From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mgamail.intel.com (mgamail.intel.com [134.134.136.20]) by gabe.freedesktop.org (Postfix) with ESMTPS id 05AA910E404 for ; Mon, 4 Dec 2023 20:47:19 +0000 (UTC) Message-ID: <79c11e97-7185-4aad-adc3-0dd36f8a1382@intel.com> Date: Mon, 4 Dec 2023 21:47:13 +0100 MIME-Version: 1.0 Content-Language: en-US To: Lukasz Laguna , igt-dev@lists.freedesktop.org References: <20231204171033.2167-1-lukasz.laguna@intel.com> <20231204171033.2167-8-lukasz.laguna@intel.com> From: Michal Wajdeczko In-Reply-To: <20231204171033.2167-8-lukasz.laguna@intel.com> Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Subject: Re: [igt-dev] [PATCH i-g-t v6 7/8] tests/sriov_basic: validate driver binding to VFs List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: igt-dev-bounces@lists.freedesktop.org Sender: "igt-dev" List-ID: On 04.12.2023 18:10, Lukasz Laguna wrote: > From: Katarzyna Dec > > Test enables VFs in range <1..totalvfs>, bind driver to all of them and > then unbind driver from all of them. > > v2: > - remove checks that are outside the scope of the test (Michal) > - change subtest run type (Michal) > > Cc: Marcin Bernatowicz > Cc: Michal Wajdeczko > Signed-off-by: Katarzyna Dec > Signed-off-by: Lukasz Laguna > --- > tests/sriov_basic.c | 49 +++++++++++++++++++++++++++++++++++++++++++++ > 1 file changed, 49 insertions(+) > > diff --git a/tests/sriov_basic.c b/tests/sriov_basic.c > index 582e644f2..c2c8e5b6b 100644 > --- a/tests/sriov_basic.c > +++ b/tests/sriov_basic.c > @@ -60,6 +60,36 @@ static void enable_vfs_autoprobe_on(int pf_fd, unsigned int num_vfs) > igt_assert(!err); > } > > +/** > + * SUBTEST: enable-vfs-bind-all-unbind-all > + * Description: > + * Verify VFs enabling, binding the driver and then unbinding it from all of them > + * Run type: FULL > + */ > +static void enable_vfs_bind_all_unbind_all(int pf_fd, unsigned int num_vfs) > +{ > + igt_debug("Testing %u VFs\n", num_vfs); > + > + igt_require(igt_sriov_get_enabled_vfs(pf_fd) == 0); > + > + igt_sriov_disable_driver_autoprobe(pf_fd); > + igt_sriov_enable_vfs(pf_fd, num_vfs); > + igt_sriov_enable_driver_autoprobe(pf_fd); > + > + for (int i = 1; i <= num_vfs; i++) { > + igt_assert(!igt_sriov_is_vf_drm_driver_probed(pf_fd, i)); > + igt_sriov_bind_vf_drm_driver(pf_fd, i); > + igt_assert(igt_sriov_is_vf_drm_driver_probed(pf_fd, i)); > + } what's the point in keeping all VFs drivers loaded ? do we really want to stress the system ? our goal should be to check that VF driver loads on each VF IMO the enable_vfs_bind_unbind_each() from 8/8 provides sufficient coverage, while this one will just consume machine-test-hours also this test duplicates enable_vfs_autoprobe_on() from 5/8 (just loads the all drivers manually instead of PCI-subsystem autoprobe) > + > + for (int i = 1; i <= num_vfs; i++) { > + igt_sriov_unbind_vf_drm_driver(pf_fd, i); > + igt_assert(!igt_sriov_is_vf_drm_driver_probed(pf_fd, i)); > + } > + > + igt_sriov_disable_vfs(pf_fd); > +} > + > igt_main > { > int pf_fd; > @@ -110,6 +140,25 @@ igt_main > } > } > > + igt_describe("Verify VFs enabling, binding the driver and then unbinding it from all of them"); > + igt_subtest_with_dynamic("enable-vfs-bind-all-unbind-all") { > + for_each_sriov_num_vfs(pf_fd, num_vfs) { > + igt_dynamic_f("numvfs-%u", num_vfs) { > + enable_vfs_bind_all_unbind_all(pf_fd, num_vfs); > + } > + } > + for_random_sriov_num_vfs(pf_fd, num_vfs) { > + igt_dynamic_f("numvfs-random") { > + enable_vfs_bind_all_unbind_all(pf_fd, num_vfs); > + } > + } > + for_max_sriov_num_vfs(pf_fd, num_vfs) { > + igt_dynamic_f("numvfs-all") { > + enable_vfs_bind_all_unbind_all(pf_fd, num_vfs); > + } > + } > + } > + > igt_fixture { > igt_sriov_disable_vfs(pf_fd); > /* abort to avoid execution of next tests with enabled VFs */