Re: [PATCH] test/event_eth_tx_adapter: skip tests if no eventdev

On Tue, Jul 21, 2026 at 02:07:23PM +0200, David Marchand wrote:
> On Tue, 21 Jul 2026 at 13:57, Bruce Richardson
> <[email protected]> wrote:
> >
> > If there is no eventdev present, and the SW eventdev driver is not
> > available, skip the eth tx adapter tests, rather than failing on event
> > creation.
> >
> > Note: We keep the failure case where we do have the event/sw driver
> > built, but fail to create an instance of it.
> >
> > Signed-off-by: Bruce Richardson <[email protected]>
> > ---
> >  app/test/test_event_eth_tx_adapter.c | 10 ++++++++--
> >  1 file changed, 8 insertions(+), 2 deletions(-)
> >
> > diff --git a/app/test/test_event_eth_tx_adapter.c b/app/test/test_event_eth_tx_adapter.c
> > index bec298a8b8..18801028ce 100644
> > --- a/app/test/test_event_eth_tx_adapter.c
> > +++ b/app/test/test_event_eth_tx_adapter.c
> > @@ -202,12 +202,13 @@ deinit_ports(void)
> >  static int
> >  testsuite_setup(void)
> >  {
> > -       const char *vdev_name = "event_sw0";
> > -
> >         int err = init_ports();
> >         TEST_ASSERT(err == 0, "Port initialization failed err %d\n", err);
> >
> > +#ifdef RTE_EVENT_SW
> >         if (rte_event_dev_count() == 0) {
> > +               const char *vdev_name = "event_sw0";
> > +
> >                 printf("Failed to find a valid event device,"
> >                         " testing with event_sw0 device\n");
> >                 err = rte_vdev_init(vdev_name, NULL);
> > @@ -215,6 +216,11 @@ testsuite_setup(void)
> >                         vdev_name, err);
> >                 event_dev_delete = 1;
> >         }
> > +#endif
> > +       if (rte_event_dev_count() == 0) {
> > +               printf("Failed to find a valid event device, skipping test\n");
> > +               return TEST_SKIPPED;
> > +       }
> 
> It seems strange to reevaluate this part here.
> 
> rte_vdev_init succeeded but we still have no eventdev object?
> That should be a failure to me.
>
This is meant to catch the case where we don't have RTE_EVENT_SW. However,
I can rework the change to avoid checking twice in the case we do have it.

/Bruce