Linux RDMA and InfiniBand development
 help / color / mirror / Atom feed
From: Gerd Rausch <gerd.rausch@oracle.com>
To: Jason Gunthorpe <jgg@ziepe.ca>, Leon Romanovsky <leon@kernel.org>
Cc: linux-rdma@vger.kernel.org
Subject: [PATCH 1/1] RDMA/addr: Refresh neighbour entries upon "rdma_resolve_addr"
Date: Mon, 06 Jun 2022 12:38:36 -0700	[thread overview]
Message-ID: <eb4e348ec730900a47caeeb08fe4aff903337675.camel@oracle.com> (raw)

Unlike with IPv[46], where "ip_finish_output2" triggers
a refresh of STALE neighbour entries via "neigh_output",
"rdma_resolve_addr" never triggers an update.

If a wrong STALE entry ever enters the cache, it'll remain
wrong forever (unless refreshed via TCP/IP, or otherwise).

Let the cache inconsistency resolve itself by triggering
an update from "rdma_resolve_addr".

Signed-off-by: Gerd Rausch <gerd.rausch@oracle.com>
---
 drivers/infiniband/core/addr.c | 40 ++++++++++++++++++++++++++++++++++
 1 file changed, 40 insertions(+)

diff --git a/drivers/infiniband/core/addr.c b/drivers/infiniband/core/addr.c
index f253295795f0..db2a86971e25 100644
--- a/drivers/infiniband/core/addr.c
+++ b/drivers/infiniband/core/addr.c
@@ -394,6 +394,8 @@ static int addr4_resolve(struct sockaddr *src_sock,
 	__be32 dst_ip = dst_in->sin_addr.s_addr;
 	struct rtable *rt;
 	struct flowi4 fl4;
+	struct net_device *dev;
+	struct neighbour *neigh;
 	int ret;
 
 	memset(&fl4, 0, sizeof(fl4));
@@ -409,6 +411,24 @@ static int addr4_resolve(struct sockaddr *src_sock,
 
 	addr->hoplimit = ip4_dst_hoplimit(&rt->dst);
 
+	/* trigger ARP-entry refresh if necessary,
+	 * the same way "ip_finish_output2" does
+	 */
+	if (addr->bound_dev_if) {
+		dev = dev_get_by_index(addr->net, addr->bound_dev_if);
+	} else {
+		dev = rt->dst.dev;
+		dev_hold(dev);
+	}
+	if (dev) {
+		rcu_read_lock_bh();
+		neigh = __ipv4_neigh_lookup_noref(dev, dst_ip);
+		if (neigh)
+			neigh_event_send(neigh, NULL);
+		rcu_read_unlock_bh();
+		dev_put(dev);
+	}
+
 	*prt = rt;
 	return 0;
 }
@@ -424,6 +444,8 @@ static int addr6_resolve(struct sockaddr *src_sock,
 				(const struct sockaddr_in6 *)dst_sock;
 	struct flowi6 fl6;
 	struct dst_entry *dst;
+	struct net_device *dev;
+	struct neighbour *neigh;
 
 	memset(&fl6, 0, sizeof fl6);
 	fl6.daddr = dst_in->sin6_addr;
@@ -439,6 +461,24 @@ static int addr6_resolve(struct sockaddr *src_sock,
 
 	addr->hoplimit = ip6_dst_hoplimit(dst);
 
+	/* trigger neighbour-entry refresh if necessary,
+	 * the same way "ip6_finish_output2" does
+	 */
+	if (addr->bound_dev_if) {
+		dev = dev_get_by_index(addr->net, addr->bound_dev_if);
+	} else {
+		dev = dst->dev;
+		dev_hold(dev);
+	}
+	if (dev) {
+		rcu_read_lock_bh();
+		neigh = __ipv6_neigh_lookup_noref(dst->dev, &dst_in->sin6_addr);
+		if (neigh)
+			neigh_event_send(neigh, NULL);
+		rcu_read_unlock_bh();
+		dev_put(dev);
+	}
+
 	*pdst = dst;
 	return 0;
 }


             reply	other threads:[~2022-06-06 19:38 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-06-06 19:38 Gerd Rausch [this message]
2022-06-07  8:21 ` [PATCH 1/1] RDMA/addr: Refresh neighbour entries upon "rdma_resolve_addr" kernel test robot
2022-06-16  7:08 ` Leon Romanovsky
2022-06-16 15:50   ` Gerd Rausch

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=eb4e348ec730900a47caeeb08fe4aff903337675.camel@oracle.com \
    --to=gerd.rausch@oracle.com \
    --cc=jgg@ziepe.ca \
    --cc=leon@kernel.org \
    --cc=linux-rdma@vger.kernel.org \
    /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