> Steven Stallion wrote: >>>> Considering that the 8029 has a number of extended features not >>>> included >>>> in the DP8390 spec (i.e. fdx, flow control, etc.) does it make sense >>>> to >>>> scrap generic support for NE2000 devices and develop specifically for >>>> the >>>> 8029? (i.e. renbrand ne as re) ? >>>> >>>> >>>> >>> IMO, it would not be a terrible idea. Generic ne2k support is probably >>> not terribly useful anymore. >>> >> >> It would make life *much* easier. The DP8390 spec is pretty sparse, and >> I >> would have had to add in vendor specific entry points anyway. >> >> I hate to admit it, but I don't like dropping generic chipset support >> (there used to be a number of other PCI ne2k devices out there), however >> if the 8029 is the only device thats deemed useful enough to emancipate, >> theres no sense in carrying along more cruft - especially if it would >> get >> in the way of integration. >> > > Support for additional chipsets would require the ability to *test* > those other parts. And the market demand for such support is > ~nonexistent. > > The 8029 is unique because it is a popular emulated device, so having it > integrated serves a real purpose. > > The other point here is, folks with other DP8390 based products can > probably use Masa's ni driver. (And that driver, because of lack of > popularity/need, is unlikely to ever be integrated if your 8029 work is > integrated.) > > -- Garrett > >
Makes sense - this should clean up the code significantly. I'll start the conversion to re tonight and shoot for full RTL8029/AF support. Steve _______________________________________________ driver-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/driver-discuss
