On Wed, Feb 19, 2025 at 06:06:51PM -0800, Jakub Kicinski wrote:
> On Wed, 19 Feb 2025 08:39:08 +0100 Kurt Kanzenbach wrote:
> > > My comment from the previous series still stands, which simply that
> > > I have no idea if the maintainers will accept changes using this API
> > > or prefer to wait until Stanislav's work [1] is completed to remove
> > > the RTNL requirement from this API altogether.
> >
> > I'd rather consider patch #2 a bugfix to restore the busy polling with
> > XDP/ZC. After commit 5ef44b3cb43b ("xsk: Bring back busy polling
> > support") it is a requirement to implement this API.
> >
> > The maintainers didn't speak up on v1, so i went along and sent v2.
> >
> > @Jakub: What's your preference? Would you accept this series or rather
> > like to wait for Stanislav's work to be finished?
>
> No strong preference. If rtnl_lock is not causing any issues
> in this driver, the we can merge as is. I haven't followed
> the past discussions, tho.
Don't mean to side-track this thread, but does this mean you've
changed your mind on the previous virtio_net thread [1] ?
Or maybe I'm just misreading your response there? And instead I
could re-spin the virtio_net but dropping the first patch and
dealing with RTNL in the code like this series is doing?
For some reason I was under the impression that the virtio_net
series and others like it (like this igb series) were being held
back until locking work Stanislav is doing is done.
[1]: https://lore.kernel.org/netdev/20250127133756.413ef...@kernel.org/
