On Wed, Nov 23, 2016 at 02:35:12PM +0200, Marius Vlad wrote: > v3: > - use igt_mean for accounting (Chris Wilson) > - make it Intel-agnostic when searching for connectors (Chris Wilson) > > v2: > - don't read cached values (Chris Wilson) > - warn on per connector, and fail per mean (Chris Wilson) > > These are synthetic: 10ms per connector, and 50ms for all. > > Signed-off-by: Marius Vlad <marius.c.vlad@xxxxxxxxx> > --- > tests/Makefile.sources | 2 +- > tests/kms_sysfs_edid_timing | 25 ----------- > tests/kms_sysfs_edid_timing.c | 96 +++++++++++++++++++++++++++++++++++++++++++ > 3 files changed, 97 insertions(+), 26 deletions(-) > delete mode 100755 tests/kms_sysfs_edid_timing > create mode 100644 tests/kms_sysfs_edid_timing.c > > diff --git a/tests/Makefile.sources b/tests/Makefile.sources > index 2d6ed8d..f835726 100644 > --- a/tests/Makefile.sources > +++ b/tests/Makefile.sources > @@ -215,6 +215,7 @@ TESTS_progs = \ > kms_sink_crc_basic \ > prime_udl \ > drv_module_reload \ > + kms_sysfs_edid_timing \ > $(NULL) > > # IMPORTANT: The ZZ_ tests need to be run last! > @@ -225,7 +226,6 @@ TESTS_scripts_M = \ > TESTS_scripts = \ > debugfs_emon_crash \ > drv_debugfs_reader \ > - kms_sysfs_edid_timing \ > sysfs_l3_parity \ > test_rte_check \ > tools_test \ > diff --git a/tests/kms_sysfs_edid_timing b/tests/kms_sysfs_edid_timing > deleted file mode 100755 > index 46ea540..0000000 > --- a/tests/kms_sysfs_edid_timing > +++ /dev/null > @@ -1,25 +0,0 @@ > -#!/bin/bash > -# > -# This check the time we take to read the content of all the possible connectors. > -# Without the edid -ENXIO patch (http://permalink.gmane.org/gmane.comp.video.dri.devel/62083), > -# we sometimes take a *really* long time. So let's just check for some reasonable timing here > -# > - > -DRM_LIB_ALLOW_NO_MASTER=1 > - > -SOURCE_DIR="$( dirname "${BASH_SOURCE[0]}" )" > -. $SOURCE_DIR/drm_lib.sh > - > -TIME1=$(date +%s%N) > -cat $(find /sys/devices/|grep drm | grep /status) > /dev/null > -TIME2=$(date +%s%N) > - > -# time in ms > -RES=$(((TIME2 - TIME1) / 1000000)) > - > -if [ $RES -gt 600 ]; then > - echo "Talking to outputs took ${RES}ms, something is wrong" > - exit $IGT_EXIT_FAILURE > -fi > - > -exit $IGT_EXIT_SUCCESS > diff --git a/tests/kms_sysfs_edid_timing.c b/tests/kms_sysfs_edid_timing.c > new file mode 100644 > index 0000000..b56a147 > --- /dev/null > +++ b/tests/kms_sysfs_edid_timing.c > @@ -0,0 +1,96 @@ > +/* > + * Copyright © 2016 Intel Corporation > + * > + * Permission is hereby granted, free of charge, to any person obtaining a > + * copy of this software and associated documentation files (the "Software"), > + * to deal in the Software without restriction, including without limitation > + * the rights to use, copy, modify, merge, publish, distribute, sublicense, > + * and/or sell copies of the Software, and to permit persons to whom the > + * Software is furnished to do so, subject to the following conditions: > + * > + * The above copyright notice and this permission notice (including the next > + * paragraph) shall be included in all copies or substantial portions of the > + * Software. > + * > + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR > + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, > + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL > + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER > + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING > + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS > + * IN THE SOFTWARE. > + */ > +#include "igt.h" > + > +#include <dirent.h> > +#include <fcntl.h> > +#include <sys/stat.h> > + > +#define THRESHOLD_PER_CONNECTOR 10 > +#define THRESHOLD_TOTAL 50 > +#define CHECK_TIMES 15 > + > +IGT_TEST_DESCRIPTION("This check the time we take to read the content of all " > + "the possible connectors. Without the edid -ENXIO patch " > + "(http://permalink.gmane.org/gmane.comp.video.dri.devel/62083), " > + "we sometimes take a *really* long time. " > + "So let's just check for some reasonable timing here"); > + > + > +igt_simple_main > +{ > + DIR *dirp; > + struct dirent *de; > + > + dirp = opendir("/sys/class/drm"); > + igt_assert(dirp != NULL); > + > + while ((de = readdir(dirp))) { > + struct igt_mean mean = {}; > + struct stat st; > + char path[PATH_MAX]; > + int i; > + > + if (*de->d_name == '.') > + continue;; > + > + snprintf(path, sizeof(path), "/sys/class/drm/%s/status", > + de->d_name); > + > + if (stat(path, &st)) > + continue; > + > + igt_mean_init(&mean); > + for (i = 0; i < CHECK_TIMES; i++) { > + struct timespec ts = {}; > + int fd; > + > + igt_nsec_elapsed(&ts); > + fd = open(path, O_WRONLY); > + igt_ignore_warn(write(fd, "detect\n", 7)); Looks good, but just rearrange this slightly: fd = open(); if (fd < 0) continue; igt_nsec_elapsed(&ts); igt_ignore_warn(write(fd, "detect\n", 7)); igt_mean_add(&mean, igt_nsec_elapsed(&ts)); That tightens the measurement to just the synchronous probe. > + > + igt_mean_add(&mean, igt_nsec_elapsed(&ts)); > + > + close(fd); > + } > + > + igt_debug("%s: mean.max %.2fns, %.2fus, %.2fms, " > + "mean.avg %.2fns, %.2fus, %.2fms\n", > + de->d_name, > + mean.max, mean.max / 1e3, mean.max / 1e6, > + mean.mean, mean.mean / 1e3, mean.mean / 1e6); For posterity, it might be nice to include a transcript of this output in the changelog. (That just helps us check the formatting comes out ok.) Longer term, we might want ministats style plotting of strange results. -Chris -- Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
