From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga12.intel.com (mga12.intel.com [192.55.52.136]) by gabe.freedesktop.org (Postfix) with ESMTPS id 3A6EA89F9F for ; Wed, 8 Jul 2020 09:30:20 +0000 (UTC) Message-ID: From: Janusz Krzysztofik Date: Wed, 08 Jul 2020 11:30:15 +0200 In-Reply-To: <0a25feb516e53edfbb3bbf6457b185d554e92b39.camel@intel.com> References: <20200707145817.17415-1-marcin.bernatowicz@intel.com> <20200707211407.GB2209461@intel.com> <0a25feb516e53edfbb3bbf6457b185d554e92b39.camel@intel.com> MIME-Version: 1.0 Subject: Re: [igt-dev] [PATCH i-g-t] tests/device_reset: Test device sysfs reset List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Errors-To: igt-dev-bounces@lists.freedesktop.org Sender: "igt-dev" To: "Bernatowicz, Marcin" , "Vivi, Rodrigo" Cc: "igt-dev@lists.freedesktop.org" , "Winiarski, Michal" List-ID: On Wed, 2020-07-08 at 08:57 +0000, Bernatowicz, Marcin wrote: > On Tue, 2020-07-07 at 14:14 -0700, Rodrigo Vivi wrote: > > On Tue, Jul 07, 2020 at 05:08:07PM +0200, Janusz Krzysztofik wrote: > > > On Tue, 2020-07-07 at 16:58 +0200, Marcin Bernatowicz wrote: > > > > Device reset is initiated by writing "1" to reset sysfs file, > > > > which should initiate PCI device Function Level Reset > > > > if supported by device. > > > > > > > > Test scenarios: > > > > 1. unbind driver from device, initiate sysfs reset, rebind driver > > > > to > > > > device > > > > 2. device reset with bound driver > > > > > > > > v2: removed unbind-rebind (duplicates > > > > core_hotunplug@unbind-rebind) > > > > added healthcheck to each test (Janusz) > > > > v3: after review corrections (renamed sysfs_fds to device_fds, > > > > corrected not closed file descriptor, removed variable length > > > > array) > > > > v4: updated description (Martin), snprintf corrections (Janusz), > > > > reset-bound added to blacklist as it hangs several gens of > > > > hosts (Tomi) > > > > > > > > Signed-off-by: Marcin Bernatowicz > > > > --- > > > > tests/device_reset.c | 281 > > > > +++++++++++++++++++++++++++++++++++ > > > > tests/intel-ci/blacklist.txt | 3 + > > > > tests/meson.build | 1 + > > > > 3 files changed, 285 insertions(+) > > > > create mode 100644 tests/device_reset.c > > > > > > > > diff --git a/tests/device_reset.c b/tests/device_reset.c > > > > new file mode 100644 > > > > index 000000000..b1181b3be > > > > --- /dev/null > > > > +++ b/tests/device_reset.c > > > > @@ -0,0 +1,281 @@ > > > > +// SPDX-License-Identifier: MIT > > > > +/* > > > > + * Copyright(c) 2020 Intel Corporation. All rights reserved. > > > > + */ > > > > +#include > > > > +#include > > > > +#include > > > > + > > > > +#include "i915/gem.h" > > > > +#include "igt.h" > > > > +#include "igt_device_scan.h" > > > > +#include "igt_sysfs.h" > > > > + > > > > +IGT_TEST_DESCRIPTION("Examine behavior of a driver on device > > > > sysfs reset"); > > > > + > > > > + > > > > +#define DEV_PATH_LEN 80 > > > > +#define DEV_BUS_ADDR_LEN 13 /* addr has form 0000:00:00.0 */ > > > > + > > > > +/** > > > > + * Helper structure containing file descriptors > > > > + * and bus address related to tested device > > > > + */ > > > > +struct device_fds { > > > > + struct { > > > > + int dev; > > > > + int dev_dir; > > > > + int drv_dir; > > > > + } fds; > > > > + char dev_bus_addr[DEV_BUS_ADDR_LEN]; > > > > +}; > > > > + > > > > +static int __open_sysfs_dir(int fd, const char* path) > > > > +{ > > > > + int sysfs; > > > > + > > > > + sysfs = igt_sysfs_open(fd); > > > > + if (sysfs < 0) { > > > > + return -1; > > > > + } > > > > + > > > > + fd = openat(sysfs, path, O_DIRECTORY); > > > > + close(sysfs); > > > > + return fd; > > > > +} > > > > + > > > > +static int open_device_sysfs_dir(int fd) > > > > +{ > > > > + return __open_sysfs_dir(fd, "device"); > > > > +} > > > > + > > > > +static int open_driver_sysfs_dir(int fd) > > > > +{ > > > > + return __open_sysfs_dir(fd, "device/driver"); > > > > +} > > > > + > > > > +/** > > > > + * device_sysfs_path: > > > > + * @fd: opened device file descriptor > > > > + * @path: buffer to store sysfs path to device directory > > > > + * > > > > + * Returns: > > > > + * On successfull path resolution sysfs path to device > > > > directory, > > > > + * NULL otherwise > > > > + */ > > > > +static char *device_sysfs_path(int fd, char *path) > > > > +{ > > > > + char sysfs[DEV_PATH_LEN]; > > > > + > > > > + if (!igt_sysfs_path(fd, sysfs, sizeof(sysfs))) > > > > + return NULL; > > > > + > > > > + if (DEV_PATH_LEN <= (strlen(sysfs) + strlen("/device"))) > > > > + return NULL; > > > > + > > > > + strcat(sysfs, "/device"); > > > > + > > > > + return realpath(sysfs, path); > > > > +} > > > > + > > > > +static void init_device_fds(struct device_fds *dev) > > > > +{ > > > > + char dev_path[PATH_MAX]; > > > > + char *addr_pos; > > > > + > > > > + igt_debug("open device\n"); > > > > + /** > > > > + * As subtests must be able to close examined devices > > > > + * completely, don't use drm_open_driver() as it keeps > > > > + * a device file descriptor open for exit handler use. > > > > + */ > > > > + dev->fds.dev = __drm_open_driver(DRIVER_ANY); > > > > + igt_assert_fd(dev->fds.dev); > > > > + if (is_i915_device(dev->fds.dev)) > > > > + igt_require_gem(dev->fds.dev); > > > > + > > > > + igt_assert(device_sysfs_path(dev->fds.dev, dev_path)); > > > > + addr_pos = strrchr(dev_path, '/'); > > > > + igt_assert(addr_pos); > > > > + igt_assert_eq(sizeof(dev->dev_bus_addr) - 1, > > > > + snprintf(dev->dev_bus_addr, sizeof(dev- > > > > > dev_bus_addr), > > > > + "%s", addr_pos + 1)); > > > > + > > > > + dev->fds.dev_dir = open_device_sysfs_dir(dev->fds.dev); > > > > + igt_assert_fd(dev->fds.dev_dir); > > > > + > > > > + dev->fds.drv_dir = open_driver_sysfs_dir(dev->fds.dev); > > > > + igt_assert_fd(dev->fds.drv_dir); > > > > +} > > > > + > > > > +static int close_if_opened(int *fd) > > > > +{ > > > > + int rc = 0; > > > > + > > > > + if (fd && *fd != -1) { > > > > + rc = close(*fd); > > > > + *fd = -1; > > > > + } > > > > + return rc; > > > > +} > > > > + > > > > +static void cleanup_device_fds(struct device_fds *dev) > > > > +{ > > > > + igt_ignore_warn(close_if_opened(&dev->fds.dev)); > > > > + igt_ignore_warn(close_if_opened(&dev->fds.dev_dir)); > > > > + igt_ignore_warn(close_if_opened(&dev->fds.drv_dir)); > > > > +} > > > > + > > > > +/** > > > > + * is_sysfs_reset_supported: > > > > + * @fd: opened device file descriptor > > > > + * > > > > + * Check if device supports reset based on sysfs file presence. > > > > + * > > > > + * Returns: > > > > + * True if device supports reset, false otherwise. > > > > + */ > > > > +static bool is_sysfs_reset_supported(int fd) > > > > +{ > > > > + struct stat st; > > > > + int rc; > > > > + int sysfs; > > > > + int reset_fd = -1; > > > > + > > > > + sysfs = igt_sysfs_open(fd); > > > > + > > > > + if (sysfs >= 0) { > > > > + reset_fd = openat(sysfs, "device/reset", O_WRONLY); > > > > + close(sysfs); > > > > + } > > > > + > > > > + if (reset_fd < 0) > > > > + return false; > > > > + > > > > + rc = fstat(reset_fd, &st); > > > > + close(reset_fd); > > > > + > > > > + if (rc || !S_ISREG(st.st_mode)) > > > > + return false; > > > > + > > > > + return true; > > > > +} > > > > + > > > > +/* Unbind the driver from the device */ > > > > +static void driver_unbind(struct device_fds *dev) > > > > +{ > > > > + igt_debug("unbind the driver from the device\n"); > > > > + igt_assert(igt_sysfs_set(dev->fds.drv_dir, "unbind", > > > > + dev->dev_bus_addr)); > > > > +} > > > > + > > > > +/* Re-bind the driver to the device */ > > > > +static void driver_bind(struct device_fds *dev) > > > > +{ > > > > + igt_debug("rebind the driver to the device\n"); > > > > + igt_abort_on_f(!igt_sysfs_set(dev->fds.drv_dir, "bind", > > > > + dev->dev_bus_addr), "driver rebind failed"); > > > > +} > > > > + > > > > +/* Initiate device reset */ > > > > +static void initiate_device_reset(struct device_fds *dev) > > > > +{ > > > > + igt_debug("reset device\n"); > > > > + igt_assert(igt_sysfs_set(dev->fds.dev_dir, "reset", "1")); > > > > +} > > > > + > > > > +static bool is_i915_wedged(int i915) > > > > +{ > > > > + int err = 0; > > > > + > > > > + if (ioctl(i915, DRM_IOCTL_I915_GEM_THROTTLE)) > > > > + err = -errno; > > > > + return err == -EIO; > > > > +} > > > > + > > > > +/** > > > > + * healthcheck: > > > > + * @dev: structure with device descriptor, if descriptor equals > > > > -1 > > > > + * the device is reopened > > > > + */ > > > > +static void healthcheck(struct device_fds *dev) > > > > +{ > > > > + if (dev->fds.dev == -1) { > > > > + /* refresh device list */ > > > > + igt_devices_scan(true); > > > > + igt_debug("reopen the device\n"); > > > > + dev->fds.dev = __drm_open_driver(DRIVER_ANY); > > > > + } > > > > + igt_assert_fd(dev->fds.dev); > > > > + > > > > + if (is_i915_device(dev->fds.dev)) > > > > + igt_assert(!is_i915_wedged(dev->fds.dev)); > > > > +} > > > > + > > > > +/** > > > > + * set_device_filter: > > > > + * > > > > + * Sets device filter to ensure subtests always reopen the same > > > > device > > > > + * > > > > + * @dev_path: path to device under tests > > > > + */ > > > > +static void set_device_filter(const char* dev_path) > > > > +{ > > > > +#define FILTER_PREFIX_LEN 4 > > > > + char filter[PATH_MAX + FILTER_PREFIX_LEN]; > > > > + > > > > + igt_assert_lt(FILTER_PREFIX_LEN, snprintf(filter, > > > > sizeof(filter), > > > > + "sys:%s", dev_path)); > > > > + igt_device_filter_free_all(); > > > > + igt_assert_eq(igt_device_filter_add(filter), 1); > > > > +} > > > > + > > > > +static void unbind_reset_rebind(struct device_fds *dev) > > > > +{ > > > > + igt_debug("close the device\n"); > > > > + close_if_opened(&dev->fds.dev); > > > > + > > > > + driver_unbind(dev); > > > > + > > > > + initiate_device_reset(dev); > > > > + > > > > + driver_bind(dev); > > > > +} > > > > + > > > > +igt_main > > > > +{ > > > > + struct device_fds dev = { .fds = {-1, -1, -1}, .dev_bus_addr = > > > > {0}}; > > > > + > > > > + igt_fixture { > > > > + char dev_path[PATH_MAX]; > > > > + > > > > + igt_debug("opening device\n"); > > > > + init_device_fds(&dev); > > > > + > > > > + /* Make sure subtests always reopen the same device */ > > > > + igt_assert(device_sysfs_path(dev.fds.dev, dev_path)); > > > > + set_device_filter(dev_path); > > > > + > > > > + igt_skip_on(!is_sysfs_reset_supported(dev.fds.dev)); > > > > + > > > > + igt_set_timeout(60, "device reset tests timed out after > > > > 60s"); > > > > + } > > > > + > > > > + igt_describe("Unbinds driver from device, initiates reset" > > > > + " then rebinds driver to device"); > > > > + igt_subtest("unbind-reset-rebind") { > > > > + unbind_reset_rebind(&dev); > > > > + healthcheck(&dev); > > > > + } > > > > + > > > > + igt_describe("Resets device with bound driver"); > > > > + igt_subtest("reset-bound") { > > > > + initiate_device_reset(&dev); > > > > + healthcheck(&dev); > > > > + } > > > > + > > > > + igt_fixture { > > > > + igt_reset_timeout(); > > > > + cleanup_device_fds(&dev); > > > > + } > > > > +} > > > > diff --git a/tests/intel-ci/blacklist.txt b/tests/intel- > > > > ci/blacklist.txt > > > > index ecbec5080..f9a57cb54 100644 > > > > --- a/tests/intel-ci/blacklist.txt > > > > +++ b/tests/intel-ci/blacklist.txt > > > > @@ -121,3 +121,6 @@ igt@perf_pmu@cpu-hotplug > > > > # Currently fails and leaves the machine in a very bad state, > > > > and > > > > # causes coverage loss for other tests. > > > > igt@core_hotunplug@.* > > > > + > > > > +# hangs several gens of hosts, and has no immediate fix > > > > +igt@device_reset@reset-bound > > > > I wonder if we should filter by platform? > > if gen >= 12? or if dgfx? > > > > or is it really random? > > I can restrict the tests on intel platforms with > > if (is_i915_device(fd)) > igt_skip_on(intel_gen(intel_get_drm_devid(fd)) < 12); That seems pointless to me unless igt@device_reset@reset-bound works on gen12+ while it doesn't work on platforms we don't care. Thanks, Janusz > > > > > \ No newline at end of file > > > > > > I'm only not sure if that missing newline will not break CI > > > scripts, > > > other than that, my R-b still applies. > > > > > > Reviewed-by: Janusz Krzysztofik > > > anyway it looks good to me > > > > Acked-by: Rodrigo Vivi > > > > > > > > > Thanks, > > > Janusz > > > > > > > > > > diff --git a/tests/meson.build b/tests/meson.build > > > > index 172d18e59..d9253f5f9 100644 > > > > --- a/tests/meson.build > > > > +++ b/tests/meson.build > > > > @@ -8,6 +8,7 @@ test_progs = [ > > > > 'core_setmaster_vs_auth', > > > > 'debugfs_test', > > > > 'dmabuf', > > > > + 'device_reset', > > > > 'drm_import_export', > > > > 'drm_mm', > > > > 'drm_read', _______________________________________________ igt-dev mailing list igt-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/igt-dev