On Wed, 2009-06-10 at 02:53 -0400, Dale Ghent wrote: > On Jun 10, 2009, at 1:52 AM, Garrett D'Amore wrote: > > Solaris suffers from ETOOMANYTUNABLES. Canonizing more of them is > > not the right way forward. > > Sure, I can buy that sentiment. How about a read-only representation > of whether some offload feature is even present? I think we allow for > read-only properties, right?
This seems like something that has been discussed in the past in the networking community; a method of discovering the set of capabilities (for some definition of "capabilities") implemented by the driver under a given link. Such a mechanism could have value to various consumers, and is probably worth investigating outside the context of this case. -Seb
