From nobody Mon Feb 9 17:59:23 2026 Delivered-To: importer@patchew.org Authentication-Results: mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of lists.xenproject.org designates 192.237.175.120 as permitted sender) smtp.mailfrom=xen-devel-bounces@lists.xenproject.org; arc=pass (i=1 dmarc=pass fromdomain=amd.com); dmarc=pass(p=quarantine dis=none) header.from=amd.com ARC-Seal: i=2; a=rsa-sha256; t=1685666980; cv=pass; d=zohomail.com; s=zohoarc; b=e/fi+VdyJYheVtlgNmRTjQLgyMKThs7/e8GrRSavYUwq7mX300azxi1htZtglIHYb9FTOqWCmPRixgN0JxcjVSgf3NK45aqYDXez05mpgH48w3dPaEGWzQOPNEshsUj3MnDla9HmEumIqyjBHUdEJak83wzH5t5Dt6QA00X3kUw= ARC-Message-Signature: i=2; a=rsa-sha256; c=relaxed/relaxed; d=zohomail.com; s=zohoarc; t=1685666980; h=Content-Type:Cc:Date:From:In-Reply-To:List-Subscribe:List-Post:List-Id:List-Help:List-Unsubscribe:MIME-Version:Message-ID:References:Sender:Subject:To; bh=d1CA76xInbIr4hE8UWeQAMGW5FdBn6s0KkNux42XsEg=; b=OodgJCt3U13Je0gnWTABGSJLXkIPoRaCVS/bdVfegcfQSpiHhX9HH3AAKanV5jPrumzHHG8aoc50nCc/NdVBV87IZW8nMU8c8IPivq6de1WzE+ve2/+rH/qXsQaZ1bCibv6FjHfN2Vw72eddAgj+E+g/KNnG5OkaopgIiDBxeJ0= ARC-Authentication-Results: i=2; mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of lists.xenproject.org designates 192.237.175.120 as permitted sender) smtp.mailfrom=xen-devel-bounces@lists.xenproject.org; arc=pass (i=1 dmarc=pass fromdomain=amd.com); dmarc=pass header.from= (p=quarantine dis=none) Return-Path: Received: from lists.xenproject.org (lists.xenproject.org [192.237.175.120]) by mx.zohomail.com with SMTPS id 1685666980540894.8565982307343; Thu, 1 Jun 2023 17:49:40 -0700 (PDT) Received: from list by lists.xenproject.org with outflank-mailman.542802.847192 (Exim 4.92) (envelope-from ) id 1q4sye-0004WK-K4; Fri, 02 Jun 2023 00:49:12 +0000 Received: by outflank-mailman (output) from mailman id 542802.847192; Fri, 02 Jun 2023 00:49:12 +0000 Received: from localhost ([127.0.0.1] helo=lists.xenproject.org) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1q4syd-0004Q4-73; Fri, 02 Jun 2023 00:49:11 +0000 Received: by outflank-mailman (input) for mailman id 542802; Fri, 02 Jun 2023 00:49:06 +0000 Received: from se1-gles-sth1-in.inumbo.com ([159.253.27.254] helo=se1-gles-sth1.inumbo.com) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1q4syX-0000if-KF for xen-devel@lists.xenproject.org; Fri, 02 Jun 2023 00:49:05 +0000 Received: from NAM10-DM6-obe.outbound.protection.outlook.com (mail-dm6nam10on2062b.outbound.protection.outlook.com [2a01:111:f400:7e88::62b]) by se1-gles-sth1.inumbo.com (Halon) with ESMTPS id 44b714f6-00df-11ee-b231-6b7b168915f2; Fri, 02 Jun 2023 02:49:03 +0200 (CEST) Received: from DM6PR06CA0076.namprd06.prod.outlook.com (2603:10b6:5:336::9) by DS0PR12MB8815.namprd12.prod.outlook.com (2603:10b6:8:14f::16) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.6455.23; Fri, 2 Jun 2023 00:49:00 +0000 Received: from DM6NAM11FT028.eop-nam11.prod.protection.outlook.com (2603:10b6:5:336:cafe::c9) by DM6PR06CA0076.outlook.office365.com (2603:10b6:5:336::9) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.6455.24 via Frontend Transport; Fri, 2 Jun 2023 00:49:00 +0000 Received: from SATLEXMB04.amd.com (165.204.84.17) by DM6NAM11FT028.mail.protection.outlook.com (10.13.173.140) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.20.6455.24 via Frontend Transport; Fri, 2 Jun 2023 00:49:00 +0000 Received: from SATLEXMB08.amd.com (10.181.40.132) 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 15.1.2375.34; Thu, 1 Jun 2023 19:48:59 -0500 Received: from SATLEXMB04.amd.com (10.181.40.145) by SATLEXMB08.amd.com (10.181.40.132) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2375.34; Thu, 1 Jun 2023 17:48:59 -0700 Received: from xsjfnuv50.xilinx.com (10.180.168.240) by SATLEXMB04.amd.com (10.181.40.145) with Microsoft SMTP Server id 15.1.2375.34 via Frontend Transport; Thu, 1 Jun 2023 19:48:58 -0500 X-Outflank-Mailman: Message body and most headers restored to incoming version X-BeenThere: xen-devel@lists.xenproject.org List-Id: Xen developer discussion List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Errors-To: xen-devel-bounces@lists.xenproject.org Precedence: list Sender: "Xen-devel" X-Inumbo-ID: 44b714f6-00df-11ee-b231-6b7b168915f2 ARC-Seal: i=1; a=rsa-sha256; s=arcselector9901; d=microsoft.com; cv=none; b=LBt9/TAmfGby1GHLGSdfMnmlFxbUXILygyrDIj0wSdeTaVLu9jLRp+G3aC8vqq5TOE8Cmtc8e1kyS6nUpygVUbtQs8bDMaTR9ifH1VMZwuDBxihlKrPZtjmauAMohmCILPhin5IFAuUqmyy8J57tkmMyr/7imirS1FEBYvW7EMipERkIFblmKfb7zpRVeWM+VHSAFYTUy6tvnv/zFCCgIF2NuUatjwrTM63YOVPyeSlafzhbtSUQNmQ9Mdzm36pdQ8qS2BghnmcptOGVoCDmeuUAN+yPCPws77apGLzhKRSw3aztRsHtPcYAYJWgnsJsF0P6GWO0Tu7rUo5IbCCL1g== 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=d1CA76xInbIr4hE8UWeQAMGW5FdBn6s0KkNux42XsEg=; b=UMIKE3Xr1I8kljc6mwrFLKsZJ+yE77EPkRIyCAv/SshEmzIlXR/J7NcKCgNTIs0IefJRc2J820G3kvDwqSPECSUZh2NgqWBT92bCaXrvfMT5WFFQT4GhjBx01+lEYWyTIy8J47OD4p0Hl6oqYmFT2d/a2+G4atdevHmY8KiS6BKg+RKv/noE+MsUhCGc7ZsS5FWZG88j5uA0e0ys7gmkuDtOek/m5rwvo0DAK6hvwd0zc0RkqqD6SM7iVlf266hztWYvZgwBbCKMkXkdYA3F0Dc/au0zEPZ24lnka2wMfNocZFQieEuisOD7NDjH65HSU4U5tjTY/hjP/UIvaVSFbg== ARC-Authentication-Results: i=1; mx.microsoft.com 1; spf=pass (sender ip is 165.204.84.17) smtp.rcpttodomain=lists.xenproject.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=d1CA76xInbIr4hE8UWeQAMGW5FdBn6s0KkNux42XsEg=; b=5YZCiXEcIcc/skE2I50ZcFa3pZBPn7CnrdRetwiyhQSK4SK5yrMBVqwAjMwA0IuoCwbyR3AIETWfD5psAGS5BKX6+MggWSGyf8hWKTi2gZBmdfxq5cm/+AJxSE9l349ualBqe4fxSijbAjL9cfSxXZYlihGmEDCuSVMF4fVAsdY= 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 (zohomail.com: domain of lists.xenproject.org designates 192.237.175.120 as permitted sender) client-ip=192.237.175.120; envelope-from=xen-devel-bounces@lists.xenproject.org; helo=lists.xenproject.org; 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; pr=C From: Vikram Garhwal To: CC: , , , , Andrew Cooper , George Dunlap , Julien Grall , Wei Liu Subject: [XEN][PATCH v7 16/19] xen/arm: Implement device tree node addition functionalities Date: Thu, 1 Jun 2023 17:48:21 -0700 Message-ID: <20230602004824.20731-17-vikram.garhwal@amd.com> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20230602004824.20731-1-vikram.garhwal@amd.com> References: <20230602004824.20731-1-vikram.garhwal@amd.com> MIME-Version: 1.0 X-EOPAttributedMessage: 0 X-MS-PublicTrafficType: Email X-MS-TrafficTypeDiagnostic: DM6NAM11FT028:EE_|DS0PR12MB8815:EE_ X-MS-Office365-Filtering-Correlation-Id: 61ed6f1f-2a60-4eda-f4e0-08db63032788 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:(13230028)(4636009)(376002)(346002)(136003)(39860400002)(396003)(451199021)(36840700001)(40470700004)(46966006)(186003)(426003)(336012)(44832011)(2616005)(1076003)(5660300002)(83380400001)(26005)(8676002)(30864003)(47076005)(8936002)(2906002)(66899021)(36860700001)(40460700003)(4326008)(54906003)(40480700001)(6916009)(41300700001)(6666004)(356005)(36756003)(82310400005)(82740400003)(81166007)(316002)(478600001)(70586007)(86362001)(70206006)(36900700001);DIR:OUT;SFP:1101; X-OriginatorOrg: amd.com X-MS-Exchange-CrossTenant-OriginalArrivalTime: 02 Jun 2023 00:49:00.1614 (UTC) X-MS-Exchange-CrossTenant-Network-Message-Id: 61ed6f1f-2a60-4eda-f4e0-08db63032788 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: DM6NAM11FT028.eop-nam11.prod.protection.outlook.com X-MS-Exchange-CrossTenant-AuthAs: Anonymous X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem X-MS-Exchange-Transport-CrossTenantHeadersStamped: DS0PR12MB8815 X-ZohoMail-DKIM: pass (identity @amd.com) X-ZM-MESSAGEID: 1685666981300100007 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update sysctl XEN_SYSCTL_dt_overlay to enable support for dtbo nodes additi= on using device tree overlay. xl dt-overlay add file.dtbo: Each time overlay nodes are added using .dtbo, a new fdt(memcpy of device_tree_flattened) is created and updated with overlay nodes. This updated fdt is further unflattened to a dt_host_new. Next, it checks if= any of the overlay nodes already exists in the dt_host. If overlay nodes do= esn't exist then find the overlay nodes in dt_host_new, find the overlay node= 's parent in dt_host and add the nodes as child under their parent in the dt_host. The node is attached as the last node under target parent. Finally, add IRQs, add device to IOMMUs, set permissions and map MMIO f= or the overlay node. When a node is added using overlay, a new entry is allocated in the overlay_track to keep the track of memory allocation due to addition of ove= rlay node. This is helpful for freeing the memory allocated when a device tree n= ode is removed. The main purpose of this to address first part of dynamic programming i.e. making xen aware of new device tree node which means updating the dt_host w= ith overlay node information. Here we are adding/removing node from dt_host, and checking/setting IOMMU and IRQ permission but never mapping them to any dom= ain. Right now, mapping/Un-mapping will happen only when a new domU is created/destroyed using "xl create". Signed-off-by: Vikram Garhwal --- Changes from v6: Fix comment style and add comment regarding false flag in irq mapping. Move malloc for nodes_full_path to handle_add_overlay_nodes. Move node_num define to start of overlay_get_nodes_info(). Remove "domain *d" from handle_add_irq_iommu(). Fix error handling for handle_add_irq_iommu(). Split handle_add_overlay_nodes to two functions. Create a separate function for freeing nodes_full_path. Fix xfree for dt_sysctl. --- xen/common/dt-overlay.c | 533 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 533 insertions(+) diff --git a/xen/common/dt-overlay.c b/xen/common/dt-overlay.c index b2a7e441df..12b6b010ef 100644 --- a/xen/common/dt-overlay.c +++ b/xen/common/dt-overlay.c @@ -33,6 +33,25 @@ static struct dt_device_node * return child_node; } =20 +/* + * Returns next node to the input node. If node has children then return + * last descendant's next node. +*/ +static struct dt_device_node * +dt_find_next_node(struct dt_device_node *dt, const struct dt_device_node *= node) +{ + struct dt_device_node *np; + + dt_for_each_device_node(dt, np) + if ( np =3D=3D node ) + break; + + if ( np->child ) + np =3D find_last_descendants_node(np); + + return np->allnext; +} + static int dt_overlay_remove_node(struct dt_device_node *device_node) { struct dt_device_node *np; @@ -106,6 +125,78 @@ static int dt_overlay_remove_node(struct dt_device_nod= e *device_node) return 0; } =20 +static int dt_overlay_add_node(struct dt_device_node *device_node, + const char *parent_node_path) +{ + struct dt_device_node *parent_node; + struct dt_device_node *next_node; + + parent_node =3D dt_find_node_by_path(parent_node_path); + + if ( parent_node =3D=3D NULL ) + { + dt_dprintk("Parent node %s not found. Overlay node will not be add= ed\n", + parent_node_path); + return -EINVAL; + } + + /* If parent has no child. */ + if ( parent_node->child =3D=3D NULL ) + { + next_node =3D parent_node->allnext; + device_node->parent =3D parent_node; + parent_node->allnext =3D device_node; + parent_node->child =3D device_node; + } + else + { + struct dt_device_node *np; + /* + * If parent has at least one child node. + * Iterate to the last child node of parent. + */ + for ( np =3D parent_node->child; np->sibling !=3D NULL; np =3D np-= >sibling ); + + /* Iterate over all child nodes of np node. */ + if ( np->child ) + { + struct dt_device_node *np_last_descendant; + + np_last_descendant =3D find_last_descendants_node(np); + + next_node =3D np_last_descendant->allnext; + np_last_descendant->allnext =3D device_node; + } + else + { + next_node =3D np->allnext; + np->allnext =3D device_node; + } + + device_node->parent =3D parent_node; + np->sibling =3D device_node; + np->sibling->sibling =3D NULL; + } + + /* Iterate over all child nodes of device_node to add children too. */ + if ( device_node->child ) + { + struct dt_device_node *device_node_last_descendant; + + device_node_last_descendant =3D find_last_descendants_node(device_= node); + + /* Plug next_node at the end of last children of device_node. */ + device_node_last_descendant->allnext =3D next_node; + } + else + { + /* Now plug next_node at the end of device_node. */ + device_node->allnext =3D next_node; + } + + return 0; +} + /* Basic sanity check for the dtbo tool stack provided to Xen. */ static int check_overlay_fdt(const void *overlay_fdt, uint32_t overlay_fdt= _size) { @@ -145,6 +236,76 @@ static unsigned int overlay_node_count(const void *ove= rlay_fdt) return num_overlay_nodes; } =20 +/* + * overlay_get_nodes_info gets full name with path for all the nodes which + * are in one level of __overlay__ tag. This is useful when checking node = for + * duplication i.e. dtbo tries to add nodes which already exists in device= tree. + */ +static int overlay_get_nodes_info(const void *fdto, char **nodes_full_path) +{ + int fragment; + unsigned int node_num =3D 0; + + fdt_for_each_subnode(fragment, fdto, 0) + { + int target; + int overlay; + int subnode; + const char *target_path; + + target =3D fdt_overlay_target_offset(device_tree_flattened, fdto, + fragment, &target_path); + if ( target < 0 ) + return target; + + if ( target_path =3D=3D NULL ) + return -EINVAL; + + overlay =3D fdt_subnode_offset(fdto, fragment, "__overlay__"); + + /* + * overlay value can be < 0. But fdt_for_each_subnode() loop check= s for + * overlay >=3D 0. So, no need for a overlay>=3D0 check here. + */ + fdt_for_each_subnode(subnode, fdto, overlay) + { + const char *node_name =3D NULL; + int node_name_len; + unsigned int target_path_len =3D strlen(target_path); + unsigned int node_full_name_len; + + node_name =3D fdt_get_name(fdto, subnode, &node_name_len); + + if ( node_name =3D=3D NULL ) + return node_name_len; + + /* + * Magic number 2 is for adding '/' and '\0'. This is done to = keep + * the node_full_path in the correct full node name format. + */ + node_full_name_len =3D target_path_len + node_name_len + 2; + + nodes_full_path[node_num] =3D xmalloc_bytes(node_full_name_len= ); + + if ( nodes_full_path[node_num] =3D=3D NULL ) + return -ENOMEM; + + memcpy(nodes_full_path[node_num], target_path, target_path_len= ); + + nodes_full_path[node_num][target_path_len] =3D '/'; + + memcpy(nodes_full_path[node_num] + target_path_len + 1, + node_name, node_name_len); + + nodes_full_path[node_num][node_full_name_len - 1] =3D '\0'; + + node_num++; + } + } + + return 0; +} + static int handle_remove_irq_iommu(struct dt_device_node *device_node) { int rc =3D 0; @@ -369,6 +530,373 @@ out: return rc; } =20 +/* + * Handles IRQ and IOMMU mapping for the overlay_node and all descendants = of the + * overlay_node. + */ +static int handle_add_irq_iommu(struct dt_device_node *overlay_node) +{ + int rc; + unsigned int naddr, i, len; + struct dt_device_node *np; + + /* + * First let's handle the interrupts. + * For now, need_mapping is set to false which means it will only perm= it IRQ + * access to hardware_domain using irq_permit_access() but will never = route + * as route_irq_to_guest() will not be called with false flag. + */ + rc =3D handle_device_interrupts(hardware_domain, overlay_node, false); + if ( rc < 0 ) + { + printk(XENLOG_ERR "Failed to retrieve interrupts configuration\n"); + return rc; + } + + /* Check if iommu property exists. */ + if ( dt_get_property(overlay_node, "iommus", &len) ) + { + /* Add device to IOMMUs. */ + rc =3D iommu_add_dt_device(overlay_node); + if ( rc < 0 ) + { + printk(XENLOG_ERR "Failed to add %s to the IOMMU\n", + dt_node_full_name(overlay_node)); + return rc; + } + } + + /* Set permissions. */ + naddr =3D dt_number_of_address(overlay_node); + + dt_dprintk("%s naddr =3D %u\n", dt_node_full_name(overlay_node), naddr= ); + + /* Give permission to map MMIOs */ + for ( i =3D 0; i < naddr; i++ ) + { + uint64_t addr, size; + + /* + * For now, we skip_mapping which means it will only permit iomem = access + * to hardware_domain using iomem_permit_access() but will never m= ap as + * map_range_p2mt() will not be called. + */ + struct map_range_data mr_data =3D { .d =3D hardware_domain, + .p2mt =3D p2m_mmio_direct_c, + .skip_mapping =3D true + }; + + rc =3D dt_device_get_address(overlay_node, i, &addr, &size); + if ( rc ) + { + printk(XENLOG_ERR "Unable to retrieve address %u for %s\n", + i, dt_node_full_name(overlay_node)); + return rc; + } + + rc =3D map_range_to_domain(overlay_node, addr, size, &mr_data); + if ( rc ) + return rc; + } + + /* Map IRQ and IOMMU for overlay_node's children. */ + for ( np =3D overlay_node->child; np !=3D NULL; np =3D np->sibling ) + { + rc =3D handle_add_irq_iommu(np); + if ( rc ) + { + printk(XENLOG_ERR "Adding IRQ and IOMMU failed\n"); + return rc; + } + } + + return rc; +} + +static void free_nodes_full_path(int num_nodes, char **nodes_full_path) +{ + int i; + + if ( nodes_full_path !=3D NULL ) + { + for ( i =3D 0; i < num_nodes && nodes_full_path[i] !=3D NULL; + i++ ) + { + xfree(nodes_full_path[i]); + } + xfree(nodes_full_path); + } + + return; +} + +static long add_nodes(struct overlay_track *tr, char **nodes_full_path) + +{ + int j, rc; + struct dt_device_node *overlay_node; + + for ( j =3D 0; j < tr->num_nodes; j++ ) + { + struct dt_device_node *prev_node, *next_node; + + dt_dprintk("Adding node: %s\n", nodes_full_path[j]); + + /* Find the newly added node in tr->dt_host_new by it's full path.= */ + overlay_node =3D device_tree_find_node_by_path(tr->dt_host_new, + nodes_full_path[j]); + if ( overlay_node =3D=3D NULL ) + { + /* Sanity check. But code will never come here. */ + ASSERT_UNREACHABLE(); + return -EFAULT; + } + + /* + * Find previous and next node to overlay_node in dt_host_new. We = will + * need these nodes to fix the dt_host_new mapping. When overlay_n= ode is + * take out of dt_host_new tree and added to dt_host, link between + * previous node and next_node is broken. We will need to refresh + * dt_host_new with correct linking for any other overlay nodes + * extraction in future. + */ + dt_for_each_device_node(tr->dt_host_new, prev_node) + if ( prev_node->allnext =3D=3D overlay_node ) + break; + + next_node =3D dt_find_next_node(tr->dt_host_new, overlay_node); + + read_lock(&dt_host->lock); + + /* Add the node to dt_host. */ + rc =3D dt_overlay_add_node(overlay_node, overlay_node->parent->ful= l_name); + if ( rc ) + { + read_unlock(&dt_host->lock); + + /* Node not added in dt_host. */ + return rc; + } + + read_unlock(&dt_host->lock); + + prev_node->allnext =3D next_node; + + overlay_node =3D dt_find_node_by_path(overlay_node->full_name); + if ( overlay_node =3D=3D NULL ) + { + /* Sanity check. But code will never come here. */ + ASSERT_UNREACHABLE(); + return -EFAULT; + } + + rc =3D handle_add_irq_iommu(overlay_node); + if ( rc ) + { + printk(XENLOG_ERR "Adding IRQ and IOMMU failed\n"); + return rc; + } + + /* Keep overlay_node address in tracker. */ + tr->nodes_address[j] =3D (unsigned long)overlay_node; + } + + return 0; +} +/* + * Adds device tree nodes under target node. + * We use tr->dt_host_new to unflatten the updated device_tree_flattened. = This + * is done to avoid the removal of device_tree generation, iomem regions m= apping + * to hardware domain done by handle_node(). + */ +static long handle_add_overlay_nodes(void *overlay_fdt, + uint32_t overlay_fdt_size) +{ + int rc, j; + struct dt_device_node *overlay_node; + struct overlay_track *tr =3D NULL; + char **nodes_full_path =3D NULL; + unsigned int new_fdt_size; + + tr =3D xzalloc(struct overlay_track); + if ( tr =3D=3D NULL ) + return -ENOMEM; + + new_fdt_size =3D fdt_totalsize(device_tree_flattened) + + fdt_totalsize(overlay_fdt); + + tr->fdt =3D xzalloc_bytes(new_fdt_size); + if ( tr->fdt =3D=3D NULL ) + { + xfree(tr); + return -ENOMEM; + } + + tr->num_nodes =3D overlay_node_count(overlay_fdt); + if ( tr->num_nodes =3D=3D 0 ) + { + xfree(tr->fdt); + xfree(tr); + return -ENOMEM; + } + + tr->nodes_address =3D xzalloc_bytes(tr->num_nodes * sizeof(unsigned lo= ng)); + if ( tr->nodes_address =3D=3D NULL ) + { + xfree(tr->fdt); + xfree(tr); + return -ENOMEM; + } + + rc =3D check_overlay_fdt(overlay_fdt, overlay_fdt_size); + if ( rc ) + { + xfree(tr->nodes_address); + xfree(tr->fdt); + xfree(tr); + return rc; + } + + /* + * Keep a copy of overlay_fdt as fdt_overlay_apply will change the inp= ut + * overlay's content(magic) when applying overlay. + */ + tr->overlay_fdt =3D xzalloc_bytes(overlay_fdt_size); + if ( tr->overlay_fdt =3D=3D NULL ) + { + xfree(tr->nodes_address); + xfree(tr->fdt); + xfree(tr); + return -ENOMEM; + } + + memcpy(tr->overlay_fdt, overlay_fdt, overlay_fdt_size); + + spin_lock(&overlay_lock); + + memcpy(tr->fdt, device_tree_flattened, + fdt_totalsize(device_tree_flattened)); + + /* Open tr->fdt with more space to accommodate the overlay_fdt. */ + rc =3D fdt_open_into(tr->fdt, tr->fdt, new_fdt_size); + if ( rc ) + { + printk(XENLOG_ERR "Increasing fdt size to accommodate overlay_fdt = failed with error %d\n", + rc); + goto err; + } + + nodes_full_path =3D xzalloc_bytes(tr->num_nodes * sizeof(char *)); + if ( nodes_full_path =3D=3D NULL ) + { + rc =3D -ENOMEM; + goto err; + } + + /* + * overlay_get_nodes_info is called to get the node information from d= tbo. + * This is done before fdt_overlay_apply() because the overlay apply w= ill + * erase the magic of overlay_fdt. + */ + rc =3D overlay_get_nodes_info(overlay_fdt, nodes_full_path); + if ( rc ) + { + printk(XENLOG_ERR "Getting nodes information failed with error %d\= n", + rc); + goto err; + } + + rc =3D fdt_overlay_apply(tr->fdt, overlay_fdt); + if ( rc ) + { + printk(XENLOG_ERR "Adding overlay node failed with error %d\n", rc= ); + goto err; + } + + /* + * Check if any of the node already exists in dt_host. If node already= exits + * we can return here as this overlay_fdt is not suitable for overlay = ops. + */ + for ( j =3D 0; j < tr->num_nodes; j++ ) + { + overlay_node =3D dt_find_node_by_path(nodes_full_path[j]); + if ( overlay_node !=3D NULL ) + { + printk(XENLOG_ERR "node %s exists in device tree\n", + nodes_full_path[j]); + rc =3D -EINVAL; + goto err; + } + } + + /* Unflatten the tr->fdt into a new dt_host. */ + rc =3D unflatten_device_tree(tr->fdt, &tr->dt_host_new); + if ( rc ) + { + printk(XENLOG_ERR "unflatten_device_tree failed with error %d\n", = rc); + goto err; + } + + rc =3D add_nodes(tr, nodes_full_path); + if ( rc ) + { + printk(XENLOG_ERR "Adding nodes failed. Removing the partially add= ed nodes.\n"); + goto remove_node; + } + + INIT_LIST_HEAD(&tr->entry); + list_add_tail(&tr->entry, &overlay_tracker); + + spin_unlock(&overlay_lock); + + free_nodes_full_path(tr->num_nodes, nodes_full_path); + + return rc; + +/* + * Failure case. We need to remove the nodes, free tracker(if tr exists) a= nd + * tr->dt_host_new. + */ +remove_node: + tr->num_nodes =3D j; + rc =3D remove_nodes(tr); + + if ( rc ) + { + /* + * User needs to provide right overlay. Incorrect node information + * example parent node doesn't exist in dt_host etc can cause memo= ry + * leaks as removing_nodes() will fail and this means nodes memory= is + * not freed from tracker. Which may cause memory leaks. Ideally, = these + * device tree related mistakes will be caught by fdt_overlay_appl= y() + * but given that we don't manage that code keeping this warning m= essage + * is better here. + */ + printk(XENLOG_ERR "Removing node failed.\n"); + spin_unlock(&overlay_lock); + + free_nodes_full_path(tr->num_nodes, nodes_full_path); + + return rc; + } + +err: + spin_unlock(&overlay_lock); + + if ( tr->dt_host_new ) + xfree(tr->dt_host_new); + + xfree(tr->overlay_fdt); + xfree(tr->nodes_address); + xfree(tr->fdt); + + free_nodes_full_path(tr->num_nodes, nodes_full_path); + + xfree(tr); + + return rc; +} + long dt_sysctl(struct xen_sysctl_dt_overlay *op) { long ret; @@ -396,6 +924,11 @@ long dt_sysctl(struct xen_sysctl_dt_overlay *op) =20 switch ( op->overlay_op ) { + case XEN_SYSCTL_DT_OVERLAY_ADD: + ret =3D handle_add_overlay_nodes(overlay_fdt, op->overlay_fdt_size= ); + + break; + case XEN_SYSCTL_DT_OVERLAY_REMOVE: ret =3D handle_remove_overlay_nodes(overlay_fdt, op->overlay_fdt_s= ize); =20 --=20 2.17.1