| Message ID | 20181121115829.13114-1-chris@chris-wilson.co.uk (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | drm/i915/selftests: Use msecs_to_jiffies() for an explicit ms timeout | expand |
Chris Wilson <chris@chris-wilson.co.uk> writes: > Make the wait for the evict kthread use an explicit timeout rather than > a handwaved jiffie value so it is consistent across all platforms. > > References: https://bugs.freedesktop.org/show_bug.cgi?id=108735 > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> ...there is one left in intel_workarounds.c -Mika > --- > drivers/gpu/drm/i915/selftests/intel_hangcheck.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c > index defe671130ab..1f76257641f2 100644 > --- a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c > +++ b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c > @@ -1171,7 +1171,7 @@ static int __igt_reset_evict_vma(struct drm_i915_private *i915, > struct igt_wedge_me w; > > /* The reset, even indirectly, should take less than 10ms. */ > - igt_wedge_on_timeout(&w, i915, HZ / 10 /* 100ms timeout*/) > + igt_wedge_on_timeout(&w, i915, msecs_to_jiffies(100)) > err = kthread_stop(tsk); > > put_task_struct(tsk); > -- > 2.19.1 > > _______________________________________________ > Intel-gfx mailing list > Intel-gfx@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoting Mika Kuoppala (2018-11-21 12:26:23) > Chris Wilson <chris@chris-wilson.co.uk> writes: > > > Make the wait for the evict kthread use an explicit timeout rather than > > a handwaved jiffie value so it is consistent across all platforms. > > > > References: https://bugs.freedesktop.org/show_bug.cgi?id=108735 > > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> > > Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> > > ...there is one left in intel_workarounds.c Fortunately, we haven't yet exceeded it. HZ/5 should be pretty stable, worst case is 20 jiffies. Even HZ/10 at 10 jiffies shouldn't be much of a problem, just worrying about why the kthread didn't wake up. Maybe I should just bump its prio. -Chris
Quoting Patchwork (2018-11-21 12:33:00) > == Series Details == > > Series: drm/i915/selftests: Use msecs_to_jiffies() for an explicit ms timeout > URL : https://patchwork.freedesktop.org/series/52820/ > State : success > > == Summary == > > = CI Bug Log - changes from CI_DRM_5180 -> Patchwork_10876 = > > == Summary - SUCCESS == > > No regressions found. > > External URL: https://patchwork.freedesktop.org/api/1.0/series/52820/revisions/1/mbox/ > > == Known issues == > > Here are the changes found in Patchwork_10876 that come from known issues: > > === IGT changes === > > ==== Issues hit ==== > > igt@gem_ctx_create@basic-files: > fi-icl-u2: PASS -> DMESG-WARN (fdo#107724) > fi-bsw-kefka: PASS -> INCOMPLETE (fdo#108714) > > igt@i915_selftest@live_coherency: > fi-gdg-551: NOTRUN -> DMESG-FAIL (fdo#107164) > > igt@i915_selftest@live_hangcheck: > fi-bwr-2160: PASS -> DMESG-FAIL (fdo#108735) Worrying proved right. Switching to 100ms isn't enough. <0> [311.332381] igt/evic-2202 0.... 311291205us : reset_ring: rcs0 request global=1326, current=1325 <0> [311.332524] kworker/-165 1.... 311394694us : __igt_wedge_me: __igt_reset_evict_vma timed out. 100ms elapsed. Why didn't you wake up? -Chris
diff --git a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c index defe671130ab..1f76257641f2 100644 --- a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c +++ b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c @@ -1171,7 +1171,7 @@ static int __igt_reset_evict_vma(struct drm_i915_private *i915, struct igt_wedge_me w; /* The reset, even indirectly, should take less than 10ms. */ - igt_wedge_on_timeout(&w, i915, HZ / 10 /* 100ms timeout*/) + igt_wedge_on_timeout(&w, i915, msecs_to_jiffies(100)) err = kthread_stop(tsk); put_task_struct(tsk);
Make the wait for the evict kthread use an explicit timeout rather than a handwaved jiffie value so it is consistent across all platforms. References: https://bugs.freedesktop.org/show_bug.cgi?id=108735 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> --- drivers/gpu/drm/i915/selftests/intel_hangcheck.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)