[Bug 1931390] Re: cpuhotplug03 in cpuhotplug from ubuntu_ltp failed after successful CPU1 offline

I have seen this issue for noble:linux-lowlatency 6.8.0-48.48.3 on
openstack:cpu2-ram4-disk20 (arm64). (SRU Cycle: 2024.09.30)

I think this test is flaky. To check if "cpuhotplug_do_spin_loop" it
uses "ps psr" command where "r" option lists the currently running
processes only. Even if a "cpuhotplug_do_spin_loop" process did migrate
to target CPU, this does not mean that it is currently running there (as
it can be preempted by availability of other tasks). And for
architectures with lower number of cpus (in my case 2), test is likely
to fail more.

Moreover, test doesn't seem to adjust "spin" tasks' priority (I am not
sure about this but it doesn't seem to do anything in this regard prior
to running the test either).

Test needs to either;
- run in isolated environment where OS threads run on a dedicated core.
- sleep some time (as done currently) and check if the spin process is migrated by the spin process itself.

Here is the log:
FO: Test start time: Fri Oct 18 20:06:59 UTC 2024
ND: /opt/ltp/bin/ltp-pan -q -e -S -a 166236 -n 166236 -f /tmp/ltp-9srhcvY2Gx/alltests -l /dev/null -C /dev/null -T /dev/null
OG File: /dev/null
ED COMMAND File: /dev/null
NF COMMAND File: /dev/null
ng tests.......
me: cpuhotplug03
te: Fri Oct 18 20:06:59 UTC 2024
sc: Do tasks get scheduled to a newly on-lined CPU?
08 CPU is 1
ER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
ot 166323 12.8 0.0 2380 1408 ? R 20:06 0:00 /bin/sh /opt/ltp/testcases/bin/cpuhotplug_do_spin_loop
ot 166324 12.8 0.0 2380 1408 ? R 20:06 0:00 /bin/sh /opt/ltp/testcases/bin/cpuhotplug_do_spin_loop
ot 166325 12.8 0.0 2380 1408 ? R 20:06 0:00 /bin/sh /opt/ltp/testcases/bin/cpuhotplug_do_spin_loop
ot 166326 12.8 0.0 2380 1408 ? R 20:06 0:00 /bin/sh /opt/ltp/testcases/bin/cpuhotplug_do_spin_loop
ot 166331 0.0 0.0 6680 1920 ? S 20:06 0:00 grep cpuhotplug_do_spin_loop
03 1 TINFO: Onlining CPU 1
0 /bin/sh /opt/ltp/testcases/bin/cpuhotplug_do_spin_loop
0 /bin/sh /opt/ltp/testcases/bin/cpuhotplug_do_spin_loop
0 /bin/sh /opt/ltp/testcases/bin/cpuhotplug_do_spin_loop
0 /bin/sh /opt/ltp/testcases/bin/cpuhotplug_do_spin_loop
03 1 TFAIL: No cpuhotplug_do_spin_loop processes found on CPU1
FO: ltp-pan reported some tests FAIL
TP Version: 20230929-874-gba610da01
FO: Test end time: Fri Oct 18 20:07:01 UTC 2024

--
You received this bug notification because you are subscribed to linux
in Ubuntu.
Matching subscriptions: Bgg, Bmail, Nb
https://bugs.launchpad.net/bugs/1931390

Title:
cpuhotplug03 in cpuhotplug from ubuntu_ltp failed after successful
CPU1 offline

Status in ubuntu-kernel-tests:
New
Status in linux package in Ubuntu:
Incomplete

Bug description:
In cpuhotplug (ubuntu_ltp) tests, offline and online of CPU1 succeeds,
but later the test fails with:

utils:0153| [stdout] Name: cpuhotplug03
utils:0153| [stdout] Date: Wed Jun 9 03:52:16 UTC 2021
utils:0153| [stdout] Desc: Do tasks get scheduled to a newly on-lined CPU?
utils:0153| [stdout]
utils:0153| [stdout] CPU is 1
utils:0153| [stdout] USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
utils:0153| [stdout] root 858611 0.0 0.0 2616 540 ? R 03:52 0:00 /bin/sh /opt/ltp/testcases/bin/cpuhotplug_do_spin_loop
...
...
...
utils:0153| [stdout] 49 /bin/sh /opt/ltp/testcases/bin/cpuhotplug_do_spin_loop
utils:0153| [stdout] 82 /bin/sh /opt/ltp/testcases/bin/cpuhotplug_do_spin_loop
utils:0153| [stdout] 18 /bin/sh /opt/ltp/testcases/bin/cpuhotplug_do_spin_loop
utils:0153| [stdout] cpuhotplug03 1 TFAIL: No cpuhotplug_do_spin_loop processes found on CPU1
utils:0153| [stdout] tag=cpuhotplug03 stime=1623210736 dur=3 exit=exited stat=1 core=no cu=22260 cs=78

IMPORTANT: The CPU1 got properly offlined and onlined. For failures of
CPU1 offline, please see lp:1836167 (reported previously as part of
lp:1836167).

Po-Hsu Lin (cypressyew) wrote on 2020-09-25:
On F-oem-5.6 with node glameow:
cpuhotplug03 1 TFAIL: No cpuhotplug_do_spin_loop processes found on CPU1

Kelsey Skunberg (kelsey-skunberg) wrote on 2020-10-06:
spotted on Focal aws : 5.4.0-1026.26 : amd64
cpuhotplug03 1 TFAIL: No cpuhotplug_do_spin_loop processes found on CPU1

Po-Hsu Lin (cypressyew) wrote on 2021-06-09:
Found on 5.11.0-1005.5 - intel
Node spitfire, passed on node bavor.
cpuhotplug03 1 TFAIL: No cpuhotplug_do_spin_loop processes found on CPU1

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu-kernel-tests/+bug/1931390/+subscriptions