[Bug 103060] [CI] igt at kms_flip@*-vs-vblank-race* - Failed assertion: end - start > 0.9 * mode_frame_time(o) && end - start < 2.1 * mode_frame_time(o)

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=103060

--- Comment #23 from Stanislav Lisovskiy <stanislav.lisovskiy at intel.com> ---
I wonder still, if this wait_for_vblanks is really that long. 
As I understand from test case we wait for 2 vblanks and then expected time
elapsed to be in ([ 1, 2] +- 0.1) * frame time.

Collected from ci buglog few last runs:

(kms_flip:1841) CRITICAL: wait for two vblanks took 35113 usec (frame time
16667.604313 usec)
(kms_flip:1836) CRITICAL: wait for two vblanks took 35024 usec (frame time
16667.604313 usec)
(kms_flip:1769) CRITICAL: wait for two vblanks took 35070 usec (frame time
16667.604313 usec)
(kms_flip:1782) CRITICAL: wait for two vblanks took 35063 usec (frame time
16667.604313 usec)
(kms_flip:1596) CRITICAL: wait for two vblanks took 35138 usec (frame time
16667.604313 usec)
(kms_flip:1777) CRITICAL: wait for two vblanks took 35060 usec (frame time
16667.604313 usec)

The frame time here is 16667.604313 which multiplied by 2.1 gives
35000.0000007(maximum value that we allow in assert)
which is only ~100 usec different from values which we currently see.

I also checked on my own ICL board - it doesn't fail as the values stay between
32000-33000 usec, however the deviation is still might be even up to 1000 usec
if I modify this test case to run it 100 times:

...
wait for two vblanks took 32139 usec (frame time 16667.604313 usec)
wait for two vblanks took 33297 usec (frame time 16667.604313 usec)
wait for two vblanks took 33289 usec (frame time 16667.604313 usec)
wait for two vblanks took 33221 usec (frame time 16667.604313 usec)
wait for two vblanks took 33330 usec (frame time 16667.604313 usec)
wait for two vblanks took 33325 usec (frame time 16667.604313 usec)
wait for two vblanks took 33334 usec (frame time 16667.604313 usec)
wait for two vblanks took 33238 usec (frame time 16667.604313 usec)
wait for two vblanks took 33352 usec (frame time 16667.604313 usec)
...

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20190618/aa5f43d4/attachment-0001.html>