On 2024/3/1 5:20, Matthew Wilcox (Oracle) wrote: > This isn't my complete collection of patches to memory-failure, but > they're the ones I have the most confidence in. I haven't done any real > testing of this beyond compiling & running on a system without trying to > do any memory failure injections. Is there a good test-suite I should > run before submitting these kinds of patches? mm_regression[1] should be a good alternative test-suite to run. It's written by Naoya Horiguchi and available at https://github.com/nhoriguchi/mm_regression. > > Matthew Wilcox (Oracle) (8): > mm/memory-failure: Remove fsdax_pgoff argument from __add_to_kill > mm/memory-failure: Pass addr to __add_to_kill() > mm: Return the address from page_mapped_in_vma() > mm/memory-failure: Convert shake_page() to shake_folio() > mm: Convert hugetlb_page_mapping_lock_write to folio > mm/memory-failure: Convert memory_failure() to use a folio > mm/memory-failure: Convert hwpoison_user_mappings to take a folio > mm/memory-failure: Add some folio conversions to unpoison_memory Thanks for your patches. I will review them as soon as possible. > > include/linux/hugetlb.h | 6 +- > include/linux/mm.h | 1 - > include/linux/rmap.h | 2 +- > mm/hugetlb.c | 6 +- > mm/hwpoison-inject.c | 11 +-- > mm/internal.h | 1 + > mm/memory-failure.c | 144 +++++++++++++++++++++------------------- > mm/migrate.c | 2 +- > mm/page_vma_mapped.c | 14 ++-- > 9 files changed, 97 insertions(+), 90 deletions(-) >
