> -----Original Message-----
> From: Thomas Monjalon <tho...@monjalon.net>
> Sent: Tuesday, March 23, 2021 7:56 PM
> To: P, Venkata Suresh Kumar <venkata.suresh.kuma...@intel.com>;
> Dumitrescu, Cristian <cristian.dumitre...@intel.com>
> Cc: dev@dpdk.org; Khangar, Churchill <churchill.khan...@intel.com>; Jangra,
> Yogesh <yogesh.jan...@intel.com>
> Subject: Re: [dpdk-dev] [PATCH] port: add file descriptor SWX port
>
> 23/03/2021 19:56, Dumitrescu, Cristian:
> > From: Thomas Monjalon <tho...@monjalon.net>
> > > 23/03/2021 19:07, P, Venkata Suresh Kumar:
> > > > Thanks a lot for reviewing the code and providing your comments.
> > > >
> > > > I have addressed below comments in V2 patch.
> > >
> > > OK thanks.
> > >
> > > What about the question about rte_trace?
> > > Opinions?
> >
> > Hi Thomas,
> >
> > All rte_swx_port ports are currently following this pattern,
> > so it makes sense to have this one do the same for now.
> >
> > I am not that familiar with the (relatively new) rte_trace mechanism, so I
> am not sure if it has any run-time performance (I am assuming that it
> doesn't). We will take the AR to take a look at rte_trace and come back with a
> patch to convert traces for all ports to rte_trace, most likely in the 21.05
> time
> frame. Is this OK for you?
>
> Yes, would be great.
> In general we should avoid such #ifdef in DPDK code.
> Maybe rte_trace can help, maybe rte_log levels would suffice.
> Note there is RTE_LOG_DP_LEVEL for logging in datapath.
>
Yes, we will look at rte_trace for 21.05 and come back with a resolution.
Thanks, Thomas!
