linux/include/drm/drm_panic.h

167 lines
5.3 KiB
C
Raw Permalink Normal View History

drm/panic: Add drm panic locking 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 complexity 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). For context the actual implementation on the drm side is by Jocelyn and this patch is meant to be combined with the overall approach in v7 (v8 is a bit less flexible, which I think is the wrong direction): https://lore.kernel.org/dri-devel/20240104160301.185915-1-jfalempe@redhat.com/ Note that the locking is very much not correct there, hence this separate rfc. Starting from v10, I (Jocelyn) have included this patch in the drm_panic series, and done the corresponding changes. v2: - fix authorship, this was all my typing - some typo oopsies - link to the drm panic work by Jocelyn for context v10: - Use spinlock_irqsave/restore (John Ogness) v11: - Use macro instead of inline functions for drm_panic_lock/unlock (John Ogness) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Jocelyn Falempe <jfalempe@redhat.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org> Cc: Lukas Wunner <lukas@wunner.de> Cc: Petr Mladek <pmladek@suse.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: John Ogness <john.ogness@linutronix.de> Cc: Sergey Senozhatsky <senozhatsky@chromium.org> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Maxime Ripard <mripard@kernel.org> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: David Airlie <airlied@gmail.com> Cc: Daniel Vetter <daniel@ffwll.ch> Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240409163432.352518-2-jfalempe@redhat.com Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2024-04-09 18:30:40 +02:00
/* SPDX-License-Identifier: GPL-2.0 or MIT */
/*
* Copyright (c) 2024 Intel
* Copyright (c) 2024 Red Hat
*/
drm/panic: Add drm panic locking 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 complexity 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). For context the actual implementation on the drm side is by Jocelyn and this patch is meant to be combined with the overall approach in v7 (v8 is a bit less flexible, which I think is the wrong direction): https://lore.kernel.org/dri-devel/20240104160301.185915-1-jfalempe@redhat.com/ Note that the locking is very much not correct there, hence this separate rfc. Starting from v10, I (Jocelyn) have included this patch in the drm_panic series, and done the corresponding changes. v2: - fix authorship, this was all my typing - some typo oopsies - link to the drm panic work by Jocelyn for context v10: - Use spinlock_irqsave/restore (John Ogness) v11: - Use macro instead of inline functions for drm_panic_lock/unlock (John Ogness) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Jocelyn Falempe <jfalempe@redhat.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org> Cc: Lukas Wunner <lukas@wunner.de> Cc: Petr Mladek <pmladek@suse.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: John Ogness <john.ogness@linutronix.de> Cc: Sergey Senozhatsky <senozhatsky@chromium.org> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Maxime Ripard <mripard@kernel.org> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: David Airlie <airlied@gmail.com> Cc: Daniel Vetter <daniel@ffwll.ch> Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240409163432.352518-2-jfalempe@redhat.com Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2024-04-09 18:30:40 +02:00
#ifndef __DRM_PANIC_H__
#define __DRM_PANIC_H__
drm/panic: Add a drm panic handler This module displays a user friendly message when a kernel panic occurs. It currently doesn't contain any debug information, but that can be added later. v2 * Use get_scanout_buffer() instead of the drm client API. (Thomas Zimmermann) * Add the panic reason to the panic message (Nerdopolis) * Add an exclamation mark (Nerdopolis) v3 * Rework the drawing functions, to write the pixels line by line and to use the drm conversion helper to support other formats. (Thomas Zimmermann) v4 * Use drm_fb_r1_to_32bit for fonts (Thomas Zimmermann) * Remove the default y to DRM_PANIC config option (Thomas Zimmermann) * Add foreground/background color config option * Fix the bottom lines not painted if the framebuffer height is not a multiple of the font height. * Automatically register the device to drm_panic, if the function get_scanout_buffer exists. (Thomas Zimmermann) v5 * Change the drawing API, use drm_fb_blit_from_r1() to draw the font. * Also add drm_fb_fill() to fill area with background color. * Add draw_pixel_xy() API for drivers that can't provide a linear buffer. * Add a flush() callback for drivers that needs to synchronize the buffer. * Add a void *private field, so drivers can pass private data to draw_pixel_xy() and flush(). v6 * Fix sparse warning for panic_msg and logo. v7 * Add select DRM_KMS_HELPER for the color conversion functions. v8 * Register directly each plane to the panic notifier (Sima) * Add raw_spinlock to properly handle concurrency (Sima) * Register plane instead of device, to avoid looping through plane list, and simplify code. * Replace get_scanout_buffer() logic with drm_panic_set_buffer() (Thomas Zimmermann) * Removed the draw_pixel_xy() API, will see later if it can be added back. v9 * Revert to using get_scanout_buffer() (Sima) * Move get_scanout_buffer() and panic_flush() to the plane helper functions (Thomas Zimmermann) * Register all planes with get_scanout_buffer() to the panic notifier * Use drm_panic_lock() to protect against race (Sima) v10 * Move blit and fill functions back in drm_panic (Thomas Zimmermann). * Simplify the text drawing functions. * Use kmsg_dumper instead of panic_notifier (Sima). v12 * Use array for map and pitch in struct drm_scanout_buffer to support multi-planar format later. (Thomas Zimmermann) * Better indent struct drm_scanout_buffer declaration. (Thomas Zimmermann) Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240409163432.352518-3-jfalempe@redhat.com Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2024-04-09 18:30:41 +02:00
#include <linux/module.h>
#include <linux/types.h>
#include <linux/iosys-map.h>
drm/panic: Add drm panic locking 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 complexity 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). For context the actual implementation on the drm side is by Jocelyn and this patch is meant to be combined with the overall approach in v7 (v8 is a bit less flexible, which I think is the wrong direction): https://lore.kernel.org/dri-devel/20240104160301.185915-1-jfalempe@redhat.com/ Note that the locking is very much not correct there, hence this separate rfc. Starting from v10, I (Jocelyn) have included this patch in the drm_panic series, and done the corresponding changes. v2: - fix authorship, this was all my typing - some typo oopsies - link to the drm panic work by Jocelyn for context v10: - Use spinlock_irqsave/restore (John Ogness) v11: - Use macro instead of inline functions for drm_panic_lock/unlock (John Ogness) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Jocelyn Falempe <jfalempe@redhat.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org> Cc: Lukas Wunner <lukas@wunner.de> Cc: Petr Mladek <pmladek@suse.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: John Ogness <john.ogness@linutronix.de> Cc: Sergey Senozhatsky <senozhatsky@chromium.org> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Maxime Ripard <mripard@kernel.org> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: David Airlie <airlied@gmail.com> Cc: Daniel Vetter <daniel@ffwll.ch> Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240409163432.352518-2-jfalempe@redhat.com Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2024-04-09 18:30:40 +02:00
#include <drm/drm_device.h>
drm/panic: Add a drm panic handler This module displays a user friendly message when a kernel panic occurs. It currently doesn't contain any debug information, but that can be added later. v2 * Use get_scanout_buffer() instead of the drm client API. (Thomas Zimmermann) * Add the panic reason to the panic message (Nerdopolis) * Add an exclamation mark (Nerdopolis) v3 * Rework the drawing functions, to write the pixels line by line and to use the drm conversion helper to support other formats. (Thomas Zimmermann) v4 * Use drm_fb_r1_to_32bit for fonts (Thomas Zimmermann) * Remove the default y to DRM_PANIC config option (Thomas Zimmermann) * Add foreground/background color config option * Fix the bottom lines not painted if the framebuffer height is not a multiple of the font height. * Automatically register the device to drm_panic, if the function get_scanout_buffer exists. (Thomas Zimmermann) v5 * Change the drawing API, use drm_fb_blit_from_r1() to draw the font. * Also add drm_fb_fill() to fill area with background color. * Add draw_pixel_xy() API for drivers that can't provide a linear buffer. * Add a flush() callback for drivers that needs to synchronize the buffer. * Add a void *private field, so drivers can pass private data to draw_pixel_xy() and flush(). v6 * Fix sparse warning for panic_msg and logo. v7 * Add select DRM_KMS_HELPER for the color conversion functions. v8 * Register directly each plane to the panic notifier (Sima) * Add raw_spinlock to properly handle concurrency (Sima) * Register plane instead of device, to avoid looping through plane list, and simplify code. * Replace get_scanout_buffer() logic with drm_panic_set_buffer() (Thomas Zimmermann) * Removed the draw_pixel_xy() API, will see later if it can be added back. v9 * Revert to using get_scanout_buffer() (Sima) * Move get_scanout_buffer() and panic_flush() to the plane helper functions (Thomas Zimmermann) * Register all planes with get_scanout_buffer() to the panic notifier * Use drm_panic_lock() to protect against race (Sima) v10 * Move blit and fill functions back in drm_panic (Thomas Zimmermann). * Simplify the text drawing functions. * Use kmsg_dumper instead of panic_notifier (Sima). v12 * Use array for map and pitch in struct drm_scanout_buffer to support multi-planar format later. (Thomas Zimmermann) * Better indent struct drm_scanout_buffer declaration. (Thomas Zimmermann) Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240409163432.352518-3-jfalempe@redhat.com Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2024-04-09 18:30:41 +02:00
#include <drm/drm_fourcc.h>
drm/panic: Add drm panic locking 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 complexity 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). For context the actual implementation on the drm side is by Jocelyn and this patch is meant to be combined with the overall approach in v7 (v8 is a bit less flexible, which I think is the wrong direction): https://lore.kernel.org/dri-devel/20240104160301.185915-1-jfalempe@redhat.com/ Note that the locking is very much not correct there, hence this separate rfc. Starting from v10, I (Jocelyn) have included this patch in the drm_panic series, and done the corresponding changes. v2: - fix authorship, this was all my typing - some typo oopsies - link to the drm panic work by Jocelyn for context v10: - Use spinlock_irqsave/restore (John Ogness) v11: - Use macro instead of inline functions for drm_panic_lock/unlock (John Ogness) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Jocelyn Falempe <jfalempe@redhat.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org> Cc: Lukas Wunner <lukas@wunner.de> Cc: Petr Mladek <pmladek@suse.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: John Ogness <john.ogness@linutronix.de> Cc: Sergey Senozhatsky <senozhatsky@chromium.org> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Maxime Ripard <mripard@kernel.org> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: David Airlie <airlied@gmail.com> Cc: Daniel Vetter <daniel@ffwll.ch> Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240409163432.352518-2-jfalempe@redhat.com Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2024-04-09 18:30:40 +02:00
drm/panic: Add a drm panic handler This module displays a user friendly message when a kernel panic occurs. It currently doesn't contain any debug information, but that can be added later. v2 * Use get_scanout_buffer() instead of the drm client API. (Thomas Zimmermann) * Add the panic reason to the panic message (Nerdopolis) * Add an exclamation mark (Nerdopolis) v3 * Rework the drawing functions, to write the pixels line by line and to use the drm conversion helper to support other formats. (Thomas Zimmermann) v4 * Use drm_fb_r1_to_32bit for fonts (Thomas Zimmermann) * Remove the default y to DRM_PANIC config option (Thomas Zimmermann) * Add foreground/background color config option * Fix the bottom lines not painted if the framebuffer height is not a multiple of the font height. * Automatically register the device to drm_panic, if the function get_scanout_buffer exists. (Thomas Zimmermann) v5 * Change the drawing API, use drm_fb_blit_from_r1() to draw the font. * Also add drm_fb_fill() to fill area with background color. * Add draw_pixel_xy() API for drivers that can't provide a linear buffer. * Add a flush() callback for drivers that needs to synchronize the buffer. * Add a void *private field, so drivers can pass private data to draw_pixel_xy() and flush(). v6 * Fix sparse warning for panic_msg and logo. v7 * Add select DRM_KMS_HELPER for the color conversion functions. v8 * Register directly each plane to the panic notifier (Sima) * Add raw_spinlock to properly handle concurrency (Sima) * Register plane instead of device, to avoid looping through plane list, and simplify code. * Replace get_scanout_buffer() logic with drm_panic_set_buffer() (Thomas Zimmermann) * Removed the draw_pixel_xy() API, will see later if it can be added back. v9 * Revert to using get_scanout_buffer() (Sima) * Move get_scanout_buffer() and panic_flush() to the plane helper functions (Thomas Zimmermann) * Register all planes with get_scanout_buffer() to the panic notifier * Use drm_panic_lock() to protect against race (Sima) v10 * Move blit and fill functions back in drm_panic (Thomas Zimmermann). * Simplify the text drawing functions. * Use kmsg_dumper instead of panic_notifier (Sima). v12 * Use array for map and pitch in struct drm_scanout_buffer to support multi-planar format later. (Thomas Zimmermann) * Better indent struct drm_scanout_buffer declaration. (Thomas Zimmermann) Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240409163432.352518-3-jfalempe@redhat.com Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2024-04-09 18:30:41 +02:00
/**
* struct drm_scanout_buffer - DRM scanout buffer
*
* This structure holds the information necessary for drm_panic to draw the
* panic screen, and display it.
*/
struct drm_scanout_buffer {
/**
* @format:
*
* drm format of the scanout buffer.
*/
const struct drm_format_info *format;
/**
* @map:
*
* Virtual address of the scanout buffer, either in memory or iomem.
* The scanout buffer should be in linear format, and can be directly
* sent to the display hardware. Tearing is not an issue for the panic
* screen.
*/
struct iosys_map map[DRM_FORMAT_MAX_PLANES];
/**
* @width: Width of the scanout buffer, in pixels.
*/
unsigned int width;
/**
* @height: Height of the scanout buffer, in pixels.
*/
unsigned int height;
/**
* @pitch: Length in bytes between the start of two consecutive lines.
*/
unsigned int pitch[DRM_FORMAT_MAX_PLANES];
/**
* @set_pixel: Optional function, to set a pixel color on the
* framebuffer. It allows to handle special tiling format inside the
* driver.
*/
void (*set_pixel)(struct drm_scanout_buffer *sb, unsigned int x,
unsigned int y, u32 color);
drm/panic: Add a drm panic handler This module displays a user friendly message when a kernel panic occurs. It currently doesn't contain any debug information, but that can be added later. v2 * Use get_scanout_buffer() instead of the drm client API. (Thomas Zimmermann) * Add the panic reason to the panic message (Nerdopolis) * Add an exclamation mark (Nerdopolis) v3 * Rework the drawing functions, to write the pixels line by line and to use the drm conversion helper to support other formats. (Thomas Zimmermann) v4 * Use drm_fb_r1_to_32bit for fonts (Thomas Zimmermann) * Remove the default y to DRM_PANIC config option (Thomas Zimmermann) * Add foreground/background color config option * Fix the bottom lines not painted if the framebuffer height is not a multiple of the font height. * Automatically register the device to drm_panic, if the function get_scanout_buffer exists. (Thomas Zimmermann) v5 * Change the drawing API, use drm_fb_blit_from_r1() to draw the font. * Also add drm_fb_fill() to fill area with background color. * Add draw_pixel_xy() API for drivers that can't provide a linear buffer. * Add a flush() callback for drivers that needs to synchronize the buffer. * Add a void *private field, so drivers can pass private data to draw_pixel_xy() and flush(). v6 * Fix sparse warning for panic_msg and logo. v7 * Add select DRM_KMS_HELPER for the color conversion functions. v8 * Register directly each plane to the panic notifier (Sima) * Add raw_spinlock to properly handle concurrency (Sima) * Register plane instead of device, to avoid looping through plane list, and simplify code. * Replace get_scanout_buffer() logic with drm_panic_set_buffer() (Thomas Zimmermann) * Removed the draw_pixel_xy() API, will see later if it can be added back. v9 * Revert to using get_scanout_buffer() (Sima) * Move get_scanout_buffer() and panic_flush() to the plane helper functions (Thomas Zimmermann) * Register all planes with get_scanout_buffer() to the panic notifier * Use drm_panic_lock() to protect against race (Sima) v10 * Move blit and fill functions back in drm_panic (Thomas Zimmermann). * Simplify the text drawing functions. * Use kmsg_dumper instead of panic_notifier (Sima). v12 * Use array for map and pitch in struct drm_scanout_buffer to support multi-planar format later. (Thomas Zimmermann) * Better indent struct drm_scanout_buffer declaration. (Thomas Zimmermann) Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240409163432.352518-3-jfalempe@redhat.com Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2024-04-09 18:30:41 +02:00
};
drm/panic: Fix uninitialized spinlock acquisition with CONFIG_DRM_PANIC=n It turns out that if you happen to have a kernel config where CONFIG_DRM_PANIC is disabled and spinlock debugging is enabled, along with KMS being enabled - we'll end up trying to acquire an uninitialized spin_lock with drm_panic_lock() when we try to do a commit: rvkms rvkms.0: [drm:drm_atomic_commit] committing 0000000068d2ade1 INFO: trying to register non-static key. The code is fine but needs lockdep annotation, or maybe you didn't initialize this object before use? turning off the locking correctness validator. CPU: 4 PID: 1347 Comm: modprobe Not tainted 6.10.0-rc1Lyude-Test+ #272 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS edk2-20240524-3.fc40 05/24/2024 Call Trace: <TASK> dump_stack_lvl+0x77/0xa0 assign_lock_key+0x114/0x120 register_lock_class+0xa8/0x2c0 __lock_acquire+0x7d/0x2bd0 ? __vmap_pages_range_noflush+0x3a8/0x550 ? drm_atomic_helper_swap_state+0x2ad/0x3a0 lock_acquire+0xec/0x290 ? drm_atomic_helper_swap_state+0x2ad/0x3a0 ? lock_release+0xee/0x310 _raw_spin_lock_irqsave+0x4e/0x70 ? drm_atomic_helper_swap_state+0x2ad/0x3a0 drm_atomic_helper_swap_state+0x2ad/0x3a0 drm_atomic_helper_commit+0xb1/0x270 drm_atomic_commit+0xaf/0xe0 ? __pfx___drm_printfn_info+0x10/0x10 drm_client_modeset_commit_atomic+0x1a1/0x250 drm_client_modeset_commit_locked+0x4b/0x180 drm_client_modeset_commit+0x27/0x50 __drm_fb_helper_restore_fbdev_mode_unlocked+0x76/0x90 drm_fb_helper_set_par+0x38/0x40 fbcon_init+0x3c4/0x690 visual_init+0xc0/0x120 do_bind_con_driver+0x409/0x4c0 do_take_over_console+0x233/0x280 do_fb_registered+0x11f/0x210 fbcon_fb_registered+0x2c/0x60 register_framebuffer+0x248/0x2a0 __drm_fb_helper_initial_config_and_unlock+0x58a/0x720 drm_fbdev_generic_client_hotplug+0x6e/0xb0 drm_client_register+0x76/0xc0 _RNvXs_CsHeezP08sTT_5rvkmsNtB4_5RvkmsNtNtCs1cdwasc6FUb_6kernel8platform6Driver5probe+0xed2/0x1060 [rvkms] ? _RNvMs_NtCs1cdwasc6FUb_6kernel8platformINtB4_7AdapterNtCsHeezP08sTT_5rvkms5RvkmsE14probe_callbackBQ_+0x2b/0x70 [rvkms] ? acpi_dev_pm_attach+0x25/0x110 ? platform_probe+0x6a/0xa0 ? really_probe+0x10b/0x400 ? __driver_probe_device+0x7c/0x140 ? driver_probe_device+0x22/0x1b0 ? __device_attach_driver+0x13a/0x1c0 ? __pfx___device_attach_driver+0x10/0x10 ? bus_for_each_drv+0x114/0x170 ? __device_attach+0xd6/0x1b0 ? bus_probe_device+0x9e/0x120 ? device_add+0x288/0x4b0 ? platform_device_add+0x75/0x230 ? platform_device_register_full+0x141/0x180 ? rust_helper_platform_device_register_simple+0x85/0xb0 ? _RNvMs2_NtCs1cdwasc6FUb_6kernel8platformNtB5_6Device13create_simple+0x1d/0x60 ? _RNvXs0_CsHeezP08sTT_5rvkmsNtB5_5RvkmsNtCs1cdwasc6FUb_6kernel6Module4init+0x11e/0x160 [rvkms] ? 0xffffffffc083f000 ? init_module+0x20/0x1000 [rvkms] ? kernfs_xattr_get+0x3e/0x80 ? do_one_initcall+0x148/0x3f0 ? __lock_acquire+0x5ef/0x2bd0 ? __lock_acquire+0x5ef/0x2bd0 ? __lock_acquire+0x5ef/0x2bd0 ? put_cpu_partial+0x51/0x1d0 ? lock_acquire+0xec/0x290 ? put_cpu_partial+0x51/0x1d0 ? lock_release+0xee/0x310 ? put_cpu_partial+0x51/0x1d0 ? fs_reclaim_acquire+0x69/0xf0 ? lock_acquire+0xec/0x290 ? fs_reclaim_acquire+0x69/0xf0 ? kfree+0x22f/0x340 ? lock_release+0xee/0x310 ? kmalloc_trace_noprof+0x48/0x340 ? do_init_module+0x22/0x240 ? kmalloc_trace_noprof+0x155/0x340 ? do_init_module+0x60/0x240 ? __se_sys_finit_module+0x2e0/0x3f0 ? do_syscall_64+0xa4/0x180 ? syscall_exit_to_user_mode+0x108/0x140 ? do_syscall_64+0xb0/0x180 ? vma_end_read+0xd0/0xe0 ? do_user_addr_fault+0x309/0x640 ? clear_bhb_loop+0x45/0xa0 ? clear_bhb_loop+0x45/0xa0 ? clear_bhb_loop+0x45/0xa0 ? entry_SYSCALL_64_after_hwframe+0x76/0x7e </TASK> Fix this by stubbing these macros out when this config option isn't enabled, along with fixing the unused variable warning that introduces. Signed-off-by: Lyude Paul <lyude@redhat.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Fixes: e2a1cda3e0c7 ("drm/panic: Add drm panic locking") Cc: <stable@vger.kernel.org> # v6.10+ Link: https://patchwork.freedesktop.org/patch/msgid/20240916230103.611490-1-lyude@redhat.com
2024-09-16 19:00:08 -04:00
#ifdef CONFIG_DRM_PANIC
drm/panic: Add drm panic locking 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 complexity 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). For context the actual implementation on the drm side is by Jocelyn and this patch is meant to be combined with the overall approach in v7 (v8 is a bit less flexible, which I think is the wrong direction): https://lore.kernel.org/dri-devel/20240104160301.185915-1-jfalempe@redhat.com/ Note that the locking is very much not correct there, hence this separate rfc. Starting from v10, I (Jocelyn) have included this patch in the drm_panic series, and done the corresponding changes. v2: - fix authorship, this was all my typing - some typo oopsies - link to the drm panic work by Jocelyn for context v10: - Use spinlock_irqsave/restore (John Ogness) v11: - Use macro instead of inline functions for drm_panic_lock/unlock (John Ogness) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Jocelyn Falempe <jfalempe@redhat.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org> Cc: Lukas Wunner <lukas@wunner.de> Cc: Petr Mladek <pmladek@suse.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: John Ogness <john.ogness@linutronix.de> Cc: Sergey Senozhatsky <senozhatsky@chromium.org> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Maxime Ripard <mripard@kernel.org> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: David Airlie <airlied@gmail.com> Cc: Daniel Vetter <daniel@ffwll.ch> Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240409163432.352518-2-jfalempe@redhat.com Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2024-04-09 18:30:40 +02:00
/**
* drm_panic_trylock - try to enter the panic printing critical section
* @dev: struct drm_device
* @flags: unsigned long irq flags you need to pass to the unlock() counterpart
*
* 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().
*
* Return:
* %0 when failing to acquire the raw spinlock, nonzero on success.
*/
#define drm_panic_trylock(dev, flags) \
raw_spin_trylock_irqsave(&(dev)->mode_config.panic_lock, flags)
/**
* drm_panic_lock - protect panic printing relevant state
* @dev: struct drm_device
* @flags: unsigned long irq flags you need to pass to the unlock() counterpart
*
* 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. It uses the irqsave/irqrestore variant, and can be
* called from irq handler. 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.
*/
#define drm_panic_lock(dev, flags) \
raw_spin_lock_irqsave(&(dev)->mode_config.panic_lock, flags)
/**
* drm_panic_unlock - end of the panic printing critical section
* @dev: struct drm_device
* @flags: irq flags that were returned when acquiring the lock
*
* Unlocks the raw spinlock acquired by either drm_panic_lock() or
* drm_panic_trylock().
*/
#define drm_panic_unlock(dev, flags) \
raw_spin_unlock_irqrestore(&(dev)->mode_config.panic_lock, flags)
drm/panic: Fix uninitialized spinlock acquisition with CONFIG_DRM_PANIC=n It turns out that if you happen to have a kernel config where CONFIG_DRM_PANIC is disabled and spinlock debugging is enabled, along with KMS being enabled - we'll end up trying to acquire an uninitialized spin_lock with drm_panic_lock() when we try to do a commit: rvkms rvkms.0: [drm:drm_atomic_commit] committing 0000000068d2ade1 INFO: trying to register non-static key. The code is fine but needs lockdep annotation, or maybe you didn't initialize this object before use? turning off the locking correctness validator. CPU: 4 PID: 1347 Comm: modprobe Not tainted 6.10.0-rc1Lyude-Test+ #272 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS edk2-20240524-3.fc40 05/24/2024 Call Trace: <TASK> dump_stack_lvl+0x77/0xa0 assign_lock_key+0x114/0x120 register_lock_class+0xa8/0x2c0 __lock_acquire+0x7d/0x2bd0 ? __vmap_pages_range_noflush+0x3a8/0x550 ? drm_atomic_helper_swap_state+0x2ad/0x3a0 lock_acquire+0xec/0x290 ? drm_atomic_helper_swap_state+0x2ad/0x3a0 ? lock_release+0xee/0x310 _raw_spin_lock_irqsave+0x4e/0x70 ? drm_atomic_helper_swap_state+0x2ad/0x3a0 drm_atomic_helper_swap_state+0x2ad/0x3a0 drm_atomic_helper_commit+0xb1/0x270 drm_atomic_commit+0xaf/0xe0 ? __pfx___drm_printfn_info+0x10/0x10 drm_client_modeset_commit_atomic+0x1a1/0x250 drm_client_modeset_commit_locked+0x4b/0x180 drm_client_modeset_commit+0x27/0x50 __drm_fb_helper_restore_fbdev_mode_unlocked+0x76/0x90 drm_fb_helper_set_par+0x38/0x40 fbcon_init+0x3c4/0x690 visual_init+0xc0/0x120 do_bind_con_driver+0x409/0x4c0 do_take_over_console+0x233/0x280 do_fb_registered+0x11f/0x210 fbcon_fb_registered+0x2c/0x60 register_framebuffer+0x248/0x2a0 __drm_fb_helper_initial_config_and_unlock+0x58a/0x720 drm_fbdev_generic_client_hotplug+0x6e/0xb0 drm_client_register+0x76/0xc0 _RNvXs_CsHeezP08sTT_5rvkmsNtB4_5RvkmsNtNtCs1cdwasc6FUb_6kernel8platform6Driver5probe+0xed2/0x1060 [rvkms] ? _RNvMs_NtCs1cdwasc6FUb_6kernel8platformINtB4_7AdapterNtCsHeezP08sTT_5rvkms5RvkmsE14probe_callbackBQ_+0x2b/0x70 [rvkms] ? acpi_dev_pm_attach+0x25/0x110 ? platform_probe+0x6a/0xa0 ? really_probe+0x10b/0x400 ? __driver_probe_device+0x7c/0x140 ? driver_probe_device+0x22/0x1b0 ? __device_attach_driver+0x13a/0x1c0 ? __pfx___device_attach_driver+0x10/0x10 ? bus_for_each_drv+0x114/0x170 ? __device_attach+0xd6/0x1b0 ? bus_probe_device+0x9e/0x120 ? device_add+0x288/0x4b0 ? platform_device_add+0x75/0x230 ? platform_device_register_full+0x141/0x180 ? rust_helper_platform_device_register_simple+0x85/0xb0 ? _RNvMs2_NtCs1cdwasc6FUb_6kernel8platformNtB5_6Device13create_simple+0x1d/0x60 ? _RNvXs0_CsHeezP08sTT_5rvkmsNtB5_5RvkmsNtCs1cdwasc6FUb_6kernel6Module4init+0x11e/0x160 [rvkms] ? 0xffffffffc083f000 ? init_module+0x20/0x1000 [rvkms] ? kernfs_xattr_get+0x3e/0x80 ? do_one_initcall+0x148/0x3f0 ? __lock_acquire+0x5ef/0x2bd0 ? __lock_acquire+0x5ef/0x2bd0 ? __lock_acquire+0x5ef/0x2bd0 ? put_cpu_partial+0x51/0x1d0 ? lock_acquire+0xec/0x290 ? put_cpu_partial+0x51/0x1d0 ? lock_release+0xee/0x310 ? put_cpu_partial+0x51/0x1d0 ? fs_reclaim_acquire+0x69/0xf0 ? lock_acquire+0xec/0x290 ? fs_reclaim_acquire+0x69/0xf0 ? kfree+0x22f/0x340 ? lock_release+0xee/0x310 ? kmalloc_trace_noprof+0x48/0x340 ? do_init_module+0x22/0x240 ? kmalloc_trace_noprof+0x155/0x340 ? do_init_module+0x60/0x240 ? __se_sys_finit_module+0x2e0/0x3f0 ? do_syscall_64+0xa4/0x180 ? syscall_exit_to_user_mode+0x108/0x140 ? do_syscall_64+0xb0/0x180 ? vma_end_read+0xd0/0xe0 ? do_user_addr_fault+0x309/0x640 ? clear_bhb_loop+0x45/0xa0 ? clear_bhb_loop+0x45/0xa0 ? clear_bhb_loop+0x45/0xa0 ? entry_SYSCALL_64_after_hwframe+0x76/0x7e </TASK> Fix this by stubbing these macros out when this config option isn't enabled, along with fixing the unused variable warning that introduces. Signed-off-by: Lyude Paul <lyude@redhat.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Fixes: e2a1cda3e0c7 ("drm/panic: Add drm panic locking") Cc: <stable@vger.kernel.org> # v6.10+ Link: https://patchwork.freedesktop.org/patch/msgid/20240916230103.611490-1-lyude@redhat.com
2024-09-16 19:00:08 -04:00
#else
static inline bool drm_panic_trylock(struct drm_device *dev, unsigned long flags)
{
return true;
}
static inline void drm_panic_lock(struct drm_device *dev, unsigned long flags) {}
static inline void drm_panic_unlock(struct drm_device *dev, unsigned long flags) {}
#endif
drm/panic: Add drm panic locking 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 complexity 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). For context the actual implementation on the drm side is by Jocelyn and this patch is meant to be combined with the overall approach in v7 (v8 is a bit less flexible, which I think is the wrong direction): https://lore.kernel.org/dri-devel/20240104160301.185915-1-jfalempe@redhat.com/ Note that the locking is very much not correct there, hence this separate rfc. Starting from v10, I (Jocelyn) have included this patch in the drm_panic series, and done the corresponding changes. v2: - fix authorship, this was all my typing - some typo oopsies - link to the drm panic work by Jocelyn for context v10: - Use spinlock_irqsave/restore (John Ogness) v11: - Use macro instead of inline functions for drm_panic_lock/unlock (John Ogness) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Jocelyn Falempe <jfalempe@redhat.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org> Cc: Lukas Wunner <lukas@wunner.de> Cc: Petr Mladek <pmladek@suse.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: John Ogness <john.ogness@linutronix.de> Cc: Sergey Senozhatsky <senozhatsky@chromium.org> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Maxime Ripard <mripard@kernel.org> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: David Airlie <airlied@gmail.com> Cc: Daniel Vetter <daniel@ffwll.ch> Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240409163432.352518-2-jfalempe@redhat.com Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2024-04-09 18:30:40 +02:00
#endif /* __DRM_PANIC_H__ */