Received-SPF: pass (google.com: domain of linux-kernel+bounces-151170-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
From: Shivansh Vij <shivanshvij@outlook.com>
To: Ryan Roberts <ryan.roberts@arm.com>, Catalin Marinas
	<catalin.marinas@arm.com>, Will Deacon <will@kernel.org>, Andrew Morton
	<akpm@linux-foundation.org>, Shuah Khan <shuah@kernel.org>, Joey Gouly
	<joey.gouly@arm.com>, Ard Biesheuvel <ardb@kernel.org>, Mark Rutland
	<mark.rutland@arm.com>, Anshuman Khandual <anshuman.khandual@arm.com>, David
 Hildenbrand <david@redhat.com>
CC: "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"linux-arm-kernel@lists.infradead.org"
	<linux-arm-kernel@lists.infradead.org>, "linux-kselftest@vger.kernel.org"
	<linux-kselftest@vger.kernel.org>
Subject: Re: [PATCH v1 0/5] arm64/mm: uffd write-protect and soft-dirty
 tracking
Thread-Topic: [PATCH v1 0/5] arm64/mm: uffd write-protect and soft-dirty
 tracking
Thread-Index: AQHaki1WdjYIu4CfuECqPGqYndfIFbFvNzIAgAAKqq0=
Date: Fri, 19 Apr 2024 08:33:22 +0000
Message-ID:
 <MW4PR12MB6875618342F088BE6F4ECBB2B90D2@MW4PR12MB6875.namprd12.prod.outlook.com>
References: <20240419074344.2643212-1-ryan.roberts@arm.com>
 <24999e38-e4f7-4616-8eae-dfdeba327558@arm.com>
In-Reply-To: <24999e38-e4f7-4616-8eae-dfdeba327558@arm.com>
Accept-Language: en-CA, en-US
Content-Language: en-CA
msip_labels:
x-tmn: [h0XkEBLu0XIa893tZci+kZDvrr4T3Siz]
x-ms-publictraffictype: Email
x-ms-traffictypediagnostic: MW4PR12MB6875:EE_|BL3PR12MB6619:EE_
x-ms-office365-filtering-correlation-id: c347403f-6cce-43b0-8e72-08dc604b5ffb
x-microsoft-antispam: BCL:0;
x-microsoft-antispam-message-info:
 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
x-ms-exchange-antispam-messagedata-chunkcount: 1
x-ms-exchange-antispam-messagedata-0:
 =?iso-8859-1?Q?rQCGhrB8V8Hg7mW7ldhuRxVunUALaNJY5zEaIaibuFb1ScLNVkDYB33YzZ?=
 =?iso-8859-1?Q?+im28LD3bYK+chApjn3dvUTsx2qplY+GmRyqdx9ZWNvbeKPcSpYUunOE71?=
 =?iso-8859-1?Q?BFahHrtwGyrpFDTxh+eA/2Ymw9AWBpxLUnF64IXBMthmR8NrisoDKZhksu?=
 =?iso-8859-1?Q?jsLJn57meRh6+IaCDDif8BihYOJ+OjWAMmDirq0IlGNeyFF7CFYDTnw/ci?=
 =?iso-8859-1?Q?1gH5mctNDmAdv8j+dKMqwKEzhZWQY+lfB3TpBNXGxjEELcg55KniQDTMDl?=
 =?iso-8859-1?Q?W6ctoQv5VDVbFbWmNGsgUIcRD7Retbtq1w31pyMThEA/HJUxfnpWyd702v?=
 =?iso-8859-1?Q?A7OnVtvO/e4dpa+4UUoz67pysnWM0en+bk61YbqqOCbRLOmu7PVZPzPRI/?=
 =?iso-8859-1?Q?qB4K2FXrM+5Ec7OxSMbyve7Muex5vWbV75JaSvu009S9jEvRusmzVika8H?=
 =?iso-8859-1?Q?unerHG2LG/2T8b4jxMgj7IATwpE9epPsMJYcymgN/9uhartDrjXMRK4Rli?=
 =?iso-8859-1?Q?JC1Wbx5PiCYM9AEX/F3tCyg1fLsLAvXlzJZwQvui+WcwBKqBwJTT7fst5J?=
 =?iso-8859-1?Q?b/i+4gvaGS7CSWTZmaNLEjf02eDn8vS3adJCeOgKt+YflRtGrNHM5pEJvv?=
 =?iso-8859-1?Q?adVjdFa+CXZyHSfLmJnULfRV287BH/QGXQljOctZ261I2uZXRZeR2UqMjd?=
 =?iso-8859-1?Q?L2Cvjq5DEj9clwtP73lVWICNAbWtNNF8yGf5sLs+7eWsi0WAEfID+SkmLT?=
 =?iso-8859-1?Q?aCpgyRCBmYUQwHLm3c4MAjPWrjlQhtXougV2PzttJEdHswdV6KNF2fqkEe?=
 =?iso-8859-1?Q?AFGR6tUr+vQ1AVeMWblnqbYkMk/6xwFaiF9B8yawQ/Wc12DPgU5q3wIC4V?=
 =?iso-8859-1?Q?SkfP3AghiJ2FzCZNWTD/QPl/cP6h3LezKVqJcJ/ImE5IstE6im5KcyFF8K?=
 =?iso-8859-1?Q?i51wR+kGyl6sMRGGcdJ2SPLUtiCkPOASjBYkj7EjZnNpiGKmw7+n86qTFf?=
 =?iso-8859-1?Q?5YYzMb/IuS7Qub05tgO3B6rkjgebNkLPP0D2iXDHcHVenccykSzsNaOJ/F?=
 =?iso-8859-1?Q?s4n9ISdPR/8UujrBrh5mrRUH76IqRT8R06jyGtCd6AYyZUn/KINm7/07BE?=
 =?iso-8859-1?Q?HpRQIsNUuKJBwJgqUtxzDITP6mi3B7qo5YftLqiNkLmVXk9iqVnzoQ41dj?=
 =?iso-8859-1?Q?2v1U6QbMWpPQ4+GY2KW8B/uod2ZXlD1vNaKUWeD1d+fYMGJS1oB/U6aSjr?=
 =?iso-8859-1?Q?qS1l/RtImIAUlorPfVTKa7lysxAY5rpdnkj6R+o/M=3D?=
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk
X-Mailing-List: linux-kernel@vger.kernel.org
List-Id: <linux-kernel.vger.kernel.org>
List-Subscribe: <mailto:linux-kernel+subscribe@vger.kernel.org>
List-Unsubscribe: <mailto:linux-kernel+unsubscribe@vger.kernel.org>
MIME-Version: 1.0
X-OriginatorOrg: outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Internal
X-MS-Exchange-CrossTenant-AuthSource: MW4PR12MB6875.namprd12.prod.outlook.com
X-MS-Exchange-CrossTenant-RMS-PersistedConsumerOrg: 00000000-0000-0000-0000-000000000000
X-MS-Exchange-CrossTenant-Network-Message-Id: c347403f-6cce-43b0-8e72-08dc604b5ffb
X-MS-Exchange-CrossTenant-rms-persistedconsumerorg: 00000000-0000-0000-0000-000000000000
X-MS-Exchange-CrossTenant-originalarrivaltime: 19 Apr 2024 08:33:22.8728
 (UTC)
X-MS-Exchange-CrossTenant-fromentityheader: Hosted
X-MS-Exchange-CrossTenant-id: 84df9e7f-e9f6-40af-b435-aaaaaaaaaaaa
X-MS-Exchange-Transport-CrossTenantHeadersStamped: BL3PR12MB6619

(Sorry about the previous HTML email, accidentally used the wrong email cli=
ent) =0A=
=0A=
Hey All,=0A=
=0A=
>On 19/04/2024 08:43, Ryan Roberts wrote:=0A=
>> Hi All,=0A=
>>=0A=
>> This series adds uffd write-protect and soft-dirty tracking support for =
arm64. I=0A=
>> consider the soft-dirty support (patches 3 and 4) as RFC - see rationale=
 below.=0A=
>>=0A=
>> Previous attempts to add these features have failed because of a perceiv=
ed lack=0A=
>> of available PTE SW bits. However it actually turns out that there are 2=
=0A=
>> available but they are hidden. PTE_PROT_NONE was previously occupying a =
SW bit,=0A=
>> but it only applies when PTE_VALID is clear, so this is moved to overlay=
 PTE_UXN=0A=
>> in patch 1, freeing up the SW bit. Bit 63 is marked as "IGNORED" in the =
Arm ARM,=0A=
>> but it does not currently indicate "reserved for SW use" like it does fo=
r the=0A=
>> other SW bits. I've confirmed with the spec owner that this is an oversi=
ght; the=0A=
>> bit is intended to be reserved for SW use and the spec will clarify this=
 in a=0A=
>> future update.=0A=
>>=0A=
>> So we have our two bits; patch 2 enables uffd-wp, patch 3 enables soft-d=
irty and=0A=
>> patches 4 and 5 sort out the selftests so that the soft-dirty tests are =
compiled=0A=
>> for, and run on arm64.=0A=
>>=0A=
>> That said, these are the last 2 SW bits and we may want to keep 1 bit in=
 reserve=0A=
>> for future use. soft-dirty is only used for CRIU to my knowledge, and it=
 is=0A=
>> thought that their use case could be solved with the more generic uffd-w=
p. So=0A=
>> unless somebody makes a clear case for the inclusion of soft-dirty suppo=
rt, we=0A=
>> are probably better off dropping patches 3 and 4 and keeping bit 63 for =
future=0A=
>> use. Although note that the most recent attempt to add soft-dirty for ar=
m64 was=0A=
>> last month [1] so I'd like to give Shivansh Vij the opportunity to make =
the=0A=
>> case.=0A=
>=0A=
> Ugh, forgot to mention that this applies on top of v6.9-rc3, and all the =
uffd-wp=0A=
> and soft-dirty tests in the mm selftests suite run and pass. And no regre=
ssions=0A=
> are observed in any of the other selftests.=0A=
=0A=
Appreciate the opportunity to provide input here. =0A=
=0A=
I personally don't know of any applications other than CRIU that make heavy=
 use of soft-dirty, and my use case is specifically focused on adding live-=
migration support to CRIU on ARM. =0A=
=0A=
Cloud providers like AWS have pretty massive discounts for ARM-based spot i=
nstances (90% last time I checked), and having live-migration in CRIU would=
 allow more applications to take advantage of that.=0A=
=0A=
As Ryan mentioned, there are two ways to achieve this - add dirty tracking =
to ARM (Patch 3/4), or tear out the existing dirty tracking code in CRIU an=
d replace it with uffd-wp. =0A=
=0A=
I picked option one (dirty tracking in arm) because it seems to be the simp=
lest way to move forward, whereas it would be a relatively heavy effort to =
add uffd-wp support to CRIU. =0A=
=0A=
From a performance perspective I am also a little worried that uffd will be=
 slower than just tracking the dirty bits asynchronously with sw dirty, but=
 maybe that's not as much of a concern with the addition of uffd-wp async. =
=0A=
=0A=
With all this being said, I'll defer to the wisdom of the crowd about which=
 approach makes more sense - after all, with this patch we should get uffd-=
wp support on arm so at least there will be _a_ way forward for CRIU (albei=
t one requiring slightly more work). =0A=
=0A=
Thanks,=0A=
Shivansh=