/proc/pid/numa_maps no longer shows "default" policy
Hi Mel,
We recently noticed that /proc/pid/numa_maps used to show default policy mappings as such:
cat /proc/self/numa_maps
00100000 default mapped=1 mapmax=339 active=0 N0=1
But now it shows them as prefer:X:
cat /proc/self/numa_maps
10000000 prefer:1 file=/usr/bin/cat mapped=1 N0=1
It looks like this was caused by 5606e387 (mm: numa: Migrate on reference policy). I'm not sure if this is expected, but we don't have CONFIG_NUMA_BALANCING enabled on ppc64 so I wasn't expecting processes to have a particular node affinity by default.
Anton
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Quoted 11 lines · Anton Blanchard
Hi Mel,
We recently noticed that
/proc/pid/numa_mapsused to show default policy mappings as such:cat
/proc/self/numa_maps
00100000 default mapped=1 mapmax=339 active=0 N0=1But now it shows them as prefer:X:
Quoted 8 lines
cat
/proc/self/numa_maps
10000000 prefer:1 file=/usr/bin/cat mapped=1 N0=1It looks like this was caused by
5606e387(mm: numa: Migrate on reference policy). I'm not sure if this is expected, but we don't haveCONFIG_NUMA_BALANCINGenabled on ppc64 so I wasn't expecting processes to have a particular node affinity by default.
https://lkml.org/lkml/2014/1/25/182
Signature · 8 lines
Mel Gorman SUSE Labs -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>