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 37517E9DE6B for ; Thu, 9 Apr 2026 08:34:14 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id C70B510E772; Thu, 9 Apr 2026 08:34:12 +0000 (UTC) Received: from rtg-sunil-navi33.amd.com (unknown [165.204.156.251]) by gabe.freedesktop.org (Postfix) with ESMTPS id C063E10E772 for ; Thu, 9 Apr 2026 08:34:10 +0000 (UTC) Received: from rtg-sunil-navi33.amd.com (localhost [127.0.0.1]) by rtg-sunil-navi33.amd.com (8.15.2/8.15.2/Debian-22ubuntu3) with ESMTP id 6398Y0Ha1326142; Thu, 9 Apr 2026 14:04:00 +0530 Received: (from sunil@localhost) by rtg-sunil-navi33.amd.com (8.15.2/8.15.2/Submit) id 6398Y0vE1326141; Thu, 9 Apr 2026 14:04:00 +0530 From: Sunil Khatri To: Alex Deucher , =?UTF-8?q?Christian=20K=C3=B6nig?= Cc: amd-gfx@lists.freedesktop.org, Sunil Khatri Subject: [PATCH v1 1/4] drm/amdgpu/userq: dont look root bo with userq_mutex held Date: Thu, 9 Apr 2026 14:03:52 +0530 Message-Id: <20260409083355.1326089-2-sunil.khatri@amd.com> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20260409083355.1326089-1-sunil.khatri@amd.com> References: <20260409083355.1326089-1-sunil.khatri@amd.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit 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" Donot hold reservation lock for root bo if userq_mutex is already held in the call flow this cause a lock issue with ttm_bo_delayed_delete. Its better to lock the vm->root.bo first and then go ahead with userq_mutex so userq_mutex threads dont get stuck until the reservation lock is held. In this case it helps in the function amdgpu_userq_buffer_vas_mapped for each queue during restore_all. Signed-off-by: Sunil Khatri --- drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c index 9f7a08a6b018..cf8c8dfde721 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c @@ -270,15 +270,13 @@ static bool amdgpu_userq_buffer_va_mapped(struct amdgpu_vm *vm, u64 addr) struct amdgpu_bo_va_mapping *mapping; bool r; - if (amdgpu_bo_reserve(vm->root.bo, false)) - return false; + dma_resv_assert_held(vm->root.bo->tbo.base.resv); mapping = amdgpu_vm_bo_lookup_mapping(vm, addr); if (!IS_ERR_OR_NULL(mapping) && atomic_read(&mapping->bo_va->userq_va_mapped)) r = true; else r = false; - amdgpu_bo_unreserve(vm->root.bo); return r; } @@ -991,10 +989,16 @@ int amdgpu_userq_ioctl(struct drm_device *dev, void *data, static int amdgpu_userq_restore_all(struct amdgpu_userq_mgr *uq_mgr) { + struct amdgpu_fpriv *fpriv = uq_mgr_to_fpriv(uq_mgr); + struct amdgpu_vm *vm = &fpriv->vm; struct amdgpu_usermode_queue *queue; unsigned long queue_id; int ret = 0, r; + + if (amdgpu_bo_reserve(vm->root.bo, false)) + return false; + mutex_lock(&uq_mgr->userq_mutex); /* Resume all the queues for this process */ xa_for_each(&uq_mgr->userq_xa, queue_id, queue) { @@ -1012,6 +1016,7 @@ amdgpu_userq_restore_all(struct amdgpu_userq_mgr *uq_mgr) } mutex_unlock(&uq_mgr->userq_mutex); + amdgpu_bo_unreserve(vm->root.bo); if (ret) drm_file_err(uq_mgr->file, -- 2.34.1