From patchwork Mon Feb 21 16:45:52 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Paneer Selvam, Arunpravin" X-Patchwork-Id: 12753958 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 D0070C433F5 for ; Mon, 21 Feb 2022 16:46:46 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 5391A10E78D; Mon, 21 Feb 2022 16:46:32 +0000 (UTC) Received: from NAM04-DM6-obe.outbound.protection.outlook.com (mail-dm6nam08on2057.outbound.protection.outlook.com [40.107.102.57]) by gabe.freedesktop.org (Postfix) with ESMTPS id D602D10E78D; Mon, 21 Feb 2022 16:46:30 +0000 (UTC) ARC-Seal: i=1; a=rsa-sha256; s=arcselector9901; d=microsoft.com; cv=none; b=LypgaizGoA/501k06SXyUZnnAhANYE/cOz5xJ/yz7UStGa1EgKUORkBx1LEP4jqs7DQVKYolvyKpaDWEacztueLQtKTZ2QgtqoZig8oQ8HLHlxlGuMVE7p27vpuXbDv86i76RyM3VvsuJYvEWpiUCcZ/w4iwcnukQnd1Fh2TIFpUjRmeRf1bf6GOpaumDJoA117KaTHcgrTynO/7Mt50/dR/n1LhyEfj9DZGW8MwbHObsI3cN34CkJu1pUbLDmBbjzsXex4imwrfjuw3ByihiOKmWSimLAGuCIRbbQrLCG2pdDu0jXePYFKnCTB9DXs4QUSzuWcyzreNi2OyHEErUw== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=microsoft.com; s=arcselector9901; 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=X699jAmhyVAJVUoq/foZpwQqlchmsl0f3XocBkCJDtc=; b=oAUqaKGINCrKf6mBgXUEa3XhtssHSPZ0BclEKD5XnimxFB0bF+47b7cyRHsAG/78WBcr2zhfad+/+DnZ9DthQALiWMksDx7du/XvzRwjl3vsakjkcy9GxklWARvLqFDoP7g65//R0VnjFHOJplQ7vDeYCfshUsgwMy+Lyvgns4G7HCwt+iyEeDiXVGadWw/1lsJVL0xKaRwb4xMVhnAs/0tWhBAmCICMRpc5fTSsgAQThv8zFQoueqZE9RIfKGlumfn0FS4hTii48BRZlEatEFnTj1Pb0drVd9B5fYt0Hyzxb2aYvt0GiTUBLr89bT9GIfYV78fBZDONIGr2UjkNWA== 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 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=X699jAmhyVAJVUoq/foZpwQqlchmsl0f3XocBkCJDtc=; b=QjGfIep69kq9NAEubBO7TFyqUeUcw3sBFxcjL6ZLS3c7MBD5qCDBt/Ro+TXMwn74wKXfSu37UVUBH0nldT+sgyIhZ2C9KpmB/f2tMqqbwqYgR97TV/G9NPSMYK34lajfknCz8T3BohZet55xYhy2T99ndlTv+TxBnNl38OvkKXc= Received: from BN9PR03CA0478.namprd03.prod.outlook.com (2603:10b6:408:139::33) by BL0PR12MB2337.namprd12.prod.outlook.com (2603:10b6:207:45::29) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.4995.24; Mon, 21 Feb 2022 16:46:26 +0000 Received: from BN8NAM11FT008.eop-nam11.prod.protection.outlook.com (2603:10b6:408:139:cafe::40) by BN9PR03CA0478.outlook.office365.com (2603:10b6:408:139::33) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.4995.16 via Frontend Transport; Mon, 21 Feb 2022 16:46:26 +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=SATLEXMB04.amd.com; Received: from SATLEXMB04.amd.com (165.204.84.17) by BN8NAM11FT008.mail.protection.outlook.com (10.13.177.95) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.20.4995.20 via Frontend Transport; Mon, 21 Feb 2022 16:46:26 +0000 Received: from rtg-amd.amd.com (10.180.168.240) by SATLEXMB04.amd.com (10.181.40.145) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2375.18; Mon, 21 Feb 2022 10:46:22 -0600 From: Arunpravin To: , , Subject: [PATCH v13 5/5] drm/amdgpu: add drm buddy support to amdgpu Date: Mon, 21 Feb 2022 22:15:52 +0530 Message-ID: <20220221164552.2434-5-Arunpravin.PaneerSelvam@amd.com> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20220221164552.2434-1-Arunpravin.PaneerSelvam@amd.com> References: <20220221164552.2434-1-Arunpravin.PaneerSelvam@amd.com> MIME-Version: 1.0 X-Originating-IP: [10.180.168.240] X-ClientProxiedBy: SATLEXMB03.amd.com (10.181.40.144) To SATLEXMB04.amd.com (10.181.40.145) X-EOPAttributedMessage: 0 X-MS-PublicTrafficType: Email X-MS-Office365-Filtering-Correlation-Id: 7a6c39a3-5b3b-4148-93e2-08d9f559b3c3 X-MS-TrafficTypeDiagnostic: BL0PR12MB2337:EE_ X-Microsoft-Antispam-PRVS: X-MS-Exchange-SenderADCheck: 1 X-MS-Exchange-AntiSpam-Relay: 0 X-Microsoft-Antispam: BCL:0; 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:CAL; SFV:NSPM; H:SATLEXMB04.amd.com; PTR:InfoDomainNonexistent; CAT:NONE; SFS:(13230001)(4636009)(46966006)(40470700004)(36840700001)(5660300002)(8936002)(26005)(16526019)(186003)(426003)(336012)(81166007)(2616005)(356005)(70586007)(82310400004)(1076003)(8676002)(4326008)(30864003)(6666004)(316002)(47076005)(70206006)(36756003)(110136005)(54906003)(83380400001)(36860700001)(86362001)(40460700003)(2906002)(508600001)(7696005)(36900700001); DIR:OUT; SFP:1101; X-OriginatorOrg: amd.com X-MS-Exchange-CrossTenant-OriginalArrivalTime: 21 Feb 2022 16:46:26.5688 (UTC) X-MS-Exchange-CrossTenant-Network-Message-Id: 7a6c39a3-5b3b-4148-93e2-08d9f559b3c3 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=[SATLEXMB04.amd.com] X-MS-Exchange-CrossTenant-AuthSource: BN8NAM11FT008.eop-nam11.prod.protection.outlook.com X-MS-Exchange-CrossTenant-AuthAs: Anonymous X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem X-MS-Exchange-Transport-CrossTenantHeadersStamped: BL0PR12MB2337 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Arunpravin , matthew.auld@intel.com, tzimmermann@suse.de, alexander.deucher@amd.com, christian.koenig@amd.com Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" - Remove drm_mm references and replace with drm buddy functionalities - Add res cursor support for drm buddy v2(Matthew Auld): - replace spinlock with mutex as we call kmem_cache_zalloc (..., GFP_KERNEL) in drm_buddy_alloc() function - lock drm_buddy_block_trim() function as it calls mark_free/mark_split are all globally visible v3(Matthew Auld): - remove trim method error handling as we address the failure case at drm_buddy_block_trim() function v4: - fix warnings reported by kernel test robot v5: - fix merge conflict issue v6: - fix warnings reported by kernel test robot v7: - remove DRM_BUDDY_RANGE_ALLOCATION flag usage v8: - keep DRM_BUDDY_RANGE_ALLOCATION flag usage - resolve conflicts created by drm/amdgpu: remove VRAM accounting v2 Signed-off-by: Arunpravin --- drivers/gpu/drm/Kconfig | 1 + .../gpu/drm/amd/amdgpu/amdgpu_res_cursor.h | 97 +++++-- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h | 7 +- drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c | 256 ++++++++++-------- 4 files changed, 229 insertions(+), 132 deletions(-) diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig index 763355330b17..019ec0440ced 100644 --- a/drivers/gpu/drm/Kconfig +++ b/drivers/gpu/drm/Kconfig @@ -279,6 +279,7 @@ config DRM_AMDGPU select HWMON select BACKLIGHT_CLASS_DEVICE select INTERVAL_TREE + select DRM_BUDDY help Choose this option if you have a recent AMD Radeon graphics card. diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_res_cursor.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_res_cursor.h index acfa207cf970..da12b4ff2e45 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_res_cursor.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_res_cursor.h @@ -30,12 +30,15 @@ #include #include +#include "amdgpu_vram_mgr.h" + /* state back for walking over vram_mgr and gtt_mgr allocations */ struct amdgpu_res_cursor { uint64_t start; uint64_t size; uint64_t remaining; - struct drm_mm_node *node; + void *node; + uint32_t mem_type; }; /** @@ -52,27 +55,63 @@ static inline void amdgpu_res_first(struct ttm_resource *res, uint64_t start, uint64_t size, struct amdgpu_res_cursor *cur) { + struct drm_buddy_block *block; + struct list_head *head, *next; struct drm_mm_node *node; - if (!res || res->mem_type == TTM_PL_SYSTEM) { - cur->start = start; - cur->size = size; - cur->remaining = size; - cur->node = NULL; - WARN_ON(res && start + size > res->num_pages << PAGE_SHIFT); - return; - } + if (!res) + goto err_out; BUG_ON(start + size > res->num_pages << PAGE_SHIFT); - node = to_ttm_range_mgr_node(res)->mm_nodes; - while (start >= node->size << PAGE_SHIFT) - start -= node++->size << PAGE_SHIFT; + cur->mem_type = res->mem_type; + + switch (cur->mem_type) { + case TTM_PL_VRAM: + head = &to_amdgpu_vram_mgr_node(res)->blocks; + + block = list_first_entry_or_null(head, + struct drm_buddy_block, + link); + if (!block) + goto err_out; + + while (start >= amdgpu_node_size(block)) { + start -= amdgpu_node_size(block); + + next = block->link.next; + if (next != head) + block = list_entry(next, struct drm_buddy_block, link); + } + + cur->start = amdgpu_node_start(block) + start; + cur->size = min(amdgpu_node_size(block) - start, size); + cur->remaining = size; + cur->node = block; + break; + case TTM_PL_TT: + node = to_ttm_range_mgr_node(res)->mm_nodes; + while (start >= node->size << PAGE_SHIFT) + start -= node++->size << PAGE_SHIFT; + + cur->start = (node->start << PAGE_SHIFT) + start; + cur->size = min((node->size << PAGE_SHIFT) - start, size); + cur->remaining = size; + cur->node = node; + break; + default: + goto err_out; + } - cur->start = (node->start << PAGE_SHIFT) + start; - cur->size = min((node->size << PAGE_SHIFT) - start, size); + return; + +err_out: + cur->start = start; + cur->size = size; cur->remaining = size; - cur->node = node; + cur->node = NULL; + WARN_ON(res && start + size > res->num_pages << PAGE_SHIFT); + return; } /** @@ -85,7 +124,9 @@ static inline void amdgpu_res_first(struct ttm_resource *res, */ static inline void amdgpu_res_next(struct amdgpu_res_cursor *cur, uint64_t size) { - struct drm_mm_node *node = cur->node; + struct drm_buddy_block *block; + struct drm_mm_node *node; + struct list_head *next; BUG_ON(size > cur->remaining); @@ -99,9 +140,27 @@ static inline void amdgpu_res_next(struct amdgpu_res_cursor *cur, uint64_t size) return; } - cur->node = ++node; - cur->start = node->start << PAGE_SHIFT; - cur->size = min(node->size << PAGE_SHIFT, cur->remaining); + switch (cur->mem_type) { + case TTM_PL_VRAM: + block = cur->node; + + next = block->link.next; + block = list_entry(next, struct drm_buddy_block, link); + + cur->node = block; + cur->start = amdgpu_node_start(block); + cur->size = min(amdgpu_node_size(block), cur->remaining); + break; + case TTM_PL_TT: + node = cur->node; + + cur->node = ++node; + cur->start = node->start << PAGE_SHIFT; + cur->size = min(node->size << PAGE_SHIFT, cur->remaining); + break; + default: + return; + } } #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h index 9120ae80ef52..8acdb2fb3a23 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h @@ -26,6 +26,7 @@ #include #include +#include #include "amdgpu.h" #define AMDGPU_PL_GDS (TTM_PL_PRIV + 0) @@ -40,11 +41,13 @@ struct amdgpu_vram_mgr { struct ttm_resource_manager manager; - struct drm_mm mm; - spinlock_t lock; + struct drm_buddy mm; + /* protects access to buffer objects */ + struct mutex lock; struct list_head reservations_pending; struct list_head reserved_pages; atomic64_t vis_usage; + u64 default_page_size; }; struct amdgpu_gtt_mgr { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c index 0a7611648573..6b2a5879e985 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c @@ -32,8 +32,11 @@ #include "atom.h" struct amdgpu_vram_reservation { + u64 start; + u64 size; + unsigned long flags; + struct list_head block; struct list_head node; - struct drm_mm_node mm_node; }; static inline struct amdgpu_vram_mgr * @@ -194,10 +197,10 @@ const struct attribute_group amdgpu_vram_mgr_attr_group = { * Calculate how many bytes of the MM node are inside visible VRAM */ static u64 amdgpu_vram_mgr_vis_size(struct amdgpu_device *adev, - struct drm_mm_node *node) + struct drm_buddy_block *block) { - uint64_t start = node->start << PAGE_SHIFT; - uint64_t end = (node->size + node->start) << PAGE_SHIFT; + u64 start = amdgpu_node_start(block); + u64 end = start + amdgpu_node_size(block); if (start >= adev->gmc.visible_vram_size) return 0; @@ -218,9 +221,9 @@ u64 amdgpu_vram_mgr_bo_visible_size(struct amdgpu_bo *bo) { struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); struct ttm_resource *res = bo->tbo.resource; - unsigned pages = res->num_pages; - struct drm_mm_node *mm; - u64 usage; + struct amdgpu_vram_mgr_node *node = to_amdgpu_vram_mgr_node(res); + struct drm_buddy_block *block; + u64 usage = 0; if (amdgpu_gmc_vram_full_visible(&adev->gmc)) return amdgpu_bo_size(bo); @@ -228,9 +231,8 @@ u64 amdgpu_vram_mgr_bo_visible_size(struct amdgpu_bo *bo) if (res->start >= adev->gmc.visible_vram_size >> PAGE_SHIFT) return 0; - mm = &container_of(res, struct ttm_range_mgr_node, base)->mm_nodes[0]; - for (usage = 0; pages; pages -= mm->size, mm++) - usage += amdgpu_vram_mgr_vis_size(adev, mm); + list_for_each_entry(block, &node->blocks, link) + usage += amdgpu_vram_mgr_vis_size(adev, block); return usage; } @@ -240,23 +242,31 @@ static void amdgpu_vram_mgr_do_reserve(struct ttm_resource_manager *man) { struct amdgpu_vram_mgr *mgr = to_vram_mgr(man); struct amdgpu_device *adev = to_amdgpu_device(mgr); - struct drm_mm *mm = &mgr->mm; + struct drm_buddy *mm = &mgr->mm; struct amdgpu_vram_reservation *rsv, *temp; + struct drm_buddy_block *block; uint64_t vis_usage; list_for_each_entry_safe(rsv, temp, &mgr->reservations_pending, node) { - if (drm_mm_reserve_node(mm, &rsv->mm_node)) + if (drm_buddy_alloc_blocks(mm, rsv->start, rsv->start + rsv->size, + rsv->size, mm->chunk_size, &rsv->block, + rsv->flags)) continue; - dev_dbg(adev->dev, "Reservation 0x%llx - %lld, Succeeded\n", - rsv->mm_node.start, rsv->mm_node.size); - - vis_usage = amdgpu_vram_mgr_vis_size(adev, &rsv->mm_node); - atomic64_add(vis_usage, &mgr->vis_usage); - spin_lock(&man->bdev->lru_lock); - man->usage += rsv->mm_node.size << PAGE_SHIFT; - spin_unlock(&man->bdev->lru_lock); - list_move(&rsv->node, &mgr->reserved_pages); + block = list_first_entry_or_null(&rsv->block, + struct drm_buddy_block, + link); + if (block) { + dev_dbg(adev->dev, "Reservation 0x%llx - %lld, Succeeded\n", + rsv->start, rsv->size); + + vis_usage = amdgpu_vram_mgr_vis_size(adev, block); + atomic64_add(vis_usage, &mgr->vis_usage); + spin_lock(&man->bdev->lru_lock); + man->usage += rsv->size; + spin_unlock(&man->bdev->lru_lock); + list_move(&rsv->node, &mgr->reserved_pages); + } } } @@ -279,13 +289,16 @@ int amdgpu_vram_mgr_reserve_range(struct amdgpu_vram_mgr *mgr, return -ENOMEM; INIT_LIST_HEAD(&rsv->node); - rsv->mm_node.start = start >> PAGE_SHIFT; - rsv->mm_node.size = size >> PAGE_SHIFT; + INIT_LIST_HEAD(&rsv->block); + + rsv->start = start; + rsv->size = size; + rsv->flags |= DRM_BUDDY_RANGE_ALLOCATION; - spin_lock(&mgr->lock); + mutex_lock(&mgr->lock); list_add_tail(&rsv->node, &mgr->reservations_pending); amdgpu_vram_mgr_do_reserve(&mgr->manager); - spin_unlock(&mgr->lock); + mutex_unlock(&mgr->lock); return 0; } @@ -307,19 +320,19 @@ int amdgpu_vram_mgr_query_page_status(struct amdgpu_vram_mgr *mgr, struct amdgpu_vram_reservation *rsv; int ret; - spin_lock(&mgr->lock); + mutex_lock(&mgr->lock); list_for_each_entry(rsv, &mgr->reservations_pending, node) { - if ((rsv->mm_node.start <= start) && - (start < (rsv->mm_node.start + rsv->mm_node.size))) { + if (rsv->start <= start && + (start < (rsv->start + rsv->size))) { ret = -EBUSY; goto out; } } list_for_each_entry(rsv, &mgr->reserved_pages, node) { - if ((rsv->mm_node.start <= start) && - (start < (rsv->mm_node.start + rsv->mm_node.size))) { + if (rsv->start <= start && + (start < (rsv->start + rsv->size))) { ret = 0; goto out; } @@ -327,32 +340,10 @@ int amdgpu_vram_mgr_query_page_status(struct amdgpu_vram_mgr *mgr, ret = -ENOENT; out: - spin_unlock(&mgr->lock); + mutex_unlock(&mgr->lock); return ret; } -/** - * amdgpu_vram_mgr_virt_start - update virtual start address - * - * @mem: ttm_resource to update - * @node: just allocated node - * - * Calculate a virtual BO start address to easily check if everything is CPU - * accessible. - */ -static void amdgpu_vram_mgr_virt_start(struct ttm_resource *mem, - struct drm_mm_node *node) -{ - unsigned long start; - - start = node->start + node->size; - if (start > mem->num_pages) - start -= mem->num_pages; - else - start = 0; - mem->start = max(mem->start, start); -} - /** * amdgpu_vram_mgr_new - allocate new ranges * @@ -368,13 +359,13 @@ static int amdgpu_vram_mgr_new(struct ttm_resource_manager *man, const struct ttm_place *place, struct ttm_resource **res) { - unsigned long lpfn, num_nodes, pages_per_node, pages_left, pages; + unsigned long lpfn, pages_per_node, pages_left, pages, n_pages; + u64 vis_usage = 0, mem_bytes, max_bytes, min_page_size; struct amdgpu_vram_mgr *mgr = to_vram_mgr(man); struct amdgpu_device *adev = to_amdgpu_device(mgr); - uint64_t vis_usage = 0, mem_bytes, max_bytes; - struct ttm_range_mgr_node *node; - struct drm_mm *mm = &mgr->mm; - enum drm_mm_insert_mode mode; + struct amdgpu_vram_mgr_node *node; + struct drm_buddy *mm = &mgr->mm; + struct drm_buddy_block *block; unsigned i; int r; @@ -387,10 +378,9 @@ static int amdgpu_vram_mgr_new(struct ttm_resource_manager *man, max_bytes -= AMDGPU_VM_RESERVED_VRAM; mem_bytes = tbo->base.size; - if (place->flags & TTM_PL_FLAG_CONTIGUOUS) { + if (place->flags & TTM_PL_FLAG_CONTIGUOUS) pages_per_node = ~0ul; - num_nodes = 1; - } else { + else { #ifdef CONFIG_TRANSPARENT_HUGEPAGE pages_per_node = HPAGE_PMD_NR; #else @@ -399,11 +389,9 @@ static int amdgpu_vram_mgr_new(struct ttm_resource_manager *man, #endif pages_per_node = max_t(uint32_t, pages_per_node, tbo->page_alignment); - num_nodes = DIV_ROUND_UP_ULL(PFN_UP(mem_bytes), pages_per_node); } - node = kvmalloc(struct_size(node, mm_nodes, num_nodes), - GFP_KERNEL | __GFP_ZERO); + node = kzalloc(sizeof(*node), GFP_KERNEL); if (!node) return -ENOMEM; @@ -415,9 +403,17 @@ static int amdgpu_vram_mgr_new(struct ttm_resource_manager *man, goto error_fini; } - mode = DRM_MM_INSERT_BEST; + INIT_LIST_HEAD(&node->blocks); + if (place->flags & TTM_PL_FLAG_TOPDOWN) - mode = DRM_MM_INSERT_HIGH; + node->flags |= DRM_BUDDY_TOPDOWN_ALLOCATION; + + if (place->fpfn || lpfn != man->size >> PAGE_SHIFT) + /* Allocate blocks in desired range */ + node->flags |= DRM_BUDDY_RANGE_ALLOCATION; + + min_page_size = mgr->default_page_size; + BUG_ON(min_page_size < mm->chunk_size); pages_left = node->base.num_pages; @@ -425,36 +421,61 @@ static int amdgpu_vram_mgr_new(struct ttm_resource_manager *man, pages = min(pages_left, 2UL << (30 - PAGE_SHIFT)); i = 0; - spin_lock(&mgr->lock); while (pages_left) { - uint32_t alignment = tbo->page_alignment; - if (pages >= pages_per_node) - alignment = pages_per_node; - - r = drm_mm_insert_node_in_range(mm, &node->mm_nodes[i], pages, - alignment, 0, place->fpfn, - lpfn, mode); - if (unlikely(r)) { - if (pages > pages_per_node) { - if (is_power_of_2(pages)) - pages = pages / 2; - else - pages = rounddown_pow_of_two(pages); - continue; - } - goto error_free; + pages = pages_per_node; + + n_pages = pages; + + if (place->flags & TTM_PL_FLAG_CONTIGUOUS) { + n_pages = roundup_pow_of_two(n_pages); + min_page_size = (u64)n_pages << PAGE_SHIFT; + + if (n_pages > lpfn) + lpfn = n_pages; } - vis_usage += amdgpu_vram_mgr_vis_size(adev, &node->mm_nodes[i]); - amdgpu_vram_mgr_virt_start(&node->base, &node->mm_nodes[i]); + mutex_lock(&mgr->lock); + r = drm_buddy_alloc_blocks(mm, (u64)place->fpfn << PAGE_SHIFT, + (u64)lpfn << PAGE_SHIFT, + (u64)n_pages << PAGE_SHIFT, + min_page_size, + &node->blocks, + node->flags); + mutex_unlock(&mgr->lock); + if (unlikely(r)) + goto error_free_blocks; + pages_left -= pages; ++i; if (pages > pages_left) pages = pages_left; } - spin_unlock(&mgr->lock); + + /* Free unused pages for contiguous allocation */ + if (place->flags & TTM_PL_FLAG_CONTIGUOUS) { + u64 actual_size = (u64)node->base.num_pages << PAGE_SHIFT; + + mutex_lock(&mgr->lock); + drm_buddy_block_trim(mm, + actual_size, + &node->blocks); + mutex_unlock(&mgr->lock); + } + + list_for_each_entry(block, &node->blocks, link) + vis_usage += amdgpu_vram_mgr_vis_size(adev, block); + + block = list_first_entry_or_null(&node->blocks, + struct drm_buddy_block, + link); + if (!block) { + r = -EINVAL; + goto error_fini; + } + + node->base.start = amdgpu_node_start(block) >> PAGE_SHIFT; if (i == 1) node->base.placement |= TTM_PL_FLAG_CONTIGUOUS; @@ -468,13 +489,13 @@ static int amdgpu_vram_mgr_new(struct ttm_resource_manager *man, *res = &node->base; return 0; -error_free: - while (i--) - drm_mm_remove_node(&node->mm_nodes[i]); - spin_unlock(&mgr->lock); +error_free_blocks: + mutex_lock(&mgr->lock); + drm_buddy_free_list(mm, &node->blocks); + mutex_unlock(&mgr->lock); error_fini: ttm_resource_fini(man, &node->base); - kvfree(node); + kfree(node); return r; } @@ -490,27 +511,26 @@ static int amdgpu_vram_mgr_new(struct ttm_resource_manager *man, static void amdgpu_vram_mgr_del(struct ttm_resource_manager *man, struct ttm_resource *res) { - struct ttm_range_mgr_node *node = to_ttm_range_mgr_node(res); + struct amdgpu_vram_mgr_node *node = to_amdgpu_vram_mgr_node(res); struct amdgpu_vram_mgr *mgr = to_vram_mgr(man); struct amdgpu_device *adev = to_amdgpu_device(mgr); + struct drm_buddy *mm = &mgr->mm; + struct drm_buddy_block *block; uint64_t vis_usage = 0; - unsigned i, pages; - spin_lock(&mgr->lock); - for (i = 0, pages = res->num_pages; pages; - pages -= node->mm_nodes[i].size, ++i) { - struct drm_mm_node *mm = &node->mm_nodes[i]; + mutex_lock(&mgr->lock); + list_for_each_entry(block, &node->blocks, link) + vis_usage += amdgpu_vram_mgr_vis_size(adev, block); - drm_mm_remove_node(mm); - vis_usage += amdgpu_vram_mgr_vis_size(adev, mm); - } amdgpu_vram_mgr_do_reserve(man); - spin_unlock(&mgr->lock); + + drm_buddy_free_list(mm, &node->blocks); + mutex_unlock(&mgr->lock); atomic64_sub(vis_usage, &mgr->vis_usage); ttm_resource_fini(man, res); - kvfree(node); + kfree(node); } /** @@ -648,13 +668,22 @@ static void amdgpu_vram_mgr_debug(struct ttm_resource_manager *man, struct drm_printer *printer) { struct amdgpu_vram_mgr *mgr = to_vram_mgr(man); + struct drm_buddy *mm = &mgr->mm; + struct drm_buddy_block *block; drm_printf(printer, " vis usage:%llu\n", amdgpu_vram_mgr_vis_usage(mgr)); - spin_lock(&mgr->lock); - drm_mm_print(&mgr->mm, printer); - spin_unlock(&mgr->lock); + mutex_lock(&mgr->lock); + drm_printf(printer, "default_page_size: %lluKiB\n", + mgr->default_page_size >> 10); + + drm_buddy_print(mm, printer); + + drm_printf(printer, "reserved:\n"); + list_for_each_entry(block, &mgr->reserved_pages, link) + drm_buddy_block_print(mm, block, printer); + mutex_unlock(&mgr->lock); } static const struct ttm_resource_manager_func amdgpu_vram_mgr_func = { @@ -674,16 +703,21 @@ int amdgpu_vram_mgr_init(struct amdgpu_device *adev) { struct amdgpu_vram_mgr *mgr = &adev->mman.vram_mgr; struct ttm_resource_manager *man = &mgr->manager; + int err; ttm_resource_manager_init(man, &adev->mman.bdev, adev->gmc.real_vram_size); man->func = &amdgpu_vram_mgr_func; - drm_mm_init(&mgr->mm, 0, man->size >> PAGE_SHIFT); - spin_lock_init(&mgr->lock); + err = drm_buddy_init(&mgr->mm, man->size, PAGE_SIZE); + if (err) + return err; + + mutex_init(&mgr->lock); INIT_LIST_HEAD(&mgr->reservations_pending); INIT_LIST_HEAD(&mgr->reserved_pages); + mgr->default_page_size = PAGE_SIZE; ttm_set_driver_manager(&adev->mman.bdev, TTM_PL_VRAM, &mgr->manager); ttm_resource_manager_set_used(man, true); @@ -711,16 +745,16 @@ void amdgpu_vram_mgr_fini(struct amdgpu_device *adev) if (ret) return; - spin_lock(&mgr->lock); + mutex_lock(&mgr->lock); list_for_each_entry_safe(rsv, temp, &mgr->reservations_pending, node) kfree(rsv); list_for_each_entry_safe(rsv, temp, &mgr->reserved_pages, node) { - drm_mm_remove_node(&rsv->mm_node); + drm_buddy_free_list(&mgr->mm, &rsv->block); kfree(rsv); } - drm_mm_takedown(&mgr->mm); - spin_unlock(&mgr->lock); + drm_buddy_fini(&mgr->mm); + mutex_unlock(&mgr->lock); ttm_resource_manager_cleanup(man); ttm_set_driver_manager(&adev->mman.bdev, TTM_PL_VRAM, NULL);