When suspending a KBL-g NUC, we got the following EDID read failure:

[  200.686086] amdgpu 0000:01:00.0: HDMI-A-1: EDID is invalid:
[  200.686088]  [00] BAD  00 ff ff ff ff ff ff 00 05 e3 cd 0c 00 00 00 00
[  200.686089]  [00] BAD  03 1b 01 03 80 3e 22 78 ea 1e c5 ae 4f 34 b1 26
[  200.686090]  [00] BAD  01 50 54 2f cf 00 d1 cf b3 00 a9 c0 95 00 81 81
[  200.686091]  [00] BAD  81 00 81 c0 d1 00 02 3a 80 18 71 38 2d 40 58 2c
[  200.686092]  [00] BAD  35 00 e0 0e 11 00 00 1a e2 68 00 a0 a0 40 2e 60
[  200.686092]  [00] BAD  30 20 36 00 80 90 21 00 00 1a 56 5e 00 a0 a0 a0
[  200.686093]  [00] BAD  29 50 30 20 36 00 80 68 21 00 00 1a 00 00 00 fc
[  200.686094]  [00] BAD  00 32 38 45 38 35 30 0a 20 20 20 20 20 20 01 b2
[  200.686117] [drm:dc_link_detect [amdgpu]] *ERROR* No EDID read.

https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_4430/fi-kbl-8809g/igt@gem_exec_suspend@basic-s3.html

This seems to have happened only once, but let's see if we can reproduce this
issue.