Re: [PATCH] mm, page_alloc: fix build_zonerefs_node()

[Juergen Gross]

On 07.04.22 11:46, David Hildenbrand wrote:
> On 07.04.22 11:32, Juergen Gross wrote:
> > Since commit 9d3be21bf9c0 ("mm, page_alloc: simplify zonelist
> > initialization") only zones with free memory are included in a built
> > zonelist. This is problematic when e.g. all memory of a zone has been
> > ballooned out.
> >
> > Use populated_zone() when building a zonelist as it has been done
> > before that commit.
> >
> > Cc: stable@xxxxxxxxxxxxxxx
> > Fixes: 9d3be21bf9c0 ("mm, page_alloc: simplify zonelist initialization")
> > Reported-by: Marek Marczykowski-Górecki <marmarek@xxxxxxxxxxxxxxxxxxxxxx>
> > Signed-off-by: Juergen Gross <jgross@xxxxxxxx>
> > ---
> >   mm/page_alloc.c | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > index bdc8f60ae462..3d0662af3289 100644
> > --- a/mm/page_alloc.c
> > +++ b/mm/page_alloc.c
> > @@ -6128,7 +6128,7 @@ static int build_zonerefs_node(pg_data_t *pgdat, struct zoneref *zonerefs)
> >   	do {
> >   		zone_type--;
> >   		zone = pgdat->node_zones + zone_type;
> > -		if (managed_zone(zone)) {
> > +		if (populated_zone(zone)) {
> >   			zoneref_set_zone(zone, &zonerefs[nr_zones++]);
> >   			check_highest_zone(zone_type);
> >   		}
>
> Let's see if we have to find another way to properly handle fadump.

TBH, I don't think this should matter here. A zone can always happen to
have no free memory left, so only handling this case when building the
zonelist can't be the full solution of that problem. It might trigger
a problem more easily, though.

> Acked-by: David Hildenbrand <david@xxxxxxxxxx>

Thanks,


Juergen

Attachment: OpenPGP_0xB0DE9DD628BF132F.asc
Description: OpenPGP public key

Attachment: OpenPGP_signature
Description: OpenPGP digital signature