Return-Path: Received: from mail-io0-f173.google.com ([209.85.223.173]:35830 "EHLO mail-io0-f173.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752388AbbJXV0j (ORCPT ); Sat, 24 Oct 2015 17:26:39 -0400 Subject: [PATCH v4 02/16] xprtrdma: Re-arm after missed events From: Chuck Lever To: Anna.Schumaker@netapp.com Cc: linux-rdma@vger.kernel.org, linux-nfs@vger.kernel.org Date: Sat, 24 Oct 2015 17:26:37 -0400 Message-ID: <20151024212637.4715.43509.stgit@manet.1015granger.net> In-Reply-To: <20151024211707.4715.61650.stgit@manet.1015granger.net> References: <20151024211707.4715.61650.stgit@manet.1015granger.net> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Sender: linux-nfs-owner@vger.kernel.org List-ID: ib_req_notify_cq(IB_CQ_REPORT_MISSED_EVENTS) returns a positive value if WCs were added to a CQ after the last completion upcall but before the CQ has been re-armed. Commit 7f23f6f6e388 ("xprtrmda: Reduce lock contention in completion handlers") assumed that when ib_req_notify_cq() returned a positive RC, the CQ had also been successfully re-armed, making it safe to return control to the provider without losing any completion signals. That is an invalid assumption. Change both completion handlers to continue polling while ib_req_notify_cq() returns a positive value. Fixes: 7f23f6f6e388 ("xprtrmda: Reduce lock contention in ...") Signed-off-by: Chuck Lever Reviewed-by: Sagi Grimberg Reviewed-by: Devesh Sharma Tested-By: Devesh Sharma --- net/sunrpc/xprtrdma/verbs.c | 66 +++++++------------------------------------ 1 file changed, 10 insertions(+), 56 deletions(-) diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c index 5502d4d..61eea73 100644 --- a/net/sunrpc/xprtrdma/verbs.c +++ b/net/sunrpc/xprtrdma/verbs.c @@ -179,38 +179,17 @@ rpcrdma_sendcq_poll(struct ib_cq *cq, struct rpcrdma_ep *ep) return 0; } -/* - * Handle send, fast_reg_mr, and local_inv completions. - * - * Send events are typically suppressed and thus do not result - * in an upcall. Occasionally one is signaled, however. This - * prevents the provider's completion queue from wrapping and - * losing a completion. +/* Handle provider send completion upcalls. */ static void rpcrdma_sendcq_upcall(struct ib_cq *cq, void *cq_context) { struct rpcrdma_ep *ep = (struct rpcrdma_ep *)cq_context; - int rc; - - rc = rpcrdma_sendcq_poll(cq, ep); - if (rc) { - dprintk("RPC: %s: ib_poll_cq failed: %i\n", - __func__, rc); - return; - } - rc = ib_req_notify_cq(cq, - IB_CQ_NEXT_COMP | IB_CQ_REPORT_MISSED_EVENTS); - if (rc == 0) - return; - if (rc < 0) { - dprintk("RPC: %s: ib_req_notify_cq failed: %i\n", - __func__, rc); - return; - } - - rpcrdma_sendcq_poll(cq, ep); + do { + rpcrdma_sendcq_poll(cq, ep); + } while (ib_req_notify_cq(cq, IB_CQ_NEXT_COMP | + IB_CQ_REPORT_MISSED_EVENTS) > 0); } static void @@ -274,42 +253,17 @@ out_schedule: return rc; } -/* - * Handle receive completions. - * - * It is reentrant but processes single events in order to maintain - * ordering of receives to keep server credits. - * - * It is the responsibility of the scheduled tasklet to return - * recv buffers to the pool. NOTE: this affects synchronization of - * connection shutdown. That is, the structures required for - * the completion of the reply handler must remain intact until - * all memory has been reclaimed. +/* Handle provider receive completion upcalls. */ static void rpcrdma_recvcq_upcall(struct ib_cq *cq, void *cq_context) { struct rpcrdma_ep *ep = (struct rpcrdma_ep *)cq_context; - int rc; - - rc = rpcrdma_recvcq_poll(cq, ep); - if (rc) { - dprintk("RPC: %s: ib_poll_cq failed: %i\n", - __func__, rc); - return; - } - rc = ib_req_notify_cq(cq, - IB_CQ_NEXT_COMP | IB_CQ_REPORT_MISSED_EVENTS); - if (rc == 0) - return; - if (rc < 0) { - dprintk("RPC: %s: ib_req_notify_cq failed: %i\n", - __func__, rc); - return; - } - - rpcrdma_recvcq_poll(cq, ep); + do { + rpcrdma_recvcq_poll(cq, ep); + } while (ib_req_notify_cq(cq, IB_CQ_NEXT_COMP | + IB_CQ_REPORT_MISSED_EVENTS) > 0); } static void