[dpdk-dev] [PATCH v2 1/2] ethdev: add buffered tx api

[Kulasek, TomaszX]

> -----Original Message-----
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> Sent: Wednesday, March 9, 2016 18:07
> To: Kulasek, TomaszX <tomaszx.kulasek at intel.com>
> Cc: dev at dpdk.org; Ananyev, Konstantin <konstantin.ananyev at intel.com>
> Subject: Re: [dpdk-dev] [PATCH v2 1/2] ethdev: add buffered tx api
> 
> 2016-03-09 16:35, Kulasek, TomaszX:
> > From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> > > > +void
> > > > +rte_eth_count_unsent_packet_callback(struct rte_mbuf **pkts,
> > > > +uint16_t
> > > unsent,
> > > > +		void *userdata);
> > >
> > > What about rte_eth_tx_buffer_default_callback as name?
> >
> > This function is used now as default to count silently dropped packets
> and update error counter in tx_buffer structure. When I remove error
> counter and set silent drop as default behavior, it's better to have two
> callbacks to choice:
> >
> > 1) silently dropping packets (set as default)
> > 2) as defined above to dropping with counter.
> >
> > Maybe better is to define two default callbacks while many
> > applications can still update it's internal error counter, So IHMO these
> names are more descriptive:
> >
> > rte_eth_tx_buffer_drop_callback
> > rte_eth_tx_buffer_count_callback
> >
> > What you think?
> 
> I think you are right about the name.
> 
> Are you sure providing a "count" callback is needed?
> Is it just to refactor the examples?

I think it's useful to have a callback which let you easily to track the overall number of packets dropped. It's handy when you want to drop packets and not leave them untracked.

It's good to have it, but it's not critical.

Changing the examples is not a problem while I've got copy-paste superpower.

Tomasz