Unregister event callback if a fatal error occurs during dispatch

If we get an I/O error in the async event callback for an RPC
client, we might not have consumed all pending data off the
wire. This could result in the callback being immediately
invoked again. At which point the same I/O might occur. And
we're invoked again. And again...And again...

Unregistering the async event callback if an error occurs is
a good safety net. The real error will be seen when the next
RPC method is invoked

* src/rpc/virnetclient.c: Unregister event callback on error

diff --git a/src/rpc/virnetclient.c b/src/rpc/virnetclient.c
index d3965c6d66162bc3a5bb737a7a879c420fc42580..9eaecfcbee0ea9960e961acdb9e7152449ad7aae 100644 (file)
--- a/src/rpc/virnetclient.c
+++ b/src/rpc/virnetclient.c
@@ -1117,8 +1117,10 @@ void virNetClientIncomingEvent(virNetSocketPtr sock,
         goto done;
     }
 
-    if (virNetClientIOHandleInput(client) < 0)
-        VIR_DEBUG("Something went wrong during async message processing");
+    if (virNetClientIOHandleInput(client) < 0) {
+        VIR_WARN("Something went wrong during async message processing");
+        virNetSocketRemoveIOCallback(sock);
+    }
 
 done:
     virNetClientUnlock(client);