On Tue, Aug 16, 2022 at 3:39 PM Alistair Popple <apopple@xxxxxxxxxx> wrote:
>
> migrate_vma_setup() has a fast path in migrate_vma_collect_pmd() that
> installs migration entries directly if it can lock the migrating page.
> When removing a dirty pte the dirty bit is supposed to be carried over
> to the underlying page to prevent it being lost.
>
> Currently migrate_vma_*() can only be used for private anonymous
> mappings. That means loss of the dirty bit usually doesn't result in
> data loss because these pages are typically not file-backed. However
> pages may be backed by swap storage which can result in data loss if an
> attempt is made to migrate a dirty page that doesn't yet have the
> PageDirty flag set.
>
> In this case migration will fail due to unexpected references but the
> dirty pte bit will be lost. If the page is subsequently reclaimed data
> won't be written back to swap storage as it is considered uptodate,
> resulting in data loss if the page is subsequently accessed.
>
> Prevent this by copying the dirty bit to the page when removing the pte
> to match what try_to_migrate_one() does.
>
> Signed-off-by: Alistair Popple <apopple@xxxxxxxxxx>
> Acked-by: Peter Xu <peterx@xxxxxxxxxx>
> Reported-by: Huang Ying <ying.huang@xxxxxxxxx>
> Fixes: 8c3328f1f36a ("mm/migrate: migrate_vma() unmap page from vma while collecting pages")
> Cc: stable@xxxxxxxxxxxxxxx
>
> ---
>
> Changes for v2:
>
> - Fixed up Reported-by tag.
> - Added Peter's Acked-by.
> - Atomically read and clear the pte to prevent the dirty bit getting
> set after reading it.
> - Added fixes tag
> ---
> mm/migrate_device.c | 21 ++++++++-------------
> 1 file changed, 8 insertions(+), 13 deletions(-)
>
> diff --git a/mm/migrate_device.c b/mm/migrate_device.c
> index 27fb37d..e2d09e5 100644
> --- a/mm/migrate_device.c
> +++ b/mm/migrate_device.c
> @@ -7,6 +7,7 @@
> #include <linux/export.h>
> #include <linux/memremap.h>
> #include <linux/migrate.h>
> +#include <linux/mm.h>
> #include <linux/mm_inline.h>
> #include <linux/mmu_notifier.h>
> #include <linux/oom.h>
> @@ -61,7 +62,7 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
> struct migrate_vma *migrate = walk->private;
> struct vm_area_struct *vma = walk->vma;
> struct mm_struct *mm = vma->vm_mm;
> - unsigned long addr = start, unmapped = 0;
> + unsigned long addr = start;
> spinlock_t *ptl;
> pte_t *ptep;
>
> @@ -193,11 +194,10 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
> bool anon_exclusive;
> pte_t swp_pte;
>
> + flush_cache_page(vma, addr, pte_pfn(*ptep));
> + pte = ptep_clear_flush(vma, addr, ptep);
Although I think it's possible to batch the TLB flushing just before
unlocking PTL. The current code looks correct.
Reviewed-by: "Huang, Ying" <ying.huang@xxxxxxxxx>
Best Regards,
Huang, Ying
> anon_exclusive = PageAnon(page) && PageAnonExclusive(page);
> if (anon_exclusive) {
> - flush_cache_page(vma, addr, pte_pfn(*ptep));
> - ptep_clear_flush(vma, addr, ptep);
> -
> if (page_try_share_anon_rmap(page)) {
> set_pte_at(mm, addr, ptep, pte);
> unlock_page(page);
> @@ -205,12 +205,14 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
> mpfn = 0;
> goto next;
> }
> - } else {
> - ptep_get_and_clear(mm, addr, ptep);
> }
>
> migrate->cpages++;
>
> + /* Set the dirty flag on the folio now the pte is gone. */
> + if (pte_dirty(pte))
> + folio_mark_dirty(page_folio(page));
> +
> /* Setup special migration page table entry */
> if (mpfn & MIGRATE_PFN_WRITE)
> entry = make_writable_migration_entry(
> @@ -242,9 +244,6 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
> */
> page_remove_rmap(page, vma, false);
> put_page(page);
> -
> - if (pte_present(pte))
> - unmapped++;
> } else {
> put_page(page);
> mpfn = 0;
> @@ -257,10 +256,6 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
> arch_leave_lazy_mmu_mode();
> pte_unmap_unlock(ptep - 1, ptl);
>
> - /* Only flush the TLB if we actually modified any entries */
> - if (unmapped)
> - flush_tlb_range(walk->vma, start, end);
> -
> return 0;
> }
>
>
> base-commit: ffcf9c5700e49c0aee42dcba9a12ba21338e8136
> --
> git-series 0.9.1
>
