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 6DA20CA600C for ; Thu, 8 Oct 2026 16:17:37 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id F12F310E599; Thu, 8 Oct 2026 16:17:36 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (1024-bit key; unprotected) header.d=amd.com header.i=@amd.com header.b="R7837uWe"; dkim-atps=neutral Received: from CY3PR05CU001.outbound.protection.outlook.com (mail-westcentralusazon11013032.outbound.protection.outlook.com [40.93.201.32]) by gabe.freedesktop.org (Postfix) with ESMTPS id 85DD110E0A4; Thu, 8 Oct 2026 16:17:35 +0000 (UTC) ARC-Seal: i=1; a=rsa-sha256; s=arcselector10001; d=microsoft.com; cv=none; b=QTtHZ/5IXxLDgahmxYtWLG9jJpxe0oexS1nX3Jibe5bDOSutd66z96OI4H60qL2fyu/pFcldQTJRsHJv9uKhTPktzzndCuPcSvJzeD2docBa8TzJ4gYZc92NnJqApokXvTnAm5lGo2RkwGn0jORGerPBDiwekpLuoEvZ29F+6PawEPnLbdrcySav2yGhcnk2alGbRvF2hBUyHzyjcjHhbG/nuOpFdMe2TcMKTe0Yi80VZ//4gPMdwwVaUE5YwCRUL3aTJ4b7BS+HSoN9XAfUgKfqteVfue0MWf10NWYjszYPQguKktfzpYFl/Ij/MufS58sj2VezgQJN7E9D89Q0kQ== 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=LOADTAPMxMPXmDXEnbe3pLYrYlGgX1aF/eSDpwqjqL4=; b=uOPRIAMbBKQjbD80fkjM6eVnX1OsuYHbf4KdmNz56ysqp9ZE17JFERKROGEx/53jcuZSx53sesTCeIeTmoBA9CiFZH4s8hTq6rzj2rxA/syJ8wE2nTTHILU7YH8A/vRmBoTYLVh2gi52637PVnYd4TeExC/fZCr47dU9eYUBqu5UV1t/DItcFYPU89AKMIVtclwN6u2ghbFrWleek+Q26YezVFSD7DTosyAk+K/TgJEGI2tyM654Hy9A72xuuwHSAEppyatC4XIiRDXRSVKPk8DJCcSpUAimdPTm/ILgu++FunXnPSNe9pRjGAzw0ifkhuJGHwCy5FoF86FhbSWhWQ== 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=LOADTAPMxMPXmDXEnbe3pLYrYlGgX1aF/eSDpwqjqL4=; b=R7837uWefNVq1uGjSDPWG1v/v6PWPnRwzSwfe/iBAzIfSca+YGDqXrCDFG/rc7ZPTWzxc6wPn7xVBL9kqVbDuqd9PBn8RcJaDByLoAa+1qe873A1laJZEaZbytJKe8pH0nstOTL2UBU7b/HqLaoVpr8GphrZQJzBqPL+xpdxgBs= Authentication-Results: mx.microsoft.com 1; dkim=none (message not signed) header.d=none;dmarc=none action=none header.from=amd.com; Received: from DSVPR12MB449095.namprd12.prod.outlook.com (2603:10b6:8:4fe::22) by CY5PR12MB6060.namprd12.prod.outlook.com (2603:10b6:930:2f::9) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.21.496.15; Thu, 8 Oct 2026 16:17:31 +0000 Received: from DSVPR12MB449095.namprd12.prod.outlook.com ([fe80::76a6:7a4f:a0a5:18c0]) by DSVPR12MB449095.namprd12.prod.outlook.com ([fe80::76a6:7a4f:a0a5:18c0%4]) with mapi id 15.21.0496.015; Thu, 8 Oct 2026 16:17:31 +0000 Message-ID: <0b8f0025-0a76-47fd-b4e4-bcab26f70366@amd.com> Date: Thu, 8 Oct 2026 12:17:28 -0400 User-Agent: Mozilla Thunderbird Subject: Re: [PATCH] drm/exec: don't immediately add prelocked obj to array of locked objs To: christian.koenig@amd.com, vitaly.prosyak@amd.com Cc: amd-gfx@lists.freedesktop.org, dri-devel@lists.freedesktop.org References: <20261007124657.9664-1-christian.koenig@amd.com> Content-Language: en-US From: vitaly prosyak In-Reply-To: <20261007124657.9664-1-christian.koenig@amd.com> Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-ClientProxiedBy: YT4PR01CA0347.CANPRD01.PROD.OUTLOOK.COM (2603:10b6:b01:fc::18) To DSVPR12MB449095.namprd12.prod.outlook.com (2603:10b6:8:4fe::22) MIME-Version: 1.0 X-MS-PublicTrafficType: Email X-MS-TrafficTypeDiagnostic: DSVPR12MB449095:EE_|CY5PR12MB6060:EE_ X-MS-Office365-Filtering-Correlation-Id: 64bcbfaf-12f3-4451-134f-08df2557a76e X-MS-Exchange-SenderADCheck: 1 X-MS-Exchange-AntiSpam-Relay: 0 X-Microsoft-Antispam: BCL:0; ARA:13230040|376014|1800799024|366016|23010399003|22082099003|18002099003|10067099003|260925021311599003|260925022911599003|260925021911599003|11063799006|56012099006; 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:DSVPR12MB449095.namprd12.prod.outlook.com; PTR:; CAT:NONE; SFS:(13230040)(376014)(1800799024)(366016)(23010399003)(22082099003)(18002099003)(10067099003)(260925021311599003)(260925022911599003)(260925021911599003)(11063799006)(56012099006); DIR:OUT; SFP:1101; X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1 X-MS-Exchange-AntiSpam-MessageData-0: =?utf-8?B?bUVaQkZKNDU4SVNubXZmOWxUODg3cjJMbkRpK21PVEdudUo1ekNwSlJSWm5n?= =?utf-8?B?SDlqQTJpRG5LbXk1Qm15Q0lXb2hGcnlSRWVxakU1dGtxckloTWp0endNNjcy?= =?utf-8?B?dDFNbHMwZXFBY1c0QlhFNkdZL1ExT3liVjhCcHBWcWxER2tFbjUva29QdWIr?= =?utf-8?B?UWJtQm5WaDBtb0UzSjZ0T0VsZC80R2ZXVkJ6THFoN2s4eDYzSkw0cU1ZQlBW?= =?utf-8?B?b0N1bURZQTNiREc4TlA5Zk4rNmhYUnVsSmNFYTd3eE1XSUNRZzdyRzA5YXcv?= =?utf-8?B?RXJud3RZYWwzd0dVSEdYRDVFSWVxVEF1M3pMbWR3NGN1K1F3dzRMV04xUzlW?= =?utf-8?B?TXA4dUM0WTlCYkoxZGtJYTBNTE5YRUh5TTgvQi9UN2M1VTBqS2FMT2hwRVpa?= =?utf-8?B?dzJQYmROV3FOOUtMUW5KOE9oWCtOVjF0TlM2SHNlMUh5c0tCcGJFUzJNL2Zu?= =?utf-8?B?aHlIbU16YTRqSjNDZ2EwdGxCVFh4MEpTaDM3cTNRSzNJc243OElZWUg2QXIy?= =?utf-8?B?eWhlWTZ6dEhBejBSL1ZOSFJmVXRiV2V1aEgvRCs1Mmp0S21td1R1SFViNVB5?= =?utf-8?B?N2ZCMUg4Z1ZhWWtTcXRQNGJ2NThPWFowZ0FxbmpueWIydHJGOFROMVZnZnB4?= =?utf-8?B?TVBjaXhoS25RU2piTUowcnNXc1JzOUpLTUErQk1sUTI0OUc4bWphc1VIa3BY?= =?utf-8?B?RjFZSkRMdHB2RzhVM1BNS0pERHNTVFZnT2hncDQwZ21mcGVDY3ZUWVpCaExL?= =?utf-8?B?dS9pS3VYNmFKWlVWMm1DUnNlZng3WnFMaTRHeVJEa1F4M0VPSFdmc3Q1TlZo?= =?utf-8?B?YlhPODB5UGpIQnZ6WHdvSE02eXNuNVMxL2JjRWE0UUtoTUpZSHZCQ1ZHWWlQ?= =?utf-8?B?VW11QVErclJGUURQajNlRFlLTTVRU29ackt2WGt1eVlpb2p1SjFIWkhRQy9m?= =?utf-8?B?UWgzM0liR2ZocmI2ZTlwVDFKM1AzTys1RHIzSGcvY3lMcnR4dzYxd2x0d1Mw?= =?utf-8?B?Q0VTRXVrSWZXL29tNFJteUExalArbGZlL0tXRlFMV0dYcHpUUWlsb3R1a2U0?= =?utf-8?B?TTlFL2hWS1NiTlg2MmdxNDNkdy9BMlU1enZubkJkU25FWGlkR21DTE9WdSsv?= =?utf-8?B?V2VYNFptMTFRTjI5TXNLc0VwTUtWZWcyV1MvWm1sVFpTTnk3MEZDVlZhQmd0?= =?utf-8?B?RTVNajIwdm81cFNWTEU3bDNDazhUTGtUb1JLTTllWWN2ZnhvcXA2djhFb3RW?= =?utf-8?B?cFozd3lhV0Z4RVpLRnRNaHJnL2VlR3ZEOXNkeG1HWkp0alZWWTRqSHVLeGlK?= =?utf-8?B?NVA2MERFOTlFa2M5Q2xKV1hOZUI5QXhwWnE3Q2l5MTNyekVtcFYyTUFTWk04?= =?utf-8?B?ZkMwZUhQa0twZWQ4QVFmL0p2Qi9rZUdtNnNlYXltcVBzd3oxMTJTc2dSWE1x?= =?utf-8?B?WjVXOWlKWmtDTnRYcmNIVDh1SVNFbU5mSlIvcFF0WEQ1TzNFWUY5SDFHN1FK?= =?utf-8?B?NEV4SFRpWDFJZDJvZmtZUHBNZUJtdFRubU9tMHE3VHUrQk1VZzNEZ1NBSXVF?= =?utf-8?B?OWN1NXVzaHVTM1E4UXhMcThUU0lFVFhHY09RYkt0dkUyRW5PVnRzNmwyaFBM?= =?utf-8?B?aEx5bi9EMW9zS3psRjVTRk9yYmZRdjliMk5pSGFXazRYaHJtT09HVjJkS21t?= =?utf-8?B?bld6T1RUcU44cDcvRDVyOWpacWVGd1pxTkpDK3Y5LzdHM1BwV1JwWHdQbXNL?= =?utf-8?B?blpNSkxFYzl5L1Y5VXVxS0pVeE9XczRQTk9XL0NGeXRqa1grUFkwd0E5S1Qx?= =?utf-8?B?d0tqRXRjSWppQ3FMSXJpZzVsc0NlS2cwWUl1a2JtaXRYVCt2MXNjV3VYc2U0?= =?utf-8?B?NkxXKy9UUTNrV09iNkFjbFYwbWNhNHIxQVY3V09iNVFBMEVpQkoyTng0d3Q2?= =?utf-8?B?NlJZOWh5VnFFYjRBL2JzN3YwMXE2NERndDRNd0VYTlcxNzRqblhXUHdJb094?= =?utf-8?B?NnFGTThHSXhXWW8rRzFYdXpXVldKK044WTMvcVh2S3psNXlkaGpnSjhNVGo0?= =?utf-8?B?RzJYUkZNem03bzdoQWZkNzMremFtN0FoUm1jd3QzL3ZEZ3BjUnJlMThHTG5H?= =?utf-8?B?KzduazZrWURvaTNqb2x5VXNqWkN1bTFkQUVBWXVMbXBHWHVqeXo2QTBqQ1gy?= =?utf-8?B?MnlQQkhVa0FzUjg3QlRDTXBqZmhBdEUzYlZ3SkhtbjZUTVNZQms3S2lJU3RN?= =?utf-8?B?TjYyZGNNbnhNdG5OUk52ZU5rSVlidmhjSjl2bU1OVXpQUEYwS2poalJIeGZ1?= =?utf-8?B?ZVptdnRxK2Nua1FscTJvTS9hNm5RUlp1RmdBc01HSm9ORFVVYzYwUT09?= X-OriginatorOrg: amd.com X-MS-Exchange-CrossTenant-Network-Message-Id: 64bcbfaf-12f3-4451-134f-08df2557a76e X-MS-Exchange-CrossTenant-AuthSource: DSVPR12MB449095.namprd12.prod.outlook.com X-MS-Exchange-CrossTenant-AuthAs: Internal X-MS-Exchange-CrossTenant-OriginalArrivalTime: 08 Oct 2026 16:17:31.3244 (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: P3FMyQVcKk79sOux774oxjR9zI3dxrQQ0yv0jdHFk82++fShqRt/A32sV3ZE3wc32YEgcLLi2U44k2ac6LxBIw== X-MS-Exchange-Transport-CrossTenantHeadersStamped: CY5PR12MB6060 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" Hi Christian, Thanks for the fix. Tested under the same conditions that previously reproduced the crash at dma-resv.c:329; the crash did not reproduce with this patch 2026-10-03T16:00:58-04:00 Intel-system kernel: kernel BUG at drivers/dma-buf/dma-resv.c:329! 2026-10-03T16:00:58-04:00 Intel-system kernel: Oops: invalid opcode: 0000 [#1] SMP NOPTI 2026-10-03T16:00:58-04:00 Intel-system kernel: CPU: 11 UID: 0 PID: 477 Comm: kworker/11:2 Not tainted 7.1.0+ #83 PREEMPT(full)  2026-10-03T16:00:58-04:00 Intel-system kernel: Hardware name: ASUS System Product Name/ROG MAXIMUS Z790 HERO, BIOS 2703 10/17/2024 2026-10-03T16:00:58-04:00 Intel-system kernel: Workqueue: events amdgpu_userq_restore_worker [amdgpu] 2026-10-03T16:00:58-04:00 Intel-system kernel: RIP: 0010:dma_resv_add_fence+0x2d6/0x2f0 2026-10-03T16:00:58-04:00 Intel-system kernel: Code: 44 8b 88 30 0a 00 00 48 05 18 0d 00 00 50 4c 8b 45 08 48 8b 75 b8 e8 f9 01 35 ff 58 41 8b 45 14 41 39 45 10 0f 82 d2 fe ff ff <0f> 0b 31 d2 45 31 ff e9 ba fe ff ff e8 c9 a5 5c 00 66 0f 1f 84 00 2026-10-03T16:00:58-04:00 Intel-system kernel: RSP: 0018:ffffcb8682e5fb28 EFLAGS: 00010246 2026-10-03T16:00:58-04:00 Intel-system kernel: RAX: 0000000000000000 RBX: 0000000000000003 RCX: 0000000000000000 2026-10-03T16:00:58-04:00 Intel-system kernel: RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000 2026-10-03T16:00:58-04:00 Intel-system kernel: RBP: ffffcb8682e5fb78 R08: 0000000000000000 R09: 0000000000000000 2026-10-03T16:00:58-04:00 Intel-system kernel: R10: 0000000000000000 R11: 0000000000000000 R12: ffff88edd40ca500 2026-10-03T16:00:58-04:00 Intel-system kernel: R13: ffff88ed91d83f00 R14: ffffffffc1c3ee60 R15: 0000000000000000 2026-10-03T16:00:58-04:00 Intel-system kernel: FS:  0000000000000000(0000) GS:ffff88f53e2bc000(0000) knlGS:0000000000000000 2026-10-03T16:00:58-04:00 Intel-system kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033 2026-10-03T16:00:58-04:00 Intel-system kernel: CR2: 0000756822dd7000 CR3: 000000012e1d4006 CR4: 0000000000f72ef0 2026-10-03T16:00:58-04:00 Intel-system kernel: PKRU: 55555554 2026-10-03T16:00:58-04:00 Intel-system kernel: Call Trace: 2026-10-03T16:00:58-04:00 Intel-system kernel:  2026-10-03T16:00:58-04:00 Intel-system kernel:  amdgpu_evf_mgr_attach_fence+0x179/0x330 [amdgpu] 2026-10-03T16:00:58-04:00 Intel-system kernel:  amdgpu_evf_mgr_rearm+0x1ff/0x2b0 [amdgpu] 2026-10-03T16:00:58-04:00 Intel-system kernel:  amdgpu_userq_vm_validate_and_restore_queue+0x8f0/0xc80 [amdgpu] 2026-10-03T16:00:58-04:00 Intel-system kernel:  amdgpu_userq_restore_worker+0x64/0xb0 [amdgpu] 2026-10-03T16:00:58-04:00 Intel-system kernel:  process_one_work+0x22e/0x780 2026-10-03T16:00:58-04:00 Intel-system kernel:  worker_thread+0x1db/0x3d0 2026-10-03T16:00:58-04:00 Intel-system kernel:  ? __pfx_worker_thread+0x10/0x10 2026-10-03T16:00:58-04:00 Intel-system kernel:  kthread+0xfe/0x140 2026-10-03T16:00:58-04:00 Intel-system kernel:  ? __pfx_kthread+0x10/0x10 2026-10-03T16:00:58-04:00 Intel-system kernel:  ret_from_fork+0x3bd/0x470 2026-10-03T16:00:58-04:00 Intel-system kernel:  ? __pfx_kthread+0x10/0x10 2026-10-03T16:00:58-04:00 Intel-system kernel:  ret_from_fork_asm+0x1a/0x30 2026-10-03T16:00:58-04:00 Intel-system kernel:  My patch preserved the fence-slot request across retries. Your approach instead excludes the implicit prelocked BO from iteration until explicitly requested again, avoiding the extra reservation-count bookkeeping. A few minor wording corrections: - "keeping the prelocked object of the array" should be   "keeping the prelocked object out of the array". - "Should the prelocked object never been mentioned again" should be   "If the prelocked object is never mentioned again". - "might never given to" should be "might never be passed to". - "unecessarily" should be "unnecessarily". Tested-by: Vitaly Prosyak Reviewed-by: Vitaly Prosyak Thanks, Vitaly On 2026-10-07 08:46, Christian König wrote: > When a contention is detected the drm_exec object grabs a reference to > the contended object, unlocks all other objects and pre-locks the > contended one before anything else. > > The problem is that it is possible that this pre-locked object was about > to be destroyed and drm_exec is now holding the last reference to it. > > So we never actually see this pre-locked object again to be explicitly > locked and so also never reserve a slot for a dma_fence object on it. > > Should the driver now use drm_exec_for_each_locked_object() to add a new > fence to all locked objects we eventually see a warning or even a > BUG_ON() to prevent random memory corruption. > > Fix this by keeping the prelocked object of the array of locked objs > until it was explicitly mentioned by the driver again. > > Should the prelocked object never been mentioned again by the driver > just unlock and drop the reference at the end of the locking loop. > > Signed-off-by: Christian König > Reported-by: Vitaly Prosyak > Fixes: 09593216bff1 ("drm: execution context for GEM buffers v7") > CC: stable@vger.kernel.org # v6.6+ > --- > drivers/gpu/drm/drm_exec.c | 73 ++++++++++++++++++++++---------------- > include/drm/drm_exec.h | 15 ++++---- > 2 files changed, 52 insertions(+), 36 deletions(-) > > diff --git a/drivers/gpu/drm/drm_exec.c b/drivers/gpu/drm/drm_exec.c > index 2453ec41360f4..a910b0ad17ce4 100644 > --- a/drivers/gpu/drm/drm_exec.c > +++ b/drivers/gpu/drm/drm_exec.c > @@ -48,6 +48,16 @@ > * See struct dma_exec for more details. > */ > > +/* Unlock and put the prelocked obj */ > +static void drm_exec_drop_prelocked(struct drm_exec *exec) > +{ > + if (exec->prelocked) { > + dma_resv_unlock(exec->prelocked->resv); > + drm_gem_object_put(exec->prelocked); > + exec->prelocked = NULL; > + } > +} > + > /* Unlock all objects and drop references */ > static void drm_exec_unlock_all(struct drm_exec *exec) > { > @@ -58,8 +68,7 @@ static void drm_exec_unlock_all(struct drm_exec *exec) > drm_gem_object_put(obj); > } > > - drm_gem_object_put(exec->prelocked); > - exec->prelocked = NULL; > + drm_exec_drop_prelocked(exec); > } > > /** > @@ -118,6 +127,14 @@ EXPORT_SYMBOL(drm_exec_fini); > bool drm_exec_cleanup(struct drm_exec *exec) > { > if (likely(!exec->contended)) { > + > + /* > + * The contended and prelocked object might never given to > + * drm_exec_lock_obj() again. This can happen when the object > + * is about to be cleaned up and we now hold the last reference > + * to it. Make sure that we don't unecessarily keep it around. > + */ > + drm_exec_drop_prelocked(exec); > ww_acquire_done(&exec->ticket); > return false; > } > @@ -175,16 +192,16 @@ static int drm_exec_lock_contended(struct drm_exec *exec) > dma_resv_lock_slow(obj->resv, &exec->ticket); > } > > - ret = drm_exec_obj_locked(exec, obj); > - if (unlikely(ret)) > - goto error_unlock; > + /* > + * It is perfectly possible that we see another contention before the > + * first one is fully handled. > + */ > + drm_exec_drop_prelocked(exec); > > + /* We move the reference from contended to prelocked here */ > exec->prelocked = obj; > return 0; > > -error_unlock: > - dma_resv_unlock(obj->resv); > - > error_dropref: > drm_gem_object_put(obj); > return ret; > @@ -212,35 +229,31 @@ int drm_exec_lock_obj(struct drm_exec *exec, struct drm_gem_object *obj) > if (exec->prelocked == obj) { > drm_gem_object_put(exec->prelocked); > exec->prelocked = NULL; > - return 0; > - } > + } else { > + if (exec->flags & DRM_EXEC_INTERRUPTIBLE_WAIT) > + ret = dma_resv_lock_interruptible(obj->resv, > + &exec->ticket); > + else > + ret = dma_resv_lock(obj->resv, &exec->ticket); > + > + if (unlikely(ret == -EDEADLK)) { > + drm_gem_object_get(obj); > + exec->contended = obj; > + return -EDEADLK; > + } > > - if (exec->flags & DRM_EXEC_INTERRUPTIBLE_WAIT) > - ret = dma_resv_lock_interruptible(obj->resv, &exec->ticket); > - else > - ret = dma_resv_lock(obj->resv, &exec->ticket); > + if (unlikely(ret == -EALREADY) && > + exec->flags & DRM_EXEC_IGNORE_DUPLICATES) > + return 0; > > - if (unlikely(ret == -EDEADLK)) { > - drm_gem_object_get(obj); > - exec->contended = obj; > - return -EDEADLK; > + if (unlikely(ret)) > + return ret; > } > > - if (unlikely(ret == -EALREADY) && > - exec->flags & DRM_EXEC_IGNORE_DUPLICATES) > - return 0; > - > - if (unlikely(ret)) > - return ret; > - > ret = drm_exec_obj_locked(exec, obj); > if (ret) > - goto error_unlock; > - > - return 0; > + dma_resv_unlock(obj->resv); > > -error_unlock: > - dma_resv_unlock(obj->resv); > return ret; > } > EXPORT_SYMBOL(drm_exec_lock_obj); > diff --git a/include/drm/drm_exec.h b/include/drm/drm_exec.h > index cc2937185a9f7..c92d69a2c793d 100644 > --- a/include/drm/drm_exec.h > +++ b/include/drm/drm_exec.h > @@ -73,13 +73,15 @@ drm_exec_obj(struct drm_exec *exec, unsigned long index) > > /* Helper for drm_exec_for_each_locked_object(). Internal use only. */ > #define __drm_exec_for_each_locked_object(exec, obj, __index) \ > - for (unsigned long __index = 0; ((obj) = drm_exec_obj(exec, __index)); ++__index) > + for (unsigned long __index = 0; ((obj) = drm_exec_obj(exec, __index)); \ > + ++__index) > /** > * drm_exec_for_each_locked_object - iterate over all the locked objects > * @exec: drm_exec object > * @obj: the current GEM object > * > - * Iterate over all the locked GEM objects inside the drm_exec object. > + * Iterate over all the explicitly locked GEM objects inside the drm_exec > + * container, except for the contended and implicit prelocked one. > */ > #define drm_exec_for_each_locked_object(exec, obj) \ > __drm_exec_for_each_locked_object(exec, obj, __UNIQUE_ID(drm_exec)) > @@ -94,12 +96,13 @@ drm_exec_obj(struct drm_exec *exec, unsigned long index) > * @exec: drm_exec object > * @obj: the current GEM object > * > - * Iterate over all the locked GEM objects inside the drm_exec object in > - * reverse locking order. Note that the internal index may wrap around, > - * but that will be caught by drm_exec_obj(), returning a NULL object. > + * Iterate over all the explicitly locked GEM objects inside the drm_exec > + * container in reverse locking order. Note that the internal index may wrap > + * around, but that will be caught by drm_exec_obj(), returning a NULL object. > */ > #define drm_exec_for_each_locked_object_reverse(exec, obj) \ > - __drm_exec_for_each_locked_object_reverse(exec, obj, __UNIQUE_ID(drm_exec)) > + __drm_exec_for_each_locked_object_reverse(exec, obj, \ > + __UNIQUE_ID(drm_exec)) > > /** > * drm_exec_until_all_locked - loop until all GEM objects are locked