Re: [PATCH v2] mm: Only re-generate demotion targets when a numa node changes its N_CPU state

From: Andrew Morton <akpm@linux-foundation.org>
To: Oscar Salvador <osalvador@suse.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>,
	"Huang, Ying" <ying.huang@intel.com>,
	Abhishek Goel <huntbag@linux.vnet.ibm.com>,
	Baolin Wang <baolin.wang@linux.alibaba.com>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2] mm: Only re-generate demotion targets when a numa node changes its N_CPU state
Date: Thu, 10 Mar 2022 18:39:51 -0800	[thread overview]
Message-ID: <20220310183951.cb713c6ae926ea6ea8489a71@linux-foundation.org> (raw)
In-Reply-To: <20220310120749.23077-1-osalvador@suse.de>

On Thu, 10 Mar 2022 13:07:49 +0100 Oscar Salvador <osalvador@suse.de> wrote:

> Abhishek reported that after patch [1], hotplug operations are
> taking ~double the expected time. [2]
> 
> The reason behind is that the CPU callbacks that migrate_on_reclaim_init()
> sets always call set_migration_target_nodes() whenever a CPU is brought
> up/down.
> But we only care about numa nodes going from having cpus to become
> cpuless, and vice versa, as that influences the demotion_target order.
> 
> We do already have two CPU callbacks (vmstat_cpu_online() and vmstat_cpu_dead())
> that check exactly that, so get rid of the CPU callbacks in
> migrate_on_reclaim_init() and only call set_migration_target_nodes() from
> vmstat_cpu_{dead,online}() whenever a numa node change its N_CPU state.

What I'm not getting here (as so often happens) is a sense of how badly
this affects our users.  Does anyone actually hotplug frequently enough
to care?

If "yes" then I'm inclined to merge this up for 5.18 with a cc:stable. 
Not for 5.17 because it's late and things are looking rather creaky
already.

And I'll add a

Fixes: 884a6e5d1f93b ("mm/migrate: update node demotion order on hotplug events")

which is that patch's fourth such bouquet.