Received-SPF: pass (google.com: domain of linux-kernel+bounces-140432-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Precedence: bulk
MIME-Version: 1.0
References: <20240408042437.10951-1-ioworker0@gmail.com> <20240408042437.10951-2-ioworker0@gmail.com>
 <38c4add8-53a2-49ca-9f1b-f62c2ee3e764@arm.com> <013334d5-62d2-4256-8045-168893a0a0cf@redhat.com>
 <a59496b6-ffdf-40ce-b030-283cc911203e@arm.com>
In-Reply-To: <a59496b6-ffdf-40ce-b030-283cc911203e@arm.com>
From: Lance Yang <ioworker0@gmail.com>
Date: Thu, 11 Apr 2024 20:23:45 +0800
Message-ID: <CAK1f24=hf2aBRMF5rSwyBxYa3pf7Cq9Y-wpuJ_HDyRxRrbM8sA@mail.gmail.com>
Subject: Re: [PATCH v5 1/2] mm/madvise: optimize lazyfreeing with mTHP in madvise_free
To: Ryan Roberts <ryan.roberts@arm.com>
Cc: David Hildenbrand <david@redhat.com>, akpm@linux-foundation.org, 21cnbao@gmail.com, 
	mhocko@suse.com, fengwei.yin@intel.com, zokeefe@google.com, 
	shy828301@gmail.com, xiehuan09@gmail.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, Apr 11, 2024 at 7:27=E2=80=AFPM Ryan Roberts <ryan.roberts@arm.com>=
 wrote:
>
> On 11/04/2024 12:20, David Hildenbrand wrote:
> > On 11.04.24 13:11, Ryan Roberts wrote:
> >> On 08/04/2024 05:24, Lance Yang wrote:
> >>> This patch optimizes lazyfreeing with PTE-mapped mTHP[1]
> >>> (Inspired by David Hildenbrand[2]). We aim to avoid unnecessary folio
> >>> splitting if the large folio is fully mapped within the target range.
> >>>
> >>> If a large folio is locked or shared, or if we fail to split it, we j=
ust
> >>> leave it in place and advance to the next PTE in the range. But note =
that
> >>> the behavior is changed; previously, any failure of this sort would c=
ause
> >>> the entire operation to give up. As large folios become more common,
> >>> sticking to the old way could result in wasted opportunities.
> >>>
> >>> On an Intel I5 CPU, lazyfreeing a 1GiB VMA backed by PTE-mapped folio=
s of
> >>> the same size results in the following runtimes for madvise(MADV_FREE=
) in
> >>> seconds (shorter is better):
> >>>
> >>> Folio Size |   Old    |   New    | Change
> >>> ------------------------------------------
> >>>        4KiB | 0.590251 | 0.590259 |    0%
> >>>       16KiB | 2.990447 | 0.185655 |  -94%
> >>>       32KiB | 2.547831 | 0.104870 |  -95%
> >>>       64KiB | 2.457796 | 0.052812 |  -97%
> >>>      128KiB | 2.281034 | 0.032777 |  -99%
> >>>      256KiB | 2.230387 | 0.017496 |  -99%
> >>>      512KiB | 2.189106 | 0.010781 |  -99%
> >>>     1024KiB | 2.183949 | 0.007753 |  -99%
> >>>     2048KiB | 0.002799 | 0.002804 |    0%
> >>>
> >>> [1] https://lkml.kernel.org/r/20231207161211.2374093-5-ryan.roberts@a=
rm.com
> >>> [2] https://lore.kernel.org/linux-mm/20240214204435.167852-1-david@re=
dhat.com
> >>>
> >>> Signed-off-by: Lance Yang <ioworker0@gmail.com>
> >>> ---
> >>>   include/linux/pgtable.h |  34 +++++++++
> >>>   mm/internal.h           |  12 +++-
> >>>   mm/madvise.c            | 149 ++++++++++++++++++++++---------------=
---
> >>>   mm/memory.c             |   4 +-
> >>>   4 files changed, 129 insertions(+), 70 deletions(-)
> >>>
> >>> diff --git a/include/linux/pgtable.h b/include/linux/pgtable.h
> >>> index 0f4b2faa1d71..4dd442787420 100644
> >>> --- a/include/linux/pgtable.h
> >>> +++ b/include/linux/pgtable.h
> >>> @@ -489,6 +489,40 @@ static inline pte_t ptep_get_and_clear(struct mm=
_struct
> >>> *mm,
> >>>   }
> >>>   #endif
> >>>   +#ifndef mkold_clean_ptes
> >>> +/**
> >>> + * mkold_clean_ptes - Mark PTEs that map consecutive pages of the sa=
me folio
> >>> + *        as old and clean.
> >>> + * @mm: Address space the pages are mapped into.
> >>> + * @addr: Address the first page is mapped at.
> >>> + * @ptep: Page table pointer for the first entry.
> >>> + * @nr: Number of entries to mark old and clean.
> >>> + *
> >>> + * May be overridden by the architecture; otherwise, implemented by
> >>> + * get_and_clear/modify/set for each pte in the range.
> >>> + *
> >>> + * Note that PTE bits in the PTE range besides the PFN can differ. F=
or example,
> >>> + * some PTEs might be write-protected.
> >>> + *
> >>> + * Context: The caller holds the page table lock.  The PTEs map cons=
ecutive
> >>> + * pages that belong to the same folio.  The PTEs are all in the sam=
e PMD.
> >>> + */
> >>> +static inline void mkold_clean_ptes(struct mm_struct *mm, unsigned l=
ong addr,
> >>> +                    pte_t *ptep, unsigned int nr)
> >>

Thanks for the suggestions, Ryan, David!

> >> Just thinking out loud, I wonder if it would be cleaner to convert mko=
ld_ptes()
> >> (which I added as part of swap-out) to something like:

Yeah, this is definitely cleaner than before.

> >>
> >> clear_young_dirty_ptes(struct mm_struct *mm, unsigned long addr,
> >>                pte_t *ptep, unsigned int nr,
> >>                bool clear_young, bool clear_dirty);
> >>
> >> Then we can use the same function for both use cases and also have the=
 ability
> >> to only clear dirty in future if we ever need it. The other advantage =
is that we
> >> only need to plumb a single function down the arm64 arch code. As it c=
urrently
> >> stands, those 2 functions would be duplicating most of their code.

Agreed. It's indeed a good idea to use a single function for both use cases=
.

> >
> > Yes. Maybe better use proper __bitwise flags, the compiler should be sm=
art
> > enough to optimize either way.

Nice. I'll use the __bitwise flags as the input.

>
> Agreed. I was also thinking perhaps it makes sense to start using output =
bitwise
> flags for folio_pte_batch() since this patch set takes us up to 3 optiona=
l bool
> pointers for different things. Might be cleaner to have input flags to te=
ll it
> what we care about and output flags to highlight those things. I guess th=
e
> compiler should be able to optimize in the same way.
>

Should I start using output bitwise flags for folio_pte_batch() in
this patch set?

Thanks,
Lance