From: Steve Best <sbest@redhat.com>
Date: Tue, 24 Aug 2010 21:21:11 -0400
Subject: [net] ibmveth: fix lost IRQ that leads to service loss
Message-id: <20100824210932.12694.16339.sendpatchset@squad5-lp1.lab.bos.redhat.com>
Patchwork-id: 27808
O-Subject: [PATCH RHEL5.6 BZ626841] ibmveth: lost IRQ while closing/opening
	device leads to service loss
Bugzilla: 626841
RH-Acked-by: Stefan Assmann <sassmann@redhat.com>
RH-Acked-by: Jiri Pirko <jpirko@redhat.com>
RH-Acked-by: Dean Nelson <dnelson@redhat.com>
RH-Acked-by: David Howells <dhowells@redhat.com>
RH-Acked-by: David S. Miller <davem@redhat.com>

RHBZ#:
------
https://bugzilla.redhat.com/show_bug.cgi?id=626841

Description:
------------
The order of freeing the IRQ and freeing the device in firmware
in ibmveth_close can cause the adapter to become unusable after a
subsequent ibmveth_open.  Only a reboot of the OS will make the
network device usable again. This is seen when cycling the adapter
up and down while there is network activity.

There is a window where an IRQ will be left unserviced (H_EOI will not
be called).  The solution is to make a VIO_IRQ_DISABLE h_call, free the
device with firmware, and then call free_irq.

Signed-off-by: Robert Jennings <rcj@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

RHEL Version Found:
-------------------
RHEL 5.4

Brew:
-----
http://brewweb.devel.redhat.com/brew/taskinfo?taskID=2704871

Upstream:
---------
http://git.kernel.org/?p-linux/kernel/git/torvalds/linux-2.6.git;
a-commit;h-ee2e6114de3bdb1c34f3910b690f990483e981ab

Test Status:
------------
Steps to reproduce.
Robert Jennings helped me test this patch.

The timing window is small; we had luck reproducing this consistently on a lab
machine owned by the customer support team (on RHEL5.4) but could not recreate
in the development lab.

There are two partitions (p1 and p2) and the following is done:
p1# nc -ul 60000 > /dev/null&
p2# dd if-/dev/zero | nc -u 10.9.47.70 60000
p1# time while true; do date; echo "bringing interface down..."; ifdown eth1;
sleep 10; echo "bringing interface up..."; ifup eth1; sleep 2; ping -c 1
10.47.65.71 || break; done

---------------------------------------------------------------
Steve Best
IBM on-site partner

Proposed Patch:
---------------

Signed-off-by: Jarod Wilson <jarod@redhat.com>

diff --git a/drivers/net/ibmveth.c b/drivers/net/ibmveth.c
index 4959fb9..2c50271 100644
--- a/drivers/net/ibmveth.c
+++ b/drivers/net/ibmveth.c
@@ -621,7 +621,7 @@ static int ibmveth_close(struct net_device *netdev)
 	if (!adapter->pool_config)
 		netif_stop_queue(netdev);
 
-	free_irq(netdev->irq, netdev);
+	h_vio_signal(adapter->vdev->unit_address, VIO_IRQ_DISABLE);
 
 	do {
 		lpar_rc = h_free_logical_lan(adapter->vdev->unit_address);
@@ -633,6 +633,8 @@ static int ibmveth_close(struct net_device *netdev)
 				     lpar_rc);
 	}
 
+	free_irq(netdev->irq, netdev);
+
 	adapter->rx_no_buffer = *(u64*)(((char*)adapter->buffer_list_addr) + 4096 - 8);
 
 	ibmveth_cleanup(adapter);