Skip to content

split vm_normal_pages for LRU and non-LRU handling

Alex Sierra amd-gfx dri-devel linux-ext4 linux-mm linux-xfs 2 versions – In review

Get series with b4
b4 am [email protected]

Versions newest first · linked by author and title

  1. v2 Latest3 patches 6 repliesIn review
  2. v13 patches 8 repliesSuperseded

Changes in v2 from the cover letter

- Changed the general description for this cover letter.
- Changed commit message for patch 1/3.
- Keep vm_normal_pages and add vm_normal_lru_pages, instead of rename both.
- Add proper kernel-doc format to new function and minimize code churn.

Patches in v2 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/3mm: add vm_normal_lru_pages for LRU handled pages onlyfs/proc/task_mmu.c include/linux/mm.h mm/gup.c +9 more+66 -256 replies
2/3tools: add more gup configs to hmm_gup selfteststools/testing/selftests/vm/hmm-tests.c+44 -210 replies
3/3tools: add selftests to hmm for COW in device memorytools/testing/selftests/vm/hmm-tests.c+80 -00 replies

Reviews trailers per patch

All 3 patches have a Reviewed-by or Acked-by.

PatchFelix Kuehling
1/3A-b
2/3A-b
3/3A-b

Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.

Read the v2 thread

Cover letter [PATCH v2 0/3] split vm_normal_pages for LRU and non-LRU handling

With DEVICE_COHERENT, we'll soon have vm_normal_pages() return device-managed anonymous pages that are not LRU pages. Although they behave like normal pages for purposes of mapping in CPU page, and for COW. They do not support LRU lists, NUMA migration or THP. The difference between new vm_normal_lru_pages vs vm_normal_pages() is, the former makes sure to return pages that are LRU handled only.

We also introduced a FOLL_LRU flag that adds the same behaviour to follow_page and related APIs, to allow callers to specify that they expect to put pages on an LRU list.

HMM tests were added to selftest to excercise these changes with device coherent pages. New test called hmm_cow_in_device, will test pages marked as COW, allocated in device zone. Also, more configurations were added into hmm_gup_test to test basic get user pages and get user pages fast paths in device zone pages.

v2:
- Changed the general description for this cover letter.
- Changed commit message for patch 1/3.
- Keep vm_normal_pages and add vm_normal_lru_pages, instead of rename both.
- Add proper kernel-doc format to new function and minimize code churn.

TODO: vm_normal_pages with pte_devmap entries still return NULL, instead of return the actual device page. The reason is page->_mapcount is never incremented for device pages that are mmap through DAX mechanism using pmem driver mounted into ext4 filesystem. When these pages are unmap, zap_pte_range is called and vm_normal_page return a valid page with page_mapcount() = 0, before page_remove_rmap is called.

Alex Sierra (3):
mm: add vm_normal_lru_pages for LRU handled pages only
tools: add more gup configs to hmm_gup selftests
tools: add selftests to hmm for COW in device memory

 fs/proc/task_mmu.c                     |   2 +-
 include/linux/mm.h                     |   9 +-
 mm/gup.c                               |   8 +-
 mm/huge_memory.c                       |   2 +-
 mm/khugepaged.c                        |   8 +-
 mm/ksm.c                               |   4 +-
 mm/madvise.c                           |   4 +-
 mm/memory.c                            |  40 ++++++-
 mm/mempolicy.c                         |   4 +-
 mm/migrate.c                           |   2 +-
 mm/mlock.c                             |   6 +-
 mm/mprotect.c                          |   2 +-
 tools/testing/selftests/vm/hmm-tests.c | 139 +++++++++++++++++++++----
 13 files changed, 187 insertions(+), 43 deletions(-)
Signature · 1 line
2.32.0