Re: [dpdk-dev] [PATCH] app/testpmd: support flow aging

[Ferruh Yigit <ferruh.yigit@intel.com>]

On 4/27/2020 4:12 PM, Matan Azrad wrote:
> Hi
> 
> From: Ferruh Yigit
>> On 4/26/2020 8:23 AM, Bill Zhou wrote:
>>>
>>>
>>>> -----Original Message-----
>>>> From: Ferruh Yigit <ferruh.yigit@intel.com>
>>>> Sent: Saturday, April 25, 2020 12:25 AM
>>>> To: Bill Zhou <dongz@mellanox.com>; wenzhuo.lu@intel.com;
>>>> jingjing.wu@intel.com; bernard.iremonger@intel.com; Ori Kam
>>>> <orika@mellanox.com>
>>>> Cc: dev@dpdk.org
>>>> Subject: Re: [dpdk-dev] [PATCH] app/testpmd: support flow aging
>>>>
>>>> On 4/24/2020 11:55 AM, Bill Zhou wrote:
>>>>> Currently, there is no way to check the aging event or to get the
>>>>> current aged flows in testpmd, this patch include those implements,
>>>>> it's
>>>> included:
>>>>> - Registering aging event based on verbose level, when set verbose > 0,
>>>>>   will register this event, otherwise, remove this event. In this event
>>>>>   only dump one line of log to user there is one aging event coming.
>>>>> - Add new command to list all aged flows, meanwhile, we can set
>>>> parameter
>>>>>   to destroy it.
>>>>
>>>> Can you please document new feature and command?
>>>>
>>>> Instead of overloading the 'verbose', what do you think having
>>>> explicit command to register aging events? ("flow aged register
>>>> <port_id>"?) I think many of the verbose usage won't really interest in
>> the flow aging.
>>>>
>>>
>>> Yes, some of the verbose usage indeed not interest in the flow aging.
>>> But If we use register or unregister event to one port, sometime, it
>>> will repeat many times for every ports.
>>> What do you think if we register this event all the time, and
>>> introduce new global var from command to control the event export to
>> application ?
>>>    for example: set aged_flow_event_print_en [0 | 1]
>>
>> I am not also sure about registering this event always, you can register for all
>> ports with same command, as we do in many commands:
>> "flow aged register <port_id>|all"
>>
>> When argument is "all" it registers for all ports, when it is a specific port_id,
>> registers for that port id.
>> And it is good to have 'unregister' command too.
> 
> As I can see in testpmd default, it all the time registers to all events..
> See register_eth_event_callback.

You are right.

> 
> I suggest to catch the age event in callback and print it only when set aged_flow_event_print_en is on.

+1

> 
> 
> Matan
> 
> 
>>
>>>
>>>>>
>>>>> Signed-off-by: Bill Zhou <dongz@mellanox.com>
>>>>
>>>> <...>
>