On Mon, Mar 11, 2019 at 6:22 PM Ferruh Yigit <ferruh.yi...@intel.com> wrote:
> On 3/4/2019 11:18 AM, David Marchand wrote: > > According to the api, the q_errors[] per queue statistic is for reception > > errors not transmit errors. > > This is a first cleanup on statistics before looking at oerrors. > > > > Yes, the patchset looks aligned with the API documentation [1]. > > What can be the solution after cleanup? We can merge this cleanup and > solution > next to each-other to not leave a gap? > 1- Different variables for Rx and Tx errors? > 2- Combine Rx & Tx into this single variable? > > It can be good to find a solution because new PMDs doing same mistake > because of > copy/paste... > Might not be feasible but how about we could introduce an internal stats structure containing the needed field for tx. pmd would use it but ethdev would translate it to the current exposed api rte_eth_dev_stats_get ? The additional field would be formatted by ethdev to be provided through the xstats api. -- David Marchand
