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 E9AEBC5DF8C for ; Fri, 21 Aug 2026 20:00:50 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 7927210F432; Fri, 21 Aug 2026 19:40:36 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (1024-bit key; unprotected) header.d=amd.com header.i=@amd.com header.b="kOjWqcGi"; dkim-atps=neutral Received: from CO1PR03CU002.outbound.protection.outlook.com (mail-westus2azon11010055.outbound.protection.outlook.com [52.101.46.55]) by gabe.freedesktop.org (Postfix) with ESMTPS id 61E0C10F432 for ; Fri, 21 Aug 2026 19:40:26 +0000 (UTC) ARC-Seal: i=1; a=rsa-sha256; s=arcselector10001; d=microsoft.com; cv=none; b=dOd1o8Nb3YQb0kxEsPtoYSLWZNQPJSG/NeGfrMrIp7BmrwVZh2X3RK4Le8purKak2mXb1LH/ETQpNI3BhftzsdNxD2kEAkjKQw4yXog2JSiS8LLSarsh3I840MZmu6vg6pzFJibLijV7kI9PO+u+0gMeDYj9vdveDinux9cvCjmzjV0kWiiBQhw1bd9KgPgIaWfG8DLGWaDyezIJJCilmrYz5sHQNRj5UryEhIMImeKt0x7cUvbsaB9qVDQK0iVccJHGzinnlph3aDINQBXJsvFKOVEdEMWzIUajqlY0SVcJSwvni5wsc5omne9JOYbm/MqVTBU8RA9Y1r2n3bRKQw== 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=30GMn799noB74wJmPP/nB74hySgZ8OHLrAT6A+Be8Io=; b=BfgEVExlaIL9L91voXB3VT2RbBAiVCUDL80pznrCfJkxX6gICHMOC9KthKWPXKJ5yhfD5ea9bGLxAJp/kwZqlhaiWvN06UybOv+hYzOjGkhnwCKCZzjXFVKXijD5UItvXAYNuHinM7sbXUeMR++FqmrmG1SIhMY1h4vqomDJDJzzebgebvZRdx7Nwiy4ONIC85VnMOrnzGhqxtE5IOIfKGOGKgootnR8cP+0G+892tw3lnfjP5kVwlTXM0cg1H0pcHDVFXzMxIV6/KBnjaOKQLIgURTHJE/mySMyn1pDWgYSjpmUTBcZ04IToKNkCS2Y/vAkqxaPwARPhULR2By7lQ== 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=30GMn799noB74wJmPP/nB74hySgZ8OHLrAT6A+Be8Io=; b=kOjWqcGiK5DaQ87zZpso1RZLGxL1qVhoAp6/4mYBytUvKzA4l6uHuFVZRNsCp3QR7iBdKuv7ujyZT9V6XdvKUEyQ++xPk3K1bgKEh3ht2IntfgipYMaRuPCI+VHdBNgwZ/WJs2isug/7Cgov6iw7oK0IDr27viQ/mxRuI9Fj//g= Received: from DS1P222CA0014.NAMP222.PROD.OUTLOOK.COM (2603:10b6:8:454::8) by CH3PR12MB9026.namprd12.prod.outlook.com (2603:10b6:610:125::15) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.21.339.11; Fri, 21 Aug 2026 19:36:04 +0000 Received: from CY4PEPF0000EDD1.namprd03.prod.outlook.com (2603:10b6:8:454:cafe::31) by DS1P222CA0014.outlook.office365.com (2603:10b6:8:454::8) 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 19:36:03 +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 CY4PEPF0000EDD1.mail.protection.outlook.com (10.167.241.197) 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 19:36:03 +0000 Received: from satlexmb10.amd.com (10.181.42.219) 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 14:36:01 -0500 Received: from satlexmb07.amd.com (10.181.42.216) by satlexmb10.amd.com (10.181.42.219) 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 14:35:59 -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 14:35:58 -0500 From: Alex Deucher To: CC: Lijo Lazar , Felix Kuehling , Alex Deucher Subject: [PATCH 78/95] drm/amdgpu: Assign accel state based on ASP config Date: Fri, 21 Aug 2026 15:34:41 -0400 Message-ID: <20260821193458.808626-79-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: CY4PEPF0000EDD1:EE_|CH3PR12MB9026:EE_ X-MS-Office365-Filtering-Correlation-Id: d4ef2b00-d290-4044-75b7-08deffbb6fc0 X-MS-Exchange-SenderADCheck: 1 X-MS-Exchange-AntiSpam-Relay: 0 X-Microsoft-Antispam: BCL:0; ARA:13230040|23010399003|82310400026|376014|1800799024|36860700016|6133799003|10067099003|11063799006|56012099006|22082099003|18002099003; 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)(23010399003)(82310400026)(376014)(1800799024)(36860700016)(6133799003)(10067099003)(11063799006)(56012099006)(22082099003)(18002099003); DIR:OUT; SFP:1101; X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1 X-MS-Exchange-AntiSpam-MessageData-0: Lc7BKWTfT+de/ACwSHi2yp1ctwGrd4WKeOFjqDpJ9x+Dj5wb4utZoy90H/EPy4L/h8cmu7h7jrnjjLxkyeSHaz10JZiVyYRNjjHihs04hIbumw6iRsITXU+m9hwr1X9VkMermLjR0uVtZnIhGTBdm/u/wvKiLB2hjFuh9GvOOMQHfvetWI+ZH5pHTPeu6xrwHdV2Ub0IpGU1K0FRiIG9HXd5fARBm5/iQAgopUwhA+iqscLxT+pG9dqagi6j3/ZFUnA16CmxnxgtWmgFIxtoyVK9tEhK58Qr5i6OFQ3JbV6tsqlmGiDtYRRpKX7Yp9DK8/3KUkZa12ZVce5arcFr561rEU1lsAMNKyMww2Qu7kZ5eYd8as4CMuoIqMC0MBo0N99ZNWbLnJRC5NLeBaDj4/z0PFQOhorUw9lw5OUCiUqhQqQ2Ka6H24UyA0nEQg/W X-OriginatorOrg: amd.com X-MS-Exchange-CrossTenant-OriginalArrivalTime: 21 Aug 2026 19:36:03.1181 (UTC) X-MS-Exchange-CrossTenant-Network-Message-Id: d4ef2b00-d290-4044-75b7-08deffbb6fc0 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: CY4PEPF0000EDD1.namprd03.prod.outlook.com X-MS-Exchange-CrossTenant-AuthAs: Anonymous X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem X-MS-Exchange-Transport-CrossTenantHeadersStamped: CH3PR12MB9026 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: Lijo Lazar Add explicit PPOD/VPOD states for clarity. Assign the state accordingly as received from ASP during initialization. UALink accel state transitions: UNCONFIGURED -> PPOD_CONFIGURED -> VPOD_CONFIGURED -> READY -> ACTIVE - PPOD: sysfs ppod commit/ ASP {UAL_CFG_PPOD} - VPOD: sysfs vpod commit/ ASP {UAL_CFG_VPOD} | {UAL_CFG_STATION} - READY: local vpod integrity + activate_accelerator() - ACTIVE: all local gpus belonging to one vpod are configured and ready. deactivate_accelerator(): READY -> PPOD_CONFIGURED invalid vpod: -> ERROR Query ualink info during late init phase. Add a local vpod integrity check before activating devices in the vpod. Move the devices to active state once integrity check passes. Signed-off-by: Lijo Lazar Reviewed-by: Felix Kuehling Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 6 +- drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h | 3 +- drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.c | 320 +++++++++++++++++---- drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.h | 3 +- 4 files changed, 278 insertions(+), 54 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c index b8def8b7e1426..4fc7aed58bfc3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c @@ -1223,7 +1223,8 @@ int psp_ual_get_interface_version(struct psp_context *psp, uint32_t *intf_ver) } int psp_ual_query_info(struct psp_context *psp, uint32_t intf_ver, - struct amdgpu_ualink_info *info) + struct amdgpu_ualink_info *info, + enum psp_gfx_ual_config_state *cfg_state) { struct psp_gfx_get_config_ual_v1 *ual_config; struct psp_gfx_cmd_resp *cmd; @@ -1269,6 +1270,9 @@ int psp_ual_query_info(struct psp_context *psp, uint32_t intf_ver, if (AMDGPU_UALINK_ACCEL_MAX > PSP_GFX_UAL_MAX_ACC_BIT_MASK*32) bitmap_clear(info->vpod.active_accel_bits, PSP_GFX_UAL_MAX_ACC_BIT_MASK*32, AMDGPU_UALINK_ACCEL_MAX - PSP_GFX_UAL_MAX_ACC_BIT_MASK*32); + + if (cfg_state) + *cfg_state = ual_config->config_state; } else if (!ret) { ret = -EINVAL; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h index bb84f3a3dc54b..2ba7ced36b7cc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h @@ -715,7 +715,8 @@ struct amdgpu_ualink_station_config; int psp_ual_get_interface_version(struct psp_context *psp, uint32_t *intf_ver); int psp_ual_query_info(struct psp_context *psp, uint32_t intf_ver, - struct amdgpu_ualink_info *info); + struct amdgpu_ualink_info *info, + enum psp_gfx_ual_config_state *cfg_state); int psp_ual_set_ppod_config(struct psp_context *psp, uint32_t intf_ver, const struct amdgpu_ualink_ppod_setup *setup); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.c index d376c3136af8a..a3758c55c0b73 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.c @@ -29,10 +29,12 @@ #include "amdgpu_ualink.h" #include "amdgpu_xgmi.h" #include "amdgpu_dma_buf.h" +#include "psp_gfx_if.h" #include #include static void deactivate_accelerator(struct amdgpu_device *adev); +static void amdgpu_ualink_activate_vpod(struct amdgpu_device *adev); static int amdgpu_ualink_remote_interrupt(struct amdgpu_device *adev, u32 remote_accel_id, u32 dw0, u32 dw1, u32 dw2, u32 dw3); @@ -95,6 +97,113 @@ static void amdgpu_ualink_object_fini(struct amdgpu_device *adev) adev->ualink.info = NULL; } +static bool __check_ppod_info(struct amdgpu_device *adev, + const struct amdgpu_ualink_info *info) +{ + const struct amdgpu_ualink_ppod_info *ppod = &info->ppod; + + if (ppod->size <= 0 || ppod->size > AMDGPU_UALINK_ACCEL_MAX) { + dev_dbg(adev->dev, "pPod size %u out of range [1..%u]\n", + ppod->size, AMDGPU_UALINK_ACCEL_MAX); + return false; + } + if (ppod->accel_id >= ppod->size) { + dev_dbg(adev->dev, + "Accelerator ID %u greater or equal pPod size %u\n", + ppod->accel_id, ppod->size); + return false; + } + + return true; +} + +static bool __check_vpod_info(struct amdgpu_device *adev, + const struct amdgpu_ualink_info *info) +{ + const struct amdgpu_ualink_ppod_info *ppod = &info->ppod; + const struct amdgpu_ualink_vpod_info *vpod = &info->vpod; + unsigned int weight; + + if (vpod->size == 0 || vpod->size > ppod->size) { + dev_dbg(adev->dev, "vPod size %u out of range [1..%u]\n", + vpod->size, ppod->size); + return false; + } + if (vpod->addr_mode >= AMDGPU_UALINK_ADDR_MODE_MAX) { + dev_dbg(adev->dev, "Invalid addr mode %u\n", vpod->id); + return false; + } + weight = + bitmap_weight(vpod->active_accel_bits, AMDGPU_UALINK_ACCEL_MAX); + if (weight != vpod->size) { + dev_dbg(adev->dev, + "vPod size doesn't match vpod_active_accels list: %u != %u\n", + vpod->size, weight); + return false; + } + if (!test_bit(ppod->accel_id, vpod->active_accel_bits)) { + dev_dbg(adev->dev, + "Accelerator ID %u not listed in vpod_active_accels\n", + ppod->accel_id); + return false; + } + + return true; +} + +static void +amdgpu_ualink_info_set_accel_state(struct amdgpu_device *adev, + struct amdgpu_ualink_info *info, + enum psp_gfx_ual_config_state cfg_state) +{ + if (!info) + return; + + switch (cfg_state) { + case UAL_CFG_IDLE: + break; + case UAL_CFG_PPOD: + if (!__check_ppod_info(adev, info)) { + info->accel_state = + AMDGPU_UALINK_ACCEL_STATE_UNCONFIGURED; + break; + } + info->accel_state = AMDGPU_UALINK_ACCEL_STATE_PPOD_CONFIGURED; + break; + case UAL_CFG_VPOD: + case UAL_CFG_STATION: + if (!__check_vpod_info(adev, info)) { + info->accel_state = AMDGPU_UALINK_ACCEL_STATE_ERROR; + dev_err(adev->dev, + "vpod configuration is invalid, setting to error state"); + break; + } + info->accel_state = AMDGPU_UALINK_ACCEL_STATE_VPOD_CONFIGURED; + break; + case UAL_CFG_COMPLETE: + info->accel_state = AMDGPU_UALINK_ACCEL_STATE_READY; + break; + default: + dev_dbg(adev->dev, "invalid configuration state %u", cfg_state); + break; + } +} + +static int amdgpu_ualink_query_info(struct amdgpu_device *adev) +{ + enum psp_gfx_ual_config_state cfg_state; + int r; + + r = psp_ual_query_info(&adev->psp, adev->ualink.psp_if_ver, + adev->ualink.info, &cfg_state); + if (r) + return r; + + amdgpu_ualink_info_set_accel_state(adev, adev->ualink.info, cfg_state); + + return 0; +} + int ualink_ip_hw_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; @@ -103,6 +212,7 @@ int ualink_ip_hw_init(struct amdgpu_ip_block *ip_block) if (!adev->ualink.info) return 0; + adev->ualink.info->accel_state = AMDGPU_UALINK_ACCEL_STATE_UNCONFIGURED; r = psp_ual_get_interface_version(&adev->psp, &adev->ualink.psp_if_ver); if (r) { adev->ualink.psp_if_ver = 0xffffffff; @@ -114,19 +224,9 @@ int ualink_ip_hw_init(struct amdgpu_ip_block *ip_block) dev_info(adev->dev, "Found UALink interface version 0x%x\n", adev->ualink.psp_if_ver); - r = psp_ual_query_info(&adev->psp, adev->ualink.psp_if_ver, - adev->ualink.info); - if (r) { - dev_info(adev->dev, - "UALink disabled, failed to query initial config: %d\n", - r); - goto disable; - } - adev->ualink.mgr_state = AMDGPU_UALINK_INIT_HW; return 0; - disable: adev->ualink.mgr_state = AMDGPU_UALINK_INIT_ERROR; return 0; @@ -140,6 +240,12 @@ int ualink_ip_late_init(struct amdgpu_ip_block *ip_block) if (adev->ualink.mgr_state != AMDGPU_UALINK_INIT_HW) return 0; + r = amdgpu_ualink_query_info(adev); + if (r) + return r; + + amdgpu_ualink_activate_vpod(adev); + r = amdgpu_ualink_drm_client_create(adev); if (r) { dev_err(adev->dev, "Failed to create UALink DRM client: %d\n", @@ -406,8 +512,15 @@ static const char * const ualink_addr_mode_values[] = { "source-aliasing", "source-identification" }; static const char * const ualink_accel_state_values[] = { - "unconfigured", "configured", "ready", "active", "error" + [AMDGPU_UALINK_ACCEL_STATE_UNCONFIGURED] = "unconfigured", + [AMDGPU_UALINK_ACCEL_STATE_PPOD_CONFIGURED] = "ppod_configured", + [AMDGPU_UALINK_ACCEL_STATE_VPOD_CONFIGURED] = "vpod_configured", + [AMDGPU_UALINK_ACCEL_STATE_READY] = "ready", + [AMDGPU_UALINK_ACCEL_STATE_ACTIVE] = "active", + [AMDGPU_UALINK_ACCEL_STATE_ERROR] = "error", }; +static_assert(ARRAY_SIZE(ualink_accel_state_values) == + AMDGPU_UALINK_ACCEL_STATE_MAX); UALINK_ENUM_SHOW(info, link_type, link_type); UALINK_VALUE_SHOW(info, accel_id, ppod.accel_id, "%u"); @@ -491,7 +604,7 @@ check_ppod_state(struct amdgpu_device *adev, setup->ppod.accel_id, setup->ppod.size); return AMDGPU_UALINK_ACCEL_STATE_UNCONFIGURED; } - return AMDGPU_UALINK_ACCEL_STATE_CONFIGURED; + return AMDGPU_UALINK_ACCEL_STATE_PPOD_CONFIGURED; } static ssize_t ualink_ppod_setup_commit_store(struct kobject *kobj, @@ -512,7 +625,7 @@ static ssize_t ualink_ppod_setup_commit_store(struct kobject *kobj, setup); if (r) return r; - r = psp_ual_query_info(&adev->psp, adev->ualink.psp_if_ver, info); + r = psp_ual_query_info(&adev->psp, adev->ualink.psp_if_ver, info, NULL); if (r) return r; @@ -581,39 +694,11 @@ UALINK_ENUM_STORE(vpod_config, addr_mode, vpod.addr_mode); static bool check_vpod_info(struct amdgpu_device *adev, const struct amdgpu_ualink_info *info) { - unsigned int weight; - - if (info->accel_state < AMDGPU_UALINK_ACCEL_STATE_CONFIGURED) { + if (info->accel_state < AMDGPU_UALINK_ACCEL_STATE_PPOD_CONFIGURED) { dev_dbg(adev->dev, "pPod is not yet configured\n"); return false; } - if (info->vpod.id >= AMDGPU_UALINK_ACCEL_MAX) { - dev_dbg(adev->dev, "vPod ID %u out of range [0..%u]\n", - info->vpod.id, AMDGPU_UALINK_ACCEL_MAX - 1); - return false; - } - if (info->vpod.size == 0 || info->vpod.size > info->ppod.size) { - dev_dbg(adev->dev, "vPod size %u out of range [1..%u]\n", - info->vpod.size, info->ppod.size); - return false; - } - if (info->vpod.addr_mode >= AMDGPU_UALINK_ADDR_MODE_MAX) { - dev_dbg(adev->dev, "Invalid addr mode %u\n", info->vpod.id); - return false; - } - weight = bitmap_weight(info->vpod.active_accel_bits, AMDGPU_UALINK_ACCEL_MAX); - if (weight != info->vpod.size) { - dev_dbg(adev->dev, "vPod size doesn't match vpod_active_accels list: %u != %u\n", - info->vpod.size, weight); - return false; - } - if (!test_bit(info->ppod.accel_id, info->vpod.active_accel_bits)) { - dev_dbg(adev->dev, "Accelerator ID %u not listed in vpod_active_accels\n", - info->ppod.accel_id); - return false; - } - - return true; + return __check_vpod_info(adev, info); } static struct amdgpu_device *find_peer_adev(unsigned int accel_id) @@ -761,7 +846,7 @@ static void activate_accelerator(struct amdgpu_device *adev) { int r; - if (adev->ualink.info->accel_state >= AMDGPU_UALINK_ACCEL_STATE_READY) + if (adev->ualink.info->accel_state >= AMDGPU_UALINK_ACCEL_STATE_ACTIVE) return; /* Enable incoming NPA address translation with NPA VMID */ @@ -785,19 +870,20 @@ static void activate_accelerator(struct amdgpu_device *adev) return; } - adev->ualink.info->accel_state = AMDGPU_UALINK_ACCEL_STATE_READY; + adev->ualink.info->accel_state = AMDGPU_UALINK_ACCEL_STATE_ACTIVE; } static void deactivate_accelerator(struct amdgpu_device *adev) { - if (adev->ualink.info->accel_state < AMDGPU_UALINK_ACCEL_STATE_READY) + if (adev->ualink.info->accel_state < AMDGPU_UALINK_ACCEL_STATE_ACTIVE) return; /* Disable incoming NPA address translation with NPA VMID */ psp_ual_set_npa_config(&adev->psp, adev->ualink.psp_if_ver, adev->vm_manager.npa_vmid, false); /* ignore return value */ - adev->ualink.info->accel_state = AMDGPU_UALINK_ACCEL_STATE_CONFIGURED; + adev->ualink.info->accel_state = + AMDGPU_UALINK_ACCEL_STATE_PPOD_CONFIGURED; amdgpu_ualink_sw_fini(adev); amdgpu_ualink_manager_stop(adev); @@ -822,6 +908,138 @@ static void activate_local_vpod(struct amdgpu_device *adev) } } +static inline bool __is_vpod_peer(struct amdgpu_ualink_info *info, + struct amdgpu_ualink_info *peer_info) +{ + return peer_info->vpod.id == info->vpod.id && + uuid_equal(&peer_info->ppod.id, &info->ppod.id); +} + +static int __check_local_vpod_integrity(struct amdgpu_device *adev) +{ + DECLARE_BITMAP(local_accel_ids, AMDGPU_UALINK_ACCEL_MAX); + struct amdgpu_ualink_info *info = adev->ualink.info; + u32 local_accels[AMDGPU_UALINK_LOCAL_ACCELS_MAX]; + struct amdgpu_ualink_info *peer_info; + struct amdgpu_device *peer_adev; + unsigned int i, n_local_accels; + unsigned int accel_id; + /* Check that all local accelerators listed in vpod_active_accels have + * matching pod IDs + */ + bitmap_zero(local_accel_ids, AMDGPU_UALINK_ACCEL_MAX); + n_local_accels = 0; + __set_bit(info->ppod.accel_id, local_accel_ids); + local_accels[n_local_accels++] = info->ppod.accel_id; + + for (i = 0; i < mgpu_info.num_gpu; i++) { + peer_adev = mgpu_info.gpu_ins[i].adev; + if (peer_adev == adev || !peer_adev->ualink.info) + continue; + + peer_info = peer_adev->ualink.info; + /* peer device ppod not configured */ + if (peer_info->accel_state < + AMDGPU_UALINK_ACCEL_STATE_PPOD_CONFIGURED) + return -EAGAIN; + + accel_id = peer_info->ppod.accel_id; + if (!test_bit(accel_id, info->vpod.active_accel_bits)) + continue; + /* peer device vpod not configured */ + if (peer_info->accel_state < + AMDGPU_UALINK_ACCEL_STATE_VPOD_CONFIGURED) + return -EAGAIN; + + if (!uuid_equal(&peer_info->ppod.id, &info->ppod.id)) { + dev_dbg(adev->dev, + "Peer %u ppod_id doesn't match: %pU != %pU", + peer_info->ppod.accel_id, &peer_info->ppod.id, + &info->ppod.id); + return -EINVAL; + } + + if (peer_info->ppod.size != info->ppod.size) { + dev_dbg(adev->dev, + "Peer %u ppod_size doesn't match: %u != %u\n", + accel_id, peer_info->ppod.size, + info->ppod.size); + return -EINVAL; + } + + if (peer_info->vpod.id != info->vpod.id) { + dev_dbg(adev->dev, + "Peer %u vpod_id doesn't match: %u != %u", + accel_id, peer_info->vpod.id, info->vpod.id); + return -EINVAL; + } + if (peer_info->vpod.size != info->vpod.size) { + dev_dbg(adev->dev, + "Peer %u vpod_size doesn't match: %u != %u\n", + accel_id, peer_info->vpod.size, + info->vpod.size); + return -EINVAL; + } + if (peer_info->vpod.addr_mode != info->vpod.addr_mode) { + dev_dbg(adev->dev, + "Peer %u addr_mode doesn't match: %u != %u\n", + accel_id, peer_info->vpod.addr_mode, + info->vpod.addr_mode); + return -EINVAL; + } + if (!bitmap_equal(peer_info->vpod.active_accel_bits, + info->vpod.active_accel_bits, + AMDGPU_UALINK_ACCEL_MAX)) { + dev_dbg(adev->dev, + "Peer %u vpod_active_accels don't match\n", + accel_id); + return -EINVAL; + } + + if (__test_and_set_bit(accel_id, local_accel_ids)) { + dev_dbg(adev->dev, + "Duplicate accel_id %u among local vpod peers\n", + accel_id); + return -EINVAL; + } + local_accels[n_local_accels++] = accel_id; + } + + for (i = 0; i < mgpu_info.num_gpu; i++) { + peer_adev = mgpu_info.gpu_ins[i].adev; + peer_info = peer_adev->ualink.info; + + if (!peer_info) + continue; + if (peer_adev != adev && !__is_vpod_peer(info, peer_info)) + continue; + + peer_info->n_local_accels = n_local_accels; + memcpy(peer_info->local_accels, local_accels, + sizeof(local_accels)); + } + + return 0; +} + +static void amdgpu_ualink_activate_vpod(struct amdgpu_device *adev) +{ + int ret; + + if (adev->ualink.info->accel_state < AMDGPU_UALINK_ACCEL_STATE_READY) + return; + mutex_lock(&mgpu_info.mutex); + ret = __check_local_vpod_integrity(adev); + if (ret && ret != -EAGAIN) { + dev_err(adev->dev, "Local vpod integrity check failed: %d\n", + ret); + return; + } + if (!ret) + activate_local_vpod(adev); + mutex_unlock(&mgpu_info.mutex); +} + static ssize_t ualink_vpod_config_commit_store(struct kobject *kobj, struct kobj_attribute *attr, const char *buf, size_t count) @@ -835,7 +1053,7 @@ static ssize_t ualink_vpod_config_commit_store(struct kobject *kobj, if (!sysfs_streq(buf, "true")) return -EINVAL; - if (info->accel_state < AMDGPU_UALINK_ACCEL_STATE_CONFIGURED) { + if (info->accel_state < AMDGPU_UALINK_ACCEL_STATE_PPOD_CONFIGURED) { dev_dbg(adev->dev, "Ualink ppod is not yet configured\n"); return -EINVAL; } @@ -844,7 +1062,7 @@ static ssize_t ualink_vpod_config_commit_store(struct kobject *kobj, config); if (r) return r; - r = psp_ual_query_info(&adev->psp, adev->ualink.psp_if_ver, info); + r = psp_ual_query_info(&adev->psp, adev->ualink.psp_if_ver, info, NULL); if (r) return r; @@ -858,7 +1076,7 @@ static ssize_t ualink_vpod_config_commit_store(struct kobject *kobj, mutex_lock(&mgpu_info.mutex); if (check_local_vpod_integrity(adev)) activate_local_vpod(adev); - else if (info->accel_state >= AMDGPU_UALINK_ACCEL_STATE_CONFIGURED) + else if (info->accel_state >= AMDGPU_UALINK_ACCEL_STATE_PPOD_CONFIGURED) deactivate_accelerator(adev); mutex_unlock(&mgpu_info.mutex); @@ -988,7 +1206,7 @@ static ssize_t ualink_station_config_commit_store(struct kobject *kobj, r = psp_ual_set_station_config(&adev->psp, adev->ualink.psp_if_ver, stations); if (r) return r; - r = psp_ual_query_info(&adev->psp, adev->ualink.psp_if_ver, info); + r = psp_ual_query_info(&adev->psp, adev->ualink.psp_if_ver, info, NULL); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.h index d2348511e61af..18096bdf1272a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ualink.h @@ -103,7 +103,8 @@ enum amdgpu_ualink_type { enum amdgpu_ualink_accel_state { AMDGPU_UALINK_ACCEL_STATE_UNCONFIGURED = 0, - AMDGPU_UALINK_ACCEL_STATE_CONFIGURED, + AMDGPU_UALINK_ACCEL_STATE_PPOD_CONFIGURED, + AMDGPU_UALINK_ACCEL_STATE_VPOD_CONFIGURED, AMDGPU_UALINK_ACCEL_STATE_READY, AMDGPU_UALINK_ACCEL_STATE_ACTIVE, AMDGPU_UALINK_ACCEL_STATE_ERROR, -- 2.55.0