[monitor/1_hdmi_PCI_ID_0x????] The test case will be failed after connect HDMI port when initial display mode is external

[Gabriel Zhi Chen]

Test image: sutton-sazan2p-trusty-amd64-20160519-0
Test machine: P50s (Sazan2p-3)
CPU: Intel(R) Core(TM) i7-6600U CPU @ 2.60GHz
GPU: 00:02.0 VGA compatible controller [0300]: Intel Corporation Device [8086:1916] (rev 07)
+
06:00.0 3D controller [0302]: NVIDIA Corporation Device [10de:137a] (rev a2)

Steps:
1. Install image sutton-sazan2p-trusty-amd64-20160519-0 on Sazan2p-3.
2. Log in and make sure system working under NVidia [10de:137a]
3. Connect HDMI cable.
4. Try to switch display mode internal to next mode (e.g. external mode)
5. Unplug HDMI cable, then plug the HDMI cable again.
6. Try to switch display mode external to next mode (e.g. extend mode)
7. Unplug HDMI cable, then plug the HDMI cable again.
8. Try to switch display mode extend to next mode (e.g. mirror mode)
9. Unplug HDMI cable, then plug the HDMI cable again.
10. Try to switch display mode mirror to internal.

Actual result:
Screen blink during the switching.

Expect result:
The modes switch smoothly.

Notes:
Trigger -- step1 -> setp2 -> setp3. When the HDMI cable conncet, if the initial display mode is external, the issue reproduces 100%; the rest of modes could make the issue occurred, but the rate is random.

The issue will lead to test case monitor/1_hdmi_PCI_ID_0x???? failed. I wander if we could add the case to verify the display of monitor.

[Pierre Equoy (pieq)]

In the steps you describe, do you do everything manually, or is it during the execution of a Checkbox job? If so, which job are you running?

From the description and the video, it looks like this is a hardware issue and not a Checkbox issue.

[Gabriel Zhi Chen (gabrielzchen)]

Actually, if the initial display mode is external, the issue reproduces 100%. If the initial display mode is extend / mirror / internal, the issue could be reproduced, but not reproduce every time (rate: 30%). I mean that maybe we could add a test case for the issue. Or we could add the case on Sutton project so that we won't be missing the issue before GM test.

[Po-Hsu Lin (cypressyew)]

Switched back to new with information provided by Gabriel.

[Sylvain Pineau (sylvain-pineau)]

I suggest to create a sutton specific test for this issue (as a regression tests for this platform). Sounds too specific to me to be part of p-p-checkbox. Set as Invalid for this project.