From patchwork Fri May 18 16:00:04 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Doug Ledford X-Patchwork-Id: 10411093 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id 858B7602CB for ; Fri, 18 May 2018 16:00:41 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 74559289FB for ; Fri, 18 May 2018 16:00:41 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 688BE28A00; Fri, 18 May 2018 16:00:41 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-7.9 required=2.0 tests=BAYES_00, MAILING_LIST_MULTI, RCVD_IN_DNSWL_HI autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id C0E7D289A9 for ; Fri, 18 May 2018 16:00:40 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751494AbeERQAj (ORCPT ); Fri, 18 May 2018 12:00:39 -0400 Received: from mx3-rdu2.redhat.com ([66.187.233.73]:34740 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1750841AbeERQAj (ORCPT ); Fri, 18 May 2018 12:00:39 -0400 Received: from smtp.corp.redhat.com (int-mx04.intmail.prod.int.rdu2.redhat.com [10.11.54.4]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id D488E818F6EF for ; Fri, 18 May 2018 16:00:38 +0000 (UTC) Received: from haswell-e.nc.xsintricity.com.com (ovpn-122-18.rdu2.redhat.com [10.10.122.18]) by smtp.corp.redhat.com (Postfix) with ESMTP id B40E02026E03; Fri, 18 May 2018 16:00:38 +0000 (UTC) From: Doug Ledford To: linux-rdma@vger.kernel.org Cc: Doug Ledford Subject: [PATCH] RDMA/ipoib: Update paths on CLIENT_REREG/SM_CHANGE events Date: Fri, 18 May 2018 12:00:04 -0400 Message-Id: <72ab5c33597f19082cd98674a204b34c59e2b772.1526658945.git.dledford@redhat.com> In-Reference-To: <20180504114921.16571-1-evgenii.smirnov@profitbricks.com> X-Scanned-By: MIMEDefang 2.78 on 10.11.54.4 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.11.55.8]); Fri, 18 May 2018 16:00:38 +0000 (UTC) X-Greylist: inspected by milter-greylist-4.5.16 (mx1.redhat.com [10.11.55.8]); Fri, 18 May 2018 16:00:38 +0000 (UTC) for IP:'10.11.54.4' DOMAIN:'int-mx04.intmail.prod.int.rdu2.redhat.com' HELO:'smtp.corp.redhat.com' FROM:'dledford@redhat.com' RCPT:'' Sender: linux-rdma-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-rdma@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP We do a light flush on CLIENT_REREG and SM_CHANGE events. This goes through and marks paths invalid. But we weren't always checking for this validity when we needed to, and so we could keep using a path marked invalid. What's more, once we establish a path with a valid ah, we put a pointer to the ah in the neigh struct directly, so even if we mark the path as invalid, as long as the neigh has a direct pointer to the ah, it keeps using the old, outdated ah. To fix this we do several things. 1) Put the valid flag in the ah instead of the path struct, so when we put the ah pointer directly in the neigh struct, we can easily check the validity of the ah on send events. 2) Check the neigh->ah and neigh->ah->valid elements in the needed places, and if we have an ah, but it's invalid, then invoke a refresh of the ah. 3) Fix the various places that check for path, but didn't check for path->valid (now path->ah && path->ah->valid). Reported-by: Evgenii Smirnov Fixes: ee1e2c82c245 ("IPoIB: Refresh paths instead of flushing them on SM change events") Signed-off-by: Doug Ledford --- drivers/infiniband/ulp/ipoib/ipoib.h | 2 +- drivers/infiniband/ulp/ipoib/ipoib_main.c | 33 ++++++++++++++++++----- 2 files changed, 28 insertions(+), 7 deletions(-) diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h index 308e0ce49289..a50b062ed13e 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib.h +++ b/drivers/infiniband/ulp/ipoib/ipoib.h @@ -415,6 +415,7 @@ struct ipoib_ah { struct list_head list; struct kref ref; unsigned last_send; + int valid; }; struct ipoib_path { @@ -431,7 +432,6 @@ struct ipoib_path { struct rb_node rb_node; struct list_head list; - int valid; }; struct ipoib_neigh { diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c index cf291f90b58f..788bb9573f1f 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_main.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c @@ -697,7 +697,8 @@ void ipoib_mark_paths_invalid(struct net_device *dev) ipoib_dbg(priv, "mark path LID 0x%08x GID %pI6 invalid\n", be32_to_cpu(sa_path_get_dlid(&path->pathrec)), path->pathrec.dgid.raw); - path->valid = 0; + if (path->ah) + path->ah->valid = 0; } spin_unlock_irq(&priv->lock); @@ -833,7 +834,7 @@ static void path_rec_completion(int status, while ((skb = __skb_dequeue(&neigh->queue))) __skb_queue_tail(&skqueue, skb); } - path->valid = 1; + path->ah->valid = 1; } path->query = NULL; @@ -926,6 +927,24 @@ static int path_rec_start(struct net_device *dev, return 0; } +static void neigh_refresh_path(struct ipoib_neigh *neigh, u8 *daddr, + struct net_device *dev) +{ + struct ipoib_dev_priv *priv = ipoib_priv(dev); + struct ipoib_path *path; + unsigned long flags; + + spin_lock_irqsave(&priv->lock, flags); + + path = __path_find(dev, daddr + 4); + if (!path) + goto out; + if (!path->query) + path_rec_start(dev, path); +out: + spin_unlock_irqrestore(&priv->lock, flags); +} + static struct ipoib_neigh *neigh_add_path(struct sk_buff *skb, u8 *daddr, struct net_device *dev) { @@ -963,7 +982,7 @@ static struct ipoib_neigh *neigh_add_path(struct sk_buff *skb, u8 *daddr, list_add_tail(&neigh->list, &path->neigh_list); - if (path->ah) { + if (path->ah && path->ah->valid) { kref_get(&path->ah->ref); neigh->ah = path->ah; @@ -1034,7 +1053,7 @@ static void unicast_arp_send(struct sk_buff *skb, struct net_device *dev, goto drop_and_unlock; path = __path_find(dev, phdr->hwaddr + 4); - if (!path || !path->valid) { + if (!path || !path->ah || !path->ah->valid) { int new_path = 0; if (!path) { @@ -1069,7 +1088,7 @@ static void unicast_arp_send(struct sk_buff *skb, struct net_device *dev, return; } - if (path->ah) { + if (path->ah && path->ah->valid) { ipoib_dbg(priv, "Send unicast ARP to %08x\n", be32_to_cpu(sa_path_get_dlid(&path->pathrec))); @@ -1161,10 +1180,12 @@ static netdev_tx_t ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev) ipoib_cm_send(dev, skb, ipoib_cm_get(neigh)); goto unref; } - } else if (neigh->ah) { + } else if (neigh->ah && neigh->ah->valid) { neigh->ah->last_send = rn->send(dev, skb, neigh->ah->ah, IPOIB_QPN(phdr->hwaddr)); goto unref; + } else if (neigh->ah) { + neigh_refresh_path(neigh, phdr->hwaddr, dev); } if (skb_queue_len(&neigh->queue) < IPOIB_MAX_PATH_REC_QUEUE) {