From patchwork Fri Nov 24 00:47:03 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: 13466858 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 DD892C61D97 for ; Fri, 24 Nov 2023 01:27:25 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 2973510E789; Fri, 24 Nov 2023 01:27:25 +0000 (UTC) Received: from fanzine2.igalia.com (fanzine.igalia.com [178.60.130.6]) by gabe.freedesktop.org (Postfix) with ESMTPS id 0C3A010E78D for ; Fri, 24 Nov 2023 01:27:17 +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=s3SdGgNJHCrwYoB2DrVu+TaAL62rMVjsQhsjuLsmVpA=; b=iFypP272yTvMGjlpLg/cOotfcm sAsuGNbNCXV/nkh5P82pgEUdc7VALYsuNS8LpFqj+Vi0o3lEJjQzFWxH+6E3y9wshgktl7uz/2D+C //bA14KeQLibQFUgfONnw5Gq1cht+Vu1rGYgzwJO8LmqraLXLxoJ6yby5si4sY0zbvl34wmTo49cN Y8qqIktBmHmI6gFCcwl8cbecJVowLHDeHhRWGomZTq2ng5v8/O8HjX2iJL38P5K4wHtnsY+aHrAt6 dSnE7pKybJD9fbORkG/QUrbk4i2sfYWLWTenRk8aeimaalsnuDDd3LS3MHSbuM56d0yJpY2ekpraI hufRq2VQ==; Received: from [177.34.168.16] (helo=morissey..) by fanzine2.igalia.com with esmtpsa (Cipher TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256) (Exim) id 1r6KyL-006Zga-Gg; Fri, 24 Nov 2023 02:27:10 +0100 From: =?utf-8?q?Ma=C3=ADra_Canal?= To: Melissa Wen , Iago Toral , David Airlie , Daniel Vetter , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Subject: [PATCH v2 07/17] drm/v3d: Add a CPU job submission Date: Thu, 23 Nov 2023 21:47:03 -0300 Message-ID: <20231124012548.772095-8-mcanal@igalia.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20231124012548.772095-1-mcanal@igalia.com> References: <20231124012548.772095-1-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?= , kernel-dev@igalia.com, 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 | 14 +++++- drivers/gpu/drm/v3d/v3d_sched.c | 57 +++++++++++++++++++++++ drivers/gpu/drm/v3d/v3d_submit.c | 79 ++++++++++++++++++++++++++++++++ include/uapi/drm/v3d_drm.h | 17 +++++++ 5 files changed, 170 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/v3d/v3d_drv.c b/drivers/gpu/drm/v3d/v3d_drv.c index 44a1ca57d6a4..3debf37e7d9b 100644 --- a/drivers/gpu/drm/v3d/v3d_drv.c +++ b/drivers/gpu/drm/v3d/v3d_drv.c @@ -91,6 +91,9 @@ 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; @@ -189,6 +192,7 @@ 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 4db9ace66024..010b146140ad 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) static inline char *v3d_queue_to_string(enum v3d_queue queue) { @@ -29,6 +29,7 @@ static inline char *v3d_queue_to_string(enum v3d_queue queue) case V3D_TFU: return "tfu"; case V3D_CSD: return "csd"; case V3D_CACHE_CLEAN: return "cache_clean"; + case V3D_CPU: return "cpu"; } return "UNKNOWN"; } @@ -122,6 +123,7 @@ 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]; @@ -312,6 +314,14 @@ 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; }; @@ -414,6 +424,8 @@ 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 fccbea2a5f2e..a32c91b94898 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -55,6 +55,12 @@ 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) { @@ -262,6 +268,42 @@ 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; + struct v3d_file_priv *file = job->base.file->driver_priv; + u64 runtime; + + 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; + } + + file->start_ns[V3D_CPU] = local_clock(); + v3d->queue[V3D_CPU].start_ns = file->start_ns[V3D_CPU]; + + v3d_cpu_job_fn[job->job_type](job); + + runtime = local_clock() - file->start_ns[V3D_CPU]; + + file->enabled_ns[V3D_CPU] += runtime; + v3d->queue[V3D_CPU].enabled_ns += runtime; + + file->jobs_sent[V3D_CPU]++; + v3d->queue[V3D_CPU].jobs_sent++; + + file->start_ns[V3D_CPU] = 0; + v3d->queue[V3D_CPU].start_ns = 0; + + return NULL; +} + static struct dma_fence * v3d_cache_clean_job_run(struct drm_sched_job *sched_job) { @@ -416,6 +458,12 @@ 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) { @@ -471,6 +519,15 @@ v3d_sched_init(struct v3d_dev *v3d) goto fail; } + ret = drm_sched_init(&v3d->queue[V3D_CPU].sched, + &v3d_cpu_sched_ops, NULL, + DRM_SCHED_PRIORITY_COUNT, + 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 ed1a310bbd2f..d4b85ab16345 100644 --- a/drivers/gpu/drm/v3d/v3d_submit.c +++ b/drivers/gpu/drm/v3d/v3d_submit.c @@ -761,3 +761,82 @@ 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 1a7d7a689de3..00abef9d0db7 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 { @@ -361,6 +365,19 @@ 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,