Re: [dpdk-dev] [PATCH v2] doc: announce ABI change for rte_eth_dev structure

[Jerin Jacob <jerin.jacob@caviumnetworks.com>]

On Thu, Jul 28, 2016 at 01:01:16PM +0000, Ananyev, Konstantin wrote:
> 
> 
> > > >
> > > > Not according to proposal. It can't be too as application has no
> > > > idea what PMD driver does with "prep" what is the implication on a
> > > > HW if application does not
> > >
> > > Why application writer wouldn't have an idea?
> > > We would document what tx_prep() supposed to do, and in what cases user don't need it.
> > 
> > But how he/she detect that on that run-time ?
> 
> By the application logic for example.
> If let say is doing the l2fwd for that group of packets, it would know
> that it doesn't need to do tx_prep().
> 
> To be honest, I don't understand what is your concern here.
> That proposed change doesn't break any existing functionality,
> doesn't introduce any new requirements to the existing API, 
> and wouldn't introduce any performance regression for existing apps.

Yes for the existing application but no for ANY application that uses tx_prep() in future,
that run on the PMD where callback is NULL(one/two PMDs vs N PMDs)

> It is a an extension, and user is free not to use it, if it doesn't fit his needs.

If it is a workaround for a specific HW then why to change the normative
"fastpath" ethdev specification. You could give your fixup as internal
PMD driver routine and be done with it. It is as simple as that.

> From other side there are users who are interested in that functionality,
> and they do have use-cases for  it.
> So what worries you?
Above things worries me, I wouldn't have cared if the changes are not comes
in fastpath and I don't think this sort of issues will never get fixed any time
soon in this community.

So I given up.

Jerin

> Konstantin
> 
> > 
> > > Then it would be up to the user:
> > > - not to use it at all (one segment per packet, no HW TX offloads)
> > 
> > We already have TX flags for that
> > 
> > > - not to use tx_prep(), and make necessary preparations himself,
> > >   that what people have to do now.
> > > - use tx_prep()
> > >
> > > Konstantin
> > >