From patchwork Fri Jul 9 19:31:33 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503347 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=Tm0DMcFe; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3BT5ZJ5z9sPf for ; Sat, 10 Jul 2021 05:29:21 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229962AbhGITcF (ORCPT ); Fri, 9 Jul 2021 15:32:05 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43726 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229506AbhGITcE (ORCPT ); Fri, 9 Jul 2021 15:32:04 -0400 Received: from mail-wr1-x436.google.com (mail-wr1-x436.google.com [IPv6:2a00:1450:4864:20::436]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 34761C0613DD for ; Fri, 9 Jul 2021 12:29:20 -0700 (PDT) Received: by mail-wr1-x436.google.com with SMTP id m2so2533237wrq.2 for ; Fri, 09 Jul 2021 12:29:20 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=HC3uZ/xCD4mCXwrSpW0A6Uy0b/o4faXsI0P7Ey8sdRo=; b=Tm0DMcFeSZ1rrp40F1HZXC/7MwGoWHYym+E+mOAYbQDolOGmXeKqeKUwYh9fGtkPFu eEED9B9km8kbV9VFDxPlMtgbKJoRfdPaS4HyoIJRDbrQJm/BrX8Qve1apNDcNkXIOg0V +aB1ObsZbgIrg1UG/GkuZ/ZBi1DwhPNXBCrMcKZLaHeFKPuIp97QiM5eVyd3KXdi8j/b ooiCfQ6FT2uSPHlNzqvxU/myTh/4e+Pom8VoSo/+KjX7aDxHdLP7VIDlKGUq+c/oKkcK Pt9YOZi4LZEBh55pvi666DE2GB/Y+ar/JR5vT7dICPdvjVKKQcoAhFpU5VGWtIMFuIOM Cahg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=HC3uZ/xCD4mCXwrSpW0A6Uy0b/o4faXsI0P7Ey8sdRo=; b=XcdNqhgVajUdcZ/irkRPeDvq8j2/OPmbrfneTy7vDYU7ED99nRONxyaMb7iJvPieqX vBKFI6SMlvZR868+bw1je92jwwBizXPn8/nE8vaXvlozViUJ9rIgB/sEYc++MEUMxnBU GQ4Jbbb7u05kQOacuPNQ/N/G+ZqJTkao2Mzr/0Sn1a6RSxd60siKGbyrex+kL5Mh14ym 1LzTyoPbTnc526w45czJj4CZ/XWwXg1UnFDv+s4WuNqGrDQq4O8I3rW6PRqmD3L+3GuF 0OJRVC39LjmGqV16nj73HVszaqBDRmK5z7+h3vV5WMuO/NFV9BSFOQuYOae+XUk7bEOj CMVQ== X-Gm-Message-State: AOAM532+abhPZwAI8p7Abub/2uuJqKgOuAUQ8A7+K5QaWqWgRBwiVw0N JkCaR5SFTumBCwEwaUgwKyk= X-Google-Smtp-Source: ABdhPJxIGDdtCachg9fY0Rgh43/bW9N0VPL6OoetPtqvrJwDxJZvI1dNytEKWP5/LKwYARHiMmmD3A== X-Received: by 2002:a5d:6b8d:: with SMTP id n13mr13516326wrx.258.1625858958746; Fri, 09 Jul 2021 12:29:18 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id f14sm12661030wmq.10.2021.07.09.12.29.17 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:17 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org, kernel test robot Subject: [PATCH v8 01/14] gpu: host1x: Add DMA fence implementation Date: Fri, 9 Jul 2021 21:31:33 +0200 Message-Id: <20210709193146.2859516-2-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen Add an implementation of dma_fences based on syncpoints. Syncpoint interrupts are used to signal fences. Additionally, after software signaling has been enabled, a 30 second timeout is started. If the syncpoint threshold is not reached within this period, the fence is signalled with an -ETIMEDOUT error code. This is to allow fences that would never reach their syncpoint threshold to be cleaned up. The timeout can potentially be removed in the future after job tracking code has been refactored. Signed-off-by: Mikko Perttunen Reported-by: kernel test robot Signed-off-by: Thierry Reding --- drivers/gpu/host1x/Makefile | 1 + drivers/gpu/host1x/fence.c | 209 ++++++++++++++++++++++++++++++++++++ drivers/gpu/host1x/fence.h | 13 +++ drivers/gpu/host1x/intr.c | 9 ++ drivers/gpu/host1x/intr.h | 2 + include/linux/host1x.h | 3 + 6 files changed, 237 insertions(+) create mode 100644 drivers/gpu/host1x/fence.c create mode 100644 drivers/gpu/host1x/fence.h diff --git a/drivers/gpu/host1x/Makefile b/drivers/gpu/host1x/Makefile index 096017b8789d..d2b6f7de0498 100644 --- a/drivers/gpu/host1x/Makefile +++ b/drivers/gpu/host1x/Makefile @@ -9,6 +9,7 @@ host1x-y = \ job.o \ debug.o \ mipi.o \ + fence.o \ hw/host1x01.o \ hw/host1x02.o \ hw/host1x04.o \ diff --git a/drivers/gpu/host1x/fence.c b/drivers/gpu/host1x/fence.c new file mode 100644 index 000000000000..06c6b86237bd --- /dev/null +++ b/drivers/gpu/host1x/fence.c @@ -0,0 +1,209 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Syncpoint dma_fence implementation + * + * Copyright (c) 2020, NVIDIA Corporation. + */ + +#include +#include +#include +#include +#include + +#include "fence.h" +#include "intr.h" +#include "syncpt.h" + +DEFINE_SPINLOCK(lock); + +struct host1x_syncpt_fence { + struct dma_fence base; + + atomic_t signaling; + + struct host1x_syncpt *sp; + u32 threshold; + + struct host1x_waitlist *waiter; + void *waiter_ref; + + struct delayed_work timeout_work; +}; + +static const char *host1x_syncpt_fence_get_driver_name(struct dma_fence *f) +{ + return "host1x"; +} + +static const char *host1x_syncpt_fence_get_timeline_name(struct dma_fence *f) +{ + return "syncpoint"; +} + +static struct host1x_syncpt_fence *to_host1x_fence(struct dma_fence *f) +{ + return container_of(f, struct host1x_syncpt_fence, base); +} + +static bool host1x_syncpt_fence_enable_signaling(struct dma_fence *f) +{ + struct host1x_syncpt_fence *sf = to_host1x_fence(f); + int err; + + if (host1x_syncpt_is_expired(sf->sp, sf->threshold)) + return false; + + dma_fence_get(f); + + /* + * The dma_fence framework requires the fence driver to keep a + * reference to any fences for which 'enable_signaling' has been + * called (and that have not been signalled). + * + * We provide a userspace API to create arbitrary syncpoint fences, + * so we cannot normally guarantee that all fences get signalled. + * As such, setup a timeout, so that long-lasting fences will get + * reaped eventually. + */ + schedule_delayed_work(&sf->timeout_work, msecs_to_jiffies(30000)); + + err = host1x_intr_add_action(sf->sp->host, sf->sp, sf->threshold, + HOST1X_INTR_ACTION_SIGNAL_FENCE, f, + sf->waiter, &sf->waiter_ref); + if (err) { + cancel_delayed_work_sync(&sf->timeout_work); + dma_fence_put(f); + return false; + } + + /* intr framework takes ownership of waiter */ + sf->waiter = NULL; + + /* + * The fence may get signalled at any time after the above call, + * so we need to initialize all state used by signalling + * before it. + */ + + return true; +} + +static void host1x_syncpt_fence_release(struct dma_fence *f) +{ + struct host1x_syncpt_fence *sf = to_host1x_fence(f); + + kfree(sf->waiter); + dma_fence_free(f); +} + +const struct dma_fence_ops host1x_syncpt_fence_ops = { + .get_driver_name = host1x_syncpt_fence_get_driver_name, + .get_timeline_name = host1x_syncpt_fence_get_timeline_name, + .enable_signaling = host1x_syncpt_fence_enable_signaling, + .release = host1x_syncpt_fence_release, +}; + +void host1x_fence_signal(struct host1x_syncpt_fence *f) +{ + if (atomic_xchg(&f->signaling, 1)) + return; + + /* + * Cancel pending timeout work - if it races, it will + * not get 'f->signaling' and return. + */ + cancel_delayed_work_sync(&f->timeout_work); + + host1x_intr_put_ref(f->sp->host, f->sp->id, f->waiter_ref, false); + + dma_fence_signal(&f->base); + dma_fence_put(&f->base); +} + +static void do_fence_timeout(struct work_struct *work) +{ + struct delayed_work *dwork = (struct delayed_work *)work; + struct host1x_syncpt_fence *f = + container_of(dwork, struct host1x_syncpt_fence, timeout_work); + + if (atomic_xchg(&f->signaling, 1)) + return; + + /* + * Cancel pending timeout work - if it races, it will + * not get 'f->signaling' and return. + */ + host1x_intr_put_ref(f->sp->host, f->sp->id, f->waiter_ref, true); + + dma_fence_set_error(&f->base, -ETIMEDOUT); + dma_fence_signal(&f->base); + dma_fence_put(&f->base); +} + +struct dma_fence *host1x_fence_create(struct host1x_syncpt *sp, u32 threshold) +{ + struct host1x_syncpt_fence *fence; + + fence = kzalloc(sizeof(*fence), GFP_KERNEL); + if (!fence) + return ERR_PTR(-ENOMEM); + + fence->waiter = kzalloc(sizeof(*fence->waiter), GFP_KERNEL); + if (!fence->waiter) + return ERR_PTR(-ENOMEM); + + fence->sp = sp; + fence->threshold = threshold; + + dma_fence_init(&fence->base, &host1x_syncpt_fence_ops, &lock, + dma_fence_context_alloc(1), 0); + + INIT_DELAYED_WORK(&fence->timeout_work, do_fence_timeout); + + return &fence->base; +} +EXPORT_SYMBOL(host1x_fence_create); + +int host1x_fence_create_fd(struct host1x_syncpt *sp, u32 threshold) +{ + struct sync_file *file; + struct dma_fence *f; + int fd; + + f = host1x_fence_create(sp, threshold); + if (IS_ERR(f)) + return PTR_ERR(f); + + fd = get_unused_fd_flags(O_CLOEXEC); + if (fd < 0) { + dma_fence_put(f); + return fd; + } + + file = sync_file_create(f); + dma_fence_put(f); + if (!file) + return -ENOMEM; + + fd_install(fd, file->file); + + return fd; +} +EXPORT_SYMBOL(host1x_fence_create_fd); + +int host1x_fence_extract(struct dma_fence *fence, u32 *id, u32 *threshold) +{ + struct host1x_syncpt_fence *f; + + if (fence->ops != &host1x_syncpt_fence_ops) + return -EINVAL; + + f = container_of(fence, struct host1x_syncpt_fence, base); + + *id = f->sp->id; + *threshold = f->threshold; + + return 0; +} +EXPORT_SYMBOL(host1x_fence_extract); diff --git a/drivers/gpu/host1x/fence.h b/drivers/gpu/host1x/fence.h new file mode 100644 index 000000000000..70c91de82f14 --- /dev/null +++ b/drivers/gpu/host1x/fence.h @@ -0,0 +1,13 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2020, NVIDIA Corporation. + */ + +#ifndef HOST1X_FENCE_H +#define HOST1X_FENCE_H + +struct host1x_syncpt_fence; + +void host1x_fence_signal(struct host1x_syncpt_fence *fence); + +#endif diff --git a/drivers/gpu/host1x/intr.c b/drivers/gpu/host1x/intr.c index 6d1f3c0fdbe7..45b6be927ec4 100644 --- a/drivers/gpu/host1x/intr.c +++ b/drivers/gpu/host1x/intr.c @@ -13,6 +13,7 @@ #include #include "channel.h" #include "dev.h" +#include "fence.h" #include "intr.h" /* Wait list management */ @@ -121,12 +122,20 @@ static void action_wakeup_interruptible(struct host1x_waitlist *waiter) wake_up_interruptible(wq); } +static void action_signal_fence(struct host1x_waitlist *waiter) +{ + struct host1x_syncpt_fence *f = waiter->data; + + host1x_fence_signal(f); +} + typedef void (*action_handler)(struct host1x_waitlist *waiter); static const action_handler action_handlers[HOST1X_INTR_ACTION_COUNT] = { action_submit_complete, action_wakeup, action_wakeup_interruptible, + action_signal_fence, }; static void run_handlers(struct list_head completed[HOST1X_INTR_ACTION_COUNT]) diff --git a/drivers/gpu/host1x/intr.h b/drivers/gpu/host1x/intr.h index 6ea55e615e3a..e4c346099273 100644 --- a/drivers/gpu/host1x/intr.h +++ b/drivers/gpu/host1x/intr.h @@ -33,6 +33,8 @@ enum host1x_intr_action { */ HOST1X_INTR_ACTION_WAKEUP_INTERRUPTIBLE, + HOST1X_INTR_ACTION_SIGNAL_FENCE, + HOST1X_INTR_ACTION_COUNT }; diff --git a/include/linux/host1x.h b/include/linux/host1x.h index 9b0487c88571..59e4a3bea0b0 100644 --- a/include/linux/host1x.h +++ b/include/linux/host1x.h @@ -170,6 +170,9 @@ u32 host1x_syncpt_base_id(struct host1x_syncpt_base *base); void host1x_syncpt_release_vblank_reservation(struct host1x_client *client, u32 syncpt_id); +struct dma_fence *host1x_fence_create(struct host1x_syncpt *sp, u32 threshold); +int host1x_fence_extract(struct dma_fence *fence, u32 *id, u32 *threshold); + /* * host1x channel */ From patchwork Fri Jul 9 19:31:34 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503348 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=Da3mU7OF; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3BX4sKMz9sPf for ; Sat, 10 Jul 2021 05:29:24 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229546AbhGITcH (ORCPT ); Fri, 9 Jul 2021 15:32:07 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43738 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229506AbhGITcH (ORCPT ); Fri, 9 Jul 2021 15:32:07 -0400 Received: from mail-wm1-x334.google.com (mail-wm1-x334.google.com [IPv6:2a00:1450:4864:20::334]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id D397DC0613DD for ; Fri, 9 Jul 2021 12:29:22 -0700 (PDT) Received: by mail-wm1-x334.google.com with SMTP id q18-20020a1ce9120000b02901f259f3a250so6943913wmc.2 for ; Fri, 09 Jul 2021 12:29:22 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=7RL+gelPRYv1G4mkEK207BAuP32P4MhtT+apt0LN9h0=; b=Da3mU7OFnOmi8aDlmuid7jJW6s5L4PZFXtbdk866ry5O3vO5nnLfK4lsscF/1BItgu GcCjVX7mJLPEsTcpsrlkzVAm+InZzN/zI5vUEuy+X25C2MocA/7pUGis8E9F162NoBmU u0cyNG/C+tcIa/d/5nN6dfL0z5VUrVt9Kvc3gwrIRp8nOEiurs4lz5fTE91uTENXKYvX P9MURKWZg5YjgChPmFiVciMwlrKS46OB7wfB7JKa2fTGXUlpr9w+t5FrjpsAThj89C8v 5dOwF6p9NGn+NeaHlT+MBMHim3lXsQ750Xt7MliEXBYzVXV+n7is/ECL8rkJZaIs/nho pxMQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=7RL+gelPRYv1G4mkEK207BAuP32P4MhtT+apt0LN9h0=; b=lKGpZ4ZmSAGezop4XPqlSHs+Mj6OrcMjK9sERE9m64mkMS6/p/RoraP6UOwVgmiUOh BLoNTzoK6HCeyH0AkDSBY22xIt3FwEEmmvaK74/Uk5r19hUtl/MC1KNf/GXJAByUSkJy q6tbQ9P9+itMRrTcA0wY2ZE0rRfsiMYmd9OvpdbJhL5b5SlECcVCv9Nlmwh86hDgYjyW GVleSZjXZxwd78/WWorAkdsx6fN4lRuuUES8srYTf9tKp9Vb1hr1Y8bXmNlMLPdmcTdH aTL638ZuSqHgo3FhSYhxo7ezKcfxcCqf7XlY0GNSCEW1Xz/xcuelC9MHz1ET7gQBVlGI EeOA== X-Gm-Message-State: AOAM53395whoDlLoCfwcFvVIkKg1eslxmJIXh7NqH0PEw8g2eW10WrY/ aRhg3IIpbokA5R3Pu3As4RtWUaXq9rw= X-Google-Smtp-Source: ABdhPJzKd3NFsuCNsaMjdB4WcFHwHYKGEX9XAfnZxsrEjVl40+MNPfQAfpCaiWa4bib0ijlduHqxrA== X-Received: by 2002:a05:600c:3541:: with SMTP id i1mr508522wmq.135.1625858961472; Fri, 09 Jul 2021 12:29:21 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id o7sm7121431wrv.72.2021.07.09.12.29.19 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:20 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org Subject: [PATCH v8 02/14] gpu: host1x: Add no-recovery mode Date: Fri, 9 Jul 2021 21:31:34 +0200 Message-Id: <20210709193146.2859516-3-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen Add a new property for jobs to enable or disable recovery i.e. CPU increments of syncpoints to max value on job timeout. This allows for a more solid model for hanged jobs, where userspace doesn't need to guess if a syncpoint increment happened because the job completed, or because job timeout was triggered. On job timeout, we stop the channel, NOP all future jobs on the channel using the same syncpoint, mark the syncpoint as locked and resume the channel from the next job, if any. The future jobs are NOPed, since because we don't do the CPU increments, the value of the syncpoint is no longer synchronized, and any waiters would become confused if a future job incremented the syncpoint. The syncpoint is marked locked to ensure that any future jobs cannot increment the syncpoint either, until the application has recognized the situation and reallocated the syncpoint. Signed-off-by: Mikko Perttunen Signed-off-by: Thierry Reding --- drivers/gpu/drm/tegra/drm.c | 1 + drivers/gpu/host1x/cdma.c | 58 ++++++++++++++++++++++++++---- drivers/gpu/host1x/hw/channel_hw.c | 2 +- drivers/gpu/host1x/job.c | 4 +++ drivers/gpu/host1x/syncpt.c | 2 ++ drivers/gpu/host1x/syncpt.h | 12 +++++++ include/linux/host1x.h | 9 +++++ 7 files changed, 81 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index f96c237b2242..739250acd498 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -201,6 +201,7 @@ int tegra_drm_submit(struct tegra_drm_context *context, job->client = client; job->class = client->class; job->serialize = true; + job->syncpt_recovery = true; /* * Track referenced BOs so that they can be unreferenced after the diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c index 6e6ca774f68d..765e5aa64eb6 100644 --- a/drivers/gpu/host1x/cdma.c +++ b/drivers/gpu/host1x/cdma.c @@ -312,10 +312,6 @@ static void update_cdma_locked(struct host1x_cdma *cdma) bool signal = false; struct host1x_job *job, *n; - /* If CDMA is stopped, queue is cleared and we can return */ - if (!cdma->running) - return; - /* * Walk the sync queue, reading the sync point registers as necessary, * to consume as many sync queue entries as possible without blocking @@ -324,7 +320,8 @@ static void update_cdma_locked(struct host1x_cdma *cdma) struct host1x_syncpt *sp = job->syncpt; /* Check whether this syncpt has completed, and bail if not */ - if (!host1x_syncpt_is_expired(sp, job->syncpt_end)) { + if (!host1x_syncpt_is_expired(sp, job->syncpt_end) && + !job->cancelled) { /* Start timer on next pending syncpt */ if (job->timeout) cdma_start_timer_locked(cdma, job); @@ -413,8 +410,11 @@ void host1x_cdma_update_sync_queue(struct host1x_cdma *cdma, else restart_addr = cdma->last_pos; + if (!job) + goto resume; + /* do CPU increments for the remaining syncpts */ - if (job) { + if (job->syncpt_recovery) { dev_dbg(dev, "%s: perform CPU incr on pending buffers\n", __func__); @@ -433,8 +433,44 @@ void host1x_cdma_update_sync_queue(struct host1x_cdma *cdma, dev_dbg(dev, "%s: finished sync_queue modification\n", __func__); + } else { + struct host1x_job *failed_job = job; + + host1x_job_dump(dev, job); + + host1x_syncpt_set_locked(job->syncpt); + failed_job->cancelled = true; + + list_for_each_entry_continue(job, &cdma->sync_queue, list) { + unsigned int i; + + if (job->syncpt != failed_job->syncpt) + continue; + + for (i = 0; i < job->num_slots; i++) { + unsigned int slot = (job->first_get/8 + i) % + HOST1X_PUSHBUFFER_SLOTS; + u32 *mapped = cdma->push_buffer.mapped; + + /* + * Overwrite opcodes with 0 word writes + * to offset 0xbad. This does nothing but + * has a easily detected signature in debug + * traces. + */ + mapped[2*slot+0] = 0x1bad0000; + mapped[2*slot+1] = 0x1bad0000; + } + + job->cancelled = true; + } + + wmb(); + + update_cdma_locked(cdma); } +resume: /* roll back DMAGET and start up channel again */ host1x_hw_cdma_resume(host1x, cdma, restart_addr); } @@ -490,6 +526,16 @@ int host1x_cdma_begin(struct host1x_cdma *cdma, struct host1x_job *job) mutex_lock(&cdma->lock); + /* + * Check if syncpoint was locked due to previous job timeout. + * This needs to be done within the cdma lock to avoid a race + * with the timeout handler. + */ + if (job->syncpt->locked) { + mutex_unlock(&cdma->lock); + return -EPERM; + } + if (job->timeout) { /* init state on first submit with timeout value */ if (!cdma->timeout.initialized) { diff --git a/drivers/gpu/host1x/hw/channel_hw.c b/drivers/gpu/host1x/hw/channel_hw.c index d4c28faf27d1..bf21512e5078 100644 --- a/drivers/gpu/host1x/hw/channel_hw.c +++ b/drivers/gpu/host1x/hw/channel_hw.c @@ -191,7 +191,7 @@ static int channel_submit(struct host1x_job *job) /* schedule a submit complete interrupt */ err = host1x_intr_add_action(host, sp, syncval, HOST1X_INTR_ACTION_SUBMIT_COMPLETE, ch, - completed_waiter, NULL); + completed_waiter, &job->waiter); completed_waiter = NULL; WARN(err, "Failed to set submit complete interrupt"); diff --git a/drivers/gpu/host1x/job.c b/drivers/gpu/host1x/job.c index adbdc225de8d..8f59b34672c2 100644 --- a/drivers/gpu/host1x/job.c +++ b/drivers/gpu/host1x/job.c @@ -79,6 +79,10 @@ static void job_free(struct kref *ref) { struct host1x_job *job = container_of(ref, struct host1x_job, ref); + if (job->waiter) + host1x_intr_put_ref(job->syncpt->host, job->syncpt->id, + job->waiter, false); + if (job->syncpt) host1x_syncpt_put(job->syncpt); diff --git a/drivers/gpu/host1x/syncpt.c b/drivers/gpu/host1x/syncpt.c index e648ebbb2027..d198a10848c6 100644 --- a/drivers/gpu/host1x/syncpt.c +++ b/drivers/gpu/host1x/syncpt.c @@ -407,6 +407,8 @@ static void syncpt_release(struct kref *ref) atomic_set(&sp->max_val, host1x_syncpt_read(sp)); + sp->locked = false; + mutex_lock(&sp->host->syncpt_mutex); host1x_syncpt_base_free(sp->base); diff --git a/drivers/gpu/host1x/syncpt.h b/drivers/gpu/host1x/syncpt.h index a6766f8d55ee..93e894677d89 100644 --- a/drivers/gpu/host1x/syncpt.h +++ b/drivers/gpu/host1x/syncpt.h @@ -40,6 +40,13 @@ struct host1x_syncpt { /* interrupt data */ struct host1x_syncpt_intr intr; + + /* + * If a submission incrementing this syncpoint fails, lock it so that + * further submission cannot be made until application has handled the + * failure. + */ + bool locked; }; /* Initialize sync point array */ @@ -115,4 +122,9 @@ static inline int host1x_syncpt_is_valid(struct host1x_syncpt *sp) return sp->id < host1x_syncpt_nb_pts(sp->host); } +static inline void host1x_syncpt_set_locked(struct host1x_syncpt *sp) +{ + sp->locked = true; +} + #endif diff --git a/include/linux/host1x.h b/include/linux/host1x.h index 59e4a3bea0b0..5fc12db94ca1 100644 --- a/include/linux/host1x.h +++ b/include/linux/host1x.h @@ -237,9 +237,15 @@ struct host1x_job { u32 syncpt_incrs; u32 syncpt_end; + /* Completion waiter ref */ + void *waiter; + /* Maximum time to wait for this job */ unsigned int timeout; + /* Job has timed out and should be released */ + bool cancelled; + /* Index and number of slots used in the push buffer */ unsigned int first_get; unsigned int num_slots; @@ -260,6 +266,9 @@ struct host1x_job { /* Add a channel wait for previous ops to complete */ bool serialize; + + /* Fast-forward syncpoint increments on job timeout */ + bool syncpt_recovery; }; struct host1x_job *host1x_job_alloc(struct host1x_channel *ch, From patchwork Fri Jul 9 19:31:35 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503349 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=KZ/5ng4l; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3BZ2ldpz9sPf for ; Sat, 10 Jul 2021 05:29:26 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229960AbhGITcJ (ORCPT ); Fri, 9 Jul 2021 15:32:09 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43750 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229506AbhGITcJ (ORCPT ); Fri, 9 Jul 2021 15:32:09 -0400 Received: from mail-wr1-x430.google.com (mail-wr1-x430.google.com [IPv6:2a00:1450:4864:20::430]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 995DDC0613E5 for ; Fri, 9 Jul 2021 12:29:24 -0700 (PDT) Received: by mail-wr1-x430.google.com with SMTP id f17so13567756wrt.6 for ; Fri, 09 Jul 2021 12:29:24 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=wqKuapy9SSK1BNEtt7I3vm19kWoM0u9dRsDQpIz91YE=; b=KZ/5ng4lI18PjBq8JWbfWGmgy39CNUaVOfbHQdGrs7D2XZlurUnpamIntEfKPvAzhP RgfC+5X7r6yH1h0txtO9UbokDtVLQL182MHeLSGXZgnAF7A6IeoV1wQn6McNWmdVRC9U JfuFaDx7n2hBIKPGKKQImW9XWx3dzjpmoZxSARp7ErQv1RyhWrapjjvE39ar0PiS2eCQ e+xTCBlbfdpuS8S5vhEeM/KwvEw1l8JFCcdHwMEZZBoQeMteJrSbq8g606GVyl0DXuxa 9mLFrdJZzLOIomlCg2NCWr4pqSjDoIjlNvoWitRwHSJLdQlRec3rujF/dwr4/1dlIf6c q/eQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=wqKuapy9SSK1BNEtt7I3vm19kWoM0u9dRsDQpIz91YE=; b=TrKovQcMxYUjcQkykciXrqKwlBuDTxikdQ+Lp54OuTV5jfF9ZJ+Lxd70fqdG9Ibc8I Lgw9U/8IY0JYfBRnCczdTkbl1MNaVaTnU84sT/HhnqrKhyvBArEv3r7XiiAOs5u/T1uo jW/bHZ1KxSnWw8Z1ey7oV3XenPeGloe2PDsR25vyprON7SK1mjbcyzWq/vNW4X4Ruofp t51Nrm7dV0w5MANsKGifRonEZra9HjeyEicuFWU5SQIrZxx1SmI2MZdlifRZ7mq25ggS fICFn6+hi3ayGyK6OU5sn07tTgGy8DiGpl6Oky0XZ+IXTXx0QGHLRDlcaaBSKwpSMXuI uV0Q== X-Gm-Message-State: AOAM532XXf2CA6UnhajQz1wMRiDvre2yUIH1LkZaJmJj+Nw1vlEX1XaJ S6eHwxiwizpQMkikAT++Mgo= X-Google-Smtp-Source: ABdhPJzpKOb3/gITc4oC81kG3kO9WX6j6UMKXAlwMta+4QKqXBJFjpGOWJAzVj4CkvL/baoaLG9ffw== X-Received: by 2002:a5d:4e08:: with SMTP id p8mr43710096wrt.425.1625858963283; Fri, 09 Jul 2021 12:29:23 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id c8sm6084190wri.91.2021.07.09.12.29.22 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:22 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org Subject: [PATCH v8 03/14] gpu: host1x: Add job release callback Date: Fri, 9 Jul 2021 21:31:35 +0200 Message-Id: <20210709193146.2859516-4-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen Add a callback field to the job structure, to be called just before the job is to be freed. This allows the job's submitter to clean up any of its own state, like decrement runtime PM refcounts. Signed-off-by: Mikko Perttunen Signed-off-by: Thierry Reding --- drivers/gpu/host1x/job.c | 3 +++ include/linux/host1x.h | 4 ++++ 2 files changed, 7 insertions(+) diff --git a/drivers/gpu/host1x/job.c b/drivers/gpu/host1x/job.c index 8f59b34672c2..09097e19c0d0 100644 --- a/drivers/gpu/host1x/job.c +++ b/drivers/gpu/host1x/job.c @@ -79,6 +79,9 @@ static void job_free(struct kref *ref) { struct host1x_job *job = container_of(ref, struct host1x_job, ref); + if (job->release) + job->release(job); + if (job->waiter) host1x_intr_put_ref(job->syncpt->host, job->syncpt->id, job->waiter, false); diff --git a/include/linux/host1x.h b/include/linux/host1x.h index 5fc12db94ca1..7eecb0aa2519 100644 --- a/include/linux/host1x.h +++ b/include/linux/host1x.h @@ -269,6 +269,10 @@ struct host1x_job { /* Fast-forward syncpoint increments on job timeout */ bool syncpt_recovery; + + /* Callback called when job is freed */ + void (*release)(struct host1x_job *job); + void *user_data; }; struct host1x_job *host1x_job_alloc(struct host1x_channel *ch, From patchwork Fri Jul 9 19:31:36 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503350 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=OXoyKMbE; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3Bc4sRlz9sPf for ; Sat, 10 Jul 2021 05:29:28 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229882AbhGITcL (ORCPT ); Fri, 9 Jul 2021 15:32:11 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43760 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229506AbhGITcL (ORCPT ); Fri, 9 Jul 2021 15:32:11 -0400 Received: from mail-wr1-x429.google.com (mail-wr1-x429.google.com [IPv6:2a00:1450:4864:20::429]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id C7DE3C0613DD for ; Fri, 9 Jul 2021 12:29:26 -0700 (PDT) Received: by mail-wr1-x429.google.com with SMTP id d2so13620897wrn.0 for ; Fri, 09 Jul 2021 12:29:26 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=Uql+uqMmHf8IV309hcTHPDwLdQ1a3ayCzh2I0ShAFPo=; b=OXoyKMbE0c3LtEeUELEjRu0A+3TdBerFbvheofFtzaYcpl5ySgx0zBcgVmhAGk4hCH H7Z6FMtLnFRYfmLxwJyuu/4qcY/R1zk+QAoO1L3lFTSGCQF6I1mtvZdLsyYZ72iZ98Th 5LMEAtc2xpGs/6YDqIjtJeF/jNwPWRv+HTBvxFeTHFT0sUFzSrExkKyk4N+crDuna4BV n0PIGizj18EQ6Xar/aiXZ5+crGHbyosZbUQoMdS1dJAzworeO/ZNvubAyYOKay0hEXTu pAnONZdkJOAKYASkgvpqyv1pKqkIgJ5SVc6k7/xAtlc/EOhIPMyMnL2ZUDvXm//lBpBX 0+bg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=Uql+uqMmHf8IV309hcTHPDwLdQ1a3ayCzh2I0ShAFPo=; b=jNcZTZ2I21IOQPuCwPhTyTl6FAk89Pi0/m8STaUmSC7Lj1oi8p1qq2GiMHQNV4cpbq Ig9LKHCjnjYSf9gmaFzOhwoAiDc5gHXRRJZbOON0qAuIyYga6lVmcyx6zocsH+YxXmgw dKuToJQE7J90jskYBOyFQ7/zlaixxVH7TEAH7WsL3TgSfHPii5kfAbi572Wm4cqeAtaE Dh/JNQXAWdj727XcEqRO+LZSv8T1O2SJEgAB9nDRYgKWm7y0vxjVs3SKjerZTcLabd5F U5gBUslQoKLE4ICItNwKWgPOgAqRwuHwZIhncDV5ngei5j34XX6tT5rS9TXAdRsRsuNI /7fQ== X-Gm-Message-State: AOAM532iMvMjh8bQb6nSrzWqZr7WIx2vdGH/tqefQ3fW8A2cUKV5nl+U brU4qM7OWaxWq5/axg5oLT1k5K5bKuI= X-Google-Smtp-Source: ABdhPJyqJM95BGHfLSfvga1mmOEMLpAW37gZvLediv1WKUGanj4SbXpqJZEQtUr/G2J1NRukbz2YFQ== X-Received: by 2002:a5d:64c5:: with SMTP id f5mr31657461wri.277.1625858965348; Fri, 09 Jul 2021 12:29:25 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id o7sm7121625wrv.72.2021.07.09.12.29.24 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:24 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org Subject: [PATCH v8 04/14] gpu: host1x: Add support for syncpoint waits in CDMA pushbuffer Date: Fri, 9 Jul 2021 21:31:36 +0200 Message-Id: <20210709193146.2859516-5-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen Add support for inserting syncpoint waits in the CDMA pushbuffer. These waits need to be done in HOST1X class, while gather submitted by the application execute in engine class. Support is added by converting the gather list of job into a command list that can include both gathers and waits. When the job is submitted, these commands are pushed as the appropriate opcodes on the CDMA pushbuffer. Also supported are waits relative to the start of the job, which are useful for jobs doing multiple things with an engine that doesn't natively support pipelining. While at it, use 32-bit waits on chips that support them. Signed-off-by: Mikko Perttunen Signed-off-by: Thierry Reding --- drivers/gpu/host1x/hw/channel_hw.c | 85 +++++++++++++++++----- drivers/gpu/host1x/hw/debug_hw.c | 9 ++- drivers/gpu/host1x/hw/hw_host1x02_uclass.h | 12 +++ drivers/gpu/host1x/hw/hw_host1x04_uclass.h | 12 +++ drivers/gpu/host1x/hw/hw_host1x05_uclass.h | 12 +++ drivers/gpu/host1x/hw/hw_host1x06_uclass.h | 12 +++ drivers/gpu/host1x/hw/hw_host1x07_uclass.h | 12 +++ drivers/gpu/host1x/job.c | 70 +++++++++++++----- drivers/gpu/host1x/job.h | 16 ++++ include/linux/host1x.h | 6 +- 10 files changed, 203 insertions(+), 43 deletions(-) diff --git a/drivers/gpu/host1x/hw/channel_hw.c b/drivers/gpu/host1x/hw/channel_hw.c index bf21512e5078..1999780a7203 100644 --- a/drivers/gpu/host1x/hw/channel_hw.c +++ b/drivers/gpu/host1x/hw/channel_hw.c @@ -47,39 +47,84 @@ static void trace_write_gather(struct host1x_cdma *cdma, struct host1x_bo *bo, } } -static void submit_gathers(struct host1x_job *job) +static void submit_wait(struct host1x_cdma *cdma, u32 id, u32 threshold, + u32 next_class) +{ +#if HOST1X_HW >= 2 + host1x_cdma_push_wide(cdma, + host1x_opcode_setclass( + HOST1X_CLASS_HOST1X, + HOST1X_UCLASS_LOAD_SYNCPT_PAYLOAD_32, + /* WAIT_SYNCPT_32 is at SYNCPT_PAYLOAD_32+2 */ + BIT(0) | BIT(2) + ), + threshold, + id, + host1x_opcode_setclass(next_class, 0, 0) + ); +#else + /* TODO add waitchk or use waitbases or other mitigation */ + host1x_cdma_push(cdma, + host1x_opcode_setclass( + HOST1X_CLASS_HOST1X, + host1x_uclass_wait_syncpt_r(), + BIT(0) + ), + host1x_class_host_wait_syncpt(id, threshold) + ); + host1x_cdma_push(cdma, + host1x_opcode_setclass(next_class, 0, 0), + HOST1X_OPCODE_NOP + ); +#endif +} + +static void submit_gathers(struct host1x_job *job, u32 job_syncpt_base) { struct host1x_cdma *cdma = &job->channel->cdma; #if HOST1X_HW < 6 struct device *dev = job->channel->dev; #endif unsigned int i; + u32 threshold; - for (i = 0; i < job->num_gathers; i++) { - struct host1x_job_gather *g = &job->gathers[i]; - dma_addr_t addr = g->base + g->offset; - u32 op2, op3; + for (i = 0; i < job->num_cmds; i++) { + struct host1x_job_cmd *cmd = &job->cmds[i]; - op2 = lower_32_bits(addr); - op3 = upper_32_bits(addr); + if (cmd->is_wait) { + if (cmd->wait.relative) + threshold = job_syncpt_base + cmd->wait.threshold; + else + threshold = cmd->wait.threshold; - trace_write_gather(cdma, g->bo, g->offset, g->words); + submit_wait(cdma, cmd->wait.id, threshold, cmd->wait.next_class); + } else { + struct host1x_job_gather *g = &cmd->gather; + + dma_addr_t addr = g->base + g->offset; + u32 op2, op3; + + op2 = lower_32_bits(addr); + op3 = upper_32_bits(addr); - if (op3 != 0) { + trace_write_gather(cdma, g->bo, g->offset, g->words); + + if (op3 != 0) { #if HOST1X_HW >= 6 - u32 op1 = host1x_opcode_gather_wide(g->words); - u32 op4 = HOST1X_OPCODE_NOP; + u32 op1 = host1x_opcode_gather_wide(g->words); + u32 op4 = HOST1X_OPCODE_NOP; - host1x_cdma_push_wide(cdma, op1, op2, op3, op4); + host1x_cdma_push_wide(cdma, op1, op2, op3, op4); #else - dev_err(dev, "invalid gather for push buffer %pad\n", - &addr); - continue; + dev_err(dev, "invalid gather for push buffer %pad\n", + &addr); + continue; #endif - } else { - u32 op1 = host1x_opcode_gather(g->words); + } else { + u32 op1 = host1x_opcode_gather(g->words); - host1x_cdma_push(cdma, op1, op2); + host1x_cdma_push(cdma, op1, op2); + } } } } @@ -126,7 +171,7 @@ static int channel_submit(struct host1x_job *job) struct host1x *host = dev_get_drvdata(ch->dev->parent); trace_host1x_channel_submit(dev_name(ch->dev), - job->num_gathers, job->num_relocs, + job->num_cmds, job->num_relocs, job->syncpt->id, job->syncpt_incrs); /* before error checks, return current max */ @@ -181,7 +226,7 @@ static int channel_submit(struct host1x_job *job) host1x_opcode_setclass(job->class, 0, 0), HOST1X_OPCODE_NOP); - submit_gathers(job); + submit_gathers(job, syncval - user_syncpt_incrs); /* end CDMA submit & stash pinned hMems into sync queue */ host1x_cdma_end(&ch->cdma, job); diff --git a/drivers/gpu/host1x/hw/debug_hw.c b/drivers/gpu/host1x/hw/debug_hw.c index ceb48229d14b..35952fd5597e 100644 --- a/drivers/gpu/host1x/hw/debug_hw.c +++ b/drivers/gpu/host1x/hw/debug_hw.c @@ -208,10 +208,15 @@ static void show_channel_gathers(struct output *o, struct host1x_cdma *cdma) job->first_get, job->timeout, job->num_slots, job->num_unpins); - for (i = 0; i < job->num_gathers; i++) { - struct host1x_job_gather *g = &job->gathers[i]; + for (i = 0; i < job->num_cmds; i++) { + struct host1x_job_gather *g; u32 *mapped; + if (job->cmds[i].is_wait) + continue; + + g = &job->cmds[i].gather; + if (job->gather_copy_mapped) mapped = (u32 *)job->gather_copy_mapped; else diff --git a/drivers/gpu/host1x/hw/hw_host1x02_uclass.h b/drivers/gpu/host1x/hw/hw_host1x02_uclass.h index 4fc51f70496b..0a2ab8f1da6f 100644 --- a/drivers/gpu/host1x/hw/hw_host1x02_uclass.h +++ b/drivers/gpu/host1x/hw/hw_host1x02_uclass.h @@ -165,5 +165,17 @@ static inline u32 host1x_uclass_indoff_rwn_read_v(void) } #define HOST1X_UCLASS_INDOFF_INDROFFSET_F(v) \ host1x_uclass_indoff_indroffset_f(v) +static inline u32 host1x_uclass_load_syncpt_payload_32_r(void) +{ + return 0x4e; +} +#define HOST1X_UCLASS_LOAD_SYNCPT_PAYLOAD_32 \ + host1x_uclass_load_syncpt_payload_32_r() +static inline u32 host1x_uclass_wait_syncpt_32_r(void) +{ + return 0x50; +} +#define HOST1X_UCLASS_WAIT_SYNCPT_32 \ + host1x_uclass_wait_syncpt_32_r() #endif diff --git a/drivers/gpu/host1x/hw/hw_host1x04_uclass.h b/drivers/gpu/host1x/hw/hw_host1x04_uclass.h index 9e84a4adca9f..60c692b92955 100644 --- a/drivers/gpu/host1x/hw/hw_host1x04_uclass.h +++ b/drivers/gpu/host1x/hw/hw_host1x04_uclass.h @@ -165,5 +165,17 @@ static inline u32 host1x_uclass_indoff_rwn_read_v(void) } #define HOST1X_UCLASS_INDOFF_INDROFFSET_F(v) \ host1x_uclass_indoff_indroffset_f(v) +static inline u32 host1x_uclass_load_syncpt_payload_32_r(void) +{ + return 0x4e; +} +#define HOST1X_UCLASS_LOAD_SYNCPT_PAYLOAD_32 \ + host1x_uclass_load_syncpt_payload_32_r() +static inline u32 host1x_uclass_wait_syncpt_32_r(void) +{ + return 0x50; +} +#define HOST1X_UCLASS_WAIT_SYNCPT_32 \ + host1x_uclass_wait_syncpt_32_r() #endif diff --git a/drivers/gpu/host1x/hw/hw_host1x05_uclass.h b/drivers/gpu/host1x/hw/hw_host1x05_uclass.h index aee5a4e32877..2fcc9a2ad3ef 100644 --- a/drivers/gpu/host1x/hw/hw_host1x05_uclass.h +++ b/drivers/gpu/host1x/hw/hw_host1x05_uclass.h @@ -165,5 +165,17 @@ static inline u32 host1x_uclass_indoff_rwn_read_v(void) } #define HOST1X_UCLASS_INDOFF_INDROFFSET_F(v) \ host1x_uclass_indoff_indroffset_f(v) +static inline u32 host1x_uclass_load_syncpt_payload_32_r(void) +{ + return 0x4e; +} +#define HOST1X_UCLASS_LOAD_SYNCPT_PAYLOAD_32 \ + host1x_uclass_load_syncpt_payload_32_r() +static inline u32 host1x_uclass_wait_syncpt_32_r(void) +{ + return 0x50; +} +#define HOST1X_UCLASS_WAIT_SYNCPT_32 \ + host1x_uclass_wait_syncpt_32_r() #endif diff --git a/drivers/gpu/host1x/hw/hw_host1x06_uclass.h b/drivers/gpu/host1x/hw/hw_host1x06_uclass.h index c4bacdb7155f..5f831438d19b 100644 --- a/drivers/gpu/host1x/hw/hw_host1x06_uclass.h +++ b/drivers/gpu/host1x/hw/hw_host1x06_uclass.h @@ -165,5 +165,17 @@ static inline u32 host1x_uclass_indoff_rwn_read_v(void) } #define HOST1X_UCLASS_INDOFF_INDROFFSET_F(v) \ host1x_uclass_indoff_indroffset_f(v) +static inline u32 host1x_uclass_load_syncpt_payload_32_r(void) +{ + return 0x4e; +} +#define HOST1X_UCLASS_LOAD_SYNCPT_PAYLOAD_32 \ + host1x_uclass_load_syncpt_payload_32_r() +static inline u32 host1x_uclass_wait_syncpt_32_r(void) +{ + return 0x50; +} +#define HOST1X_UCLASS_WAIT_SYNCPT_32 \ + host1x_uclass_wait_syncpt_32_r() #endif diff --git a/drivers/gpu/host1x/hw/hw_host1x07_uclass.h b/drivers/gpu/host1x/hw/hw_host1x07_uclass.h index c74070f3f203..8cd2ef087d5d 100644 --- a/drivers/gpu/host1x/hw/hw_host1x07_uclass.h +++ b/drivers/gpu/host1x/hw/hw_host1x07_uclass.h @@ -165,5 +165,17 @@ static inline u32 host1x_uclass_indoff_rwn_read_v(void) } #define HOST1X_UCLASS_INDOFF_INDROFFSET_F(v) \ host1x_uclass_indoff_indroffset_f(v) +static inline u32 host1x_uclass_load_syncpt_payload_32_r(void) +{ + return 0x4e; +} +#define HOST1X_UCLASS_LOAD_SYNCPT_PAYLOAD_32 \ + host1x_uclass_load_syncpt_payload_32_r() +static inline u32 host1x_uclass_wait_syncpt_32_r(void) +{ + return 0x50; +} +#define HOST1X_UCLASS_WAIT_SYNCPT_32 \ + host1x_uclass_wait_syncpt_32_r() #endif diff --git a/drivers/gpu/host1x/job.c b/drivers/gpu/host1x/job.c index 09097e19c0d0..32619b73a2fc 100644 --- a/drivers/gpu/host1x/job.c +++ b/drivers/gpu/host1x/job.c @@ -38,7 +38,7 @@ struct host1x_job *host1x_job_alloc(struct host1x_channel *ch, total = sizeof(struct host1x_job) + (u64)num_relocs * sizeof(struct host1x_reloc) + (u64)num_unpins * sizeof(struct host1x_job_unpin_data) + - (u64)num_cmdbufs * sizeof(struct host1x_job_gather) + + (u64)num_cmdbufs * sizeof(struct host1x_job_cmd) + (u64)num_unpins * sizeof(dma_addr_t) + (u64)num_unpins * sizeof(u32 *); if (total > ULONG_MAX) @@ -57,8 +57,8 @@ struct host1x_job *host1x_job_alloc(struct host1x_channel *ch, mem += num_relocs * sizeof(struct host1x_reloc); job->unpins = num_unpins ? mem : NULL; mem += num_unpins * sizeof(struct host1x_job_unpin_data); - job->gathers = num_cmdbufs ? mem : NULL; - mem += num_cmdbufs * sizeof(struct host1x_job_gather); + job->cmds = num_cmdbufs ? mem : NULL; + mem += num_cmdbufs * sizeof(struct host1x_job_cmd); job->addr_phys = num_unpins ? mem : NULL; job->reloc_addr_phys = job->addr_phys; @@ -101,22 +101,38 @@ EXPORT_SYMBOL(host1x_job_put); void host1x_job_add_gather(struct host1x_job *job, struct host1x_bo *bo, unsigned int words, unsigned int offset) { - struct host1x_job_gather *gather = &job->gathers[job->num_gathers]; + struct host1x_job_gather *gather = &job->cmds[job->num_cmds].gather; gather->words = words; gather->bo = bo; gather->offset = offset; - job->num_gathers++; + job->num_cmds++; } EXPORT_SYMBOL(host1x_job_add_gather); +void host1x_job_add_wait(struct host1x_job *job, u32 id, u32 thresh, + bool relative, u32 next_class) +{ + struct host1x_job_cmd *cmd = &job->cmds[job->num_cmds]; + + cmd->is_wait = true; + cmd->wait.id = id; + cmd->wait.threshold = thresh; + cmd->wait.next_class = next_class; + cmd->wait.relative = relative; + + job->num_cmds++; +} +EXPORT_SYMBOL(host1x_job_add_wait); + static unsigned int pin_job(struct host1x *host, struct host1x_job *job) { struct host1x_client *client = job->client; struct device *dev = client->dev; struct host1x_job_gather *g; struct iommu_domain *domain; + struct sg_table *sgt; unsigned int i; int err; @@ -126,7 +142,6 @@ static unsigned int pin_job(struct host1x *host, struct host1x_job *job) for (i = 0; i < job->num_relocs; i++) { struct host1x_reloc *reloc = &job->relocs[i]; dma_addr_t phys_addr, *phys; - struct sg_table *sgt; reloc->target.bo = host1x_bo_get(reloc->target.bo); if (!reloc->target.bo) { @@ -202,17 +217,20 @@ static unsigned int pin_job(struct host1x *host, struct host1x_job *job) if (IS_ENABLED(CONFIG_TEGRA_HOST1X_FIREWALL)) return 0; - for (i = 0; i < job->num_gathers; i++) { + for (i = 0; i < job->num_cmds; i++) { size_t gather_size = 0; struct scatterlist *sg; - struct sg_table *sgt; dma_addr_t phys_addr; unsigned long shift; struct iova *alloc; dma_addr_t *phys; unsigned int j; - g = &job->gathers[i]; + if (job->cmds[i].is_wait) + continue; + + g = &job->cmds[i].gather; + g->bo = host1x_bo_get(g->bo); if (!g->bo) { err = -EINVAL; @@ -545,8 +563,13 @@ static inline int copy_gathers(struct device *host, struct host1x_job *job, fw.num_relocs = job->num_relocs; fw.class = job->class; - for (i = 0; i < job->num_gathers; i++) { - struct host1x_job_gather *g = &job->gathers[i]; + for (i = 0; i < job->num_cmds; i++) { + struct host1x_job_gather *g; + + if (job->cmds[i].is_wait) + continue; + + g = &job->cmds[i].gather; size += g->words * sizeof(u32); } @@ -568,10 +591,14 @@ static inline int copy_gathers(struct device *host, struct host1x_job *job, job->gather_copy_size = size; - for (i = 0; i < job->num_gathers; i++) { - struct host1x_job_gather *g = &job->gathers[i]; + for (i = 0; i < job->num_cmds; i++) { + struct host1x_job_gather *g; void *gather; + if (job->cmds[i].is_wait) + continue; + g = &job->cmds[i].gather; + /* Copy the gather */ gather = host1x_bo_mmap(g->bo); memcpy(job->gather_copy_mapped + offset, gather + g->offset, @@ -614,8 +641,12 @@ int host1x_job_pin(struct host1x_job *job, struct device *dev) } /* patch gathers */ - for (i = 0; i < job->num_gathers; i++) { - struct host1x_job_gather *g = &job->gathers[i]; + for (i = 0; i < job->num_cmds; i++) { + struct host1x_job_gather *g; + + if (job->cmds[i].is_wait) + continue; + g = &job->cmds[i].gather; /* process each gather mem only once */ if (g->handled) @@ -625,10 +656,11 @@ int host1x_job_pin(struct host1x_job *job, struct device *dev) if (!IS_ENABLED(CONFIG_TEGRA_HOST1X_FIREWALL)) g->base = job->gather_addr_phys[i]; - for (j = i + 1; j < job->num_gathers; j++) { - if (job->gathers[j].bo == g->bo) { - job->gathers[j].handled = true; - job->gathers[j].base = g->base; + for (j = i + 1; j < job->num_cmds; j++) { + if (!job->cmds[j].is_wait && + job->cmds[j].gather.bo == g->bo) { + job->cmds[j].gather.handled = true; + job->cmds[j].gather.base = g->base; } } diff --git a/drivers/gpu/host1x/job.h b/drivers/gpu/host1x/job.h index 94bc2e4ae241..b4428c5495c9 100644 --- a/drivers/gpu/host1x/job.h +++ b/drivers/gpu/host1x/job.h @@ -18,6 +18,22 @@ struct host1x_job_gather { bool handled; }; +struct host1x_job_wait { + u32 id; + u32 threshold; + u32 next_class; + bool relative; +}; + +struct host1x_job_cmd { + bool is_wait; + + union { + struct host1x_job_gather gather; + struct host1x_job_wait wait; + }; +}; + struct host1x_job_unpin_data { struct host1x_bo *bo; struct sg_table *sgt; diff --git a/include/linux/host1x.h b/include/linux/host1x.h index 7eecb0aa2519..9b6784708f2e 100644 --- a/include/linux/host1x.h +++ b/include/linux/host1x.h @@ -219,8 +219,8 @@ struct host1x_job { struct host1x_client *client; /* Gathers and their memory */ - struct host1x_job_gather *gathers; - unsigned int num_gathers; + struct host1x_job_cmd *cmds; + unsigned int num_cmds; /* Array of handles to be pinned & unpinned */ struct host1x_reloc *relocs; @@ -279,6 +279,8 @@ struct host1x_job *host1x_job_alloc(struct host1x_channel *ch, u32 num_cmdbufs, u32 num_relocs); void host1x_job_add_gather(struct host1x_job *job, struct host1x_bo *bo, unsigned int words, unsigned int offset); +void host1x_job_add_wait(struct host1x_job *job, u32 id, u32 thresh, + bool relative, u32 next_class); struct host1x_job *host1x_job_get(struct host1x_job *job); void host1x_job_put(struct host1x_job *job); int host1x_job_pin(struct host1x_job *job, struct device *dev); From patchwork Fri Jul 9 19:31:37 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503351 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=WVFtFfqB; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3Bd3Trpz9sRN for ; Sat, 10 Jul 2021 05:29:29 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229971AbhGITcM (ORCPT ); Fri, 9 Jul 2021 15:32:12 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43770 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229506AbhGITcM (ORCPT ); Fri, 9 Jul 2021 15:32:12 -0400 Received: from mail-wm1-x32c.google.com (mail-wm1-x32c.google.com [IPv6:2a00:1450:4864:20::32c]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id A874CC0613DD for ; Fri, 9 Jul 2021 12:29:28 -0700 (PDT) Received: by mail-wm1-x32c.google.com with SMTP id u8-20020a7bcb080000b02901e44e9caa2aso6952054wmj.4 for ; Fri, 09 Jul 2021 12:29:28 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=35f9uUm3ArgMp1cLHX0Pf0HqrWQHxue9iozfkzujAeI=; b=WVFtFfqBM7xmsOGH6rhkPw2CDdjE9M7kwkiAhdNx+cmgnqqDwTlnqRwjzW7ud8zjyG wD9lih2wk3yQmwjFMQ/SR1+0AJ8Ci15nrelRD6hiVpl8dziOIU6ziJ9Z/uAGu1EIO6wF 3v/G8neZ7kvFHuRhF2jIrFQlfjypNNXo9zK91m0DpzkB6Ea0P7JJvfTzHiW07kWKjUd4 UOLlZPBJMr2qwLXF8Bj2CuoZOPXO1BKFws+3NGTUAxpSPS89mUvhYsNcIlSC43QvMnUB bqG0By90z9AHXy7sAbpZYb95EsV1Fb2ifbwDlFVbNNBtt/6nMh+rA9OvE1604YlFrGTY l5Bg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=35f9uUm3ArgMp1cLHX0Pf0HqrWQHxue9iozfkzujAeI=; b=CK4V/MJSGLhRADx5z6ph8tM0NP/VGetzSWCBTt+m11ZPWPj0MkDXpnQmcm8UUkm3pE kxqBB3NIbcF9vPC/tkb1eBmtwq3M6ObBtKycXYYbFsCaTEOIgO/ojc9S0bQGHVz4fbTu 9q7+CRG7olibT1iTSE9c9NfenSljbZ00/vomcElQkKJw5rozZtkRhX09ItZHKJNqWcqq iAHYNipnQi1LGL2zqVzAHvZyTIpZ/12OfxY/LOLLBbV9Gk3OlWwHpSEjMLfx25Kq8v4B OXz9ahEpHTfTEZSt20ZGkVMOm6aLYDXLlo86unfSMXpBQtIClYCyLV9AZF/IyBB1qjAZ y6aw== X-Gm-Message-State: AOAM531Q9l5XGoWiwjctrjbX+Q0501Cugb+afV7CzGFf42KUL1rlzypl wGIKUJU65LrTz+Z2lSi0BA8= X-Google-Smtp-Source: ABdhPJybQBin29CwVDlAxvB2J6f2tMQwe/ss+S6TkbXMpAA49yLabyuNbfkYReGhISaKkmygI2pfQw== X-Received: by 2002:a7b:ca50:: with SMTP id m16mr535392wml.140.1625858967292; Fri, 09 Jul 2021 12:29:27 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id p8sm5681757wre.76.2021.07.09.12.29.26 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:26 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org Subject: [PATCH v8 05/14] drm/tegra: Extract tegra_gem_lookup Date: Fri, 9 Jul 2021 21:31:37 +0200 Message-Id: <20210709193146.2859516-6-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen The static function host1x_bo_lookup in drm.c is also useful elsewhere. Extract it as tegra_gem_lookup in gem.c. Signed-off-by: Mikko Perttunen Signed-off-by: Thierry Reding --- drivers/gpu/drm/tegra/drm.c | 20 +++----------------- drivers/gpu/drm/tegra/gem.c | 13 +++++++++++++ drivers/gpu/drm/tegra/gem.h | 2 ++ 3 files changed, 18 insertions(+), 17 deletions(-) diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 739250acd498..87954e69ab6c 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -107,20 +107,6 @@ static void tegra_drm_context_free(struct tegra_drm_context *context) kfree(context); } -static struct host1x_bo * -host1x_bo_lookup(struct drm_file *file, u32 handle) -{ - struct drm_gem_object *gem; - struct tegra_bo *bo; - - gem = drm_gem_object_lookup(file, handle); - if (!gem) - return NULL; - - bo = to_tegra_bo(gem); - return &bo->base; -} - static int host1x_reloc_copy_from_user(struct host1x_reloc *dest, struct drm_tegra_reloc __user *src, struct drm_device *drm, @@ -151,11 +137,11 @@ static int host1x_reloc_copy_from_user(struct host1x_reloc *dest, dest->flags = HOST1X_RELOC_READ | HOST1X_RELOC_WRITE; - dest->cmdbuf.bo = host1x_bo_lookup(file, cmdbuf); + dest->cmdbuf.bo = tegra_gem_lookup(file, cmdbuf); if (!dest->cmdbuf.bo) return -ENOENT; - dest->target.bo = host1x_bo_lookup(file, target); + dest->target.bo = tegra_gem_lookup(file, target); if (!dest->target.bo) return -ENOENT; @@ -238,7 +224,7 @@ int tegra_drm_submit(struct tegra_drm_context *context, goto fail; } - bo = host1x_bo_lookup(file, cmdbuf.handle); + bo = tegra_gem_lookup(file, cmdbuf.handle); if (!bo) { err = -ENOENT; goto fail; diff --git a/drivers/gpu/drm/tegra/gem.c b/drivers/gpu/drm/tegra/gem.c index 26af8daa9a16..6ec598f5d5b3 100644 --- a/drivers/gpu/drm/tegra/gem.c +++ b/drivers/gpu/drm/tegra/gem.c @@ -707,3 +707,16 @@ struct drm_gem_object *tegra_gem_prime_import(struct drm_device *drm, return &bo->gem; } + +struct host1x_bo *tegra_gem_lookup(struct drm_file *file, u32 handle) +{ + struct drm_gem_object *gem; + struct tegra_bo *bo; + + gem = drm_gem_object_lookup(file, handle); + if (!gem) + return NULL; + + bo = to_tegra_bo(gem); + return &bo->base; +} diff --git a/drivers/gpu/drm/tegra/gem.h b/drivers/gpu/drm/tegra/gem.h index c15fd99d6cb2..cb5146a67668 100644 --- a/drivers/gpu/drm/tegra/gem.h +++ b/drivers/gpu/drm/tegra/gem.h @@ -80,4 +80,6 @@ struct dma_buf *tegra_gem_prime_export(struct drm_gem_object *gem, struct drm_gem_object *tegra_gem_prime_import(struct drm_device *drm, struct dma_buf *buf); +struct host1x_bo *tegra_gem_lookup(struct drm_file *file, u32 handle); + #endif From patchwork Fri Jul 9 19:31:38 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503352 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=pqelCgeX; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3Bh210cz9sPf for ; Sat, 10 Jul 2021 05:29:32 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230155AbhGITcP (ORCPT ); Fri, 9 Jul 2021 15:32:15 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43782 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229506AbhGITcO (ORCPT ); Fri, 9 Jul 2021 15:32:14 -0400 Received: from mail-wr1-x42e.google.com (mail-wr1-x42e.google.com [IPv6:2a00:1450:4864:20::42e]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 99E97C0613DD for ; Fri, 9 Jul 2021 12:29:30 -0700 (PDT) Received: by mail-wr1-x42e.google.com with SMTP id i94so13552904wri.4 for ; Fri, 09 Jul 2021 12:29:30 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=t/PPBeundHHV5YTZRwwjwy8TkUTmGmzjIPGpViip3II=; b=pqelCgeXolsAnxuHmLV3Ct1tvYPFaB4LJOy08eu2LIO50Rlo1R7XtQ/RAfRB5qzEhA NyVr36XsiX4JEreMf6Ud8BUSWPnKUb4JMQnoZnuPNV1RocIEdMTlivkzRBoyoFILjsji t36FTGeghKAiKd5AfJsEnRiU/GuhdJjH0X3LbGzpWTCu/79nAPoq2zC0jIMXQDkqLWhV QFajYZ40adUkD0KyfKrOwADHd0+KZFbf9P2pFArgfK12oDjR9Ys1yFx5HiT6YGcVQN3G q+OrjhAQfK+MvLfDMflebtgocSQTrvsEtF8dzDToZ2CMAs8hKDbBB4wA3npLco/Eji9E Z3Dw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=t/PPBeundHHV5YTZRwwjwy8TkUTmGmzjIPGpViip3II=; b=aFTWnorKWW9U1TobGhmA0Q6o5HTawcjYSak15jSxaZp3j/1R5tH8tgAaESDVa9i4XI 2LXz65G5AJpZFYrCe6JaXpfskM4uulS4U5uoKv2wzKeXXfM6OUt1govUfHEf1Kn0YW4v o3Be/JTt/JPwazolnfikQ+EN913XOOtl66m0yBn6iYQgAbgoQcsox316/lnCW1wVTeCL QSlFReyae6tqY6P0WYg5oPy+mk92+TmsvcJqD3U7RESDPgf0LHnrl/PIKry1AQROImGk JxFFO0gQdTRaK90lmISH1BrSgTcJKH6xKh9r/kpOtS0TN4FAqlHmu5o7a9Xh0n9GQsXq gK0Q== X-Gm-Message-State: AOAM533imeeAKPwwV5YPiHn33qocVLQQXEH+7NXF09FY3CThrDsjyi7g Ly+v+qsPP1MkOC88Pz+2bGk= X-Google-Smtp-Source: ABdhPJxGLU+b4+22T+GiDmeH6f5tdlfWvnNAU6bBt9+dvM17X93TRCmjTMT9zcq1yruAWPjocE86fA== X-Received: by 2002:a5d:634e:: with SMTP id b14mr42593504wrw.96.1625858969210; Fri, 09 Jul 2021 12:29:29 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id n5sm5988765wri.31.2021.07.09.12.29.28 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:28 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org Subject: [PATCH v8 06/14] drm/tegra: Add new UAPI to header Date: Fri, 9 Jul 2021 21:31:38 +0200 Message-Id: <20210709193146.2859516-7-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen Update the tegra_drm.h UAPI header, adding the new proposed UAPI. The old staging UAPI is left in for now, with minor modification to avoid name collisions. Signed-off-by: Mikko Perttunen Signed-off-by: Thierry Reding --- Changes in v8: - renamed hardware_flags to capabilities (flags sounds like they can be user-specified) - renamed hardware_version to just version - consistently use "context" to refer to channels contexts - allow channel mappings to be read-only or write-only (or neither) - consistently use "mapping" (instead of "mapping_id") - rename BLOCKLINEAR relocation flag to SECTOR_LAYOUT for a more accurate description of what it does - cleanup submission syncpoint structure: - drop _incr suffix for conciseness - rename fence_value to just value - rename num_incrs to increments - rename syncpoint threshold to value for consistency - add padding to make all structures 64-bit aligned include/uapi/drm/tegra_drm.h | 425 +++++++++++++++++++++++++++++++++-- 1 file changed, 402 insertions(+), 23 deletions(-) diff --git a/include/uapi/drm/tegra_drm.h b/include/uapi/drm/tegra_drm.h index c4df3c3668b3..94cfc306d50a 100644 --- a/include/uapi/drm/tegra_drm.h +++ b/include/uapi/drm/tegra_drm.h @@ -1,24 +1,5 @@ -/* - * Copyright (c) 2012-2013, NVIDIA CORPORATION. All rights reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - */ +/* SPDX-License-Identifier: MIT */ +/* Copyright (c) 2012-2020 NVIDIA Corporation */ #ifndef _UAPI_TEGRA_DRM_H_ #define _UAPI_TEGRA_DRM_H_ @@ -29,6 +10,8 @@ extern "C" { #endif +/* Tegra DRM legacy UAPI. Only enabled with STAGING */ + #define DRM_TEGRA_GEM_CREATE_TILED (1 << 0) #define DRM_TEGRA_GEM_CREATE_BOTTOM_UP (1 << 1) @@ -649,8 +632,8 @@ struct drm_tegra_gem_get_flags { #define DRM_TEGRA_SYNCPT_READ 0x02 #define DRM_TEGRA_SYNCPT_INCR 0x03 #define DRM_TEGRA_SYNCPT_WAIT 0x04 -#define DRM_TEGRA_OPEN_CHANNEL 0x05 -#define DRM_TEGRA_CLOSE_CHANNEL 0x06 +#define DRM_TEGRA_OPEN_CHANNEL 0x05 +#define DRM_TEGRA_CLOSE_CHANNEL 0x06 #define DRM_TEGRA_GET_SYNCPT 0x07 #define DRM_TEGRA_SUBMIT 0x08 #define DRM_TEGRA_GET_SYNCPT_BASE 0x09 @@ -674,6 +657,402 @@ struct drm_tegra_gem_get_flags { #define DRM_IOCTL_TEGRA_GEM_SET_FLAGS DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_GEM_SET_FLAGS, struct drm_tegra_gem_set_flags) #define DRM_IOCTL_TEGRA_GEM_GET_FLAGS DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_GEM_GET_FLAGS, struct drm_tegra_gem_get_flags) +/* New Tegra DRM UAPI */ + +/* + * Reported by the driver in the `capabilities` field. + * + * DRM_TEGRA_CHANNEL_CAP_CACHE_COHERENT: If set, the engine is cache coherent + * with regard to the system memory. + */ +#define DRM_TEGRA_CHANNEL_CAP_CACHE_COHERENT (1 << 0) + +struct drm_tegra_channel_open { + /** + * @host1x_class: [in] + * + * Host1x class of the engine that will be programmed using this + * channel. + */ + __u32 host1x_class; + + /** + * @flags: [in] + * + * Flags. + */ + __u32 flags; + + /** + * @context: [out] + * + * Opaque identifier corresponding to the opened channel. + */ + __u32 context; + + /** + * @version: [out] + * + * Version of the engine hardware. This can be used by userspace + * to determine how the engine needs to be programmed. + */ + __u32 version; + + /** + * @capabilities: [out] + * + * Flags describing the hardware capabilities. + */ + __u32 capabilities; + __u32 padding; +}; + +struct drm_tegra_channel_close { + /** + * @context: [in] + * + * Identifier of the channel to close. + */ + __u32 context; + __u32 padding; +}; + +/* + * Mapping flags that can be used to influence how the mapping is created. + * + * DRM_TEGRA_CHANNEL_MAP_READ: create mapping that allows HW read access + * DRM_TEGRA_CHANNEL_MAP_WRITE: create mapping that allows HW write access + */ +#define DRM_TEGRA_CHANNEL_MAP_READ (1 << 0) +#define DRM_TEGRA_CHANNEL_MAP_WRITE (1 << 1) +#define DRM_TEGRA_CHANNEL_MAP_READ_WRITE (DRM_TEGRA_CHANNEL_MAP_READ | \ + DRM_TEGRA_CHANNEL_MAP_WRITE) + +struct drm_tegra_channel_map { + /** + * @context: [in] + * + * Identifier of the channel to which make memory available for. + */ + __u32 context; + + /** + * @handle: [in] + * + * GEM handle of the memory to map. + */ + __u32 handle; + + /** + * @flags: [in] + * + * Flags. + */ + __u32 flags; + + /** + * @mapping: [out] + * + * Identifier corresponding to the mapping, to be used for + * relocations or unmapping later. + */ + __u32 mapping; +}; + +struct drm_tegra_channel_unmap { + /** + * @context: [in] + * + * Channel identifier of the channel to unmap memory from. + */ + __u32 context; + + /** + * @mapping: [in] + * + * Mapping identifier of the memory mapping to unmap. + */ + __u32 mapping; +}; + +/* Submission */ + +/** + * Specify that bit 39 of the patched-in address should be set to switch + * swizzling between Tegra and non-Tegra sector layout on systems that store + * surfaces in system memory in non-Tegra sector layout. + */ +#define DRM_TEGRA_SUBMIT_RELOC_SECTOR_LAYOUT (1 << 0) + +struct drm_tegra_submit_buf { + /** + * @mapping: [in] + * + * Identifier of the mapping to use in the submission. + */ + __u32 mapping; + + /** + * @flags: [in] + * + * Flags. + */ + __u32 flags; + + /** + * Information for relocation patching. + */ + struct { + /** + * @target_offset: [in] + * + * Offset from the start of the mapping of the data whose + * address is to be patched into the gather. + */ + __u64 target_offset; + + /** + * @gather_offset_words: [in] + * + * Offset in words from the start of the gather data to + * where the address should be patched into. + */ + __u32 gather_offset_words; + + /** + * @shift: [in] + * + * Number of bits the address should be shifted right before + * patching in. + */ + __u32 shift; + } reloc; +}; + +/** + * Execute `words` words of Host1x opcodes specified in the `gather_data_ptr` + * buffer. Each GATHER_UPTR command uses successive words from the buffer. + */ +#define DRM_TEGRA_SUBMIT_CMD_GATHER_UPTR 0 +/** + * Wait for a syncpoint to reach a value before continuing with further + * commands. + */ +#define DRM_TEGRA_SUBMIT_CMD_WAIT_SYNCPT 1 +/** + * Wait for a syncpoint to reach a value before continuing with further + * commands. The threshold is calculated relative to the start of the job. + */ +#define DRM_TEGRA_SUBMIT_CMD_WAIT_SYNCPT_RELATIVE 2 + +struct drm_tegra_submit_cmd_gather_uptr { + __u32 words; + __u32 reserved[3]; +}; + +struct drm_tegra_submit_cmd_wait_syncpt { + __u32 id; + __u32 value; + __u32 reserved[2]; +}; + +struct drm_tegra_submit_cmd { + /** + * @type: [in] + * + * Command type to execute. One of the DRM_TEGRA_SUBMIT_CMD* + * defines. + */ + __u32 type; + + /** + * @flags: [in] + * + * Flags. + */ + __u32 flags; + + union { + struct drm_tegra_submit_cmd_gather_uptr gather_uptr; + struct drm_tegra_submit_cmd_wait_syncpt wait_syncpt; + __u32 reserved[4]; + }; +}; + +struct drm_tegra_submit_syncpt { + /** + * @id: [in] + * + * ID of the syncpoint that the job will increment. + */ + __u32 id; + + /** + * @flags: [in] + * + * Flags. + */ + __u32 flags; + + /** + * @increments: [in] + * + * Number of times the job will increment this syncpoint. + */ + __u32 increments; + + /** + * @value: [out] + * + * Value the syncpoint will have once the job has completed all + * its specified syncpoint increments. + * + * Note that the kernel may increment the syncpoint before or after + * the job. These increments are not reflected in this field. + * + * If the job hangs or times out, not all of the increments may + * get executed. + */ + __u32 value; +}; + +struct drm_tegra_channel_submit { + /** + * @context: [in] + * + * Identifier of the channel to submit this job to. + */ + __u32 context; + + /** + * @num_bufs: [in] + * + * Number of elements in the `bufs_ptr` array. + */ + __u32 num_bufs; + + /** + * @num_cmds: [in] + * + * Number of elements in the `cmds_ptr` array. + */ + __u32 num_cmds; + + /** + * @gather_data_words: [in] + * + * Number of 32-bit words in the `gather_data_ptr` array. + */ + __u32 gather_data_words; + + /** + * @bufs_ptr: [in] + * + * Pointer to an array of drm_tegra_submit_buf structures. + */ + __u64 bufs_ptr; + + /** + * @cmds_ptr: [in] + * + * Pointer to an array of drm_tegra_submit_cmd structures. + */ + __u64 cmds_ptr; + + /** + * @gather_data_ptr: [in] + * + * Pointer to an array of Host1x opcodes to be used by GATHER_UPTR + * commands. + */ + __u64 gather_data_ptr; + + /** + * @syncobj_in: [in] + * + * Handle for DRM syncobj that will be waited before submission. + * Ignored if zero. + */ + __u32 syncobj_in; + + /** + * @syncobj_out: [in] + * + * Handle for DRM syncobj that will have its fence replaced with + * the job's completion fence. Ignored if zero. + */ + __u32 syncobj_out; + + /** + * @syncpt_incr: [in,out] + * + * Information about the syncpoint the job will increment. + */ + struct drm_tegra_submit_syncpt syncpt; +}; + +struct drm_tegra_syncpoint_allocate { + /** + * @id: [out] + * + * ID of allocated syncpoint. + */ + __u32 id; + __u32 padding; +}; + +struct drm_tegra_syncpoint_free { + /** + * @id: [in] + * + * ID of syncpoint to free. + */ + __u32 id; + __u32 padding; +}; + +struct drm_tegra_syncpoint_wait { + /** + * @timeout: [in] + * + * Absolute timestamp at which the wait will time out. + */ + __s64 timeout_ns; + + /** + * @id: [in] + * + * ID of syncpoint to wait on. + */ + __u32 id; + + /** + * @threshold: [in] + * + * Threshold to wait for. + */ + __u32 threshold; + + /** + * @value: [out] + * + * Value of the syncpoint upon wait completion. + */ + __u32 value; + + __u32 padding; +}; + +#define DRM_IOCTL_TEGRA_CHANNEL_OPEN DRM_IOWR(DRM_COMMAND_BASE + 0x10, struct drm_tegra_channel_open) +#define DRM_IOCTL_TEGRA_CHANNEL_CLOSE DRM_IOWR(DRM_COMMAND_BASE + 0x11, struct drm_tegra_channel_close) +#define DRM_IOCTL_TEGRA_CHANNEL_MAP DRM_IOWR(DRM_COMMAND_BASE + 0x12, struct drm_tegra_channel_map) +#define DRM_IOCTL_TEGRA_CHANNEL_UNMAP DRM_IOWR(DRM_COMMAND_BASE + 0x13, struct drm_tegra_channel_unmap) +#define DRM_IOCTL_TEGRA_CHANNEL_SUBMIT DRM_IOWR(DRM_COMMAND_BASE + 0x14, struct drm_tegra_channel_submit) + +#define DRM_IOCTL_TEGRA_SYNCPOINT_ALLOCATE DRM_IOWR(DRM_COMMAND_BASE + 0x20, struct drm_tegra_syncpoint_allocate) +#define DRM_IOCTL_TEGRA_SYNCPOINT_FREE DRM_IOWR(DRM_COMMAND_BASE + 0x21, struct drm_tegra_syncpoint_free) +#define DRM_IOCTL_TEGRA_SYNCPOINT_WAIT DRM_IOWR(DRM_COMMAND_BASE + 0x22, struct drm_tegra_syncpoint_wait) + #if defined(__cplusplus) } #endif From patchwork Fri Jul 9 19:31:39 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503354 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=gt8znewZ; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3Bm0BK6z9sRN for ; Sat, 10 Jul 2021 05:29:36 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230173AbhGITcT (ORCPT ); Fri, 9 Jul 2021 15:32:19 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43800 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230166AbhGITcS (ORCPT ); Fri, 9 Jul 2021 15:32:18 -0400 Received: from mail-wm1-x335.google.com (mail-wm1-x335.google.com [IPv6:2a00:1450:4864:20::335]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id A90A5C0613E5 for ; Fri, 9 Jul 2021 12:29:32 -0700 (PDT) Received: by mail-wm1-x335.google.com with SMTP id k16-20020a05600c1c90b02901f4ed0fcfe7so6950290wms.5 for ; Fri, 09 Jul 2021 12:29:32 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=fqNBjbH3rPHiYRwYwGOI/SzSmc1V4SyfxtAqmW62gNs=; b=gt8znewZZK6XKs/w5uBhJkXn5qEczq7wmCaJqFvE//zZfq7/PpR2j7JBETWRtGshtv zzXKZGutLrTgCDwiTYcMyimNt4UF1bg/yVMLb8g5nLC+B5OGek3TqCCKjagugR9dZny3 RP/0G6Ns3ExkBDWMamvWkg8cFjASRM/U7J7H20A0lXXH6Ms2fwgpqNzOO7Q2ribongXx exs3BwZwJcH0c9YQ9+E3WQri3INd+P6ruLDeHn8AI5+9U/Mzchh89VB5G0AYfVEw+0QB ZIp9uIpWfx/8xr6ZDI1LwR5BrnsgDNQxU2OSK4DaJQnnIi2hk7FKrPLVre65OQyJrWuV ayXg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=fqNBjbH3rPHiYRwYwGOI/SzSmc1V4SyfxtAqmW62gNs=; b=OoLI9Ddu/Na9/aonefarfwQtiM9PMpXONyK/hvfFhdIOkySxmt9pvKiWyS1Saf806u HrTCQjt4cc4NxCeAXH34x7WI2elApbYvz6FBPwfQXfp8BF7vVYBWjhVB5TZcO4S2f0HF 92WrbRnbNjd22bOHz+wNxlwH5sB55w9CTnr3DrBrmDftXNscU40gl9OWNUncjeUzHYjd 7o3wQp9gQQD6XSNpqNK7Xcgxg4+LL5MKZLOPsF3ICbl+kCP9ApXH+dCiZYo39fHk11Cg kxQpcy8BB38c4/aMl9qEW/H6GrzjrUC6oUHU7U9KoAx4Lt9bf040lZUXxjDU/H51DHKr U+xA== X-Gm-Message-State: AOAM531txV8AtXLY7FIr28NKCCpauLrS8fr2371rpH4USokCef7oIxtF 4uMw/91MOM+Zl2ZFfiT2zAE= X-Google-Smtp-Source: ABdhPJyyto22V4g+Vao8QFiXSt0LgkQ5bjzQPG4kbHF8Dw0WdX6zsdVNkGxmNxqReH81ilWmbfA1Tw== X-Received: by 2002:a7b:c30f:: with SMTP id k15mr548591wmj.128.1625858971280; Fri, 09 Jul 2021 12:29:31 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id f9sm6079953wrm.48.2021.07.09.12.29.30 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:30 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org Subject: [PATCH v8 07/14] drm/tegra: Boot VIC during runtime PM resume Date: Fri, 9 Jul 2021 21:31:39 +0200 Message-Id: <20210709193146.2859516-8-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen With the new UAPI implementation, engines are powered on and off when there are active jobs, and the core code handles channel allocation. To accommodate that, boot the engine as part of runtime PM instead of using the open_channel callback, which is not used by the new submit path. Signed-off-by: Mikko Perttunen Signed-off-by: Thierry Reding --- drivers/gpu/drm/tegra/vic.c | 112 ++++++++++++++++-------------------- 1 file changed, 51 insertions(+), 61 deletions(-) diff --git a/drivers/gpu/drm/tegra/vic.c b/drivers/gpu/drm/tegra/vic.c index c9d55a9a3180..c02010ff2b7f 100644 --- a/drivers/gpu/drm/tegra/vic.c +++ b/drivers/gpu/drm/tegra/vic.c @@ -29,7 +29,6 @@ struct vic_config { struct vic { struct falcon falcon; - bool booted; void __iomem *regs; struct tegra_drm_client client; @@ -52,48 +51,6 @@ static void vic_writel(struct vic *vic, u32 value, unsigned int offset) writel(value, vic->regs + offset); } -static int vic_runtime_resume(struct device *dev) -{ - struct vic *vic = dev_get_drvdata(dev); - int err; - - err = clk_prepare_enable(vic->clk); - if (err < 0) - return err; - - usleep_range(10, 20); - - err = reset_control_deassert(vic->rst); - if (err < 0) - goto disable; - - usleep_range(10, 20); - - return 0; - -disable: - clk_disable_unprepare(vic->clk); - return err; -} - -static int vic_runtime_suspend(struct device *dev) -{ - struct vic *vic = dev_get_drvdata(dev); - int err; - - err = reset_control_assert(vic->rst); - if (err < 0) - return err; - - usleep_range(2000, 4000); - - clk_disable_unprepare(vic->clk); - - vic->booted = false; - - return 0; -} - static int vic_boot(struct vic *vic) { #ifdef CONFIG_IOMMU_API @@ -103,9 +60,6 @@ static int vic_boot(struct vic *vic) void *hdr; int err = 0; - if (vic->booted) - return 0; - #ifdef CONFIG_IOMMU_API if (vic->config->supports_sid && spec) { u32 value; @@ -168,8 +122,6 @@ static int vic_boot(struct vic *vic) return err; } - vic->booted = true; - return 0; } @@ -323,35 +275,74 @@ static int vic_load_firmware(struct vic *vic) return err; } -static int vic_open_channel(struct tegra_drm_client *client, - struct tegra_drm_context *context) + +static int vic_runtime_resume(struct device *dev) { - struct vic *vic = to_vic(client); + struct vic *vic = dev_get_drvdata(dev); int err; - err = pm_runtime_resume_and_get(vic->dev); + err = clk_prepare_enable(vic->clk); if (err < 0) return err; + usleep_range(10, 20); + + err = reset_control_deassert(vic->rst); + if (err < 0) + goto disable; + + usleep_range(10, 20); + err = vic_load_firmware(vic); if (err < 0) - goto rpm_put; + goto assert; err = vic_boot(vic); if (err < 0) - goto rpm_put; + goto assert; + + return 0; + +assert: + reset_control_assert(vic->rst); +disable: + clk_disable_unprepare(vic->clk); + return err; +} + +static int vic_runtime_suspend(struct device *dev) +{ + struct vic *vic = dev_get_drvdata(dev); + int err; + + err = reset_control_assert(vic->rst); + if (err < 0) + return err; + + usleep_range(2000, 4000); + + clk_disable_unprepare(vic->clk); + + return 0; +} + +static int vic_open_channel(struct tegra_drm_client *client, + struct tegra_drm_context *context) +{ + struct vic *vic = to_vic(client); + int err; + + err = pm_runtime_resume_and_get(vic->dev); + if (err < 0) + return err; context->channel = host1x_channel_get(vic->channel); if (!context->channel) { - err = -ENOMEM; - goto rpm_put; + pm_runtime_put(vic->dev); + return -ENOMEM; } return 0; - -rpm_put: - pm_runtime_put(vic->dev); - return err; } static void vic_close_channel(struct tegra_drm_context *context) @@ -359,7 +350,6 @@ static void vic_close_channel(struct tegra_drm_context *context) struct vic *vic = to_vic(context->client); host1x_channel_put(context->channel); - pm_runtime_put(vic->dev); } From patchwork Fri Jul 9 19:31:40 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503353 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=OTDWKUWG; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3Bl50Xxz9sPf for ; Sat, 10 Jul 2021 05:29:35 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230168AbhGITcS (ORCPT ); Fri, 9 Jul 2021 15:32:18 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43802 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229506AbhGITcS (ORCPT ); Fri, 9 Jul 2021 15:32:18 -0400 Received: from mail-wr1-x430.google.com (mail-wr1-x430.google.com [IPv6:2a00:1450:4864:20::430]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id D7327C0613DD for ; Fri, 9 Jul 2021 12:29:34 -0700 (PDT) Received: by mail-wr1-x430.google.com with SMTP id k4so7272086wrc.8 for ; Fri, 09 Jul 2021 12:29:34 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=nO2VGF0GrX8eq1kIPmNF9NDL8unf6/wwgmYeLn0z60A=; b=OTDWKUWGMl6zGndK9eFgIokkIOVkQtjxt4yGNj4M4Od9i5+nNyrhzGyizKj5jeBTB2 8+V7ngN/unevaFcHVs73EPGa5RSfW+XZWlryVS/Kn6P4pQoUPLE0rMvSJ88d4ywqi12r zAcgWkgaK6mDWL36JqHrdoPvoZmqBJTRh19YGkk142Y90VFWB/P6P/ika13nbxNWsX68 LAviBBmvafT74ur/doA7CjqRCSTZdWJsNhmjJDOoM8uK1PxNts11+W3N7SFwb7NkVlxB CI+uFsUhH9ivqUsSmGNd9yYVklIsic3L0w8CzpFNVCJiXZy/t8AIkKoTQD4qILsJ4/Oh 9iWQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=nO2VGF0GrX8eq1kIPmNF9NDL8unf6/wwgmYeLn0z60A=; b=ME6+/SL7ITIkd9Y74NWCnhfHxyzu10fxc5o/Ka0p/xyn+3P562B2md9ysHKwcT+y3k NZjdzp1cQ5quN5QtvDrmOJJbgdwF037P+cgSiqMnlnfDL1MuuQUaTOLnNaG+ZOSbC8EA iyhCeEg/c+S4YDOwE7GLICcjrlpcuhxSIQdwPwjE28MfqTc6GylC8rgsWZ661+x5lZ0j Z8hyB3XntRbsc6enL8Fjg5aTNNIbJpBmU0pbNgaI2vs6H/4CbYgSdy1r+orXxrPfdPBE gMXJyzCkdlNIwqBGzC+tIMEaArAxJBa3maNP68sRGP2jPAT6MAKYEN2G+t+gF19HwoUf MLNA== X-Gm-Message-State: AOAM530OsnqcVpCuba/+CqFEnSo+AulIZofhvj9Yq4zJSJWQ/J5yK0na 9rLlfeX2kz1WqfBEPDHG6uQ= X-Google-Smtp-Source: ABdhPJwexKMid8Y7ZFKxgXv6gNWN+bjD3QttXFUzwM+raTEJLtzfVRBbUUVNPsA8qfmKvt/mmduqaQ== X-Received: by 2002:a05:6000:1867:: with SMTP id d7mr36571354wri.263.1625858973547; Fri, 09 Jul 2021 12:29:33 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id t6sm6144348wru.75.2021.07.09.12.29.32 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:32 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org Subject: [PATCH v8 08/14] drm/tegra: Allocate per-engine channel in core code Date: Fri, 9 Jul 2021 21:31:40 +0200 Message-Id: <20210709193146.2859516-9-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen To avoid code duplication, allocate the per-engine shared channel in the core code instead. This is the usual channel that all jobs are submitted to when MLOCKing is not in use. Once MLOCKs are implemented on Host1x side, we can also update this to avoid allocating a shared channel when MLOCKs are enabled. Signed-off-by: Mikko Perttunen Signed-off-by: Thierry Reding --- drivers/gpu/drm/tegra/drm.c | 11 +++++++++++ drivers/gpu/drm/tegra/drm.h | 2 ++ 2 files changed, 13 insertions(+) diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 87954e69ab6c..cddee6425461 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -870,6 +870,14 @@ static const struct drm_driver tegra_drm_driver = { int tegra_drm_register_client(struct tegra_drm *tegra, struct tegra_drm_client *client) { + /* + * When MLOCKs are implemented, change to allocate a shared channel + * only when MLOCKs are disabled. + */ + client->shared_channel = host1x_channel_request(&client->base); + if (!client->shared_channel) + return -EBUSY; + mutex_lock(&tegra->clients_lock); list_add_tail(&client->list, &tegra->clients); client->drm = tegra; @@ -886,6 +894,9 @@ int tegra_drm_unregister_client(struct tegra_drm *tegra, client->drm = NULL; mutex_unlock(&tegra->clients_lock); + if (client->shared_channel) + host1x_channel_put(client->shared_channel); + return 0; } diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h index 0cb868065348..536861017d24 100644 --- a/drivers/gpu/drm/tegra/drm.h +++ b/drivers/gpu/drm/tegra/drm.h @@ -91,7 +91,9 @@ struct tegra_drm_client { struct host1x_client base; struct list_head list; struct tegra_drm *drm; + struct host1x_channel *shared_channel; + /* Set by driver */ unsigned int version; const struct tegra_drm_client_ops *ops; }; From patchwork Fri Jul 9 19:31:41 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503355 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=EAW3cdmm; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3Bp54gkz9sPf for ; Sat, 10 Jul 2021 05:29:38 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230166AbhGITcW (ORCPT ); Fri, 9 Jul 2021 15:32:22 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43816 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229506AbhGITcV (ORCPT ); Fri, 9 Jul 2021 15:32:21 -0400 Received: from mail-wm1-x329.google.com (mail-wm1-x329.google.com [IPv6:2a00:1450:4864:20::329]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 0A16DC0613DD for ; Fri, 9 Jul 2021 12:29:37 -0700 (PDT) Received: by mail-wm1-x329.google.com with SMTP id n33so6988643wms.1 for ; Fri, 09 Jul 2021 12:29:36 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=oia4tcEzeIofOiob91SyDfZH5klf9jnj07ifF12omHU=; b=EAW3cdmmaJigYehNSd92YwenxPmUjVVc0NEJ8WAV12VMBPAVlpqJpo6MV7ZJVMqPQ4 k3cMaAlU90j/RnhbQ5JZnPqixErnsALeQrtf/HDR00CABRZzHZSVx5Zo1cYA7/OJhd3p M2Na3Y43pPouwTI4GOPd7FPrFN7zrlLJBe/eLDnMIWutK8+Hx84cXyY+i5IUGBBYMlwy VgubVk5HmwtE09mOU4zchYxsiK4o1cLL8IxrPy6rOWyZ4i9bZUaxrCxqWuCqVzU1BzsC 4+1CyqkcBvkeGCnkihLTmhGi28G5QgzZxzFnJ+nPTmv1IJm9eFbztSqWKwsFWq4ugmx1 n0/w== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=oia4tcEzeIofOiob91SyDfZH5klf9jnj07ifF12omHU=; b=elziHgi2vkg4oUzclgxMTXbUUQa5ucy6gKXtHS1pqmheX+z/6prZGrqXxU5lNA28fJ AFAFNXyIsrhlYA7wWau3lCUSREWkmjJLNt/sDEziQNGLu/eqKh7EhF+Gy2uUFF3bn5ro gV14EQ2f804t78niAVbQpQWzRlB6oXBw7TaN9sSR/t9CfeYejA5XwWIuykWUh3BkXyOu RU2lyfXyjPnTyHhUbMcL++RS/CReKWZFDJwDWQBcb22Xt3oAZe054xlCbY3E3RL4rFvv sdj2qthmCT49ykht6tYRXA/zlOVBzycO3ToIeh0I8nfWpw5NymHw0d6JJm9CTIuMx1oM f6Jg== X-Gm-Message-State: AOAM531OOqOJ9VN0jk+FEls7xCgLwr5EkLz9wQfxYEk4FK8ENoUbmFyN VI1djU2Y7hcKw3u3ACS+lCg= X-Google-Smtp-Source: ABdhPJyK7hm/SDPsm39+6JZIwtvobK2HqAZBV7EmYOaslcQFUvbaqCXtRilI1T6peOEXhcMIitKKxw== X-Received: by 2002:a1c:b403:: with SMTP id d3mr41258282wmf.79.1625858975541; Fri, 09 Jul 2021 12:29:35 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id p5sm6346396wrd.25.2021.07.09.12.29.34 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:34 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org Subject: [PATCH v8 09/14] drm/tegra: Implement new UAPI Date: Fri, 9 Jul 2021 21:31:41 +0200 Message-Id: <20210709193146.2859516-10-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen Implement the non-submission parts of the new UAPI, including channel management and memory mapping. The UAPI is under the CONFIG_DRM_TEGRA_STAGING config flag for now. Signed-off-by: Mikko Perttunen Signed-off-by: Thierry Reding --- Changes in v8: - rebase on top of latest UABI changes drivers/gpu/drm/tegra/Makefile | 1 + drivers/gpu/drm/tegra/drm.c | 37 ++-- drivers/gpu/drm/tegra/drm.h | 10 ++ drivers/gpu/drm/tegra/uapi.c | 313 +++++++++++++++++++++++++++++++++ drivers/gpu/drm/tegra/uapi.h | 51 ++++++ 5 files changed, 396 insertions(+), 16 deletions(-) create mode 100644 drivers/gpu/drm/tegra/uapi.c create mode 100644 drivers/gpu/drm/tegra/uapi.h diff --git a/drivers/gpu/drm/tegra/Makefile b/drivers/gpu/drm/tegra/Makefile index d6cf202414f0..783475ffd943 100644 --- a/drivers/gpu/drm/tegra/Makefile +++ b/drivers/gpu/drm/tegra/Makefile @@ -3,6 +3,7 @@ ccflags-$(CONFIG_DRM_TEGRA_DEBUG) += -DDEBUG tegra-drm-y := \ drm.o \ + uapi.o \ gem.o \ fb.o \ dp.o \ diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index cddee6425461..6ee08e49ec57 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -21,6 +21,7 @@ #include #include +#include "uapi.h" #include "drm.h" #include "gem.h" @@ -34,11 +35,6 @@ #define CARVEOUT_SZ SZ_64M #define CDMA_GATHER_FETCHES_MAX_NB 16383 -struct tegra_drm_file { - struct idr contexts; - struct mutex lock; -}; - static int tegra_atomic_check(struct drm_device *drm, struct drm_atomic_state *state) { @@ -94,7 +90,8 @@ static int tegra_drm_open(struct drm_device *drm, struct drm_file *filp) if (!fpriv) return -ENOMEM; - idr_init_base(&fpriv->contexts, 1); + idr_init_base(&fpriv->legacy_contexts, 1); + xa_init_flags(&fpriv->contexts, XA_FLAGS_ALLOC1); mutex_init(&fpriv->lock); filp->driver_priv = fpriv; @@ -419,7 +416,7 @@ static int tegra_client_open(struct tegra_drm_file *fpriv, if (err < 0) return err; - err = idr_alloc(&fpriv->contexts, context, 1, 0, GFP_KERNEL); + err = idr_alloc(&fpriv->legacy_contexts, context, 1, 0, GFP_KERNEL); if (err < 0) { client->ops->close_channel(context); return err; @@ -474,13 +471,13 @@ static int tegra_close_channel(struct drm_device *drm, void *data, mutex_lock(&fpriv->lock); - context = idr_find(&fpriv->contexts, args->context); + context = idr_find(&fpriv->legacy_contexts, args->context); if (!context) { err = -EINVAL; goto unlock; } - idr_remove(&fpriv->contexts, context->id); + idr_remove(&fpriv->legacy_contexts, context->id); tegra_drm_context_free(context); unlock: @@ -499,7 +496,7 @@ static int tegra_get_syncpt(struct drm_device *drm, void *data, mutex_lock(&fpriv->lock); - context = idr_find(&fpriv->contexts, args->context); + context = idr_find(&fpriv->legacy_contexts, args->context); if (!context) { err = -ENODEV; goto unlock; @@ -528,7 +525,7 @@ static int tegra_submit(struct drm_device *drm, void *data, mutex_lock(&fpriv->lock); - context = idr_find(&fpriv->contexts, args->context); + context = idr_find(&fpriv->legacy_contexts, args->context); if (!context) { err = -ENODEV; goto unlock; @@ -553,7 +550,7 @@ static int tegra_get_syncpt_base(struct drm_device *drm, void *data, mutex_lock(&fpriv->lock); - context = idr_find(&fpriv->contexts, args->context); + context = idr_find(&fpriv->legacy_contexts, args->context); if (!context) { err = -ENODEV; goto unlock; @@ -722,10 +719,17 @@ static int tegra_gem_get_flags(struct drm_device *drm, void *data, static const struct drm_ioctl_desc tegra_drm_ioctls[] = { #ifdef CONFIG_DRM_TEGRA_STAGING - DRM_IOCTL_DEF_DRV(TEGRA_GEM_CREATE, tegra_gem_create, + DRM_IOCTL_DEF_DRV(TEGRA_CHANNEL_OPEN, tegra_drm_ioctl_channel_open, + DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(TEGRA_CHANNEL_CLOSE, tegra_drm_ioctl_channel_close, DRM_RENDER_ALLOW), - DRM_IOCTL_DEF_DRV(TEGRA_GEM_MMAP, tegra_gem_mmap, + DRM_IOCTL_DEF_DRV(TEGRA_CHANNEL_MAP, tegra_drm_ioctl_channel_map, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(TEGRA_CHANNEL_UNMAP, tegra_drm_ioctl_channel_unmap, + DRM_RENDER_ALLOW), + + DRM_IOCTL_DEF_DRV(TEGRA_GEM_CREATE, tegra_gem_create, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(TEGRA_GEM_MMAP, tegra_gem_mmap, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(TEGRA_SYNCPT_READ, tegra_syncpt_read, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(TEGRA_SYNCPT_INCR, tegra_syncpt_incr, @@ -779,10 +783,11 @@ static void tegra_drm_postclose(struct drm_device *drm, struct drm_file *file) struct tegra_drm_file *fpriv = file->driver_priv; mutex_lock(&fpriv->lock); - idr_for_each(&fpriv->contexts, tegra_drm_context_cleanup, NULL); + idr_for_each(&fpriv->legacy_contexts, tegra_drm_context_cleanup, NULL); + tegra_drm_uapi_close_file(fpriv); mutex_unlock(&fpriv->lock); - idr_destroy(&fpriv->contexts); + idr_destroy(&fpriv->legacy_contexts); mutex_destroy(&fpriv->lock); kfree(fpriv); } diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h index 536861017d24..8b28327c931c 100644 --- a/drivers/gpu/drm/tegra/drm.h +++ b/drivers/gpu/drm/tegra/drm.h @@ -64,12 +64,22 @@ struct tegra_drm { struct tegra_display_hub *hub; }; +static inline struct host1x *tegra_drm_to_host1x(struct tegra_drm *tegra) +{ + return dev_get_drvdata(tegra->drm->dev->parent); +} + struct tegra_drm_client; struct tegra_drm_context { struct tegra_drm_client *client; struct host1x_channel *channel; + + /* Only used by legacy UAPI. */ unsigned int id; + + /* Only used by new UAPI. */ + struct xarray mappings; }; struct tegra_drm_client_ops { diff --git a/drivers/gpu/drm/tegra/uapi.c b/drivers/gpu/drm/tegra/uapi.c new file mode 100644 index 000000000000..48e872c768a9 --- /dev/null +++ b/drivers/gpu/drm/tegra/uapi.c @@ -0,0 +1,313 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* Copyright (c) 2020 NVIDIA Corporation */ + +#include +#include +#include + +#include +#include + +#include "drm.h" +#include "uapi.h" + +static void tegra_drm_mapping_release(struct kref *ref) +{ + struct tegra_drm_mapping *mapping = + container_of(ref, struct tegra_drm_mapping, ref); + + if (mapping->sgt) + dma_unmap_sgtable(mapping->dev, mapping->sgt, + mapping->direction, DMA_ATTR_SKIP_CPU_SYNC); + + host1x_bo_unpin(mapping->dev, mapping->bo, mapping->sgt); + host1x_bo_put(mapping->bo); + + kfree(mapping); +} + +void tegra_drm_mapping_put(struct tegra_drm_mapping *mapping) +{ + kref_put(&mapping->ref, tegra_drm_mapping_release); +} + +static void tegra_drm_channel_context_close(struct tegra_drm_context *context) +{ + struct tegra_drm_mapping *mapping; + unsigned long id; + + xa_for_each(&context->mappings, id, mapping) + tegra_drm_mapping_put(mapping); + + xa_destroy(&context->mappings); + + host1x_channel_put(context->channel); + + kfree(context); +} + +void tegra_drm_uapi_close_file(struct tegra_drm_file *file) +{ + struct tegra_drm_context *context; + unsigned long id; + + xa_for_each(&file->contexts, id, context) + tegra_drm_channel_context_close(context); + + xa_destroy(&file->contexts); +} + +static struct tegra_drm_client *tegra_drm_find_client(struct tegra_drm *tegra, + u32 class) +{ + struct tegra_drm_client *client; + + list_for_each_entry(client, &tegra->clients, list) + if (client->base.class == class) + return client; + + return NULL; +} + +int tegra_drm_ioctl_channel_open(struct drm_device *drm, void *data, + struct drm_file *file) +{ + struct tegra_drm_file *fpriv = file->driver_priv; + struct tegra_drm *tegra = drm->dev_private; + struct drm_tegra_channel_open *args = data; + struct tegra_drm_client *client = NULL; + struct tegra_drm_context *context; + int err; + + if (args->flags) + return -EINVAL; + + context = kzalloc(sizeof(*context), GFP_KERNEL); + if (!context) + return -ENOMEM; + + client = tegra_drm_find_client(tegra, args->host1x_class); + if (!client) { + err = -ENODEV; + goto free; + } + + if (client->shared_channel) { + context->channel = host1x_channel_get(client->shared_channel); + } else { + context->channel = host1x_channel_request(&client->base); + if (!context->channel) { + err = -EBUSY; + goto free; + } + } + + err = xa_alloc(&fpriv->contexts, &args->context, context, + XA_LIMIT(1, U32_MAX), GFP_KERNEL); + if (err < 0) + goto put_channel; + + context->client = client; + xa_init_flags(&context->mappings, XA_FLAGS_ALLOC1); + + args->version = client->version; + args->capabilities = 0; + + if (device_get_dma_attr(client->base.dev) == DEV_DMA_COHERENT) + args->capabilities |= DRM_TEGRA_CHANNEL_CAP_CACHE_COHERENT; + + return 0; + +put_channel: + host1x_channel_put(context->channel); +free: + kfree(context); + + return err; +} + +int tegra_drm_ioctl_channel_close(struct drm_device *drm, void *data, + struct drm_file *file) +{ + struct tegra_drm_file *fpriv = file->driver_priv; + struct drm_tegra_channel_close *args = data; + struct tegra_drm_context *context; + + mutex_lock(&fpriv->lock); + + context = xa_load(&fpriv->contexts, args->context); + if (!context) { + mutex_unlock(&fpriv->lock); + return -EINVAL; + } + + xa_erase(&fpriv->contexts, args->context); + + mutex_unlock(&fpriv->lock); + + tegra_drm_channel_context_close(context); + + return 0; +} + +int tegra_drm_ioctl_channel_map(struct drm_device *drm, void *data, + struct drm_file *file) +{ + struct tegra_drm_file *fpriv = file->driver_priv; + struct drm_tegra_channel_map *args = data; + struct tegra_drm_mapping *mapping; + struct tegra_drm_context *context; + int err = 0; + + if (args->flags & ~DRM_TEGRA_CHANNEL_MAP_READ_WRITE) + return -EINVAL; + + mutex_lock(&fpriv->lock); + + context = xa_load(&fpriv->contexts, args->context); + if (!context) { + mutex_unlock(&fpriv->lock); + return -EINVAL; + } + + mapping = kzalloc(sizeof(*mapping), GFP_KERNEL); + if (!mapping) { + err = -ENOMEM; + goto unlock; + } + + kref_init(&mapping->ref); + + mapping->dev = context->client->base.dev; + mapping->bo = tegra_gem_lookup(file, args->handle); + if (!mapping->bo) { + err = -EINVAL; + goto unlock; + } + + if (context->client->base.group) { + /* IOMMU domain managed directly using IOMMU API */ + host1x_bo_pin(mapping->dev, mapping->bo, &mapping->iova); + } else { + switch (args->flags & DRM_TEGRA_CHANNEL_MAP_READ_WRITE) { + case DRM_TEGRA_CHANNEL_MAP_READ_WRITE: + mapping->direction = DMA_BIDIRECTIONAL; + break; + + case DRM_TEGRA_CHANNEL_MAP_WRITE: + mapping->direction = DMA_FROM_DEVICE; + break; + + case DRM_TEGRA_CHANNEL_MAP_READ: + mapping->direction = DMA_TO_DEVICE; + break; + + default: + return -EINVAL; + } + + mapping->sgt = host1x_bo_pin(mapping->dev, mapping->bo, NULL); + if (IS_ERR(mapping->sgt)) { + err = PTR_ERR(mapping->sgt); + goto put_gem; + } + + err = dma_map_sgtable(mapping->dev, mapping->sgt, + mapping->direction, + DMA_ATTR_SKIP_CPU_SYNC); + if (err) + goto unpin; + + mapping->iova = sg_dma_address(mapping->sgt->sgl); + } + + mapping->iova_end = mapping->iova + host1x_to_tegra_bo(mapping->bo)->size; + + err = xa_alloc(&context->mappings, &args->mapping, mapping, + XA_LIMIT(1, U32_MAX), GFP_KERNEL); + if (err < 0) + goto unmap; + + mutex_unlock(&fpriv->lock); + + return 0; + +unmap: + if (mapping->sgt) { + dma_unmap_sgtable(mapping->dev, mapping->sgt, + mapping->direction, DMA_ATTR_SKIP_CPU_SYNC); + } +unpin: + host1x_bo_unpin(mapping->dev, mapping->bo, mapping->sgt); +put_gem: + host1x_bo_put(mapping->bo); + kfree(mapping); +unlock: + mutex_unlock(&fpriv->lock); + return err; +} + +int tegra_drm_ioctl_channel_unmap(struct drm_device *drm, void *data, + struct drm_file *file) +{ + struct tegra_drm_file *fpriv = file->driver_priv; + struct drm_tegra_channel_unmap *args = data; + struct tegra_drm_mapping *mapping; + struct tegra_drm_context *context; + + mutex_lock(&fpriv->lock); + + context = xa_load(&fpriv->contexts, args->context); + if (!context) { + mutex_unlock(&fpriv->lock); + return -EINVAL; + } + + mapping = xa_erase(&context->mappings, args->mapping); + + mutex_unlock(&fpriv->lock); + + if (mapping) { + tegra_drm_mapping_put(mapping); + return 0; + } else { + return -EINVAL; + } +} + +int tegra_drm_ioctl_gem_create(struct drm_device *drm, void *data, + struct drm_file *file) +{ + struct drm_tegra_gem_create *args = data; + struct tegra_bo *bo; + + if (args->flags) + return -EINVAL; + + bo = tegra_bo_create_with_handle(file, drm, args->size, args->flags, + &args->handle); + if (IS_ERR(bo)) + return PTR_ERR(bo); + + return 0; +} + +int tegra_drm_ioctl_gem_mmap(struct drm_device *drm, void *data, + struct drm_file *file) +{ + struct drm_tegra_gem_mmap *args = data; + struct drm_gem_object *gem; + struct tegra_bo *bo; + + gem = drm_gem_object_lookup(file, args->handle); + if (!gem) + return -EINVAL; + + bo = to_tegra_bo(gem); + + args->offset = drm_vma_node_offset_addr(&bo->gem.vma_node); + + drm_gem_object_put(gem); + + return 0; +} diff --git a/drivers/gpu/drm/tegra/uapi.h b/drivers/gpu/drm/tegra/uapi.h new file mode 100644 index 000000000000..fbef39726c29 --- /dev/null +++ b/drivers/gpu/drm/tegra/uapi.h @@ -0,0 +1,51 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* Copyright (c) 2020 NVIDIA Corporation */ + +#ifndef _TEGRA_DRM_UAPI_H +#define _TEGRA_DRM_UAPI_H + +#include +#include +#include +#include + +#include + +struct drm_file; +struct drm_device; + +struct tegra_drm_file { + /* Legacy UAPI state */ + struct idr legacy_contexts; + struct mutex lock; + + /* New UAPI state */ + struct xarray contexts; +}; + +struct tegra_drm_mapping { + struct kref ref; + + struct device *dev; + struct host1x_bo *bo; + struct sg_table *sgt; + enum dma_data_direction direction; + dma_addr_t iova; + dma_addr_t iova_end; +}; + +int tegra_drm_ioctl_channel_open(struct drm_device *drm, void *data, + struct drm_file *file); +int tegra_drm_ioctl_channel_close(struct drm_device *drm, void *data, + struct drm_file *file); +int tegra_drm_ioctl_channel_map(struct drm_device *drm, void *data, + struct drm_file *file); +int tegra_drm_ioctl_channel_unmap(struct drm_device *drm, void *data, + struct drm_file *file); +int tegra_drm_ioctl_channel_submit(struct drm_device *drm, void *data, + struct drm_file *file); + +void tegra_drm_uapi_close_file(struct tegra_drm_file *file); +void tegra_drm_mapping_put(struct tegra_drm_mapping *mapping); + +#endif From patchwork Fri Jul 9 19:31:42 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503356 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=ahNE4pzY; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3Bq4GFBz9sRN for ; Sat, 10 Jul 2021 05:29:39 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229506AbhGITcW (ORCPT ); Fri, 9 Jul 2021 15:32:22 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43826 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230191AbhGITcW (ORCPT ); Fri, 9 Jul 2021 15:32:22 -0400 Received: from mail-wr1-x429.google.com (mail-wr1-x429.google.com [IPv6:2a00:1450:4864:20::429]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id B706CC0613E6 for ; Fri, 9 Jul 2021 12:29:38 -0700 (PDT) Received: by mail-wr1-x429.google.com with SMTP id f9so7887275wrq.11 for ; Fri, 09 Jul 2021 12:29:38 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=fleDkTlpZCnkAhAKVFvno1InuM+Dhrnib08wE4hr/VE=; b=ahNE4pzYVKxUy5kYAdYGrmDRiXtYKG8+n/QkPjISNglNBPwhedxhJceud0S53OCDvk 45/UouVs/lLKpxLH06WEr7N8oGyQHAuISGa4yxLGJouAX8J109dwt8abgTslDz7BM4nU 2NYomJOCCBymOI85A7m9qk4Hr+/StRKUqjdz5nUEClkuGbpvMI2gBolcqVYx77d6xM8j Fa3bnmssnut4WOh+j9eTdaNDVeuyJRgPujoFef3AuXJXvLGy8SISoSMtXxhoJTtxwR2l RmL2FzDT6CSbL093KoKrzvSEO0niCMZxO8+6SIFhuIs0b6Q9zjtokVyL5aPbeWy8GF6c oB1w== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=fleDkTlpZCnkAhAKVFvno1InuM+Dhrnib08wE4hr/VE=; b=GF3DZ1vbzhUhW3jKq7G9lTwbZDqVxg910t/T0lhtW0N+3iupO4NvhAK3w8ZbbaZoFj TMOi7ReTl2pApQEBUdHjT2onbORFrY71UTeWVd8Dvjt4Cb6P2CPsBWrEgVXLT5A4h7Hi lwMmK8M7FRhANrWcdWZUMFBs9INZIbzKpZeisyhGyg/N65fhs/t+fIqz8LQZt6FMoSqp 0VycTUU/i3H0bOcmFSaNuJal3LaGtGj7twi/GE9uj7BecdCGDcL/TSSsY4ME+G2qp+1u 9ycEgaoyyeISYAP3hI3MhvaRDFy7FMzS4CicDwlLvVhqX4IgZsn7tfpRKrAkopcqUH37 11zg== X-Gm-Message-State: AOAM531iqOqvj1i7PcFakR8+pyr5xbwkqlmkjpWIq+BmY4hYSgs/bOYD 2ruhTv1XC5UnUfdgfLPiwXk= X-Google-Smtp-Source: ABdhPJwwp2qzAkEDPdrGaHdEuk1+ij56XzDLU6J2OkRvUXUHJiMq7oJ8Bx+sSgrrTYeHduW/xdbasw== X-Received: by 2002:a5d:6c67:: with SMTP id r7mr37346415wrz.111.1625858977406; Fri, 09 Jul 2021 12:29:37 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id b12sm6096749wrx.60.2021.07.09.12.29.36 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:36 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org Subject: [PATCH v8 10/14] drm/tegra: Implement syncpoint management UAPI Date: Fri, 9 Jul 2021 21:31:42 +0200 Message-Id: <20210709193146.2859516-11-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen Implement TegraDRM IOCTLs for allocating and freeing syncpoints. Signed-off-by: Mikko Perttunen Signed-off-by: Thierry Reding --- drivers/gpu/drm/tegra/drm.c | 5 ++++ drivers/gpu/drm/tegra/uapi.c | 52 ++++++++++++++++++++++++++++++++++++ drivers/gpu/drm/tegra/uapi.h | 5 ++++ 3 files changed, 62 insertions(+) diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 6ee08e49ec57..18aee825a1ff 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -92,6 +92,7 @@ static int tegra_drm_open(struct drm_device *drm, struct drm_file *filp) idr_init_base(&fpriv->legacy_contexts, 1); xa_init_flags(&fpriv->contexts, XA_FLAGS_ALLOC1); + xa_init(&fpriv->syncpoints); mutex_init(&fpriv->lock); filp->driver_priv = fpriv; @@ -727,6 +728,10 @@ static const struct drm_ioctl_desc tegra_drm_ioctls[] = { DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(TEGRA_CHANNEL_UNMAP, tegra_drm_ioctl_channel_unmap, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(TEGRA_SYNCPOINT_ALLOCATE, tegra_drm_ioctl_syncpoint_allocate, + DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(TEGRA_SYNCPOINT_FREE, tegra_drm_ioctl_syncpoint_free, + DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(TEGRA_GEM_CREATE, tegra_gem_create, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(TEGRA_GEM_MMAP, tegra_gem_mmap, DRM_RENDER_ALLOW), diff --git a/drivers/gpu/drm/tegra/uapi.c b/drivers/gpu/drm/tegra/uapi.c index 48e872c768a9..e91394e7264e 100644 --- a/drivers/gpu/drm/tegra/uapi.c +++ b/drivers/gpu/drm/tegra/uapi.c @@ -49,12 +49,17 @@ static void tegra_drm_channel_context_close(struct tegra_drm_context *context) void tegra_drm_uapi_close_file(struct tegra_drm_file *file) { struct tegra_drm_context *context; + struct host1x_syncpt *sp; unsigned long id; xa_for_each(&file->contexts, id, context) tegra_drm_channel_context_close(context); + xa_for_each(&file->syncpoints, id, sp) + host1x_syncpt_put(sp); + xa_destroy(&file->contexts); + xa_destroy(&file->syncpoints); } static struct tegra_drm_client *tegra_drm_find_client(struct tegra_drm *tegra, @@ -311,3 +316,50 @@ int tegra_drm_ioctl_gem_mmap(struct drm_device *drm, void *data, return 0; } + +int tegra_drm_ioctl_syncpoint_allocate(struct drm_device *drm, void *data, + struct drm_file *file) +{ + struct host1x *host1x = tegra_drm_to_host1x(drm->dev_private); + struct tegra_drm_file *fpriv = file->driver_priv; + struct drm_tegra_syncpoint_allocate *args = data; + struct host1x_syncpt *sp; + int err; + + if (args->id) + return -EINVAL; + + sp = host1x_syncpt_alloc(host1x, HOST1X_SYNCPT_CLIENT_MANAGED, + current->comm); + if (!sp) + return -EBUSY; + + args->id = host1x_syncpt_id(sp); + + err = xa_insert(&fpriv->syncpoints, args->id, sp, GFP_KERNEL); + if (err) { + host1x_syncpt_put(sp); + return err; + } + + return 0; +} + +int tegra_drm_ioctl_syncpoint_free(struct drm_device *drm, void *data, + struct drm_file *file) +{ + struct tegra_drm_file *fpriv = file->driver_priv; + struct drm_tegra_syncpoint_allocate *args = data; + struct host1x_syncpt *sp; + + mutex_lock(&fpriv->lock); + sp = xa_erase(&fpriv->syncpoints, args->id); + mutex_unlock(&fpriv->lock); + + if (!sp) + return -EINVAL; + + host1x_syncpt_put(sp); + + return 0; +} diff --git a/drivers/gpu/drm/tegra/uapi.h b/drivers/gpu/drm/tegra/uapi.h index fbef39726c29..e4e498facf61 100644 --- a/drivers/gpu/drm/tegra/uapi.h +++ b/drivers/gpu/drm/tegra/uapi.h @@ -21,6 +21,7 @@ struct tegra_drm_file { /* New UAPI state */ struct xarray contexts; + struct xarray syncpoints; }; struct tegra_drm_mapping { @@ -44,6 +45,10 @@ int tegra_drm_ioctl_channel_unmap(struct drm_device *drm, void *data, struct drm_file *file); int tegra_drm_ioctl_channel_submit(struct drm_device *drm, void *data, struct drm_file *file); +int tegra_drm_ioctl_syncpoint_allocate(struct drm_device *drm, void *data, + struct drm_file *file); +int tegra_drm_ioctl_syncpoint_free(struct drm_device *drm, void *data, + struct drm_file *file); void tegra_drm_uapi_close_file(struct tegra_drm_file *file); void tegra_drm_mapping_put(struct tegra_drm_mapping *mapping); From patchwork Fri Jul 9 19:31:43 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503357 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=cyPTUDOI; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3Bt1gmnz9sPf for ; Sat, 10 Jul 2021 05:29:42 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230196AbhGITcZ (ORCPT ); Fri, 9 Jul 2021 15:32:25 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43834 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230191AbhGITcZ (ORCPT ); Fri, 9 Jul 2021 15:32:25 -0400 Received: from mail-wm1-x335.google.com (mail-wm1-x335.google.com [IPv6:2a00:1450:4864:20::335]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 97F6EC0613DD for ; Fri, 9 Jul 2021 12:29:40 -0700 (PDT) Received: by mail-wm1-x335.google.com with SMTP id a5-20020a7bc1c50000b02901e3bbe0939bso6979150wmj.0 for ; Fri, 09 Jul 2021 12:29:40 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=AdLcTRIWBNsI6B28xgdufoW2tJ7+hChXLZ6LrUe2WGw=; b=cyPTUDOINofC9gk2QIHwy4LVVdGvEaXD8Qdygp0jgNILRlWaANqbKToIOMWfkkSASs 6DvyaYpvZPAa1B21oa9zeasMrixPKK/EmrfiN0YA32xJC9FVrJEgOEGK/QKOX691WCBz xA6rz387EFwPquer+pRl67YWpsi3YnprcjbNTQLCEsfzarnZ+aRvbOuBCvsOvn3vpE04 It/whpV46PX/ksxUG0Uk5PUPphOctGEANZQxibRgRlgra5qL4PomSrIvUjy3v+5Atpk1 7KCo7yQd78J2wXBR5OQBgheXn87mnnpgGBoqEZWybp+Vq3mwBrjARhKEDBFsSJ5+If0i 2ibQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=AdLcTRIWBNsI6B28xgdufoW2tJ7+hChXLZ6LrUe2WGw=; b=fTfBRzlKaErOrMYBIWAxD9YLmyQgt5QgaAaVCmuO11OZ/Xqk6A/WtqERJUA+kQvjnz 4+ngmsRp2cv2L+H5CSfIpFIGRXHlBmcY8uabg/FxrPQm7qtrtN7xU+R74bQWJBNRFv56 glFNjddZtm/sNMXxIlPlm1ghjFe3+OcLTA968+6q6FiT8B8CcX8R6N9jxA1QJTmotIfi 7dTHzOD2NzHXtmujid+aFJ+T01Y7VJoBFOJp1gpVUvAhfUjNufK0c2rOYN31khxk0M6w c16Qm2HtlPcV5C0DCwvQjZiNxYUu85OkexTxbPXQANExp7ZHcEwiSFAsZc9v8nEiNYF7 y6EA== X-Gm-Message-State: AOAM531l1MB+2X5QiKm/Gk7oASsLXyiDbyPguwFhZv57GiWhobaEqibB W9/NvIbqr3TSIEUrjncWF4Q= X-Google-Smtp-Source: ABdhPJz3OWweGn3OR8zd7DEwqR9tFVenHuC7VNywboQIBKxV8uSe7j9G8yOGGb3BgQh1TBRwWh7xeQ== X-Received: by 2002:a7b:ca50:: with SMTP id m16mr536046wml.140.1625858979234; Fri, 09 Jul 2021 12:29:39 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id s24sm6426081wra.33.2021.07.09.12.29.38 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:38 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org Subject: [PATCH v8 11/14] drm/tegra: Implement syncpoint wait UAPI Date: Fri, 9 Jul 2021 21:31:43 +0200 Message-Id: <20210709193146.2859516-12-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen Implement new syncpoint wait UAPI. This is different from the legacy one in taking an absolute timestamp in line with modern DRM conventions. Signed-off-by: Mikko Perttunen Signed-off-by: Thierry Reding --- Changes in v8: - rebase on top of latest UABI changes drivers/gpu/drm/tegra/drm.c | 2 ++ drivers/gpu/drm/tegra/uapi.c | 22 ++++++++++++++++++++++ drivers/gpu/drm/tegra/uapi.h | 2 ++ 3 files changed, 26 insertions(+) diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 18aee825a1ff..082d520bdffc 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -732,6 +732,8 @@ static const struct drm_ioctl_desc tegra_drm_ioctls[] = { DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(TEGRA_SYNCPOINT_FREE, tegra_drm_ioctl_syncpoint_free, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(TEGRA_SYNCPOINT_WAIT, tegra_drm_ioctl_syncpoint_wait, + DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(TEGRA_GEM_CREATE, tegra_gem_create, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(TEGRA_GEM_MMAP, tegra_gem_mmap, DRM_RENDER_ALLOW), diff --git a/drivers/gpu/drm/tegra/uapi.c b/drivers/gpu/drm/tegra/uapi.c index e91394e7264e..0ba57697706b 100644 --- a/drivers/gpu/drm/tegra/uapi.c +++ b/drivers/gpu/drm/tegra/uapi.c @@ -7,6 +7,7 @@ #include #include +#include #include "drm.h" #include "uapi.h" @@ -363,3 +364,24 @@ int tegra_drm_ioctl_syncpoint_free(struct drm_device *drm, void *data, return 0; } + +int tegra_drm_ioctl_syncpoint_wait(struct drm_device *drm, void *data, + struct drm_file *file) +{ + struct host1x *host1x = tegra_drm_to_host1x(drm->dev_private); + struct drm_tegra_syncpoint_wait *args = data; + signed long timeout_jiffies; + struct host1x_syncpt *sp; + + if (args->padding != 0) + return -EINVAL; + + sp = host1x_syncpt_get_by_id_noref(host1x, args->id); + if (!sp) + return -EINVAL; + + timeout_jiffies = drm_timeout_abs_to_jiffies(args->timeout_ns); + + return host1x_syncpt_wait(sp, args->threshold, timeout_jiffies, + &args->value); +} diff --git a/drivers/gpu/drm/tegra/uapi.h b/drivers/gpu/drm/tegra/uapi.h index e4e498facf61..12adad770ad3 100644 --- a/drivers/gpu/drm/tegra/uapi.h +++ b/drivers/gpu/drm/tegra/uapi.h @@ -49,6 +49,8 @@ int tegra_drm_ioctl_syncpoint_allocate(struct drm_device *drm, void *data, struct drm_file *file); int tegra_drm_ioctl_syncpoint_free(struct drm_device *drm, void *data, struct drm_file *file); +int tegra_drm_ioctl_syncpoint_wait(struct drm_device *drm, void *data, + struct drm_file *file); void tegra_drm_uapi_close_file(struct tegra_drm_file *file); void tegra_drm_mapping_put(struct tegra_drm_mapping *mapping); From patchwork Fri Jul 9 19:31:44 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503358 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=YS2BBhSm; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3Bw2lgkz9sPf for ; Sat, 10 Jul 2021 05:29:44 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230221AbhGITc1 (ORCPT ); Fri, 9 Jul 2021 15:32:27 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43848 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230191AbhGITc1 (ORCPT ); Fri, 9 Jul 2021 15:32:27 -0400 Received: from mail-wm1-x32e.google.com (mail-wm1-x32e.google.com [IPv6:2a00:1450:4864:20::32e]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 9593EC0613E5 for ; Fri, 9 Jul 2021 12:29:42 -0700 (PDT) Received: by mail-wm1-x32e.google.com with SMTP id j34so6967489wms.5 for ; Fri, 09 Jul 2021 12:29:42 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=/CQ3Y84vBWk9hYr6q0rTSMwSq3Rh7cgxgqd+VyRI3+4=; b=YS2BBhSmD313F0I0iD/BtkPXhLY6+a/gV0ow9O7DtzVvFqIaVorXMJR9PPLa2lCxEb +uT+nYLk0WQNh8maP2iNWq+JZdwD5q6o7DZaZ2HE9dFn3djhmsj+gf3NrvwCNzw4gGHW LMXWkSMyAyYfmDFDi/H0MxSxFhhgmC+D2tMplG8C/0JAG4bjRD9wrDOTvvvwAtiV/e3r SKZOoIJLiPvT1tfBb0iet/gnZCqBiZ4wNxHMjjn0hMrGgqX2jA/j3Q0Ak12ZwxZZC0ER jAOctMKrv8Bb8R277E5WImtxuZfbroxMnX7b28OZpm918yOggSfe6dzb/Dc3K05U3qMV /Syg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=/CQ3Y84vBWk9hYr6q0rTSMwSq3Rh7cgxgqd+VyRI3+4=; b=CL2VgivT86i3J0FTPCWOW6AMqJ0ZbM0yVbAesurVXjNrzj6ZbL3uMDbN5BLijvMdlN Sr0J3oAszNzrVxGygxBaxKByCGF1m7ttvZcNeOlU1gB9Hi9PNr+0G/FDfrPZpISR4v/R kzHXLDL06JOEauc+Ns/a2uAmb8mSlnBlWtTZ2K9bAJK7ivLXIva4jNp/EeANN6VBcl0f Uqr0FK+epBhoHFASHxC0euAD8BSpTFnKYua+jLGkwSmdOp/qgEOwgnUmQ9yo7/Yu3T9o LuweIwfi2miCxZmphj9kIbjPIChUuz1ZoWQJTBcf5XaPuoJHkQf8y1/yVAiVvE2oh+4J /L5A== X-Gm-Message-State: AOAM532ezdiJFwqbV1p7I011pF10vIoIa7EkKdzFsYwGJ03yNcIjFQ39 NIkSWEltq4poO8EiPc6euW2QtX6Jqo8= X-Google-Smtp-Source: ABdhPJxo+oHvmnVb+UZNxIKX77Qko1gpIjRPJPpojlQ8TKVkm+QavhlCLVxQ7XsCjjv+g8lkg3RLmA== X-Received: by 2002:a05:600c:3209:: with SMTP id r9mr521669wmp.79.1625858981115; Fri, 09 Jul 2021 12:29:41 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id s13sm2032391wrm.13.2021.07.09.12.29.40 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:40 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org Subject: [PATCH v8 12/14] drm/tegra: Implement job submission part of new UAPI Date: Fri, 9 Jul 2021 21:31:44 +0200 Message-Id: <20210709193146.2859516-13-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen Implement the job submission IOCTL with a minimum feature set. Signed-off-by: Mikko Perttunen Signed-off-by: Thierry Reding --- Changes in v8: - rebased on top of latest UABI changes drivers/gpu/drm/tegra/Makefile | 2 + drivers/gpu/drm/tegra/drm.c | 4 +- drivers/gpu/drm/tegra/gather_bo.c | 81 +++++ drivers/gpu/drm/tegra/gather_bo.h | 22 ++ drivers/gpu/drm/tegra/submit.c | 521 ++++++++++++++++++++++++++++++ drivers/gpu/drm/tegra/submit.h | 17 + 6 files changed, 646 insertions(+), 1 deletion(-) create mode 100644 drivers/gpu/drm/tegra/gather_bo.c create mode 100644 drivers/gpu/drm/tegra/gather_bo.h create mode 100644 drivers/gpu/drm/tegra/submit.c create mode 100644 drivers/gpu/drm/tegra/submit.h diff --git a/drivers/gpu/drm/tegra/Makefile b/drivers/gpu/drm/tegra/Makefile index 783475ffd943..ab4289d1c991 100644 --- a/drivers/gpu/drm/tegra/Makefile +++ b/drivers/gpu/drm/tegra/Makefile @@ -4,6 +4,8 @@ ccflags-$(CONFIG_DRM_TEGRA_DEBUG) += -DDEBUG tegra-drm-y := \ drm.o \ uapi.o \ + submit.o \ + gather_bo.o \ gem.o \ fb.o \ dp.o \ diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 082d520bdffc..ae9dafc32c2b 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -728,6 +728,8 @@ static const struct drm_ioctl_desc tegra_drm_ioctls[] = { DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(TEGRA_CHANNEL_UNMAP, tegra_drm_ioctl_channel_unmap, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(TEGRA_CHANNEL_SUBMIT, tegra_drm_ioctl_channel_submit, + DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(TEGRA_SYNCPOINT_ALLOCATE, tegra_drm_ioctl_syncpoint_allocate, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(TEGRA_SYNCPOINT_FREE, tegra_drm_ioctl_syncpoint_free, @@ -852,7 +854,7 @@ static void tegra_debugfs_init(struct drm_minor *minor) static const struct drm_driver tegra_drm_driver = { .driver_features = DRIVER_MODESET | DRIVER_GEM | - DRIVER_ATOMIC | DRIVER_RENDER, + DRIVER_ATOMIC | DRIVER_RENDER | DRIVER_SYNCOBJ, .open = tegra_drm_open, .postclose = tegra_drm_postclose, .lastclose = drm_fb_helper_lastclose, diff --git a/drivers/gpu/drm/tegra/gather_bo.c b/drivers/gpu/drm/tegra/gather_bo.c new file mode 100644 index 000000000000..755c2da62361 --- /dev/null +++ b/drivers/gpu/drm/tegra/gather_bo.c @@ -0,0 +1,81 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* Copyright (c) 2020 NVIDIA Corporation */ + +#include +#include + +#include "gather_bo.h" + +static struct host1x_bo *gather_bo_get(struct host1x_bo *host_bo) +{ + struct gather_bo *bo = container_of(host_bo, struct gather_bo, base); + + kref_get(&bo->ref); + + return host_bo; +} + +static void gather_bo_release(struct kref *ref) +{ + struct gather_bo *bo = container_of(ref, struct gather_bo, ref); + + kfree(bo->gather_data); + kfree(bo); +} + +void gather_bo_put(struct host1x_bo *host_bo) +{ + struct gather_bo *bo = container_of(host_bo, struct gather_bo, base); + + kref_put(&bo->ref, gather_bo_release); +} + +static struct sg_table * +gather_bo_pin(struct device *dev, struct host1x_bo *host_bo, dma_addr_t *phys) +{ + struct gather_bo *bo = container_of(host_bo, struct gather_bo, base); + struct sg_table *sgt; + int err; + + sgt = kzalloc(sizeof(*sgt), GFP_KERNEL); + if (!sgt) + return ERR_PTR(-ENOMEM); + + err = sg_alloc_table(sgt, 1, GFP_KERNEL); + if (err) { + kfree(sgt); + return ERR_PTR(err); + } + + sg_init_one(sgt->sgl, bo->gather_data, bo->gather_data_words*4); + + return sgt; +} + +static void gather_bo_unpin(struct device *dev, struct sg_table *sgt) +{ + if (sgt) { + sg_free_table(sgt); + kfree(sgt); + } +} + +static void *gather_bo_mmap(struct host1x_bo *host_bo) +{ + struct gather_bo *bo = container_of(host_bo, struct gather_bo, base); + + return bo->gather_data; +} + +static void gather_bo_munmap(struct host1x_bo *host_bo, void *addr) +{ +} + +const struct host1x_bo_ops gather_bo_ops = { + .get = gather_bo_get, + .put = gather_bo_put, + .pin = gather_bo_pin, + .unpin = gather_bo_unpin, + .mmap = gather_bo_mmap, + .munmap = gather_bo_munmap, +}; diff --git a/drivers/gpu/drm/tegra/gather_bo.h b/drivers/gpu/drm/tegra/gather_bo.h new file mode 100644 index 000000000000..6b4c9d83ac91 --- /dev/null +++ b/drivers/gpu/drm/tegra/gather_bo.h @@ -0,0 +1,22 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* Copyright (c) 2020 NVIDIA Corporation */ + +#ifndef _TEGRA_DRM_SUBMIT_GATHER_BO_H +#define _TEGRA_DRM_SUBMIT_GATHER_BO_H + +#include +#include + +struct gather_bo { + struct host1x_bo base; + + struct kref ref; + + u32 *gather_data; + size_t gather_data_words; +}; + +extern const struct host1x_bo_ops gather_bo_ops; +void gather_bo_put(struct host1x_bo *host_bo); + +#endif diff --git a/drivers/gpu/drm/tegra/submit.c b/drivers/gpu/drm/tegra/submit.c new file mode 100644 index 000000000000..0225706e3163 --- /dev/null +++ b/drivers/gpu/drm/tegra/submit.c @@ -0,0 +1,521 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* Copyright (c) 2020 NVIDIA Corporation */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include + +#include "drm.h" +#include "gather_bo.h" +#include "gem.h" +#include "submit.h" +#include "uapi.h" + +#define SUBMIT_ERR(context, fmt, ...) \ + dev_err_ratelimited(context->client->base.dev, \ + "%s: job submission failed: " fmt "\n", \ + current->comm __VA_OPT__(,) __VA_ARGS__) + +static struct tegra_drm_mapping * +tegra_drm_mapping_get(struct tegra_drm_context *context, u32 id) +{ + struct tegra_drm_mapping *mapping; + + xa_lock(&context->mappings); + mapping = xa_load(&context->mappings, id); + if (mapping) + kref_get(&mapping->ref); + xa_unlock(&context->mappings); + + return mapping; +} + +static void *alloc_copy_user_array(void __user *from, size_t count, size_t size) +{ + size_t copy_len; + void *data; + + if (check_mul_overflow(count, size, ©_len)) + return ERR_PTR(-EINVAL); + + if (copy_len > 0x4000) + return ERR_PTR(-E2BIG); + + data = kvmalloc(copy_len, GFP_KERNEL); + if (!data) + return ERR_PTR(-ENOMEM); + + if (copy_from_user(data, from, copy_len)) { + kvfree(data); + return ERR_PTR(-EFAULT); + } + + return data; +} + +static int submit_copy_gather_data(struct gather_bo **pbo, struct tegra_drm_context *context, + struct drm_tegra_channel_submit *args) +{ + struct gather_bo *bo; + size_t copy_len; + + if (args->gather_data_words == 0) { + SUBMIT_ERR(context, "gather_data_words cannot be zero"); + return -EINVAL; + } + + if (check_mul_overflow((size_t)args->gather_data_words, (size_t)4, ©_len)) { + SUBMIT_ERR(context, "gather_data_words is too large"); + return -EINVAL; + } + + bo = kzalloc(sizeof(*bo), GFP_KERNEL); + if (!bo) { + SUBMIT_ERR(context, "failed to allocate memory for bo info"); + return -ENOMEM; + } + + kref_init(&bo->ref); + host1x_bo_init(&bo->base, &gather_bo_ops); + + bo->gather_data = kmalloc(copy_len, GFP_KERNEL | __GFP_NOWARN); + if (!bo->gather_data) { + SUBMIT_ERR(context, "failed to allocate memory for gather data"); + kfree(bo); + return -ENOMEM; + } + + if (copy_from_user(bo->gather_data, u64_to_user_ptr(args->gather_data_ptr), copy_len)) { + SUBMIT_ERR(context, "failed to copy gather data from userspace"); + kfree(bo->gather_data); + kfree(bo); + return -EFAULT; + } + + bo->gather_data_words = args->gather_data_words; + + *pbo = bo; + + return 0; +} + +static int submit_write_reloc(struct tegra_drm_context *context, struct gather_bo *bo, + struct drm_tegra_submit_buf *buf, struct tegra_drm_mapping *mapping) +{ + /* TODO check that target_offset is within bounds */ + dma_addr_t iova = mapping->iova + buf->reloc.target_offset; + u32 written_ptr; + +#ifdef CONFIG_ARCH_DMA_ADDR_T_64BIT + if (buf->flags & DRM_TEGRA_SUBMIT_RELOC_SECTOR_LAYOUT) + iova |= BIT_ULL(39); +#endif + + written_ptr = iova >> buf->reloc.shift; + + if (buf->reloc.gather_offset_words >= bo->gather_data_words) { + SUBMIT_ERR(context, "relocation has too large gather offset (%u vs gather length %zu)", + buf->reloc.gather_offset_words, bo->gather_data_words); + return -EINVAL; + } + + buf->reloc.gather_offset_words = array_index_nospec(buf->reloc.gather_offset_words, + bo->gather_data_words); + + bo->gather_data[buf->reloc.gather_offset_words] = written_ptr; + + return 0; +} + +static int submit_process_bufs(struct tegra_drm_context *context, struct gather_bo *bo, + struct drm_tegra_channel_submit *args, + struct tegra_drm_submit_data *job_data) +{ + struct tegra_drm_used_mapping *mappings; + struct drm_tegra_submit_buf *bufs; + int err; + u32 i; + + bufs = alloc_copy_user_array(u64_to_user_ptr(args->bufs_ptr), args->num_bufs, + sizeof(*bufs)); + if (IS_ERR(bufs)) { + SUBMIT_ERR(context, "failed to copy bufs array from userspace"); + return PTR_ERR(bufs); + } + + mappings = kcalloc(args->num_bufs, sizeof(*mappings), GFP_KERNEL); + if (!mappings) { + SUBMIT_ERR(context, "failed to allocate memory for mapping info"); + err = -ENOMEM; + goto done; + } + + for (i = 0; i < args->num_bufs; i++) { + struct drm_tegra_submit_buf *buf = &bufs[i]; + struct tegra_drm_mapping *mapping; + + if (buf->flags & ~DRM_TEGRA_SUBMIT_RELOC_SECTOR_LAYOUT) { + SUBMIT_ERR(context, "invalid flag specified for buf"); + err = -EINVAL; + goto drop_refs; + } + + mapping = tegra_drm_mapping_get(context, buf->mapping); + if (!mapping) { + SUBMIT_ERR(context, "invalid mapping ID '%u' for buffer", buf->mapping); + err = -EINVAL; + goto drop_refs; + } + + err = submit_write_reloc(context, bo, buf, mapping); + if (err) { + tegra_drm_mapping_put(mapping); + goto drop_refs; + } + + mappings[i].mapping = mapping; + mappings[i].flags = buf->flags; + } + + job_data->used_mappings = mappings; + job_data->num_used_mappings = i; + + err = 0; + + goto done; + +drop_refs: + while (i--) + tegra_drm_mapping_put(mappings[i].mapping); + + kfree(mappings); + job_data->used_mappings = NULL; + +done: + kvfree(bufs); + + return err; +} + +static int submit_get_syncpt(struct tegra_drm_context *context, struct host1x_job *job, + struct xarray *syncpoints, struct drm_tegra_channel_submit *args) +{ + struct host1x_syncpt *sp; + + if (args->syncpt.flags) { + SUBMIT_ERR(context, "invalid flag specified for syncpt"); + return -EINVAL; + } + + /* Syncpt ref will be dropped on job release */ + sp = xa_load(syncpoints, args->syncpt.id); + if (!sp) { + SUBMIT_ERR(context, "syncpoint specified in syncpt was not allocated"); + return -EINVAL; + } + + job->syncpt = host1x_syncpt_get(sp); + job->syncpt_incrs = args->syncpt.increments; + + return 0; +} + +static int submit_job_add_gather(struct host1x_job *job, struct tegra_drm_context *context, + struct drm_tegra_submit_cmd_gather_uptr *cmd, + struct gather_bo *bo, u32 *offset, + struct tegra_drm_submit_data *job_data) +{ + u32 next_offset; + + if (cmd->reserved[0] || cmd->reserved[1] || cmd->reserved[2]) { + SUBMIT_ERR(context, "non-zero reserved field in GATHER_UPTR command"); + return -EINVAL; + } + + /* Check for maximum gather size */ + if (cmd->words > 16383) { + SUBMIT_ERR(context, "too many words in GATHER_UPTR command"); + return -EINVAL; + } + + if (check_add_overflow(*offset, cmd->words, &next_offset)) { + SUBMIT_ERR(context, "too many total words in job"); + return -EINVAL; + } + + if (next_offset > bo->gather_data_words) { + SUBMIT_ERR(context, "GATHER_UPTR command overflows gather data"); + return -EINVAL; + } + + host1x_job_add_gather(job, &bo->base, cmd->words, *offset * 4); + + *offset = next_offset; + + return 0; +} + +static struct host1x_job * +submit_create_job(struct tegra_drm_context *context, struct gather_bo *bo, + struct drm_tegra_channel_submit *args, struct tegra_drm_submit_data *job_data, + struct xarray *syncpoints) +{ + struct drm_tegra_submit_cmd *cmds; + u32 i, gather_offset = 0, class; + struct host1x_job *job; + int err; + + /* Set initial class for firewall. */ + class = context->client->base.class; + + cmds = alloc_copy_user_array(u64_to_user_ptr(args->cmds_ptr), args->num_cmds, + sizeof(*cmds)); + if (IS_ERR(cmds)) { + SUBMIT_ERR(context, "failed to copy cmds array from userspace"); + return ERR_CAST(cmds); + } + + job = host1x_job_alloc(context->channel, args->num_cmds, 0); + if (!job) { + SUBMIT_ERR(context, "failed to allocate memory for job"); + job = ERR_PTR(-ENOMEM); + goto done; + } + + err = submit_get_syncpt(context, job, syncpoints, args); + if (err < 0) + goto free_job; + + job->client = &context->client->base; + job->class = context->client->base.class; + job->serialize = true; + + for (i = 0; i < args->num_cmds; i++) { + struct drm_tegra_submit_cmd *cmd = &cmds[i]; + + if (cmd->flags) { + SUBMIT_ERR(context, "unknown flags given for cmd"); + err = -EINVAL; + goto free_job; + } + + if (cmd->type == DRM_TEGRA_SUBMIT_CMD_GATHER_UPTR) { + err = submit_job_add_gather(job, context, &cmd->gather_uptr, bo, + &gather_offset, job_data); + if (err) + goto free_job; + } else if (cmd->type == DRM_TEGRA_SUBMIT_CMD_WAIT_SYNCPT) { + if (cmd->wait_syncpt.reserved[0] || cmd->wait_syncpt.reserved[1]) { + SUBMIT_ERR(context, "non-zero reserved value"); + err = -EINVAL; + goto free_job; + } + + host1x_job_add_wait(job, cmd->wait_syncpt.id, cmd->wait_syncpt.value, + false, class); + } else if (cmd->type == DRM_TEGRA_SUBMIT_CMD_WAIT_SYNCPT_RELATIVE) { + if (cmd->wait_syncpt.reserved[0] || cmd->wait_syncpt.reserved[1]) { + SUBMIT_ERR(context, "non-zero reserved value"); + err = -EINVAL; + goto free_job; + } + + if (cmd->wait_syncpt.id != args->syncpt.id) { + SUBMIT_ERR(context, "syncpoint ID in CMD_WAIT_SYNCPT_RELATIVE is not used by the job"); + err = -EINVAL; + goto free_job; + } + + host1x_job_add_wait(job, cmd->wait_syncpt.id, cmd->wait_syncpt.value, + true, class); + } else { + SUBMIT_ERR(context, "unknown cmd type"); + err = -EINVAL; + goto free_job; + } + } + + if (gather_offset == 0) { + SUBMIT_ERR(context, "job must have at least one gather"); + err = -EINVAL; + goto free_job; + } + + goto done; + +free_job: + host1x_job_put(job); + job = ERR_PTR(err); + +done: + kvfree(cmds); + + return job; +} + +static void release_job(struct host1x_job *job) +{ + struct tegra_drm_client *client = container_of(job->client, struct tegra_drm_client, base); + struct tegra_drm_submit_data *job_data = job->user_data; + u32 i; + + for (i = 0; i < job_data->num_used_mappings; i++) + tegra_drm_mapping_put(job_data->used_mappings[i].mapping); + + kfree(job_data->used_mappings); + kfree(job_data); + + if (pm_runtime_enabled(client->base.dev)) + pm_runtime_put_autosuspend(client->base.dev); +} + +int tegra_drm_ioctl_channel_submit(struct drm_device *drm, void *data, + struct drm_file *file) +{ + struct tegra_drm_file *fpriv = file->driver_priv; + struct drm_tegra_channel_submit *args = data; + struct tegra_drm_submit_data *job_data; + struct drm_syncobj *syncobj = NULL; + struct tegra_drm_context *context; + struct host1x_job *job; + struct gather_bo *bo; + u32 i; + int err; + + mutex_lock(&fpriv->lock); + + context = xa_load(&fpriv->contexts, args->context); + if (!context) { + mutex_unlock(&fpriv->lock); + pr_err_ratelimited("%s: %s: invalid channel context '%#x'", __func__, + current->comm, args->context); + return -EINVAL; + } + + if (args->syncobj_in) { + struct dma_fence *fence; + + err = drm_syncobj_find_fence(file, args->syncobj_in, 0, 0, &fence); + if (err) { + SUBMIT_ERR(context, "invalid syncobj_in '%#x'", args->syncobj_in); + goto unlock; + } + + err = dma_fence_wait_timeout(fence, true, msecs_to_jiffies(10000)); + dma_fence_put(fence); + if (err) { + SUBMIT_ERR(context, "wait for syncobj_in timed out"); + goto unlock; + } + } + + if (args->syncobj_out) { + syncobj = drm_syncobj_find(file, args->syncobj_out); + if (!syncobj) { + SUBMIT_ERR(context, "invalid syncobj_out '%#x'", args->syncobj_out); + err = -ENOENT; + goto unlock; + } + } + + /* Allocate gather BO and copy gather words in. */ + err = submit_copy_gather_data(&bo, context, args); + if (err) + goto unlock; + + job_data = kzalloc(sizeof(*job_data), GFP_KERNEL); + if (!job_data) { + SUBMIT_ERR(context, "failed to allocate memory for job data"); + err = -ENOMEM; + goto put_bo; + } + + /* Get data buffer mappings and do relocation patching. */ + err = submit_process_bufs(context, bo, args, job_data); + if (err) + goto free_job_data; + + /* Allocate host1x_job and add gathers and waits to it. */ + job = submit_create_job(context, bo, args, job_data, &fpriv->syncpoints); + if (IS_ERR(job)) { + err = PTR_ERR(job); + goto free_job_data; + } + + /* Map gather data for Host1x. */ + err = host1x_job_pin(job, context->client->base.dev); + if (err) + goto put_job; + + /* Boot engine. */ + if (pm_runtime_enabled(context->client->base.dev)) { + err = pm_runtime_resume_and_get(context->client->base.dev); + if (err < 0) { + SUBMIT_ERR(context, "could not power up engine: %d", err); + goto unpin_job; + } + } + + job->user_data = job_data; + job->release = release_job; + job->timeout = 10000; + + /* + * job_data is now part of job reference counting, so don't release + * it from here. + */ + job_data = NULL; + + /* Submit job to hardware. */ + err = host1x_job_submit(job); + if (err) { + SUBMIT_ERR(context, "host1x job submission failed: %d", err); + goto unpin_job; + } + + /* Return postfences to userspace and add fences to DMA reservations. */ + args->syncpt.value = job->syncpt_end; + + if (syncobj) { + struct dma_fence *fence = host1x_fence_create(job->syncpt, job->syncpt_end); + if (IS_ERR(fence)) + err = PTR_ERR(fence); + + drm_syncobj_replace_fence(syncobj, fence); + } + + goto put_job; + +unpin_job: + host1x_job_unpin(job); +put_job: + host1x_job_put(job); +free_job_data: + if (job_data && job_data->used_mappings) { + for (i = 0; i < job_data->num_used_mappings; i++) + tegra_drm_mapping_put(job_data->used_mappings[i].mapping); + + kfree(job_data->used_mappings); + } + + if (job_data) + kfree(job_data); +put_bo: + gather_bo_put(&bo->base); +unlock: + if (syncobj) + drm_syncobj_put(syncobj); + + mutex_unlock(&fpriv->lock); + return err; +} diff --git a/drivers/gpu/drm/tegra/submit.h b/drivers/gpu/drm/tegra/submit.h new file mode 100644 index 000000000000..0a165e9e4bda --- /dev/null +++ b/drivers/gpu/drm/tegra/submit.h @@ -0,0 +1,17 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* Copyright (c) 2020 NVIDIA Corporation */ + +#ifndef _TEGRA_DRM_UAPI_SUBMIT_H +#define _TEGRA_DRM_UAPI_SUBMIT_H + +struct tegra_drm_used_mapping { + struct tegra_drm_mapping *mapping; + u32 flags; +}; + +struct tegra_drm_submit_data { + struct tegra_drm_used_mapping *used_mappings; + u32 num_used_mappings; +}; + +#endif From patchwork Fri Jul 9 19:31:45 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503359 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=uiV620pm; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3By0kt1z9sPf for ; Sat, 10 Jul 2021 05:29:46 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230233AbhGITc3 (ORCPT ); Fri, 9 Jul 2021 15:32:29 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43856 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230191AbhGITc3 (ORCPT ); Fri, 9 Jul 2021 15:32:29 -0400 Received: from mail-wm1-x32d.google.com (mail-wm1-x32d.google.com [IPv6:2a00:1450:4864:20::32d]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 4379DC0613DD for ; Fri, 9 Jul 2021 12:29:44 -0700 (PDT) Received: by mail-wm1-x32d.google.com with SMTP id a5-20020a7bc1c50000b02901e3bbe0939bso6979251wmj.0 for ; Fri, 09 Jul 2021 12:29:44 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=Isvi3/x44UzMyQwAu2oXCJWsW3MfoQvAQpXOQZ41AUA=; b=uiV620pmGLtybo+gE7xDLFgFV34/JFByTec9u6HCLoPrAt0HWpTTsEISpQ9J0xaD2Q lYCSnURP8d0w9quFtH1RIH7Uwv9xjh+CGXvzUsPSB0RJqg3UpIw0qHRSSAfv9531oKrv 2f3qUsjRIl60e7t0DUhdFyzKjQCYkgoGHhzUdnvTK+AUWQRu/kV9W4mZo34FH/c5kGEV ccVeQOJzDJXZxShkXkwVyPdK79oW9JIa+YqM/WnrMtIW/ntHWY51dk2aTdoj5zRXCA0D vXCxMB+MEh7BdngtVTN6ejR7t/PexZOei1nbwvQBttFhomOm10I+OweorRJu3lx1UAPE UmvA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=Isvi3/x44UzMyQwAu2oXCJWsW3MfoQvAQpXOQZ41AUA=; b=Z5np/yM4pCnenn1MlEhPd4yS9dzGN87g/kC4nAwVXQ9Tel1B8qlpIbcR31H3mdmjgi dqbcgT/q4do/V8dnqce6lGtDl4zsMuEkOFrhYnY+LFsZ5igK5Ndo+DTbo5kH4GaJ2jVw RVMbJ37VdaqcXOy7n8ctgcVwBthu8nilhX9kwjq97IjsiZpX7dvEV0+5EeoRnXGDzAr3 K/LC/VLq+0ma4uM/OxrvX7wdDazjTZjHv3a3lPjI2LqhFJ4rDxBDFu+ujaY35D//rmh4 3CuemPMrzTrNORpsN/udmApXLyM9Po4UDuC7pfxd4vmm0zgmlRQjyrmad9NXnDqy13DQ cUnw== X-Gm-Message-State: AOAM531I1iBckk61iqhk0rQiym5Vhh7B276AYRw5t1bYtJhunhPkKE4O pVpAO9BcPxsQ8dj/SUAgiyo= X-Google-Smtp-Source: ABdhPJyXAYkiZEydo/FIWDVxRj8d46lfTsg0O8Mk5qP0lQx+6fPmMa/pK3ikpz8FbevTlJBev+a7AQ== X-Received: by 2002:a05:600c:1c19:: with SMTP id j25mr544594wms.166.1625858982840; Fri, 09 Jul 2021 12:29:42 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id y23sm5739410wmi.28.2021.07.09.12.29.41 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:42 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org Subject: [PATCH v8 13/14] drm/tegra: Add job firewall Date: Fri, 9 Jul 2021 21:31:45 +0200 Message-Id: <20210709193146.2859516-14-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen Add a firewall that validates jobs before submission to ensure they don't do anything they aren't allowed to do, like accessing memory they should not access. The firewall is functionality-wise a copy of the firewall already implemented in gpu/host1x. It is copied here as it makes more sense for it to live on the DRM side, as it is only needed for userspace job submissions, and generally the data it needs to do its job is easier to access here. In the future, the other implementation will be removed. Signed-off-by: Mikko Perttunen Signed-off-by: Thierry Reding --- drivers/gpu/drm/tegra/Makefile | 1 + drivers/gpu/drm/tegra/firewall.c | 254 +++++++++++++++++++++++++++++++ drivers/gpu/drm/tegra/submit.c | 10 +- drivers/gpu/drm/tegra/submit.h | 4 + 4 files changed, 267 insertions(+), 2 deletions(-) create mode 100644 drivers/gpu/drm/tegra/firewall.c diff --git a/drivers/gpu/drm/tegra/Makefile b/drivers/gpu/drm/tegra/Makefile index ab4289d1c991..5d2039f0c734 100644 --- a/drivers/gpu/drm/tegra/Makefile +++ b/drivers/gpu/drm/tegra/Makefile @@ -6,6 +6,7 @@ tegra-drm-y := \ uapi.o \ submit.o \ gather_bo.o \ + firewall.o \ gem.o \ fb.o \ dp.o \ diff --git a/drivers/gpu/drm/tegra/firewall.c b/drivers/gpu/drm/tegra/firewall.c new file mode 100644 index 000000000000..1824d2db0e2c --- /dev/null +++ b/drivers/gpu/drm/tegra/firewall.c @@ -0,0 +1,254 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* Copyright (c) 2010-2020 NVIDIA Corporation */ + +#include "drm.h" +#include "submit.h" +#include "uapi.h" + +struct tegra_drm_firewall { + struct tegra_drm_submit_data *submit; + struct tegra_drm_client *client; + u32 *data; + u32 pos; + u32 end; + u32 class; +}; + +static int fw_next(struct tegra_drm_firewall *fw, u32 *word) +{ + if (fw->pos == fw->end) + return -EINVAL; + + *word = fw->data[fw->pos++]; + + return 0; +} + +static bool fw_check_addr_valid(struct tegra_drm_firewall *fw, u32 offset) +{ + u32 i; + + for (i = 0; i < fw->submit->num_used_mappings; i++) { + struct tegra_drm_mapping *m = fw->submit->used_mappings[i].mapping; + + if (offset >= m->iova && offset <= m->iova_end) + return true; + } + + return false; +} + +static int fw_check_reg(struct tegra_drm_firewall *fw, u32 offset) +{ + bool is_addr; + u32 word; + int err; + + err = fw_next(fw, &word); + if (err) + return err; + + if (!fw->client->ops->is_addr_reg) + return 0; + + is_addr = fw->client->ops->is_addr_reg(fw->client->base.dev, fw->class, + offset); + + if (!is_addr) + return 0; + + if (!fw_check_addr_valid(fw, word)) + return -EINVAL; + + return 0; +} + +static int fw_check_regs_seq(struct tegra_drm_firewall *fw, u32 offset, + u32 count, bool incr) +{ + u32 i; + + for (i = 0; i < count; i++) { + if (fw_check_reg(fw, offset)) + return -EINVAL; + + if (incr) + offset++; + } + + return 0; +} + +static int fw_check_regs_mask(struct tegra_drm_firewall *fw, u32 offset, + u16 mask) +{ + unsigned long bmask = mask; + unsigned int bit; + + for_each_set_bit(bit, &bmask, 16) { + if (fw_check_reg(fw, offset+bit)) + return -EINVAL; + } + + return 0; +} + +static int fw_check_regs_imm(struct tegra_drm_firewall *fw, u32 offset) +{ + bool is_addr; + + is_addr = fw->client->ops->is_addr_reg(fw->client->base.dev, fw->class, + offset); + if (is_addr) + return -EINVAL; + + return 0; +} + +static int fw_check_class(struct tegra_drm_firewall *fw, u32 class) +{ + if (!fw->client->ops->is_valid_class) { + if (class == fw->client->base.class) + return 0; + else + return -EINVAL; + } + + if (!fw->client->ops->is_valid_class(class)) + return -EINVAL; + + return 0; +} + +enum { + HOST1X_OPCODE_SETCLASS = 0x00, + HOST1X_OPCODE_INCR = 0x01, + HOST1X_OPCODE_NONINCR = 0x02, + HOST1X_OPCODE_MASK = 0x03, + HOST1X_OPCODE_IMM = 0x04, + HOST1X_OPCODE_RESTART = 0x05, + HOST1X_OPCODE_GATHER = 0x06, + HOST1X_OPCODE_SETSTRMID = 0x07, + HOST1X_OPCODE_SETAPPID = 0x08, + HOST1X_OPCODE_SETPYLD = 0x09, + HOST1X_OPCODE_INCR_W = 0x0a, + HOST1X_OPCODE_NONINCR_W = 0x0b, + HOST1X_OPCODE_GATHER_W = 0x0c, + HOST1X_OPCODE_RESTART_W = 0x0d, + HOST1X_OPCODE_EXTEND = 0x0e, +}; + +int tegra_drm_fw_validate(struct tegra_drm_client *client, u32 *data, u32 start, + u32 words, struct tegra_drm_submit_data *submit, + u32 *job_class) +{ + struct tegra_drm_firewall fw = { + .submit = submit, + .client = client, + .data = data, + .pos = start, + .end = start+words, + .class = *job_class, + }; + bool payload_valid = false; + u32 payload; + int err; + + while (fw.pos != fw.end) { + u32 word, opcode, offset, count, mask, class; + + err = fw_next(&fw, &word); + if (err) + return err; + + opcode = (word & 0xf0000000) >> 28; + + switch (opcode) { + case HOST1X_OPCODE_SETCLASS: + offset = word >> 16 & 0xfff; + mask = word & 0x3f; + class = (word >> 6) & 0x3ff; + err = fw_check_class(&fw, class); + fw.class = class; + *job_class = class; + if (!err) + err = fw_check_regs_mask(&fw, offset, mask); + if (err) + dev_warn(client->base.dev, + "illegal SETCLASS(offset=0x%x, mask=0x%x, class=0x%x) at word %u", + offset, mask, class, fw.pos-1); + break; + case HOST1X_OPCODE_INCR: + offset = (word >> 16) & 0xfff; + count = word & 0xffff; + err = fw_check_regs_seq(&fw, offset, count, true); + if (err) + dev_warn(client->base.dev, + "illegal INCR(offset=0x%x, count=%u) in class 0x%x at word %u", + offset, count, fw.class, fw.pos-1); + break; + case HOST1X_OPCODE_NONINCR: + offset = (word >> 16) & 0xfff; + count = word & 0xffff; + err = fw_check_regs_seq(&fw, offset, count, false); + if (err) + dev_warn(client->base.dev, + "illegal NONINCR(offset=0x%x, count=%u) in class 0x%x at word %u", + offset, count, fw.class, fw.pos-1); + break; + case HOST1X_OPCODE_MASK: + offset = (word >> 16) & 0xfff; + mask = word & 0xffff; + err = fw_check_regs_mask(&fw, offset, mask); + if (err) + dev_warn(client->base.dev, + "illegal MASK(offset=0x%x, mask=0x%x) in class 0x%x at word %u", + offset, mask, fw.class, fw.pos-1); + break; + case HOST1X_OPCODE_IMM: + /* IMM cannot reasonably be used to write a pointer */ + offset = (word >> 16) & 0xfff; + err = fw_check_regs_imm(&fw, offset); + if (err) + dev_warn(client->base.dev, + "illegal IMM(offset=0x%x) in class 0x%x at word %u", + offset, fw.class, fw.pos-1); + break; + case HOST1X_OPCODE_SETPYLD: + payload = word & 0xffff; + payload_valid = true; + break; + case HOST1X_OPCODE_INCR_W: + if (!payload_valid) + return -EINVAL; + + offset = word & 0x3fffff; + err = fw_check_regs_seq(&fw, offset, payload, true); + if (err) + dev_warn(client->base.dev, + "illegal INCR_W(offset=0x%x) in class 0x%x at word %u", + offset, fw.class, fw.pos-1); + break; + case HOST1X_OPCODE_NONINCR_W: + if (!payload_valid) + return -EINVAL; + + offset = word & 0x3fffff; + err = fw_check_regs_seq(&fw, offset, payload, false); + if (err) + dev_warn(client->base.dev, + "illegal NONINCR(offset=0x%x) in class 0x%x at word %u", + offset, fw.class, fw.pos-1); + break; + default: + dev_warn(client->base.dev, "illegal opcode at word %u", + fw.pos-1); + return -EINVAL; + } + + if (err) + return err; + } + + return 0; +} diff --git a/drivers/gpu/drm/tegra/submit.c b/drivers/gpu/drm/tegra/submit.c index 0225706e3163..c53b7207c478 100644 --- a/drivers/gpu/drm/tegra/submit.c +++ b/drivers/gpu/drm/tegra/submit.c @@ -233,7 +233,7 @@ static int submit_get_syncpt(struct tegra_drm_context *context, struct host1x_jo static int submit_job_add_gather(struct host1x_job *job, struct tegra_drm_context *context, struct drm_tegra_submit_cmd_gather_uptr *cmd, struct gather_bo *bo, u32 *offset, - struct tegra_drm_submit_data *job_data) + struct tegra_drm_submit_data *job_data, u32 *class) { u32 next_offset; @@ -258,6 +258,12 @@ static int submit_job_add_gather(struct host1x_job *job, struct tegra_drm_contex return -EINVAL; } + if (tegra_drm_fw_validate(context->client, bo->gather_data, *offset, + cmd->words, job_data, class)) { + SUBMIT_ERR(context, "job was rejected by firewall"); + return -EINVAL; + } + host1x_job_add_gather(job, &bo->base, cmd->words, *offset * 4); *offset = next_offset; @@ -311,7 +317,7 @@ submit_create_job(struct tegra_drm_context *context, struct gather_bo *bo, if (cmd->type == DRM_TEGRA_SUBMIT_CMD_GATHER_UPTR) { err = submit_job_add_gather(job, context, &cmd->gather_uptr, bo, - &gather_offset, job_data); + &gather_offset, job_data, &class); if (err) goto free_job; } else if (cmd->type == DRM_TEGRA_SUBMIT_CMD_WAIT_SYNCPT) { diff --git a/drivers/gpu/drm/tegra/submit.h b/drivers/gpu/drm/tegra/submit.h index 0a165e9e4bda..cf6a2f0a29fc 100644 --- a/drivers/gpu/drm/tegra/submit.h +++ b/drivers/gpu/drm/tegra/submit.h @@ -14,4 +14,8 @@ struct tegra_drm_submit_data { u32 num_used_mappings; }; +int tegra_drm_fw_validate(struct tegra_drm_client *client, u32 *data, u32 start, + u32 words, struct tegra_drm_submit_data *submit, + u32 *job_class); + #endif From patchwork Fri Jul 9 19:31:46 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thierry Reding X-Patchwork-Id: 1503360 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org (client-ip=23.128.96.18; helo=vger.kernel.org; envelope-from=linux-tegra-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20161025 header.b=D6l4qCob; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by ozlabs.org (Postfix) with ESMTP id 4GM3Bz60nyz9sPf for ; Sat, 10 Jul 2021 05:29:47 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230242AbhGITcb (ORCPT ); Fri, 9 Jul 2021 15:32:31 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43868 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230191AbhGITca (ORCPT ); Fri, 9 Jul 2021 15:32:30 -0400 Received: from mail-wr1-x432.google.com (mail-wr1-x432.google.com [IPv6:2a00:1450:4864:20::432]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 12136C0613DD for ; Fri, 9 Jul 2021 12:29:46 -0700 (PDT) Received: by mail-wr1-x432.google.com with SMTP id f17so13568734wrt.6 for ; Fri, 09 Jul 2021 12:29:45 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=woytsdeMsSLEt6Q2LUUCjM1OTzgLZo6zg5hihbFLDiA=; b=D6l4qCobI+52GX7y18gix33GA/A4mXjgZoy/KsRqGv6+4RZcGjjkUwDtQ/vyf2NPXr rO/BHD5bKf87107ikpOQLXQcG2igOsQ4/Tu5y8fqAgxodx/cIELEmmTQymOSw9WGOPR5 AHk8KmoDNXSvA7P41NwGR2Sr8i8BsB5rRnBXyTmO30Uo6/8tYnZ3iT+YHc6d54MJ3jTe 7j7Ds7bfBf0SGk06amNKYmh7hBCLbLyiLd7+H5EXL/lAfOyJIKwzwR4Vl0dEuZ/bejAv 0RaKRMzvpAnBLBsnxmff3z6dUoMRpDOo5fK/8mFp5RqfsI7zhI1exDtgz3c/bD2BMHH7 btIw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=woytsdeMsSLEt6Q2LUUCjM1OTzgLZo6zg5hihbFLDiA=; b=Lu7Nzt4nD0g8tVOxoA8/xedmDfmK53PVF2NP+Rr+dWFwqRRfFJYUWSO9WmDn81qvDs tbNZJKqFvlQGwLCHUfVLSUOpwJ8Enibl+ZMaprySnu0fu/KqUZh1yxb6J6bugaB5Zpzs dYG7Q/FEGu5rqlzbCSMO5oR8dAEOLEs0+vMXXOcwFZLLfVJnnmKGnNIoIkGDgD+DqN7Q e8dEMkvq/GAFhtM+B5mJnFN+pm65ppeXrQaz2kvwnWf+ESzNI/vGlt46jR02dJ9zYnAU kwzA2IMXqKM7GjdKWuOnE3OKn8S2MyT6npsSqOr2mjQr/r2WZtcMUGk1wwoZdlXKbqfB y+Tw== X-Gm-Message-State: AOAM533++MTaGAFFkwhF3cdrHbgilFX09PcxUZOqquDaM3o7NHFCo1bJ dGORvhr8RcE0csJYF5c5Gzw= X-Google-Smtp-Source: ABdhPJwqz5ruPMQXGdLS4uAiwY75l0FJB+5rb4XzspO0b/yWlo3SGmCRxY9ed1mUsqelmHYbCsSeFA== X-Received: by 2002:a5d:4d01:: with SMTP id z1mr838536wrt.34.1625858984638; Fri, 09 Jul 2021 12:29:44 -0700 (PDT) Received: from localhost ([62.96.65.119]) by smtp.gmail.com with ESMTPSA id x17sm6256027wru.6.2021.07.09.12.29.43 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 09 Jul 2021 12:29:43 -0700 (PDT) From: Thierry Reding To: Thierry Reding Cc: Mikko Perttunen , Dmitry Osipenko , dri-devel@lists.freedesktop.org, linux-tegra@vger.kernel.org Subject: [PATCH v8 14/14] drm/tegra: Bump driver version Date: Fri, 9 Jul 2021 21:31:46 +0200 Message-Id: <20210709193146.2859516-15-thierry.reding@gmail.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20210709193146.2859516-1-thierry.reding@gmail.com> References: <20210709193146.2859516-1-thierry.reding@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-tegra@vger.kernel.org From: Mikko Perttunen Bump driver version to 1.0.0 to allow userspace to detect availability of new interfaces. Signed-off-by: Mikko Perttunen Signed-off-by: Thierry Reding --- drivers/gpu/drm/tegra/drm.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index ae9dafc32c2b..b20fd0833661 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -28,7 +28,7 @@ #define DRIVER_NAME "tegra" #define DRIVER_DESC "NVIDIA Tegra graphics" #define DRIVER_DATE "20120330" -#define DRIVER_MAJOR 0 +#define DRIVER_MAJOR 1 #define DRIVER_MINOR 0 #define DRIVER_PATCHLEVEL 0