Dennis Clarke wrote: > n.b.: topic changed > >>> It is when you consider that pcn does not work and the ae driver does. >> As Alan mentioned, a number of folks have been looking into "ae". One >> issue that several folks identified (at least on the Indiana prototype) >> is that "ae" seemed to lock up on bulk transfers while the "pcn" driver >> did not: >> >> http://mail.opensolaris.org/pipermail/indiana-discuss/2008-February/004712.html >> http://mail.opensolaris.org/pipermail/indiana-discuss/2008-February/004714.html >> http://mail.opensolaris.org/pipermail/indiana-discuss/2008-February/004722.html > > I see nothing in those messages that refers to real hardware. I think that > Virtual Box should be regarded in the same light as VMware; not real. If I > have the real thing, actual PCI network cards from Hewlett Packard with the > AMD PCNet chips on them and they fail to work with the pcn driver but work > perfectly with ae then I think my test reports and bug reports will carry > more validity than virtual non-real hardware. >
Perhaps, except that the principal "driver" for making this substitution is because it's what VirtualBox emulates, and there'll be a lot more users of the driver that way than in real hardware these days. Perhaps the problem is VirtualBox's, which wouldn't be all that surprising, but getting to the bottom of it would be useful so we could make an informed decision about what to do. Dave _______________________________________________ opensolaris-discuss mailing list [email protected]
