[PATCH] (0/8) replace brlock with RCU

[Stephen Hemminger]

> estionable).
> 
> It's entirely possible that the current user could be replaced by RCU 
> and/or seqlocks, and we could get rid of brlocks entirely.
> 
> 		Linus

The following sequence of patches replaces the remaining use of brlock
with RCU.  Most of this is fairly straightforward. The unregister functions
use synchronize_kernel(), perhaps there should be a special version to
indicate sychronizing with network BH. 

Tested the normal network path and IPV4. 
For the others all the pieces compile. But don't have the test setup to 
give VLAN, SNAP, IPV6, and all the netfilter code a real test.

Comments?