Fw: [Bugme-new] [Bug 3992] New: Bondig. Not correct work function ARP Monitoring. Broken link.

Fw: [Bugme-new] [Bug 3992] New: Bondig. Not correct work function ARP Monitoring. Broken link.

[Andrew Morton]

Begin forwarded message:

Date: Wed, 5 Jan 2005 07:59:17 -0800
From: bugme-daemon@osdl.org
To: bugme-new@lists.osdl.org
Subject: [Bugme-new] [Bug 3992] New: Bondig. Not correct work function ARP Monitoring. Broken link.


http://bugme.osdl.org/show_bug.cgi?id=3992

           Summary: Bondig. Not correct work function ARP Monitoring. Broken
                    link.
    Kernel Version: 2.6.10
            Status: NEW
          Severity: normal
             Owner: jgarzik@pobox.com
         Submitter: stanislav@muhachev.petro.ru


Distribution:
Hardware Environment:
Vmware GSX machine
Software Environment:
Gentoo
Problem Description:
bonding interface going down and down slave link


Steps to reproduce:

2 virtual pc
identical gentoo linux (copy)
1 nic for both virtual pc
up 2 openvpn link trouch nic to another virtual pc (tap0 and tap1)[ethernet 
bridge]
link both vpn1 and vpn2 - ok!(independing link like:192.168.1.1-192.168.1.2 
and 192.168.2.1-192.168.2.2)

bonding vpn1 & vpn2 òî bond0 (bonding default setup -> nothing failover 
setings)
link îê!(192.168.100.1-192.168.100.2)

setting arp monitor in bonding (TUN/TAP driver not support Mii status)
link down!

arp request go from bond0(machine1)[from tap0 & tap1] to bond0(machine2) - ok.
arp answer bond0(machine2) interface - ok!
!!! bond0(machine2) -> tap0(tap1)(machine2) - noting!!!(arp answer broken)
resultat arp request not complite

situation analog both side

for monitoring use TCPDUMP

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

Re: Fw: [Bugme-new] [Bug 3992] New: Bondig. Not correct work function ARP Monitoring. Broken link.

[Tommy Christensen]

[-- Attachment #1: Type: text/plain, Size: 986 bytes --]

> http://bugme.osdl.org/show_bug.cgi?id=3992
> 
>            Summary: Bondig. Not correct work function ARP Monitoring. Broken
>                     link.
>     Kernel Version: 2.6.10
>             Status: NEW
>           Severity: normal
>              Owner: jgarzik@pobox.com
>          Submitter: stanislav@muhachev.petro.ru
> 
...
> 
> bonding vpn1 & vpn2 òî bond0 (bonding default setup -> nothing failover 
> setings)
> link îê!(192.168.100.1-192.168.100.2)
> 
> setting arp monitor in bonding (TUN/TAP driver not support Mii status)
> link down!

I am not sure I understand your setup completely ...
But as stated in bonding.txt, the ARP monitor requires the underlying
driver to update dev->trans_start and dev->last_rx.

Since the TUN/TAP driver doesn't maintain these, it could very well
explain the behavior you are seeing.

The patch below adds the required functionality to the TUN/TAP driver.
Please test if this helps in your case.

-Tommy

[-- Attachment #2: tun.c.patch --]
[-- Type: text/plain, Size: 508 bytes --]

--- linux-2.6.10-bk14/drivers/net/tun.c	2005-01-10 16:28:48.000000000 +0100
+++ linux-2.6.10-work/drivers/net/tun.c	2005-01-11 23:05:56.759691345 +0100
@@ -92,6 +92,7 @@
 			goto drop;
 	}
 	skb_queue_tail(&tun->readq, skb);
+	dev->trans_start = jiffies;
 
 	/* Notify and wake up reader process */
 	if (tun->flags & TUN_FASYNC)
@@ -240,6 +241,7 @@
 		skb->ip_summed = CHECKSUM_UNNECESSARY;
  
 	netif_rx_ni(skb);
+	tun->dev->last_rx = jiffies;
    
 	tun->stats.rx_packets++;
 	tun->stats.rx_bytes += len;