mbox series

[v3,0/3] Various memory tiering fixes

Message ID 20240724130115.793641-1-ziy@nvidia.com (mailing list archive)
Headers show
Series Various memory tiering fixes | expand

Message

Zi Yan July 24, 2024, 1:01 p.m. UTC
Changelog
===

From v2:
1. Renamed folio_has_cpupid() to folio_use_access_time() and reverse the
code logic, since folio_has_cpupid() = !folio_use_access_time().
2. Wrapped folio_use_access_time() with CONFIG_NUMA_BALANCING to avoid the
function redefinition compilation error.

From v1:
1. Added a memory tiering runtime check in migrate_misplaced_folio() to
correct PGPROMOTE_SUCCESS counting.
2. Moved folio_has_cpupid() from memory-tiers.h to mm.h.

From RFC:
1. Added Reported-by and Closes flags to first patch

Zi Yan (3):
  memory tiering: read last_cpupid correctly in do_huge_pmd_numa_page()
  memory tiering: introduce folio_use_access_time() check
  memory tiering: count PGPROMOTE_SUCCESS when mem tiering is enabled.

 include/linux/mm.h  |  6 ++++++
 kernel/sched/fair.c |  3 +--
 mm/huge_memory.c    |  5 ++---
 mm/memory-tiers.c   | 19 +++++++++++++++++++
 mm/memory.c         |  3 +--
 mm/migrate.c        |  4 +++-
 mm/mprotect.c       |  3 +--
 7 files changed, 33 insertions(+), 10 deletions(-)