[libvirt] [PATCH 2/4] rpc: Discard non-blocking calls only when necessary

[Jiri Denemark]

Currently, non-blocking calls are either sent immediately or discarded
in case sending would block. This was implemented based on the
assumption that the non-blocking keepalive call is not needed as there
are other calls in the queue which would keep the connection alive.
However, if those calls are no-reply calls (such as those carrying
stream data), the remote party knows the connection is alive but since
we don't get any reply from it, we think the connection is dead.

This is most visible in tunnelled migration. If it happens to be longer
than keepalive timeout (30s by default), it may be unexpectedly aborted
because the connection is considered to be dead.

With this patch, we only discard non-blocking calls when the last call
with a thread is completed and thus there is no thread left to keep
sending the remaining non-blocking calls.
---
 src/rpc/virnetclient.c |   21 +++++++++------------
 1 files changed, 9 insertions(+), 12 deletions(-)

diff --git a/src/rpc/virnetclient.c b/src/rpc/virnetclient.c
index 33b7701..8952dd4 100644
--- a/src/rpc/virnetclient.c
+++ b/src/rpc/virnetclient.c
@@ -1257,6 +1257,13 @@ static void virNetClientIOEventLoopPassTheBuck(virNetClientPtr client, virNetCli
     }
     client->haveTheBuck = false;
 
+    /* Remove non-blocking calls from the dispatch list since there is no
+     * call with a thread in the list which could take care of them.
+     */
+    virNetClientCallRemovePredicate(&client->waitDispatch,
+                                    virNetClientIOEventLoopRemoveNonBlocking,
+                                    thiscall);
+
     VIR_DEBUG("No thread to pass the buck to");
     if (client->wantClose) {
         virNetClientCloseLocked(client);
@@ -1300,12 +1307,9 @@ static int virNetClientIOEventLoop(virNetClientPtr client,
         if (virNetSocketHasCachedData(client->sock) || client->wantClose)
             timeout = 0;
 
-        /* If there are any non-blocking calls in the queue,
-         * then we don't want to sleep in poll()
+        /* If we are non-blocking, we don't want to sleep in poll()
          */
-        if (virNetClientCallMatchPredicate(client->waitDispatch,
-                                           virNetClientIOEventLoopWantNonBlock,
-                                           NULL))
+        if (thiscall->nonBlock)
             timeout = 0;
 
         fds[0].events = fds[0].revents = 0;
@@ -1410,13 +1414,6 @@ static int virNetClientIOEventLoop(virNetClientPtr client,
                                         virNetClientIOEventLoopRemoveDone,
                                         thiscall);
 
-        /* Iterate through waiting calls and if any are
-         * non-blocking, remove them from the dispatch list...
-         */
-        virNetClientCallRemovePredicate(&client->waitDispatch,
-                                        virNetClientIOEventLoopRemoveNonBlocking,
-                                        thiscall);
-
         /* Now see if *we* are done */
         if (thiscall->mode == VIR_NET_CLIENT_MODE_COMPLETE) {
             virNetClientCallRemove(&client->waitDispatch, thiscall);
-- 
1.7.8.5