From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id ADB3BCDE001 for ; Thu, 25 Jun 2026 13:06:31 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id AC4406B0088; Thu, 25 Jun 2026 09:06:30 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id A9C626B0093; Thu, 25 Jun 2026 09:06:30 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 98B686B0095; Thu, 25 Jun 2026 09:06:30 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (smtprelay0010.hostedemail.com [216.40.44.10]) by kanga.kvack.org (Postfix) with ESMTP id 675CF6B0088 for ; Thu, 25 Jun 2026 09:06:30 -0400 (EDT) Received: from smtpin14.hostedemail.com (lb01a-stub [10.200.18.249]) by unirelay10.hostedemail.com (Postfix) with ESMTP id E58FEC23E5 for ; Thu, 25 Jun 2026 13:06:29 +0000 (UTC) X-FDA: 84918458898.14.433AF54 Received: from BL2PR02CU003.outbound.protection.outlook.com (mail-eastusazon11011070.outbound.protection.outlook.com [52.101.52.70]) by imf05.hostedemail.com (Postfix) with ESMTP id DFB32100013 for ; Thu, 25 Jun 2026 13:06:26 +0000 (UTC) Authentication-Results: imf05.hostedemail.com; dkim=pass header.d=amd.com header.s=selector1 header.b=ILnSkech; spf=pass (imf05.hostedemail.com: domain of Christian.Koenig@amd.com designates 52.101.52.70 as permitted sender) smtp.mailfrom=Christian.Koenig@amd.com; dmarc=pass (policy=quarantine) header.from=amd.com; arc=pass ("microsoft.com:s=arcselector10001:i=1") ARC-Seal: i=2; a=rsa-sha256; d=hostedemail.com; s=arc-20220608; cv=pass; t=1782392787; b=lNMZ9mWhbYGimmbnP8VT+oyeJr+e656dD2o4bT5C83hOZXF85ydcsV2Ud8ZxTqxL9I/eeE 8CPY5/mTSueztvoaosfXz7ImpWJ3M5BZlHdvuWnoO6+3cQRHG+mym78BwQ2VNn0VNsOlAL t0Ph5arOwj/AI8LHF1eQtW/d/xuu5hg= ARC-Message-Signature: i=2; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1782392787; h=from:from:sender:reply-to:subject:subject:date:date: message-id:message-id:to:to:cc:cc:mime-version:mime-version: content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references:dkim-signature; bh=+7L0x7xdBg47yhgFzeWgQ1a1i8LcxW2uyIVEmpYszCo=; b=lQa4LcfGDVcH+N7FuGHxOTGGRUrfbXJncRkfjCJFcSCriQMg1h+ZXy74hSfESz81FY50id wCFYv1Z69herprVG0PaMYejAij5IheVcod24N8YimURN0klERqTSrXaPz4Ovc4rcUSj+PO Tc1Tl3zwi+p+O+I0lbnx8CDYC1LfVDk= ARC-Authentication-Results: i=2; imf05.hostedemail.com; dkim=pass header.d=amd.com header.s=selector1 header.b=ILnSkech; spf=pass (imf05.hostedemail.com: domain of Christian.Koenig@amd.com designates 52.101.52.70 as permitted sender) smtp.mailfrom=Christian.Koenig@amd.com; dmarc=pass (policy=quarantine) header.from=amd.com; arc=pass ("microsoft.com:s=arcselector10001:i=1") ARC-Seal: i=1; a=rsa-sha256; s=arcselector10001; d=microsoft.com; cv=none; b=os0aqVdgMN/UhPIlMt8nyFBW/GN3iSbZqk/dilpsv7Ms5iOF9Q/Qu6aaJ5FMQc1gQ0N9Jr0IJpNvTURdsaAH5RvlA29Llkc+ELEPkv2PvvWgjunDUWjhKAMbmdlG9TtyhihlxNC3SYU2XfILApl1+FbAJAZM8g7C+YyfTOL3cMYZIN2zJ5rc6ov4P8ICKDuWsThB8SsNlhiTazbsfcDjkdtU5LWoK463CTTSgqPPcrn50qA8S9DSIwO1DhEG4cGtVs8xJ8FafMdYAojZsi70FCa5JqqF7s70OUmSd2GxFeiOS7Ruy0so21BzS/dL/x5gVgkkFL7kFvkF1t/6JL24MQ== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=microsoft.com; s=arcselector10001; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-AntiSpam-MessageData-ChunkCount:X-MS-Exchange-AntiSpam-MessageData-0:X-MS-Exchange-AntiSpam-MessageData-1; bh=+7L0x7xdBg47yhgFzeWgQ1a1i8LcxW2uyIVEmpYszCo=; b=c07XWp9vtVgDAjYy+uYZ9A9KukTbykdUBZocq/3AFsLv7sa93i8hAYAwP5i26EbilokVz5due79Xn5Aaq00kwPxKLkKk3pGphHfaYK/EbMmqpAq3Mi/8Y7s2r2cJ27M1lexWauKuGnPq2lEeQJCFYomZZ0Uq7DM6IWYAQjqnbAchZmhAk6g9MmpoLf9mAAJ4A0XSpf/hqxUZqkvIRZhP+IMMvQK55g38TGgD9LKRk5/o/j8IxIvicnabEilThXHjuIAtMjfms/DtXeAWnx0fUu4LdiVQs4Xk4DIFIaBZ/K25ervZQyOxNddge5Vud3oC+qjuwezBBUYyC7DGLdQVgA== ARC-Authentication-Results: i=1; mx.microsoft.com 1; spf=pass smtp.mailfrom=amd.com; dmarc=pass action=none header.from=amd.com; dkim=pass header.d=amd.com; arc=none DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=amd.com; s=selector1; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-SenderADCheck; bh=+7L0x7xdBg47yhgFzeWgQ1a1i8LcxW2uyIVEmpYszCo=; b=ILnSkechU6A9PhHbg9fIAOE8kVqmzgBTWMHcQ8Szw26BUGng/pFkOtDIL024er1Wz0eXEef6rSKkFTCA0WKFBG4C+M0j4rIEWsoAzjQYhyyTjvRvCbID3NyOefwE4bXY5vP02Pn9RPQC88UH5MX3FXOQDjHxrDN49CbnV5ARYAg= Received: from SJ0PR12MB5673.namprd12.prod.outlook.com (2603:10b6:a03:42b::13) by SJ0PR12MB7458.namprd12.prod.outlook.com (2603:10b6:a03:48d::13) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.21.159.16; Thu, 25 Jun 2026 13:06:23 +0000 Received: from SJ0PR12MB5673.namprd12.prod.outlook.com ([fe80::c3e5:48f8:beb6:ea68]) by SJ0PR12MB5673.namprd12.prod.outlook.com ([fe80::c3e5:48f8:beb6:ea68%5]) with mapi id 15.21.0159.007; Thu, 25 Jun 2026 13:06:22 +0000 Message-ID: Date: Thu, 25 Jun 2026 15:06:15 +0200 User-Agent: Mozilla Thunderbird Subject: =?UTF-8?B?UmU6IOWbnuWkjTogW1BBVENIIDAvM10gbW0vbW11X25vdGlmaWVyLCBk?= =?UTF-8?Q?rm/amdgpu=3A_block_THP_for_GPU_user_mappings?= To: =?UTF-8?B?6JKLIOS6pumfrA==?= , Alex Deucher , David Airlie , Simona Vetter , Felix Kuehling , Andrew Morton , David Hildenbrand , Lorenzo Stoakes , "Yang, Philip" Cc: Zi Yan , Baolin Wang , "Liam R . Howlett" , Nico Pache , Ryan Roberts , Dev Jain , Barry Song , Lance Yang , Vlastimil Babka , Mike Rapoport , Suren Baghdasaryan , Michal Hocko , Jann Horn , "amd-gfx@lists.freedesktop.org" , "dri-devel@lists.freedesktop.org" , "linux-kernel@vger.kernel.org" , "linux-mm@kvack.org" References: <8f68ce2f-c33d-49f5-a671-7a4ab2f4f3d3@amd.com> Content-Language: en-US From: =?UTF-8?Q?Christian_K=C3=B6nig?= In-Reply-To: Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable X-ClientProxiedBy: FR3P281CA0166.DEUP281.PROD.OUTLOOK.COM (2603:10a6:d10:a0::9) To PH7PR12MB5685.namprd12.prod.outlook.com (2603:10b6:510:13c::22) MIME-Version: 1.0 X-MS-PublicTrafficType: Email X-MS-TrafficTypeDiagnostic: SJ0PR12MB5673:EE_|SJ0PR12MB7458:EE_ X-MS-Office365-Filtering-Correlation-Id: 4bf9f905-3e6d-402f-70ae-08ded2ba8dec X-MS-Exchange-SenderADCheck: 1 X-MS-Exchange-AntiSpam-Relay: 0 X-Microsoft-Antispam: BCL:0;ARA:13230040|23010399003|7416014|376014|1800799024|366016|921020|6133799003|20046099003|11063799006|5023799004|56012099006|4143699003|22082099003|18002099003; X-Microsoft-Antispam-Message-Info: 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 X-Forefront-Antispam-Report: CIP:255.255.255.255;CTRY:;LANG:en;SCL:1;SRV:;IPV:NLI;SFV:NSPM;H:SJ0PR12MB5673.namprd12.prod.outlook.com;PTR:;CAT:NONE;SFS:(13230040)(23010399003)(7416014)(376014)(1800799024)(366016)(921020)(6133799003)(20046099003)(11063799006)(5023799004)(56012099006)(4143699003)(22082099003)(18002099003);DIR:OUT;SFP:1101; X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1 X-MS-Exchange-AntiSpam-MessageData-0: =?utf-8?B?ZmtXWWwwQy9qLzIva0Nyc29NL1VqbTNQeC8wUXo5TjhMb1BXUXN2VC9tOUMy?= =?utf-8?B?Si8wS2VnM1p1RDlycnA4KzJXdDZyR0tOVEgyY2ZtY0tyS2h5blN1WTVxRmtr?= =?utf-8?B?eU1PWjVFV0tpZ2JLLy9mVEtXTnpYc24xeGV3amsxR3VneUl6aEF1cnQwRXVo?= =?utf-8?B?QWtQbWs0c015UXY1dWwxRFFsK084M3ZpUWlJRnJ5L1d5ZTFpaFpkbEUxUDdl?= =?utf-8?B?OWhjd003K0diSEpNQ3B6RG43Q01ERXdEcEdCR0dycjlUTGJzMXczT2FCeHFa?= =?utf-8?B?U2hQMWZxczNQMzdiV2FHM1RBVWhWSUNDZE96ZzdMd1FIMjErOGtEWFZLYmFT?= =?utf-8?B?djBJZ1lOVlAva1ZyTHVyRDF4aTVOSHZjYmhrQVhZMWtXeDc5TzVJUjd0SnR2?= =?utf-8?B?Yk5wT3RpeUpOM1duWmlXWlBWZ0oxck1GQzl2VXdIdmRkUWQyM1cwenpEUzRw?= =?utf-8?B?SitYRTV0YWFhd2kyZXZtbWIycmFXNmdMMlhXMlVUbE1hNlpaSk5NdnpqWFFV?= =?utf-8?B?ZEtKcTIrQVhHWjBEbEJoeUVDOTRkWExQZzBqRFZUVWxVcy96Y1pYVUdxTE1D?= =?utf-8?B?R0VHdldkK2VjU0JkQ3FOMlhqQ3RDcnNUMm02eWw1b3JHNnUrQytqWmlUZi83?= =?utf-8?B?ektSLzZBZE9ETWY5YVNmbGJuWjBPemd0cDYwc1JrWFpHbWwrMkNlSDZHdGNv?= =?utf-8?B?UUFUUUdEQkZ1OTFKbW93UWNmcHp4S3RFYmdvUXAwaWNKZTE2S3ZYME12dmdX?= =?utf-8?B?M1lIOU13djhhMWlRdTh5QllqVitFeXZWT1FZQzVYZ1VPR21rb1ArNVIrbkx3?= =?utf-8?B?WnRpbWNza3dqczUvRGg0NkxsQXZXcVlqczVDWW1meUZ2ZnBBQmdYSnJkM3h2?= =?utf-8?B?cVkzU0xvOHRCK1FEc2MzY1pEcXVYYXU4RU11VTVBZjRPNDYrVTFFVkRmRWN4?= =?utf-8?B?RHRoSjBjWS96L0hJZkV4bFNJdjg3T1NOeWNjLzlCNHkxZDJya3Zka0wvZ3hF?= =?utf-8?B?cTB1cE5aVWJLdk51dURIbzhjeWJwU2dudlJsRk1iRzlzeWgvbHNhZlhHdGxt?= =?utf-8?B?WFVsalQyVWZXbzV5Q0Y1UDE0dEFpdllXRk9vckJnZTQ0VTdvUXhJeEZxQ1JT?= =?utf-8?B?QmpHaXc0WUh2TmlNMEJKOUtzMlphRDNiRlBjNE8xT0EveUZ3a0l3MiswZHNG?= =?utf-8?B?ZDNrWkJWdzNnVXUxYTJPV09iQ0lobTZmdmxjQzFGdys0VURQckhYQlBacjha?= =?utf-8?B?RmgxK1crckd4aUtQR0ZkdjVZSVY5L2djWmgrU3lpbEcrNXQ5ZTFsYW5vRkNL?= =?utf-8?B?M0NXNk9xcVk4WUJScDB5dkFmSnFIdGFiTDhiR2huNTZtTVZHOGQrRnNNUERa?= =?utf-8?B?a1lSVUljUWVOd3ZpbTNHMU5vT3JHMUFVeGRHS2pEWG8xQ3JkcHcxRTcwaDE1?= =?utf-8?B?QVRpaEE3RmZwQUh2NFNrbUhNTStIVlg4Z2ppSXloaUxYNUR6OWNWYXFFRnRZ?= =?utf-8?B?akdacjJpVjZ0eFJvSDE0L2ZBUFp0OTdMMVRobEYxWnNLS29CYVhRelpNaC8y?= =?utf-8?B?MUFWSXRVWjdQbjlpNUNVakJYSjVIRnlQb0V1Qm4xWmd6OEk5MUZaYVVEWEx4?= =?utf-8?B?S01jckhDSi8zaDhWVU1IaEpBWVZGM25mUlp1TkxwWllhS3J2VEpYUjNtaWt3?= =?utf-8?B?c2JpdUh4UFJycnZKUERzYVFlY2QvR20wOHNuNndZcW1Ob0MvRThhdU1Fb1RX?= =?utf-8?B?ZjUyQlp5UjU5dUpDN1NkSWdUMWx4aFZvQ3lQeDZVRVJzbm93dkIyb3k5SUZ2?= =?utf-8?B?TWJKNlZBQUI5cFQ1N3l4Sk04VWlaMWRQREI1a3dXWE80NnU3NmNHemRxa3Yw?= =?utf-8?B?ZTd0ZkRYOTNXL0U5bkxMYXdRRXowZ2hhTTZWMUdWdFZuN2RUMXBOZmVBT3Br?= =?utf-8?B?cXpXeGhRUlVnbG4zdlRXTFlOVVZ0V0gwWFhPeWlZSVdFQkNVV2pIaCt2eWZh?= =?utf-8?B?Yjg0bEFRVm81ZU14eDRrbkJyb0hETXZvclBlQTh4TU9QQ1ZadHFOcEZYWHNZ?= =?utf-8?B?NnIvZnBpQ0FqZzlJMTFEMlM3YmdvZW1aRXV4UzFLc0ZEZ2pOTHFndVM2bWFI?= =?utf-8?B?NkVXSUNJaE9OTWpJUW01VXFPVmhiR3Y5ZUlGOGJXblJoQnZrRDFBbC9GQWRX?= =?utf-8?B?dnhWblhSdU9sMC9JMHBUbkhVbFp6QnprQWRDLzFmYWFJY1ppdTgwdGk1bDdO?= =?utf-8?B?YndoMkJIUzJmdWpsRXdLek11RW1yWVBoc2R3amtVa0I1NkdlRlU1aEdzSHFZ?= =?utf-8?Q?dYqEJSd5xsLA0c9NbP?= X-OriginatorOrg: amd.com X-MS-Exchange-CrossTenant-Network-Message-Id: 4bf9f905-3e6d-402f-70ae-08ded2ba8dec X-MS-Exchange-CrossTenant-AuthSource: PH7PR12MB5685.namprd12.prod.outlook.com X-MS-Exchange-CrossTenant-AuthAs: Internal X-MS-Exchange-CrossTenant-OriginalArrivalTime: 25 Jun 2026 13:06:22.5840 (UTC) X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted X-MS-Exchange-CrossTenant-Id: 3dd8961f-e488-4e60-8e11-a82d994e183d X-MS-Exchange-CrossTenant-MailboxType: HOSTED X-MS-Exchange-CrossTenant-UserPrincipalName: +s5x3TqTTiDvGuqOOEP1Q93V0xadwu6SxtnCp+gRfW1nxLuMz6VutCaQM8Py5P4X X-MS-Exchange-Transport-CrossTenantHeadersStamped: SJ0PR12MB7458 X-Rspam-User: X-Stat-Signature: x5ppk4j9gtcxdnfosdthtakj338e6kgg X-Rspamd-Server: rspam10 X-Rspamd-Queue-Id: DFB32100013 X-HE-Tag: 1782392786-561562 X-HE-Meta: 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 HV2iHfqD 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 1zmue7zC Ypp8RNZJwFP/DJNov3QgaMt7YN1VsiXDyLO57Zp3ygnICLfg6XQ788JO5nt5LaT0+qt33KLX1gCE0bTZmT3DkVY/2+HH/dAgvFKR0O+WY1I= Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: List-Subscribe: List-Unsubscribe: Hi Yitao, adding Philip Yang. Thanks for the investigation, that sounds like some kind of bug in the KFD = SVM handling. The driver should be perfectly capable of handling this. I strongly suggest to open up a bug report for ROCm and describe how to rep= roduce this, Philip can probably point you to the right location for that. Regards, Christian. On 6/25/26 15:01, =E8=92=8B =E4=BA=A6=E9=9F=AC wrote: > Hi Christian, >=20 > I agree that my previous approach was wrong. Sorry about that. Please let= me clarify the problem I was seeing and how I ended up with that incorrect= conclusion. >=20 > The original problem was not a synthetic THP test. I was running ROCm/PyT= orch ML training on an AMD Radeon 780M system, and the workload frequently = failed with asynchronous HIP kernel launch failures. The userspace error us= ually surfaced later in PyTorch, for example around a copy/to_device/SetDev= ice path, but the kernel log showed GPU resets and KFD/MES queue eviction f= ailures. >=20 > The relevant kernel messages I repeatedly saw were along these lines: >=20 > =C2=A0 MES failed to respond to msg=3DREMOVE_QUEUE > =C2=A0 MES failed to respond to msg=3DSUSPEND > =C2=A0 failed to suspend all gangs > =C2=A0 failed to remove hardware queue from MES > =C2=A0 Failed to evict queue > =C2=A0 Failed to evict process queues > =C2=A0 GPU reset begin >=20 > While trying to reduce the issue, I saw memory invalidations and THP-rela= ted page-table/backing-page activity driving the AMDGPU/KFD path through SV= M eviction. On this system, the path I was looking at was roughly: >=20 > =C2=A0 svm_range_cpu_invalidate_pagetables() > =C2=A0 =C2=A0 -> svm_range_evict() > =C2=A0 =C2=A0 -> kgd2kfd_quiesce_mm() > =C2=A0 =C2=A0 -> KFD process queue eviction > =C2=A0 =C2=A0 -> MES REMOVE_QUEUE / SUSPEND >=20 > One thing that misled me was the XNACK-disabled path. Since the issue app= eared on an XNACK-disabled APU, and that path requires queue eviction/quies= ce when CPU page table invalidations affect GPU mappings, I incorrectly tho= ught the backing-page change itself was something the driver had to prevent= . >=20 > Another thing that misled me was that the application was not intentional= ly asking for THP behavior. From the workload=E2=80=99s point of view, thes= e page transitions looked unrelated to the model computation. I therefore i= ncorrectly assumed that userspace should not be able to change backing-page= characteristics in a way that affects a driver mapping already registered = with MMU interval notifiers. I now understand from the MM feedback that thi= s is expected behavior, and that the notifier user must handle unmap/remap = correctly. >=20 > So the more precise problem is that THP/remap is only one way to trigger = the invalidation path. What is failing for my workload is the AMDGPU/KFD/ME= S queue quiesce/eviction path during those invalidations. When that fails, = the GPU resets, and userspace later observes an asynchronous HIP failure. >=20 > Please allow me to continue investigating a more appropriate fix for this= problem. I will try to keep the fix boundary within AMDGPU/KFD/MES and avo= id changing MM-core or THP policy semantics. >=20 > Regards, > Yitao > -------------------------------------------------------------------------= ---------------------------------------------------------------------------= ---------------------------------------------------------------------------= ---------------------------------------------------------------------------= ---------------------------------------------------------------------------= ---------------------------------------------------------------------------= ---------------------------------------------------------------------------= ---------------------------------------------------------------------------= ---------------------------------------------------------------------------= ---------------------------------------------------------------------------= ---------------------------------------------------------------------------= ---------------------------------------------------------------------------= ---------------------------------------------------------------------------= ----------------- > *=E5=8F=91=E4=BB=B6=E4=BA=BA:* Christian K=C3=B6nig > *=E5=8F=91=E9=80=81=E6=97=B6=E9=97=B4:* 2026=E5=B9=B46=E6=9C=8825=E6=97= =A5 8:35 > *=E6=94=B6=E4=BB=B6=E4=BA=BA:* Yitao Jiang ; Al= ex Deucher ; David Airlie ; S= imona Vetter ; Felix Kuehling ; An= drew Morton ; David Hildenbrand ; Lorenzo Stoakes > *=E6=8A=84=E9=80=81:* Zi Yan ; Baolin Wang ; Liam R . Howlett ; Nico Pache ; Ryan Roberts ; Dev Jain ; Barry Song ; Lance Yang ; Vl= astimil Babka ; Mike Rapoport ; Suren B= aghdasaryan ; Michal Hocko ; Jann Horn = ; amd-gfx@lists.freedesktop.org ; dri-devel@lists.freedesktop.org ; = linux-kernel@vger.kernel.org ; linux-mm@kvack= .org > *=E4=B8=BB=E9=A2=98:* Re: [PATCH 0/3] mm/mmu_notifier, drm/amdgpu: block = THP for GPU user mappings > =C2=A0 > On 6/25/26 12:59, Yitao Jiang wrote: >> Hi, >>=20 >> This series fixes a THP policy problem I found while debugging >> frequent ROCm GPU failures on an AMD Radeon 780M system during ML >> training. >>=20 >> Some AMDGPU/KFD user mappings are registered through interval >> notifiers and cannot safely tolerate the backing VMA changing from base >> pages to a transparent huge page after registration. >=20 > That's certainly not correct. This is a must have for a whole lot of use = cases. >=20 > Why exactly isn't that working for your use case? >=20 > Regards, > Christian. >=20 >> Userspace can >> still apply MADV_HUGEPAGE or MADV_COLLAPSE, and khugepaged can also >> collapse the range, after the GPU mapping has been registered. >>=20 >> On my system this showed up as asynchronous ROCm/HIP kernel launch >> failures, often reported later at a synchronization or copy point. I >> expect the issue to be relevant to AMDGPU/KFD mappings on >> XNACK-disabled GPUs more generally, because those mappings cannot rely >> on replayable GPU faults after a CPU-side THP remap. I have validated >> the failure and fix on AMD Radeon 780M / gfx1103. >>=20 >> Patch 1 adds MMU_INTERVAL_NOTIFIER_BLOCK_THP so interval notifier >> users can ask the MM core to keep the covered VMA range out of THP >> while the notifier is active. The MM core applies VM_NOHUGEPAGE and >> clears VM_HUGEPAGE under mmap_lock for write. A later MADV_HUGEPAGE >> over an active opt-in range is treated as an ignored hint, and >> MADV_COLLAPSE is rejected by the existing VM_NOHUGEPAGE checks. >>=20 >> Patches 2 and 3 opt in the AMDGPU/KFD paths that need this behavior: >> HSA userptr BOs, KFD SVM ranges when XNACK is disabled, and >> GPU_ALWAYS_MAPPED SVM ranges. Other interval notifier users keep their >> current behavior. >>=20 >> This does not disable THP globally and does not add work to GPU >> command submission or kernel launch paths. Additional work is limited >> to opt-in notifier registration, opt-in notifier flag transitions, and >> MADV_HUGEPAGE attempts that overlap an active opt-in range. >>=20 >> I tested this on top of torvalds/linux commit ab9de95c9cf9 with: >>=20 >>=C2=A0=C2=A0 - scripts/checkpatch.pl --strict --no-tree >>=C2=A0=C2=A0 - git apply --check >>=C2=A0=C2=A0 - x86_64 defconfig build with TRANSPARENT_HUGEPAGE=3Dy, >>=C2=A0=C2=A0=C2=A0=C2=A0 DRM_AMDGPU=3Dm, and HSA_AMD=3Dy for mm/ and AMDG= PU/KFD objects >>=C2=A0=C2=A0 - standalone HSA/HIP reproducers and the ROCm/PyTorch worklo= ad that >>=C2=A0=C2=A0=C2=A0=C2=A0 originally exposed the failure on my Radeon 780M= system >>=20 >> The standalone reproducers depend on ROCm userspace libraries, so I >> have not included them in this series. I can send them separately if >> useful. >>=20 >> This series was prepared with assistance from OpenAI Codex (GPT-5.5). >> I reviewed the resulting code and take responsibility for the >> submission. >>=20 >> Yitao Jiang (3): >>=C2=A0=C2=A0 mm/mmu_notifier: let interval notifiers block THP >>=C2=A0=C2=A0 drm/amdgpu: block THP for HSA userptr notifiers >>=C2=A0=C2=A0 drm/amdkfd: block THP for non-replayable SVM ranges >>=20 >>=C2=A0 drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c |=C2=A0 25 ++- >>=C2=A0 drivers/gpu/drm/amd/amdkfd/kfd_svm.c=C2=A0=C2=A0=C2=A0 |=C2=A0 36 = ++++- >>=C2=A0 include/linux/huge_mm.h=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0 5 +- >>=C2=A0 include/linux/mmu_notifier.h=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0 28 ++++ >>=C2=A0 mm/khugepaged.c=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0 9 +- >>=C2=A0 mm/madvise.c=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0 3 +- >>=C2=A0 mm/mmu_notifier.c=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0 | 204 +++++++++++++++++++++++- >>=C2=A0 7 files changed, 286 insertions(+), 24 deletions(-) >>=20 >>=20 >> base-commit: ab9de95c9cf952332ab79453b4b5d1bfca8e514f >> -- >> 2.53.0 >=20