eeh-basic.sh from powerpc in ubuntu_kernel_selftests break with B-5.4

[Po-Hsu Lin]

Issue found on Power9 node baltar (5.4.0-48.52~18.04.1) passed with P8 node

The whole sru-misc suite which contains the ubuntu_kernel_selftests breaks at this test:

 TAP version 13
 1..1
 # selftests: powerpc/eeh: eeh-basic.sh
 # 0000:00:00.0, Skipped: bridge
 # 0001:00:00.0, Skipped: bridge
 # 0002:00:00.0, Skipped: bridge
 # 0002:01:00.0, Added
 # 0003:00:00.0, Skipped: bridge
 # 0003:01:00.0, Added
 # 0004:00:00.0, Skipped: bridge
 # 0004:01:00.0, Skipped: bridge
 # 0004:02:00.0, Added
 # 0005:00:00.0, Skipped: bridge
 # 0005:01:00.0, Added
 # 0030:00:00.0, Skipped: bridge
 # 0031:00:00.0, Skipped: bridge
 # 0032:00:00.0, Skipped: bridge
 # 0033:00:00.0, Skipped: bridge
 # Found 4 breakable devices...
 # Breaking 0002:01:00.0...
 # 0002:01:00.0, waited 0/60
 # 0002:01:00.0, waited 1/60
 # 0002:01:00.0, waited 2/60
 # 0002:01:00.0, waited 3/60
 # 0002:01:00.0, waited 4/60
 # 0002:01:00.0, waited 5/60
 # 0002:01:00.0, waited 6/60
 # 0002:01:00.0, Recovered after 7 seconds
 # Breaking 0003:01:00.0...

 Due to the above error(s), this script is unable to continue and is terminating.

This is making the test report unable to populate.

[Po-Hsu Lin (cypressyew)]

I tried to reproduce this manually with the following combination on this node baltar but no luck.

Kernel 5.4.0-48.52~18.04.1 + Source Code 5.4.0-48.52~18.04.1
Kernel 5.4.0-48.52~18.04.1 + Source Code 5.4.0-47.51~18.04.1
Kernel 5.4.0-47.51~18.04.1 + Source Code 5.4.0-47.51~18.04.1

[Po-Hsu Lin (cypressyew)]

Device 0003:01:00.0 on node baltar is:
0003:01:00.0 Serial Attached SCSI controller [0107]: Adaptec Series 8 12G SAS/PCIe 3 [9005:028d] (rev 01)

And it works with 5.8.0-25-generic #26~20.04.1-Ubuntu

From the most recent test log, this issue does not exist in 5.4.0-45.49 and 5.4.0-52.57~18.04.1 - generic with P9, I will close this bug.