Nicolas Droux wrote:
Sebastien Roy wrote:
the mac_handle_t would be passed through by the plugin, and the
notification would make it up to the MAC clients.
I don't understand how this can be made to work. How does the MAC
notification get translated to some DLPI notification? How do MAC
clients in the GLDv3 framework know what MAC notifications to register
for if such notifications are completely transparent to the framework
Upper layers (dld and up) would still have to be changed to properly
process the new notifications. However the rest of the framework could
be left untouched.
That's true.
My concern with the current proposal is that new plugin types might
require new entry points to be added to the MAC driver and client
interfaces of the common framework.
That's true, because there may be things that can be viewed as common
that are missing from the common framework. I don't see destination
addresses as plugin specific. They are no more plugin specific than
broadcast addresses or source addresses.
Sure. Given the limited number of plugins that will be implemented and
since there are no plans to make the plugin interface public anytime
soon, I'm fine if this issue is put on the back burner for now.
Okay.
Thanks for taking the time to review this, your comments have resulted
in some good changes.
-Seb
_______________________________________________
networking-discuss mailing list
[email protected]
