> Perhaps this should be a key item to identify for drivers during code > review -- use of anything other than the mc_setprop()/getprop() API > should require justification. If the problem is a driver defect, then > it should be fixable. (If the problem is a requirement to retrieve the > property at start-of-day, during attach, because the hardware needs it > to initialize properly... that's a different case. Hopefully there > aren't too many of those around.)
Perhaps. This is getting into a gray area, where we're like "here's the ring, Frodo, but", you know, hopefully being increasingly open (source) about all things Solaris would help. -Artem _______________________________________________ networking-discuss mailing list [email protected]
