linux-toradex.git/drivers/net/bonding, branch v2.6.31.3

bonding: clean muticast addresses when device changes type

2009-07-17T01:05:23+00:00

Bonding device forbids slave device of different types under the same
master.

However, it is possible for a bonding master to change type during its
lifetime.  This can be either from ARPHRD_ETHER to ARPHRD_INFINIBAND
or the other way arround.  The change of type requires device level
multicast address cleanup because device level multicast addresses
depend on the device type.

The patch adds a call to dev_close() before the bonding master changes
type and dev_open() just after that.

In the example below I enslaved an IPoIB device (ib0) under
bond0. Since each bonding master starts as device of type ARPHRD_ETHER
by default, a change of type occurs when ib0 is enslaved.

This is how /proc/net/dev_mcast looks like without the patch

5    bond0           1     0     00ffffffff12601bffff000000000001ff96ca05
5    bond0           1     0     01005e000116
5    bond0           1     0     01005e7ffffd
5    bond0           1     0     01005e000001
5    bond0           1     0     333300000001
6    ib0             1     0     00ffffffff12601bffff000000000001ff96ca05
6    ib0             1     0     333300000001
6    ib0             1     0     01005e000001
6    ib0             1     0     01005e7ffffd
6    ib0             1     0     01005e000116
6    ib0             1     0     00ffffffff12401bffff00000000000000000001
6    ib0             1     0     00ffffffff12601bffff00000000000000000001

and this is how it looks like after the patch.

5    bond0           1     0     00ffffffff12601bffff000000000001ff96ca05
5    bond0           1     0     00ffffffff12601bffff00000000000000000001
5    bond0           1     0     00ffffffff12401bffff0000000000000ffffffd
5    bond0           1     0     00ffffffff12401bffff00000000000000000116
5    bond0           1     0     00ffffffff12401bffff00000000000000000001
6    ib0             1     0     00ffffffff12601bffff000000000001ff96ca05
6    ib0             1     0     00ffffffff12401bffff00000000000000000116
6    ib0             1     0     00ffffffff12401bffff0000000000000ffffffd
6    ib0             2     0     00ffffffff12401bffff00000000000000000001
6    ib0             2     0     00ffffffff12601bffff00000000000000000001

Signed-off-by: Moni Shoua 
Signed-off-by: Jay Vosburgh 
Signed-off-by: David S. Miller

bonding: initialization rework

2009-06-14T06:29:04+00:00

Need to rework how bonding devices are initialized to make it more
amenable to creating bonding devices via netlink.

Signed-off-by: Stephen Hemminger 
Signed-off-by: David S. Miller

bonding: use is_zero_ether_addr

2009-06-14T06:29:03+00:00

Remove bogus non-portable possibly unaligned way of testing
for zero addres..

Signed-off-by: Stephen Hemminger 
Signed-off-by: David S. Miller

bonding: network device names are case sensative

2009-06-14T06:29:01+00:00

The bonding device acts unlike all other Linux network device functions
in that it ignores case of device names. The developer must have come
from windows!

Cleanup the management of names and use standard routines where possible.
Flag places where bonding device still doesn't work right with network
namespaces.

Signed-off-by: Stephen Hemminger 
Signed-off-by: David S. Miller

bonding: elminate bad refcount code

2009-06-14T06:29:00+00:00

The "expected_refcount" stuff in bonding sysfs module is a mistake.
Sysfs does proper refcounting, and it is okay to remove a bond device
that has some user process holding the file open.

Signed-off-by: Stephen Hemminger 
Signed-off-by: David S. Miller

bonding: fix style issues

2009-06-14T06:28:57+00:00

Resolve some of the complaints from checkpatch, and remove "magic emacs format"
comments, and useless MODULE_SUPPORTED_DEVICE(). But should not
change actual code.

Signed-off-by: Stephen Hemminger 
Signed-off-by: David S. Miller

bonding: fix destructor

2009-06-14T06:28:56+00:00

It is not safe to use a network device destructor that is a function in
the module, since it can be called after module is unloaded if sysfs
handle is open.

When eventually using netlink, the device cleanup code needs to be done
via uninit function.

Signed-off-by: Stephen Hemminger 
Signed-off-by: David S. Miller

bonding: remove bonding read/write semaphore

2009-06-14T06:28:54+00:00

The whole read/write semaphore locking can be removed. It doesn't add any
protection that isn't already done by using the RTNL mutex properly.

Signed-off-by: Stephen Hemminger 
Signed-off-by: David S. Miller

bonding: initialize before registration

2009-06-14T06:28:52+00:00

Avoid a unnecessary carrier state transistion that happens when device
is registered.
Lockdep works better if initialization is done before registration as well.

Signed-off-by: Stephen Hemminger 
Signed-off-by: David S. Miller

bonding: bond_create always called with default parameters

2009-06-14T06:28:51+00:00

bond_create() is always called with same parameters so move the argument
down.

Signed-off-by: Stephen Hemminger 
Signed-off-by: David S. Miller