kernel-packages team mailing list archive

[bugproxy <bugproxy@xxxxxxxxxx>]

------- Comment From a.efremkin@xxxxxxxxxx 2016-03-23 12:01 EDT-------
On kernel 4.4.0-15:
CONFIG_NUMA_BALANCING_DEFAULT_ENABLED is still y, though requested "disabled"

root@s83lp41:~# cat /boot/config-4.4.0-15-generic | grep _NUMA
CONFIG_ARCH_SUPPORTS_NUMA_BALANCING=y
CONFIG_NUMA_BALANCING=y
CONFIG_NUMA_BALANCING_DEFAULT_ENABLED=y
CONFIG_NUMA=y
CONFIG_NUMA_EMU=y

root@s83lp41:~# cat /boot/config-4.4.0-15-generic | grep _EMU_SIZE
CONFIG_EMU_SIZE=0x10000000

root@s83lp41:~# uname -a
Linux s83lp41 4.4.0-15-generic #31-Ubuntu SMP Fri Mar 18 19:07:12 UTC 2016 s390x s390x s390x GNU/Linux

-- 
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/1557690

Title:
  s390/kconfig: CONFIG_NUMA without CONFIG_NUMA_EMU does not make any
  sense on s390x

Status in linux package in Ubuntu:
  Fix Released
Status in linux source package in Xenial:
  Fix Released

Bug description:
  == Comment: #0 - Hendrik Brueckner - 2016-03-15 06:33:03 ==
  For NUMA, ensure the setting of these kernel configuration for s390x:

  # CONFIG_NUMA_BALANCING_DEFAULT_ENABLED is not set
  CONFIG_NUMA_EMU=y
  CONFIG_EMU_SIZE=0x10000000

  
  CONFIG_NUMA without CONFIG_NUMA_EMU does not make any sense. s390 does not provide hardware topology information. However, we do support fake NUMA to improve performance issues seen with very large memory configurations. Therefore, CONFIG_NUMA_EMU has to be enabled.

  Given that we only support fake NUMA
  CONFIG_NUMA_BALANCING_DEFAULT_ENABLED should be disabled, since it
  moves pages without any benefit but costs additional cpu cycles.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1557690/+subscriptions