From patchwork Mon Mar 9 23:35:58 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: jeff.mcgee@intel.com X-Patchwork-Id: 5972351 Return-Path: X-Original-To: patchwork-intel-gfx@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork2.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork2.web.kernel.org (Postfix) with ESMTP id 03C21BF440 for ; Mon, 9 Mar 2015 23:14:44 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 15A9D20437 for ; Mon, 9 Mar 2015 23:14:43 +0000 (UTC) Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) by mail.kernel.org (Postfix) with ESMTP id AA96B203F4 for ; Mon, 9 Mar 2015 23:14:41 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id EA7BD6E62C; Mon, 9 Mar 2015 16:14:40 -0700 (PDT) X-Original-To: intel-gfx@lists.freedesktop.org Delivered-To: intel-gfx@lists.freedesktop.org Received: from mga11.intel.com (mga11.intel.com [192.55.52.93]) by gabe.freedesktop.org (Postfix) with ESMTP id B9ABB6E62C; Mon, 9 Mar 2015 16:14:39 -0700 (PDT) Received: from orsmga003.jf.intel.com ([10.7.209.27]) by fmsmga102.fm.intel.com with ESMTP; 09 Mar 2015 16:14:39 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.11,370,1422950400"; d="scan'208";a="538345947" Received: from jeffdesk.fm.intel.com ([10.19.123.159]) by orsmga003.jf.intel.com with ESMTP; 09 Mar 2015 16:14:09 -0700 From: jeff.mcgee@intel.com To: beignet@lists.freedesktop.org Date: Mon, 9 Mar 2015 16:35:58 -0700 Message-Id: <1425944158-28223-1-git-send-email-jeff.mcgee@intel.com> X-Mailer: git-send-email 2.3.0 In-Reply-To: <1425339759-19027-2-git-send-email-jeff.mcgee@intel.com> References: <1425339759-19027-2-git-send-email-jeff.mcgee@intel.com> Cc: intel-gfx@lists.freedesktop.org, dri-devel@lists.freedesktop.org Subject: [Intel-gfx] [PATCH 2/2 v2] Query the driver directly for compute units and subslice X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.18 Precedence: list List-Id: Intel graphics driver community testing & development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" X-Spam-Status: No, score=-4.2 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_MED, T_RP_MATCHES_RCVD, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP From: Jeff McGee Values of device max compute units and max subslice obtained directly from the driver should be more accurate than our own ID-based lookup values. This is particularly important when a single device ID may encompass more than one configuration. If the driver cannot provide a valid value for the given device, we fallback on the ID-based lookup value. This query requires libdrm 2.4.60. For now we will consider the use of this query to be optional and exclude it from compilation when building against older libdrm. Later we may want to consider requiring the query or at least warning more strongly when it is not supported. v2: Make feature use conditional on libdrm version (Zhigang). Signed-off-by: Jeff McGee Reviewed-by: Zhigang Gong --- CMakeLists.txt | 9 +++++++++ src/CMakeLists.txt | 10 ++++++++++ src/intel/intel_driver.c | 25 +++++++++++++++++++++---- 3 files changed, 40 insertions(+), 4 deletions(-) diff --git a/CMakeLists.txt b/CMakeLists.txt index 65f2c70..bb03566 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -131,6 +131,15 @@ IF(DRM_INTEL_FOUND) ELSE(DRM_INTEL_VERSION VERSION_GREATER 2.4.57) MESSAGE(STATUS "Disable userptr support") ENDIF(DRM_INTEL_VERSION VERSION_GREATER 2.4.57) + IF(DRM_INTEL_VERSION VERSION_GREATER 2.4.59) + MESSAGE(STATUS "Enable EU total query support") + SET(DRM_INTEL_EU_TOTAL "enable") + MESSAGE(STATUS "Enable subslice total query support") + SET(DRM_INTEL_SUBSLICE_TOTAL "enable") + ELSE(DRM_INTEL_VERSION VERSION_GREATER 2.4.59) + MESSAGE(STATUS "Disable EU total query support") + MESSAGE(STATUS "Disable subslice total query support") + ENDIF(DRM_INTEL_VERSION VERSION_GREATER 2.4.59) ELSE(DRM_INTEL_FOUND) MESSAGE(FATAL_ERROR "Looking for DRM Intel (>= 2.4.52) - not found") ENDIF(DRM_INTEL_FOUND) diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt index d4181d8..464765f 100644 --- a/src/CMakeLists.txt +++ b/src/CMakeLists.txt @@ -118,6 +118,16 @@ SET(CMAKE_CXX_FLAGS "-DHAS_USERPTR ${CMAKE_CXX_FLAGS}") SET(CMAKE_C_FLAGS "-DHAS_USERPTR ${CMAKE_C_FLAGS}") endif (DRM_INTEL_USERPTR) +if (DRM_INTEL_EU_TOTAL) +SET(CMAKE_CXX_FLAGS "-DHAS_EU_TOTAL ${CMAKE_CXX_FLAGS}") +SET(CMAKE_C_FLAGS "-DHAS_EU_TOTAL ${CMAKE_C_FLAGS}") +endif (DRM_INTEL_EU_TOTAL) + +if (DRM_INTEL_SUBSLICE_TOTAL) +SET(CMAKE_CXX_FLAGS "-DHAS_SUBSLICE_TOTAL ${CMAKE_CXX_FLAGS}") +SET(CMAKE_C_FLAGS "-DHAS_SUBSLICE_TOTAL ${CMAKE_C_FLAGS}") +endif (DRM_INTEL_SUBSLICE_TOTAL) + set(GIT_SHA1 "git_sha1.h") add_custom_target(${GIT_SHA1} ALL COMMAND chmod +x ${CMAKE_CURRENT_SOURCE_DIR}/git_sha1.sh diff --git a/src/intel/intel_driver.c b/src/intel/intel_driver.c index d61988c..755ab6b 100644 --- a/src/intel/intel_driver.c +++ b/src/intel/intel_driver.c @@ -757,10 +757,7 @@ static int intel_buffer_set_tiling(cl_buffer bo, static void intel_update_device_info(cl_device_id device) { -#ifdef HAS_USERPTR intel_driver_t *driver; - const size_t sz = 4096; - void *host_ptr; driver = intel_driver_new(); assert(driver != NULL); @@ -769,6 +766,10 @@ intel_update_device_info(cl_device_id device) return; } +#ifdef HAS_USERPTR + const size_t sz = 4096; + void *host_ptr; + host_ptr = cl_aligned_malloc(sz, 4096); if (host_ptr != NULL) { cl_buffer bo = intel_buffer_alloc_userptr((cl_buffer_mgr)driver->bufmgr, @@ -781,12 +782,28 @@ intel_update_device_info(cl_device_id device) } else device->host_unified_memory = CL_FALSE; +#endif + +#ifdef HAS_EU_TOTAL + unsigned int eu_total; + + /* Prefer driver-queried max compute units if supported */ + if (!drm_intel_get_eu_total(driver->fd, &eu_total)) + device->max_compute_unit = eu_total; +#endif + +#ifdef HAS_SUBSLICE_TOTAL + unsigned int subslice_total; + + /* Prefer driver-queried subslice count if supported */ + if (!drm_intel_get_subslice_total(driver->fd, &subslice_total)) + device->sub_slice_count = subslice_total; +#endif intel_driver_context_destroy(driver); intel_driver_close(driver); intel_driver_terminate(driver); intel_driver_delete(driver); -#endif } LOCAL void