From patchwork Thu May 23 11:46:19 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Lionel Landwerlin X-Patchwork-Id: 10957457 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id ECA6A924 for ; Thu, 23 May 2019 11:46:30 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id D4AC428438 for ; Thu, 23 May 2019 11:46:30 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id C8A652844C; Thu, 23 May 2019 11:46:30 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-5.2 required=2.0 tests=BAYES_00,MAILING_LIST_MULTI, RCVD_IN_DNSWL_MED autolearn=unavailable version=3.3.1 Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher DHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.wl.linuxfoundation.org (Postfix) with ESMTPS id 4EA2C28438 for ; Thu, 23 May 2019 11:46:30 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id E067D89DFD; Thu, 23 May 2019 11:46:27 +0000 (UTC) X-Original-To: intel-gfx@lists.freedesktop.org Delivered-To: intel-gfx@lists.freedesktop.org Received: from mga03.intel.com (mga03.intel.com [134.134.136.65]) by gabe.freedesktop.org (Postfix) with ESMTPS id F29BF89DFF; Thu, 23 May 2019 11:46:24 +0000 (UTC) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga004.jf.intel.com ([10.7.209.38]) by orsmga103.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 23 May 2019 04:46:24 -0700 X-ExtLoop1: 1 Received: from delly.ld.intel.com ([10.103.238.204]) by orsmga004.jf.intel.com with ESMTP; 23 May 2019 04:46:23 -0700 From: Lionel Landwerlin To: intel-gfx@lists.freedesktop.org Date: Thu, 23 May 2019 12:46:19 +0100 Message-Id: <20190523114620.19335-2-lionel.g.landwerlin@intel.com> X-Mailer: git-send-email 2.21.0.392.gf8f6787159e In-Reply-To: <20190523114620.19335-1-lionel.g.landwerlin@intel.com> References: <20190523114620.19335-1-lionel.g.landwerlin@intel.com> MIME-Version: 1.0 Subject: [Intel-gfx] [PATCH 1/2] drm/syncobj: add an output syncobj parameter to find_fence X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: Intel graphics driver community testing & development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: David Zhou , DRI-Devel , Christian Koenig Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" X-Virus-Scanned: ClamAV using ClamSMTP We would like to get both the fence & the syncobj in i915 rather than doing 2 calls to drm_syncobj_find() & drm_syncobj_find_fence(). Signed-off-by: Lionel Landwerlin Cc: Christian Koenig Cc: David(ChunMing) Zhou Cc: Eric Anholt CC: DRI-Devel Signed-off-by: Lionel Landwerlin Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
--- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 4 ++- drivers/gpu/drm/drm_syncobj.c | 45 +++++++++++++++++--------- drivers/gpu/drm/v3d/v3d_gem.c | 5 ++- include/drm/drm_syncobj.h | 1 + 4 files changed, 38 insertions(+), 17 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 2f6239b6be6f..09fde3c73a2c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -1124,10 +1124,11 @@ static int amdgpu_syncobj_lookup_and_add_to_sync(struct amdgpu_cs_parser *p, uint32_t handle, u64 point, u64 flags) { + struct drm_syncobj *syncobj; struct dma_fence *fence; int r; - r = drm_syncobj_find_fence(p->filp, handle, point, flags, &fence); + r = drm_syncobj_find_fence(p->filp, handle, point, flags, &syncobj, &fence); if (r) { DRM_ERROR("syncobj %u failed to find fence @ %llu (%d)!\n", handle, point, r); @@ -1136,6 +1137,7 @@ static int amdgpu_syncobj_lookup_and_add_to_sync(struct amdgpu_cs_parser *p, r = amdgpu_sync_fence(p->adev, &p->job->sync, fence, true); dma_fence_put(fence); + drm_syncobj_put(syncobj); return r; } diff --git a/drivers/gpu/drm/drm_syncobj.c b/drivers/gpu/drm/drm_syncobj.c index 3d400905100b..f2fd0c1fb1d3 100644 --- a/drivers/gpu/drm/drm_syncobj.c +++ b/drivers/gpu/drm/drm_syncobj.c @@ -222,29 +222,32 @@ static void drm_syncobj_assign_null_handle(struct drm_syncobj *syncobj) * @handle: sync object handle to lookup. * @point: timeline point * @flags: DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT or not + * @syncobj: out parameter for the syncobj * @fence: out parameter for the fence * * This is just a convenience function that combines drm_syncobj_find() and * drm_syncobj_fence_get(). * - * Returns 0 on success or a negative error value on failure. On success @fence - * contains a reference to the fence, which must be released by calling - * dma_fence_put(). + * Returns 0 on success or a negative error value on failure. On + * success @syncobj and @fence contains a reference respectively to + * the syncobj and to the fence, which must be released by calling + * respectively drm_syncobj_put() and dma_fence_put(). */ int drm_syncobj_find_fence(struct drm_file *file_private, u32 handle, u64 point, u64 flags, + struct drm_syncobj **syncobj, struct dma_fence **fence) { - struct drm_syncobj *syncobj = drm_syncobj_find(file_private, handle); struct syncobj_wait_entry wait; u64 timeout = nsecs_to_jiffies64(DRM_SYNCOBJ_WAIT_FOR_SUBMIT_TIMEOUT); int ret; - if (!syncobj) + *syncobj = drm_syncobj_find(file_private, handle); + + if (!(*syncobj)) return -ENOENT; - *fence = drm_syncobj_fence_get(syncobj); - drm_syncobj_put(syncobj); + *fence = drm_syncobj_fence_get(*syncobj); if (*fence) { ret = dma_fence_chain_find_seqno(fence, point); @@ -255,13 +258,15 @@ int drm_syncobj_find_fence(struct drm_file *file_private, ret = -EINVAL; } - if (!(flags & DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT)) + if (!(flags & DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT)) { + drm_syncobj_put(*syncobj); return ret; + } memset(&wait, 0, sizeof(wait)); wait.task = current; wait.point = point; - drm_syncobj_fence_add_wait(syncobj, &wait); + drm_syncobj_fence_add_wait(*syncobj, &wait); do { set_current_state(TASK_INTERRUPTIBLE); @@ -286,7 +291,10 @@ int drm_syncobj_find_fence(struct drm_file *file_private, *fence = wait.fence; if (wait.node.next) - drm_syncobj_remove_wait(syncobj, &wait); + drm_syncobj_remove_wait(*syncobj, &wait); + + if (ret) + drm_syncobj_put(*syncobj); return ret; } @@ -531,6 +539,7 @@ static int drm_syncobj_export_sync_file(struct drm_file *file_private, int handle, int *p_fd) { int ret; + struct drm_syncobj *syncobj; struct dma_fence *fence; struct sync_file *sync_file; int fd = get_unused_fd_flags(O_CLOEXEC); @@ -538,13 +547,14 @@ static int drm_syncobj_export_sync_file(struct drm_file *file_private, if (fd < 0) return fd; - ret = drm_syncobj_find_fence(file_private, handle, 0, 0, &fence); + ret = drm_syncobj_find_fence(file_private, handle, 0, 0, &syncobj, &fence); if (ret) goto err_put_fd; sync_file = sync_file_create(fence); dma_fence_put(fence); + drm_syncobj_put(syncobj); if (!sync_file) { ret = -EINVAL; @@ -682,7 +692,8 @@ drm_syncobj_fd_to_handle_ioctl(struct drm_device *dev, void *data, static int drm_syncobj_transfer_to_timeline(struct drm_file *file_private, struct drm_syncobj_transfer *args) { - struct drm_syncobj *timeline_syncobj = NULL; + struct drm_syncobj *timeline_syncobj; + struct drm_syncobj *src_syncobj; struct dma_fence *fence; struct dma_fence_chain *chain; int ret; @@ -693,7 +704,7 @@ static int drm_syncobj_transfer_to_timeline(struct drm_file *file_private, } ret = drm_syncobj_find_fence(file_private, args->src_handle, args->src_point, args->flags, - &fence); + &src_syncobj, &fence); if (ret) goto err; chain = kzalloc(sizeof(struct dma_fence_chain), GFP_KERNEL); @@ -704,6 +715,7 @@ static int drm_syncobj_transfer_to_timeline(struct drm_file *file_private, drm_syncobj_add_point(timeline_syncobj, chain, fence, args->dst_point); err1: dma_fence_put(fence); + drm_syncobj_put(src_syncobj); err: drm_syncobj_put(timeline_syncobj); @@ -714,7 +726,8 @@ static int drm_syncobj_transfer_to_binary(struct drm_file *file_private, struct drm_syncobj_transfer *args) { - struct drm_syncobj *binary_syncobj = NULL; + struct drm_syncobj *binary_syncobj; + struct drm_syncobj *src_syncobj; struct dma_fence *fence; int ret; @@ -722,11 +735,13 @@ drm_syncobj_transfer_to_binary(struct drm_file *file_private, if (!binary_syncobj) return -ENOENT; ret = drm_syncobj_find_fence(file_private, args->src_handle, - args->src_point, args->flags, &fence); + args->src_point, args->flags, + &src_syncobj, &fence); if (ret) goto err; drm_syncobj_replace_fence(binary_syncobj, fence); dma_fence_put(fence); + drm_syncobj_put(src_syncobj); err: drm_syncobj_put(binary_syncobj); diff --git a/drivers/gpu/drm/v3d/v3d_gem.c b/drivers/gpu/drm/v3d/v3d_gem.c index 27e0f87075d9..26bd3a2e39ca 100644 --- a/drivers/gpu/drm/v3d/v3d_gem.c +++ b/drivers/gpu/drm/v3d/v3d_gem.c @@ -431,6 +431,7 @@ v3d_job_init(struct v3d_dev *v3d, struct drm_file *file_priv, struct v3d_job *job, void (*free)(struct kref *ref), u32 in_sync) { + struct drm_syncobj *in_syncobj = NULL; struct dma_fence *in_fence = NULL; int ret; @@ -443,10 +444,12 @@ v3d_job_init(struct v3d_dev *v3d, struct drm_file *file_priv, xa_init_flags(&job->deps, XA_FLAGS_ALLOC); - ret = drm_syncobj_find_fence(file_priv, in_sync, 0, 0, &in_fence); + ret = drm_syncobj_find_fence(file_priv, in_sync, 0, 0, &syncobj, &in_fence); if (ret == -EINVAL) goto fail; + drm_syncobj_put(in_sync); + ret = drm_gem_fence_array_add(&job->deps, in_fence); if (ret) goto fail; diff --git a/include/drm/drm_syncobj.h b/include/drm/drm_syncobj.h index 6cf7243a1dc5..08eca690f783 100644 --- a/include/drm/drm_syncobj.h +++ b/include/drm/drm_syncobj.h @@ -121,6 +121,7 @@ void drm_syncobj_replace_fence(struct drm_syncobj *syncobj, struct dma_fence *fence); int drm_syncobj_find_fence(struct drm_file *file_private, u32 handle, u64 point, u64 flags, + struct drm_syncobj **syncobj, struct dma_fence **fence); void drm_syncobj_free(struct kref *kref); int drm_syncobj_create(struct drm_syncobj **out_syncobj, uint32_t flags, From patchwork Thu May 23 11:46:20 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Lionel Landwerlin X-Patchwork-Id: 10957459 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 352F9924 for ; Thu, 23 May 2019 11:46:32 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 1BBAB2843C for ; Thu, 23 May 2019 11:46:32 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 0ECBB28438; Thu, 23 May 2019 11:46:32 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-5.2 required=2.0 tests=BAYES_00,MAILING_LIST_MULTI, RCVD_IN_DNSWL_MED autolearn=ham version=3.3.1 Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher DHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.wl.linuxfoundation.org (Postfix) with ESMTPS id 4879028438 for ; Thu, 23 May 2019 11:46:31 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 8C40089E05; Thu, 23 May 2019 11:46:28 +0000 (UTC) X-Original-To: intel-gfx@lists.freedesktop.org Delivered-To: intel-gfx@lists.freedesktop.org Received: from mga03.intel.com (mga03.intel.com [134.134.136.65]) by gabe.freedesktop.org (Postfix) with ESMTPS id 25AA289E01 for ; Thu, 23 May 2019 11:46:26 +0000 (UTC) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga004.jf.intel.com ([10.7.209.38]) by orsmga103.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 23 May 2019 04:46:26 -0700 X-ExtLoop1: 1 Received: from delly.ld.intel.com ([10.103.238.204]) by orsmga004.jf.intel.com with ESMTP; 23 May 2019 04:46:25 -0700 From: Lionel Landwerlin To: intel-gfx@lists.freedesktop.org Date: Thu, 23 May 2019 12:46:20 +0100 Message-Id: <20190523114620.19335-3-lionel.g.landwerlin@intel.com> X-Mailer: git-send-email 2.21.0.392.gf8f6787159e In-Reply-To: <20190523114620.19335-1-lionel.g.landwerlin@intel.com> References: <20190523114620.19335-1-lionel.g.landwerlin@intel.com> MIME-Version: 1.0 Subject: [Intel-gfx] [PATCH 2/2] drm/i915: add syncobj timeline support X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: Intel graphics driver community testing & development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" X-Virus-Scanned: ClamAV using ClamSMTP Introduces a new parameters to execbuf so that we can specify syncobj handles as well as timeline points. This is needed for the submission side of the Vulkan timeline semaphore (VK_KHR_timeline_semaphore extension). Signed-off-by: Lionel Landwerlin --- drivers/gpu/drm/i915/i915_drv.c | 4 +- drivers/gpu/drm/i915/i915_gem_execbuffer.c | 190 ++++++++++++++++----- include/uapi/drm/i915_drm.h | 38 ++++- 3 files changed, 182 insertions(+), 50 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 83d2eb9e74cb..d62ddf2fa5b5 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -444,6 +444,7 @@ static int i915_getparam_ioctl(struct drm_device *dev, void *data, case I915_PARAM_HAS_EXEC_BATCH_FIRST: case I915_PARAM_HAS_EXEC_FENCE_ARRAY: case I915_PARAM_HAS_EXEC_SUBMIT_FENCE: + case I915_PARAM_HAS_EXEC_FENCE_ARRAY2: /* For the time being all of these are always true; * if some supported hardware does not have one of these * features this value needs to be provided from @@ -3175,7 +3176,8 @@ static struct drm_driver driver = { */ .driver_features = DRIVER_GEM | DRIVER_PRIME | - DRIVER_RENDER | DRIVER_MODESET | DRIVER_ATOMIC | DRIVER_SYNCOBJ, + DRIVER_RENDER | DRIVER_MODESET | DRIVER_ATOMIC | DRIVER_SYNCOBJ | + DRIVER_SYNCOBJ_TIMELINE, .release = i915_driver_release, .open = i915_driver_open, .lastclose = i915_driver_lastclose, diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index 8b85c91c3ea4..f681533c085a 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -229,6 +229,13 @@ enum { * the batchbuffer in trusted mode, otherwise the ioctl is rejected. */ +struct i915_drm_dma_fences { + struct drm_syncobj *syncobj; /* Use with ptr_mask_bits() */ + struct dma_fence *dma_fence; + u64 value; + struct dma_fence_chain *chain_fence; +}; + struct i915_execbuffer { struct drm_i915_private *i915; /** i915 backpointer */ struct drm_file *file; /** per-file lookup tables and limits */ @@ -1932,7 +1939,7 @@ static bool i915_gem_check_execbuffer(struct drm_i915_gem_execbuffer2 *exec) return false; /* Kernel clipping was a DRI1 misfeature */ - if (!(exec->flags & I915_EXEC_FENCE_ARRAY)) { + if (!(exec->flags & (I915_EXEC_FENCE_ARRAY | I915_EXEC_FENCE_ARRAY2))) { if (exec->num_cliprects || exec->cliprects_ptr) return false; } @@ -2182,25 +2189,30 @@ eb_select_engine(struct i915_execbuffer *eb, } static void -__free_fence_array(struct drm_syncobj **fences, unsigned int n) +__free_fence_array(struct i915_drm_dma_fences *fences, unsigned int n) { - while (n--) - drm_syncobj_put(ptr_mask_bits(fences[n], 2)); + while (n--) { + drm_syncobj_put(ptr_mask_bits(fences[n].syncobj, 2)); + dma_fence_put(fences[n].dma_fence); + kfree(fences[n].chain_fence); + } kvfree(fences); } -static struct drm_syncobj ** +static struct i915_drm_dma_fences * get_fence_array(struct drm_i915_gem_execbuffer2 *args, struct drm_file *file) { const unsigned long nfences = args->num_cliprects; struct drm_i915_gem_exec_fence __user *user; - struct drm_syncobj **fences; + struct drm_i915_gem_exec_fence2 __user *user2; + struct i915_drm_dma_fences *fences; unsigned long n; int err; - if (!(args->flags & I915_EXEC_FENCE_ARRAY)) - return NULL; + if ((args->flags & (I915_EXEC_FENCE_ARRAY | I915_EXEC_FENCE_ARRAY2)) == + (I915_EXEC_FENCE_ARRAY | I915_EXEC_FENCE_ARRAY2)) + return ERR_PTR(-EINVAL); /* Check multiplication overflow for access_ok() and kvmalloc_array() */ BUILD_BUG_ON(sizeof(size_t) > sizeof(unsigned long)); @@ -2209,40 +2221,121 @@ get_fence_array(struct drm_i915_gem_execbuffer2 *args, SIZE_MAX / sizeof(*fences))) return ERR_PTR(-EINVAL); - user = u64_to_user_ptr(args->cliprects_ptr); - if (!access_ok(user, nfences * sizeof(*user))) - return ERR_PTR(-EFAULT); + if (args->flags & I915_EXEC_FENCE_ARRAY2) { + user2 = u64_to_user_ptr(args->cliprects_ptr); + if (!access_ok(user, nfences * sizeof(*user2))) + return ERR_PTR(-EFAULT); + } else { + user = u64_to_user_ptr(args->cliprects_ptr); + if (!access_ok(user, nfences * sizeof(*user))) + return ERR_PTR(-EFAULT); + } fences = kvmalloc_array(nfences, sizeof(*fences), __GFP_NOWARN | GFP_KERNEL); if (!fences) return ERR_PTR(-ENOMEM); - for (n = 0; n < nfences; n++) { - struct drm_i915_gem_exec_fence fence; - struct drm_syncobj *syncobj; + BUILD_BUG_ON(~(ARCH_KMALLOC_MINALIGN - 1) & + ~__I915_EXEC_FENCE_UNKNOWN_FLAGS); - if (__copy_from_user(&fence, user++, sizeof(fence))) { - err = -EFAULT; - goto err; - } + if (args->flags & I915_EXEC_FENCE_ARRAY2) { + for (n = 0; n < nfences; n++) { + struct drm_i915_gem_exec_fence2 user_fence; + struct drm_syncobj *syncobj; + struct dma_fence *fence = NULL; - if (fence.flags & __I915_EXEC_FENCE_UNKNOWN_FLAGS) { - err = -EINVAL; - goto err; - } + if (__copy_from_user(&user_fence, user2++, sizeof(user_fence))) { + err = -EFAULT; + goto err; + } - syncobj = drm_syncobj_find(file, fence.handle); - if (!syncobj) { - DRM_DEBUG("Invalid syncobj handle provided\n"); - err = -ENOENT; - goto err; + if (user_fence.flags & __I915_EXEC_FENCE_UNKNOWN_FLAGS) { + err = -EINVAL; + goto err; + } + + if (user_fence.flags & I915_EXEC_FENCE_WAIT) { + err = drm_syncobj_find_fence( + file, user_fence.handle, user_fence.value, + DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT, + &syncobj, &fence); + if (err) { + DRM_DEBUG("Invalid syncobj handle or timeline value provided\n"); + goto err; + } + } else { + syncobj = drm_syncobj_find( + file, user_fence.handle); + if (!syncobj) { + err = -ENOENT; + DRM_DEBUG("Invalid syncobj handle provided\n"); + goto err; + } + } + + if (user_fence.value != 0 && user_fence.flags & I915_EXEC_FENCE_SIGNAL) { + fences[n].chain_fence = + kmalloc(sizeof(*fences[n].chain_fence), + GFP_KERNEL); + if (!fences[n].chain_fence) { + dma_fence_put(fence); + drm_syncobj_put(syncobj); + err = -ENOMEM; + DRM_DEBUG("Unable to alloc chain_fence\n"); + goto err; + } + } else { + fences[n].chain_fence = NULL; + } + + fences[n].syncobj = ptr_pack_bits(syncobj, user_fence.flags, 2); + fences[n].dma_fence = fence; + fences[n].value = user_fence.value; } + } else { + for (n = 0; n < nfences; n++) { + struct drm_i915_gem_exec_fence user_fence; + struct drm_syncobj *syncobj; + struct dma_fence *fence; + + if (__copy_from_user(&user_fence, user++, sizeof(user_fence))) { + err = -EFAULT; + goto err; + } + + if (user_fence.flags & __I915_EXEC_FENCE_UNKNOWN_FLAGS) { + err = -EINVAL; + goto err; + } - BUILD_BUG_ON(~(ARCH_KMALLOC_MINALIGN - 1) & - ~__I915_EXEC_FENCE_UNKNOWN_FLAGS); + /* If we're just signaling a syncobj, no need to get + * the fence. + */ + if (user_fence.flags & I915_EXEC_FENCE_WAIT) { + err = drm_syncobj_find_fence( + file, user_fence.handle, 0, 0, &syncobj, &fence); + if (err) { + DRM_DEBUG("Invalid syncobj handle provided\n"); + goto err; + } + } else { + syncobj = drm_syncobj_find(file, user_fence.handle); + if (!syncobj) { + DRM_DEBUG("Invalid syncobj handle provided\n"); + goto err; + } + fence = NULL; + } + + BUILD_BUG_ON(~(ARCH_KMALLOC_MINALIGN - 1) & + ~__I915_EXEC_FENCE_UNKNOWN_FLAGS); - fences[n] = ptr_pack_bits(syncobj, fence.flags, 2); + fences[n].syncobj = ptr_pack_bits(syncobj, user_fence.flags, 2); + fences[n].dma_fence = fence; + fences[n].value = 0; + fences[n].chain_fence = NULL; + } } return fences; @@ -2254,7 +2347,7 @@ get_fence_array(struct drm_i915_gem_execbuffer2 *args, static void put_fence_array(struct drm_i915_gem_execbuffer2 *args, - struct drm_syncobj **fences) + struct i915_drm_dma_fences *fences) { if (fences) __free_fence_array(fences, args->num_cliprects); @@ -2262,7 +2355,7 @@ put_fence_array(struct drm_i915_gem_execbuffer2 *args, static int await_fence_array(struct i915_execbuffer *eb, - struct drm_syncobj **fences) + struct i915_drm_dma_fences *fences) { const unsigned int nfences = eb->args->num_cliprects; unsigned int n; @@ -2270,19 +2363,14 @@ await_fence_array(struct i915_execbuffer *eb, for (n = 0; n < nfences; n++) { struct drm_syncobj *syncobj; - struct dma_fence *fence; unsigned int flags; - syncobj = ptr_unpack_bits(fences[n], &flags, 2); + syncobj = ptr_unpack_bits(fences[n].syncobj, &flags, 2); if (!(flags & I915_EXEC_FENCE_WAIT)) continue; - fence = drm_syncobj_fence_get(syncobj); - if (!fence) - return -EINVAL; - - err = i915_request_await_dma_fence(eb->request, fence); - dma_fence_put(fence); + err = i915_request_await_dma_fence(eb->request, + fences[n].dma_fence); if (err < 0) return err; } @@ -2292,7 +2380,7 @@ await_fence_array(struct i915_execbuffer *eb, static void signal_fence_array(struct i915_execbuffer *eb, - struct drm_syncobj **fences) + struct i915_drm_dma_fences *fences) { const unsigned int nfences = eb->args->num_cliprects; struct dma_fence * const fence = &eb->request->fence; @@ -2302,11 +2390,21 @@ signal_fence_array(struct i915_execbuffer *eb, struct drm_syncobj *syncobj; unsigned int flags; - syncobj = ptr_unpack_bits(fences[n], &flags, 2); + syncobj = ptr_unpack_bits(fences[n].syncobj, &flags, 2); if (!(flags & I915_EXEC_FENCE_SIGNAL)) continue; - drm_syncobj_replace_fence(syncobj, fence); + if (fences[n].chain_fence) { + drm_syncobj_add_point(syncobj, fences[n].chain_fence, + fence, fences[n].value); + /* + * The chain's ownership is transfered to the + * timeline. + */ + fences[n].chain_fence = NULL; + } else { + drm_syncobj_replace_fence(syncobj, fence); + } } } @@ -2315,7 +2413,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, struct drm_file *file, struct drm_i915_gem_execbuffer2 *args, struct drm_i915_gem_exec_object2 *exec, - struct drm_syncobj **fences) + struct i915_drm_dma_fences *fences) { struct i915_execbuffer eb; struct dma_fence *in_fence = NULL; @@ -2705,7 +2803,7 @@ i915_gem_execbuffer2_ioctl(struct drm_device *dev, void *data, { struct drm_i915_gem_execbuffer2 *args = data; struct drm_i915_gem_exec_object2 *exec2_list; - struct drm_syncobj **fences = NULL; + struct i915_drm_dma_fences *fences = NULL; const size_t count = args->buffer_count; int err; @@ -2733,7 +2831,7 @@ i915_gem_execbuffer2_ioctl(struct drm_device *dev, void *data, return -EFAULT; } - if (args->flags & I915_EXEC_FENCE_ARRAY) { + if (args->flags & (I915_EXEC_FENCE_ARRAY | I915_EXEC_FENCE_ARRAY2)) { fences = get_fence_array(args, file); if (IS_ERR(fences)) { kvfree(exec2_list); diff --git a/include/uapi/drm/i915_drm.h b/include/uapi/drm/i915_drm.h index 328d05e77d9f..eaf7b89360fd 100644 --- a/include/uapi/drm/i915_drm.h +++ b/include/uapi/drm/i915_drm.h @@ -610,6 +610,12 @@ typedef struct drm_i915_irq_wait { * See I915_EXEC_FENCE_OUT and I915_EXEC_FENCE_SUBMIT. */ #define I915_PARAM_HAS_EXEC_SUBMIT_FENCE 53 + +/* Query whether DRM_I915_GEM_EXECBUFFER2 supports supplying an array of + * drm_i915_gem_exec_fence2 structures. See I915_EXEC_FENCE_ARRAY. + */ +#define I915_PARAM_HAS_EXEC_FENCE_ARRAY2 54 + /* Must be kept compact -- no holes and well documented */ typedef struct drm_i915_getparam { @@ -1006,6 +1012,24 @@ struct drm_i915_gem_exec_fence { __u32 flags; }; +struct drm_i915_gem_exec_fence2 { + /** + * User's handle for a drm_syncobj to wait on or signal. + */ + __u32 handle; + + /** + * Same flags as drm_i915_gem_exec_fence. + */ + __u32 flags; + + /** + * A point for a timeline drm_syncobj to wait on or signal. Must be 0 + * for a binary drm_syncobj. + */ + __u64 value; +}; + struct drm_i915_gem_execbuffer2 { /** * List of gem_exec_object2 structs @@ -1022,8 +1046,10 @@ struct drm_i915_gem_execbuffer2 { __u32 num_cliprects; /** * This is a struct drm_clip_rect *cliprects if I915_EXEC_FENCE_ARRAY - * is not set. If I915_EXEC_FENCE_ARRAY is set, then this is a - * struct drm_i915_gem_exec_fence *fences. + * & I915_EXEC_FENCE_ARRAY2 are not set. If I915_EXEC_FENCE_ARRAY is + * set, then this is a struct drm_i915_gem_exec_fence *fences. If + * I915_EXEC_FENCE_ARRAY2 is set, then this is a struct + * drm_i915_gem_exec_fence2 *fences. */ __u64 cliprects_ptr; #define I915_EXEC_RING_MASK (0x3f) @@ -1141,7 +1167,13 @@ struct drm_i915_gem_execbuffer2 { */ #define I915_EXEC_FENCE_SUBMIT (1 << 20) -#define __I915_EXEC_UNKNOWN_FLAGS (-(I915_EXEC_FENCE_SUBMIT << 1)) +/* Setting I915_FENCE_ARRAY2 implies that num_cliprects and cliprects_ptr + * define an array of i915_gem_exec_fence2 structures which specify a set of + * dma fences to wait upon or signal. + */ +#define I915_EXEC_FENCE_ARRAY2 (1<<22) + +#define __I915_EXEC_UNKNOWN_FLAGS (-(I915_EXEC_FENCE_ARRAY2<<1)) #define I915_EXEC_CONTEXT_ID_MASK (0xffffffff) #define i915_execbuffer2_set_context_id(eb2, context) \