On (10/30/08 15:37), James Carlson wrote:
> Part of the Rapid Spanning Tree protocol requires information about
> the duplex setting on the ports attached to the bridge.  To get this
> information, I have at least two choices: one is to go through
> libdladm to get the "duplex" property, which is unfortunately a text
> string I have to parse, or I can go through kstats to get the integer
> value.
> 
> What's the future direction here?  Is relying on libdladm (despite its
> warts) the better way to go?  Are kstats going to stay in place?
> 
> (I'm planning to use libdladm despite the problems, but I'm curious
> about the longer-term direction as well.)

that's a well-timed question.

I was actually working on cleaning up some of the interfaces 
surrounding 'dladm show-ether' and ran into a similary issue
with getting link state. 

See
  http://cr.opensolaris.org/~sowmini/show-ether-cleanup/

(Review comments invited!)

So this presents a couple of choices:
- use the dladm_ether_info() interface and use the returned 
  le_attr[CURRENT] to get the info

and/or

- we could provide a separate set of libdladm interfaces to get 
  the link_duplex_t info (but then we would have to do something similar
  for every property, in the interest of consistency).

thoughts? 

For your immediate need, the webrev posted may suffice..

--Sowmini





Reply via email to