enosys test in ubuntu_stress_smoke will failed on Artful

Bug #1750787 reported by Po-Hsu Lin
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
Fix Released
High
Colin Ian King
stress-ng (Ubuntu)
Fix Released
High
Colin Ian King

Bug Description

The ubuntu_stress_smoke test will fail with enosys on node "gonzo" with Artful kernel (only on this node)

This can be reproduced in 4.13.0-32 and 4.13.0-36

enosys STARTING
enosys RETURNED 2
enosys FAILED
stress-ng: debug: [22679] 6 processors online, 6 processors configured
stress-ng: info: [22679] dispatching hogs: 4 enosys
stress-ng: debug: [22679] cache allocate: default cache size: 8192K
stress-ng: debug: [22679] starting stressors
stress-ng: debug: [22680] stress-ng-enosys: started [22680] (instance 0)
stress-ng: debug: [22681] stress-ng-enosys: started [22681] (instance 1)
stress-ng: debug: [22679] 4 stressors spawned
stress-ng: debug: [22682] stress-ng-enosys: started [22682] (instance 2)
stress-ng: debug: [22683] stress-ng-enosys: started [22683] (instance 3)
stress-ng: debug: [22682] stress-ng-enosys: exited [22682] (instance 2)
stress-ng: debug: [22680] stress-ng-enosys: exited [22680] (instance 0)
stress-ng: debug: [22679] process [22680] terminated
stress-ng: debug: [22679] process [22681] (stress-ng-enosys) terminated on signal: 9 (Killed)
stress-ng: debug: [22679] process [22681] terminated
stress-ng: debug: [22679] process [22682] terminated
stress-ng: debug: [22679] process [22683] (stress-ng-enosys) terminated on signal: 9 (Killed)
stress-ng: debug: [22679] process [22683] terminated
stress-ng: info: [22679] unsuccessful run completed in 7.40s

ProblemType: Bug
DistroRelease: Ubuntu 17.10
Package: linux-image-4.13.0-32-generic 4.13.0-32.35
ProcVersionSignature: User Name 4.13.0-32.35-generic 4.13.13
Uname: Linux 4.13.0-32-generic x86_64
AlsaDevices:
 total 0
 crw-rw---- 1 root audio 116, 1 Feb 21 10:59 seq
 crw-rw---- 1 root audio 116, 33 Feb 21 10:59 timer
AlsaVersion: Advanced Linux Sound Architecture Driver Version k4.13.0-32-generic.
AplayDevices: Error: [Errno 2] No such file or directory: 'aplay': 'aplay'
ApportVersion: 2.20.7-0ubuntu3.7
Architecture: amd64
ArecordDevices: Error: [Errno 2] No such file or directory: 'arecord': 'arecord'
AudioDevicesInUse: Error: command ['fuser', '-v', '/dev/snd/seq', '/dev/snd/timer'] failed with exit code 1:
Date: Wed Feb 21 11:03:28 2018
IwConfig: Error: [Errno 2] No such file or directory: 'iwconfig': 'iwconfig'
MachineType: Dell Inc. PowerEdge R415
PciMultimedia:

ProcEnviron:
 TERM=xterm-256color
 PATH=(custom, no user)
 XDG_RUNTIME_DIR=<set>
 LANG=C.UTF-8
 SHELL=/bin/bash
ProcFB: 0 mgadrmfb
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-4.13.0-32-generic root=UUID=3e0e190e-d425-4d00-a9a9-deb0c452399b ro
RelatedPackageVersions:
 linux-restricted-modules-4.13.0-32-generic N/A
 linux-backports-modules-4.13.0-32-generic N/A
 linux-firmware 1.169.3
RfKill: Error: [Errno 2] No such file or directory: 'rfkill': 'rfkill'
SourcePackage: linux
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 04/26/2012
dmi.bios.vendor: Dell Inc.
dmi.bios.version: 1.9.3
dmi.board.name: 08WNM9
dmi.board.vendor: Dell Inc.
dmi.board.version: A02
dmi.chassis.type: 23
dmi.chassis.vendor: Dell Inc.
dmi.modalias: dmi:bvnDellInc.:bvr1.9.3:bd04/26/2012:svnDellInc.:pnPowerEdgeR415:pvr:rvnDellInc.:rn08WNM9:rvrA02:cvnDellInc.:ct23:cvr:
dmi.product.name: PowerEdge R415
dmi.sys.vendor: Dell Inc.

Revision history for this message
Po-Hsu Lin (cypressyew) wrote :
Changed in linux (Ubuntu):
assignee: nobody → Colin Ian King (colin-king)
Revision history for this message
Ubuntu Kernel Bot (ubuntu-kernel-bot) wrote : Status changed to Confirmed

This change was made by a bot.

Changed in linux (Ubuntu):
status: New → Confirmed
Revision history for this message
Colin Ian King (colin-king) wrote :

Looks like the enosys stressor was being prematurely killed by the OOM killer, I'll add some extra wrapping to detect this and try to re-start the stressor when it gets OOM'd. This does not appear to be a kernel bug per se.

Changed in linux (Ubuntu):
importance: Undecided → High
status: Confirmed → In Progress
Changed in stress-ng (Ubuntu):
status: New → In Progress
importance: Undecided → High
assignee: nobody → Colin Ian King (colin-king)
Revision history for this message
Colin Ian King (colin-king) wrote :
Changed in stress-ng (Ubuntu):
status: In Progress → Fix Committed
Revision history for this message
Colin Ian King (colin-king) wrote :

Do you mind re-running the test to see if my fix addressed the issue?

Revision history for this message
Po-Hsu Lin (cypressyew) wrote :

The very same node has passed the enosys test with Artful kernel.
Thanks!

Revision history for this message
Colin Ian King (colin-king) wrote :

Fix will land in the stress-ng 0.09.17 release in the next 24 hours

Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package stress-ng - 0.09.17-1

---------------
stress-ng (0.09.17-1) unstable; urgency=medium

  * Makefile: bump version
  * Manual: update date
  * stress-rawdev: make some functions static
  * stress-funcall: fix -Wmissing-declarations warning message
  * stress-file-ioctl: remove extraneous shadowed variable 'ret'
  * stress-madvise: make ret non-clobberable for gcc 4.2.1
  * Add minimum GCC version for Pragma push/pop
  * shim: older gcc required pragmas outside of functions
  * stress-enosys: older gcc required pragmas outside of functions
  * Check for clang version when using pragmas
  * stress-file-ioctl: remove debug printf
  * Add new file-ioctl stressor
  * shim: disable warnings for brk and sbrk for OS X
  * Swap ordering on compiler detection on pragmas
  * stress-enosys: turn off warning on syscall, cleans up OS X warning
  * Add no warning pragma macros
  * Makefile: force sync after build in case reboot loses executable
  * stress-enosys: add syscall7 shim
  * stress-dirdeep: only print info message on instance #0
  * stress-enosys: wrap with OOM catcher (LP: #1750787)
  * stress-dev: add bug reference to hpet failure
  * stress-enosys: exercise more arguments in syscall
  * Manual: udp-lite should be for ipv6 as well as ipv4
  * Manual: update arch specific limitations on lockbus and memthrash
    stressors
  * Return EXIT_NOT_IMPLEMENTED when system call returns ENOSYS
  * stress-userfaultfd: don't fail with an error if ENOSYS is returned
  * stress-membarrier: don't fail with an error if ENOSYS is returned
  * stress-getrandom: don't fail with an error if ENOSYS is returned
  * helper: make zero more obscure so that optimizer won't optimize
    zero out
  * stress-filename: remove shadowed redundant variable ret
  * Makefile: build perf-event with $(CC) rather gcc
  * stress-wcstr: fix warning on systems without wcs functions
  * job: make fp non-clobberable, cleans up build warning on gcc 4.8.4
  * stress-softlock: clean up clobbering warning on gcc 4.9.2
  * mwc: add more information into the reseed mix
  * stress-filename: don't closedir on a NULL dir pointer
  * Move LIB_DL headers to stress-ng.h
  * stress-dynlib: try to read the function memory to force page fault
  * Manual: fix enosys-ops entry, missing e from enosys-ops
  * Add new dynlib shared library dynamic loading stressor
  * stress-tree: make nodes align naturally with padding
  * stress-tree: move data to start of struct, ensures natural alignment
  * stress-tree: clean splay nodes at removal time
  * stress-tree: ensure splay tree data is cleared
  * stress-exec: drop all capabilities before exec'ing
  * stack-mmap: add MAP_STACK flag

 -- Colin King <email address hidden> Thu, 22 Feb 2018 14:04:11 +0000

Changed in stress-ng (Ubuntu):
status: Fix Committed → Fix Released
Changed in linux (Ubuntu):
status: In Progress → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

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