Hi Steve, Steven Stallion píše v Čt 09. 07. 2009 v 13:36 -0500: > On Thu, 09 Jul 2009 16:26:55 +0200, Milan Jurik <[email protected]> > wrote: > > Nothing against it if we have some driver available. There are not many > > tulip cards flying around probably (I have one or two which are used for > > testing before RTIs) but it is also used as virtual device in one > > virtualization tool (Hyper-V), so it is still important driver. > > > > The question is - how long will it take to have good replacement? Will > > it be more stable and in better quality? > > > > One thing I never understood - did you dnet updates brought new > > regressions or were you trying to solve all dnet bugs (which is not > > possible probably)? > > No regressions at all - in fact, a number of existing regressions were > fixed ;) > > I was trying to get all of the bugs worked out as best I could - some of > those were contributing to occaisional vlan and aggr failures in NICDRV. >
Tulip cards were and are buggy. That is reality. You can create simpler driver probably which will work on non-buggy tulip card (does such exist these days?) and Hyper-V. But then somebody could complain "my card stopped to work and it worked with dnet" and it can bring the old workarounds back. And our actual set of cards is limited for which we can do testing. Please, integrate your work, do not send it to trash. We can do additional testing on my card and some Hyper-V installation also if you want. I hope your testing should satisfy CRT advocate. And then you can work on new, great, clean implementation. But do not EOL dnet till time you will have good-enough replacement. For now there were mentioned 2 different starting points but no-one seems to be ready now. Best regards, Milan _______________________________________________ driver-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/driver-discuss
