Re: driver uABI review list? (was: Re: [PATCH/RFC net-next 0/3] nfp: support VF multi-queues configuration)

[Leon Romanovsky <leon@kernel.org>]

On Thu, Sep 22, 2022 at 07:37:08PM +0200, Andrew Lunn wrote:
> On Thu, Sep 22, 2022 at 07:04:19PM +0300, Leon Romanovsky wrote:
> > On Wed, Sep 21, 2022 at 06:34:48AM -0700, Jakub Kicinski wrote:
> > > On Tue, 20 Sep 2022 16:14:16 +0100 Simon Horman wrote:
> > > > this short series adds the max_vf_queue generic devlink device parameter,
> > > > the intention of this is to allow configuration of the number of queues
> > > > associated with VFs, and facilitates having VFs with different queue
> > > > counts.
> > 
> > <...>
> > 
> > > Would it be helpful for participation if we had a separate mailing 
> > > list for discussing driver uAPI introduction which would hopefully 
> > > be lower traffic?
> > 
> > Please don't. It will cause to an opposite situation where UAPI
> > discussions will be hidden from most people. IMHO, every net vendor
> > should be registered to netdev mailing list and read, review and
> > participate.
> 
> Good in theory, but how often do you really see it happen?

I agree that the situation in netdev is not ideal, but it can be
improved by slightly changing acceptance criteria. 

As a rough idea (influenced by DRM subsystem), require cross-vendor
review prior merge. It doesn't need to be universal, and can be
applicable only to most active companies. If reviews are not happening
in sensible time frame, there are ways "to punish" vendor that was
asked to review.

Thanks