ubuntu-kernel-tests

ubuntu_kernel_selftests: ./cpu-on-off-test.sh: line 94: echo: write error: Device or resource busy

Bug #1923114 reported by Marcelo Cerri on 2021-04-08

This bug affects 1 person

	Status	Importance	Assigned to
ubuntu-kernel-tests	In Progress	Undecided	Krzysztof Kozlowski
linux-azure (Ubuntu)	New	Undecided	Unassigned
Trusty	New	Undecided	Unassigned
Xenial	New	Undecided	Unassigned
Bionic	New	Undecided	Unassigned
Groovy	Won't Fix	Undecided	Unassigned
linux-azure-4.15 (Ubuntu)	New	Undecided	Unassigned
Trusty	New	Undecided	Unassigned
Xenial	New	Undecided	Unassigned
Bionic	New	Undecided	Unassigned
Groovy	Won't Fix	Undecided	Unassigned

Bug Description

Test cpu-hotplug from ubuntu_kernel_selftests failed with bionic:linux-azure-4.15 running on a Basic A2 with 2 cores (besides other instance types):

selftests: cpu-on-off-test.sh
========================================
pid 28041's current affinity mask: 3
pid 28041's new affinity mask: 1
CPU online/offline summary:
present_cpus = 0-1 present_max = 1
Cpus in online state: 0-1
Cpus in offline state: 0
Limited scope test: one hotplug cpu
(leaves cpu in the original state):
online to offline to online: cpu 1
not ok 1..1 selftests: cpu-on-off-test.sh [FAIL]
./cpu-on-off-test.sh: line 94: echo: write error: Device or resource busy
offline_cpu_expect_success 1: unexpected fail

http://10.246.72.46/4.15.0-1112.124~16.04.1-azure/xenial-linux-azure-azure-amd64-4.15.0-Basic_A2-ubuntu_kernel_selftests/ubuntu_kernel_selftests/results/ubuntu_kernel_selftests.cpu-hotplug/debug/ubuntu_kernel_selftests.cpu-hotplug.DEBUG.html

The problem happens at "autotest-client-tests/ubuntu_kernel_selftests/cpu-on-off-test.sh" when executing:

echo 0 > $SYSFS/devices/system/cpu/cpu$1/online

Tags:

Marcelo Cerri (mhcerri) on 2021-04-09

tags:

added: 4.15 amd64 azure bionic kqa-blocker sru-20210315 trusty ubuntu-kernel-selftests xenial

Revision history for this message

Marcelo Cerri (mhcerri) wrote on 2021-04-09:

Also seen in linux-azure 5.8.0-1027.29 (and later) for groovy

tags:

added: 5.8 groovy

Revision history for this message

Marcelo Cerri (mhcerri) wrote on 2021-04-09:

Also seen in bionic linux-azure-fips 4.15.0-2024.27 for sru-20210315

tags:

added: azure-fips

Kleber Sacilotto de Souza (kleber-souza) on 2021-05-04

tags:

added: sru-20210412

Revision history for this message

Ian May (ian-may) wrote on 2021-05-04:

bionic/linux-azure-fips: 4.15.0-2026.29

Revision history for this message

Ian May (ian-may) wrote on 2021-05-04:

bionic/linux-azure-4.15: 4.15.0-1114.127

Revision history for this message

Ian May (ian-may) wrote on 2021-05-27:

Found on bionic/linux-azure-4.15: 4.15.0-1115.128

Revision history for this message

Ian May (ian-may) wrote on 2021-05-31:

Found on bionic/linux-azure-fips: 4.15.0-2027.30

Revision history for this message

Krzysztof Kozlowski (krzk) wrote on 2021-07-01:

Found on bionic/linux-azure: 4.15.0-1119.132

tags:

added: sru-20210621

Krzysztof Kozlowski (krzk) on 2021-07-01

Changed in ubuntu-kernel-tests:
assignee:	nobody → Krzysztof Kozlowski (krzk)
status:	New → In Progress

Revision history for this message

Thadeu Lima de Souza Cascardo (cascardo) wrote on 2021-09-10:

While looking for some improvements on cpu-on-off-test.sh, I investigated the path that is taken on azure. One good tool to debug these failures is the cpuhp tracing.

echo 1 > /sys/kernel/debug/tracing/events/cpuhp/enable
echo 1 > /sys/kernel/debug/tracing/tracing_on
cat /sys/kernel/debug/tracing/trace_pipe &
echo 0 > /sys/devices/system/cpu/cpu1/online

&nb cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 cpuhp/1-14 sp; bash-2660 [001] .... 10306.404476: cpuhp_enter: cpu: 0001 target: 143 step: 213 (cpuhp_kick_ap_work)
[001] .... 10306.404506: cpuhp_enter: cpu: 0001 target: 143 step: 212 (sched_cpu_deactivate)
[001] .... 10306.420476: cpuhp_exit: cpu: 0001 state: 211 step: 212 ret: 0
[001] .... 10306.420502: cpuhp_enter: cpu: 0001 target: 143 step: 189 (msr_device_destroy [msr])
[001] .... 10306.420623: cpuhp_exit: cpu: 0001 state: 188 step: 189 ret: 0
[001] .... 10306.420629: cpuhp_enter: cpu: 0001 target: 143 step: 187 (mce_cpu_pre_down)
[001] .... 10306.420653: cpuhp_exit: cpu: 0001 state: 186 step: 187 ret: 0
[001] .... 10306.420654: cpuhp_enter: cpu: 0001 target: 143 step: 184 (hv_synic_cleanup)
[001] .... 10306.420655: cpuhp_exit: cpu: 0001 state: 183 step: 184 ret: -16
[001] .... 10306.426523: cpuhp_enter: cpu: 0001 target: 213 step: 185 (compute_batch_value)
[001] .... 10306.426541: cpuhp_exit: cpu: 0001 state: 185 step: 185 ret: 0
[001] .... 10306.426546: cpuhp_enter: cpu: 0001 target: 213 step: 186 (acpi_soft_cpu_online)
[001] .... 10306.426553: cpuhp_exit: cpu: 0001 state: 186 step: 186 ret: 0
[001] .... 10306.426553: cpuhp_enter: cpu: 0001 target: 213 step: 187 (mce_cpu_online)
[001] .... 10306.426597: cpuhp_exit: cpu: 0001 state: 187 step: 187 ret: 0
[001] .... 10306.426606: cpuhp_enter: cpu: 0001 target: 213 step: 188 (console_cpu_notify)
[001] .... 10306.426607: cpuhp_exit: cpu: 0001 state: 188 step: 188 ret: 0
[001] .... 10306.426611: cpuhp_enter: cpu: 0001 target: 213 step: 189 (msr_device_create [msr])
[001] .... 10306.426665: cpuhp_exit: cpu: 0001 state: 189 step: 189 ret: 0
[001] .... 10306.426667: cpuhp_enter: cpu: 0001 target: 213 step: 212 (sched_cpu_activate)
[001] .... 10306.426670: cpuhp_exit: cpu: 0001 state: 212 step: 212 ret: 0

Notice the -16 failure on hv_synic_cleanup. It can fail for two reasons: the cpu is VMBUS_CONNECT_CPU, which is 0, not the case here, or the cpu is the target_cpu for a vmbus_channel.

# grep 1 /sys/bus/vmbus/devices/*/channels/*/cpu
/sys/bus/vmbus/devices/00000000-0001-8899-0000-000000000000/channels/3/cpu:1
/sys/bus/vmbus/devices/000d3a6e-002e-000d-3a6e-002e000d3a6e/channels/15/cpu:1
/sys/bus/vmbus/devices/f8b3781b-1e82-4818-a1c3-63d806ec15bb/channels/13/cpu:1

While looking for some improvements on cpu-on-off-test.sh, I investigated the path that is taken on azure. One good tool to debug these failures is the cpuhp tracing.

echo 1 > /sys/kernel/debug/tracing/events/cpuhp/enable
echo 1 > /sys/kernel/debug/tracing/tracing_on
cat /sys/kernel/debug/tracing/trace_pipe &
echo 0 > /sys/devices/system/cpu/cpu1/online

bash-2660    [001] .... 10306.404476: cpuhp_enter: cpu: 0001 target: 143 step: 213 (cpuhp_kick_ap_work)
         cpuhp/1-14      [001] .... 10306.404506: cpuhp_enter: cpu: 0001 target: 143 step: 212 (sched_cpu_deactivate)
         cpuhp/1-14      [001] .... 10306.420476: cpuhp_exit:  cpu: 0001  state: 211 step: 212 ret: 0
         cpuhp/1-14      [001] .... 10306.420502: cpuhp_enter: cpu: 0001 target: 143 step: 189 (msr_device_destroy [msr])
         cpuhp/1-14      [001] .... 10306.420623: cpuhp_exit:  cpu: 0001  state: 188 step: 189 ret: 0
         cpuhp/1-14      [001] .... 10306.420629: cpuhp_enter: cpu: 0001 target: 143 step: 187 (mce_cpu_pre_down)
         cpuhp/1-14      [001] .... 10306.420653: cpuhp_exit:  cpu: 0001  state: 186 step: 187 ret: 0
         cpuhp/1-14      [001] .... 10306.420654: cpuhp_enter: cpu: 0001 target: 143 step: 184 (hv_synic_cleanup)
         cpuhp/1-14      [001] .... 10306.420655: cpuhp_exit:  cpu: 0001  state: 183 step: 184 ret: -16
         cpuhp/1-14      [001] .... 10306.426523: cpuhp_enter: cpu: 0001 target: 213 step: 185 (compute_batch_value)
         cpuhp/1-14      [001] .... 10306.426541: cpuhp_exit:  cpu: 0001  state: 185 step: 185 ret: 0
         cpuhp/1-14      [001] .... 10306.426546: cpuhp_enter: cpu: 0001 target: 213 step: 186 (acpi_soft_cpu_online)
         cpuhp/1-14      [001] .... 10306.426553: cpuhp_exit:  cpu: 0001  state: 186 step: 186 ret: 0
         cpuhp/1-14      [001] .... 10306.426553: cpuhp_enter: cpu: 0001 target: 213 step: 187 (mce_cpu_online)
         cpuhp/1-14      [001] .... 10306.426597: cpuhp_exit:  cpu: 0001  state: 187 step: 187 ret: 0
         cpuhp/1-14      [001] .... 10306.426606: cpuhp_enter: cpu: 0001 target: 213 step: 188 (console_cpu_notify)
         cpuhp/1-14      [001] .... 10306.426607: cpuhp_exit:  cpu: 0001  state: 188 step: 188 ret: 0
         cpuhp/1-14      [001] .... 10306.426611: cpuhp_enter: cpu: 0001 target: 213 step: 189 (msr_device_create [msr])
         cpuhp/1-14      [001] .... 10306.426665: cpuhp_exit:  cpu: 0001  state: 189 step: 189 ret: 0
         cpuhp/1-14      [001] .... 10306.426667: cpuhp_enter: cpu: 0001 target: 213 step: 212 (sched_cpu_activate)
         cpuhp/1-14      [001] .... 10306.426670: cpuhp_exit:  cpu: 0001  state: 212 step: 212 ret: 0

Notice the -16 failure on hv_synic_cleanup. It can fail for two reasons: the cpu is VMBUS_CONNECT_CPU, which is 0, not the case here, or the cpu is the target_cpu for a vmbus_channel.

Francis Ginther (fginther) on 2022-08-08

tags:

added: 5.4

Revision history for this message

Brian Murray (brian-murray) wrote on 2024-07-26:

Ubuntu 20.10 (Groovy Gorilla) has reached end of life, so this bug will not be fixed for that specific release.

Changed in linux-azure (Ubuntu Groovy):
status:	New → Won't Fix
Changed in linux-azure-4.15 (Ubuntu Groovy):
status:	New → Won't Fix

Juerg Haefliger (juergh) on 2025-04-18

tags:

added: kernel-daily-bug

Report a bug

This report contains Public information

Everyone can see this information.

You are

Subscribing...

Edit bug mail

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.