Re: [PATCH 0/3] replace use of EAL logtype

[David Marchand <david.marchand@redhat.com>]

On Mon, Dec 11, 2023 at 8:19 PM Tyler Retzlaff
<roretzla@linux.microsoft.com> wrote:
>
> On Mon, Dec 11, 2023 at 09:23:22AM -0800, Stephen Hemminger wrote:
> > There are some places EAL logtype is being used in testpmd
> > and examples where it should not be. Lets reserve EAL
> > logtype to only be used by DPDK internals.
> >
> > Stephen Hemminger (3):
> >   testpmd: replace EAL logtype with fprintf
> >   examples/l2fwd-keepalive: don't use EAL logtype
> >   examples/vm_power_manager: do not use EAL logtype
> >
> >  app/test-pmd/testpmd.c           | 33 +++++++++++---------------------
> >  examples/l2fwd-keepalive/shm.c   | 21 +++++++-------------
> >  examples/vm_power_manager/main.c | 11 ++++-------
>
> are the log types under ALLOW_INTERNAL_API? if not should they be? and
> if they are should examples be defining ALLOW_INTERNAL_API?

ALLOW_INTERNAL_API and ALLOW_EXPERIMENTAL_API tags have a transitive aspect.
So it would prevent applications from calling inline helpers and
macros with such logs in them.

Like for example:
lib/eal/include/rte_test.h:             RTE_LOG(ERR, EAL, "Test assert
%s line %d failed: "           \

The EAL logtype is "easy" to fix, but others like ETHDEV are more tricky.


-- 
David Marchand