[dpdk-dev] [PATCH 1/2] net/i40e: queue region set and flush

[Ferruh Yigit]

On 9/20/2017 4:20 AM, Zhao1, Wei wrote:
> Hi, Ferruh
> 
>> -----Original Message-----
>> From: Yigit, Ferruh
>> Sent: Friday, September 15, 2017 7:01 PM
>> To: Zhao1, Wei <wei.zhao1 at intel.com>; dev at dpdk.org
>> Subject: Re: [dpdk-dev] [PATCH 1/2] net/i40e: queue region set and flush
>>
>> On 9/6/2017 10:11 AM, Ferruh Yigit wrote:
>>> On 9/1/2017 3:38 AM, Zhao1, Wei wrote:
>>>> HI,  Ferruh
>>>>
>>>>> -----Original Message-----
>>>>> From: Yigit, Ferruh
>>>>> Sent: Friday, September 1, 2017 12:18 AM
>>>>> To: Zhao1, Wei <wei.zhao1 at intel.com>; dev at dpdk.org
>>>>> Subject: Re: [dpdk-dev] [PATCH 1/2] net/i40e: queue region set and
>>>>> flush
>>>>>
>>>>> On 8/24/2017 4:26 AM, Wei Zhao wrote:
>>>>>> This feature enable queue regions configuration for RSS in PF/VF,
>>>>>> so that different traffic classes or different packet
>>>>>> classification types can be separated to different queues in
>>>>>> different queue regions.This patch can set queue region range, it
>>>>>> include queue number in a region and the index of first queue.
>>>>>> This patch enable mapping between different priorities (UP) and
>>>>>> different traffic classes.It also enable mapping between a region
>>>>>> index and a sepcific flowtype(PCTYPE).It also provide the solution
>>>>>> of flush all configuration about queue region the above described.
>>>>>>
>>>>>> Signed-off-by: Wei Zhao <wei.zhao1 at intel.com>
>>>>>
>>>>> Hi Wei,
>>>>>
>>>>> I wonder if this can be implemented using rte_flow, instead of PMD
>>>>> specific API?
>>>>>
>>>>> And if not, what is missing in rte_flow for this?
>>>>
>>>> Yes, at first I have plan to use rte_flow to implement this feature.
>>>> But there are many opposition for this, for example, this  feature is
>>>> only support by i40e but not igb/ixgbe From hardware capacity, not all NIC.
>> So, private API is selected.
>>>
>>> I guess rte_flow headers needs to be updated for this support, how big
>>> that update, what is missing in rte_flow for this?
>>
>> Hi Wei,
>>
>> Would you mind answering this?
> 
> If we want rte_flow also to support this feature,
> although there will be many new CLI command need to be Add to testpmd app,
> and some code change in ethdev, for example  add new member in enum rte_filter_type and so on.

new CLI part is side effect of the rte_flow, to demonstrate how to use
flow modes. I was more interested in how much change required in
rte_flow header.

> But the biggest problem is to get recognize and support from expert and authority in DPDK community for this new change.

Is this is referring previous push back for rte_filter_type update, I
think this can be evaluated case by case. As far as I remember
previously it is rejected in favor of improving rte_flow, this time
update is required to be able to use rte_flow, so I assume can be OK.
I believe this can be discussed out of this patch scope.

> By now, rte_flow has no support for rss config application, but only filter configuration. But maybe this can be a work in next release.

At least it is good that this effort can be re-used later for rte_flow,
I still believe this option worth investigating for further releases.

> 
> 
>>
>> Thanks,
>> ferruh
>>
>>>
>>> Even this is only for i40e, rte_flow can be an option. I believe this
>>> increases the usability of the feature for the driver.
>>>
>>>>
>>>>>
>>>>> Thanks,
>>>>> ferruh
>>>
>