From patchwork Mon Sep 4 17:38:49 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374234 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 B9451C71153 for ; Mon, 4 Sep 2023 17:51:16 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id E604E10E3BA; Mon, 4 Sep 2023 17:51:15 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id A0CAB10E15A for ; Mon, 4 Sep 2023 17:51:14 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=sZTNbm/Fp0+1ZsHY5TRMflwtPPUmX+vlibf8omYEz9c=; b=XPCq/M5IKghZsnxW94PePpQBQ/ DgakGpAk3RsCTartoaagQBUmA5aAle2o9AXqygitYOQjcm19huouzTeMNLD0Z4KsvLLKRTWMBjp5x vAG0cElFZ9SsB+yBw6cDUdwjHRYjNh57SzgG/oP7zIsRrtRtBdKAlRqP5K5BRvIXsl8P781l+9qs4 JoTvbxoo2sR0a22/2O644pk0n/rDEZJ2WoSusxc9Kj+B+7xAInnnwtnfNcsb7PP3MwvxNNLKvNYfG ZE5kexxtL90hjr0wMdfYTsF1HI1rUWnUkvw85NUIw3hOB3QoL2x0X4w0CEOLAQOxtXnSbFPzOzRJY uJ+LPWbQ==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDj1-0028nB-AV; Mon, 04 Sep 2023 19:50:59 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 01/17] drm/v3d: Remove unused function header Date: Mon, 4 Sep 2023 14:38:49 -0300 Message-ID: <20230904175019.1172713-2-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Melissa Wen v3d_mmu_get_offset header was added but the function was never defined. Just remove it. Signed-off-by: Melissa Wen Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_drv.h | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/gpu/drm/v3d/v3d_drv.h b/drivers/gpu/drm/v3d/v3d_drv.h index 7f664a4b2a75..e8e45a9c705f 100644 --- a/drivers/gpu/drm/v3d/v3d_drv.h +++ b/drivers/gpu/drm/v3d/v3d_drv.h @@ -393,8 +393,6 @@ void v3d_irq_disable(struct v3d_dev *v3d); void v3d_irq_reset(struct v3d_dev *v3d); /* v3d_mmu.c */ -int v3d_mmu_get_offset(struct drm_file *file_priv, struct v3d_bo *bo, - u32 *offset); int v3d_mmu_set_page_table(struct v3d_dev *v3d); void v3d_mmu_insert_ptes(struct v3d_bo *bo); void v3d_mmu_remove_ptes(struct v3d_bo *bo); From patchwork Mon Sep 4 17:38:50 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374235 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 BF0CFC71153 for ; Mon, 4 Sep 2023 17:51:21 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 6B86810E3BB; Mon, 4 Sep 2023 17:51:16 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id A132510E3BA for ; Mon, 4 Sep 2023 17:51:14 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=kLznYpGmudP86IXpMuG57/6/Q2cv3NvzkHXbRWUAHb0=; b=UWl1938QNdjeqxR8dGvyIdl7h7 +Ne6+BKTQHNKlYUzxn4db0k8I0tZsoJ+44LGjGUgjK7J12RykvfbR9VlU9oxjAPAXFBm8rOEuwUCX 5mO4GuOzt8XPGXisV606NbMzfFJpLm3EtM3B8h18LehLfgoNC4b5xfyZ10RjbwSl3TdSoGCSBwO3b LUb/ViLeYRFLseBlEnMhoHIUaWlh1wFQRbuZHv00J+bM2/bYOPghVndi0vwiY1LBPkYnYszgKy21A V0PyGMeqZu/j28Dpk+FUddL4s36SShFYQT3fYRWOwyyb4nK4SuVxh5CxyI0Ofg7jQP6IVy9dpNsvK F0d+h47A==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDj5-0028nB-Gt; Mon, 04 Sep 2023 19:51:04 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 02/17] drm/v3d: Move wait BO ioctl to the v3d_bo file Date: Mon, 4 Sep 2023 14:38:50 -0300 Message-ID: <20230904175019.1172713-3-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Melissa Wen IOCTLs related to BO operations reside on the file v3d_bo.c. The wait BO ioctl is the only IOCTL regarding BOs that is placed in a different file. So, move it to the v3d_bo.c file. Signed-off-by: Melissa Wen Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_bo.c | 33 +++++++++++++++++++++++++++++++++ drivers/gpu/drm/v3d/v3d_drv.h | 4 ++-- drivers/gpu/drm/v3d/v3d_gem.c | 33 --------------------------------- 3 files changed, 35 insertions(+), 35 deletions(-) diff --git a/drivers/gpu/drm/v3d/v3d_bo.c b/drivers/gpu/drm/v3d/v3d_bo.c index 8b3229a37c6d..357a0da7e16a 100644 --- a/drivers/gpu/drm/v3d/v3d_bo.c +++ b/drivers/gpu/drm/v3d/v3d_bo.c @@ -233,3 +233,36 @@ int v3d_get_bo_offset_ioctl(struct drm_device *dev, void *data, drm_gem_object_put(gem_obj); return 0; } + +int +v3d_wait_bo_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + int ret; + struct drm_v3d_wait_bo *args = data; + ktime_t start = ktime_get(); + u64 delta_ns; + unsigned long timeout_jiffies = + nsecs_to_jiffies_timeout(args->timeout_ns); + + if (args->pad != 0) + return -EINVAL; + + ret = drm_gem_dma_resv_wait(file_priv, args->handle, + true, timeout_jiffies); + + /* Decrement the user's timeout, in case we got interrupted + * such that the ioctl will be restarted. + */ + delta_ns = ktime_to_ns(ktime_sub(ktime_get(), start)); + if (delta_ns < args->timeout_ns) + args->timeout_ns -= delta_ns; + else + args->timeout_ns = 0; + + /* Asked to wait beyond the jiffie/scheduler precision? */ + if (ret == -ETIME && args->timeout_ns) + ret = -EAGAIN; + + return ret; +} diff --git a/drivers/gpu/drm/v3d/v3d_drv.h b/drivers/gpu/drm/v3d/v3d_drv.h index e8e45a9c705f..8257fb64506c 100644 --- a/drivers/gpu/drm/v3d/v3d_drv.h +++ b/drivers/gpu/drm/v3d/v3d_drv.h @@ -358,6 +358,8 @@ int v3d_mmap_bo_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); int v3d_get_bo_offset_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); +int v3d_wait_bo_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); struct drm_gem_object *v3d_prime_import_sg_table(struct drm_device *dev, struct dma_buf_attachment *attach, struct sg_table *sgt); @@ -378,8 +380,6 @@ int v3d_submit_tfu_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); int v3d_submit_csd_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); -int v3d_wait_bo_ioctl(struct drm_device *dev, void *data, - struct drm_file *file_priv); void v3d_job_cleanup(struct v3d_job *job); void v3d_job_put(struct v3d_job *job); void v3d_reset(struct v3d_dev *v3d); diff --git a/drivers/gpu/drm/v3d/v3d_gem.c b/drivers/gpu/drm/v3d/v3d_gem.c index 2e94ce788c71..e6f4b7ffd958 100644 --- a/drivers/gpu/drm/v3d/v3d_gem.c +++ b/drivers/gpu/drm/v3d/v3d_gem.c @@ -363,39 +363,6 @@ void v3d_job_put(struct v3d_job *job) kref_put(&job->refcount, job->free); } -int -v3d_wait_bo_ioctl(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - int ret; - struct drm_v3d_wait_bo *args = data; - ktime_t start = ktime_get(); - u64 delta_ns; - unsigned long timeout_jiffies = - nsecs_to_jiffies_timeout(args->timeout_ns); - - if (args->pad != 0) - return -EINVAL; - - ret = drm_gem_dma_resv_wait(file_priv, args->handle, - true, timeout_jiffies); - - /* Decrement the user's timeout, in case we got interrupted - * such that the ioctl will be restarted. - */ - delta_ns = ktime_to_ns(ktime_sub(ktime_get(), start)); - if (delta_ns < args->timeout_ns) - args->timeout_ns -= delta_ns; - else - args->timeout_ns = 0; - - /* Asked to wait beyond the jiffie/scheduler precision? */ - if (ret == -ETIME && args->timeout_ns) - ret = -EAGAIN; - - return ret; -} - static int v3d_job_init(struct v3d_dev *v3d, struct drm_file *file_priv, void **container, size_t size, void (*free)(struct kref *ref), From patchwork Mon Sep 4 17:38:51 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374238 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 9F99BC83F2C for ; Mon, 4 Sep 2023 17:51:26 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 6E21210E3BD; Mon, 4 Sep 2023 17:51:19 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id 98DDF10E3BD for ; Mon, 4 Sep 2023 17:51:16 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=NCq95LsAtABUcGp0NbvBMhFJ9OazHSTy0mzolFGUA8w=; b=Xc9ZETx9xtkY5BZj252AB6FAPt T+ehmLugQPgpSa2rVwGWyT3TQTQ/AnD/BpYSQpjLJSTRdiOLiM8xhbrmJ6JgQMQHtmZ108vMyHhl5 gwMrhF0PqThXmD6HLaiJvIdN5yvAmHUz0C+fsNpzR6RdBJ4eiLh7JifbJNtZnAMbGqOpnq7VF98IQ /B/m+h3pUJWKf2SFXPDBPzxMyC+tAfT0sP1fWS7JjEqIR9WMUzCjpfTMtqirBcT+VYJ7QbFQhVH6C iB2EBneGY0lutgVo8Ht6j5x2wOZq2rxomhJd3P0JIr66cKBInufg81vXGPCDU8hKjGaFYhTFSfV3Q McUy6WcQ==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDj9-0028nB-MZ; Mon, 04 Sep 2023 19:51:08 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 03/17] drm/v3d: Detach job submissions IOCTLs to a new specific file Date: Mon, 4 Sep 2023 14:38:51 -0300 Message-ID: <20230904175019.1172713-4-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Melissa Wen We will include a new job submission type, the CPU job submission. For readability and maintability, separate the job submission IOCTLs and related operations from v3d_gem.c. Minor fix in the CSD submission kernel doc: CSD (texture formatting) -> CSD (compute shader). Signed-off-by: Melissa Wen Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/Makefile | 3 +- drivers/gpu/drm/v3d/v3d_drv.h | 12 +- drivers/gpu/drm/v3d/v3d_gem.c | 734 ------------------------------ drivers/gpu/drm/v3d/v3d_submit.c | 743 +++++++++++++++++++++++++++++++ 4 files changed, 752 insertions(+), 740 deletions(-) create mode 100644 drivers/gpu/drm/v3d/v3d_submit.c -- 2.41.0 diff --git a/drivers/gpu/drm/v3d/Makefile b/drivers/gpu/drm/v3d/Makefile index e8b314137020..f3b15a014ee4 100644 --- a/drivers/gpu/drm/v3d/Makefile +++ b/drivers/gpu/drm/v3d/Makefile @@ -11,7 +11,8 @@ v3d-y := \ v3d_mmu.o \ v3d_perfmon.o \ v3d_trace_points.o \ - v3d_sched.o + v3d_sched.o \ + v3d_submit.o v3d-$(CONFIG_DEBUG_FS) += v3d_debugfs.o diff --git a/drivers/gpu/drm/v3d/v3d_drv.h b/drivers/gpu/drm/v3d/v3d_drv.h index 8257fb64506c..722a627e0a6e 100644 --- a/drivers/gpu/drm/v3d/v3d_drv.h +++ b/drivers/gpu/drm/v3d/v3d_drv.h @@ -374,17 +374,19 @@ struct dma_fence *v3d_fence_create(struct v3d_dev *v3d, enum v3d_queue queue); /* v3d_gem.c */ int v3d_gem_init(struct drm_device *dev); void v3d_gem_destroy(struct drm_device *dev); +void v3d_reset(struct v3d_dev *v3d); +void v3d_invalidate_caches(struct v3d_dev *v3d); +void v3d_clean_caches(struct v3d_dev *v3d); + +/* v3d_submit.c */ +void v3d_job_cleanup(struct v3d_job *job); +void v3d_job_put(struct v3d_job *job); int v3d_submit_cl_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); int v3d_submit_tfu_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); int v3d_submit_csd_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); -void v3d_job_cleanup(struct v3d_job *job); -void v3d_job_put(struct v3d_job *job); -void v3d_reset(struct v3d_dev *v3d); -void v3d_invalidate_caches(struct v3d_dev *v3d); -void v3d_clean_caches(struct v3d_dev *v3d); /* v3d_irq.c */ int v3d_irq_init(struct v3d_dev *v3d); diff --git a/drivers/gpu/drm/v3d/v3d_gem.c b/drivers/gpu/drm/v3d/v3d_gem.c index e6f4b7ffd958..63378004272b 100644 --- a/drivers/gpu/drm/v3d/v3d_gem.c +++ b/drivers/gpu/drm/v3d/v3d_gem.c @@ -11,8 +11,6 @@ #include #include -#include -#include #include "v3d_drv.h" #include "v3d_regs.h" @@ -241,738 +239,6 @@ v3d_invalidate_caches(struct v3d_dev *v3d) v3d_invalidate_slices(v3d, 0); } -/* Takes the reservation lock on all the BOs being referenced, so that - * at queue submit time we can update the reservations. - * - * We don't lock the RCL the tile alloc/state BOs, or overflow memory - * (all of which are on exec->unref_list). They're entirely private - * to v3d, so we don't attach dma-buf fences to them. - */ -static int -v3d_lock_bo_reservations(struct v3d_job *job, - struct ww_acquire_ctx *acquire_ctx) -{ - int i, ret; - - ret = drm_gem_lock_reservations(job->bo, job->bo_count, acquire_ctx); - if (ret) - return ret; - - for (i = 0; i < job->bo_count; i++) { - ret = dma_resv_reserve_fences(job->bo[i]->resv, 1); - if (ret) - goto fail; - - ret = drm_sched_job_add_implicit_dependencies(&job->base, - job->bo[i], true); - if (ret) - goto fail; - } - - return 0; - -fail: - drm_gem_unlock_reservations(job->bo, job->bo_count, acquire_ctx); - return ret; -} - -/** - * v3d_lookup_bos() - Sets up job->bo[] with the GEM objects - * referenced by the job. - * @dev: DRM device - * @file_priv: DRM file for this fd - * @job: V3D job being set up - * @bo_handles: GEM handles - * @bo_count: Number of GEM handles passed in - * - * The command validator needs to reference BOs by their index within - * the submitted job's BO list. This does the validation of the job's - * BO list and reference counting for the lifetime of the job. - * - * Note that this function doesn't need to unreference the BOs on - * failure, because that will happen at v3d_exec_cleanup() time. - */ -static int -v3d_lookup_bos(struct drm_device *dev, - struct drm_file *file_priv, - struct v3d_job *job, - u64 bo_handles, - u32 bo_count) -{ - job->bo_count = bo_count; - - if (!job->bo_count) { - /* See comment on bo_index for why we have to check - * this. - */ - DRM_DEBUG("Rendering requires BOs\n"); - return -EINVAL; - } - - return drm_gem_objects_lookup(file_priv, - (void __user *)(uintptr_t)bo_handles, - job->bo_count, &job->bo); -} - -static void -v3d_job_free(struct kref *ref) -{ - struct v3d_job *job = container_of(ref, struct v3d_job, refcount); - int i; - - if (job->bo) { - for (i = 0; i < job->bo_count; i++) - drm_gem_object_put(job->bo[i]); - kvfree(job->bo); - } - - dma_fence_put(job->irq_fence); - dma_fence_put(job->done_fence); - - if (job->perfmon) - v3d_perfmon_put(job->perfmon); - - kfree(job); -} - -static void -v3d_render_job_free(struct kref *ref) -{ - struct v3d_render_job *job = container_of(ref, struct v3d_render_job, - base.refcount); - struct v3d_bo *bo, *save; - - list_for_each_entry_safe(bo, save, &job->unref_list, unref_head) { - drm_gem_object_put(&bo->base.base); - } - - v3d_job_free(ref); -} - -void v3d_job_cleanup(struct v3d_job *job) -{ - if (!job) - return; - - drm_sched_job_cleanup(&job->base); - v3d_job_put(job); -} - -void v3d_job_put(struct v3d_job *job) -{ - kref_put(&job->refcount, job->free); -} - -static int -v3d_job_init(struct v3d_dev *v3d, struct drm_file *file_priv, - void **container, size_t size, void (*free)(struct kref *ref), - u32 in_sync, struct v3d_submit_ext *se, enum v3d_queue queue) -{ - struct v3d_file_priv *v3d_priv = file_priv->driver_priv; - struct v3d_job *job; - bool has_multisync = se && (se->flags & DRM_V3D_EXT_ID_MULTI_SYNC); - int ret, i; - - *container = kcalloc(1, size, GFP_KERNEL); - if (!*container) { - DRM_ERROR("Cannot allocate memory for v3d job."); - return -ENOMEM; - } - - job = *container; - job->v3d = v3d; - job->free = free; - - ret = drm_sched_job_init(&job->base, &v3d_priv->sched_entity[queue], - v3d_priv); - if (ret) - goto fail; - - if (has_multisync) { - if (se->in_sync_count && se->wait_stage == queue) { - struct drm_v3d_sem __user *handle = u64_to_user_ptr(se->in_syncs); - - for (i = 0; i < se->in_sync_count; i++) { - struct drm_v3d_sem in; - - if (copy_from_user(&in, handle++, sizeof(in))) { - ret = -EFAULT; - DRM_DEBUG("Failed to copy wait dep handle.\n"); - goto fail_deps; - } - ret = drm_sched_job_add_syncobj_dependency(&job->base, file_priv, in.handle, 0); - - // TODO: Investigate why this was filtered out for the IOCTL. - if (ret && ret != -ENOENT) - goto fail_deps; - } - } - } else { - ret = drm_sched_job_add_syncobj_dependency(&job->base, file_priv, in_sync, 0); - - // TODO: Investigate why this was filtered out for the IOCTL. - if (ret && ret != -ENOENT) - goto fail_deps; - } - - kref_init(&job->refcount); - - return 0; - -fail_deps: - drm_sched_job_cleanup(&job->base); -fail: - kfree(*container); - *container = NULL; - - return ret; -} - -static void -v3d_push_job(struct v3d_job *job) -{ - drm_sched_job_arm(&job->base); - - job->done_fence = dma_fence_get(&job->base.s_fence->finished); - - /* put by scheduler job completion */ - kref_get(&job->refcount); - - drm_sched_entity_push_job(&job->base); -} - -static void -v3d_attach_fences_and_unlock_reservation(struct drm_file *file_priv, - struct v3d_job *job, - struct ww_acquire_ctx *acquire_ctx, - u32 out_sync, - struct v3d_submit_ext *se, - struct dma_fence *done_fence) -{ - struct drm_syncobj *sync_out; - bool has_multisync = se && (se->flags & DRM_V3D_EXT_ID_MULTI_SYNC); - int i; - - for (i = 0; i < job->bo_count; i++) { - /* XXX: Use shared fences for read-only objects. */ - dma_resv_add_fence(job->bo[i]->resv, job->done_fence, - DMA_RESV_USAGE_WRITE); - } - - drm_gem_unlock_reservations(job->bo, job->bo_count, acquire_ctx); - - /* Update the return sync object for the job */ - /* If it only supports a single signal semaphore*/ - if (!has_multisync) { - sync_out = drm_syncobj_find(file_priv, out_sync); - if (sync_out) { - drm_syncobj_replace_fence(sync_out, done_fence); - drm_syncobj_put(sync_out); - } - return; - } - - /* If multiple semaphores extension is supported */ - if (se->out_sync_count) { - for (i = 0; i < se->out_sync_count; i++) { - drm_syncobj_replace_fence(se->out_syncs[i].syncobj, - done_fence); - drm_syncobj_put(se->out_syncs[i].syncobj); - } - kvfree(se->out_syncs); - } -} - -static void -v3d_put_multisync_post_deps(struct v3d_submit_ext *se) -{ - unsigned int i; - - if (!(se && se->out_sync_count)) - return; - - for (i = 0; i < se->out_sync_count; i++) - drm_syncobj_put(se->out_syncs[i].syncobj); - kvfree(se->out_syncs); -} - -static int -v3d_get_multisync_post_deps(struct drm_file *file_priv, - struct v3d_submit_ext *se, - u32 count, u64 handles) -{ - struct drm_v3d_sem __user *post_deps; - int i, ret; - - if (!count) - return 0; - - se->out_syncs = (struct v3d_submit_outsync *) - kvmalloc_array(count, - sizeof(struct v3d_submit_outsync), - GFP_KERNEL); - if (!se->out_syncs) - return -ENOMEM; - - post_deps = u64_to_user_ptr(handles); - - for (i = 0; i < count; i++) { - struct drm_v3d_sem out; - - if (copy_from_user(&out, post_deps++, sizeof(out))) { - ret = -EFAULT; - DRM_DEBUG("Failed to copy post dep handles\n"); - goto fail; - } - - se->out_syncs[i].syncobj = drm_syncobj_find(file_priv, - out.handle); - if (!se->out_syncs[i].syncobj) { - ret = -EINVAL; - goto fail; - } - } - se->out_sync_count = count; - - return 0; - -fail: - for (i--; i >= 0; i--) - drm_syncobj_put(se->out_syncs[i].syncobj); - kvfree(se->out_syncs); - - return ret; -} - -/* Get data for multiple binary semaphores synchronization. Parse syncobj - * to be signaled when job completes (out_sync). - */ -static int -v3d_get_multisync_submit_deps(struct drm_file *file_priv, - struct drm_v3d_extension __user *ext, - void *data) -{ - struct drm_v3d_multi_sync multisync; - struct v3d_submit_ext *se = data; - int ret; - - if (copy_from_user(&multisync, ext, sizeof(multisync))) - return -EFAULT; - - if (multisync.pad) - return -EINVAL; - - ret = v3d_get_multisync_post_deps(file_priv, data, multisync.out_sync_count, - multisync.out_syncs); - if (ret) - return ret; - - se->in_sync_count = multisync.in_sync_count; - se->in_syncs = multisync.in_syncs; - se->flags |= DRM_V3D_EXT_ID_MULTI_SYNC; - se->wait_stage = multisync.wait_stage; - - return 0; -} - -/* Whenever userspace sets ioctl extensions, v3d_get_extensions parses data - * according to the extension id (name). - */ -static int -v3d_get_extensions(struct drm_file *file_priv, - u64 ext_handles, - void *data) -{ - struct drm_v3d_extension __user *user_ext; - int ret; - - user_ext = u64_to_user_ptr(ext_handles); - while (user_ext) { - struct drm_v3d_extension ext; - - if (copy_from_user(&ext, user_ext, sizeof(ext))) { - DRM_DEBUG("Failed to copy submit extension\n"); - return -EFAULT; - } - - switch (ext.id) { - case DRM_V3D_EXT_ID_MULTI_SYNC: - ret = v3d_get_multisync_submit_deps(file_priv, user_ext, data); - if (ret) - return ret; - break; - default: - DRM_DEBUG_DRIVER("Unknown extension id: %d\n", ext.id); - return -EINVAL; - } - - user_ext = u64_to_user_ptr(ext.next); - } - - return 0; -} - -/** - * v3d_submit_cl_ioctl() - Submits a job (frame) to the V3D. - * @dev: DRM device - * @data: ioctl argument - * @file_priv: DRM file for this fd - * - * This is the main entrypoint for userspace to submit a 3D frame to - * the GPU. Userspace provides the binner command list (if - * applicable), and the kernel sets up the render command list to draw - * to the framebuffer described in the ioctl, using the command lists - * that the 3D engine's binner will produce. - */ -int -v3d_submit_cl_ioctl(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - struct v3d_dev *v3d = to_v3d_dev(dev); - struct v3d_file_priv *v3d_priv = file_priv->driver_priv; - struct drm_v3d_submit_cl *args = data; - struct v3d_submit_ext se = {0}; - struct v3d_bin_job *bin = NULL; - struct v3d_render_job *render = NULL; - struct v3d_job *clean_job = NULL; - struct v3d_job *last_job; - struct ww_acquire_ctx acquire_ctx; - int ret = 0; - - trace_v3d_submit_cl_ioctl(&v3d->drm, args->rcl_start, args->rcl_end); - - if (args->pad) - return -EINVAL; - - if (args->flags && - args->flags & ~(DRM_V3D_SUBMIT_CL_FLUSH_CACHE | - DRM_V3D_SUBMIT_EXTENSION)) { - DRM_INFO("invalid flags: %d\n", args->flags); - return -EINVAL; - } - - if (args->flags & DRM_V3D_SUBMIT_EXTENSION) { - ret = v3d_get_extensions(file_priv, args->extensions, &se); - if (ret) { - DRM_DEBUG("Failed to get extensions.\n"); - return ret; - } - } - - ret = v3d_job_init(v3d, file_priv, (void *)&render, sizeof(*render), - v3d_render_job_free, args->in_sync_rcl, &se, V3D_RENDER); - if (ret) - goto fail; - - render->start = args->rcl_start; - render->end = args->rcl_end; - INIT_LIST_HEAD(&render->unref_list); - - if (args->bcl_start != args->bcl_end) { - ret = v3d_job_init(v3d, file_priv, (void *)&bin, sizeof(*bin), - v3d_job_free, args->in_sync_bcl, &se, V3D_BIN); - if (ret) - goto fail; - - bin->start = args->bcl_start; - bin->end = args->bcl_end; - bin->qma = args->qma; - bin->qms = args->qms; - bin->qts = args->qts; - bin->render = render; - } - - if (args->flags & DRM_V3D_SUBMIT_CL_FLUSH_CACHE) { - ret = v3d_job_init(v3d, file_priv, (void *)&clean_job, sizeof(*clean_job), - v3d_job_free, 0, NULL, V3D_CACHE_CLEAN); - if (ret) - goto fail; - - last_job = clean_job; - } else { - last_job = &render->base; - } - - ret = v3d_lookup_bos(dev, file_priv, last_job, - args->bo_handles, args->bo_handle_count); - if (ret) - goto fail; - - ret = v3d_lock_bo_reservations(last_job, &acquire_ctx); - if (ret) - goto fail; - - if (args->perfmon_id) { - render->base.perfmon = v3d_perfmon_find(v3d_priv, - args->perfmon_id); - - if (!render->base.perfmon) { - ret = -ENOENT; - goto fail_perfmon; - } - } - - mutex_lock(&v3d->sched_lock); - if (bin) { - bin->base.perfmon = render->base.perfmon; - v3d_perfmon_get(bin->base.perfmon); - v3d_push_job(&bin->base); - - ret = drm_sched_job_add_dependency(&render->base.base, - dma_fence_get(bin->base.done_fence)); - if (ret) - goto fail_unreserve; - } - - v3d_push_job(&render->base); - - if (clean_job) { - struct dma_fence *render_fence = - dma_fence_get(render->base.done_fence); - ret = drm_sched_job_add_dependency(&clean_job->base, - render_fence); - if (ret) - goto fail_unreserve; - clean_job->perfmon = render->base.perfmon; - v3d_perfmon_get(clean_job->perfmon); - v3d_push_job(clean_job); - } - - mutex_unlock(&v3d->sched_lock); - - v3d_attach_fences_and_unlock_reservation(file_priv, - last_job, - &acquire_ctx, - args->out_sync, - &se, - last_job->done_fence); - - if (bin) - v3d_job_put(&bin->base); - v3d_job_put(&render->base); - if (clean_job) - v3d_job_put(clean_job); - - return 0; - -fail_unreserve: - mutex_unlock(&v3d->sched_lock); -fail_perfmon: - drm_gem_unlock_reservations(last_job->bo, - last_job->bo_count, &acquire_ctx); -fail: - v3d_job_cleanup((void *)bin); - v3d_job_cleanup((void *)render); - v3d_job_cleanup(clean_job); - v3d_put_multisync_post_deps(&se); - - return ret; -} - -/** - * v3d_submit_tfu_ioctl() - Submits a TFU (texture formatting) job to the V3D. - * @dev: DRM device - * @data: ioctl argument - * @file_priv: DRM file for this fd - * - * Userspace provides the register setup for the TFU, which we don't - * need to validate since the TFU is behind the MMU. - */ -int -v3d_submit_tfu_ioctl(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - struct v3d_dev *v3d = to_v3d_dev(dev); - struct drm_v3d_submit_tfu *args = data; - struct v3d_submit_ext se = {0}; - struct v3d_tfu_job *job = NULL; - struct ww_acquire_ctx acquire_ctx; - int ret = 0; - - trace_v3d_submit_tfu_ioctl(&v3d->drm, args->iia); - - if (args->flags && !(args->flags & DRM_V3D_SUBMIT_EXTENSION)) { - DRM_DEBUG("invalid flags: %d\n", args->flags); - return -EINVAL; - } - - if (args->flags & DRM_V3D_SUBMIT_EXTENSION) { - ret = v3d_get_extensions(file_priv, args->extensions, &se); - if (ret) { - DRM_DEBUG("Failed to get extensions.\n"); - return ret; - } - } - - ret = v3d_job_init(v3d, file_priv, (void *)&job, sizeof(*job), - v3d_job_free, args->in_sync, &se, V3D_TFU); - if (ret) - goto fail; - - job->base.bo = kcalloc(ARRAY_SIZE(args->bo_handles), - sizeof(*job->base.bo), GFP_KERNEL); - if (!job->base.bo) { - ret = -ENOMEM; - goto fail; - } - - job->args = *args; - - for (job->base.bo_count = 0; - job->base.bo_count < ARRAY_SIZE(args->bo_handles); - job->base.bo_count++) { - struct drm_gem_object *bo; - - if (!args->bo_handles[job->base.bo_count]) - break; - - bo = drm_gem_object_lookup(file_priv, args->bo_handles[job->base.bo_count]); - if (!bo) { - DRM_DEBUG("Failed to look up GEM BO %d: %d\n", - job->base.bo_count, - args->bo_handles[job->base.bo_count]); - ret = -ENOENT; - goto fail; - } - job->base.bo[job->base.bo_count] = bo; - } - - ret = v3d_lock_bo_reservations(&job->base, &acquire_ctx); - if (ret) - goto fail; - - mutex_lock(&v3d->sched_lock); - v3d_push_job(&job->base); - mutex_unlock(&v3d->sched_lock); - - v3d_attach_fences_and_unlock_reservation(file_priv, - &job->base, &acquire_ctx, - args->out_sync, - &se, - job->base.done_fence); - - v3d_job_put(&job->base); - - return 0; - -fail: - v3d_job_cleanup((void *)job); - v3d_put_multisync_post_deps(&se); - - return ret; -} - -/** - * v3d_submit_csd_ioctl() - Submits a CSD (texture formatting) job to the V3D. - * @dev: DRM device - * @data: ioctl argument - * @file_priv: DRM file for this fd - * - * Userspace provides the register setup for the CSD, which we don't - * need to validate since the CSD is behind the MMU. - */ -int -v3d_submit_csd_ioctl(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - struct v3d_dev *v3d = to_v3d_dev(dev); - struct v3d_file_priv *v3d_priv = file_priv->driver_priv; - struct drm_v3d_submit_csd *args = data; - struct v3d_submit_ext se = {0}; - struct v3d_csd_job *job = NULL; - struct v3d_job *clean_job = NULL; - struct ww_acquire_ctx acquire_ctx; - int ret; - - trace_v3d_submit_csd_ioctl(&v3d->drm, args->cfg[5], args->cfg[6]); - - if (args->pad) - return -EINVAL; - - if (!v3d_has_csd(v3d)) { - DRM_DEBUG("Attempting CSD submit on non-CSD hardware\n"); - return -EINVAL; - } - - if (args->flags && !(args->flags & DRM_V3D_SUBMIT_EXTENSION)) { - DRM_INFO("invalid flags: %d\n", args->flags); - return -EINVAL; - } - - if (args->flags & DRM_V3D_SUBMIT_EXTENSION) { - ret = v3d_get_extensions(file_priv, args->extensions, &se); - if (ret) { - DRM_DEBUG("Failed to get extensions.\n"); - return ret; - } - } - - ret = v3d_job_init(v3d, file_priv, (void *)&job, sizeof(*job), - v3d_job_free, args->in_sync, &se, V3D_CSD); - if (ret) - goto fail; - - ret = v3d_job_init(v3d, file_priv, (void *)&clean_job, sizeof(*clean_job), - v3d_job_free, 0, NULL, V3D_CACHE_CLEAN); - if (ret) - goto fail; - - job->args = *args; - - ret = v3d_lookup_bos(dev, file_priv, clean_job, - args->bo_handles, args->bo_handle_count); - if (ret) - goto fail; - - ret = v3d_lock_bo_reservations(clean_job, &acquire_ctx); - if (ret) - goto fail; - - if (args->perfmon_id) { - job->base.perfmon = v3d_perfmon_find(v3d_priv, - args->perfmon_id); - if (!job->base.perfmon) { - ret = -ENOENT; - goto fail_perfmon; - } - } - - mutex_lock(&v3d->sched_lock); - v3d_push_job(&job->base); - - ret = drm_sched_job_add_dependency(&clean_job->base, - dma_fence_get(job->base.done_fence)); - if (ret) - goto fail_unreserve; - - v3d_push_job(clean_job); - mutex_unlock(&v3d->sched_lock); - - v3d_attach_fences_and_unlock_reservation(file_priv, - clean_job, - &acquire_ctx, - args->out_sync, - &se, - clean_job->done_fence); - - v3d_job_put(&job->base); - v3d_job_put(clean_job); - - return 0; - -fail_unreserve: - mutex_unlock(&v3d->sched_lock); -fail_perfmon: - drm_gem_unlock_reservations(clean_job->bo, clean_job->bo_count, - &acquire_ctx); -fail: - v3d_job_cleanup((void *)job); - v3d_job_cleanup(clean_job); - v3d_put_multisync_post_deps(&se); - - return ret; -} - int v3d_gem_init(struct drm_device *dev) { diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c new file mode 100644 index 000000000000..0886b3ec9aef --- /dev/null +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -0,0 +1,743 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* + * Copyright (C) 2014-2018 Broadcom + * Copyright (C) 2023 Raspberry Pi + */ + +#include + +#include "v3d_drv.h" +#include "v3d_regs.h" +#include "v3d_trace.h" + +/* Takes the reservation lock on all the BOs being referenced, so that + * at queue submit time we can update the reservations. + * + * We don't lock the RCL the tile alloc/state BOs, or overflow memory + * (all of which are on exec->unref_list). They're entirely private + * to v3d, so we don't attach dma-buf fences to them. + */ +static int +v3d_lock_bo_reservations(struct v3d_job *job, + struct ww_acquire_ctx *acquire_ctx) +{ + int i, ret; + + ret = drm_gem_lock_reservations(job->bo, job->bo_count, acquire_ctx); + if (ret) + return ret; + + for (i = 0; i < job->bo_count; i++) { + ret = dma_resv_reserve_fences(job->bo[i]->resv, 1); + if (ret) + goto fail; + + ret = drm_sched_job_add_implicit_dependencies(&job->base, + job->bo[i], true); + if (ret) + goto fail; + } + + return 0; + +fail: + drm_gem_unlock_reservations(job->bo, job->bo_count, acquire_ctx); + return ret; +} + +/** + * v3d_lookup_bos() - Sets up job->bo[] with the GEM objects + * referenced by the job. + * @dev: DRM device + * @file_priv: DRM file for this fd + * @job: V3D job being set up + * @bo_handles: GEM handles + * @bo_count: Number of GEM handles passed in + * + * The command validator needs to reference BOs by their index within + * the submitted job's BO list. This does the validation of the job's + * BO list and reference counting for the lifetime of the job. + * + * Note that this function doesn't need to unreference the BOs on + * failure, because that will happen at v3d_exec_cleanup() time. + */ +static int +v3d_lookup_bos(struct drm_device *dev, + struct drm_file *file_priv, + struct v3d_job *job, + u64 bo_handles, + u32 bo_count) +{ + job->bo_count = bo_count; + + if (!job->bo_count) { + /* See comment on bo_index for why we have to check + * this. + */ + DRM_DEBUG("Rendering requires BOs\n"); + return -EINVAL; + } + + return drm_gem_objects_lookup(file_priv, + (void __user *)(uintptr_t)bo_handles, + job->bo_count, &job->bo); +} + +static void +v3d_job_free(struct kref *ref) +{ + struct v3d_job *job = container_of(ref, struct v3d_job, refcount); + int i; + + if (job->bo) { + for (i = 0; i < job->bo_count; i++) + drm_gem_object_put(job->bo[i]); + kvfree(job->bo); + } + + dma_fence_put(job->irq_fence); + dma_fence_put(job->done_fence); + + if (job->perfmon) + v3d_perfmon_put(job->perfmon); + + kfree(job); +} + +static void +v3d_render_job_free(struct kref *ref) +{ + struct v3d_render_job *job = container_of(ref, struct v3d_render_job, + base.refcount); + struct v3d_bo *bo, *save; + + list_for_each_entry_safe(bo, save, &job->unref_list, unref_head) { + drm_gem_object_put(&bo->base.base); + } + + v3d_job_free(ref); +} + +void v3d_job_cleanup(struct v3d_job *job) +{ + if (!job) + return; + + drm_sched_job_cleanup(&job->base); + v3d_job_put(job); +} + +void v3d_job_put(struct v3d_job *job) +{ + kref_put(&job->refcount, job->free); +} + +static int +v3d_job_init(struct v3d_dev *v3d, struct drm_file *file_priv, + void **container, size_t size, void (*free)(struct kref *ref), + u32 in_sync, struct v3d_submit_ext *se, enum v3d_queue queue) +{ + struct v3d_file_priv *v3d_priv = file_priv->driver_priv; + struct v3d_job *job; + bool has_multisync = se && (se->flags & DRM_V3D_EXT_ID_MULTI_SYNC); + int ret, i; + + *container = kcalloc(1, size, GFP_KERNEL); + if (!*container) { + DRM_ERROR("Cannot allocate memory for v3d job."); + return -ENOMEM; + } + + job = *container; + job->v3d = v3d; + job->free = free; + + ret = drm_sched_job_init(&job->base, &v3d_priv->sched_entity[queue], + v3d_priv); + if (ret) + goto fail; + + if (has_multisync) { + if (se->in_sync_count && se->wait_stage == queue) { + struct drm_v3d_sem __user *handle = u64_to_user_ptr(se->in_syncs); + + for (i = 0; i < se->in_sync_count; i++) { + struct drm_v3d_sem in; + + if (copy_from_user(&in, handle++, sizeof(in))) { + ret = -EFAULT; + DRM_DEBUG("Failed to copy wait dep handle.\n"); + goto fail_deps; + } + ret = drm_sched_job_add_syncobj_dependency(&job->base, file_priv, in.handle, 0); + + // TODO: Investigate why this was filtered out for the IOCTL. + if (ret && ret != -ENOENT) + goto fail_deps; + } + } + } else { + ret = drm_sched_job_add_syncobj_dependency(&job->base, file_priv, in_sync, 0); + + // TODO: Investigate why this was filtered out for the IOCTL. + if (ret && ret != -ENOENT) + goto fail_deps; + } + + kref_init(&job->refcount); + + return 0; + +fail_deps: + drm_sched_job_cleanup(&job->base); +fail: + kfree(*container); + *container = NULL; + + return ret; +} + +static void +v3d_push_job(struct v3d_job *job) +{ + drm_sched_job_arm(&job->base); + + job->done_fence = dma_fence_get(&job->base.s_fence->finished); + + /* put by scheduler job completion */ + kref_get(&job->refcount); + + drm_sched_entity_push_job(&job->base); +} + +static void +v3d_attach_fences_and_unlock_reservation(struct drm_file *file_priv, + struct v3d_job *job, + struct ww_acquire_ctx *acquire_ctx, + u32 out_sync, + struct v3d_submit_ext *se, + struct dma_fence *done_fence) +{ + struct drm_syncobj *sync_out; + bool has_multisync = se && (se->flags & DRM_V3D_EXT_ID_MULTI_SYNC); + int i; + + for (i = 0; i < job->bo_count; i++) { + /* XXX: Use shared fences for read-only objects. */ + dma_resv_add_fence(job->bo[i]->resv, job->done_fence, + DMA_RESV_USAGE_WRITE); + } + + drm_gem_unlock_reservations(job->bo, job->bo_count, acquire_ctx); + + /* Update the return sync object for the job */ + /* If it only supports a single signal semaphore*/ + if (!has_multisync) { + sync_out = drm_syncobj_find(file_priv, out_sync); + if (sync_out) { + drm_syncobj_replace_fence(sync_out, done_fence); + drm_syncobj_put(sync_out); + } + return; + } + + /* If multiple semaphores extension is supported */ + if (se->out_sync_count) { + for (i = 0; i < se->out_sync_count; i++) { + drm_syncobj_replace_fence(se->out_syncs[i].syncobj, + done_fence); + drm_syncobj_put(se->out_syncs[i].syncobj); + } + kvfree(se->out_syncs); + } +} + +static void +v3d_put_multisync_post_deps(struct v3d_submit_ext *se) +{ + unsigned int i; + + if (!(se && se->out_sync_count)) + return; + + for (i = 0; i < se->out_sync_count; i++) + drm_syncobj_put(se->out_syncs[i].syncobj); + kvfree(se->out_syncs); +} + +static int +v3d_get_multisync_post_deps(struct drm_file *file_priv, + struct v3d_submit_ext *se, + u32 count, u64 handles) +{ + struct drm_v3d_sem __user *post_deps; + int i, ret; + + if (!count) + return 0; + + se->out_syncs = (struct v3d_submit_outsync *) + kvmalloc_array(count, + sizeof(struct v3d_submit_outsync), + GFP_KERNEL); + if (!se->out_syncs) + return -ENOMEM; + + post_deps = u64_to_user_ptr(handles); + + for (i = 0; i < count; i++) { + struct drm_v3d_sem out; + + if (copy_from_user(&out, post_deps++, sizeof(out))) { + ret = -EFAULT; + DRM_DEBUG("Failed to copy post dep handles\n"); + goto fail; + } + + se->out_syncs[i].syncobj = drm_syncobj_find(file_priv, + out.handle); + if (!se->out_syncs[i].syncobj) { + ret = -EINVAL; + goto fail; + } + } + se->out_sync_count = count; + + return 0; + +fail: + for (i--; i >= 0; i--) + drm_syncobj_put(se->out_syncs[i].syncobj); + kvfree(se->out_syncs); + + return ret; +} + +/* Get data for multiple binary semaphores synchronization. Parse syncobj + * to be signaled when job completes (out_sync). + */ +static int +v3d_get_multisync_submit_deps(struct drm_file *file_priv, + struct drm_v3d_extension __user *ext, + void *data) +{ + struct drm_v3d_multi_sync multisync; + struct v3d_submit_ext *se = data; + int ret; + + if (copy_from_user(&multisync, ext, sizeof(multisync))) + return -EFAULT; + + if (multisync.pad) + return -EINVAL; + + ret = v3d_get_multisync_post_deps(file_priv, data, multisync.out_sync_count, + multisync.out_syncs); + if (ret) + return ret; + + se->in_sync_count = multisync.in_sync_count; + se->in_syncs = multisync.in_syncs; + se->flags |= DRM_V3D_EXT_ID_MULTI_SYNC; + se->wait_stage = multisync.wait_stage; + + return 0; +} + +/* Whenever userspace sets ioctl extensions, v3d_get_extensions parses data + * according to the extension id (name). + */ +static int +v3d_get_extensions(struct drm_file *file_priv, + u64 ext_handles, + void *data) +{ + struct drm_v3d_extension __user *user_ext; + int ret; + + user_ext = u64_to_user_ptr(ext_handles); + while (user_ext) { + struct drm_v3d_extension ext; + + if (copy_from_user(&ext, user_ext, sizeof(ext))) { + DRM_DEBUG("Failed to copy submit extension\n"); + return -EFAULT; + } + + switch (ext.id) { + case DRM_V3D_EXT_ID_MULTI_SYNC: + ret = v3d_get_multisync_submit_deps(file_priv, user_ext, data); + if (ret) + return ret; + break; + default: + DRM_DEBUG_DRIVER("Unknown extension id: %d\n", ext.id); + return -EINVAL; + } + + user_ext = u64_to_user_ptr(ext.next); + } + + return 0; +} + +/** + * v3d_submit_cl_ioctl() - Submits a job (frame) to the V3D. + * @dev: DRM device + * @data: ioctl argument + * @file_priv: DRM file for this fd + * + * This is the main entrypoint for userspace to submit a 3D frame to + * the GPU. Userspace provides the binner command list (if + * applicable), and the kernel sets up the render command list to draw + * to the framebuffer described in the ioctl, using the command lists + * that the 3D engine's binner will produce. + */ +int +v3d_submit_cl_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct v3d_dev *v3d = to_v3d_dev(dev); + struct v3d_file_priv *v3d_priv = file_priv->driver_priv; + struct drm_v3d_submit_cl *args = data; + struct v3d_submit_ext se = {0}; + struct v3d_bin_job *bin = NULL; + struct v3d_render_job *render = NULL; + struct v3d_job *clean_job = NULL; + struct v3d_job *last_job; + struct ww_acquire_ctx acquire_ctx; + int ret = 0; + + trace_v3d_submit_cl_ioctl(&v3d->drm, args->rcl_start, args->rcl_end); + + if (args->pad) + return -EINVAL; + + if (args->flags && + args->flags & ~(DRM_V3D_SUBMIT_CL_FLUSH_CACHE | + DRM_V3D_SUBMIT_EXTENSION)) { + DRM_INFO("invalid flags: %d\n", args->flags); + return -EINVAL; + } + + if (args->flags & DRM_V3D_SUBMIT_EXTENSION) { + ret = v3d_get_extensions(file_priv, args->extensions, &se); + if (ret) { + DRM_DEBUG("Failed to get extensions.\n"); + return ret; + } + } + + ret = v3d_job_init(v3d, file_priv, (void *)&render, sizeof(*render), + v3d_render_job_free, args->in_sync_rcl, &se, V3D_RENDER); + if (ret) + goto fail; + + render->start = args->rcl_start; + render->end = args->rcl_end; + INIT_LIST_HEAD(&render->unref_list); + + if (args->bcl_start != args->bcl_end) { + ret = v3d_job_init(v3d, file_priv, (void *)&bin, sizeof(*bin), + v3d_job_free, args->in_sync_bcl, &se, V3D_BIN); + if (ret) + goto fail; + + bin->start = args->bcl_start; + bin->end = args->bcl_end; + bin->qma = args->qma; + bin->qms = args->qms; + bin->qts = args->qts; + bin->render = render; + } + + if (args->flags & DRM_V3D_SUBMIT_CL_FLUSH_CACHE) { + ret = v3d_job_init(v3d, file_priv, (void *)&clean_job, sizeof(*clean_job), + v3d_job_free, 0, NULL, V3D_CACHE_CLEAN); + if (ret) + goto fail; + + last_job = clean_job; + } else { + last_job = &render->base; + } + + ret = v3d_lookup_bos(dev, file_priv, last_job, + args->bo_handles, args->bo_handle_count); + if (ret) + goto fail; + + ret = v3d_lock_bo_reservations(last_job, &acquire_ctx); + if (ret) + goto fail; + + if (args->perfmon_id) { + render->base.perfmon = v3d_perfmon_find(v3d_priv, + args->perfmon_id); + + if (!render->base.perfmon) { + ret = -ENOENT; + goto fail_perfmon; + } + } + + mutex_lock(&v3d->sched_lock); + if (bin) { + bin->base.perfmon = render->base.perfmon; + v3d_perfmon_get(bin->base.perfmon); + v3d_push_job(&bin->base); + + ret = drm_sched_job_add_dependency(&render->base.base, + dma_fence_get(bin->base.done_fence)); + if (ret) + goto fail_unreserve; + } + + v3d_push_job(&render->base); + + if (clean_job) { + struct dma_fence *render_fence = + dma_fence_get(render->base.done_fence); + ret = drm_sched_job_add_dependency(&clean_job->base, + render_fence); + if (ret) + goto fail_unreserve; + clean_job->perfmon = render->base.perfmon; + v3d_perfmon_get(clean_job->perfmon); + v3d_push_job(clean_job); + } + + mutex_unlock(&v3d->sched_lock); + + v3d_attach_fences_and_unlock_reservation(file_priv, + last_job, + &acquire_ctx, + args->out_sync, + &se, + last_job->done_fence); + + if (bin) + v3d_job_put(&bin->base); + v3d_job_put(&render->base); + if (clean_job) + v3d_job_put(clean_job); + + return 0; + +fail_unreserve: + mutex_unlock(&v3d->sched_lock); +fail_perfmon: + drm_gem_unlock_reservations(last_job->bo, + last_job->bo_count, &acquire_ctx); +fail: + v3d_job_cleanup((void *)bin); + v3d_job_cleanup((void *)render); + v3d_job_cleanup(clean_job); + v3d_put_multisync_post_deps(&se); + + return ret; +} + +/** + * v3d_submit_tfu_ioctl() - Submits a TFU (texture formatting) job to the V3D. + * @dev: DRM device + * @data: ioctl argument + * @file_priv: DRM file for this fd + * + * Userspace provides the register setup for the TFU, which we don't + * need to validate since the TFU is behind the MMU. + */ +int +v3d_submit_tfu_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct v3d_dev *v3d = to_v3d_dev(dev); + struct drm_v3d_submit_tfu *args = data; + struct v3d_submit_ext se = {0}; + struct v3d_tfu_job *job = NULL; + struct ww_acquire_ctx acquire_ctx; + int ret = 0; + + trace_v3d_submit_tfu_ioctl(&v3d->drm, args->iia); + + if (args->flags && !(args->flags & DRM_V3D_SUBMIT_EXTENSION)) { + DRM_DEBUG("invalid flags: %d\n", args->flags); + return -EINVAL; + } + + if (args->flags & DRM_V3D_SUBMIT_EXTENSION) { + ret = v3d_get_extensions(file_priv, args->extensions, &se); + if (ret) { + DRM_DEBUG("Failed to get extensions.\n"); + return ret; + } + } + + ret = v3d_job_init(v3d, file_priv, (void *)&job, sizeof(*job), + v3d_job_free, args->in_sync, &se, V3D_TFU); + if (ret) + goto fail; + + job->base.bo = kcalloc(ARRAY_SIZE(args->bo_handles), + sizeof(*job->base.bo), GFP_KERNEL); + if (!job->base.bo) { + ret = -ENOMEM; + goto fail; + } + + job->args = *args; + + for (job->base.bo_count = 0; + job->base.bo_count < ARRAY_SIZE(args->bo_handles); + job->base.bo_count++) { + struct drm_gem_object *bo; + + if (!args->bo_handles[job->base.bo_count]) + break; + + bo = drm_gem_object_lookup(file_priv, args->bo_handles[job->base.bo_count]); + if (!bo) { + DRM_DEBUG("Failed to look up GEM BO %d: %d\n", + job->base.bo_count, + args->bo_handles[job->base.bo_count]); + ret = -ENOENT; + goto fail; + } + job->base.bo[job->base.bo_count] = bo; + } + + ret = v3d_lock_bo_reservations(&job->base, &acquire_ctx); + if (ret) + goto fail; + + mutex_lock(&v3d->sched_lock); + v3d_push_job(&job->base); + mutex_unlock(&v3d->sched_lock); + + v3d_attach_fences_and_unlock_reservation(file_priv, + &job->base, &acquire_ctx, + args->out_sync, + &se, + job->base.done_fence); + + v3d_job_put(&job->base); + + return 0; + +fail: + v3d_job_cleanup((void *)job); + v3d_put_multisync_post_deps(&se); + + return ret; +} + +/** + * v3d_submit_csd_ioctl() - Submits a CSD (compute shader) job to the V3D. + * @dev: DRM device + * @data: ioctl argument + * @file_priv: DRM file for this fd + * + * Userspace provides the register setup for the CSD, which we don't + * need to validate since the CSD is behind the MMU. + */ +int +v3d_submit_csd_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct v3d_dev *v3d = to_v3d_dev(dev); + struct v3d_file_priv *v3d_priv = file_priv->driver_priv; + struct drm_v3d_submit_csd *args = data; + struct v3d_submit_ext se = {0}; + struct v3d_csd_job *job = NULL; + struct v3d_job *clean_job = NULL; + struct ww_acquire_ctx acquire_ctx; + int ret; + + trace_v3d_submit_csd_ioctl(&v3d->drm, args->cfg[5], args->cfg[6]); + + if (args->pad) + return -EINVAL; + + if (!v3d_has_csd(v3d)) { + DRM_DEBUG("Attempting CSD submit on non-CSD hardware\n"); + return -EINVAL; + } + + if (args->flags && !(args->flags & DRM_V3D_SUBMIT_EXTENSION)) { + DRM_INFO("invalid flags: %d\n", args->flags); + return -EINVAL; + } + + if (args->flags & DRM_V3D_SUBMIT_EXTENSION) { + ret = v3d_get_extensions(file_priv, args->extensions, &se); + if (ret) { + DRM_DEBUG("Failed to get extensions.\n"); + return ret; + } + } + + ret = v3d_job_init(v3d, file_priv, (void *)&job, sizeof(*job), + v3d_job_free, args->in_sync, &se, V3D_CSD); + if (ret) + goto fail; + + ret = v3d_job_init(v3d, file_priv, (void *)&clean_job, sizeof(*clean_job), + v3d_job_free, 0, NULL, V3D_CACHE_CLEAN); + if (ret) + goto fail; + + job->args = *args; + + ret = v3d_lookup_bos(dev, file_priv, clean_job, + args->bo_handles, args->bo_handle_count); + if (ret) + goto fail; + + ret = v3d_lock_bo_reservations(clean_job, &acquire_ctx); + if (ret) + goto fail; + + if (args->perfmon_id) { + job->base.perfmon = v3d_perfmon_find(v3d_priv, + args->perfmon_id); + if (!job->base.perfmon) { + ret = -ENOENT; + goto fail_perfmon; + } + } + + mutex_lock(&v3d->sched_lock); + v3d_push_job(&job->base); + + ret = drm_sched_job_add_dependency(&clean_job->base, + dma_fence_get(job->base.done_fence)); + if (ret) + goto fail_unreserve; + + v3d_push_job(clean_job); + mutex_unlock(&v3d->sched_lock); + + v3d_attach_fences_and_unlock_reservation(file_priv, + clean_job, + &acquire_ctx, + args->out_sync, + &se, + clean_job->done_fence); + + v3d_job_put(&job->base); + v3d_job_put(clean_job); + + return 0; + +fail_unreserve: + mutex_unlock(&v3d->sched_lock); +fail_perfmon: + drm_gem_unlock_reservations(clean_job->bo, clean_job->bo_count, + &acquire_ctx); +fail: + v3d_job_cleanup((void *)job); + v3d_job_cleanup(clean_job); + v3d_put_multisync_post_deps(&se); + + return ret; +} From patchwork Mon Sep 4 17:38:52 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374237 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 9B279C83F3F for ; Mon, 4 Sep 2023 17:51:28 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id F276610E3C0; Mon, 4 Sep 2023 17:51:22 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id 198D810E3C0 for ; Mon, 4 Sep 2023 17:51:21 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=cXsrpJcfgn8WqALuyrPzgsWgpeIMLJDt3E/KYtPzpMg=; b=TJfXBxJ3tCnMO0GniYlD6ksfRU aItptzeU/7cbUA6Fd6hjtdEQNwbj4GwWHGbwd9L1HkJZKMupK7Jk7WHlFRi19zpeG1npuxG/T4p/L oD23oJvwojiR+jyuq9SN8xdJDfxbIFTU16u1waf9lNyVLVlnvCS78Jh7PN45GlJRPoB9qy0BWn1rh EuF8zP1IruARdBTcFXHRD//sEVNJsxhzjpiaaykIamo8RWIj0T8MM/D1qh1pq29zlAN9zswjOB48e AbFkYGfVX97rwmo2UUdHxvD5rxIytdufvHQdk/v33a/+WocfTCriPyvBVeUPxm2PdafizAzXPTxLV jUClJaGg==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDjE-0028nB-4O; Mon, 04 Sep 2023 19:51:12 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 04/17] drm/v3d: Simplify job refcount handling Date: Mon, 4 Sep 2023 14:38:52 -0300 Message-ID: <20230904175019.1172713-5-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Melissa Wen Instead of checking if the job is NULL every time we call the function, check it inside the function. Signed-off-by: Melissa Wen Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_submit.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index 0886b3ec9aef..b6be34060d4f 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -129,6 +129,9 @@ void v3d_job_cleanup(struct v3d_job *job) void v3d_job_put(struct v3d_job *job) { + if (!job) + return; + kref_put(&job->refcount, job->free); } @@ -516,11 +519,9 @@ v3d_submit_cl_ioctl(struct drm_device *dev, void *data, &se, last_job->done_fence); - if (bin) - v3d_job_put(&bin->base); - v3d_job_put(&render->base); - if (clean_job) - v3d_job_put(clean_job); + v3d_job_put((void *)bin); + v3d_job_put((void *)render); + v3d_job_put((void *)clean_job); return 0; @@ -620,7 +621,7 @@ v3d_submit_tfu_ioctl(struct drm_device *dev, void *data, &se, job->base.done_fence); - v3d_job_put(&job->base); + v3d_job_put((void *)job); return 0; @@ -724,7 +725,7 @@ v3d_submit_csd_ioctl(struct drm_device *dev, void *data, &se, clean_job->done_fence); - v3d_job_put(&job->base); + v3d_job_put((void *)job); v3d_job_put(clean_job); return 0; From patchwork Mon Sep 4 17:38:53 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374239 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 747B1CA0FE3 for ; Mon, 4 Sep 2023 17:51:30 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 7C0E810E3BE; Mon, 4 Sep 2023 17:51:27 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id 2282810E3C2 for ; Mon, 4 Sep 2023 17:51:25 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=UWwfXvnq5gNOxmmva2w4VKJSyzXdaCzY+LazedhAQu0=; b=C2IPptIjDIqGkO3l5Twp7imjki tlQqndwkD2MUw7isPauyIeV7OMdpmeowE4cYV8MjIBXDFDdJ2VQPBK7gyd6xnhlVeii2ChXeFtVNo 41dqaDrYE+n8usuhUYgiNpj7XJ3C5cJueT/Sv5b4qJKmwb3bQpJhEctx+AU0Su9jUmYKHxFSTfJda xlj0aifpoQQEJOAwtjlRb8TpOA0+8MstInVVl50H5rAxrXBo6CejJn46qAxuyRTW0U9gtC8Ztmhsr s9LaHgzU2BHFbjmAyVt3vUb6tvdg1a9qypAs+Ej7DaTnGMrlMi12ASESQ9FTqBDEPhg/Invy7vWM3 QQfOhvCw==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDjI-0028nB-Ad; Mon, 04 Sep 2023 19:51:16 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 05/17] drm/v3d: Don't allow two multisync extensions in the same job Date: Mon, 4 Sep 2023 14:38:53 -0300 Message-ID: <20230904175019.1172713-6-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" Currently, two multisync extensions can be added to the same job and only the last multisync extension will be used. To avoid this vulnerability, don't allow two multisync extensions in the same job. Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_submit.c | 5 +++++ 1 file changed, 5 insertions(+) -- 2.41.0 diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index b6be34060d4f..2c39e2acf01b 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -328,6 +328,11 @@ v3d_get_multisync_submit_deps(struct drm_file *file_priv, struct v3d_submit_ext *se = data; int ret; + if (se->in_sync_count || se->out_sync_count) { + DRM_DEBUG("Two multisync extensions were added to the same job."); + return -EINVAL; + } + if (copy_from_user(&multisync, ext, sizeof(multisync))) return -EFAULT; From patchwork Mon Sep 4 17:38:54 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374240 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 E4466C83F3F for ; Mon, 4 Sep 2023 17:51:31 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id DFC3F10E3C2; Mon, 4 Sep 2023 17:51:30 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id 00B4810E3BC for ; Mon, 4 Sep 2023 17:51:28 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=Nm8GA70bZRuilSNAI6bS9zTz2fhlHnGN270RDw5vE5o=; b=Ohqx44BvUptdxVH+Y5JT5AXSlz oYVC9pFwUJ3ZYDkU/5377yR+4uW55fCc7HeOtqx2JEe9Y2n8Ggf58GCFO0BZqtp/D1nnHjKNglsCR 89CsrWZOxdm2IjYeDvD2z9/hA1YsCC5fd+k2RFLb6+BKxTVn2OQj/LJscnRcz/lJ735wh78ejBrls 4qNJhGmCF/0/SuLnXRUSlzcLhgyMS0XMXvZTUOoy3Ye0v9ZdNtvW2EmIAGDR0Px9XLQE1kI6L1hTV kOmpvoiQLj1yrirKSyNqOv84V6BftzpQcw2qXmnPW128bqen+cM/tog6Nfb/o7ly5uIetlH3FRuOP aD6vPxnA==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDjM-0028nB-FM; Mon, 04 Sep 2023 19:51:20 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 06/17] drm/v3d: Decouple job allocation from job initiation Date: Mon, 4 Sep 2023 14:38:54 -0300 Message-ID: <20230904175019.1172713-7-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" We want to allow the IOCTLs to allocate the job without initiating it. This will be useful for the CPU job submission IOCTL, as the CPU job has the need to use information from the user extensions. Currently, the user extensions are parsed before the job allocation, making it impossible to fill the CPU job when parsing the user extensions. Therefore, decouple the job allocation from the job initiation. Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_submit.c | 23 ++++++++++++++++++----- 1 file changed, 18 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index 2c39e2acf01b..dff4525e6fde 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -135,6 +135,21 @@ void v3d_job_put(struct v3d_job *job) kref_put(&job->refcount, job->free); } +static int +v3d_job_allocate(void **container, size_t size) +{ + if (*container) + return 0; + + *container = kcalloc(1, size, GFP_KERNEL); + if (!*container) { + DRM_ERROR("Cannot allocate memory for V3D job.\n"); + return -ENOMEM; + } + + return 0; +} + static int v3d_job_init(struct v3d_dev *v3d, struct drm_file *file_priv, void **container, size_t size, void (*free)(struct kref *ref), @@ -145,11 +160,9 @@ v3d_job_init(struct v3d_dev *v3d, struct drm_file *file_priv, bool has_multisync = se && (se->flags & DRM_V3D_EXT_ID_MULTI_SYNC); int ret, i; - *container = kcalloc(1, size, GFP_KERNEL); - if (!*container) { - DRM_ERROR("Cannot allocate memory for v3d job."); - return -ENOMEM; - } + ret = v3d_job_allocate(container, size); + if (ret) + return ret; job = *container; job->v3d = v3d; From patchwork Mon Sep 4 17:38:55 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374241 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 DF3A9C71153 for ; Mon, 4 Sep 2023 17:51:34 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 4B89410E3C3; Mon, 4 Sep 2023 17:51:34 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id 7FCEA10E3C3 for ; Mon, 4 Sep 2023 17:51:32 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=rFF2cvRQpWk4qr6kCxWTLND/sRV4ctfuwxD55Rzr4zY=; b=VSA6mfkojkYw0GE+kZPckuUJOG SqdLiZKMjKP3CAyz4QXfu8CT2JXCR/Bk4dVGG+7AedOZpt6mTEo4f83D3/CMtevTqLuKJsni153aX b/3i1fENZ7eo/IkpsOddJK+NvsAHg2NWZtEHwx9XioHTTE1Bw9evgNRCB1YGoQ3A1udL8GreBWwHG Utvs1NLVOQzSH/FJgfXi5mbVoi/iGNBdlqULorCksVA+Mg1TSLsodQpbyFjMPWSHOOYwyQvnupeZK s5BLr5fqFsQAu3Hx0YdPf/kcGeSrZtrWPA+aqfyeDpciP3QO1NFaT82JNeav8WccCkrAOFZ9Gjh2w t9fZKQYg==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDjQ-0028nB-Ln; Mon, 04 Sep 2023 19:51:25 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 07/17] drm/v3d: Add a CPU job submission Date: Mon, 4 Sep 2023 14:38:55 -0300 Message-ID: <20230904175019.1172713-8-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Melissa Wen Create a new type of job, a CPU job. A CPU job is a type of job that performs operations that requires CPU intervention. The overall idea is to use user extensions to enable different types of CPU job, allowing the CPU job to perform different operations according to the type of user externsion. The user extension ID identify the type of CPU job that must be dealt. Having a CPU job is interesting for synchronization purposes as a CPU job has a queue like any other V3D job and can be synchoronized by the multisync extension. Signed-off-by: Melissa Wen Co-developed-by: Maíra Canal Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_drv.c | 4 ++ drivers/gpu/drm/v3d/v3d_drv.h | 13 +++++- drivers/gpu/drm/v3d/v3d_sched.c | 40 ++++++++++++++++ drivers/gpu/drm/v3d/v3d_submit.c | 79 ++++++++++++++++++++++++++++++++ include/uapi/drm/v3d_drm.h | 17 +++++++ 5 files changed, 152 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/v3d/v3d_drv.c b/drivers/gpu/drm/v3d/v3d_drv.c index ffbbe9d527d3..6f6ef5af2bd0 100644 --- a/drivers/gpu/drm/v3d/v3d_drv.c +++ b/drivers/gpu/drm/v3d/v3d_drv.c @@ -90,6 +90,9 @@ static int v3d_get_param_ioctl(struct drm_device *dev, void *data, case DRM_V3D_PARAM_SUPPORTS_MULTISYNC_EXT: args->value = 1; return 0; + case DRM_V3D_PARAM_SUPPORTS_CPU_QUEUE: + args->value = 1; + return 0; default: DRM_DEBUG("Unknown parameter %d\n", args->param); return -EINVAL; @@ -156,6 +159,7 @@ static const struct drm_ioctl_desc v3d_drm_ioctls[] = { DRM_IOCTL_DEF_DRV(V3D_PERFMON_CREATE, v3d_perfmon_create_ioctl, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(V3D_PERFMON_DESTROY, v3d_perfmon_destroy_ioctl, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(V3D_PERFMON_GET_VALUES, v3d_perfmon_get_values_ioctl, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(V3D_SUBMIT_CPU, v3d_submit_cpu_ioctl, DRM_RENDER_ALLOW | DRM_AUTH), }; static const struct drm_driver v3d_drm_driver = { diff --git a/drivers/gpu/drm/v3d/v3d_drv.h b/drivers/gpu/drm/v3d/v3d_drv.h index 722a627e0a6e..2a3f3beb272c 100644 --- a/drivers/gpu/drm/v3d/v3d_drv.h +++ b/drivers/gpu/drm/v3d/v3d_drv.h @@ -19,7 +19,7 @@ struct reset_control; #define GMP_GRANULARITY (128 * 1024) -#define V3D_MAX_QUEUES (V3D_CACHE_CLEAN + 1) +#define V3D_MAX_QUEUES (V3D_CPU + 1) struct v3d_queue_state { struct drm_gpu_scheduler sched; @@ -106,6 +106,7 @@ struct v3d_dev { struct v3d_render_job *render_job; struct v3d_tfu_job *tfu_job; struct v3d_csd_job *csd_job; + struct v3d_cpu_job *cpu_job; struct v3d_queue_state queue[V3D_MAX_QUEUES]; @@ -285,6 +286,14 @@ struct v3d_csd_job { struct drm_v3d_submit_csd args; }; +enum v3d_cpu_job_type {}; + +struct v3d_cpu_job { + struct v3d_job base; + + enum v3d_cpu_job_type job_type; +}; + struct v3d_submit_outsync { struct drm_syncobj *syncobj; }; @@ -387,6 +396,8 @@ int v3d_submit_tfu_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); int v3d_submit_csd_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); +int v3d_submit_cpu_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); /* v3d_irq.c */ int v3d_irq_init(struct v3d_dev *v3d); diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c index 06238e6d7f5c..88c483da360c 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -54,6 +54,12 @@ to_csd_job(struct drm_sched_job *sched_job) return container_of(sched_job, struct v3d_csd_job, base.base); } +static struct v3d_cpu_job * +to_cpu_job(struct drm_sched_job *sched_job) +{ + return container_of(sched_job, struct v3d_cpu_job, base.base); +} + static void v3d_sched_job_free(struct drm_sched_job *sched_job) { @@ -241,6 +247,26 @@ v3d_csd_job_run(struct drm_sched_job *sched_job) return fence; } +static struct dma_fence * +v3d_cpu_job_run(struct drm_sched_job *sched_job) +{ + struct v3d_cpu_job *job = to_cpu_job(sched_job); + struct v3d_dev *v3d = job->base.v3d; + + void (*v3d_cpu_job_fn[])(struct v3d_cpu_job *job) = { }; + + v3d->cpu_job = job; + + if (job->job_type >= ARRAY_SIZE(v3d_cpu_job_fn)) { + DRM_DEBUG_DRIVER("Unknown CPU job: %d\n", job->job_type); + return NULL; + } + + v3d_cpu_job_fn[job->job_type](job); + + return NULL; +} + static struct dma_fence * v3d_cache_clean_job_run(struct drm_sched_job *sched_job) { @@ -379,6 +405,12 @@ static const struct drm_sched_backend_ops v3d_cache_clean_sched_ops = { .free_job = v3d_sched_job_free }; +static const struct drm_sched_backend_ops v3d_cpu_sched_ops = { + .run_job = v3d_cpu_job_run, + .timedout_job = v3d_generic_job_timedout, + .free_job = v3d_sched_job_free +}; + int v3d_sched_init(struct v3d_dev *v3d) { @@ -429,6 +461,14 @@ v3d_sched_init(struct v3d_dev *v3d) goto fail; } + ret = drm_sched_init(&v3d->queue[V3D_CPU].sched, + &v3d_cpu_sched_ops, + 1, job_hang_limit, + msecs_to_jiffies(hang_limit_ms), NULL, + NULL, "v3d_cpu", v3d->drm.dev); + if (ret) + goto fail; + return 0; fail: diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index dff4525e6fde..40880b758071 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -760,3 +760,82 @@ v3d_submit_csd_ioctl(struct drm_device *dev, void *data, return ret; } + +/** + * v3d_submit_cpu_ioctl() - Submits a CPU job to the V3D. + * @dev: DRM device + * @data: ioctl argument + * @file_priv: DRM file for this fd + * + * Userspace specifies the CPU job type and data required to perform its + * operations through the drm_v3d_extension struct. + */ +int +v3d_submit_cpu_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct v3d_dev *v3d = to_v3d_dev(dev); + struct drm_v3d_submit_cpu *args = data; + struct v3d_submit_ext se = {0}; + struct v3d_cpu_job *cpu_job = NULL; + struct ww_acquire_ctx acquire_ctx; + int ret; + + if (args->flags && !(args->flags & DRM_V3D_SUBMIT_EXTENSION)) { + DRM_INFO("invalid flags: %d\n", args->flags); + return -EINVAL; + } + + ret = v3d_job_allocate((void *)&cpu_job, sizeof(*cpu_job)); + if (ret) + return ret; + + if (args->flags & DRM_V3D_SUBMIT_EXTENSION) { + ret = v3d_get_extensions(file_priv, args->extensions, &se); + if (ret) { + DRM_DEBUG("Failed to get extensions.\n"); + goto fail; + } + } + + /* Every CPU job must have a CPU job user extension */ + if (!cpu_job->job_type) { + DRM_DEBUG("CPU job must have a CPU job user extension.\n"); + goto fail; + } + + ret = v3d_job_init(v3d, file_priv, (void *)&cpu_job, sizeof(*cpu_job), + v3d_job_free, 0, &se, V3D_CPU); + if (ret) + goto fail; + + if (args->bo_handle_count) { + ret = v3d_lookup_bos(dev, file_priv, &cpu_job->base, + args->bo_handles, args->bo_handle_count); + if (ret) + goto fail; + + ret = v3d_lock_bo_reservations(&cpu_job->base, &acquire_ctx); + if (ret) + goto fail; + } + + mutex_lock(&v3d->sched_lock); + v3d_push_job(&cpu_job->base); + mutex_unlock(&v3d->sched_lock); + + v3d_attach_fences_and_unlock_reservation(file_priv, + &cpu_job->base, + &acquire_ctx, 0, + NULL, cpu_job->base.done_fence); + + v3d_job_put((void *)cpu_job); + + return 0; + +fail: + v3d_job_cleanup((void *)cpu_job); + v3d_put_multisync_post_deps(&se); + + return ret; +} diff --git a/include/uapi/drm/v3d_drm.h b/include/uapi/drm/v3d_drm.h index 3dfc0af8756a..d63044ebd5e0 100644 --- a/include/uapi/drm/v3d_drm.h +++ b/include/uapi/drm/v3d_drm.h @@ -41,6 +41,7 @@ extern "C" { #define DRM_V3D_PERFMON_CREATE 0x08 #define DRM_V3D_PERFMON_DESTROY 0x09 #define DRM_V3D_PERFMON_GET_VALUES 0x0a +#define DRM_V3D_SUBMIT_CPU 0x0b #define DRM_IOCTL_V3D_SUBMIT_CL DRM_IOWR(DRM_COMMAND_BASE + DRM_V3D_SUBMIT_CL, struct drm_v3d_submit_cl) #define DRM_IOCTL_V3D_WAIT_BO DRM_IOWR(DRM_COMMAND_BASE + DRM_V3D_WAIT_BO, struct drm_v3d_wait_bo) @@ -56,6 +57,7 @@ extern "C" { struct drm_v3d_perfmon_destroy) #define DRM_IOCTL_V3D_PERFMON_GET_VALUES DRM_IOWR(DRM_COMMAND_BASE + DRM_V3D_PERFMON_GET_VALUES, \ struct drm_v3d_perfmon_get_values) +#define DRM_IOCTL_V3D_SUBMIT_CPU DRM_IOW(DRM_COMMAND_BASE + DRM_V3D_SUBMIT_CPU, struct drm_v3d_submit_cpu) #define DRM_V3D_SUBMIT_CL_FLUSH_CACHE 0x01 #define DRM_V3D_SUBMIT_EXTENSION 0x02 @@ -93,6 +95,7 @@ enum v3d_queue { V3D_TFU, V3D_CSD, V3D_CACHE_CLEAN, + V3D_CPU, }; /** @@ -276,6 +279,7 @@ enum drm_v3d_param { DRM_V3D_PARAM_SUPPORTS_CACHE_FLUSH, DRM_V3D_PARAM_SUPPORTS_PERFMON, DRM_V3D_PARAM_SUPPORTS_MULTISYNC_EXT, + DRM_V3D_PARAM_SUPPORTS_CPU_QUEUE, }; struct drm_v3d_get_param { @@ -356,6 +360,19 @@ struct drm_v3d_submit_csd { __u32 pad; }; +struct drm_v3d_submit_cpu { + /* Pointer to a u32 array of the BOs that are referenced by the job. */ + __u64 bo_handles; + + /* Number of BO handles passed in (size is that times 4). */ + __u32 bo_handle_count; + + __u32 flags; + + /* Pointer to an array of ioctl extensions*/ + __u64 extensions; +}; + enum { V3D_PERFCNT_FEP_VALID_PRIMTS_NO_PIXELS, V3D_PERFCNT_FEP_VALID_PRIMS, From patchwork Mon Sep 4 17:38:56 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374242 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 4B0CBC83F2C for ; Mon, 4 Sep 2023 17:51:40 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 7F68410E3BC; Mon, 4 Sep 2023 17:51:39 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id 01A0610E3C4 for ; Mon, 4 Sep 2023 17:51:37 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=v9YxCx8GoRqIlIuDd7fj0bn4SSx2n0N+8MT0Lj2MDWo=; b=af1izxyIYJhBIyNyoJIfkIJck0 17tp0p8/GvQ2GSinsJxpbMxKPPcloukm9K6zwcAIk0vjP1JX5rrcFmSddijXx1dBz4lamDBmaDVlU zHkwR/1i5CSG0QSopCXt4rB1M1H/pumjuVWLSlBTWucDwWPVe/GP4E+acQkL8nSXvD/NSqfsaMWed z5+ss9lWXBKZWObhI9wgI6yu/C9tPH5N917iWcS/kiulfBC/cmAgyfTMBgdxUNqu8RDlGBH9+g1Mk 4Jh+1Q6l7SCbG91XNGKHee697ouArOEYC1EKo7btvdXxN++CxAk3/L6MR6RCOeZmEODNY0qRYvIVA HZvCWu8Q==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDjU-0028nB-Qa; Mon, 04 Sep 2023 19:51:29 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 08/17] drm/v3d: Use v3d_get_extensions() to parse CPU job data Date: Mon, 4 Sep 2023 14:38:56 -0300 Message-ID: <20230904175019.1172713-9-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" Currently, v3d_get_extensions() only parses multisync data and assigns it to the `struct v3d_submit_ext`. But, to implement the CPU job with user extensions, we want v3d_get_extensions() to be able to parse CPU job data and assign it to the `struct v3d_cpu_job`. Therefore, allow the function v3d_get_extensions() to use `struct v3d_cpu_job *` as a parameter. If the `struct v3d_cpu_job *` is assigned to NULL, it means that the job is a GPU job and CPU job extensions should be rejected. Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_submit.c | 23 ++++++++++++----------- 1 file changed, 12 insertions(+), 11 deletions(-) diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index 40880b758071..ff8a77a4e2b0 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -335,10 +335,9 @@ v3d_get_multisync_post_deps(struct drm_file *file_priv, static int v3d_get_multisync_submit_deps(struct drm_file *file_priv, struct drm_v3d_extension __user *ext, - void *data) + struct v3d_submit_ext *se) { struct drm_v3d_multi_sync multisync; - struct v3d_submit_ext *se = data; int ret; if (se->in_sync_count || se->out_sync_count) { @@ -352,7 +351,7 @@ v3d_get_multisync_submit_deps(struct drm_file *file_priv, if (multisync.pad) return -EINVAL; - ret = v3d_get_multisync_post_deps(file_priv, data, multisync.out_sync_count, + ret = v3d_get_multisync_post_deps(file_priv, se, multisync.out_sync_count, multisync.out_syncs); if (ret) return ret; @@ -371,7 +370,8 @@ v3d_get_multisync_submit_deps(struct drm_file *file_priv, static int v3d_get_extensions(struct drm_file *file_priv, u64 ext_handles, - void *data) + struct v3d_submit_ext *se, + struct v3d_cpu_job *job) { struct drm_v3d_extension __user *user_ext; int ret; @@ -387,15 +387,16 @@ v3d_get_extensions(struct drm_file *file_priv, switch (ext.id) { case DRM_V3D_EXT_ID_MULTI_SYNC: - ret = v3d_get_multisync_submit_deps(file_priv, user_ext, data); - if (ret) - return ret; + ret = v3d_get_multisync_submit_deps(file_priv, user_ext, se); break; default: DRM_DEBUG_DRIVER("Unknown extension id: %d\n", ext.id); return -EINVAL; } + if (ret) + return ret; + user_ext = u64_to_user_ptr(ext.next); } @@ -442,7 +443,7 @@ v3d_submit_cl_ioctl(struct drm_device *dev, void *data, } if (args->flags & DRM_V3D_SUBMIT_EXTENSION) { - ret = v3d_get_extensions(file_priv, args->extensions, &se); + ret = v3d_get_extensions(file_priv, args->extensions, &se, NULL); if (ret) { DRM_DEBUG("Failed to get extensions.\n"); return ret; @@ -585,7 +586,7 @@ v3d_submit_tfu_ioctl(struct drm_device *dev, void *data, } if (args->flags & DRM_V3D_SUBMIT_EXTENSION) { - ret = v3d_get_extensions(file_priv, args->extensions, &se); + ret = v3d_get_extensions(file_priv, args->extensions, &se, NULL); if (ret) { DRM_DEBUG("Failed to get extensions.\n"); return ret; @@ -688,7 +689,7 @@ v3d_submit_csd_ioctl(struct drm_device *dev, void *data, } if (args->flags & DRM_V3D_SUBMIT_EXTENSION) { - ret = v3d_get_extensions(file_priv, args->extensions, &se); + ret = v3d_get_extensions(file_priv, args->extensions, &se, NULL); if (ret) { DRM_DEBUG("Failed to get extensions.\n"); return ret; @@ -791,7 +792,7 @@ v3d_submit_cpu_ioctl(struct drm_device *dev, void *data, return ret; if (args->flags & DRM_V3D_SUBMIT_EXTENSION) { - ret = v3d_get_extensions(file_priv, args->extensions, &se); + ret = v3d_get_extensions(file_priv, args->extensions, &se, cpu_job); if (ret) { DRM_DEBUG("Failed to get extensions.\n"); goto fail; From patchwork Mon Sep 4 17:38:57 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374243 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 804EFC71153 for ; Mon, 4 Sep 2023 17:51:44 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id AD90810E3CE; Mon, 4 Sep 2023 17:51:43 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id DA68F10E3CA for ; Mon, 4 Sep 2023 17:51:41 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=6oBzeutVeCkn956eUnxIc5d5ac49gLhcNqAKwTWO3ZM=; b=fO7cnov6eOo3uCSWMC3X96TQvc o4w/Cl7A2hK9dl2nBJmtyHVEbODoCxW53igSz01Q5zpifkzOKa+gj8G4fZpPDMREt/lAyAXs/Lhyd Z5uTy1zsIBkdf0vIQA/AWmLHHgrGCGOus0PIY4lmd77EbeoDNsseicnHsE24APMWuev23JR6eLSq9 MTRvegS1M+Hn6AMdfqK3Ewkm03mKGspT+qb3UmBh1oQMaZeQ4xB9uqosdiqS7gnos0BHjqPf8Xpc3 KeRUE5J8K5JhxNYfH0SL5RqzDIClUMq9XNyppvYuyyKfVoEGInswBVPwcWxx6whhhptCGlWb4CPYJ 7sJ+lVMw==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDjZ-0028nB-0p; Mon, 04 Sep 2023 19:51:33 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 09/17] drm/v3d: Create tracepoints to track the CPU job Date: Mon, 4 Sep 2023 14:38:57 -0300 Message-ID: <20230904175019.1172713-10-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" Create tracepoints to track the three major events of a CPU job lifetime: 1. Submission of a `v3d_submit_cpu` IOCTL 2. Beginning of the execution of a CPU job 3. Ending of the execution of a CPU job Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_sched.c | 4 +++ drivers/gpu/drm/v3d/v3d_submit.c | 2 ++ drivers/gpu/drm/v3d/v3d_trace.h | 57 ++++++++++++++++++++++++++++++++ 3 files changed, 63 insertions(+) diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c index 88c483da360c..85c11e0fe057 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -262,8 +262,12 @@ v3d_cpu_job_run(struct drm_sched_job *sched_job) return NULL; } + trace_v3d_cpu_job_begin(&v3d->drm, job->job_type); + v3d_cpu_job_fn[job->job_type](job); + trace_v3d_cpu_job_end(&v3d->drm, job->job_type); + return NULL; } diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index ff8a77a4e2b0..5402d8aacb71 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -805,6 +805,8 @@ v3d_submit_cpu_ioctl(struct drm_device *dev, void *data, goto fail; } + trace_v3d_submit_cpu_ioctl(&v3d->drm, cpu_job->job_type); + ret = v3d_job_init(v3d, file_priv, (void *)&cpu_job, sizeof(*cpu_job), v3d_job_free, 0, &se, V3D_CPU); if (ret) diff --git a/drivers/gpu/drm/v3d/v3d_trace.h b/drivers/gpu/drm/v3d/v3d_trace.h index 7aa8dc356e54..06086ece6e9e 100644 --- a/drivers/gpu/drm/v3d/v3d_trace.h +++ b/drivers/gpu/drm/v3d/v3d_trace.h @@ -225,6 +225,63 @@ TRACE_EVENT(v3d_submit_csd, __entry->seqno) ); +TRACE_EVENT(v3d_submit_cpu_ioctl, + TP_PROTO(struct drm_device *dev, enum v3d_cpu_job_type job_type), + TP_ARGS(dev, job_type), + + TP_STRUCT__entry( + __field(u32, dev) + __field(enum v3d_cpu_job_type, job_type) + ), + + TP_fast_assign( + __entry->dev = dev->primary->index; + __entry->job_type = job_type; + ), + + TP_printk("dev=%u, job_type=%d", + __entry->dev, + __entry->job_type) +); + +TRACE_EVENT(v3d_cpu_job_begin, + TP_PROTO(struct drm_device *dev, enum v3d_cpu_job_type job_type), + TP_ARGS(dev, job_type), + + TP_STRUCT__entry( + __field(u32, dev) + __field(enum v3d_cpu_job_type, job_type) + ), + + TP_fast_assign( + __entry->dev = dev->primary->index; + __entry->job_type = job_type; + ), + + TP_printk("dev=%u, job_type=%d", + __entry->dev, + __entry->job_type) +); + +TRACE_EVENT(v3d_cpu_job_end, + TP_PROTO(struct drm_device *dev, enum v3d_cpu_job_type job_type), + TP_ARGS(dev, job_type), + + TP_STRUCT__entry( + __field(u32, dev) + __field(enum v3d_cpu_job_type, job_type) + ), + + TP_fast_assign( + __entry->dev = dev->primary->index; + __entry->job_type = job_type; + ), + + TP_printk("dev=%u, job_type=%d", + __entry->dev, + __entry->job_type) +); + TRACE_EVENT(v3d_cache_clean_begin, TP_PROTO(struct drm_device *dev), TP_ARGS(dev), From patchwork Mon Sep 4 17:38:58 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374244 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 24B3AC83F2C for ; Mon, 4 Sep 2023 17:51:49 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 4AFC010E3D1; Mon, 4 Sep 2023 17:51:48 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id D424F10E3C8 for ; Mon, 4 Sep 2023 17:51:45 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=K48fL1OJrSt/3NkK7o6WRqVZPeK/n/GipA+lTpt/Kek=; b=BIl3Tejwv4wvhx0zDge/gWNaWI XjTksqNyevxK6xNoFIZoEeVoWbOTR2ebLnphPI5zcVXyUXA+ugjPTQUCzyhemcupvf/2efPXb9WOY K6Cc6XxisTRv2aCvDImBoGxeYvMuZlmfHxSvvXGlZKkgs8hKKrcqAbBRm8tROYeJhPp7nO2ccg9gw 85p8H185Tcd6dOVxVf4odkqZ7mm+AgswRCT/vcD0qOTU2GZy40saKfhqHlh/4imNZ8hY/MwmMAJU/ GdsmNSHQjvJ80WUUa1X2Vk+N7bs3cniSfSsqokMVyE+82u/pbf4c3CJMbEp3JAAFrdmxlqU+c7LGh DmIEudXQ==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDjd-0028nB-67; Mon, 04 Sep 2023 19:51:37 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 10/17] drm/v3d: Detach the CSD job BO setup Date: Mon, 4 Sep 2023 14:38:58 -0300 Message-ID: <20230904175019.1172713-11-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Melissa Wen Detach CSD job setup from CSD submission ioctl to reuse it in CPU submission ioctl for indirect CSD job. Signed-off-by: Melissa Wen Co-developed-by: Maíra Canal Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_submit.c | 52 +++++++++++++++++++++----------- 1 file changed, 34 insertions(+), 18 deletions(-) diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index 5402d8aacb71..72f006738132 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -268,6 +268,37 @@ v3d_attach_fences_and_unlock_reservation(struct drm_file *file_priv, } } +static int +v3d_setup_csd_jobs_and_bos(struct drm_file *file_priv, + struct v3d_dev *v3d, + struct drm_v3d_submit_csd *args, + struct v3d_csd_job **job, + struct v3d_job **clean_job, + struct v3d_submit_ext *se, + struct ww_acquire_ctx *acquire_ctx) +{ + int ret; + + ret = v3d_job_init(v3d, file_priv, (void *)job, sizeof(**job), + v3d_job_free, args->in_sync, se, V3D_CSD); + if (ret) + return ret; + + ret = v3d_job_init(v3d, file_priv, (void *)clean_job, sizeof(**clean_job), + v3d_job_free, 0, NULL, V3D_CACHE_CLEAN); + if (ret) + return ret; + + (*job)->args = *args; + + ret = v3d_lookup_bos(&v3d->drm, file_priv, *clean_job, + args->bo_handles, args->bo_handle_count); + if (ret) + return ret; + + return v3d_lock_bo_reservations(*clean_job, acquire_ctx); +} + static void v3d_put_multisync_post_deps(struct v3d_submit_ext *se) { @@ -696,24 +727,9 @@ v3d_submit_csd_ioctl(struct drm_device *dev, void *data, } } - ret = v3d_job_init(v3d, file_priv, (void *)&job, sizeof(*job), - v3d_job_free, args->in_sync, &se, V3D_CSD); - if (ret) - goto fail; - - ret = v3d_job_init(v3d, file_priv, (void *)&clean_job, sizeof(*clean_job), - v3d_job_free, 0, NULL, V3D_CACHE_CLEAN); - if (ret) - goto fail; - - job->args = *args; - - ret = v3d_lookup_bos(dev, file_priv, clean_job, - args->bo_handles, args->bo_handle_count); - if (ret) - goto fail; - - ret = v3d_lock_bo_reservations(clean_job, &acquire_ctx); + ret = v3d_setup_csd_jobs_and_bos(file_priv, v3d, args, + &job, &clean_job, &se, + &acquire_ctx); if (ret) goto fail; From patchwork Mon Sep 4 17:38:59 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374245 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 2F3A4C83F2C for ; Mon, 4 Sep 2023 17:51:53 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 8540B10E3DB; Mon, 4 Sep 2023 17:51:52 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id 1100510E3D5 for ; Mon, 4 Sep 2023 17:51:50 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=MInRjyjscEBOnZfW5d5L3p863wPuEBOVo5Iv/be4Zk8=; b=So+ueg8QrmUDwFkZALHXS8bGIc PorGNQgazPp0N9zeLlYiEryZdW2t6gEB1f4tQ5QY+V+V7+FUf8orODe7tUQ4pBuk53hgWVELmOWr1 S+QBzZzmMj3K6m3BO1zASrpQLYdXB9t8+rIl4ij+QuoQfWK3TbrBBQh1/wX0uJ5G9FCiMiYMWs6M+ 5r/hHKJMVxUrqn9jEYABGqtMZz3MK5X+O7dovtg/06GqI46aoOxW7Lqo5Rm1bZR9H6ixU7lvx7c4D GmAXYGlvdUoByHhfbF0FA7NZ060GyqSF4kXQKScP7FGM2kmhgvBxufpUqKTQE+2BQ7i8Tr2vVzKH6 V4/tLCEw==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDjh-0028nB-CW; Mon, 04 Sep 2023 19:51:41 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 11/17] drm/v3d: Enable BO mapping Date: Mon, 4 Sep 2023 14:38:59 -0300 Message-ID: <20230904175019.1172713-12-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" For the indirect CSD CPU job, we will need to access the internal contents of the BO with the dispatch parameters. Therefore, create methods to allow the mapping and unmapping of the BO. Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_bo.c | 18 ++++++++++++++++++ drivers/gpu/drm/v3d/v3d_drv.h | 4 ++++ 2 files changed, 22 insertions(+) diff --git a/drivers/gpu/drm/v3d/v3d_bo.c b/drivers/gpu/drm/v3d/v3d_bo.c index 357a0da7e16a..1bdfac8beafd 100644 --- a/drivers/gpu/drm/v3d/v3d_bo.c +++ b/drivers/gpu/drm/v3d/v3d_bo.c @@ -33,6 +33,9 @@ void v3d_free_object(struct drm_gem_object *obj) struct v3d_dev *v3d = to_v3d_dev(obj->dev); struct v3d_bo *bo = to_v3d_bo(obj); + if (bo->vaddr) + v3d_put_bo_vaddr(bo); + v3d_mmu_remove_ptes(bo); mutex_lock(&v3d->bo_lock); @@ -134,6 +137,7 @@ struct v3d_bo *v3d_bo_create(struct drm_device *dev, struct drm_file *file_priv, if (IS_ERR(shmem_obj)) return ERR_CAST(shmem_obj); bo = to_v3d_bo(&shmem_obj->base); + bo->vaddr = NULL; ret = v3d_bo_create_finish(&shmem_obj->base); if (ret) @@ -167,6 +171,20 @@ v3d_prime_import_sg_table(struct drm_device *dev, return obj; } +void v3d_get_bo_vaddr(struct v3d_bo *bo) +{ + struct drm_gem_shmem_object *obj = &bo->base; + + bo->vaddr = vmap(obj->pages, obj->base.size >> PAGE_SHIFT, VM_MAP, + pgprot_writecombine(PAGE_KERNEL)); +} + +void v3d_put_bo_vaddr(struct v3d_bo *bo) +{ + vunmap(bo->vaddr); + bo->vaddr = NULL; +} + int v3d_create_bo_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { diff --git a/drivers/gpu/drm/v3d/v3d_drv.h b/drivers/gpu/drm/v3d/v3d_drv.h index 2a3f3beb272c..7940b6b1efd0 100644 --- a/drivers/gpu/drm/v3d/v3d_drv.h +++ b/drivers/gpu/drm/v3d/v3d_drv.h @@ -179,6 +179,8 @@ struct v3d_bo { * v3d_render_job->unref_list */ struct list_head unref_head; + + void *vaddr; }; static inline struct v3d_bo * @@ -361,6 +363,8 @@ struct drm_gem_object *v3d_create_object(struct drm_device *dev, size_t size); void v3d_free_object(struct drm_gem_object *gem_obj); struct v3d_bo *v3d_bo_create(struct drm_device *dev, struct drm_file *file_priv, size_t size); +void v3d_get_bo_vaddr(struct v3d_bo *bo); +void v3d_put_bo_vaddr(struct v3d_bo *bo); int v3d_create_bo_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); int v3d_mmap_bo_ioctl(struct drm_device *dev, void *data, From patchwork Mon Sep 4 17:39:00 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374246 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 B1977C71153 for ; Mon, 4 Sep 2023 17:51:55 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id C2ECB10E3C8; Mon, 4 Sep 2023 17:51:54 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id 42BE010E3C8 for ; Mon, 4 Sep 2023 17:51:53 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=Xpk0Kb3qEpNu/FCAwvUOJFSN7smUYntlerddsrCQq1A=; b=D33yCZOIdVuK3xTCwyLgJ8DFTm RAYUjne/rFU26QxcabpKA274/5RKQtwhUAeCl7hWFWigVOIemb9Ijhpe5cS7qcIVA3fq6IzpVdY1Y NvkmW6vrUO44gkjqR56pepQ2tWT2/LJ/2XeTTIF4Tmn+nLubnUzOh7d4gGtZzaBxSPm1k6DQ8Z29U S0PoKmlRxrJPszrCSXkwbQ+QhL07Ypm4wA5l176RAsFUxmgl6w/xoLeEdBQzZKl7mB9RfVI4Xy6Ba 6bDYy68hzjM1X5zDGAlff8Z1GMNEgkO0ifzmLjT/sWrKZzP8PolEvi70aptt/bLpr/vzyqWaUIp+Z jor9wiIg==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDjl-0028nB-HU; Mon, 04 Sep 2023 19:51:46 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 12/17] drm/v3d: Create a CPU job extension for a indirect CSD job Date: Mon, 4 Sep 2023 14:39:00 -0300 Message-ID: <20230904175019.1172713-13-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" A CPU job is a type of job that performs operations that requires CPU intervention. An indirect CSD job is a job that, when executed in the queue, will map the indirect buffer, read the dispatch parameters, and submit a regular dispatch. Therefore, it is a job that needs CPU intervention. So, create a user extension for the CPU job that enables the creation of an indirect CSD. This user extension will allow the creation of a CSD job linked to a CPU job. The CPU job will wait for the indirect CSD job dependencies and, once they are signaled, it will update the CSD job parameters. Co-developed-by: Melissa Wen Signed-off-by: Melissa Wen Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_drv.h | 31 +++++++++- drivers/gpu/drm/v3d/v3d_sched.c | 41 ++++++++++++- drivers/gpu/drm/v3d/v3d_submit.c | 100 ++++++++++++++++++++++++++++++- include/uapi/drm/v3d_drm.h | 37 +++++++++++- 4 files changed, 205 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/v3d/v3d_drv.h b/drivers/gpu/drm/v3d/v3d_drv.h index 7940b6b1efd0..5e44542e7899 100644 --- a/drivers/gpu/drm/v3d/v3d_drv.h +++ b/drivers/gpu/drm/v3d/v3d_drv.h @@ -288,12 +288,41 @@ struct v3d_csd_job { struct drm_v3d_submit_csd args; }; -enum v3d_cpu_job_type {}; +enum v3d_cpu_job_type { + V3D_CPU_JOB_TYPE_INDIRECT_CSD = 1, +}; + +struct v3d_indirect_csd_info { + /* Indirect CSD */ + struct v3d_csd_job *job; + + /* Clean cache job associated to the Indirect CSD job */ + struct v3d_job *clean_job; + + /* Offset within the BO where the workgroup counts are stored */ + u32 offset; + + /* Workgroups size */ + u32 wg_size; + + /* Indices of the uniforms with the workgroup dispatch counts + * in the uniform stream. + */ + u32 wg_uniform_offsets[3]; + + /* Indirect BO */ + struct drm_gem_object *indirect; + + /* Context of the Indirect CSD job */ + struct ww_acquire_ctx acquire_ctx; +}; struct v3d_cpu_job { struct v3d_job base; enum v3d_cpu_job_type job_type; + + struct v3d_indirect_csd_info indirect_csd; }; struct v3d_submit_outsync { diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c index 85c11e0fe057..7d3567c237dc 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -24,6 +24,8 @@ #include "v3d_regs.h" #include "v3d_trace.h" +#define V3D_CSD_CFG012_WG_COUNT_SHIFT 16 + static struct v3d_job * to_v3d_job(struct drm_sched_job *sched_job) { @@ -247,13 +249,50 @@ v3d_csd_job_run(struct drm_sched_job *sched_job) return fence; } +static void +v3d_rewrite_csd_job_wg_counts_from_indirect(struct v3d_cpu_job *job) +{ + struct v3d_indirect_csd_info *indirect_csd = &job->indirect_csd; + struct v3d_bo *bo = to_v3d_bo(job->base.bo[0]); + struct v3d_bo *indirect = to_v3d_bo(indirect_csd->indirect); + struct drm_v3d_submit_csd *args = &indirect_csd->job->args; + u32 *wg_counts; + + v3d_get_bo_vaddr(bo); + v3d_get_bo_vaddr(indirect); + + wg_counts = (uint32_t *) (bo->vaddr + indirect_csd->offset); + + if (wg_counts[0] == 0 || wg_counts[1] == 0 || wg_counts[2] == 0) + return; + + args->cfg[0] = wg_counts[0] << V3D_CSD_CFG012_WG_COUNT_SHIFT; + args->cfg[1] = wg_counts[1] << V3D_CSD_CFG012_WG_COUNT_SHIFT; + args->cfg[2] = wg_counts[2] << V3D_CSD_CFG012_WG_COUNT_SHIFT; + args->cfg[4] = DIV_ROUND_UP(indirect_csd->wg_size, 16) * + (wg_counts[0] * wg_counts[1] * wg_counts[2]) - 1; + + for (int i = 0; i < 3; i++) { + /* 0xffffffff indicates that the uniform rewrite is not needed */ + if (indirect_csd->wg_uniform_offsets[i] != 0xffffffff) { + u32 uniform_idx = indirect_csd->wg_uniform_offsets[i]; + ((uint32_t *) indirect->vaddr)[uniform_idx] = wg_counts[i]; + } + } + + v3d_put_bo_vaddr(indirect); + v3d_put_bo_vaddr(bo); +} + static struct dma_fence * v3d_cpu_job_run(struct drm_sched_job *sched_job) { struct v3d_cpu_job *job = to_cpu_job(sched_job); struct v3d_dev *v3d = job->base.v3d; - void (*v3d_cpu_job_fn[])(struct v3d_cpu_job *job) = { }; + void (*v3d_cpu_job_fn[])(struct v3d_cpu_job *job) = { + [V3D_CPU_JOB_TYPE_INDIRECT_CSD] = v3d_rewrite_csd_job_wg_counts_from_indirect, + }; v3d->cpu_job = job; diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index 72f006738132..cb62d5752d6d 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -395,6 +395,48 @@ v3d_get_multisync_submit_deps(struct drm_file *file_priv, return 0; } +/* Get data for the indirect CSD job submission. */ +static int +v3d_get_cpu_indirect_csd_params(struct drm_file *file_priv, + struct drm_v3d_extension __user *ext, + struct v3d_cpu_job *job) +{ + struct v3d_file_priv *v3d_priv = file_priv->driver_priv; + struct v3d_dev *v3d = v3d_priv->v3d; + struct drm_v3d_indirect_csd indirect_csd; + struct v3d_indirect_csd_info *info = &job->indirect_csd; + + if (!job) { + DRM_DEBUG("CPU job extension was attached to a GPU job.\n"); + return -EINVAL; + } + + if (job->job_type) { + DRM_DEBUG("Two CPU job extensions were added to the same CPU job.\n"); + return -EINVAL; + } + + if (copy_from_user(&indirect_csd, ext, sizeof(indirect_csd))) + return -EFAULT; + + if (!v3d_has_csd(v3d)) { + DRM_DEBUG("Attempting CSD submit on non-CSD hardware.\n"); + return -EINVAL; + } + + job->job_type = V3D_CPU_JOB_TYPE_INDIRECT_CSD; + info->offset = indirect_csd.offset; + info->wg_size = indirect_csd.wg_size; + memcpy(&info->wg_uniform_offsets, &indirect_csd.wg_uniform_offsets, + sizeof(indirect_csd.wg_uniform_offsets)); + + info->indirect = drm_gem_object_lookup(file_priv, indirect_csd.indirect); + + return v3d_setup_csd_jobs_and_bos(file_priv, v3d, &indirect_csd.submit, + &info->job, &info->clean_job, + NULL, &info->acquire_ctx); +} + /* Whenever userspace sets ioctl extensions, v3d_get_extensions parses data * according to the extension id (name). */ @@ -420,6 +462,9 @@ v3d_get_extensions(struct drm_file *file_priv, case DRM_V3D_EXT_ID_MULTI_SYNC: ret = v3d_get_multisync_submit_deps(file_priv, user_ext, se); break; + case DRM_V3D_EXT_ID_CPU_INDIRECT_CSD: + ret = v3d_get_cpu_indirect_csd_params(file_priv, user_ext, job); + break; default: DRM_DEBUG_DRIVER("Unknown extension id: %d\n", ext.id); return -EINVAL; @@ -794,7 +839,10 @@ v3d_submit_cpu_ioctl(struct drm_device *dev, void *data, struct v3d_dev *v3d = to_v3d_dev(dev); struct drm_v3d_submit_cpu *args = data; struct v3d_submit_ext se = {0}; + struct v3d_submit_ext *out_se = NULL; struct v3d_cpu_job *cpu_job = NULL; + struct v3d_csd_job *csd_job = NULL; + struct v3d_job *clean_job = NULL; struct ww_acquire_ctx acquire_ctx; int ret; @@ -828,6 +876,9 @@ v3d_submit_cpu_ioctl(struct drm_device *dev, void *data, if (ret) goto fail; + clean_job = cpu_job->indirect_csd.clean_job; + csd_job = cpu_job->indirect_csd.job; + if (args->bo_handle_count) { ret = v3d_lookup_bos(dev, file_priv, &cpu_job->base, args->bo_handles, args->bo_handle_count); @@ -841,19 +892,66 @@ v3d_submit_cpu_ioctl(struct drm_device *dev, void *data, mutex_lock(&v3d->sched_lock); v3d_push_job(&cpu_job->base); + + switch (cpu_job->job_type) { + case V3D_CPU_JOB_TYPE_INDIRECT_CSD: + ret = drm_sched_job_add_dependency(&csd_job->base.base, + dma_fence_get(cpu_job->base.done_fence)); + if (ret) + goto fail_unreserve; + + v3d_push_job(&csd_job->base); + + ret = drm_sched_job_add_dependency(&clean_job->base, + dma_fence_get(csd_job->base.done_fence)); + if (ret) + goto fail_unreserve; + + v3d_push_job(clean_job); + + break; + default: + break; + } mutex_unlock(&v3d->sched_lock); + out_se = (cpu_job->job_type == V3D_CPU_JOB_TYPE_INDIRECT_CSD) ? NULL : &se; + v3d_attach_fences_and_unlock_reservation(file_priv, &cpu_job->base, &acquire_ctx, 0, - NULL, cpu_job->base.done_fence); + out_se, cpu_job->base.done_fence); + + switch (cpu_job->job_type) { + case V3D_CPU_JOB_TYPE_INDIRECT_CSD: + v3d_attach_fences_and_unlock_reservation(file_priv, + clean_job, + &cpu_job->indirect_csd.acquire_ctx, + 0, &se, clean_job->done_fence); + break; + default: + break; + } v3d_job_put((void *)cpu_job); + v3d_job_put((void *)csd_job); + v3d_job_put(clean_job); return 0; +fail_unreserve: + mutex_unlock(&v3d->sched_lock); + + drm_gem_unlock_reservations(cpu_job->base.bo, cpu_job->base.bo_count, + &acquire_ctx); + + drm_gem_unlock_reservations(clean_job->bo, clean_job->bo_count, + &cpu_job->indirect_csd.acquire_ctx); + fail: v3d_job_cleanup((void *)cpu_job); + v3d_job_cleanup((void *)csd_job); + v3d_job_cleanup(clean_job); v3d_put_multisync_post_deps(&se); return ret; diff --git a/include/uapi/drm/v3d_drm.h b/include/uapi/drm/v3d_drm.h index d63044ebd5e0..a7aae8115263 100644 --- a/include/uapi/drm/v3d_drm.h +++ b/include/uapi/drm/v3d_drm.h @@ -71,7 +71,8 @@ extern "C" { struct drm_v3d_extension { __u64 next; __u32 id; -#define DRM_V3D_EXT_ID_MULTI_SYNC 0x01 +#define DRM_V3D_EXT_ID_MULTI_SYNC 0x01 +#define DRM_V3D_EXT_ID_CPU_INDIRECT_CSD 0x02 __u32 flags; /* mbz */ }; @@ -360,6 +361,40 @@ struct drm_v3d_submit_csd { __u32 pad; }; +/** + * struct drm_v3d_indirect_csd - ioctl extension for the CPU job to create an + * indirect CSD + * + * When an extension of DRM_V3D_EXT_ID_CPU_INDIRECT_CSD id is defined, it + * points to this extension to define a indirect CSD submission. It creates a + * CPU job linked to a CSD job. The CPU job waits for the indirect CSD + * dependencies and, once they are signaled, it updates the CSD job config + * before allowing the CSD job execution. + */ +struct drm_v3d_indirect_csd { + struct drm_v3d_extension base; + + /* Indirect CSD */ + struct drm_v3d_submit_csd submit; + + /* Handle of the indirect BO, that should be also attached to the + * indirect CSD. + */ + __u32 indirect; + + /* Offset within the BO where the workgroup counts are stored */ + __u32 offset; + + /* Workgroups size */ + __u32 wg_size; + + /* Indices of the uniforms with the workgroup dispatch counts + * in the uniform stream. If the uniform rewrite is not needed, + * the offset must be 0xffffffff. + */ + __u32 wg_uniform_offsets[3]; +}; + struct drm_v3d_submit_cpu { /* Pointer to a u32 array of the BOs that are referenced by the job. */ __u64 bo_handles; From patchwork Mon Sep 4 17:39:01 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374247 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 5A32FC83F2C for ; Mon, 4 Sep 2023 17:52:02 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id B9F7410E3C4; Mon, 4 Sep 2023 17:52:01 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id F1A0410E3C4 for ; Mon, 4 Sep 2023 17:51:58 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=sU6IwTGYjfJwAGireGFIYaj+zkTOFsWOGUIIFG4iuVc=; b=oL+hOoYVICohq93xXeDn+wWN1U tGt/d6SehMFmaTnYBCmy9InCjzA7e5peWmNaMFE+9KRM5SjI+XMkh52KYw8NGzF8Twoivq2qzjIBw BnLMy28ks+wG97UF07Qfe/NsqgbGN5V7xV+Oam9AT32znaoHzduEmYp/LqR/EpzCR1vh6iJVP/bCw 7hJowFyZ4wxyl2xF1TK6a2g2Yvdbskck+3YPUO7K7sNyMfvVor0AY4QzrwEdp3siEvzUKKqOs8l8q YPQRNuNNRaNlZiKv3O4hXBZYjlBCTPnCivi5fQhxj2phM7/KNJAzB1OYZ6iDu48RP9Ft8mFLpXUq5 EfdFCgng==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDjp-0028nB-NI; Mon, 04 Sep 2023 19:51:50 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 13/17] drm/v3d: Create a CPU job extension for the timestamp query job Date: Mon, 4 Sep 2023 14:39:01 -0300 Message-ID: <20230904175019.1172713-14-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" A CPU job is a type of job that performs operations that requires CPU intervention. A timestamp query job is a job that calculates the query timestamp and updates the query availability by signaling a syncobj. As V3D doesn't provide any mechanism to obtain a timestamp from the GPU, it is a job that needs CPU intervention. So, create a user extension for the CPU job that enables the creation of a timestamp query job. This user extension will allow the creation of a CPU job that performs the timestamp query calculation and updates the timestamp BO with the proper value. Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_drv.h | 17 +++++++++ drivers/gpu/drm/v3d/v3d_sched.c | 39 +++++++++++++++++++- drivers/gpu/drm/v3d/v3d_submit.c | 62 ++++++++++++++++++++++++++++++++ include/uapi/drm/v3d_drm.h | 27 ++++++++++++++ 4 files changed, 144 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/v3d/v3d_drv.h b/drivers/gpu/drm/v3d/v3d_drv.h index 5e44542e7899..2733f9c9f5f8 100644 --- a/drivers/gpu/drm/v3d/v3d_drv.h +++ b/drivers/gpu/drm/v3d/v3d_drv.h @@ -290,6 +290,15 @@ struct v3d_csd_job { enum v3d_cpu_job_type { V3D_CPU_JOB_TYPE_INDIRECT_CSD = 1, + V3D_CPU_JOB_TYPE_TIMESTAMP_QUERY, +}; + +struct v3d_timestamp_query { + /* Offset of this query in the timestamp BO for its value. */ + u32 offset; + + /* Syncobj that indicates the timestamp availability */ + struct drm_syncobj *syncobj; }; struct v3d_indirect_csd_info { @@ -317,12 +326,20 @@ struct v3d_indirect_csd_info { struct ww_acquire_ctx acquire_ctx; }; +struct v3d_timestamp_query_info { + struct v3d_timestamp_query *queries; + + u32 count; +}; + struct v3d_cpu_job { struct v3d_job base; enum v3d_cpu_job_type job_type; struct v3d_indirect_csd_info indirect_csd; + + struct v3d_timestamp_query_info timestamp_query; }; struct v3d_submit_outsync { diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c index 7d3567c237dc..f2fd7ffc8704 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -18,6 +18,7 @@ * semaphores to interlock between them. */ +#include #include #include "v3d_drv.h" @@ -70,6 +71,21 @@ v3d_sched_job_free(struct drm_sched_job *sched_job) v3d_job_cleanup(job); } +static void +v3d_cpu_job_free(struct drm_sched_job *sched_job) +{ + struct v3d_cpu_job *job = to_cpu_job(sched_job); + struct v3d_timestamp_query_info *timestamp_query = &job->timestamp_query; + + if (timestamp_query->queries) { + for (int i = 0; i < timestamp_query->count; i++) + drm_syncobj_put(timestamp_query->queries[i].syncobj); + kvfree(timestamp_query->queries); + } + + v3d_job_cleanup(&job->base); +} + static void v3d_switch_perfmon(struct v3d_dev *v3d, struct v3d_job *job) { @@ -284,6 +300,26 @@ v3d_rewrite_csd_job_wg_counts_from_indirect(struct v3d_cpu_job *job) v3d_put_bo_vaddr(bo); } +static void +v3d_timestamp_query(struct v3d_cpu_job *job) +{ + struct v3d_timestamp_query_info *timestamp_query = &job->timestamp_query; + struct v3d_bo *bo = to_v3d_bo(job->base.bo[0]); + u8 *value_addr; + + v3d_get_bo_vaddr(bo); + + for (int i = 0; i < timestamp_query->count; i++) { + value_addr = ((u8 *) bo->vaddr) + timestamp_query->queries[i].offset; + *((u64 *) value_addr) = i == 0 ? ktime_get_ns() : 0ull; + + drm_syncobj_replace_fence(timestamp_query->queries[i].syncobj, + job->base.done_fence); + } + + v3d_put_bo_vaddr(bo); +} + static struct dma_fence * v3d_cpu_job_run(struct drm_sched_job *sched_job) { @@ -292,6 +328,7 @@ v3d_cpu_job_run(struct drm_sched_job *sched_job) void (*v3d_cpu_job_fn[])(struct v3d_cpu_job *job) = { [V3D_CPU_JOB_TYPE_INDIRECT_CSD] = v3d_rewrite_csd_job_wg_counts_from_indirect, + [V3D_CPU_JOB_TYPE_TIMESTAMP_QUERY] = v3d_timestamp_query, }; v3d->cpu_job = job; @@ -451,7 +488,7 @@ static const struct drm_sched_backend_ops v3d_cache_clean_sched_ops = { static const struct drm_sched_backend_ops v3d_cpu_sched_ops = { .run_job = v3d_cpu_job_run, .timedout_job = v3d_generic_job_timedout, - .free_job = v3d_sched_job_free + .free_job = v3d_cpu_job_free }; int diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index cb62d5752d6d..5bf55276c4f0 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -437,6 +437,64 @@ v3d_get_cpu_indirect_csd_params(struct drm_file *file_priv, NULL, &info->acquire_ctx); } +/* Get data for the query timestamp job submission. */ +static int +v3d_get_cpu_timestamp_query_params(struct drm_file *file_priv, + struct drm_v3d_extension __user *ext, + struct v3d_cpu_job *job) +{ + u32 __user *offsets, *syncs; + struct drm_v3d_timestamp_query timestamp; + + if (!job) { + DRM_DEBUG("CPU job extension was attached to a GPU job.\n"); + return -EINVAL; + } + + if (job->job_type) { + DRM_DEBUG("Two CPU job extensions were added to the same CPU job.\n"); + return -EINVAL; + } + + if (copy_from_user(×tamp, ext, sizeof(timestamp))) + return -EFAULT; + + if (timestamp.pad) + return -EINVAL; + + job->job_type = V3D_CPU_JOB_TYPE_TIMESTAMP_QUERY; + + job->timestamp_query.queries = kvmalloc_array(timestamp.count, + sizeof(struct v3d_timestamp_query), + GFP_KERNEL); + if (!job->timestamp_query.queries) + return -ENOMEM; + + offsets = u64_to_user_ptr(timestamp.offsets); + syncs = u64_to_user_ptr(timestamp.syncs); + + for (int i = 0; i < timestamp.count; i++) { + u32 offset, sync; + + if (copy_from_user(&offset, offsets++, sizeof(offset))) { + kvfree(job->timestamp_query.queries); + return -EFAULT; + } + + job->timestamp_query.queries[i].offset = offset; + + if (copy_from_user(&sync, syncs++, sizeof(sync))) { + kvfree(job->timestamp_query.queries); + return -EFAULT; + } + + job->timestamp_query.queries[i].syncobj = drm_syncobj_find(file_priv, sync); + } + job->timestamp_query.count = timestamp.count; + + return 0; +} + /* Whenever userspace sets ioctl extensions, v3d_get_extensions parses data * according to the extension id (name). */ @@ -465,6 +523,9 @@ v3d_get_extensions(struct drm_file *file_priv, case DRM_V3D_EXT_ID_CPU_INDIRECT_CSD: ret = v3d_get_cpu_indirect_csd_params(file_priv, user_ext, job); break; + case DRM_V3D_EXT_ID_CPU_TIMESTAMP_QUERY: + ret = v3d_get_cpu_timestamp_query_params(file_priv, user_ext, job); + break; default: DRM_DEBUG_DRIVER("Unknown extension id: %d\n", ext.id); return -EINVAL; @@ -953,6 +1014,7 @@ v3d_submit_cpu_ioctl(struct drm_device *dev, void *data, v3d_job_cleanup((void *)csd_job); v3d_job_cleanup(clean_job); v3d_put_multisync_post_deps(&se); + kvfree(cpu_job->timestamp_query.queries); return ret; } diff --git a/include/uapi/drm/v3d_drm.h b/include/uapi/drm/v3d_drm.h index a7aae8115263..b94edfd275a7 100644 --- a/include/uapi/drm/v3d_drm.h +++ b/include/uapi/drm/v3d_drm.h @@ -73,6 +73,7 @@ struct drm_v3d_extension { __u32 id; #define DRM_V3D_EXT_ID_MULTI_SYNC 0x01 #define DRM_V3D_EXT_ID_CPU_INDIRECT_CSD 0x02 +#define DRM_V3D_EXT_ID_CPU_TIMESTAMP_QUERY 0x03 __u32 flags; /* mbz */ }; @@ -395,6 +396,32 @@ struct drm_v3d_indirect_csd { __u32 wg_uniform_offsets[3]; }; +/** + * struct drm_v3d_timestamp_query - ioctl extension for the CPU job to calculate + * a timestamp query + * + * When an extension DRM_V3D_EXT_ID_TIMESTAMP_QUERY is defined, it points to + * this extension to define a timestamp query submission. This CPU job will + * calculate the timestamp query and update the query value within the + * timestamp BO. Moreover, it will signal the timestamp syncobj to indicate + * query availability. + */ +struct drm_v3d_timestamp_query { + struct drm_v3d_extension base; + + /* Array of queries' offsets within the timestamp BO for their value */ + __u64 offsets; + + /* Array of timestamp's syncobjs to indicate its availability */ + __u64 syncs; + + /* Number of queries */ + __u32 count; + + /* mbz */ + __u32 pad; +}; + struct drm_v3d_submit_cpu { /* Pointer to a u32 array of the BOs that are referenced by the job. */ __u64 bo_handles; From patchwork Mon Sep 4 17:39:02 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374248 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 58469C71153 for ; Mon, 4 Sep 2023 17:52:05 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 726A110E3CA; Mon, 4 Sep 2023 17:52:04 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id 9596E10E3C7 for ; Mon, 4 Sep 2023 17:52:02 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=BAEHVVUOPfzAbAcuqwPvUuLzZpLyIX3OVhHL8pv+Bbg=; b=iIV485TbD1mNhlZJC0hNMOeSPZ 9IlGkNC5QPydPAVgzoSC23kgazwonjk3N+bYrImLqfopBFLY8ViHqkKblsIHbWTPSySAEUjtutN4n v/EFxYt2ykU+wETTJt//lDqtYGLWBiyvcAK1VQ/DKWGq9u8fxozDOoDoSMxea54uvAkYFozGyRvMy TZqsEOfzRyrlseGjZIzoaQeV5n1/1hxDIwjCKhXroEY/6XmjbamHrMktk6gcpyoB04JQ/YVNO6Bgl /LHYQDxPLQEh72jZMJyoX1lqgKjZV9F2PcbqQEbvsuXxmxiBF21MQr5Y9x2mPL17zmOgeBCGEXT8X kkpmUukQ==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDjt-0028nB-Tk; Mon, 04 Sep 2023 19:51:54 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 14/17] drm/v3d: Create a CPU job extension for the reset timestamp job Date: Mon, 4 Sep 2023 14:39:02 -0300 Message-ID: <20230904175019.1172713-15-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" A CPU job is a type of job that performs operations that requires CPU intervention. A reset timestamp job is a job that resets the timestamp queries based on the value offset of the first query. As V3D doesn't provide any mechanism to obtain a timestamp from the GPU, it is a job that needs CPU intervention. So, create a user extension for the CPU job that enables the creation of a reset timestamp job. This user extension will allow the creation of a CPU job that resets the timestamp value in the timestamp BO and resets the availability syncobj. Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_drv.h | 1 + drivers/gpu/drm/v3d/v3d_sched.c | 21 +++++++++++++ drivers/gpu/drm/v3d/v3d_submit.c | 51 ++++++++++++++++++++++++++++++++ include/uapi/drm/v3d_drm.h | 24 +++++++++++++++ 4 files changed, 97 insertions(+) diff --git a/drivers/gpu/drm/v3d/v3d_drv.h b/drivers/gpu/drm/v3d/v3d_drv.h index 2733f9c9f5f8..0ffcc8db155e 100644 --- a/drivers/gpu/drm/v3d/v3d_drv.h +++ b/drivers/gpu/drm/v3d/v3d_drv.h @@ -291,6 +291,7 @@ struct v3d_csd_job { enum v3d_cpu_job_type { V3D_CPU_JOB_TYPE_INDIRECT_CSD = 1, V3D_CPU_JOB_TYPE_TIMESTAMP_QUERY, + V3D_CPU_JOB_TYPE_RESET_TIMESTAMP_QUERY, }; struct v3d_timestamp_query { diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c index f2fd7ffc8704..da8a709ee590 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -320,6 +320,26 @@ v3d_timestamp_query(struct v3d_cpu_job *job) v3d_put_bo_vaddr(bo); } +static void +v3d_reset_timestamp_queries(struct v3d_cpu_job *job) +{ + struct v3d_timestamp_query_info *timestamp_query = &job->timestamp_query; + struct v3d_timestamp_query *queries = timestamp_query->queries; + struct v3d_bo *bo = to_v3d_bo(job->base.bo[0]); + u8 *value_addr; + + v3d_get_bo_vaddr(bo); + + for (int i = 0; i < timestamp_query->count; i++) { + value_addr = ((u8 *) bo->vaddr) + queries[i].offset; + *((u64 *) value_addr) = 0; + + drm_syncobj_replace_fence(queries[i].syncobj, NULL); + } + + v3d_put_bo_vaddr(bo); +} + static struct dma_fence * v3d_cpu_job_run(struct drm_sched_job *sched_job) { @@ -329,6 +349,7 @@ v3d_cpu_job_run(struct drm_sched_job *sched_job) void (*v3d_cpu_job_fn[])(struct v3d_cpu_job *job) = { [V3D_CPU_JOB_TYPE_INDIRECT_CSD] = v3d_rewrite_csd_job_wg_counts_from_indirect, [V3D_CPU_JOB_TYPE_TIMESTAMP_QUERY] = v3d_timestamp_query, + [V3D_CPU_JOB_TYPE_RESET_TIMESTAMP_QUERY] = v3d_reset_timestamp_queries, }; v3d->cpu_job = job; diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index 5bf55276c4f0..abd182b20021 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -495,6 +495,54 @@ v3d_get_cpu_timestamp_query_params(struct drm_file *file_priv, return 0; } +static int +v3d_get_cpu_reset_timestamp_params(struct drm_file *file_priv, + struct drm_v3d_extension __user *ext, + struct v3d_cpu_job *job) +{ + u32 __user *syncs; + struct drm_v3d_reset_timestamp_query reset; + + if (!job) { + DRM_DEBUG("CPU job extension was attached to a GPU job.\n"); + return -EINVAL; + } + + if (job->job_type) { + DRM_DEBUG("Two CPU job extensions were added to the same CPU job.\n"); + return -EINVAL; + } + + if (copy_from_user(&reset, ext, sizeof(reset))) + return -EFAULT; + + job->job_type = V3D_CPU_JOB_TYPE_RESET_TIMESTAMP_QUERY; + + job->timestamp_query.queries = kvmalloc_array(reset.count, + sizeof(struct v3d_timestamp_query), + GFP_KERNEL); + if (!job->timestamp_query.queries) + return -ENOMEM; + + syncs = u64_to_user_ptr(reset.syncs); + + for (int i = 0; i < reset.count; i++) { + u32 sync; + + job->timestamp_query.queries[i].offset = reset.offset + 8 * i; + + if (copy_from_user(&sync, syncs++, sizeof(sync))) { + kvfree(job->timestamp_query.queries); + return -EFAULT; + } + + job->timestamp_query.queries[i].syncobj = drm_syncobj_find(file_priv, sync); + } + job->timestamp_query.count = reset.count; + + return 0; +} + /* Whenever userspace sets ioctl extensions, v3d_get_extensions parses data * according to the extension id (name). */ @@ -526,6 +574,9 @@ v3d_get_extensions(struct drm_file *file_priv, case DRM_V3D_EXT_ID_CPU_TIMESTAMP_QUERY: ret = v3d_get_cpu_timestamp_query_params(file_priv, user_ext, job); break; + case DRM_V3D_EXT_ID_CPU_RESET_TIMESTAMP_QUERY: + ret = v3d_get_cpu_reset_timestamp_params(file_priv, user_ext, job); + break; default: DRM_DEBUG_DRIVER("Unknown extension id: %d\n", ext.id); return -EINVAL; diff --git a/include/uapi/drm/v3d_drm.h b/include/uapi/drm/v3d_drm.h index b94edfd275a7..c93104121d8d 100644 --- a/include/uapi/drm/v3d_drm.h +++ b/include/uapi/drm/v3d_drm.h @@ -74,6 +74,7 @@ struct drm_v3d_extension { #define DRM_V3D_EXT_ID_MULTI_SYNC 0x01 #define DRM_V3D_EXT_ID_CPU_INDIRECT_CSD 0x02 #define DRM_V3D_EXT_ID_CPU_TIMESTAMP_QUERY 0x03 +#define DRM_V3D_EXT_ID_CPU_RESET_TIMESTAMP_QUERY 0x04 __u32 flags; /* mbz */ }; @@ -422,6 +423,29 @@ struct drm_v3d_timestamp_query { __u32 pad; }; +/** + * struct drm_v3d_reset_timestamp_query - ioctl extension for the CPU job to + * reset timestamp queries + * + * When an extension DRM_V3D_EXT_ID_CPU_RESET_TIMESTAMP_QUERY is defined, it + * points to this extension to define a reset timestamp submission. This CPU + * job will reset the timestamp queries based on value offset of the first + * query. Moreover, it will reset the timestamp syncobj to reset query + * availability. + */ +struct drm_v3d_reset_timestamp_query { + struct drm_v3d_extension base; + + /* Array of timestamp's syncobjs to indicate its availability */ + __u64 syncs; + + /* Offset of the first query within the timestamp BO for its value */ + __u32 offset; + + /* Number of queries */ + __u32 count; +}; + struct drm_v3d_submit_cpu { /* Pointer to a u32 array of the BOs that are referenced by the job. */ __u64 bo_handles; From patchwork Mon Sep 4 17:39:03 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374249 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 840A9C71153 for ; Mon, 4 Sep 2023 17:52:08 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id E9CE110E3CB; Mon, 4 Sep 2023 17:52:07 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id 9C16910E3CB for ; Mon, 4 Sep 2023 17:52:05 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=bwp4skxz4wLxqzupwrfBYlS8N2IkhphsSiwq+p/Erhc=; b=o/FLfh7KkNmpO4DVzdgVAnP2Y1 xrhLLUQqa8NadnNRwwQpzP5ULCEORqUiWx8enbraR8o00JIjUIkwMtWIbJSjuTbCB+JqRDUBK41RB WYa7LRWuFj2JHgsTMO7zbg8dgSKSw8ADQaKf1e2sW8uJ7OG/lrH+oi3rY6mAdjrfrcZrIdEXiZfkN uSnzs4duy1gfdYDXT2LYznFwkYSOluVmDhxjM0XjdD2lBvDx5er6Fqbf4d9dNB5M6Wowj6vJbUta/ PksS/2AmqZ2P+fbXqKahovYyGXSAdeoL0UgScOEvMQvS/o71znZOWTshvBSAViNhK+cIRklSo32yJ BU5J4DuQ==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDjy-0028nB-48; Mon, 04 Sep 2023 19:51:58 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 15/17] drm/v3d: Create a CPU job extension to copy timestamp query to a buffer Date: Mon, 4 Sep 2023 14:39:03 -0300 Message-ID: <20230904175019.1172713-16-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" A CPU job is a type of job that performs operations that requires CPU intervention. A copy timestamp query job is a job that copy the complete or partial result of a query to a buffer. As V3D doesn't provide any mechanism to obtain a timestamp from the GPU, it is a job that needs CPU intervention. So, create a user extension for the CPU job that enables the creation of a copy timestamp query job. This user extension will allow the creation of a CPU job that copy the results of a timestamp query to a BO with the possibility to indicate the timestamp availability with a availability bit. Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_drv.h | 20 ++++++++++ drivers/gpu/drm/v3d/v3d_sched.c | 54 +++++++++++++++++++++++++ drivers/gpu/drm/v3d/v3d_submit.c | 68 ++++++++++++++++++++++++++++++++ include/uapi/drm/v3d_drm.h | 41 +++++++++++++++++++ 4 files changed, 183 insertions(+) diff --git a/drivers/gpu/drm/v3d/v3d_drv.h b/drivers/gpu/drm/v3d/v3d_drv.h index 0ffcc8db155e..0cb629b116f1 100644 --- a/drivers/gpu/drm/v3d/v3d_drv.h +++ b/drivers/gpu/drm/v3d/v3d_drv.h @@ -292,6 +292,7 @@ enum v3d_cpu_job_type { V3D_CPU_JOB_TYPE_INDIRECT_CSD = 1, V3D_CPU_JOB_TYPE_TIMESTAMP_QUERY, V3D_CPU_JOB_TYPE_RESET_TIMESTAMP_QUERY, + V3D_CPU_JOB_TYPE_COPY_TIMESTAMP_QUERY, }; struct v3d_timestamp_query { @@ -333,6 +334,23 @@ struct v3d_timestamp_query_info { u32 count; }; +struct v3d_copy_query_results_info { + /* Define if should write to buffer using 64 or 32 bits */ + bool do_64bit; + + /* Define if it can write to buffer even if the query is not available */ + bool do_partial; + + /* Define if it should write availability bit to buffer */ + bool availability_bit; + + /* Offset of the copy buffer in the BO */ + u32 offset; + + /* Stride of the copy buffer in the BO */ + u32 stride; +}; + struct v3d_cpu_job { struct v3d_job base; @@ -341,6 +359,8 @@ struct v3d_cpu_job { struct v3d_indirect_csd_info indirect_csd; struct v3d_timestamp_query_info timestamp_query; + + struct v3d_copy_query_results_info copy; }; struct v3d_submit_outsync { diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c index da8a709ee590..b3664fb5aeef 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -340,6 +340,59 @@ v3d_reset_timestamp_queries(struct v3d_cpu_job *job) v3d_put_bo_vaddr(bo); } +static void +write_to_buffer(void *dst, u32 idx, bool do_64bit, u64 value) +{ + if (do_64bit) { + u64 *dst64 = (u64 *) dst; + dst64[idx] = value; + } else { + u32 *dst32 = (u32 *) dst; + dst32[idx] = (u32) value; + } +} + +static void +v3d_copy_query_results(struct v3d_cpu_job *job) +{ + struct v3d_timestamp_query_info *timestamp_query = &job->timestamp_query; + struct v3d_timestamp_query *queries = timestamp_query->queries; + struct v3d_bo *bo = to_v3d_bo(job->base.bo[0]); + struct v3d_bo *timestamp = to_v3d_bo(job->base.bo[1]); + struct v3d_copy_query_results_info *copy = &job->copy; + struct dma_fence *fence; + u8 *query_addr; + bool available, write_result; + u8 *data; + int i; + + v3d_get_bo_vaddr(bo); + v3d_get_bo_vaddr(timestamp); + + data = ((u8 *) bo->vaddr) + copy->offset; + + for (i = 0; i < timestamp_query->count; i++) { + fence = drm_syncobj_fence_get(queries[i].syncobj); + available = fence ? dma_fence_is_signaled(fence) : false; + + write_result = available || copy->do_partial; + if (write_result) { + query_addr = ((u8 *) timestamp->vaddr) + queries[i].offset; + write_to_buffer(data, 0, copy->do_64bit, *((u64 *) query_addr)); + } + + if (copy->availability_bit) + write_to_buffer(data, 1, copy->do_64bit, available ? 1u : 0u); + + data += copy->stride; + + dma_fence_put(fence); + } + + v3d_put_bo_vaddr(timestamp); + v3d_put_bo_vaddr(bo); +} + static struct dma_fence * v3d_cpu_job_run(struct drm_sched_job *sched_job) { @@ -350,6 +403,7 @@ v3d_cpu_job_run(struct drm_sched_job *sched_job) [V3D_CPU_JOB_TYPE_INDIRECT_CSD] = v3d_rewrite_csd_job_wg_counts_from_indirect, [V3D_CPU_JOB_TYPE_TIMESTAMP_QUERY] = v3d_timestamp_query, [V3D_CPU_JOB_TYPE_RESET_TIMESTAMP_QUERY] = v3d_reset_timestamp_queries, + [V3D_CPU_JOB_TYPE_COPY_TIMESTAMP_QUERY] = v3d_copy_query_results, }; v3d->cpu_job = job; diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index abd182b20021..9a41e8044011 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -543,6 +543,71 @@ v3d_get_cpu_reset_timestamp_params(struct drm_file *file_priv, return 0; } +/* Get data for the copy timestamp query results job submission. */ +static int +v3d_get_cpu_copy_query_results_params(struct drm_file *file_priv, + struct drm_v3d_extension __user *ext, + struct v3d_cpu_job *job) +{ + u32 __user *offsets, *syncs; + struct drm_v3d_copy_timestamp_query copy; + int i; + + if (!job) { + DRM_DEBUG("CPU job extension was attached to a GPU job.\n"); + return -EINVAL; + } + + if (job->job_type) { + DRM_DEBUG("Two CPU job extensions were added to the same CPU job.\n"); + return -EINVAL; + } + + if (copy_from_user(©, ext, sizeof(copy))) + return -EFAULT; + + if (copy.pad) + return -EINVAL; + + job->job_type = V3D_CPU_JOB_TYPE_COPY_TIMESTAMP_QUERY; + + job->timestamp_query.queries = kvmalloc_array(copy.count, + sizeof(struct v3d_timestamp_query), + GFP_KERNEL); + if (!job->timestamp_query.queries) + return -ENOMEM; + + offsets = u64_to_user_ptr(copy.offsets); + syncs = u64_to_user_ptr(copy.syncs); + + for (i = 0; i < copy.count; i++) { + u32 offset, sync; + + if (copy_from_user(&offset, offsets++, sizeof(offset))) { + kvfree(job->timestamp_query.queries); + return -EFAULT; + } + + job->timestamp_query.queries[i].offset = offset; + + if (copy_from_user(&sync, syncs++, sizeof(sync))) { + kvfree(job->timestamp_query.queries); + return -EFAULT; + } + + job->timestamp_query.queries[i].syncobj = drm_syncobj_find(file_priv, sync); + } + job->timestamp_query.count = copy.count; + + job->copy.do_64bit = copy.do_64bit; + job->copy.do_partial = copy.do_partial; + job->copy.availability_bit = copy.availability_bit; + job->copy.offset = copy.offset; + job->copy.stride = copy.stride; + + return 0; +} + /* Whenever userspace sets ioctl extensions, v3d_get_extensions parses data * according to the extension id (name). */ @@ -577,6 +642,9 @@ v3d_get_extensions(struct drm_file *file_priv, case DRM_V3D_EXT_ID_CPU_RESET_TIMESTAMP_QUERY: ret = v3d_get_cpu_reset_timestamp_params(file_priv, user_ext, job); break; + case DRM_V3D_EXT_ID_CPU_COPY_TIMESTAMP_QUERY: + ret = v3d_get_cpu_copy_query_results_params(file_priv, user_ext, job); + break; default: DRM_DEBUG_DRIVER("Unknown extension id: %d\n", ext.id); return -EINVAL; diff --git a/include/uapi/drm/v3d_drm.h b/include/uapi/drm/v3d_drm.h index c93104121d8d..c6b5a2549302 100644 --- a/include/uapi/drm/v3d_drm.h +++ b/include/uapi/drm/v3d_drm.h @@ -75,6 +75,7 @@ struct drm_v3d_extension { #define DRM_V3D_EXT_ID_CPU_INDIRECT_CSD 0x02 #define DRM_V3D_EXT_ID_CPU_TIMESTAMP_QUERY 0x03 #define DRM_V3D_EXT_ID_CPU_RESET_TIMESTAMP_QUERY 0x04 +#define DRM_V3D_EXT_ID_CPU_COPY_TIMESTAMP_QUERY 0x05 __u32 flags; /* mbz */ }; @@ -446,6 +447,46 @@ struct drm_v3d_reset_timestamp_query { __u32 count; }; +/** + * struct drm_v3d_copy_timestamp_query - ioctl extension for the CPU job to copy + * query results to a buffer + * + * When an extension DRM_V3D_EXT_ID_CPU_COPY_TIMESTAMP_QUERY is defined, it + * points to this extension to define a copy timestamp query submission. This + * CPU job will copy the timestamp queries results to a BO with the offset + * and stride defined in the extension. + */ +struct drm_v3d_copy_timestamp_query { + struct drm_v3d_extension base; + + /* Define if should write to buffer using 64 or 32 bits */ + __u8 do_64bit; + + /* Define if it can write to buffer even if the query is not available */ + __u8 do_partial; + + /* Define if it should write availability bit to buffer */ + __u8 availability_bit; + + /* mbz */ + __u8 pad; + + /* Offset of the buffer in the BO */ + __u32 offset; + + /* Stride of the buffer in the BO */ + __u32 stride; + + /* Number of queries */ + __u32 count; + + /* Array of queries' offsets within the timestamp BO for their value */ + __u64 offsets; + + /* Array of timestamp's syncobjs to indicate its availability */ + __u64 syncs; +}; + struct drm_v3d_submit_cpu { /* Pointer to a u32 array of the BOs that are referenced by the job. */ __u64 bo_handles; From patchwork Mon Sep 4 17:39:04 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374250 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 07C3CC71153 for ; Mon, 4 Sep 2023 17:52:13 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 3E9BE10E3CC; Mon, 4 Sep 2023 17:52:12 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id 80F6C10E3CC for ; Mon, 4 Sep 2023 17:52:10 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=ho2T5F2oF3K2oB08862Wy90G/XI6JTpb6knJXvs+p/A=; b=V1I/brDD0C8bpHweiyKKQNQri0 lpwlIp2e9HhBzvNha/uOrKgNIwhbMWPY6dWMbk27PKAFjdDcXcdxz1YRy/75d58xvggNjJr2sGss+ De3XWpbTlL4Vv2Za84SiZTJVccIcXypc+2C1C6azcrBCeCOWOFuJLKvJgTB+ZWNkrBbU6gaAOKspm a/bgmXNnw9dsrnVtj0QLIjhonzuNc3GQrN4gI2uSFefAcjhZFvTIlNTJX2XYksGDW/uFGJS6AdTTt n/PgnHlvFOY1YX5dHEqvbU68gP0etcNXV41VQ5N8uNP8BOSSJX7edTcfQiL2k2Tza2zvymhm05tZK 3UgRqyoQ==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDk2-0028nB-8m; Mon, 04 Sep 2023 19:52:02 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 16/17] drm/v3d: Create a CPU job extension for the reset performance query job Date: Mon, 4 Sep 2023 14:39:04 -0300 Message-ID: <20230904175019.1172713-17-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" A CPU job is a type of job that performs operations that requires CPU intervention. A reset performance query job is a job that resets the performance queries by resetting the values of the perfmons. Moreover, we also reset the syncobjs related to the availability of the query. So, create a user extension for the CPU job that enables the creation of a reset performance job. This user extension will allow the creation of a CPU job that resets the perfmons values and resets the availability syncobj. Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_drv.h | 31 ++++++++++++++ drivers/gpu/drm/v3d/v3d_sched.c | 36 ++++++++++++++++ drivers/gpu/drm/v3d/v3d_submit.c | 73 ++++++++++++++++++++++++++++++++ include/uapi/drm/v3d_drm.h | 27 ++++++++++++ 4 files changed, 167 insertions(+) diff --git a/drivers/gpu/drm/v3d/v3d_drv.h b/drivers/gpu/drm/v3d/v3d_drv.h index 0cb629b116f1..0da88dcea01a 100644 --- a/drivers/gpu/drm/v3d/v3d_drv.h +++ b/drivers/gpu/drm/v3d/v3d_drv.h @@ -241,6 +241,9 @@ struct v3d_job { */ struct v3d_perfmon *perfmon; + /* File descriptor of the process that submitted the job */ + struct drm_file *file; + /* Callback for the freeing of the job on refcount going to 0. */ void (*free)(struct kref *ref); }; @@ -293,6 +296,7 @@ enum v3d_cpu_job_type { V3D_CPU_JOB_TYPE_TIMESTAMP_QUERY, V3D_CPU_JOB_TYPE_RESET_TIMESTAMP_QUERY, V3D_CPU_JOB_TYPE_COPY_TIMESTAMP_QUERY, + V3D_CPU_JOB_TYPE_RESET_PERFORMANCE_QUERY, }; struct v3d_timestamp_query { @@ -303,6 +307,18 @@ struct v3d_timestamp_query { struct drm_syncobj *syncobj; }; +/* Number of perfmons required to handle all supported performance counters */ +#define V3D_MAX_PERFMONS DIV_ROUND_UP(V3D_PERFCNT_NUM, \ + DRM_V3D_MAX_PERF_COUNTERS) + +struct v3d_performance_query { + /* Performance monitor IDs for this query */ + u32 kperfmon_ids[V3D_MAX_PERFMONS]; + + /* Syncobj that indicates the query availability */ + struct drm_syncobj *syncobj; +}; + struct v3d_indirect_csd_info { /* Indirect CSD */ struct v3d_csd_job *job; @@ -334,6 +350,19 @@ struct v3d_timestamp_query_info { u32 count; }; +struct v3d_performance_query_info { + struct v3d_performance_query *queries; + + /* Number of performance queries */ + u32 count; + + /* Number of performance monitors related to that query pool */ + u32 nperfmons; + + /* Number of performance counters related to that query pool */ + u32 ncounters; +}; + struct v3d_copy_query_results_info { /* Define if should write to buffer using 64 or 32 bits */ bool do_64bit; @@ -361,6 +390,8 @@ struct v3d_cpu_job { struct v3d_timestamp_query_info timestamp_query; struct v3d_copy_query_results_info copy; + + struct v3d_performance_query_info performance_query; }; struct v3d_submit_outsync { diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c index b3664fb5aeef..b1964bc75d02 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -76,6 +76,7 @@ v3d_cpu_job_free(struct drm_sched_job *sched_job) { struct v3d_cpu_job *job = to_cpu_job(sched_job); struct v3d_timestamp_query_info *timestamp_query = &job->timestamp_query; + struct v3d_performance_query_info *performance_query = &job->performance_query; if (timestamp_query->queries) { for (int i = 0; i < timestamp_query->count; i++) @@ -83,6 +84,12 @@ v3d_cpu_job_free(struct drm_sched_job *sched_job) kvfree(timestamp_query->queries); } + if (performance_query->queries) { + for (int i = 0; i < performance_query->count; i++) + drm_syncobj_put(performance_query->queries[i].syncobj); + kvfree(performance_query->queries); + } + v3d_job_cleanup(&job->base); } @@ -393,6 +400,34 @@ v3d_copy_query_results(struct v3d_cpu_job *job) v3d_put_bo_vaddr(bo); } +static void +v3d_reset_performance_queries(struct v3d_cpu_job *job) +{ + struct v3d_performance_query_info *performance_query = &job->performance_query; + struct v3d_file_priv *v3d_priv = job->base.file->driver_priv; + struct v3d_dev *v3d = job->base.v3d; + struct v3d_perfmon *perfmon; + + for (int i = 0; i < performance_query->count; i++) { + for (int j = 0; j < performance_query->nperfmons; j++) { + perfmon = v3d_perfmon_find(v3d_priv, + performance_query->queries[i].kperfmon_ids[j]); + if (!perfmon) { + DRM_DEBUG("Failed to find perfmon."); + continue; + } + + v3d_perfmon_stop(v3d, perfmon, false); + + memset(perfmon->values, 0, perfmon->ncounters * sizeof(u64)); + + v3d_perfmon_put(perfmon); + } + + drm_syncobj_replace_fence(performance_query->queries[i].syncobj, NULL); + } +} + static struct dma_fence * v3d_cpu_job_run(struct drm_sched_job *sched_job) { @@ -404,6 +439,7 @@ v3d_cpu_job_run(struct drm_sched_job *sched_job) [V3D_CPU_JOB_TYPE_TIMESTAMP_QUERY] = v3d_timestamp_query, [V3D_CPU_JOB_TYPE_RESET_TIMESTAMP_QUERY] = v3d_reset_timestamp_queries, [V3D_CPU_JOB_TYPE_COPY_TIMESTAMP_QUERY] = v3d_copy_query_results, + [V3D_CPU_JOB_TYPE_RESET_PERFORMANCE_QUERY] = v3d_reset_performance_queries, }; v3d->cpu_job = job; diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index 9a41e8044011..fb4c68352f4f 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -167,6 +167,7 @@ v3d_job_init(struct v3d_dev *v3d, struct drm_file *file_priv, job = *container; job->v3d = v3d; job->free = free; + job->file = file_priv; ret = drm_sched_job_init(&job->base, &v3d_priv->sched_entity[queue], v3d_priv); @@ -608,6 +609,74 @@ v3d_get_cpu_copy_query_results_params(struct drm_file *file_priv, return 0; } +static int +v3d_get_cpu_reset_performance_params(struct drm_file *file_priv, + struct drm_v3d_extension __user *ext, + struct v3d_cpu_job *job) +{ + u32 __user *syncs; + u64 __user *kperfmon_ids; + struct drm_v3d_reset_performance_query reset; + + if (!job) { + DRM_DEBUG("CPU job extension was attached to a GPU job.\n"); + return -EINVAL; + } + + if (job->job_type) { + DRM_DEBUG("Two CPU job extensions were added to the same CPU job.\n"); + return -EINVAL; + } + + if (copy_from_user(&reset, ext, sizeof(reset))) + return -EFAULT; + + job->job_type = V3D_CPU_JOB_TYPE_RESET_PERFORMANCE_QUERY; + + job->performance_query.queries = kvmalloc_array(reset.count, + sizeof(struct v3d_performance_query), + GFP_KERNEL); + if (!job->performance_query.queries) + return -ENOMEM; + + syncs = u64_to_user_ptr(reset.syncs); + kperfmon_ids = u64_to_user_ptr(reset.kperfmon_ids); + + for (int i = 0; i < reset.count; i++) { + u32 sync; + u64 ids; + u32 __user *ids_pointer; + u32 id; + + if (copy_from_user(&sync, syncs++, sizeof(sync))) { + kvfree(job->performance_query.queries); + return -EFAULT; + } + + job->performance_query.queries[i].syncobj = drm_syncobj_find(file_priv, sync); + + if (copy_from_user(&ids, kperfmon_ids++, sizeof(ids))) { + kvfree(job->performance_query.queries); + return -EFAULT; + } + + ids_pointer = u64_to_user_ptr(ids); + + for (int j = 0; j < reset.nperfmons; j++) { + if (copy_from_user(&id, ids_pointer++, sizeof(id))) { + kvfree(job->performance_query.queries); + return -EFAULT; + } + + job->performance_query.queries[i].kperfmon_ids[j] = id; + } + } + job->performance_query.count = reset.count; + job->performance_query.nperfmons = reset.nperfmons; + + return 0; +} + /* Whenever userspace sets ioctl extensions, v3d_get_extensions parses data * according to the extension id (name). */ @@ -645,6 +714,9 @@ v3d_get_extensions(struct drm_file *file_priv, case DRM_V3D_EXT_ID_CPU_COPY_TIMESTAMP_QUERY: ret = v3d_get_cpu_copy_query_results_params(file_priv, user_ext, job); break; + case DRM_V3D_EXT_ID_CPU_RESET_PERFORMANCE_QUERY: + ret = v3d_get_cpu_reset_performance_params(file_priv, user_ext, job); + break; default: DRM_DEBUG_DRIVER("Unknown extension id: %d\n", ext.id); return -EINVAL; @@ -1134,6 +1206,7 @@ v3d_submit_cpu_ioctl(struct drm_device *dev, void *data, v3d_job_cleanup(clean_job); v3d_put_multisync_post_deps(&se); kvfree(cpu_job->timestamp_query.queries); + kvfree(cpu_job->performance_query.queries); return ret; } diff --git a/include/uapi/drm/v3d_drm.h b/include/uapi/drm/v3d_drm.h index c6b5a2549302..6d5303413758 100644 --- a/include/uapi/drm/v3d_drm.h +++ b/include/uapi/drm/v3d_drm.h @@ -76,6 +76,7 @@ struct drm_v3d_extension { #define DRM_V3D_EXT_ID_CPU_TIMESTAMP_QUERY 0x03 #define DRM_V3D_EXT_ID_CPU_RESET_TIMESTAMP_QUERY 0x04 #define DRM_V3D_EXT_ID_CPU_COPY_TIMESTAMP_QUERY 0x05 +#define DRM_V3D_EXT_ID_CPU_RESET_PERFORMANCE_QUERY 0x06 __u32 flags; /* mbz */ }; @@ -487,6 +488,32 @@ struct drm_v3d_copy_timestamp_query { __u64 syncs; }; +/** + * struct drm_v3d_reset_performance_query - ioctl extension for the CPU job to + * reset performance queries + * + * When an extension DRM_V3D_EXT_ID_CPU_RESET_PERFORMANCE_QUERY is defined, it + * points to this extension to define a reset performance submission. This CPU + * job will reset the performance queries by resetting the values of the + * performance monitors. Moreover, it will reset the syncobj to reset query + * availability. + */ +struct drm_v3d_reset_performance_query { + struct drm_v3d_extension base; + + /* Array of performance queries's syncobjs to indicate its availability */ + __u64 syncs; + + /* Number of queries */ + __u32 count; + + /* Number of performance monitors */ + __u32 nperfmons; + + /* Array of u64 user-pointers that point to an array of kperfmon_ids */ + __u64 kperfmon_ids; +}; + struct drm_v3d_submit_cpu { /* Pointer to a u32 array of the BOs that are referenced by the job. */ __u64 bo_handles; From patchwork Mon Sep 4 17:39:05 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?Ma=C3=ADra_Canal?= X-Patchwork-Id: 13374251 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 96CA7C83F2C for ; Mon, 4 Sep 2023 17:52:17 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 0D15010E3C6; Mon, 4 Sep 2023 17:52:17 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine2.igalia.com [213.97.179.56]) by gabe.freedesktop.org (Postfix) with ESMTPS id 9DC9810E3D0 for ; Mon, 4 Sep 2023 17:52:14 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=igalia.com; s=20170329; h=Content-Transfer-Encoding:Content-Type:MIME-Version:References: In-Reply-To:Message-ID:Date:Subject:Cc:To:From:Sender:Reply-To:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=ZB9+cCi/d/S5uS91D1+Bf3nvzYEBLJy08A/X+YcNzi4=; b=eLK7vmaPX56LVU2iituMfvkFV9 f919feWSNQHcgTDx2U3MQfxmlJ7LEVMZrxFkMeQFJD5esiMzUhsk+x3H7eXvDTVKDoMC/8aOhF5s/ ck0y03edmxt9uQiCFUxktOl+pcV8BWBYX+189ixZgaN/CSYPlmzw/IHm3+GizI0oIjQ1tugm/bEqX gbXtkOL5GjwKWqB6OxoBWeOw02eLiRCMGeto0zFo5LggPQTve1w8QkmYJHh2cNCy511aMSOYFb2+L r1xpgvAJ4hVjl+3iwc6rs7bFkaOwZOeGQz3iOge5JS74BS5TLoSvg1w70RXiSzwkQ5iijQatvQ/LQ tQafnLaw==; Received: from [187.36.235.191] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1qdDk6-0028nB-FM; Mon, 04 Sep 2023 19:52:07 +0200 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Emma Anholt , Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH 17/17] drm/v3d: Create a CPU job extension for the copy performance query job Date: Mon, 4 Sep 2023 14:39:05 -0300 Message-ID: <20230904175019.1172713-18-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230904175019.1172713-1-mcanal@igalia.com> References: <20230904175019.1172713-1-mcanal@igalia.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: =?utf-8?q?Ma=C3=ADra_Canal?= , Boris Brezillon , dri-devel@lists.freedesktop.org, Faith Ekstrand Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" A CPU job is a type of job that performs operations that requires CPU intervention. A copy performance query job is a job that copy the complete or partial result of a query to a buffer. In order to copy the result of a performance query to a buffer, we need to get the values from the performance monitors. So, create a user extension for the CPU job that enables the creation of a copy performance query job. This user extension will allow the creation of a CPU job that copy the results of a performance query to a BO with the possibility to indicate the availability with a availability bit. Signed-off-by: Maíra Canal --- drivers/gpu/drm/v3d/v3d_drv.h | 1 + drivers/gpu/drm/v3d/v3d_sched.c | 66 ++++++++++++++++++++++++++ drivers/gpu/drm/v3d/v3d_submit.c | 81 ++++++++++++++++++++++++++++++++ include/uapi/drm/v3d_drm.h | 47 ++++++++++++++++++ 4 files changed, 195 insertions(+) diff --git a/drivers/gpu/drm/v3d/v3d_drv.h b/drivers/gpu/drm/v3d/v3d_drv.h index 0da88dcea01a..1852b144e737 100644 --- a/drivers/gpu/drm/v3d/v3d_drv.h +++ b/drivers/gpu/drm/v3d/v3d_drv.h @@ -297,6 +297,7 @@ enum v3d_cpu_job_type { V3D_CPU_JOB_TYPE_RESET_TIMESTAMP_QUERY, V3D_CPU_JOB_TYPE_COPY_TIMESTAMP_QUERY, V3D_CPU_JOB_TYPE_RESET_PERFORMANCE_QUERY, + V3D_CPU_JOB_TYPE_COPY_PERFORMANCE_QUERY, }; struct v3d_timestamp_query { diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c index b1964bc75d02..2ca714e0df40 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -428,6 +428,71 @@ v3d_reset_performance_queries(struct v3d_cpu_job *job) } } +static void +v3d_write_performance_query_result(struct v3d_cpu_job *job, void *data, u32 query) +{ + struct v3d_performance_query_info *performance_query = &job->performance_query; + struct v3d_copy_query_results_info *copy = &job->copy; + struct v3d_file_priv *v3d_priv = job->base.file->driver_priv; + struct v3d_dev *v3d = job->base.v3d; + struct v3d_perfmon *perfmon; + u64 counter_values[V3D_PERFCNT_NUM]; + + for (int i = 0; i < performance_query->nperfmons; i++) { + perfmon = v3d_perfmon_find(v3d_priv, + performance_query->queries[query].kperfmon_ids[i]); + if (!perfmon) { + DRM_DEBUG("Failed to find perfmon."); + continue; + } + + v3d_perfmon_stop(v3d, perfmon, true); + + memcpy(&counter_values[i * DRM_V3D_MAX_PERF_COUNTERS], perfmon->values, + perfmon->ncounters * sizeof(u64)); + + v3d_perfmon_put(perfmon); + } + + for (int i = 0; i < performance_query->ncounters; i++) + write_to_buffer(data, i, copy->do_64bit, counter_values[i]); +} + + +static void +v3d_copy_performance_query(struct v3d_cpu_job *job) +{ + struct v3d_performance_query_info *performance_query = &job->performance_query; + struct v3d_copy_query_results_info *copy = &job->copy; + struct v3d_bo *bo = to_v3d_bo(job->base.bo[0]); + struct dma_fence *fence; + bool available, write_result; + u8 *data; + + v3d_get_bo_vaddr(bo); + + data = ((u8 *) bo->vaddr) + copy->offset; + + for (int i = 0; i < performance_query->count; i++) { + fence = drm_syncobj_fence_get(performance_query->queries[i].syncobj); + available = fence ? dma_fence_is_signaled(fence) : false; + + write_result = available || copy->do_partial; + if (write_result) + v3d_write_performance_query_result(job, data, i); + + if (copy->availability_bit) + write_to_buffer(data, performance_query->ncounters, + copy->do_64bit, available ? 1u : 0u); + + data += copy->stride; + + dma_fence_put(fence); + } + + v3d_put_bo_vaddr(bo); +} + static struct dma_fence * v3d_cpu_job_run(struct drm_sched_job *sched_job) { @@ -440,6 +505,7 @@ v3d_cpu_job_run(struct drm_sched_job *sched_job) [V3D_CPU_JOB_TYPE_RESET_TIMESTAMP_QUERY] = v3d_reset_timestamp_queries, [V3D_CPU_JOB_TYPE_COPY_TIMESTAMP_QUERY] = v3d_copy_query_results, [V3D_CPU_JOB_TYPE_RESET_PERFORMANCE_QUERY] = v3d_reset_performance_queries, + [V3D_CPU_JOB_TYPE_COPY_PERFORMANCE_QUERY] = v3d_copy_performance_query, }; v3d->cpu_job = job; diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c index fb4c68352f4f..9f8d2caa69c7 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -677,6 +677,84 @@ v3d_get_cpu_reset_performance_params(struct drm_file *file_priv, return 0; } +static int +v3d_get_cpu_copy_performance_query_params(struct drm_file *file_priv, + struct drm_v3d_extension __user *ext, + struct v3d_cpu_job *job) +{ + u32 __user *syncs; + u64 __user *kperfmon_ids; + struct drm_v3d_copy_performance_query copy; + + if (!job) { + DRM_DEBUG("CPU job extension was attached to a GPU job.\n"); + return -EINVAL; + } + + if (job->job_type) { + DRM_DEBUG("Two CPU job extensions were added to the same CPU job.\n"); + return -EINVAL; + } + + if (copy_from_user(©, ext, sizeof(copy))) + return -EFAULT; + + if (copy.pad) + return -EINVAL; + + job->job_type = V3D_CPU_JOB_TYPE_COPY_PERFORMANCE_QUERY; + + job->performance_query.queries = kvmalloc_array(copy.count, + sizeof(struct v3d_performance_query), + GFP_KERNEL); + if (!job->performance_query.queries) + return -ENOMEM; + + syncs = u64_to_user_ptr(copy.syncs); + kperfmon_ids = u64_to_user_ptr(copy.kperfmon_ids); + + for (int i = 0; i < copy.count; i++) { + u32 sync; + u64 ids; + u32 __user *ids_pointer; + u32 id; + + if (copy_from_user(&sync, syncs++, sizeof(sync))) { + kvfree(job->performance_query.queries); + return -EFAULT; + } + + job->performance_query.queries[i].syncobj = drm_syncobj_find(file_priv, sync); + + if (copy_from_user(&ids, kperfmon_ids++, sizeof(ids))) { + kvfree(job->performance_query.queries); + return -EFAULT; + } + + ids_pointer = u64_to_user_ptr(ids); + + for (int j = 0; j < copy.nperfmons; j++) { + if (copy_from_user(&id, ids_pointer++, sizeof(id))) { + kvfree(job->performance_query.queries); + return -EFAULT; + } + + job->performance_query.queries[i].kperfmon_ids[j] = id; + } + } + job->performance_query.count = copy.count; + job->performance_query.nperfmons = copy.nperfmons; + job->performance_query.ncounters = copy.ncounters; + + job->copy.do_64bit = copy.do_64bit; + job->copy.do_partial = copy.do_partial; + job->copy.availability_bit = copy.availability_bit; + job->copy.offset = copy.offset; + job->copy.stride = copy.stride; + + return 0; +} + /* Whenever userspace sets ioctl extensions, v3d_get_extensions parses data * according to the extension id (name). */ @@ -717,6 +795,9 @@ v3d_get_extensions(struct drm_file *file_priv, case DRM_V3D_EXT_ID_CPU_RESET_PERFORMANCE_QUERY: ret = v3d_get_cpu_reset_performance_params(file_priv, user_ext, job); break; + case DRM_V3D_EXT_ID_CPU_COPY_PERFORMANCE_QUERY: + ret = v3d_get_cpu_copy_performance_query_params(file_priv, user_ext, job); + break; default: DRM_DEBUG_DRIVER("Unknown extension id: %d\n", ext.id); return -EINVAL; diff --git a/include/uapi/drm/v3d_drm.h b/include/uapi/drm/v3d_drm.h index 6d5303413758..0cde24e10b8d 100644 --- a/include/uapi/drm/v3d_drm.h +++ b/include/uapi/drm/v3d_drm.h @@ -77,6 +77,7 @@ struct drm_v3d_extension { #define DRM_V3D_EXT_ID_CPU_RESET_TIMESTAMP_QUERY 0x04 #define DRM_V3D_EXT_ID_CPU_COPY_TIMESTAMP_QUERY 0x05 #define DRM_V3D_EXT_ID_CPU_RESET_PERFORMANCE_QUERY 0x06 +#define DRM_V3D_EXT_ID_CPU_COPY_PERFORMANCE_QUERY 0x07 __u32 flags; /* mbz */ }; @@ -514,6 +515,52 @@ struct drm_v3d_reset_performance_query { __u64 kperfmon_ids; }; +/** + * struct drm_v3d_copy_performance_query - ioctl extension for the CPU job to copy + * performance query results to a buffer + * + * When an extension DRM_V3D_EXT_ID_CPU_COPY_PERFORMANCE_QUERY is defined, it + * points to this extension to define a copy performance query submission. This + * CPU job will copy the performance queries results to a BO with the offset + * and stride defined in the extension. + */ +struct drm_v3d_copy_performance_query { + struct drm_v3d_extension base; + + /* Define if should write to buffer using 64 or 32 bits */ + __u8 do_64bit; + + /* Define if it can write to buffer even if the query is not available */ + __u8 do_partial; + + /* Define if it should write availability bit to buffer */ + __u8 availability_bit; + + /* mbz */ + __u8 pad; + + /* Offset of the buffer in the BO */ + __u32 offset; + + /* Stride of the buffer in the BO */ + __u32 stride; + + /* Number of performance monitors */ + __u32 nperfmons; + + /* Number of performance counters related to this query pool */ + __u32 ncounters; + + /* Number of queries */ + __u32 count; + + /* Array of performance queries's syncobjs to indicate its availability */ + __u64 syncs; + + /* Array of u64 user-pointers that point to an array of kperfmon_ids */ + __u64 kperfmon_ids; +}; + struct drm_v3d_submit_cpu { /* Pointer to a u32 array of the BOs that are referenced by the job. */ __u64 bo_handles;