Re: [dpdk-dev] [RFC 1/2] ethdev: move queue stats to xstats

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

On 10/12/2020 10:53 PM, Thomas Monjalon wrote:
> 12/10/2020 18:46, Ferruh Yigit:
>> Queue stats are stored in 'struct rte_eth_stats' as array and array size
>> is defined by 'RTE_ETHDEV_QUEUE_STAT_CNTRS' compile time flag.
>>
>> As a result of technical board discussion, decided to remove the queue
>> statistics from 'struct rte_eth_stats' in the long term.
>>
>> Instead PMDs should represent the queue statistics via xstats, this
>> gives more flexibility on the number of the queues supported.
> 
> Yes
> 
>> Currently queue stats in the xstats are filled by ethdev layer, using
> 
> "some basic" queue stats
> 
>> the basic stats, when queue stats removed from basic stats the
>> responsibility to fill the relevant xstats will be pushed to the PMDs.
>>
>> During the switch period, a temporary
>> 'RTE_ETH_DEV_QUEUE_STATS_IN_XSTATS' device flag is created. The PMDs
>> providing the queue stats in the xstats should set this flag to bypass
>> the relevant part in ethdev layer.
>>
>> When all PMDs switch to the xstats for the queue stats, queue stats
>> related fields from 'struct rte_eth_stats' will be removed, as well as
>> 'RTE_ETH_DEV_QUEUE_STATS_IN_XSTATS' flag.
>> Later 'RTE_ETHDEV_QUEUE_STAT_CNTRS' compile time flag also can be
>> removed.
>>
>> Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
>> ---
>> --- a/lib/librte_ethdev/rte_ethdev.h
>> +++ b/lib/librte_ethdev/rte_ethdev.h
>> +/** Device provides queue stats in xstats */
>> +#define RTE_ETH_DEV_QUEUE_STATS_IN_XSTATS 0x0040
> 
> Not exact wording here and in the title.
> We should try to convey the idea that the basic queue stats
> are not automatically converted from the array to some xstats.
> Something like RTE_ETH_DEV_NO_AUTOFILL_QUEUE_XSTATS
> 
> Or we can define the opposite flag and set it by default
> in all drivers which fill the basic queue stats.
> I suggest RTE_ETH_DEV_AUTOFILL_QUEUE_XSTATS
> 

Agree that opposite flag makes it easier to traces the PMDs missing 
implementation, will send another RFC with it.