Re: [PATCH iproute2-next v2] tc: add skip_hw and skip_sw to control action offload

[Roi Dayan <roid@nvidia.com>]

On 2022-02-16 2:51 PM, Jamal Hadi Salim wrote:
> On 2022-02-11 05:01, Baowen Zheng wrote:
>> Hi Jamal:
>> Sorry for the delay of the reply.
>>
> 
> I guess it is my turn to say sorry for the latency ;->
> 
>> On February 2, 2022 7:47 PM, Jamal wrote:
>>> On 2022-02-02 04:37, Baowen Zheng wrote:
>>>> Hi Roi:
>>>> Thanks for bring this to us, please see the inline comments.
>>>>
> 
> [..]
>>>
>>> Probably the language usage is causing the confusion and I missed 
>>> this detail
>>> in the output as well. Let me see if i can break this down.
>>>
>>> Either both action and  filter are in h/w or they are not. i.e
>>>
>>> action in h/w  + filter in h/w == GOOD
>>> action in h/w  + filter in s/w == BAD
>>> action in s/w  + filter in h/w == BAD
>>> action in s/w  + filter in s/w == GOOD
>>>
>>> The kernel patches did have those rules in place - and Baowen added 
>>> tdc tests
>>> to check for this.
>>>
>>> Now on the workflow:
>>> 1) If you add an action independently to offload before you add a 
>>> filter when
>>> you dump actions it should say "skip_sw, ref 1 bind 0"
>>> i.e information is sufficient here to know that the action is 
>>> offloaded but there
>>> is no filter attached.
>>>
>>> 2) If you bind this action after to a filter which _has to be offloaded_
>>> (otherwise the filter will be rejected) then when you dump the 
>>> actions you
>>> should see "skip_sw ref 2 bind 1"; when you dump the filter you 
>>> should see
>>> the same on the filter.
>>>
>>> 3) If you create a skip_sw filter without step #1 then when you dump you
>>> should see "skip_sw ref 1 bind 1" both when dumping in IOW, the 
>>> not_in_hw
>>> is really unnecessary.
>>>
>>> So why not just stick with skip_sw and not add some new language?
>>>
>> If I do not misunderstand, you mean we just show the skip_sw flag and 
>> do not show other information(in_hw, not_in_hw and in_hw_count), I 
>> think it is reasonable to show the action information as your 
>> suggestion if the action is dumped along with the filters.
>>
> 
> Yes, thats what i am saying - it maintains the existing semantics people
> are aware of for usability.
> 
>> But as we discussed previously, we added the flags of skip_hw, 
>> skip_sw, in_hw_count mainly for the action dump command(tc -s -d 
>> actions list action xxx).
>> We know that the action can be created with three flags case: skip_sw, 
>> skip_hw and no flag.
>> Then when the actions are dumped independently, the information of 
>> skip_hw, skip_sw, in_hw_count will become important for the user to 
>> distinguish if the action is offloaded or not.
>>
>> So does that mean we need to show different item when the action is 
>> dumped independent or along with the filter?
>>
> 
> I see your point. I am trying to visualize how we deal with the
> tri-state  in filters and we never considered what you are suggesting.
> Most people either skip_sw or skip_hw in presence of offloadable hw.
> In absence of hardware nobody specifies a flag, so nothing is displayed.
> My eyes are used to how filters look like. Not sure anymore tbh. Roi?
> 

Hi,

Is the question here if to show different information
when actions are dumped independently or with a filter?

then I think yes. when actions are dumped as part of the filter
skip showing skip_sw/skip_hw/in_hw/not_in_hw flags as it's redundant and
it's always whatever the filter state is.

I also noticed we can improve extack msgs when a user will try to mix
the state like adding a filter without skip_hw flag but use action index
that is created with skip_hw.
I noticed currently there is no informative extack msg back to the user.

Thanks,
Roi


> cheers,
> jamal
> 
>