From patchwork Tue Oct 6 14:59:56 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Chuck Lever X-Patchwork-Id: 7335641 Return-Path: X-Original-To: patchwork-linux-rdma@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork1.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork1.web.kernel.org (Postfix) with ESMTP id 8C4209F1B9 for ; Tue, 6 Oct 2015 15:00:03 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 1EC9C20654 for ; Tue, 6 Oct 2015 15:00:02 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 7D1152064E for ; Tue, 6 Oct 2015 15:00:00 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752933AbbJFO77 (ORCPT ); Tue, 6 Oct 2015 10:59:59 -0400 Received: from mail-ig0-f179.google.com ([209.85.213.179]:37793 "EHLO mail-ig0-f179.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751906AbbJFO76 (ORCPT ); Tue, 6 Oct 2015 10:59:58 -0400 Received: by igcpe7 with SMTP id pe7so26282218igc.0; Tue, 06 Oct 2015 07:59:58 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=sender:subject:from:to:date:message-id:in-reply-to:references :user-agent:mime-version:content-type:content-transfer-encoding; bh=EgynzQqRTHdwM4goQXy4YXV4LSjKxgD6qISJyIhxSeo=; b=bMr4bZskiQeuwPpTt9dZzZOSfpMWVGpHlMmvSBgNZlQhuB0RfDXF6chDGuh4LKD/u/ qTqTD4Y5SQN7hyRrYe3dYuvQ02lwp7o4R2sd16nfTn9MGKCU4H23542Da9b8u+8a2axm j+/14rBWyVlZUBpxxvb79IaK3vHZB17glL02jdEukVRFZMzWFiTyqfH94bGbSXBE+ouA BodqM7+Y/bQdcJ2ArSBsFPaLuvL97dUQrMw0rt5ASfJ4fzOf26d6qSNS12bD+bEvUqKq 9TRiePDtMwGNN2NE/z1dM8Jgr13S98E185w8oqXiOzJ0dtdu54YzCe8hmXvb+94NZsgh JBlQ== X-Received: by 10.50.131.138 with SMTP id om10mr16877964igb.47.1444143597986; Tue, 06 Oct 2015 07:59:57 -0700 (PDT) Received: from manet.1015granger.net ([2604:8800:100:81fc:82ee:73ff:fe43:d64f]) by smtp.gmail.com with ESMTPSA id g2sm6933440igi.15.2015.10.06.07.59.57 (version=TLSv1.2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Tue, 06 Oct 2015 07:59:57 -0700 (PDT) Subject: [PATCH v2 10/16] xprtrdma: Pre-allocate backward rpc_rqst and send/receive buffers From: Chuck Lever To: linux-rdma@vger.kernel.org, linux-nfs@vger.kernel.org Date: Tue, 06 Oct 2015 10:59:56 -0400 Message-ID: <20151006145956.11788.29903.stgit@manet.1015granger.net> In-Reply-To: <20151006142430.11788.42604.stgit@manet.1015granger.net> References: <20151006142430.11788.42604.stgit@manet.1015granger.net> User-Agent: StGit/0.17.1-3-g7d0f MIME-Version: 1.0 Sender: linux-rdma-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-rdma@vger.kernel.org X-Spam-Status: No, score=-6.8 required=5.0 tests=BAYES_00,DKIM_SIGNED, RCVD_IN_DNSWL_HI,T_DKIM_INVALID,T_RP_MATCHES_RCVD,UNPARSEABLE_RELAY autolearn=ham version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP xprtrdma's backward direction send and receive buffers are the same size as the forechannel's inline threshold, and must be pre- registered. The consumer has no control over which receive buffer the adapter chooses to catch an incoming backwards-direction call. Any receive buffer can be used for either a forward reply or a backward call. Thus both types of RPC message must all be the same size. Signed-off-by: Chuck Lever --- net/sunrpc/xprtrdma/Makefile | 1 net/sunrpc/xprtrdma/backchannel.c | 206 +++++++++++++++++++++++++++++++++++++ net/sunrpc/xprtrdma/transport.c | 7 + net/sunrpc/xprtrdma/verbs.c | 87 ++++++++++++++-- net/sunrpc/xprtrdma/xprt_rdma.h | 20 ++++ 5 files changed, 309 insertions(+), 12 deletions(-) create mode 100644 net/sunrpc/xprtrdma/backchannel.c -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html diff --git a/net/sunrpc/xprtrdma/Makefile b/net/sunrpc/xprtrdma/Makefile index 48913de..33f99d3 100644 --- a/net/sunrpc/xprtrdma/Makefile +++ b/net/sunrpc/xprtrdma/Makefile @@ -5,3 +5,4 @@ rpcrdma-y := transport.o rpc_rdma.o verbs.o \ svc_rdma.o svc_rdma_transport.o \ svc_rdma_marshal.o svc_rdma_sendto.o svc_rdma_recvfrom.o \ module.o +rpcrdma-$(CONFIG_SUNRPC_BACKCHANNEL) += backchannel.o diff --git a/net/sunrpc/xprtrdma/backchannel.c b/net/sunrpc/xprtrdma/backchannel.c new file mode 100644 index 0000000..3d01b32 --- /dev/null +++ b/net/sunrpc/xprtrdma/backchannel.c @@ -0,0 +1,206 @@ +/* + * Copyright (c) 2015 Oracle. All rights reserved. + * + * Support for backward direction RPCs on RPC/RDMA. + */ + +#include + +#include "xprt_rdma.h" + +#if IS_ENABLED(CONFIG_SUNRPC_DEBUG) +# define RPCDBG_FACILITY RPCDBG_TRANS +#endif + +static void rpcrdma_bc_free_rqst(struct rpcrdma_xprt *r_xprt, + struct rpc_rqst *rqst) +{ + struct rpcrdma_buffer *buf = &r_xprt->rx_buf; + struct rpcrdma_req *req = rpcr_to_rdmar(rqst); + + spin_lock(&buf->rb_reqslock); + list_del(&req->rl_all); + spin_unlock(&buf->rb_reqslock); + + rpcrdma_destroy_req(&r_xprt->rx_ia, req); + + kfree(rqst); +} + +static int rpcrdma_bc_setup_rqst(struct rpcrdma_xprt *r_xprt, + struct rpc_rqst *rqst) +{ + struct rpcrdma_ia *ia = &r_xprt->rx_ia; + struct rpcrdma_regbuf *rb; + struct rpcrdma_req *req; + struct xdr_buf *buf; + size_t size; + + req = rpcrdma_create_req(r_xprt); + if (!req) + return -ENOMEM; + req->rl_backchannel = true; + + size = RPCRDMA_INLINE_WRITE_THRESHOLD(rqst); + rb = rpcrdma_alloc_regbuf(ia, size, GFP_KERNEL); + if (IS_ERR(rb)) + goto out_fail; + req->rl_rdmabuf = rb; + + size += RPCRDMA_INLINE_READ_THRESHOLD(rqst); + rb = rpcrdma_alloc_regbuf(ia, size, GFP_KERNEL); + if (IS_ERR(rb)) + goto out_fail; + rb->rg_owner = req; + req->rl_sendbuf = rb; + /* so that rpcr_to_rdmar works when receiving a request */ + rqst->rq_buffer = (void *)req->rl_sendbuf->rg_base; + + buf = &rqst->rq_snd_buf; + buf->head[0].iov_base = rqst->rq_buffer; + buf->head[0].iov_len = 0; + buf->tail[0].iov_base = NULL; + buf->tail[0].iov_len = 0; + buf->page_len = 0; + buf->len = 0; + buf->buflen = size; + + return 0; + +out_fail: + rpcrdma_bc_free_rqst(r_xprt, rqst); + return -ENOMEM; +} + +/* Allocate and add receive buffers to the rpcrdma_buffer's + * existing list of rep's. These are released when the + * transport is destroyed. + */ +static int rpcrdma_bc_setup_reps(struct rpcrdma_xprt *r_xprt, + unsigned int count) +{ + struct rpcrdma_buffer *buffers = &r_xprt->rx_buf; + struct rpcrdma_rep *rep; + unsigned long flags; + int rc = 0; + + while (count--) { + rep = rpcrdma_create_rep(r_xprt); + if (IS_ERR(rep)) { + pr_err("RPC: %s: reply buffer alloc failed\n", + __func__); + rc = PTR_ERR(rep); + break; + } + + spin_lock_irqsave(&buffers->rb_lock, flags); + list_add(&rep->rr_list, &buffers->rb_recv_bufs); + spin_unlock_irqrestore(&buffers->rb_lock, flags); + } + + return rc; +} + +/** + * xprt_rdma_bc_setup - Pre-allocate resources for handling backchannel requests + * @xprt: transport associated with these backchannel resources + * @reqs: number of concurrent incoming requests to expect + * + * Returns 0 on success; otherwise a negative errno + */ +int xprt_rdma_bc_setup(struct rpc_xprt *xprt, unsigned int reqs) +{ + struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt); + struct rpcrdma_buffer *buffer = &r_xprt->rx_buf; + struct rpc_rqst *rqst; + unsigned int i; + int rc; + + /* The backchannel reply path returns each rpc_rqst to the + * bc_pa_list _after_ the reply is sent. If the server is + * faster than the client, it can send another backward + * direction request before the rpc_rqst is returned to the + * list. The client rejects the request in this case. + * + * Twice as many rpc_rqsts are prepared to ensure there is + * always an rpc_rqst available as soon as a reply is sent. + */ + for (i = 0; i < (reqs << 1); i++) { + rqst = kzalloc(sizeof(*rqst), GFP_KERNEL); + if (!rqst) { + pr_err("RPC: %s: Failed to create bc rpc_rqst\n", + __func__); + goto out_free; + } + + rqst->rq_xprt = &r_xprt->rx_xprt; + INIT_LIST_HEAD(&rqst->rq_list); + INIT_LIST_HEAD(&rqst->rq_bc_list); + + if (rpcrdma_bc_setup_rqst(r_xprt, rqst)) + goto out_free; + + spin_lock_bh(&xprt->bc_pa_lock); + list_add(&rqst->rq_bc_pa_list, &xprt->bc_pa_list); + spin_unlock_bh(&xprt->bc_pa_lock); + } + + rc = rpcrdma_bc_setup_reps(r_xprt, reqs); + if (rc) + goto out_free; + + rc = rpcrdma_ep_post_extra_recv(r_xprt, reqs); + if (rc) + goto out_free; + + buffer->rb_bc_srv_max_requests = reqs; + request_module("svcrdma"); + + return 0; + +out_free: + xprt_rdma_bc_destroy(xprt, reqs); + + pr_err("RPC: %s: setup backchannel transport failed\n", __func__); + return -ENOMEM; +} + +/** + * xprt_rdma_bc_destroy - Release resources for handling backchannel requests + * @xprt: transport associated with these backchannel resources + * @reqs: number of incoming requests to destroy; ignored + */ +void xprt_rdma_bc_destroy(struct rpc_xprt *xprt, unsigned int reqs) +{ + struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt); + struct rpc_rqst *rqst, *tmp; + + spin_lock_bh(&xprt->bc_pa_lock); + list_for_each_entry_safe(rqst, tmp, &xprt->bc_pa_list, rq_bc_pa_list) { + list_del(&rqst->rq_bc_pa_list); + spin_unlock_bh(&xprt->bc_pa_lock); + + rpcrdma_bc_free_rqst(r_xprt, rqst); + + spin_lock_bh(&xprt->bc_pa_lock); + } + spin_unlock_bh(&xprt->bc_pa_lock); +} + +/** + * xprt_rdma_bc_free_rqst - Release a backchannel rqst + * @rqst: request to release + */ +void xprt_rdma_bc_free_rqst(struct rpc_rqst *rqst) +{ + struct rpc_xprt *xprt = rqst->rq_xprt; + + smp_mb__before_atomic(); + WARN_ON_ONCE(!test_bit(RPC_BC_PA_IN_USE, &rqst->rq_bc_pa_state)); + clear_bit(RPC_BC_PA_IN_USE, &rqst->rq_bc_pa_state); + smp_mb__after_atomic(); + + spin_lock_bh(&xprt->bc_pa_lock); + list_add_tail(&rqst->rq_bc_pa_list, &xprt->bc_pa_list); + spin_unlock_bh(&xprt->bc_pa_lock); +} diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c index 897a2f3..845278e 100644 --- a/net/sunrpc/xprtrdma/transport.c +++ b/net/sunrpc/xprtrdma/transport.c @@ -705,7 +705,12 @@ static struct rpc_xprt_ops xprt_rdma_procs = { .print_stats = xprt_rdma_print_stats, .enable_swap = xprt_rdma_enable_swap, .disable_swap = xprt_rdma_disable_swap, - .inject_disconnect = xprt_rdma_inject_disconnect + .inject_disconnect = xprt_rdma_inject_disconnect, +#if defined(CONFIG_SUNRPC_BACKCHANNEL) + .bc_setup = xprt_rdma_bc_setup, + .bc_free_rqst = xprt_rdma_bc_free_rqst, + .bc_destroy = xprt_rdma_bc_destroy, +#endif }; static struct xprt_class xprt_rdma = { diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c index 2dca18d..d920367 100644 --- a/net/sunrpc/xprtrdma/verbs.c +++ b/net/sunrpc/xprtrdma/verbs.c @@ -833,7 +833,21 @@ retry: } rc = ep->rep_connected; } else { + struct rpcrdma_xprt *r_xprt; + unsigned int extras; + dprintk("RPC: %s: connected\n", __func__); + + r_xprt = container_of(ia, struct rpcrdma_xprt, rx_ia); + extras = r_xprt->rx_buf.rb_bc_srv_max_requests; + + if (extras) { + rc = rpcrdma_ep_post_extra_recv(r_xprt, extras); + if (rc) + pr_warn("%s: rpcrdma_ep_post_extra_recv: %i\n", + __func__, rc); + rc = 0; + } } out: @@ -870,20 +884,25 @@ rpcrdma_ep_disconnect(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia) } } -static struct rpcrdma_req * +struct rpcrdma_req * rpcrdma_create_req(struct rpcrdma_xprt *r_xprt) { + struct rpcrdma_buffer *buffer = &r_xprt->rx_buf; struct rpcrdma_req *req; req = kzalloc(sizeof(*req), GFP_KERNEL); if (req == NULL) return ERR_PTR(-ENOMEM); + INIT_LIST_HEAD(&req->rl_free); + spin_lock(&buffer->rb_reqslock); + list_add(&req->rl_all, &buffer->rb_allreqs); + spin_unlock(&buffer->rb_reqslock); req->rl_buffer = &r_xprt->rx_buf; return req; } -static struct rpcrdma_rep * +struct rpcrdma_rep * rpcrdma_create_rep(struct rpcrdma_xprt *r_xprt) { struct rpcrdma_create_data_internal *cdata = &r_xprt->rx_data; @@ -922,6 +941,7 @@ rpcrdma_buffer_create(struct rpcrdma_xprt *r_xprt) int i, rc; buf->rb_max_requests = r_xprt->rx_data.max_requests; + buf->rb_bc_srv_max_requests = 0; spin_lock_init(&buf->rb_lock); rc = ia->ri_ops->ro_init(r_xprt); @@ -929,6 +949,8 @@ rpcrdma_buffer_create(struct rpcrdma_xprt *r_xprt) goto out; INIT_LIST_HEAD(&buf->rb_send_bufs); + INIT_LIST_HEAD(&buf->rb_allreqs); + spin_lock_init(&buf->rb_reqslock); for (i = 0; i < buf->rb_max_requests; i++) { struct rpcrdma_req *req; @@ -939,6 +961,7 @@ rpcrdma_buffer_create(struct rpcrdma_xprt *r_xprt) rc = PTR_ERR(req); goto out; } + req->rl_backchannel = false; list_add(&req->rl_free, &buf->rb_send_bufs); } @@ -987,19 +1010,13 @@ rpcrdma_buffer_get_rep_locked(struct rpcrdma_buffer *buf) static void rpcrdma_destroy_rep(struct rpcrdma_ia *ia, struct rpcrdma_rep *rep) { - if (!rep) - return; - rpcrdma_free_regbuf(ia, rep->rr_rdmabuf); kfree(rep); } -static void +void rpcrdma_destroy_req(struct rpcrdma_ia *ia, struct rpcrdma_req *req) { - if (!req) - return; - rpcrdma_free_regbuf(ia, req->rl_sendbuf); rpcrdma_free_regbuf(ia, req->rl_rdmabuf); kfree(req); @@ -1017,12 +1034,19 @@ rpcrdma_buffer_destroy(struct rpcrdma_buffer *buf) rpcrdma_destroy_rep(ia, rep); } - while (!list_empty(&buf->rb_send_bufs)) { + spin_lock(&buf->rb_reqslock); + while (!list_empty(&buf->rb_allreqs)) { struct rpcrdma_req *req; - req = rpcrdma_buffer_get_req_locked(buf); + req = list_first_entry(&buf->rb_allreqs, + struct rpcrdma_req, rl_all); + list_del(&req->rl_all); + + spin_unlock(&buf->rb_reqslock); rpcrdma_destroy_req(ia, req); + spin_lock(&buf->rb_reqslock); } + spin_unlock(&buf->rb_reqslock); ia->ri_ops->ro_destroy(buf); } @@ -1290,6 +1314,47 @@ rpcrdma_ep_post_recv(struct rpcrdma_ia *ia, return rc; } +/** + * rpcrdma_ep_post_extra_recv - Post buffers for incoming backchannel requests + * @r_xprt: transport associated with these backchannel resources + * @min_reqs: minimum number of incoming requests expected + * + * Returns zero if all requested buffers were posted, or a negative errno. + */ +int +rpcrdma_ep_post_extra_recv(struct rpcrdma_xprt *r_xprt, unsigned int count) +{ + struct rpcrdma_buffer *buffers = &r_xprt->rx_buf; + struct rpcrdma_ia *ia = &r_xprt->rx_ia; + struct rpcrdma_ep *ep = &r_xprt->rx_ep; + struct rpcrdma_rep *rep; + unsigned long flags; + int rc; + + while (count--) { + spin_lock_irqsave(&buffers->rb_lock, flags); + if (list_empty(&buffers->rb_recv_bufs)) + goto out_reqbuf; + rep = rpcrdma_buffer_get_rep_locked(buffers); + spin_unlock_irqrestore(&buffers->rb_lock, flags); + + rc = rpcrdma_ep_post_recv(ia, ep, rep); + if (rc) + goto out_rc; + } + + return 0; + +out_reqbuf: + spin_unlock_irqrestore(&buffers->rb_lock, flags); + pr_warn("%s: no extra receive buffers\n", __func__); + return -ENOMEM; + +out_rc: + rpcrdma_recv_buffer_put(rep); + return rc; +} + /* How many chunk list items fit within our inline buffers? */ unsigned int diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h index 6ea1dbe..1eb86c79f 100644 --- a/net/sunrpc/xprtrdma/xprt_rdma.h +++ b/net/sunrpc/xprtrdma/xprt_rdma.h @@ -263,6 +263,9 @@ struct rpcrdma_req { struct rpcrdma_regbuf *rl_rdmabuf; struct rpcrdma_regbuf *rl_sendbuf; struct rpcrdma_mr_seg rl_segments[RPCRDMA_MAX_SEGS]; + + struct list_head rl_all; + bool rl_backchannel; }; static inline struct rpcrdma_req * @@ -291,6 +294,10 @@ struct rpcrdma_buffer { struct list_head rb_send_bufs; struct list_head rb_recv_bufs; u32 rb_max_requests; + + u32 rb_bc_srv_max_requests; + spinlock_t rb_reqslock; /* protect rb_allreqs */ + struct list_head rb_allreqs; }; #define rdmab_to_ia(b) (&container_of((b), struct rpcrdma_xprt, rx_buf)->rx_ia) @@ -411,6 +418,9 @@ int rpcrdma_ep_post_recv(struct rpcrdma_ia *, struct rpcrdma_ep *, /* * Buffer calls - xprtrdma/verbs.c */ +struct rpcrdma_req *rpcrdma_create_req(struct rpcrdma_xprt *); +struct rpcrdma_rep *rpcrdma_create_rep(struct rpcrdma_xprt *); +void rpcrdma_destroy_req(struct rpcrdma_ia *, struct rpcrdma_req *); int rpcrdma_buffer_create(struct rpcrdma_xprt *); void rpcrdma_buffer_destroy(struct rpcrdma_buffer *); @@ -427,6 +437,7 @@ void rpcrdma_free_regbuf(struct rpcrdma_ia *, struct rpcrdma_regbuf *); unsigned int rpcrdma_max_segments(struct rpcrdma_xprt *); +int rpcrdma_ep_post_extra_recv(struct rpcrdma_xprt *, unsigned int); int frwr_alloc_recovery_wq(void); void frwr_destroy_recovery_wq(void); @@ -494,6 +505,15 @@ int rpcrdma_marshal_req(struct rpc_rqst *); int xprt_rdma_init(void); void xprt_rdma_cleanup(void); +/* Backchannel calls - xprtrdma/backchannel.c + */ +#if defined(CONFIG_SUNRPC_BACKCHANNEL) +int xprt_rdma_bc_setup(struct rpc_xprt *, unsigned int); +int rpcrdma_bc_post_recv(struct rpcrdma_xprt *, unsigned int); +void xprt_rdma_bc_free_rqst(struct rpc_rqst *); +void xprt_rdma_bc_destroy(struct rpc_xprt *, unsigned int); +#endif /* CONFIG_SUNRPC_BACKCHANNEL */ + /* Temporary NFS request map cache. Created in svc_rdma.c */ extern struct kmem_cache *svc_rdma_map_cachep; /* WR context cache. Created in svc_rdma.c */