Test on instance loop in ftracetest from ubuntu_kernel_selftests/ftrace make P8 node dryden stop responding on I-5.13

On P8 node dryden with Impish 5.13.0-30

This ftracetest will take about 45 minutes to run through the first 80 tests and disconnects at the first "instance" test:
02:50:47 INFO | START ubuntu_kernel_selftests.ftrace:ftracetest ubuntu_kernel_selftests.ftrace:ftracetest timestamp=1646707847 timeout=10800 localtime=Mar 08 02:50:47
....
03:35:24 DEBUG| [stdout] # [80] event trigger - test traceon/off trigger [PASS]
(system disconnects here)

The 81st test caused this issue, which should be:
[81] (instance) Basic test for tracers

Nothing interesting from dmesg / syslog, a console watching dmesg with `sudo demsg -w` didn't print any info about this and the console just stop responding here.

This is the last few lines in both dmesg and syslog:
[ 3231.082626] trace_kprobe: Could not probe notrace function do_break
[ 3231.084505] trace_kprobe: Could not probe notrace function msr_check_and_set
[ 3231.093811] trace_kprobe: Could not probe notrace function restore_math
[ 3231.096524] trace_kprobe: Could not probe notrace function __switch_to_tm
[ 3231.110229] trace_kprobe: Could not probe notrace function __ppc64_runlatch_on
[ 3231.110514] trace_kprobe: Could not probe notrace function __ppc64_runlatch_off
client_loop: send disconnect: Broken pipe

This issue is 100% reproducible, you can modify the ftracetest script and make it skip the "# Main loop" section, just run the "# Test on instance loop" part.

Please don't use this bug on other HW, unless it has been manually verified with the same issue.