> -----Original Message-----
> From: Jakub Kicinski <k...@kernel.org>
> Sent: Wednesday, April 29, 2020 9:39 AM
> To: Keller, Jacob E <jacob.e.kel...@intel.com>
> Cc: Jiri Pirko <j...@resnulli.us>; da...@davemloft.net;
> netdev@vger.kernel.org;
> kernel-t...@fb.com
> Subject: Re: [PATCH net-next] devlink: let kernel allocate region snapshot id
>
> On Wed, 29 Apr 2020 15:34:30 +0000 Keller, Jacob E wrote:
> > > How the output is going to looks like it this or any of the follow-up
> > > calls in this function are going to fail?
> > >
> > > I guess it is going to be handled gracefully in the userspace app,
> > > right?
> >
> > I'm wondering what the issue is with just waiting to send the
> > snapshot id back until after this succeeds. Is it just easier to keep
> > it near the allocation?
>
> I just wasn't happy with the fact that the response send may fail.
> So it just seems like better protocol from the start to expect user
> space to pay attention to the error code at the end, before it takes
> action on the response.
Ok that seems reasonable to me.
Can we get documentation updates for this?
Thanks,
Jake
