From patchwork Fri Mar 1 10:05:16 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Daniel Vetter X-Patchwork-Id: 13578276 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 70D40C5478C for ; Fri, 1 Mar 2024 10:05:25 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 8652F10E1CA; Fri, 1 Mar 2024 10:05:24 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (1024-bit key; secure) header.d=ffwll.ch header.i=@ffwll.ch header.b="eiHgy1Us"; dkim-atps=neutral Received: from mail-lf1-f50.google.com (mail-lf1-f50.google.com [209.85.167.50]) by gabe.freedesktop.org (Postfix) with ESMTPS id A1A0B10E1CA for ; Fri, 1 Mar 2024 10:05:22 +0000 (UTC) Received: by mail-lf1-f50.google.com with SMTP id 2adb3069b0e04-50e4e36c09cso630437e87.1 for ; Fri, 01 Mar 2024 02:05:22 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=ffwll.ch; s=google; t=1709287520; x=1709892320; darn=lists.freedesktop.org; h=content-transfer-encoding:mime-version:message-id:date:subject:cc :to:from:from:to:cc:subject:date:message-id:reply-to; bh=wQGNsd9pf3Q0pJgkYTHLsj6DvJ2Trk9t60cOcnToq+A=; b=eiHgy1UssSioTv3p6eloSXFZI63u3/oio+TYT3rvvc2XW5WFPbp8ycPTTukVVVH63A V3gA17W4chnKjY0Jtpbqwb79dstKBpIj3nsiicTn+fHSXfboLaFzm1Kyz7qd5s+IaFFa D8GDObJ9nNK3Zl6toUT0Q1sbcF/lyew6xT4Lo= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1709287520; x=1709892320; h=content-transfer-encoding:mime-version:message-id:date:subject:cc :to:from:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=wQGNsd9pf3Q0pJgkYTHLsj6DvJ2Trk9t60cOcnToq+A=; b=gkB6mx6S8hBww01FD+ZLXaTlA5loze34kjv1EuQUMdenyNefOSoovtH5lMZb6zl45P 4SZ5dygbZW8OhfBbHFKgNjBB7nfRo64lz307USQmZATEsrKxcbU/4ZY2mRWRKNGDePmW /Tjfb0kBzCuKSEtG4rCXSI8QVZFjJhD+VtuAaZ2sC85yzzNOulqrm4rJEUQXO0pt+hNU 9B1h81CwouQOaGXWy8pCF+CgVBhj80XvdLFKW76fdRFtvth8w9PzxUfT2nplFHQBgxru vInG6jJ6LeIvaViL28hXw8y+JyFnbzsWcluFfWrYj4Idde8YzceidrEQa2YOHnU7tYt5 bElg== X-Gm-Message-State: AOJu0Yxh0Y3bAwy3/zUGQNx8lzvgWJ2ofggJ0T7kD8PUTzBOCg2+2T2H CwMxy01E2XU15yuFVxH5lBND6a+VK8gS3zU959EUuIRpc2QWtVzzGTOgrT5GygXTsPRWycs22v6 G X-Google-Smtp-Source: AGHT+IGrJCEmEmHGxfNfvl9FgRwQCRJTk01GBtqjkFgg71Me8R+hir23GDTbTBt3mKAYS3Sep2uEjg== X-Received: by 2002:a05:6512:514:b0:513:22d5:3ab5 with SMTP id o20-20020a056512051400b0051322d53ab5mr654363lfb.5.1709287520336; Fri, 01 Mar 2024 02:05:20 -0800 (PST) Received: from phenom.ffwll.local ([2a02:168:57f4:0:efd0:b9e5:5ae6:c2fa]) by smtp.gmail.com with ESMTPSA id h12-20020adf9ccc000000b0033e03a6b1ecsm4144127wre.18.2024.03.01.02.05.18 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 01 Mar 2024 02:05:19 -0800 (PST) From: Daniel Vetter To: DRI Development Cc: LKML , Jocelyn Falempe , Daniel Vetter , Andrew Morton , "Peter Zijlstra (Intel)" , Lukas Wunner , Petr Mladek , Steven Rostedt , John Ogness , Sergey Senozhatsky , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann , David Airlie , Daniel Vetter Subject: [RFC] drm/panic: Add drm panic locking Date: Fri, 1 Mar 2024 11:05:16 +0100 Message-ID: <20240301100516.2516297-1-daniel.vetter@ffwll.ch> X-Mailer: git-send-email 2.43.0 MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Jocelyn Falempe Rough sketch for the locking of drm panic printing code. The upshot of this approach is that we can pretty much entirely rely on the atomic commit flow, with the pair of raw_spin_lock/unlock providing any barriers we need, without having to create really big critical sections in code. This also avoids the need that drivers must explicitly update the panic handler state, which they might forget to do, or not do consistently, and then we blow up in the worst possible times. It is somewhat racy against a concurrent atomic update, and we might write into a buffer which the hardware will never display. But there's fundamentally no way to avoid that - if we do the panic state update explicitly after writing to the hardware, we might instead write to an old buffer that the user will barely ever see. Note that an rcu protected deference of plane->state would give us the the same guarantees, but it has the downside that we then need to protect the plane state freeing functions with call_rcu too. Which would very widely impact a lot of code and therefore doesn't seem worth the it compared to a raw spinlock with very tiny critical sections. Plus rcu cannot be used to protect access to peek/poke registers anyway, so we'd still need it for those cases. Peek/poke registers for vram access (or a gart pte reserved just for panic code) are also the reason I've gone with a per-device and not per-plane spinlock, since usually these things are global for the entire display. Going with per-plane locks would mean drivers for such hardware would need additional locks, which we don't want, since it deviates from the per-console takeoverlocks design. Longer term it might be useful if the panic notifiers grow a bit more structure than just the absolute bare EXPORT_SYMBOL(panic_notifier_list) - somewhat aside, why is that not EXPORT_SYMBOL_GPL ... If panic notifiers would be more like console drivers with proper register/unregister interfaces we could perhaps reuse the very fancy console lock with all it's check and takeover semantics that John Ogness is developing to fix the console_lock mess. But for the initial cut of a drm panic printing support I don't think we need that, because the critical sections are extremely small and only happen once per display refresh. So generally just 60 tiny locked sections per second, which is nothing compared to a serial console running a 115kbaud doing really slow mmio writes for each byte. So for now the raw spintrylock in drm panic notifier callback should be good enough. Another benefit of making panic notifiers more like full blown consoles (that are used in panics only) would be that we get the two stage design, where first all the safe outputs are used. And then the dangerous takeover tricks are deployed (where for display drivers we also might try to intercept any in-flight display buffer flips, which if we race and misprogram fifos and watermarks can hang the memory controller on some hw). Signed-off-by: Daniel Vetter Cc: Jocelyn Falempe Cc: Andrew Morton Cc: "Peter Zijlstra (Intel)" Cc: Lukas Wunner Cc: Petr Mladek Cc: Steven Rostedt Cc: John Ogness Cc: Sergey Senozhatsky Cc: Maarten Lankhorst Cc: Maxime Ripard Cc: Thomas Zimmermann Cc: David Airlie Cc: Daniel Vetter --- drivers/gpu/drm/drm_atomic_helper.c | 3 + include/drm/drm_mode_config.h | 10 +++ include/drm/drm_panic.h | 99 +++++++++++++++++++++++++++++ 3 files changed, 112 insertions(+) create mode 100644 include/drm/drm_panic.h diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index 40c2bd3e62e8..5a908c186037 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -38,6 +38,7 @@ #include #include #include +#include #include #include #include @@ -3086,6 +3087,7 @@ int drm_atomic_helper_swap_state(struct drm_atomic_state *state, } } + drm_panic_lock(state->dev); for_each_oldnew_plane_in_state(state, plane, old_plane_state, new_plane_state, i) { WARN_ON(plane->state != old_plane_state); @@ -3095,6 +3097,7 @@ int drm_atomic_helper_swap_state(struct drm_atomic_state *state, state->planes[i].state = old_plane_state; plane->state = new_plane_state; } + drm_panic_unlock(state->dev); for_each_oldnew_private_obj_in_state(state, obj, old_obj_state, new_obj_state, i) { WARN_ON(obj->state != old_obj_state); diff --git a/include/drm/drm_mode_config.h b/include/drm/drm_mode_config.h index 973119a9176b..92a390379e85 100644 --- a/include/drm/drm_mode_config.h +++ b/include/drm/drm_mode_config.h @@ -505,6 +505,16 @@ struct drm_mode_config { */ struct list_head plane_list; + /** + * @panic_lock: + * + * Raw spinlock used to protect critical sections of code that access + * the display hardware or modeset software, which the panic printing + * code must be protected against. See drm_panic_trylock(), + * drm_panic_lock() and drm_panic_unlock(). + */ + struct raw_spinlock panic_lock; + /** * @num_crtc: * diff --git a/include/drm/drm_panic.h b/include/drm/drm_panic.h new file mode 100644 index 000000000000..caa12f60beae --- /dev/null +++ b/include/drm/drm_panic.h @@ -0,0 +1,99 @@ +/* SPDX-License-Identifier: GPL-2.0 or MIT */ +#ifndef __DRM_PANIC_H__ +#define __DRM_PANIC_H__ + +#include +/* + * Copyright (c) 2024 Intel + */ + +/** + * drm_panic_trylock - try to enter the panic printing critical section + * @dev: struct drm_device + * + * This function must be called by any panic printing code. The panic printing + * attempt must be aborted if the trylock fails. + * + * Panic printing code can make the following assumptions while holding the + * panic lock: + * + * - Anything protected by drm_panic_lock() and drm_panic_unlock() pairs is safe + * to access. + * + * - Furthermore the panic printing code only registers in drm_dev_unregister() + * and gets removed in drm_dev_unregister(). This allows the panic code to + * safely access any state which is invariant in between these two function + * calls, like the list of planes drm_mode_config.plane_list or most of the + * struct drm_plane structure. + * + * Specifically thanks to the protection around plane updates in + * drm_atomic_helper_swap_state() the following additional guarantees hold: + * + * - It is safe to deference the drm_plane.state pointer. + * + * - Anything in struct drm_plane_state or the driver's subclass thereof which + * stays invariant after the atomic check code has finished is safe to access. + * Specifically this includes the reference counted pointers to framebuffer + * and buffer objects. + * + * - Anything set up by drm_plane_helper_funcs.fb_prepare and cleaned up + * drm_plane_helper_funcs.fb_cleanup is safe to access, as long as it stays + * invariant between these two calls. This also means that for drivers using + * dynamic buffer management the framebuffer is pinned, and therefer all + * relevant datastructures can be accessed without taking any further locks + * (which would be impossible in panic context anyway). + * + * - Importantly, software and hardware state set up by + * drm_plane_helper_funcs.begin_fb_access and + * drm_plane_helper_funcs.end_fb_access is not safe to access. + * + * Drivers must not make any assumptions about the actual state of the hardware, + * unless they explicitly protected these hardware access with drm_panic_lock() + * and drm_panic_unlock(). + * + * Returns: + * + * 0 when failing to acquire the raw spinlock, nonzero on success.:w + */ +static inline int drm_panic_trylock(struct drm_device *dev) +{ + return raw_spin_trylock(&dev->mode_config.panic_lock); +} + +/** + * drm_panic_lock - protect panic printing relevant state + * @dev: struct drm_device + * + * This function must be called to protect software and hardware state that the + * panic printing code must be able to rely on. The protected sections must be + * as small as possible. Examples include: + * + * - Access to peek/poke or other similar registers, if that is the way the + * driver prints the pixels into the scanout buffer at panic time. + * + * - Updates to pointers like drm_plane.state, allowing the panic handler to + * safely deference these. This is done in drm_atomic_helper_swap_state(). + * + * - An state that isn't invariant and that the driver must be able to access + * during panic printing. + * + * Call drm_panic_unlock() to unlock the locked spinlock. + */ +static inline void drm_panic_lock(struct drm_device *dev) +{ + return raw_spin_lock(&dev->mode_config.panic_lock); +} + +/** + * drm_panic_unlock - end of the panic printing critical section + * @dev: struct drm_device + * + * Unlocks the raw spinlock acquired by either drm_panic_lock() or + * drm_panic_trylock(). + */ +static inline void drm_panic_unlock(struct drm_device *dev) +{ + raw_spin_unlock(&dev->mode_config.panic_lock); +} + +#endif /* __DRM_PANIC_H__ */