Linux Ethernet Bridging - Jan 2008 - [Bridge] Bridge not functional after disconnect / connect

Hi,

I have two NICs on my machine, one connected to a router and is always up, and
the other connected back-to-back to a laptop, and this one I disconnect and
connect all the time. I created a bridge on top of these two interfaces, and it
functions fine, except when I disconnect and connect the laptop, I have to
remove it from the bridge and add it back before it starts bridging! In other
words, I need to do "brctl delif br0 eth0; brctl addif br0 eth0".

The output of "brctl showmacs br0" and "brctl showstp br0"
looks the same and perfectly correct before the removal/adding and after it.
From playing around with Wireshark it seems to me that before delif/addif only
broadcasts are received from the laptop; unicast packets do not appear there. I
got all zeros in /proc/sys/net/bridge/*. It looks like everything is configured
correctly.

Any idea at all why I see this behavior? Any tips as to how I might debug it?

The kernel I'm running is 2.6.18 from Debian Etch.

Please, if any more information is required, let me know. I don't want to
spam the list with long dumps of everything if it's not needed.

Thanks in advance,
--Alex





     
____________________________________________________________________________________
Be a better friend, newshound, and 
know-it-all with Yahoo! Mobile.  Try it now. 
http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ

On Wed, 30 Jan 2008 14:15:11 -0800 (PST)
Alex Shnitman <alexta69@yahoo.com> wrote:
> Hi,
> 
> I have two NICs on my machine, one connected to a router and is always up,
and the other connected back-to-back to a laptop, and this one I disconnect and
connect all the time. I created a bridge on top of these two interfaces, and it
functions fine, except when I disconnect and connect the laptop, I have to
remove it from the bridge and add it back before it starts bridging! In other
words, I need to do "brctl delif br0 eth0; brctl addif br0 eth0".
> 
> The output of "brctl showmacs br0" and "brctl showstp
br0" looks the same and perfectly correct before the removal/adding and
after it. From playing around with Wireshark it seems to me that before
delif/addif only broadcasts are received from the laptop; unicast packets do not
appear there. I got all zeros in /proc/sys/net/bridge/*. It looks like
everything is configured correctly.
> 
> Any idea at all why I see this behavior? Any tips as to how I might debug
it?
> 
> The kernel I'm running is 2.6.18 from Debian Etch.
What hardware? There were many fixes since that old kernel.

-- 
Stephen Hemminger <stephen.hemminger@vyatta.com>

Hi,

lspci gives this:

02:05.0 Ethernet controller: Marvell Technology Group Ltd. 88E8001 Gigabit
Ethernet Controller (rev 13)
02:09.0 Ethernet controller: Realtek Semiconductor Co., Ltd.
RTL-8139/8139C/8139C+ (rev 10)

The first one is the one having the trouble. Is this hardware known to be
problematic?

Thanks,
--Alex


----- Original Message ----
From: Stephen Hemminger <shemminger@linux-foundation.org>
To: Alex Shnitman <alexta69@yahoo.com>
Cc: bridge@lists.linux-foundation.org
Sent: Thursday, January 31, 2008 12:22:22 AM
Subject: Re: [Bridge] Bridge not functional after disconnect / connect


On 
Wed, 
30 
Jan 
2008 
14:15:11 
-0800 
(PST)
Alex 
Shnitman 
<alexta69@yahoo.com> 
wrote:
> 
Hi,
> 
> 
I 
have 
two 
NICs 
on 
my 
machine, 
one 
connected 
to 
a 
router 
and 
is 
always 
up, 
and 
the 
other 
connected 
back-to-back 
to 
a 
laptop, 
and 
this 
one 
I 
disconnect 
and 
connect 
all 
the 
time. 
I 
created 
a 
bridge 
on 
top 
of 
these 
two 
interfaces, 
and 
it 
functions 
fine, 
except 
when 
I 
disconnect 
and 
connect 
the 
laptop, 
I 
have 
to 
remove 
it 
from 
the 
bridge 
and 
add 
it 
back 
before 
it 
starts 
bridging! 
In 
other 
words, 
I 
need 
to 
do 
"brctl 
delif 
br0 
eth0; 
brctl 
addif 
br0 
eth0".
> 
> 
The 
output 
of 
"brctl 
showmacs 
br0" 
and 
"brctl 
showstp 
br0" 
looks 
the 
same 
and 
perfectly 
correct 
before 
the 
removal/adding 
and 
after 
it. 
>From 
playing 
around 
with 
Wireshark 
it 
seems 
to 
me 
that 
before 
delif/addif 
only 
broadcasts 
are 
received 
from 
the 
laptop; 
unicast 
packets 
do 
not 
appear 
there. 
I 
got 
all 
zeros 
in 
/proc/sys/net/bridge/*. 
It 
looks 
like 
everything 
is 
configured 
correctly.
> 
> 
Any 
idea 
at 
all 
why 
I 
see 
this 
behavior? 
Any 
tips 
as 
to 
how 
I 
might 
debug 
it?
> 
> 
The 
kernel 
I'm 
running 
is 
2.6.18 
from 
Debian 
Etch.

What 
hardware? 
There 
were 
many 
fixes 
since 
that 
old 
kernel.

-- 
Stephen 
Hemminger 
<stephen.hemminger@vyatta.com>





     
____________________________________________________________________________________
Looking for last minute shopping deals?  
Find them fast with Yahoo! Search. 
http://tools.search.yahoo.com/newsearch/category.php?category=shopping

At 00:44 today, Alex Shnitman wrote:
> 02:05.0 Ethernet controller: Marvell Technology Group Ltd. 88E8001 Gigabit
Ethernet Controller (rev 13)
Is that the sky2 driver?  That is in my experience pretty buggy until about 
2.6.22.

-- 
Malcolm Scott
Research Assistant
University of Cambridge Computer Laboratory

In fact, it's the skge driver. I'll try to upgrade the kernel to 2.6.22
(the latest one available in etch backports) and report the results.

Thanks,
--Alex


----- Original Message ----
From: Malcolm Scott <Malcolm.Scott@cl.cam.ac.uk>
To: Alex Shnitman <alexta69@yahoo.com>
Cc: Stephen Hemminger <shemminger@linux-foundation.org>;
bridge@lists.linux-foundation.org
Sent: Thursday, January 31, 2008 11:59:34 AM
Subject: Re: [Bridge] Bridge not functional after disconnect / connect


At 
00:44 
today, 
Alex 
Shnitman 
wrote:
> 
02:05.0 
Ethernet 
controller: 
Marvell 
Technology 
Group 
Ltd. 
88E8001 
Gigabit 
Ethernet 
Controller 
(rev 
13)

Is 
that 
the 
sky2 
driver?  
That 
is 
in 
my 
experience 
pretty 
buggy 
until 
about 
2.6.22.

-- 
Malcolm 
Scott
Research 
Assistant
University 
of 
Cambridge 
Computer 
Laboratory






     
____________________________________________________________________________________
Never miss a thing.  Make Yahoo your home page. 
http://www.yahoo.com/r/hs

There were some fixes to carrier handling that went into 2.6.24.

commit 21d7f67700ad7a4523d35d43ce95755e40eae5b7
Author: Stephen Hemminger <shemminger@linux-foundation.org>
Date:   Mon Nov 26 11:54:51 2007 -0800

    skge: fiber link up/down fix
    
    The driver would not work over fibre if other end when down then
    came back up (would require reloading driver). The correct way
    to manage the link the same way for both TP and fibre.
    
    Resloves problem described in: 	 http://lkml.org/lkml/2007/11/6/395
    
    Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
    Signed-off-by: Jeff Garzik <jeff@garzik.org>

diff --git a/drivers/net/skge.c b/drivers/net/skge.c
index b680cb0..73a4246 100644
--- a/drivers/net/skge.c
+++ b/drivers/net/skge.c
@@ -1095,16 +1095,9 @@ static void xm_link_down(struct skge_hw *hw, int port)
 {
 	struct net_device *dev = hw->dev[port];
 	struct skge_port *skge = netdev_priv(dev);
-	u16 cmd = xm_read16(hw, port, XM_MMU_CMD);
 
 	xm_write16(hw, port, XM_IMSK, XM_IMSK_DISABLE);
 
-	cmd &= ~(XM_MMU_ENA_RX | XM_MMU_ENA_TX);
-	xm_write16(hw, port, XM_MMU_CMD, cmd);
-
-	/* dummy read to ensure writing */
-	xm_read16(hw, port, XM_MMU_CMD);
-
 	if (netif_carrier_ok(dev))
 		skge_link_down(skge);
 }
@@ -1194,6 +1187,7 @@ static void genesis_init(struct skge_hw *hw)
 static void genesis_reset(struct skge_hw *hw, int port)
 {
 	const u8 zero[8]  = { 0 };
+	u32 reg;
 
 	skge_write8(hw, SK_REG(port, GMAC_IRQ_MSK), 0);
 
@@ -1209,6 +1203,11 @@ static void genesis_reset(struct skge_hw *hw, int port)
 		xm_write16(hw, port, PHY_BCOM_INT_MASK, 0xffff);
 
 	xm_outhash(hw, port, XM_HSM, zero);
+
+	/* Flush TX and RX fifo */
+	reg = xm_read32(hw, port, XM_MODE);
+	xm_write32(hw, port, XM_MODE, reg | XM_MD_FTF);
+	xm_write32(hw, port, XM_MODE, reg | XM_MD_FRF);
 }
 
 
@@ -1714,6 +1713,12 @@ static void genesis_stop(struct skge_port *skge)
 	struct skge_hw *hw = skge->hw;
 	int port = skge->port;
 	unsigned retries = 1000;
+	u16 cmd;
+
+ 	/* Disable Tx and Rx */
+	cmd = xm_read16(hw, port, XM_MMU_CMD);
+	cmd &= ~(XM_MMU_ENA_RX | XM_MMU_ENA_TX);
+	xm_write16(hw, port, XM_MMU_CMD, cmd);
 
 	genesis_reset(hw, port);