speedtch testatm.c,1.4,1.5

[Roman Kagan]

On Fri, Apr 08, 2005 at 12:18:32PM +0200, Duncan Sands wrote:
> > Can't it be because you release all interfaces including the one
> > disconnect is called with?
> 
> Sure, but the point is that that's supposed to be OK.

Are you certain?  I rather expected that the number of probe +
claim_interface should equal the number of disconnect +
release_interface...  But I'll try to check with the source.

Roman.