mm / virtio-mem: support ZONE

From:		David Hildenbrand <david-AT-redhat.com>
To:		linux-kernel-AT-vger.kernel.org
Subject:		[PATCH v5 0/6] mm / virtio-mem: support ZONE_MOVABLE
Date:		Sun, 16 Aug 2020 14:53:27 +0200
Message-ID:		<20200816125333.7434-1-david@redhat.com>
Cc:		virtualization-AT-lists.linux-foundation.org, linux-mm-AT-kvack.org, David Hildenbrand <david-AT-redhat.com>, Andrew Morton <akpm-AT-linux-foundation.org>, Baoquan He <bhe-AT-redhat.com>, Jason Wang <jasowang-AT-redhat.com>, "Michael S. Tsirkin" <mst-AT-redhat.com>, Michal Hocko <mhocko-AT-suse.com>, Mike Kravetz <mike.kravetz-AT-oracle.com>, Mike Rapoport <rppt-AT-kernel.org>, Mike Rapoport <rppt-AT-linux.ibm.com>, Pankaj Gupta <pankaj.gupta.linux-AT-gmail.com>, Qian Cai <cai-AT-lca.pw>
Archive-link:		Article

For 5.10. Patch #1-#4,#6 have RBs or ACKs, patch #5 is virtio-mem stuff
maintained by me. This should go via the -mm tree.

---

When introducing virtio-mem, the semantics of ZONE_MOVABLE were rather
unclear, which is why we special-cased ZONE_MOVABLE such that partially
plugged blocks would never end up in ZONE_MOVABLE.

Now that the semantics are much clearer (and are documented in patch #6),
let's support partially plugged memory blocks in ZONE_MOVABLE, allowing
partially plugged memory blocks to be online to ZONE_MOVABLE and also
unplugging from such memory blocks. This avoids surprises when onlining
of memory blocks suddenly fails, just because they are not completely
populated by virtio-mem (yet).

This is especially helpful for testing, but also paves the way for
virtio-mem optimizations, allowing more memory to get reliably unplugged.

Cleanup has_unmovable_pages() and set_migratetype_isolate(), providing
better documentation of how ZONE_MOVABLE interacts with different kind of
unmovable pages (memory offlining vs. alloc_contig_range()).

v4 -> v5:
- Rename "mm/page_isolation: don't dump_page(NULL) in
  set_migratetype_isolate()" to "mm/page_isolation: exit early when
  pageblock is isolated in set_migratetype_isolate()" as I was messing
  up things while reshuffling patches (dump_page(NULL) could never happen,
  only the WARN_ON_ONCE())
-- Clarify in the description that this is currently a cleanup only
- "mm: document semantics of ZONE_MOVABLE"
-- Clarified in the memory hole case, that kernelcore/movablecore is
   required to create such rare special cases

v3 -> v4:
- "mm/page_isolation: drop WARN_ON_ONCE() in set_migratetype_isolate()"
-- Fix typo in description
- "virtio-mem: don't special-case ZONE_MOVABLE"
-- Add more details why we initialli special-cased ZONE_MOVABLE (via MST)
- "mm: document semantics of ZONE_MOVABLE"
-- Rephrase some parts of documentation (via Mike)

v2 -> v3:
- "mm: document semantics of ZONE_MOVABLE"
-- Fix a typo

v1 -> v2:
- "mm/page_isolation: don't dump_page(NULL) in set_migratetype_isolate()"
-- Move to position 1, add Fixes: tag
-- Drop unused "out:" label
- "mm/page_isolation: drop WARN_ON_ONCE() in set_migratetype_isolate()"
-- Keep curly braces on "else" case
- Replace "[PATCH v1 5/6] mm/page_alloc: restrict ZONE_MOVABLE optimization
           in has_unmovable_pages() to memory offlining"
  by "mm: document semantics of ZONE_MOVABLE"
-- Brain dump of what I know about ZONE_MOVABLE

David Hildenbrand (6):
  mm/page_alloc: tweak comments in has_unmovable_pages()
  mm/page_isolation: exit early when pageblock is isolated in
    set_migratetype_isolate()
  mm/page_isolation: drop WARN_ON_ONCE() in set_migratetype_isolate()
  mm/page_isolation: cleanup set_migratetype_isolate()
  virtio-mem: don't special-case ZONE_MOVABLE
  mm: document semantics of ZONE_MOVABLE

 drivers/virtio/virtio_mem.c | 47 +++++++------------------------------
 include/linux/mmzone.h      | 35 +++++++++++++++++++++++++++
 mm/page_alloc.c             | 22 +++++------------
 mm/page_isolation.c         | 39 ++++++++++++++----------------
 4 files changed, 66 insertions(+), 77 deletions(-)

-- 
2.26.2

mm / virtio-mem: support ZONE_MOVABLE