Re: [PATCH v5 3/3] mm/mempolicy: Support memory hotplug in weighted interleave

[Gregory Price]

On Wed, Apr 02, 2025 at 01:15:48AM -0400, Gregory Price wrote:
> On Tue, Apr 01, 2025 at 09:34:39PM -0700, Andrew Morton wrote:
> > 
> > Do we need to backport all three patches into -stable?  If so, all three
> > should have Fixes:.  Preferably they all have the same Fixes: so we
> > aren't backporting untested patch combinations.
> >
> 
> Patches 2 and 3 aren't fixes, they're desired behavioral changes
> to the feature. An interface being confusing doesn't mean it's broken.
>

Just some added clarity:

The behavioral change in sysfs change goes from N_POSSIBLE nodes always
being exposed - to sysfs entries coming and going as nodes transition
between having memory and not having memory.

I can see the argument for wanting this backported as a fix to ensure
userspace software is consistent across LTS.

So for backporting: all or nothing - and all the fixes tags should be
the same.

~Gregory