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 gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id C0C43C5DF7D for ; Fri, 21 Aug 2026 19:53:07 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 62D6C10F483; Fri, 21 Aug 2026 19:53:07 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (1024-bit key; unprotected) header.d=amd.com header.i=@amd.com header.b="SYTPdKNm"; dkim-atps=neutral Received: from CY7PR03CU001.outbound.protection.outlook.com (mail-westcentralusazon11010000.outbound.protection.outlook.com [40.93.198.0]) by gabe.freedesktop.org (Postfix) with ESMTPS id 84C9810F483 for ; Fri, 21 Aug 2026 19:53:05 +0000 (UTC) ARC-Seal: i=1; a=rsa-sha256; s=arcselector10001; d=microsoft.com; cv=none; b=Ml+IKoXsHZceYG/XuBr50SGlFCqafwCJEPfaLwnPHocWVV1lZYQ2W2kErYeo+KwryjN3yk4kVu5efpHjTTi0qGkh0Q/vW5o8IAPWlQAdqx35epOkJzQNRbjF95ntND1rIvNhKkCllfZ+TE9U/FvfFMEFkX/ZZVMR8K73H6AZE2AVvqsLKJ6414E9Sqe9Bgu18XiIDYHD2+aRS9+9A5q2ozyj8iLP2xHO52G8yI/4t8xDZalm/rwy+4oa4GjwJ9F8gzTgQgeKVcIMX9URp7xKCsnuoaW8VsoA/vZxMlXQ9V/2uVGwQwSmpUQcN5PfYWAGtvp4fHAjZijbRQJ8IK7ryw== 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=LXdY2nVrfUOMK3e1cKCwqXpw/G85++Aa7GmorS3fU3g=; b=hFhYXXpmoWvCsCab9KtjA0MQjZ7CMs3ZCQXsam0A2IqW6BN9DRCtwvMWPwoXkZ01mkNGrkgb2S68bBK2/yeV9p/c44kD8S6G3gEsmBvYpfCwSZXdJoTsHwteOmsYcjtC719IJwghK0VPOD3KWmeFsEmn1qTOJG6qoTckLGfQVttHgTVzjZXb/zZTIxidciUGyhfLZBSKL5fpSrVFtu0OGqLhV388IEx5Ouw/eJ/dZzggdXagDDsLITgshOCQMTp4+bZ7FyVUuADLlf99TAtrGmcl6MlL+RH8XEEhCS33UAhQb30OR0LNcwkUHmBz6ncIBKFtpAW6pfzOeDkkUvTsfQ== ARC-Authentication-Results: i=1; mx.microsoft.com 1; spf=pass (sender ip is 165.204.84.17) smtp.rcpttodomain=lists.freedesktop.org smtp.mailfrom=amd.com; dmarc=pass (p=quarantine sp=quarantine pct=100) action=none header.from=amd.com; dkim=none (message not signed); arc=none (0) 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=LXdY2nVrfUOMK3e1cKCwqXpw/G85++Aa7GmorS3fU3g=; b=SYTPdKNmbmKpwT/LR090nVz+wnpvp/NazNgRw+5pTXUMArOABQi3xUWRJY2GhfNi0F1cwdKG3rAmNn3zZEN+hBYbL3lN/X6pE0a91BLAtdguYLpa777hyhFzXiVrw73sKvcyTlkIzg7Az0q/K7GnpGs92pBD9JYobUPEILhDwbI= Received: from MW4PR04CA0037.namprd04.prod.outlook.com (2603:10b6:303:6a::12) by PH7PR12MB9255.namprd12.prod.outlook.com (2603:10b6:510:30c::19) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.21.339.10; Fri, 21 Aug 2026 19:35:20 +0000 Received: from CO1PEPF000066EA.namprd05.prod.outlook.com (2603:10b6:303:6a:cafe::8b) by MW4PR04CA0037.outlook.office365.com (2603:10b6:303:6a::12) with Microsoft SMTP Server (version=TLS1_3, cipher=TLS_AES_256_GCM_SHA384) id 15.21.339.11 via Frontend Transport; Fri, 21 Aug 2026 19:35:18 +0000 X-MS-Exchange-Authentication-Results: spf=pass (sender IP is 165.204.84.17) smtp.mailfrom=amd.com; dkim=none (message not signed) header.d=none;dmarc=pass action=none header.from=amd.com; Received-SPF: Pass (protection.outlook.com: domain of amd.com designates 165.204.84.17 as permitted sender) receiver=protection.outlook.com; client-ip=165.204.84.17; helo=satlexmb08.amd.com; pr=C Received: from satlexmb08.amd.com (165.204.84.17) by CO1PEPF000066EA.mail.protection.outlook.com (10.167.249.5) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.21.360.3 via Frontend Transport; Fri, 21 Aug 2026 19:35:17 +0000 Received: from Satlexmb09.amd.com (10.181.42.218) by satlexmb08.amd.com (10.181.42.217) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.2.2562.45; Fri, 21 Aug 2026 14:35:16 -0500 Received: from satlexmb07.amd.com (10.181.42.216) by satlexmb09.amd.com (10.181.42.218) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.2.2562.45; Fri, 21 Aug 2026 14:35:15 -0500 Received: from p8.amd.com (10.180.168.240) by satlexmb07.amd.com (10.181.42.216) with Microsoft SMTP Server id 15.2.2562.45 via Frontend Transport; Fri, 21 Aug 2026 14:35:15 -0500 From: Alex Deucher To: CC: Mukul Joshi , Felix Kuehling , Alex Deucher Subject: [PATCH 07/95] drm/amdgpu: Add UALink manager core infrastructure Date: Fri, 21 Aug 2026 15:33:30 -0400 Message-ID: <20260821193458.808626-8-alexander.deucher@amd.com> X-Mailer: git-send-email 2.55.0 In-Reply-To: <20260821193458.808626-1-alexander.deucher@amd.com> References: <20260821193458.808626-1-alexander.deucher@amd.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Content-Type: text/plain X-EOPAttributedMessage: 0 X-MS-PublicTrafficType: Email X-MS-TrafficTypeDiagnostic: CO1PEPF000066EA:EE_|PH7PR12MB9255:EE_ X-MS-Office365-Filtering-Correlation-Id: 672613a2-c36b-406b-caa7-08deffbb54ab X-MS-Exchange-SenderADCheck: 1 X-MS-Exchange-AntiSpam-Relay: 0 X-Microsoft-Antispam: BCL:0; ARA:13230040|376014|23010399003|1800799024|36860700016|82310400026|11063799006|10067099003|56012099006|18002099003|22082099003; X-Microsoft-Antispam-Message-Info: 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 X-Forefront-Antispam-Report: CIP:165.204.84.17; CTRY:US; LANG:en; SCL:1; SRV:; IPV:NLI; SFV:NSPM; H:satlexmb08.amd.com; PTR:InfoDomainNonexistent; CAT:NONE; SFS:(13230040)(376014)(23010399003)(1800799024)(36860700016)(82310400026)(11063799006)(10067099003)(56012099006)(18002099003)(22082099003); DIR:OUT; SFP:1101; X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1 X-MS-Exchange-AntiSpam-MessageData-0: JT7SqiE7Ps6FxVA+UfONJLVwUhzKnvY133azLQYeFD8AuykBuIzq4ZaVwQyhqkhzTrX1yWlojtBZjXYI4S0kWeA4QgXLURAuWHx+sMLZ8Msmuf86FPh0GUUGkOl4sjkwO/MsJQBBCSowmX56ulHaK/7k9HWjC4hCZs9u8VOtNrQFRHzrEcxAvrzx9Jdqqk7PadeqsYtAsTDDkYdnP034UfD/mNlSx5xTjSeAZFBLuxAj0P6PGNC88GgoMfksAMJYV3gvl9Dwj+FMRjqjHjGmHtYODW0/mn/ElmHP/mYZPQHYCf6k07AVILemHtWpOIKXClqQzFgBaFBlJDJ1uFTqmtOI2i0QCWROTlOJP0UwRKu2GvMyNomqPMApUP7BYRRmZT2LROUMjQH/6tjd593y2TIJpkmOe2DSDrizMN/v1jHGW4i4XeGz0tvSy6Tx4BwQ X-OriginatorOrg: amd.com X-MS-Exchange-CrossTenant-OriginalArrivalTime: 21 Aug 2026 19:35:17.6367 (UTC) X-MS-Exchange-CrossTenant-Network-Message-Id: 672613a2-c36b-406b-caa7-08deffbb54ab X-MS-Exchange-CrossTenant-Id: 3dd8961f-e488-4e60-8e11-a82d994e183d X-MS-Exchange-CrossTenant-OriginalAttributedTenantConnectingIp: TenantId=3dd8961f-e488-4e60-8e11-a82d994e183d; Ip=[165.204.84.17]; Helo=[satlexmb08.amd.com] X-MS-Exchange-CrossTenant-AuthSource: CO1PEPF000066EA.namprd05.prod.outlook.com X-MS-Exchange-CrossTenant-AuthAs: Anonymous X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem X-MS-Exchange-Transport-CrossTenantHeadersStamped: PH7PR12MB9255 X-BeenThere: amd-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Discussion list for AMD gfx List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: amd-gfx-bounces@lists.freedesktop.org Sender: "amd-gfx" From: Mukul Joshi This patch introduces AMDGPU UALink manager core infrastructure which will be responsible for sharing buffer objects across multiple GPUs, with each GPU running their own OS, in a rack scale setup. The core infrastructure of the ualink manager will include: - Store BOs exported from a GPU. - Store BOs imported by a GPU. - Store connection state with different GPUs. Signed-off-by: Mukul Joshi Reviewed-by: Felix Kuehling Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.c | 42 +++++++++++++++++++++- drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.h | 41 ++++++++++++++++++++- 2 files changed, 81 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.c index 4f50040a99256..f5ff9f1e9ce0c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.c @@ -22,9 +22,10 @@ * */ +#include +#include "amdgpu.h" #include "amdgpu_ualink.h" #include "amdgpu_xgmi.h" -#include "amdgpu.h" #include #include @@ -958,3 +959,42 @@ void amdgpu_ualink_sysfs_fini(struct amdgpu_device *adev) adev->ualink.info = NULL; } } + +int amdgpu_ualink_manager_start(struct amdgpu_device *adev) +{ + int i, r; + + adev->ualink.npa_wq = alloc_workqueue("NPA WQ", WQ_UNBOUND, 0); + if (unlikely(!adev->ualink.npa_wq)) { + dev_err(adev->dev, "Failed to allocate NPA WQ\n"); + return -ENOMEM; + } + + xa_init_flags(&adev->ualink.exp_xa, XA_FLAGS_LOCK_BH); + xa_init_flags(&adev->ualink.imp_xa, XA_FLAGS_LOCK_BH); + xa_init_flags(&adev->ualink.handle_invalid_xa, XA_FLAGS_LOCK_BH); + + for (i = 0; i < AMDGPU_UALINK_ACCEL_MAX; i++) { + init_completion(&adev->ualink.conn_state[i].hello_done); + mutex_init(&adev->ualink.conn_state[i].lock); + adev->ualink.conn_state[i].state = AMDGPU_UALINK_CONN_NOT_READY; + INIT_LIST_HEAD(&adev->ualink.exp_handles_list[i]); + INIT_LIST_HEAD(&adev->ualink.imp_handles_list[i]); + } + + return 0; +} + +void amdgpu_ualink_manager_stop(struct amdgpu_device *adev) +{ + int i; + + xa_destroy(&adev->ualink.exp_xa); + xa_destroy(&adev->ualink.imp_xa); + xa_destroy(&adev->ualink.handle_invalid_xa); + + for (i = 0; i < AMDGPU_UALINK_ACCEL_MAX; i++) + mutex_destroy(&adev->ualink.conn_state[i].lock); + + destroy_workqueue(adev->ualink.npa_wq); +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.h index 384959fbb36fb..f2be71232813d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.h @@ -31,6 +31,12 @@ #define AMDGPU_UALINK_LOCAL_ACCELS_MAX 8 #define AMDGPU_UALINK_STATIONS_MAX 64 +enum amdgpu_ualink_conn_state { + AMDGPU_UALINK_CONN_NOT_READY = 0, + AMDGPU_UALINK_CONN_IN_PROGRESS = 1, + AMDGPU_UALINK_CONN_ESTABLISHED = 2 +}; + enum amdgpu_ualink_type { AMDGPU_UALINK_TYPE_UALOE = 0, AMDGPU_UALINK_TYPE_UALINK = 1, @@ -115,14 +121,47 @@ struct amdgpu_ualink_station_config { }; #define to_ualink_station_config(ko) container_of(ko, struct amdgpu_ualink_station_config, kobj) +struct amdgpu_ualink_connection { + struct completion hello_done; + struct mutex lock; + u32 generation_count; + enum amdgpu_ualink_conn_state state; +}; + struct amdgpu_ualink_mgr { struct amdgpu_ualink_info *info; struct amdgpu_ualink_ppod_setup *setup; struct amdgpu_ualink_vpod_config *config; struct amdgpu_ualink_station_config *stations; + + /* Xarray to store info about exported BOs */ + struct xarray exp_xa; + + /* Xarray to store info about imported ualink handles */ + struct xarray imp_xa; + + /* Xarray to store handles to be deleted */ + struct xarray handle_invalid_xa; + + /* Array to store connection state per remote GPU */ + struct amdgpu_ualink_connection conn_state[AMDGPU_UALINK_ACCEL_MAX]; + + /* List of exported ualink handles per GPU. Used to quickly traverse + * the list of all handles exported to a remote GPU. + */ + struct list_head exp_handles_list[AMDGPU_UALINK_ACCEL_MAX]; + + /* List of imported ualink handles per GPU. Used to quickly traverse + * the list of all handles imported from a remote GPU. + */ + struct list_head imp_handles_list[AMDGPU_UALINK_ACCEL_MAX]; + + /* WQ to manage revocation of exported memory. */ + struct workqueue_struct *npa_wq; }; int amdgpu_ualink_sysfs_init(struct amdgpu_device *adev); void amdgpu_ualink_sysfs_fini(struct amdgpu_device *adev); - +int amdgpu_ualink_manager_start(struct amdgpu_device *adev); +void amdgpu_ualink_manager_stop(struct amdgpu_device *adev); #endif -- 2.55.0