wlroots/include/backend/drm/drm.h
Johan Malm 175c20d954 Revert removal of automatic drm reset on VT switch
...on the basis that it reduces user-space breakage with respect to
layer-shell clients disappearing on suspend/resume and on VT change.

Revert the following two commits:

- 1edd5e2  "backend/drm: Remove reset from interface"
- 0f255b4  "backend/drm: Remove automatic reset on VT switch"

This it not intended to be reverted on the `master` branch. Instead, the
next wlroots release is anticipated to handle the situation differently,
possibly by splitting some objects and leaving output-related wl_globals
which have been announced to clients. See discussions on [issue-3944].

[issue-3944]: https://gitlab.freedesktop.org/wlroots/wlroots/-/issues/3944#note_3030632

Background:

With [MR-4878] on wlroots-0.19.0, the DRM backend destroys/recreates
outputs on VT switch and in some cases on suspend/resume too. The reason
for this change was that (i) the KMS state is undefined when a VT is
switched away; and (ii) the previous outputs had issues with restoration,
particularly when the output configuration had changed whilst switched
away. This change causes two issues for users:

- Some layer-shell clients do not re-appear on output re-connection, or
  may appear on a different output. Whilst this has always been the case,
  it will now also happen in said situations. It is technically possible
  for layer-shell clients to deal with this more thoughtfully by handling
  the new-output and surface-destroy signals to achieve desired
  behaviours, but such changes take time and meanwhile Desktop Environment
  teams and other users consider this a serious regression.

- Some Gtk clients issue critical warnings on some compositors as they
  assume that at least one output is always available. This will be fixed
  in `Gtk-3.24.50` and is believed to be a harmless warning,

[MR-4878]: https://gitlab.freedesktop.org/wlroots/wlroots/-/merge_requests/4878

Testing:

1. This reversion has been tested with [conky] and [GlassyClock]. With the
reversion applied, the clients remain visible after VT switching.
2. It was also tested with labwc on tty1; and sway on tty2. No adverse
affects were observed when changing output scales on either compositor.

[conky]: https://github.com/brndnmtthws/conky
[GlassyClock]: https://github.com/tsujan/GlassyClock
2025-08-03 13:01:45 +01:00

256 lines
6.8 KiB
C

#ifndef BACKEND_DRM_DRM_H
#define BACKEND_DRM_DRM_H
#include <stdbool.h>
#include <stddef.h>
#include <stdint.h>
#include <wayland-server-core.h>
#include <wayland-util.h>
#include <wlr/backend/drm.h>
#include <wlr/backend/session.h>
#include <wlr/render/drm_format_set.h>
#include <wlr/types/wlr_output_layer.h>
#include <xf86drmMode.h>
#include "backend/drm/iface.h"
#include "backend/drm/properties.h"
#include "backend/drm/renderer.h"
struct wlr_drm_viewport {
struct wlr_fbox src_box;
struct wlr_box dst_box;
};
struct wlr_drm_plane {
uint32_t type;
uint32_t id;
/* Only initialized on multi-GPU setups */
struct wlr_drm_surface mgpu_surf;
/* Buffer submitted to the kernel, will be presented on next vblank */
struct wlr_drm_fb *queued_fb;
/* Buffer currently displayed on screen */
struct wlr_drm_fb *current_fb;
/* Viewport belonging to the last committed fb */
struct wlr_drm_viewport viewport;
struct wlr_drm_format_set formats;
struct wlr_output_cursor_size *cursor_sizes;
size_t cursor_sizes_len;
struct wlr_drm_plane_props props;
uint32_t initial_crtc_id;
struct liftoff_plane *liftoff;
struct liftoff_layer *liftoff_layer;
};
struct wlr_drm_layer {
struct wlr_output_layer *wlr;
struct liftoff_layer *liftoff;
struct wlr_addon addon; // wlr_output_layer.addons
struct wl_list link; // wlr_drm_crtc.layers
/* Buffer to be submitted to the kernel on the next page-flip */
struct wlr_drm_fb *pending_fb;
/* Buffer submitted to the kernel, will be presented on next vblank */
struct wlr_drm_fb *queued_fb;
/* Buffer currently displayed on screen */
struct wlr_drm_fb *current_fb;
// One entry per wlr_drm_backend.planes
bool *candidate_planes;
};
struct wlr_drm_crtc {
uint32_t id;
struct wlr_drm_lease *lease;
struct liftoff_output *liftoff;
struct liftoff_layer *liftoff_composition_layer;
struct wl_list layers; // wlr_drm_layer.link
// Atomic modesetting only
bool own_mode_id;
uint32_t mode_id;
uint32_t gamma_lut;
// Legacy only
int legacy_gamma_size;
struct wlr_drm_plane *primary;
struct wlr_drm_plane *cursor;
struct wlr_drm_crtc_props props;
};
struct wlr_drm_backend {
struct wlr_backend backend;
struct wlr_drm_backend *parent;
const struct wlr_drm_interface *iface;
bool addfb2_modifiers;
int fd;
char *name;
struct wlr_device *dev;
struct liftoff_device *liftoff;
size_t num_crtcs;
struct wlr_drm_crtc *crtcs;
size_t num_planes;
struct wlr_drm_plane *planes;
struct wl_event_source *drm_event;
struct wl_listener session_destroy;
struct wl_listener session_active;
struct wl_listener parent_destroy;
struct wl_listener dev_change;
struct wl_listener dev_remove;
struct wl_list fbs; // wlr_drm_fb.link
struct wl_list connectors; // wlr_drm_connector.link
struct wl_list page_flips; // wlr_drm_page_flip.link
/* Only initialized on multi-GPU setups */
struct wlr_drm_renderer mgpu_renderer;
struct wlr_session *session;
uint64_t cursor_width, cursor_height;
struct wlr_drm_format_set mgpu_formats;
bool supports_tearing_page_flips;
};
struct wlr_drm_mode {
struct wlr_output_mode wlr_mode;
drmModeModeInfo drm_mode;
};
struct wlr_drm_device_state {
bool modeset;
bool nonblock;
struct wlr_drm_connector_state *connectors;
size_t connectors_len;
};
struct wlr_drm_connector_state {
struct wlr_drm_connector *connector;
const struct wlr_output_state *base;
bool active;
drmModeModeInfo mode;
struct wlr_drm_fb *primary_fb;
struct wlr_drm_viewport primary_viewport;
struct wlr_drm_fb *cursor_fb;
struct wlr_drm_syncobj_timeline *wait_timeline;
uint64_t wait_point;
// used by atomic
uint32_t mode_id;
uint32_t gamma_lut;
uint32_t fb_damage_clips;
int primary_in_fence_fd, out_fence_fd;
bool vrr_enabled;
};
/**
* Per-page-flip tracking struct.
*
* We've asked for a state change in the kernel, and yet to receive a
* notification for its completion. Currently, the kernel only has a queue
* length of 1, and no way to modify your submissions after they're sent.
*
* However, we might have multiple in-flight page-flip events, for instance
* when performing a non-blocking commit followed by a blocking commit. In
* that case, conn will be set to NULL on the non-blocking commit to indicate
* that it's been superseded.
*/
struct wlr_drm_page_flip {
struct wl_list link; // wlr_drm_connector.page_flips
struct wlr_drm_page_flip_connector *connectors;
size_t connectors_len;
// True if DRM_MODE_PAGE_FLIP_ASYNC was set
bool async;
};
struct wlr_drm_page_flip_connector {
uint32_t crtc_id;
struct wlr_drm_connector *connector; // may be NULL
};
struct wlr_drm_connector {
struct wlr_output output; // only valid if status != DISCONNECTED
struct wlr_drm_backend *backend;
char name[24];
drmModeConnection status;
uint32_t id;
uint64_t max_bpc_bounds[2];
struct wlr_drm_lease *lease;
struct wlr_drm_crtc *crtc;
uint32_t possible_crtcs;
struct wlr_drm_connector_props props;
bool cursor_enabled;
int cursor_x, cursor_y;
int cursor_width, cursor_height;
int cursor_hotspot_x, cursor_hotspot_y;
/* Buffer to be submitted to the kernel on the next page-flip */
struct wlr_drm_fb *cursor_pending_fb;
struct wl_list link; // wlr_drm_backend.connectors
// Last committed page-flip
struct wlr_drm_page_flip *pending_page_flip;
int32_t refresh;
};
struct wlr_drm_backend *get_drm_backend_from_backend(
struct wlr_backend *wlr_backend);
bool check_drm_features(struct wlr_drm_backend *drm);
bool init_drm_resources(struct wlr_drm_backend *drm);
void finish_drm_resources(struct wlr_drm_backend *drm);
void scan_drm_connectors(struct wlr_drm_backend *state,
struct wlr_device_hotplug_event *event);
void scan_drm_leases(struct wlr_drm_backend *drm);
bool commit_drm_device(struct wlr_drm_backend *drm,
const struct wlr_backend_output_state *states, size_t states_len, bool test_only);
void restore_drm_device(struct wlr_drm_backend *drm);
int handle_drm_event(int fd, uint32_t mask, void *data);
void destroy_drm_connector(struct wlr_drm_connector *conn);
bool drm_connector_is_cursor_visible(struct wlr_drm_connector *conn);
size_t drm_crtc_get_gamma_lut_size(struct wlr_drm_backend *drm,
struct wlr_drm_crtc *crtc);
void drm_lease_destroy(struct wlr_drm_lease *lease);
void drm_page_flip_destroy(struct wlr_drm_page_flip *page_flip);
struct wlr_drm_layer *get_drm_layer(struct wlr_drm_backend *drm,
struct wlr_output_layer *layer);
#if __STDC_VERSION__ >= 202311L
#define wlr_drm_conn_log(conn, verb, fmt, ...) \
wlr_log(verb, "connector %s: " fmt, conn->name __VA_OPT__(,) __VA_ARGS__)
#define wlr_drm_conn_log_errno(conn, verb, fmt, ...) \
wlr_log_errno(verb, "connector %s: " fmt, conn->name __VA_OPT__(,) __VA_ARGS__)
#else
#define wlr_drm_conn_log(conn, verb, fmt, ...) \
wlr_log(verb, "connector %s: " fmt, conn->name, ##__VA_ARGS__)
#define wlr_drm_conn_log_errno(conn, verb, fmt, ...) \
wlr_log_errno(verb, "connector %s: " fmt, conn->name, ##__VA_ARGS__)
#endif
#endif