I don't have time to do a code review, but I have comments on the nwamd README
file:
- expand acronyms -- what's ENM, NCU, NCP, ...?
Or tell readers to read libnwam's README first.
- 256-257:
256 ... To accommodate this,
257 WiFi links of the same or lower (more preferred) priority group will always
258 be trying to connect (and thus be offline* if they cannot).
Trying to connect... actively, or passively (i.e., by scanning for
them)?
The difference is important. I would not want nwamd to leak
information about the user's laptop.
- 142-153
Your description of doors is not really correct, and anyways, also
not necessary. I'd say: remove it.
("The way doors work is to provide a mechanism for another process to
execute code in your process space" is right, on some level, but it
gives the impression that one can make a door server execute
arbitrary code.
A better description would be: Solaris doors IPC is a method of IPC
that resembles RPC, and, most importantly, is synchronous, which
allows the system to switch contexts from the door calling thread to
a server thread directly, without passing through a scheduler queue.)
- Design comment: if NWAM object states closely resemble SMF service
states... why not have NWAM NCU/NCP/Location/WLAN FMRIs? Eventually
SMF services could even depend on these...
- Is the fmri arg in nwam_enm_create() the FMRI of the ENM object,
or an SMF FMRI that is the external network modifier?
- SMF is not mentioned as a source of events. Should it be?
Nico
--
_______________________________________________
networking-discuss mailing list
[email protected]
