Re: [dpdk-dev] [PATCH] telemetry: fix error and warning printfs

[Bruce Richardson <bruce.richardson@intel.com>]

On Fri, May 22, 2020 at 10:04:44PM +0530, Jerin Jacob wrote:
> On Fri, May 22, 2020 at 7:21 PM Ciara Power <ciara.power@intel.com> wrote:
> >
> > Initially, printf was used to indicate and error/warning resulting from
> > telemetry initialisation. This is now fixed to use EAL logs for
> > warnings, and the unnecessary printf for an error is removed.
> >
> > Fixes: eeb486f3ba65 ("eal: add telemetry as dependency")
> > Fixes: dd6275a424ac ("telemetry: fix error log output")
> > Cc: david.marchand@redhat.com
> >
> > Signed-off-by: Ciara Power <ciara.power@intel.com>
> > ---
> >  lib/librte_eal/freebsd/eal.c         | 4 +++-
> >  lib/librte_eal/linux/eal.c           | 4 +++-
> >  lib/librte_telemetry/rte_telemetry.h | 9 +++++++++
> >  lib/librte_telemetry/telemetry.c     | 2 --
> >  4 files changed, 15 insertions(+), 4 deletions(-)
> >
> > diff --git a/lib/librte_eal/freebsd/eal.c b/lib/librte_eal/freebsd/eal.c
> > index 14b52168e..07456059e 100644
> > --- a/lib/librte_eal/freebsd/eal.c
> > +++ b/lib/librte_eal/freebsd/eal.c
> > @@ -956,13 +956,15 @@ rte_eal_init(int argc, char **argv)
> >                 return -1;
> >         }
> >         if (!internal_config.no_telemetry) {
> > -               const char *error_str;
> > +               const char *error_str = NULL;
> >                 if (rte_telemetry_init(rte_eal_get_runtime_dir(),
> >                                 &internal_config.ctrl_cpuset, &error_str)
> >                                 != 0) {
> >                         rte_eal_init_alert(error_str);
> >                         return -1;
> >                 }
> > +               if (error_str != NULL)
> > +                       RTE_LOG(WARNING, EAL, "%s\n", error_str);
> >         }
> >
> >         eal_mcfg_complete();
> > diff --git a/lib/librte_eal/linux/eal.c b/lib/librte_eal/linux/eal.c
> > index 9620d2544..10317a660 100644
> > --- a/lib/librte_eal/linux/eal.c
> > +++ b/lib/librte_eal/linux/eal.c
> > @@ -1294,13 +1294,15 @@ rte_eal_init(int argc, char **argv)
> >                 return -1;
> >         }
> >         if (!internal_config.no_telemetry) {
> > -               const char *error_str;
> > +               const char *error_str = NULL;
> >                 if (rte_telemetry_init(rte_eal_get_runtime_dir(),
> >                                 &internal_config.ctrl_cpuset, &error_str)
> >                                 != 0) {
> >                         rte_eal_init_alert(error_str);
> >                         return -1;
> >                 }
> > +               if (error_str != NULL)
> > +                       RTE_LOG(WARNING, EAL, "%s\n", error_str);
> 
> "EAL: No legacy callbacks, legacy socket not created" comes all the times now.
> Should we treat this as a warning? as it comes for with for the
> built-in applications?
> If it is important why not have a default handler in the EAL library?
> 
Perhaps the warning level could be changed to INFO.
BTW: if you compile with jansson library available, the legacy callbacks
will be available and this will be silenced.

/Bruce