Re: [PATCH net-next 0/3] nfp: support VF multi-queues configuration

[Leon Romanovsky <leon@kernel.org>]

On Thu, Oct 27, 2022 at 02:11:55AM +0000, Yinjun Zhang wrote:
> On Wed, 26 Oct 2022 15:22:21 +0100, Saeed Mahameed wrote:
> > On 25 Oct 11:39, Yinjun Zhang wrote:
> > >On Date: Tue, 25 Oct 2022 12:05:14 +0100, Saeed Mahameed wrote:

<...>

> > >```
> > >another is from queue's perspective, several class is supported, not very
> > flexible:
> > >```
> > >pci/xxxx:xx:xx.x:
> > >  name max_q_class size 128 unit entry
> > >    resources:
> > >      # means how many VFs possess max-q-number of 16/8/..1 respectively
> > >      name _16 size 0 unit entry size_min 0 size_max 128 size_gran 1
> > >      name _8 size 0 unit entry size_min 0 size_max 128 size_gran 1
> > >      ...
> > >      name _1 size 0 unit entry size_min 0 size_max 128 size_gran 1
> > >```
> > 
> > weirder.
> 
> Yes, kind of obscure. The intention is to avoid configuring one by one, especially
> when there're thousands of VFs. Any better idea is welcomed.

In parallel to netdev world, we extended netlink UAPI to receive ranges
and gave an option for users to write something like this: "cmd ... 1-100 ...",
which in kernel was translated to loop from 1 to 100.

Of course, we are talking about very specific fields (IDs) which can receive range.

It is just an idea how devlink can be extended to support batch configuration
of same values.

Thanks