ubuntu_ftrace_smoke_test crash on B-azure-5.4

[Po-Hsu Lin]

Issue found with Bionic Azure 5.4.0-1088.93~18.04.1

It looks like test will get interrupted across all Azure instances.

Test log:
 START ---- ---- timestamp=1658735692 localtime=Jul 25 07:54:52
 Persistent state client._record_indent now set to 1
 Test has timeout: 900 sec.
        START ubuntu_ftrace_smoke_test.ftrace-smoke-test ubuntu_ftrace_smoke_test.ftrace-smoke-test timestamp=1658735692 timeout=900 localtime=Jul 25 07:54:52
 Persistent state client._record_indent now set to 2
 Persistent state client.unexpected_reboot now set to ('ubuntu_ftrace_smoke_test.ftrace-smoke-test', 'ubuntu_ftrace_smoke_test.ftrace-smoke-test')
 Waiting for pid 3735 for 900 seconds
 Running '/home/azure/autotest/client/tests/ubuntu_ftrace_smoke_test/ubuntu_ftrace_smoke_test.sh'
 PASSED (CONFIG_FUNCTION_TRACER=y in /boot/config-5.4.0-1088-azure)
 PASSED (CONFIG_FUNCTION_GRAPH_TRACER=y in /boot/config-5.4.0-1088-azure)
 PASSED (CONFIG_STACK_TRACER=y in /boot/config-5.4.0-1088-azure)
 PASSED (CONFIG_DYNAMIC_FTRACE=y in /boot/config-5.4.0-1088-azure)
 PASSED all expected /sys/kernel/debug/tracing files exist
 PASSED (function_graph in /sys/kernel/debug/tracing/available_tracers)
 PASSED (function in /sys/kernel/debug/tracing/available_tracers)
 PASSED (nop in /sys/kernel/debug/tracing/available_tracers)
 PASSED (tracer function can be enabled)
(Test stops here, system stop responding)

dmesg, syslog and console log didn't provide useful information.

dmesg:
[ 8.389305] hv_netvsc 000d3ac3-4ce3-000d-3ac3-4ce3000d3ac3 eth0: Data path switched from VF: enP18310s1
[ 8.800147] mlx5_core 4786:00:02.0 enP18310s1: Link up
[ 8.801218] hv_netvsc 000d3ac3-4ce3-000d-3ac3-4ce3000d3ac3 eth0: Data path switched to VF: enP18310s1
[ 10.961896] EXT4-fs (sda1): mounted filesystem with ordered data mode. Opts: (null)
[ 11.021630] new mount options do not match the existing superblock, will be ignored
[ 53.490220] hv_balloon: Max. dynamic memory size: 65536 MB
(system stop responding after test starts)

syslog:
Jul 25 07:54:48 b-laz54-az-5-4-stdf32sv2-u-ftrace-smk-test systemd[1]: Started Session 12 of user azure.
Jul 25 07:54:49 b-laz54-az-5-4-stdf32sv2-u-ftrace-smk-test dbus-daemon[1522]: [system] Activating via systemd: service name='org.freedesktop.hostname1' unit='dbus-org.freedesktop.hostname1.service' requested by ':1.29' (uid=0 pid=3330 comm="systemd-analyze plot " label="unconfined")
Jul 25 07:54:49 b-laz54-az-5-4-stdf32sv2-u-ftrace-smk-test systemd[1]: Starting Hostname Service...
Jul 25 07:54:49 b-laz54-az-5-4-stdf32sv2-u-ftrace-smk-test dbus-daemon[1522]: [system] Successfully activated service 'org.freedesktop.hostname1'
Jul 25 07:54:49 b-laz54-az-5-4-stdf32sv2-u-ftrace-smk-test systemd[1]: Started Hostname Service.
(system stop responding after test starts)

Please find attachment for the console log.

[Thadeu Lima de Souza Cascardo (cascardo)]

This happens with 5.4.0-1086 as well. Just echo function_graph > current_tracer.

[Thadeu Lima de Souza Cascardo (cascardo)]

Reproduced on a Standard_A2_v2 instance. Could not reproduce on a local VM.