cifssmb.c @@ -43,6 +43,9 @@ #include "cifs_unicode.h" #include "cifs_debug.h" #include "fscache.h" +#ifdef CONFIG_CIFS_SMB_DIRECT +#include "smbdirect.h" +#endif #ifdef CONFIG_CIFS_POSIX static struct { @@ -1911,6 +1914,12 @@ cifs_writedata_release(struct kref *refcount) { struct cifs_writedata *wdata = container_of(refcount, struct cifs_writedata, refcount); +#ifdef CONFIG_CIFS_SMB_DIRECT + if (wdata->mr) { + smbd_deregister_mr(wdata->mr); + wdata->mr = NULL; + } +#endif if (wdata->cfile) cifsFileInfo_put(wdata->cfile); diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c index bab3da6..0841aba 100644 --- a/fs/cifs/smb2pdu.c +++ b/fs/cifs/smb2pdu.c @@ -48,6 +48,9 @@ #include "smb2glob.h" #include "cifspdu.h" #include "cifs_spnego.h" +#ifdef CONFIG_CIFS_SMB_DIRECT +#include "smbdirect.h" +#endif /* * The following table defines the expected "StructureSize" of SMB2 requests @@ -2652,7 +2655,19 @@ smb2_writev_callback(struct mid_q_entry *mid) wdata->result = -EIO; break; } - +#ifdef CONFIG_CIFS_SMB_DIRECT + /* + * If this wdata has a memory registered, the MR can be freed + * The number of MRs available is limited, it's important to recover + * used MR as soon as I/O is finished. Hold MR longer in the later + * I/O process can possibly result in I/O deadlock due to lack of MR + * to send request on I/O retry + */ + if (wdata->mr) { + smbd_deregister_mr(wdata->mr); + wdata->mr = NULL; + } +#endif if (wdata->result) cifs_stats_fail_inc(tcon, SMB2_WRITE_HE); @@ -2703,7 +2718,42 @@ smb2_async_writev(struct cifs_writedata *wdata, req->DataOffset = cpu_to_le16( offsetof(struct smb2_write_req, Buffer) - 4); req->RemainingBytes = 0; - +#ifdef CONFIG_CIFS_SMB_DIRECT + /* + * 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 = + (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; + } +#endif /* 4 for rfc1002 length field and 1 for Buffer */ iov[0].iov_len = 4; iov[0].iov_base = req; @@ -2716,12 +2766,22 @@ smb2_async_writev(struct cifs_writedata *wdata, rqst.rq_npages = wdata->nr_pages; rqst.rq_pagesz = wdata->pagesz; rqst.rq_tailsz = wdata->tailsz; - +#ifdef CONFIG_CIFS_SMB_DIRECT + if (wdata->mr) { + iov[1].iov_len += sizeof(struct smbd_buffer_descriptor_v1); + rqst.rq_npages = 0; + } +#endif cifs_dbg(FYI, "async write at %llu %u bytes\n", wdata->offset, wdata->bytes); +#ifdef CONFIG_CIFS_SMB_DIRECT + /* For RDMA read, I/O size is in RemainingBytes not in Length */ + if (!wdata->mr) + req->Length = cpu_to_le32(wdata->bytes); +#else req->Length = cpu_to_le32(wdata->bytes); - +#endif inc_rfc1001_len(&req->hdr, wdata->bytes - 1 /* Buffer */); if (wdata->credits) {