Received-SPF: pass (google.com: domain of linux-kernel+bounces-160663-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
From: Zi Yan <ziy@nvidia.com>
To: David Hildenbrand <david@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>, linux-mm@kvack.org,
 "Matthew Wilcox (Oracle)" <willy@infradead.org>,
 Yang Shi <shy828301@gmail.com>, Ryan Roberts <ryan.roberts@arm.com>,
 Barry Song <21cnbao@gmail.com>, Lance Yang <ioworker0@gmail.com>,
 linux-kernel@vger.kernel.org
Subject: Re: [PATCH v5] mm/rmap: do not add fully unmapped large folio to
 deferred split list
Date: Fri, 26 Apr 2024 16:22:38 -0400
Message-ID: <AFA6A945-21AC-4665-A410-E906C7EB3EBF@nvidia.com>
In-Reply-To: <fb58cd88-d5f3-42ea-a710-0cff3764cd32@redhat.com>
References: <20240426190253.541419-1-zi.yan@sent.com>
 <e274ad3b-bbeb-49b0-81bf-24fd33106f05@redhat.com>
 <CB058583-C99A-4897-B3AE-11C56D6E3A59@nvidia.com>
 <fb58cd88-d5f3-42ea-a710-0cff3764cd32@redhat.com>
Content-Type: multipart/signed;
 boundary="=_MailMate_33548C9A-C040-4AF4-94DC-F7A4FE8A5C87_=";
 micalg=pgp-sha512; protocol="application/pgp-signature"
Precedence: bulk
MIME-Version: 1.0
X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1
X-MS-Exchange-AntiSpam-MessageData-0:
	=?us-ascii?Q?OqFylFbxxaUu1PIxg0xHr7D5tSBYpgPK+/YhJ1htU/kRlV1JF79Cay+sUsCj?=
 =?us-ascii?Q?80rW79ql8F92G1tFlmPLmME1yj3UJn86Rwg2fznKzQLegibEO1LhsFESvje8?=
 =?us-ascii?Q?pir2HPYltwH2iFipcudXYURjEwh2ySJEAl9hSaIL3jnUxJnnkhOTmsJh6vN4?=
 =?us-ascii?Q?uH8By2BRJgm/ov62b10d8AfYl1+qtpor4YqyrZZW7gnZAkB9J5sxcHbcF9BE?=
 =?us-ascii?Q?i7mziCqqQq/V8Ra4uRq5vE3a2/c+OP8x0xq/FSB6e5IkOAor3rzO22qsx9uX?=
 =?us-ascii?Q?ZuFYKlYQxX6oKFjpj6GWaQw+LUcnnamAWpqT4397nBnL+L1qeVzRPSME9QF+?=
 =?us-ascii?Q?gW82J+0DgtlZvrgPCVfXO8KC6FRdYd/gih2LWhR8G20jBHxaXh7DV/WC6ble?=
 =?us-ascii?Q?DnSjHGgZkzgImxFvZCQlvX7mBiMnjpoIBW+/FEbgj7sYSatOJ6+1031Yt0QO?=
 =?us-ascii?Q?tR9kJn+2qEijq8Abvziqhu3EH/c1lIIznoblKJlTF+a2j27RuDTfCw1WN//i?=
 =?us-ascii?Q?TiAWMhigWVC+ODdBmP6wrkn3+Lyi2bgzuOK5+lk7Lvzq0YSLac9zmvPaegAP?=
 =?us-ascii?Q?ytIlta878MGq/XWyIDk2dhuy2CtUtYIOpa5BR2DXSDgWbACVNrhlWeddzyg5?=
 =?us-ascii?Q?oz62I0UDyf+/E1KGlbH7ijpu/007rSuF400y9yFXox40HHmoFyUTowsM3Z5R?=
 =?us-ascii?Q?YFkmHpiCVjTx6eVuCSkKI2QJxLA+veT4X+OcyB1VY9zGkBzp1Qqr/5ypER1M?=
 =?us-ascii?Q?qxs4Q+x/oIqegai24S3OhWAmZFDbd0kRyx8R1C7TR/XmZnM3oi1Y04OImN0o?=
 =?us-ascii?Q?qt3gTglYUW+IhTKTl6zJkeuTTzTJ5na6E21MMbvH+u7Aynuruu2tKiLbhyOx?=
 =?us-ascii?Q?9TLP+QIzXwIthmIL+G5IudLMnLXLrIv+zi7jm1XTyvYCOgTSIllMVFl0Vt9L?=
 =?us-ascii?Q?2Qd/Ly3baafs0sDlr+oywM5OKblJ2vKTdyWgEYal8IYfu32mumsbXNu9FoTo?=
 =?us-ascii?Q?IwijexJTsqvvjNlcEw1edqxR4hGwUZIAUnvbKJ7hrd+UsikBDfhxDTOjD4D8?=
 =?us-ascii?Q?WFC8h73shR0WZ5TKKwQvCrYoyu0kxuwO2oosTjATFaW3XMft5/4ovflNX3HX?=
 =?us-ascii?Q?yViN33jE1mI8jKb9F5FAJexSgNGfG8TR6rOH0TKKzdaK6KZsV21RsJSmPTBY?=
 =?us-ascii?Q?Q9KRAgJID53QD+UBbW0XTGKI/dduraMp+iS4/bB4yEVDkmHHHCbpmBaVLEC8?=
 =?us-ascii?Q?aLwiHhy0XBJv7bpWAxusa/hQ8dMQMSPQHbiZywzH8DZVifd116Pv2h1Dasbn?=
 =?us-ascii?Q?TzalfQRL+AA44MMMPXtOBb56iRg4WvFzuG0op39WJS0NEqu9M5dBfHu8qnzb?=
 =?us-ascii?Q?TFRxHfvggQVRd1qNfel71sexPLK0+IQTMdQ0p5apmj/+k62UCcZLGiQnm6iT?=
 =?us-ascii?Q?aLCZuCYkrEgbb+ZLEG03V9diURxjMA3c/oGs/C/nMwEP29HX1MSmBLv96Me5?=
 =?us-ascii?Q?bNoiQHZKY+U5/utmPry4TeQ5/TltiEs+DwxPuaxwBaDi+YZXtDsg3UJ/mXpH?=
 =?us-ascii?Q?HOHGHNYn6mHnRmBSjjM=3D?=
X-OriginatorOrg: Nvidia.com
X-MS-Exchange-CrossTenant-Network-Message-Id: 897da066-3f31-494a-a6b8-08dc662e9f5f
X-MS-Exchange-CrossTenant-AuthSource: DS7PR12MB5744.namprd12.prod.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Internal
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 26 Apr 2024 20:22:40.9439
 (UTC)
X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted
X-MS-Exchange-CrossTenant-Id: 43083d15-7273-40c1-b7db-39efd9ccc17a
X-MS-Exchange-CrossTenant-MailboxType: HOSTED
X-MS-Exchange-CrossTenant-UserPrincipalName: qJyA3I4D++lvBN4Z259XQ7B3w9sDlz5cCtp6THsuLQUuZXg0GeqLFxbgt/7tLopT
X-MS-Exchange-Transport-CrossTenantHeadersStamped: PH0PR12MB7488

--=_MailMate_33548C9A-C040-4AF4-94DC-F7A4FE8A5C87_=
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On 26 Apr 2024, at 16:15, David Hildenbrand wrote:

> On 26.04.24 21:20, Zi Yan wrote:
>> On 26 Apr 2024, at 15:08, David Hildenbrand wrote:
>>
>>> On 26.04.24 21:02, Zi Yan wrote:
>>>> From: Zi Yan <ziy@nvidia.com>
>>>>
>>>> In __folio_remove_rmap(), a large folio is added to deferred split l=
ist
>>>> if any page in a folio loses its final mapping. But it is possible t=
hat
>>>> the folio is fully unmapped and adding it to deferred split list is
>>>> unnecessary.
>>>>
>>>> For PMD-mapped THPs, that was not really an issue, because removing =
the
>>>> last PMD mapping in the absence of PTE mappings would not have added=
 the
>>>> folio to the deferred split queue.
>>>>
>>>> However, for PTE-mapped THPs, which are now more prominent due to mT=
HP,
>>>> they are always added to the deferred split queue. One side effect
>>>> is that the THP_DEFERRED_SPLIT_PAGE stat for a PTE-mapped folio can =
be
>>>> unintentionally increased, making it look like there are many partia=
lly
>>>> mapped folios -- although the whole folio is fully unmapped stepwise=
=2E
>>>>
>>>> Core-mm now tries batch-unmapping consecutive PTEs of PTE-mapped THP=
s
>>>> where possible starting from commit b06dc281aa99 ("mm/rmap: introduc=
e
>>>> folio_remove_rmap_[pte|ptes|pmd]()"). When it happens, a whole PTE-m=
apped
>>>> folio is unmapped in one go and can avoid being added to deferred sp=
lit
>>>> list, reducing the THP_DEFERRED_SPLIT_PAGE noise. But there will sti=
ll be
>>>> noise when we cannot batch-unmap a complete PTE-mapped folio in one =
go
>>>> -- or where this type of batching is not implemented yet, e.g., migr=
ation.
>>>>
>>>> To avoid the unnecessary addition, folio->_nr_pages_mapped is checke=
d
>>>> to tell if the whole folio is unmapped. If the folio is already on
>>>> deferred split list, it will be skipped, too.
>>>>
>>>> Note: commit 98046944a159 ("mm: huge_memory: add the missing
>>>> folio_test_pmd_mappable() for THP split statistics") tried to exclud=
e
>>>> mTHP deferred split stats from THP_DEFERRED_SPLIT_PAGE, but it does =
not
>>>> fix the above issue. A fully unmapped PTE-mapped order-9 THP was sti=
ll
>>>> added to deferred split list and counted as THP_DEFERRED_SPLIT_PAGE,=

>>>> since nr is 512 (non zero), level is RMAP_LEVEL_PTE, and inside
>>>> deferred_split_folio() the order-9 folio is folio_test_pmd_mappable(=
).
>>>>
>>>> Suggested-by: David Hildenbrand <david@redhat.com>
>>>> Signed-off-by: Zi Yan <ziy@nvidia.com>
>>>> ---
>>>>    mm/rmap.c | 12 +++++++++---
>>>>    1 file changed, 9 insertions(+), 3 deletions(-)
>>>>
>>>> diff --git a/mm/rmap.c b/mm/rmap.c
>>>> index 2608c40dffad..a9bd64ebdd9a 100644
>>>> --- a/mm/rmap.c
>>>> +++ b/mm/rmap.c
>>>> @@ -1495,6 +1495,7 @@ static __always_inline void __folio_remove_rma=
p(struct folio *folio,
>>>>    {
>>>>    	atomic_t *mapped =3D &folio->_nr_pages_mapped;
>>>>    	int last, nr =3D 0, nr_pmdmapped =3D 0;
>>>> +	bool partially_mapped =3D false;
>>>>    	enum node_stat_item idx;
>>>>     	__folio_rmap_sanity_checks(folio, page, nr_pages, level);
>>>> @@ -1515,6 +1516,8 @@ static __always_inline void __folio_remove_rma=
p(struct folio *folio,
>>>>    					nr++;
>>>>    			}
>>>>    		} while (page++, --nr_pages > 0);
>>>> +
>>>> +		partially_mapped =3D !!nr && !!atomic_read(mapped);
>>>
>>> Nit: The && should remove the need for both !!.
>>
>> My impression was that !! is needed to convert from int to bool and I =
do
>> find "!!int && !!int" use in the kernel.
>
> I might be wrong about this, but if you wouldn't write
>
> 	if (!!nr && !!atomic_read(mapped))
>
> then
>
> bool partially_mapped =3D nr && atomic_read(mapped);
>
> is sufficient.
>
> && would make sure that the result is either 0 or 1, which
> you can store safely in a bool, no matter which underlying type
> is used to store that value.
>
> But I *think* nowdays, the compiler will always handle that
> correctly, even without the "&&" (ever since C99 added _Bool).
>
> Likely, also
>
> 	bool partially_mapped =3D nr & atomic_read(mapped);
>
> Would nowadays work, but looks stupid.
>
>
> Related: https://lkml.org/lkml/2013/8/31/138
>
> ---
> #include <stdio.h>
> #include <stdbool.h>
> #include <stdint.h>
> #include <inttypes.h>
>
> volatile uint64_t a =3D 0x8000000000000000ull;
>
> void main (void) {
>         printf("uint64_t a =3D a: 0x%" PRIx64 "\n", a);
>
>         int i1 =3D a;
>         printf("int i1 =3D a: %d\n", i1);
>
>         int i2 =3D !!a;
>         printf("int i2 =3D !!a: %d\n", i2);
>
>         bool b1 =3D a;
>         printf("bool b1 =3D a: %d\n", b1);
>
>         bool b2 =3D !!a;
>         printf("bool b2 =3D !!a: %d\n", b2);
> }
> ---
> $ ./test
> uint64_t a =3D a: 0x8000000000000000
> int i1 =3D a: 0
> int i2 =3D !!a: 1
> bool b1 =3D a: 1
> bool b2 =3D !!a: 1
> ---
>
> Note that if bool would be defined as "int", you would need the !!, oth=
erwise you
> would lose information.

Thank you for all above. Really appreciate it! And you are right about &&=
 and !!.
My memory on !! must be from the old days and now is refreshed. :)

>
> But even for b1, the gcc generates now:
>
>  40118c:       48 8b 05 7d 2e 00 00    mov    0x2e7d(%rip),%rax        =
# 404010 <a>
>  401193:       48 85 c0                test   %rax,%rax
>  401196:       0f 95 c0                setne  %al
>
>
> My stdbool.h contains
>
> #define bool	_Bool
>
> And I think C99 added _Bool that makes that work.
>
> But I didn't read the standard, and it's time for the weekend :)

Have a good weekend!

--
Best Regards,
Yan, Zi

--=_MailMate_33548C9A-C040-4AF4-94DC-F7A4FE8A5C87_=
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename=signature.asc
Content-Type: application/pgp-signature; name=signature.asc

-----BEGIN PGP SIGNATURE-----
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=tQFj
-----END PGP SIGNATURE-----

--=_MailMate_33548C9A-C040-4AF4-94DC-F7A4FE8A5C87_=--