FWD under bridged enviroment...

[Luigi Rizzo]

On Tue, Aug 17, 2004 at 06:07:25PM -0300, Patrick Tracanelli wrote:
...
> Btw, (maybe Rizzo can comment) what's the reason it has not been 
> commited when the patch was posted? :) It's a feature that is usually 

as probably Kelly pointed out,
my patch had a locking bug -- bridging works at
splimp(), and FWD plays with data structures that are only
protected by splnet(), i.e. a layer2-fwd packet might interrupt
a section of code running at splnet() and modify the
data structures.

Kelly's solution was to go through the ipintrq.

	cheers
	luigi