[ovs-discuss] [NIC-20 08/11] secchan: Fix behavior when a network device is renamed.

Ben Pfaff blp at nicira.com
Wed Aug 5 22:37:11 UTC 2009


update_port() deals with the case where we have been notified that a
network device with a given name, that is part of the datapath, has changed
in some way.  In particular it breaks the problem space up into ports that
have been added, deleted, or modified.

But the code here deals badly with the case where the only change is that
the network device associated with a port has been renamed (which is
reported to it with 'devname' as the network device's new named): it
looks up devname in the ofproto's index by name and doesn't find it, then
it looks up the port number assigned to the netdev in the ofproto's index
by datapath index and sees that there already is one.  This makes it
think that it's a new port, but with a port number that conflicts with an
existing port (under the old name for the port), which makes it discard
the notification and keep the old netdev name, and so afterward nothing
on the netdev will work since it still has the old netdev name.

This rewrite fixes the problem and simplifies the code.
---
 secchan/ofproto.c |   92 +++++++++++++++++++++++++++++++---------------------
 1 files changed, 55 insertions(+), 37 deletions(-)

diff --git a/secchan/ofproto.c b/secchan/ofproto.c
index 9f3a849..efa5c9b 100644
--- a/secchan/ofproto.c
+++ b/secchan/ofproto.c
@@ -1213,53 +1213,71 @@ static void
 update_port(struct ofproto *p, const char *devname)
 {
     struct odp_port odp_port;
-    struct ofport *ofport;
+    struct ofport *old_ofport;
+    struct ofport *new_ofport;
     int error;
 
     COVERAGE_INC(ofproto_update_port);
-    ofport = shash_find_data(&p->port_by_name, devname);
+
+    /* Query the datapath for port information. */
     error = dpif_port_query_by_name(&p->dpif, devname, &odp_port);
-    if (!error) {
-        if (!ofport) {
-            /* New port. */
-            if (!ofport_conflicts(p, &odp_port)) {
-                ofport = make_ofport(&odp_port);
-                if (ofport) {
-                    ofport_install(p, ofport);
-                    send_port_status(p, ofport, OFPPR_ADD);
-                }
-            }
-        } else {
-            /* Modified port. */
-            struct ofport *new_ofport = make_ofport(&odp_port);
-            if (!new_ofport) {
-                return;
-            }
 
-            new_ofport->opp.config &= OFPPC_PORT_DOWN;
-            new_ofport->opp.config |= ofport->opp.config & ~OFPPC_PORT_DOWN;
-            if (ofport_equal(ofport, new_ofport)) {
-                /* False alarm--no change. */
-                ofport_free(new_ofport);
-            } else {
-                ofport_remove(p, ofport);
-                ofport_install(p, new_ofport);
-                ofport_free(ofport);
-                send_port_status(p, new_ofport, OFPPR_MODIFY);
-            }
-        }
-    } else if (error == ENOENT || error == ENODEV) {
-        /* Deleted port. */
-        if (ofport) {
-            send_port_status(p, ofport, OFPPR_DELETE);
-            ofport_remove(p, ofport);
-            ofport_free(ofport);
+    /* Find the old ofport. */
+    old_ofport = shash_find_data(&p->port_by_name, devname);
+    if (!error) {
+        if (!old_ofport) {
+            /* There's no port named 'devname' but there might be a port with
+             * the same port number.  This could happen if a port is deleted
+             * and then a new one added in its place very quickly, or if a port
+             * is renamed.  In the former case we want to send an OFPPR_DELETE
+             * and an OFPPR_ADD, and in the latter case we want to send a
+             * single OFPPR_MODIFY.  We can distinguish the cases by comparing
+             * the old port's ifindex against the new port, or perhaps less
+             * reliably but more portably by comparing the old port's MAC
+             * against the new port's MAC.  However, this code isn't that smart
+             * and always sends an OFPPR_MODIFY (XXX). */
+            old_ofport = port_array_get(&p->ports, odp_port.port);
         }
-    } else {
+    } else if (error != ENOENT && error != ENODEV) {
         VLOG_WARN_RL(&rl, "dpif_port_query_by_name returned unexpected error "
                      "%s", strerror(error));
         return;
     }
+
+    /* Create a new ofport. */
+    new_ofport = !error ? make_ofport(&odp_port) : NULL;
+
+    /* Eliminate a few pathological cases. */
+    if (!old_ofport && !new_ofport) {
+        return;
+    } else if (old_ofport && new_ofport) {
+        /* Most of the 'config' bits are OpenFlow soft state, but
+         * OFPPC_PORT_DOWN is maintained the kernel.  So transfer the OpenFlow
+         * bits from old_ofport.  (make_ofport() only sets OFPPC_PORT_DOWN and
+         * leaves the other bits 0.)  */
+        new_ofport->opp.config |= old_ofport->opp.config & ~OFPPC_PORT_DOWN;
+
+        if (ofport_equal(old_ofport, new_ofport)) {
+            /* False alarm--no change. */
+            ofport_free(new_ofport);
+            return;
+        }
+    }
+
+    /* Now deal with the normal cases. */
+    if (old_ofport) {
+        ofport_remove(p, old_ofport);
+    }
+    if (new_ofport) {
+        ofport_install(p, new_ofport);
+    }
+    send_port_status(p, new_ofport ? new_ofport : old_ofport,
+                     (!old_ofport ? OFPPR_ADD
+                      : !new_ofport ? OFPPR_DELETE
+                      : OFPPR_MODIFY));
+    ofport_free(old_ofport);
+
+    /* Update port groups. */
     refresh_port_groups(p);
 }
 
-- 
1.6.3.3





More information about the discuss mailing list