sysfs test in ubuntu_stress_smoke_test will hang on a X-HWE lowlatency node
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
ubuntu-kernel-tests |
Fix Released
|
Undecided
|
Unassigned | ||
stress-ng (Ubuntu) |
Fix Released
|
High
|
Colin Ian King |
Bug Description
ubuntu@pepe:~$ uname -a
Linux pepe 4.15.0-
The sysfs test in ubuntu_
Sep 7 06:50:01 pepe stress-ng: invoked with './stress-n' by user 0
Sep 7 06:50:01 pepe stress-ng: system: 'pepe' Linux 4.15.0-
Sep 7 06:50:01 pepe stress-ng: memory (MB): total 3989.54, free 3595.64, shared 14.70, buffer 92.63, swap 1023.99, free swap 940.58
Sep 7 06:50:01 pepe stress-ng: info: [27744] dispatching hogs: 4 sysinfo
Sep 7 06:50:06 pepe stress-ng: info: [27744] successful run completed in 5.01s
Sep 7 06:50:06 pepe stress-ng: invoked with './stress-n' by user 0
Sep 7 06:50:06 pepe stress-ng: system: 'pepe' Linux 4.15.0-
Sep 7 06:50:06 pepe stress-ng: memory (MB): total 3989.54, free 3594.25, shared 14.70, buffer 92.64, swap 1023.99, free swap 940.58
Sep 7 06:50:06 pepe stress-ng: info: [27763] dispatching hogs: 4 sysfs
Sep 7 06:50:09 pepe kernel: [ 1014.787741] WARNING! power/level is deprecated; use power/control instead
There are many repetitive message in dmesg right after triggering the test with:
$ ./stress-ng -t 5 --sysfs 4 --ignite-cpu --syslog --verbose --verify --oomable
[ 1655.642421] mpt2sas_cm0: _ctl_BRM_
[ 1655.642429] mpt2sas_cm0: _ctl_BRM_
[ 1655.642440] mpt2sas_cm0: _ctl_BRM_
[ 1655.642454] mpt2sas_cm0: _ctl_BRM_
[ 1655.642463] mpt2sas_cm0: _ctl_BRM_
ProblemType: Bug
DistroRelease: Ubuntu 16.04
Package: stress-ng (not installed)
ProcVersionSign
Uname: Linux 4.15.0-
ApportVersion: 2.20.1-0ubuntu2.18
Architecture: i386
Date: Fri Sep 7 06:57:16 2018
SourcePackage: stress-ng
UpgradeStatus: No upgrade log present (probably fresh install)
Hangs when accessing:
/sys/devices/ pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/version_ product pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/prot_ guard_type pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/ioc_ reset_count pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/version_ bios pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/sg_ tablesize pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/device_ delay pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/board_ tracer pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/host_ sas_address pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/diag_ trigger_ mpi pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/logging_ level pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/fwfault_ debug pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/prot_ capabilities pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/unchecked _isa_dma pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/reply_ queue_count pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/sg_ prot_tablesize pci0000: 00/0000: 00:05.0/ 0000:05: 00.0/host4/ scsi_host/ host4/host_ trace_buffer
/sys/devices/
/sys/devices/
/sys/devices/
/sys/devices/
/sys/devices/
/sys/devices/
/sys/devices/
/sys/devices/
/sys/devices/
/sys/devices/
/sys/devices/
/sys/devices/
/sys/devices/
/sys/devices/
/sys/devices/
..so probably the host_trace_buffer is the problematic sysfs file.