Re: [PATCH RESEND] netvsc: transfer lower device max tso size

[Stephen Hemminger]

On Mon, 28 Jul 2025 08:19:07 -0700
Jakub Kicinski <kuba@xxxxxxxxxx> wrote:

> On Sun, 27 Jul 2025 20:01:26 -0700 Stephen Hemminger wrote:
> > On Wed, 23 Jul 2025 15:16:22 -0700
> > Jakub Kicinski <kuba@xxxxxxxxxx> wrote:  
> > >  
> > > > Actually, we had used the common bonding driver 9 years ago. But it's
> > > > replaced by this kernel/netvsc based "transparent" bonding mode. See
> > > > the patches listed below.
> > > > 
> > > > The user mode bonding scripts were unstable, and difficult to deliver
> > > > & update for various distros. So Stephen developed the new "transparent"
> > > > bonding mode, which greatly improves the situation.      
> > > 
> > > I specifically highlighted systemd-networkd as the change in the user
> > > space landscape.    
> > 
> > Haiyang tried valiantly but getting every distro to do the right thing
> > with VF's bonding and hot plug was impossible to support.  
> 
> I understand, but I also don't want it to be an upstream Linux problem.
> 
> Again, no other cloud provider seems to have this issue, AFAIU.

The problem is that other cloud providers don't expose the VF, the hide it in HW firmware.
The userspace world is a mess, with systemd, netplan, cloud init, and the SuSe stuff.
And custom appliances that assume that there is a default eth0 device on boot.
Yes, there were users that expect to see eth0 all the time.