The comment removed in this patch originally belonged to the
build_zonelists_in_zone_order() function, which was introduced by commit
f0c0b2b808f2 ("change zonelist order: zonelist order selection logic").

Later, commit c9bff3eebc09 ("mm, page_alloc: rip out ZONELIST_ORDER_ZONE")
removed build_zonelists_in_zone_order() but left its comment behind.

Subsequently, commit 9d3be21bf9c0 ("mm, page_alloc: simplify zonelist
initialization") moved the node_order variable into build_zonelists(),
making the comment originally belonged to build_zonelists_in_zone_order()
appear as if it were part of build_zonelists().

Remove this misleading comment.

Signed-off-by: Yuntao Wang <yuntao.wang@linux.dev>
---
 mm/page_alloc.c | 7 -------
 1 file changed, 7 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index cae7b93864c2..ff23296e30dc 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -5161,13 +5161,6 @@ static void build_thisnode_zonelists(pg_data_t *pgdat)
 	zonerefs->zone_idx = 0;
 }
 
-/*
- * Build zonelists ordered by zone and nodes within zones.
- * This results in conserving DMA zone[s] until all Normal memory is
- * exhausted, but results in overflowing to remote node while memory
- * may still exist in local DMA zone.
- */
-
 static void build_zonelists(pg_data_t *pgdat)
 {
 	static int node_order[MAX_NUMNODES];
-- 
2.47.0

On Wed, 15 Jan 2025 12:16:34 +0800 Yuntao Wang <yuntao.wang@linux.dev> wrote:

> The comment removed in this patch originally belonged to the
> build_zonelists_in_zone_order() function, which was introduced by commit
> f0c0b2b808f2 ("change zonelist order: zonelist order selection logic").
> 
> Later, commit c9bff3eebc09 ("mm, page_alloc: rip out ZONELIST_ORDER_ZONE")
> removed build_zonelists_in_zone_order() but left its comment behind.
> 
> Subsequently, commit 9d3be21bf9c0 ("mm, page_alloc: simplify zonelist
> initialization") moved the node_order variable into build_zonelists(),
> making the comment originally belonged to build_zonelists_in_zone_order()
> appear as if it were part of build_zonelists().
> 
> ...
>
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -5161,13 +5161,6 @@ static void build_thisnode_zonelists(pg_data_t *pgdat)
>  	zonerefs->zone_idx = 0;
>  }
>  
> -/*
> - * Build zonelists ordered by zone and nodes within zones.
> - * This results in conserving DMA zone[s] until all Normal memory is
> - * exhausted, but results in overflowing to remote node while memory
> - * may still exist in local DMA zone.
> - */
> -

Yes, it's not obvious which code this comment is telling us about.

I do wonder if the information in this comment could be updated and
moved somewhere more appropriate, rather than simply removing it.  Oh
well, not very important.