Garrett, First round of hardware testing is complete, and I am putting the finishing touches on the link notification bits for dnet.
Something I noticed while checking dladm is that the ifspeed kstat for dnet was always reporting 0 for this reference board. Upon further inspection, it seems that dnetp->speed is only available if set via driver properties for non-mii phy's. (At this point I am assuming that mii phy's behave correctly) Essentially the question is this: should this behavior be fixed now or should it wait until after this first putback? If this is to be fixed now, I can make a modification to the dnet_periodic cyclic function which was added for link notification (it would be pretty trivial and involves checking the same registers needed for the link notification check). Also, when checking for link notification (and potentially link speed) would an mii phy affect how I would check for these values? Essentially for the 21143, it is a simple poll on the CSR5 and CSR12 registers. Regards, Steve -- Yet magic and hierarchy arise from the same source, and this source has a null pointer. Reference the NULL within NULL, it is the gateway to all wizardry. _______________________________________________ driver-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/driver-discuss
