Boris Popov writes:
> > I took a look through a whole bunch of kernel code (hoping to do a MFC on my
> > own -STABLE machines), but it looks like the fix for this will involve much
> > of the code to be updated to revs that include "First round implementation
> > of a fine grain enhanced module to
> > If I did kldload if_rl it would pick up the kingston but not the
> > realtek, if I loaded if_dc it would pick up the kingston and
> > then I would try to load if_rl and it would tell me its already
> > loaded and wouldn't pick up the realtek, but I could not get
> > both nic's to work until I
> loaded statically in the kernel, and the others are loaded dynamically.
The
> dynamic loader doesn't see that the module is already present, and tries
to
> load it again. (For PPPoE, the problem manifests when you only have
> 'options netgraph' in your kernel. This forces ppp to try and load
> I did not have to statically compile pppoe support in the kernel. Here
are
> the modules I loaded
>
> ng_pppoe
> ng_ppp
> ng_socket
> ng_ether
> netgraph
>
> it all seemed to work good.
Did you kldload all of these from the command prompt? The problem seems to
crop up when one (or more) mod