Re: [PATCH v3 01/12] mm/khugepaged: record SCAN_PMD_MAPPED when scan_pmd() finds THP

["Zach O'Keefe" <zokeefe@google.com>]

Thanks for taking the time to review, Peter!

On Tue, Apr 26, 2022 at 5:26 PM Peter Xu <peterx@redhat.com> wrote:
>
> Hi, Zach,
>
> On Tue, Apr 26, 2022 at 07:44:01AM -0700, Zach O'Keefe wrote:
> > When scanning an anon pmd to see if it's eligible for collapse, return
> > SCAN_PMD_MAPPED if the pmd already maps a THP.  Note that
> > SCAN_PMD_MAPPED is different from SCAN_PAGE_COMPOUND used in the
> > file-collapse path, since the latter might identify pte-mapped compound
> > pages.  This is required by MADV_COLLAPSE which necessarily needs to
> > know what hugepage-aligned/sized regions are already pmd-mapped.
> >
> > Signed-off-by: Zach O'Keefe <zokeefe@google.com>
> > Reported-by: kernel test robot <lkp@intel.com>
>
> IIUC we don't need to attach this reported-by if this is not a bugfix.  I
> think you can simply fix all issues reported by the test bot and only
> attach the line if the patch is fixing the problem that the bot was
> reporting explicitly.
>

Ya, I wasn't entirely sure what to do here, but including seems to not
be without precedent, e.g. commit 92bbef67d459 ("mm: make
alloc_contig_range work at pageblock granularity"), and likewise just
wanted to give credit where I thought it was due. Though, I suppose
folks who catch bugs in the review process aren't ack'd similarly, so
perhaps it does make sense to remove this.

> > ---
> >  include/trace/events/huge_memory.h |  3 ++-
> >  mm/internal.h                      |  1 +
> >  mm/khugepaged.c                    | 30 ++++++++++++++++++++++++++----
> >  mm/rmap.c                          | 15 +++++++++++++--
> >  4 files changed, 42 insertions(+), 7 deletions(-)
> >
> > diff --git a/include/trace/events/huge_memory.h b/include/trace/events/huge_memory.h
> > index d651f3437367..9faa678e0a5b 100644
> > --- a/include/trace/events/huge_memory.h
> > +++ b/include/trace/events/huge_memory.h
> > @@ -33,7 +33,8 @@
> >       EM( SCAN_ALLOC_HUGE_PAGE_FAIL,  "alloc_huge_page_failed")       \
> >       EM( SCAN_CGROUP_CHARGE_FAIL,    "ccgroup_charge_failed")        \
> >       EM( SCAN_TRUNCATED,             "truncated")                    \
> > -     EMe(SCAN_PAGE_HAS_PRIVATE,      "page_has_private")             \
> > +     EM( SCAN_PAGE_HAS_PRIVATE,      "page_has_private")             \
> > +     EMe(SCAN_PMD_MAPPED,            "page_pmd_mapped")              \
>
> Nit: IMHO it can be put even in the middle so we don't need to touch the
> EMe() every time. :)
>
> Apart from that, it does sound proper to me to put SCAN_PMD_MAPPED to be
> right after SCAN_PMD_NULL anyway.
>

Makes sense to me. Done.

> >
> >  #undef EM
> >  #undef EMe
> > diff --git a/mm/internal.h b/mm/internal.h
> > index 0667abd57634..51ae9f71a2a3 100644
> > --- a/mm/internal.h
> > +++ b/mm/internal.h
> > @@ -172,6 +172,7 @@ extern void reclaim_throttle(pg_data_t *pgdat, enum vmscan_throttle_state reason
> >  /*
> >   * in mm/rmap.c:
> >   */
> > +pmd_t *mm_find_pmd_raw(struct mm_struct *mm, unsigned long address);
> >  extern pmd_t *mm_find_pmd(struct mm_struct *mm, unsigned long address);
> >
> >  /*
> > diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> > index ba8dbd1825da..2933b13fc975 100644
> > --- a/mm/khugepaged.c
> > +++ b/mm/khugepaged.c
> > @@ -51,6 +51,7 @@ enum scan_result {
> >       SCAN_CGROUP_CHARGE_FAIL,
> >       SCAN_TRUNCATED,
> >       SCAN_PAGE_HAS_PRIVATE,
> > +     SCAN_PMD_MAPPED,
> >  };
> >
> >  #define CREATE_TRACE_POINTS
> > @@ -987,6 +988,29 @@ static int hugepage_vma_revalidate(struct mm_struct *mm, unsigned long address,
> >       return 0;
> >  }
> >
> > +static int find_pmd_or_thp_or_none(struct mm_struct *mm,
> > +                                unsigned long address,
> > +                                pmd_t **pmd)
> > +{
> > +     pmd_t pmde;
> > +
> > +     *pmd = mm_find_pmd_raw(mm, address);
> > +     if (!*pmd)
> > +             return SCAN_PMD_NULL;
> > +
> > +     pmde = pmd_read_atomic(*pmd);
> > +
> > +#ifdef CONFIG_TRANSPARENT_HUGEPAGE
> > +     /* See comments in pmd_none_or_trans_huge_or_clear_bad() */
> > +     barrier();
> > +#endif
> > +     if (!pmd_present(pmde) || pmd_none(pmde))
>
> Could we drop the pmd_none() check?  I assume !pmd_present() should have
> covered that case already?
>

I opted for safety here since I didn't know if pmd_present() always
implied !pmd_none() on all archs, but given mm_find_pmd() elides the
check, perhaps it's safe to do so here. Thanks for the suggestion.


> > +             return SCAN_PMD_NULL;
> > +     if (pmd_trans_huge(pmde))
> > +             return SCAN_PMD_MAPPED;
> > +     return SCAN_SUCCEED;
> > +}
> > +
> >  /*
> >   * Bring missing pages in from swap, to complete THP collapse.
> >   * Only done if khugepaged_scan_pmd believes it is worthwhile.
> > @@ -1238,11 +1262,9 @@ static int khugepaged_scan_pmd(struct mm_struct *mm,
> >
> >       VM_BUG_ON(address & ~HPAGE_PMD_MASK);
> >
> > -     pmd = mm_find_pmd(mm, address);
> > -     if (!pmd) {
> > -             result = SCAN_PMD_NULL;
> > +     result = find_pmd_or_thp_or_none(mm, address, &pmd);
> > +     if (result != SCAN_SUCCEED)
> >               goto out;
> > -     }
> >
> >       memset(khugepaged_node_load, 0, sizeof(khugepaged_node_load));
> >       pte = pte_offset_map_lock(mm, pmd, address, &ptl);
> > diff --git a/mm/rmap.c b/mm/rmap.c
> > index 61e63db5dc6f..49817f35e65c 100644
> > --- a/mm/rmap.c
> > +++ b/mm/rmap.c
> > @@ -759,13 +759,12 @@ unsigned long page_address_in_vma(struct page *page, struct vm_area_struct *vma)
> >       return vma_address(page, vma);
> >  }
> >
> > -pmd_t *mm_find_pmd(struct mm_struct *mm, unsigned long address)
> > +pmd_t *mm_find_pmd_raw(struct mm_struct *mm, unsigned long address)
> >  {
> >       pgd_t *pgd;
> >       p4d_t *p4d;
> >       pud_t *pud;
> >       pmd_t *pmd = NULL;
> > -     pmd_t pmde;
> >
> >       pgd = pgd_offset(mm, address);
> >       if (!pgd_present(*pgd))
> > @@ -780,6 +779,18 @@ pmd_t *mm_find_pmd(struct mm_struct *mm, unsigned long address)
> >               goto out;
> >
> >       pmd = pmd_offset(pud, address);
> > +out:
> > +     return pmd;
> > +}
> > +
> > +pmd_t *mm_find_pmd(struct mm_struct *mm, unsigned long address)
> > +{
> > +     pmd_t pmde;
> > +     pmd_t *pmd;
> > +
> > +     pmd = mm_find_pmd_raw(mm, address);
> > +     if (!pmd)
> > +             goto out;
> >       /*
> >        * Some THP functions use the sequence pmdp_huge_clear_flush(), set_pmd_at()
> >        * without holding anon_vma lock for write.  So when looking for a
> > --
> > 2.36.0.rc2.479.g8af0fa9b8e-goog
> >
>
> --
> Peter Xu
>