Thanks for spotting this. I'll move the call to rte_cryptodev_close into the function testsuite_teardown and send v3.
Adam > -----Original Message----- > From: Akhil Goyal <gak...@marvell.com> > Sent: Thursday, 8 April, 2021 14:16 > To: Dybkowski, AdamX <adamx.dybkow...@intel.com>; dev@dpdk.org; > Doherty, Declan <declan.dohe...@intel.com>; Kusztal, ArkadiuszX > <arkadiuszx.kusz...@intel.com> > Subject: RE: [dpdk-dev] [PATCH v2 1/2] test/crypto: close PMD after tests > > > The pmd is initialized in the setup function. > > This patch adds one extra step inside the teardown function: the call > > to the rte_cryptodev_close apart of the call to rte_cryptodev_stop > > function that existed before. > > > > I don't see any sense of calling the stop function inside the setup, > > in my opinion it's much better to do it during the teardown. > > > > Sorry, I wanted to refer to testsuite_teardown() instead of ut_teardown(). > The reason is that, vdevs are initialized in testsuite_setup and not ut_setup. > Hence corresponding reverse function should be in testsuite_teardown(). > > -akhil > > > > > -----Original Message----- > > > From: Akhil Goyal <gak...@marvell.com> > > > Sent: Monday, 5 April, 2021 20:59 > > > To: Akhil Goyal <gak...@marvell.com>; Dybkowski, AdamX > > > <adamx.dybkow...@intel.com>; dev@dpdk.org; Doherty, Declan > > > <declan.dohe...@intel.com>; Kusztal, ArkadiuszX > > > <arkadiuszx.kusz...@intel.com> > > > Subject: RE: [dpdk-dev] [PATCH v2 1/2] test/crypto: close PMD after > > > tests > > > > > > Hi Adam/Arek, > > > > > > Could you please reply to the below query. > > > > > > > > This patch adds closing of the PMD after running the tests. > > > > > > > > > > Signed-off-by: Adam Dybkowski <adamx.dybkow...@intel.com> > > > > > --- > > > > > app/test/test_cryptodev.c | 6 +++++- > > > > > 1 file changed, 5 insertions(+), 1 deletion(-) > > > > > > > > > > diff --git a/app/test/test_cryptodev.c > > > > > b/app/test/test_cryptodev.c index f91debc16..ea965a64a 100644 > > > > > --- a/app/test/test_cryptodev.c > > > > > +++ b/app/test/test_cryptodev.c > > > > > @@ -928,6 +928,7 @@ ut_teardown(void) > > > > > struct crypto_testsuite_params *ts_params = > &testsuite_params; > > > > > struct crypto_unittest_params *ut_params = > &unittest_params; > > > > > struct rte_cryptodev_stats stats; > > > > > + int res; > > > > > > > > > > /* free crypto session structure */ #ifdef RTE_LIB_SECURITY > > > > > @@ > > > > > -976,8 +977,11 @@ ut_teardown(void) > > > > > > > > > > rte_cryptodev_stats_get(ts_params->valid_devs[0], > &stats); > > > > > > > > > > - /* Stop the device */ > > > > > + /* Stop and close the device */ > > > > > rte_cryptodev_stop(ts_params->valid_devs[0]); > > > > > + res = rte_cryptodev_close(ts_params->valid_devs[0]); > > > > > + if (res) > > > > > + RTE_LOG(ERR, USER1, "Crypto device close error > %d\n", > > > res); > > > > > > > > Shouldn't this be part of testsuite_setup() instead of ut_teardown()? > > > > In cases of vdev, devices are initialized as part of testsuite_setup(). > > > > > > > > Should we also call rte_cryptodev_queue_pair_release from > > > ut_teardown? > > > > > > > > Regards, > > > > Akhil
