hv: do not lose pending heartbeat vmbus packets
authorLong Li <longli@microsoft.com>
Wed, 5 Oct 2016 23:57:46 +0000 (16:57 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 25 Oct 2016 06:52:10 +0000 (08:52 +0200)
The host keeps sending heartbeat packets independent of the
guest responding to them.  Even though we respond to the heartbeat messages at
interrupt level, we can have situations where there maybe multiple heartbeat
messages pending that have not been responded to. For instance this occurs when the
VM is paused and the host continues to send the heartbeat messages.
Address this issue by draining and responding to all
the heartbeat messages that maybe pending.

Signed-off-by: Long Li <longli@microsoft.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
CC: Stable <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/hv/hv_util.c

index 4aa3cb63fd41f4506254187608c995e22359aa28..bcd06306f3e894a379603a4216cadab475be6b69 100644 (file)
@@ -314,10 +314,14 @@ static void heartbeat_onchannelcallback(void *context)
        u8 *hbeat_txf_buf = util_heartbeat.recv_buffer;
        struct icmsg_negotiate *negop = NULL;
 
-       vmbus_recvpacket(channel, hbeat_txf_buf,
-                        PAGE_SIZE, &recvlen, &requestid);
+       while (1) {
+
+               vmbus_recvpacket(channel, hbeat_txf_buf,
+                                PAGE_SIZE, &recvlen, &requestid);
+
+               if (!recvlen)
+                       break;
 
-       if (recvlen > 0) {
                icmsghdrp = (struct icmsg_hdr *)&hbeat_txf_buf[
                                sizeof(struct vmbuspipe_hdr)];