[dpdk-dev] [PATCH] doc: update new ethdev offload API description

[Bruce Richardson]

On Wed, Mar 21, 2018 at 03:40:43PM +0100, Thomas Monjalon wrote:
> 21/03/2018 15:28, Ferruh Yigit:
> > On 3/21/2018 2:08 PM, Thomas Monjalon wrote:
> > > 21/03/2018 11:54, Ferruh Yigit:
> > >> On 3/21/2018 9:47 AM, Andrew Rybchenko wrote:
> > >>> IMHO, it should be allowed to specify queue offloads on port level.
> > >>> It should simply enable these offloads on all queues. Also it will
> > >>> match dev_info [rt]x_offload_capa which include both port and queue
> > >>> offloads.
> > >>>
> > >>> Yes, we lose possibility to enable on port level, but disable on queue
> > >>> level by suggested changes, but I think it is OK - if you don't need
> > >>> it for all queues, just control separately on queue level.
> > >>
> > >> What I understand was queue offload can only enable more, but it seems it can
> > >> both enable or disable.
> > > 
> > > Yes, queue offload should only enable more.
> > > An offload enabled at port level, cannot be disabled at queue level.
> > 
> > Agree an offload enabled at port level can't be disabled at queue level, but why
> > not have the ability to disable a queue level offload with another queue setup call.
> 
> Yes it should be possible to disable a queue offload:
> 1/ enable offload in queue configuration
> 2/ later disable this offload in queue configuration
> 
> So, when I say "only enable more", I mean queue config should only enable
> more than port config. In other words, a port-level offload cannot be
> disabled at queue level.
> 
> > > A port offload can be repeated in queue configuration.
> > > If a port offload is not repeated in queue configuration, there should be
> > > no impact: it is still in the port configuration, thus applying to all queues.
> > 
> > This was a requirement, this patch targets removing the requirement to repeat
> > the port offload in queue config.
> 
> Understood, and I agree with this change if it is well explained
> in sphinx and doxygen.
> It is important to say that queue configuration has no impact on offloads
> already enabled at port level.
> 
> > > About capabilities, the queue offloads must be a subset of port offloads.
> > > The queue capabilities show which offloads can be enabled per queue.
> 
> 
>

Why not abandon port-level config entirely? Then you just have queue-level
configs, with the restriction that on some NICs all queues must be
configured the same way. It can be up to the NIC drivers - or possibly
ethdev layer - to identify and report an error in such cases.

/Bruce