mm: fix finish_fault() handling for large folios

JIRA: https://issues.redhat.com/browse/RHEL-83854
Tested: by me

This one was a hard one to track. Trinity would trigger this but
sometimes taking multiple runs. We might need
3685024ed backported in the future too, but I prefer not to mix up both
in the same ticket.

Signed-off-by: Aristeu Rozanski arozansk@redhat.com

Merge request reports

Loading