Received-SPF: pass (google.com: domain of linux-kernel+bounces-203598-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Precedence: bulk
MIME-Version: 1.0
References: <20240521040244.48760-1-ioworker0@gmail.com> <20240521040244.48760-3-ioworker0@gmail.com>
 <fd16b219-bc46-484a-8581-a21240440fa6@redhat.com> <CAK1f24kwf4gDwK=8X4z1bM9-H6_M9QKy6-ko9pTUZij-W=40wg@mail.gmail.com>
 <d319f00e-9dfb-43b1-ae81-a2e2afdf36c4@redhat.com> <8580a462-eadc-4fa5-b01a-c0b8c3ae644d@redhat.com>
 <CAK1f24=7=QqDFoh=joC+o1eZfyJ8onyc9ELBZBzn1CFr4uPSwQ@mail.gmail.com>
 <7f2ab112-5916-422c-b29f-343cc0d6d754@redhat.com> <CAK1f24nDH2UxCskNsFM84=5uGTUb_bAUivgmQJCcE+H1dJHDxw@mail.gmail.com>
 <c1975e82-3442-441a-a9ce-c612c2d3f70f@redhat.com>
In-Reply-To: <c1975e82-3442-441a-a9ce-c612c2d3f70f@redhat.com>
From: Lance Yang <ioworker0@gmail.com>
Date: Thu, 6 Jun 2024 11:57:01 +0800
Message-ID: <CAK1f24n-VscL_uBOpiPfE_HJ2NM+DwoWP2c6eg5a7=oqPRy2dQ@mail.gmail.com>
Subject: Re: [PATCH v6 2/3] mm/rmap: integrate PMD-mapped folio splitting into
 pagewalk loop
To: David Hildenbrand <david@redhat.com>
Cc: Yin Fengwei <fengwei.yin@intel.com>, akpm@linux-foundation.org, willy@infradead.org, 
	sj@kernel.org, baolin.wang@linux.alibaba.com, maskray@google.com, 
	ziy@nvidia.com, ryan.roberts@arm.com, 21cnbao@gmail.com, mhocko@suse.com, 
	zokeefe@google.com, shy828301@gmail.com, xiehuan09@gmail.com, 
	libang.li@antgroup.com, wangkefeng.wang@huawei.com, songmuchun@bytedance.com, 
	peterx@redhat.com, minchan@kernel.org, linux-mm@kvack.org, 
	linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Thu, Jun 6, 2024 at 12:16=E2=80=AFAM David Hildenbrand <david@redhat.com=
> wrote:
>
> On 05.06.24 17:43, Lance Yang wrote:
> > On Wed, Jun 5, 2024 at 11:03=E2=80=AFPM David Hildenbrand <david@redhat=
.com> wrote:
> >>
> >> On 05.06.24 16:57, Lance Yang wrote:
> >>> On Wed, Jun 5, 2024 at 10:39=E2=80=AFPM David Hildenbrand <david@redh=
at.com> wrote:
> >>>>
> >>>> On 05.06.24 16:28, David Hildenbrand wrote:
> >>>>> On 05.06.24 16:20, Lance Yang wrote:
> >>>>>> Hi David,
> >>>>>>
> >>>>>> On Wed, Jun 5, 2024 at 8:46=E2=80=AFPM David Hildenbrand <david@re=
dhat.com> wrote:
> >>>>>>>
> >>>>>>> On 21.05.24 06:02, Lance Yang wrote:
> >>>>>>>> In preparation for supporting try_to_unmap_one() to unmap PMD-ma=
pped
> >>>>>>>> folios, start the pagewalk first, then call split_huge_pmd_addre=
ss() to
> >>>>>>>> split the folio.
> >>>>>>>>
> >>>>>>>> Since TTU_SPLIT_HUGE_PMD will no longer perform immediately, we =
might
> >>>>>>>> encounter a PMD-mapped THP missing the mlock in the VM_LOCKED ra=
nge during
> >>>>>>>> the page walk. It=E2=80=99s probably necessary to mlock this THP=
 to prevent it from
> >>>>>>>> being picked up during page reclaim.
> >>>>>>>>
> >>>>>>>> Suggested-by: David Hildenbrand <david@redhat.com>
> >>>>>>>> Suggested-by: Baolin Wang <baolin.wang@linux.alibaba.com>
> >>>>>>>> Signed-off-by: Lance Yang <ioworker0@gmail.com>
> >>>>>>>> ---
> >>>>>>>
> >>>>>>> [...] again, sorry for the late review.
> >>>>>>
> >>>>>> No worries at all, thanks for taking time to review!
> >>>>>>
> >>>>>>>
> >>>>>>>> diff --git a/mm/rmap.c b/mm/rmap.c
> >>>>>>>> index ddffa30c79fb..08a93347f283 100644
> >>>>>>>> --- a/mm/rmap.c
> >>>>>>>> +++ b/mm/rmap.c
> >>>>>>>> @@ -1640,9 +1640,6 @@ static bool try_to_unmap_one(struct folio =
*folio, struct vm_area_struct *vma,
> >>>>>>>>           if (flags & TTU_SYNC)
> >>>>>>>>                   pvmw.flags =3D PVMW_SYNC;
> >>>>>>>>
> >>>>>>>> -     if (flags & TTU_SPLIT_HUGE_PMD)
> >>>>>>>> -             split_huge_pmd_address(vma, address, false, folio)=
;
> >>>>>>>> -
> >>>>>>>>           /*
> >>>>>>>>            * For THP, we have to assume the worse case ie pmd fo=
r invalidation.
> >>>>>>>>            * For hugetlb, it could be much worse if we need to d=
o pud
> >>>>>>>> @@ -1668,20 +1665,35 @@ static bool try_to_unmap_one(struct foli=
o *folio, struct vm_area_struct *vma,
> >>>>>>>>           mmu_notifier_invalidate_range_start(&range);
> >>>>>>>>
> >>>>>>>>           while (page_vma_mapped_walk(&pvmw)) {
> >>>>>>>> -             /* Unexpected PMD-mapped THP? */
> >>>>>>>> -             VM_BUG_ON_FOLIO(!pvmw.pte, folio);
> >>>>>>>> -
> >>>>>>>>                   /*
> >>>>>>>>                    * If the folio is in an mlock()d vma, we must=
 not swap it out.
> >>>>>>>>                    */
> >>>>>>>>                   if (!(flags & TTU_IGNORE_MLOCK) &&
> >>>>>>>>                       (vma->vm_flags & VM_LOCKED)) {
> >>>>>>>>                           /* Restore the mlock which got missed =
*/
> >>>>>>>> -                     if (!folio_test_large(folio))
> >>>>>>>> +                     if (!folio_test_large(folio) ||
> >>>>>>>> +                         (!pvmw.pte && (flags & TTU_SPLIT_HUGE_=
PMD)))
> >>>>>>>>                                   mlock_vma_folio(folio, vma);
> >>>>>>>
> >>>>>>> Can you elaborate why you think this would be required? If we wou=
ld have
> >>>>>>> performed the  split_huge_pmd_address() beforehand, we would stil=
l be
> >>>>>>> left with a large folio, no?
> >>>>>>
> >>>>>> Yep, there would still be a large folio, but it wouldn't be PMD-ma=
pped.
> >>>>>>
> >>>>>> After Weifeng's series[1], the kernel supports mlock for PTE-mappe=
d large
> >>>>>> folio, but there are a few scenarios where we don't mlock a large =
folio, such
> >>>>>> as when it crosses a VM_LOCKed VMA boundary.
> >>>>>>
> >>>>>>      -                     if (!folio_test_large(folio))
> >>>>>>      +                     if (!folio_test_large(folio) ||
> >>>>>>      +                         (!pvmw.pte && (flags & TTU_SPLIT_HU=
GE_PMD)))
> >>>>>>
> >>>>>> And this check is just future-proofing and likely unnecessary. If =
encountering a
> >>>>>> PMD-mapped THP missing the mlock for some reason, we can mlock thi=
s
> >>>>>> THP to prevent it from being picked up during page reclaim, since =
it is fully
> >>>>>> mapped and doesn't cross the VMA boundary, IIUC.
> >>>>>>
> >>>>>> What do you think?
> >>>>>> I would appreciate any suggestions regarding this check ;)
> >>>>>
> >>>>> Reading this patch only, I wonder if this change makes sense in the
> >>>>> context here.
> >>>>>
> >>>>> Before this patch, we would have PTE-mapped the PMD-mapped THP befo=
re
> >>>>> reaching this call and skipped it due to "!folio_test_large(folio)"=
.
> >>>>>
> >>>>> After this patch, we either
> >>>>>
> >>>>> a) PTE-remap the THP after this check, but retry and end-up here ag=
ain,
> >>>>> whereby we would skip it due to "!folio_test_large(folio)".
> >>>>>
> >>>>> b) Discard the PMD-mapped THP due to lazyfree directly. Can that
> >>>>> co-exist with mlock and what would be the problem here with mlock?
> >>>>>
> >>>>>
> >>>
> >>> Thanks a lot for clarifying!
> >>>
> >>>>> So if the check is required in this patch, we really have to unders=
tand
> >>>>> why. If not, we should better drop it from this patch.
> >>>>>
> >>>>> At least my opinion, still struggling to understand why it would be
> >>>>> required (I have 0 knowledge about mlock interaction with large fol=
ios :) ).
> >>>>>
> >>>>
> >>>> Looking at that series, in folio_references_one(), we do
> >>>>
> >>>>                           if (!folio_test_large(folio) || !pvmw.pte)=
 {
> >>>>                                   /* Restore the mlock which got mis=
sed */
> >>>>                                   mlock_vma_folio(folio, vma);
> >>>>                                   page_vma_mapped_walk_done(&pvmw);
> >>>>                                   pra->vm_flags |=3D VM_LOCKED;
> >>>>                                   return false; /* To break the loop=
 */
> >>>>                           }
> >>>>
> >>>> I wonder if we want that here as well now: in case of lazyfree we
> >>>> would not back off, right?
> >>>>
> >>>> But I'm not sure if lazyfree in mlocked areas are even possible.
> >>>>
> >>>> Adding the "!pvmw.pte" would be much clearer to me than the flag che=
ck.
> >>>
> >>> Hmm... How about we drop it from this patch for now, and add it back =
if needed
> >>> in the future?
> >>
> >> If we can rule out that MADV_FREE + mlock() keeps working as expected =
in
> >> the PMD-mapped case, we're good.
> >>
> >> Can we rule that out? (especially for MADV_FREE followed by mlock())
> >
> > Perhaps we don't worry about that.
> >
> > IIUC, without that check, MADV_FREE + mlock() still works as expected i=
n
> > the PMD-mapped case, since if encountering a large folio in a VM_LOCKED
> > VMA range, we will stop the page walk immediately.
>
>
> Can you point me at the code (especially considering patch #3?)

Yep, please see my other mail ;)

Thanks,
Lance

>
> --
> Cheers,
>
> David / dhildenb
>