On 7 January 2016 at 05:53, Ashok Kumar <ash...@broadcom.com> wrote: > On Wed, Jan 06, 2016 at 06:10:15PM +0000, Peter Maydell wrote: >> I guess from the patch that this is adding support for >> the legacy '-net' way of configuring networking, but do >> we need that if we never supported it in the first place? >> (If virt is the only PCI machine which doesn't support >> -net syntax that would probably be a strong argument for >> supporting it.) > Fine with me. But there are some documentation for e.g [1] which says "-net" > is > still supported. > > [1] http://www.linux-kvm.org/page/Networking#Compatibility
I've cc'd the networking maintainer and some other people who might have an opinion. Basically the question is "should we support the legacy -net syntax for new board models which didn't exist back when -net was the only option and have never supported -net" ? The argument for "don't support it" is that -net is legacy and if there aren't any legacy command lines to support then we should avoid extending the scope of legacy behaviour. The argument for "do support it" is consistency -- we don't document that -net is only for certain boards, and users will expect that network config options that work on some machines (including x86 PC and many of the ARM embedded board models) will also work on the ARM virt board. I think my current inclination is to say that virt should support -net, because I would prefer to avoid having yet another speedbump in the path of people trying to move to using KVM-on-ARM based on their previous experience with KVM-on-x86. thanks -- PMM