On Wed, Jun 17, 2020 at 3:56 PM Andrew Rybchenko
<arybche...@solarflare.com> wrote:
>
> On 6/17/20 1:00 PM, Thomas Monjalon wrote:
> > 17/06/2020 08:30, jer...@marvell.com:
> >> From: Jerin Jacob <jer...@marvell.com>
> >>
> >> Introducing the RTE_LOG_REGISTER macro to avoid the code duplication
> >> in the log registration process.
> >>
> >> It is a wrapper macro for declaring the logtype, register the log and sets
> >> it's level in the constructor context.
> >>
> >> Signed-off-by: Jerin Jacob <jer...@marvell.com>
> >> ---
> >> --- a/lib/librte_eal/include/rte_log.h
> >> +++ b/lib/librte_eal/include/rte_log.h
> >> +#define RTE_LOG_REGISTER(type, name, level) \
> >> +int type; \
> >> +RTE_INIT(__##type) \
> >> +{ \
> >> + type = rte_log_register(RTE_STR(name)); \
> >> + if (type >= 0) \
> >> + rte_log_set_level(type, RTE_LOG_##level); \
> >> +}
> >
> > It should use rte_log_register_type_and_pick_level()
> > which works for drivers loaded later in the init sequence.
>
> Yes, rte_log_register_type_and_pick_level() should be used.
> Otherwise, if a driver is loaded (the constructor is called)
> after --log-level options processing, the level set using
> the option is not applied.
OK. I will change it in v2.
>
> > rte_log_register() should be deprecated.
>
> +1
>
