On Tue, Apr 22, 2014 at 12:27 PM, Stefan Baur <[email protected]> wrote: > > If Adam is willing to donate his spare card to you dev folks, and maybe > Volker buys a Gold Membership (in case he doesn't have one already), > would that significantly increase the chances of having a working hme > driver in a future release? :-) >
It's just not practical. Kernel developers are expensive, in order for it to make any sense for us to put one of our developers' time towards it, it'd require more money than a brand new Intel quad port gig NIC would cost that'll work right out of the box. On ebay you can find a slew of used quad port gigabit Intel cards for under $100 USD for PCI-e and under $50 for PCI-X, including shipping. You could buy a stock pile of those cheaper than fixing a driver issue. The economics of fixing a driver issue on a 16 year old NIC just aren't there unless you can find a driver developer who's willing to do all the work for free, or you have a significant number of them in production where it becomes financially viable. There may be hope in reporting it upstream and hoping for the best, but don't hold your breath. It may also already be fixed in FreeBSD 10.x, so trying 2.2 snapshots first would be worthwhile. That'd be a requirement anyway before anyone upstream would pay any attention. _______________________________________________ List mailing list [email protected] https://lists.pfsense.org/mailman/listinfo/list
