sungmg/mutter-performance-source
1
0
Fork 0
Code Activity
mutter-performance-source/src/core/monitor.c

1187 lines
36 KiB
C
Raw Normal View History

Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
/* -*- mode: C; c-file-style: "gnu"; indent-tabs-mode: nil; -*- */
/*
* Copyright (C) 2001, 2002 Havoc Pennington
* Copyright (C) 2002, 2003 Red Hat Inc.
* Some ICCCM manager selection code derived from fvwm2,
* Copyright (C) 2001 Dominik Vogt, Matthias Clasen, and fvwm2 team
* Copyright (C) 2003 Rob Adams
* Copyright (C) 2004-2006 Elijah Newren
* Copyright (C) 2013 Red Hat Inc.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License as
* published by the Free Software Foundation; either version 2 of the
* License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
* 02111-1307, USA.
*/
#include "config.h"
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
#include <string.h>
#include <math.h>
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
#include <clutter/clutter.h>
#ifdef HAVE_RANDR
#include <X11/extensions/Xrandr.h>
#endif
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
#include <meta/main.h>
#include <meta/util.h>
DisplayConfig: add the write side of the API Implement ApplyConfiguration in terms of XRandR calls. Error checking is done before actually committing the configuration. If mutter is using one of the other monitor config backends, an error is reported and nothing happens. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-22 11:31:10 +00:00
#include <meta/errors.h>
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
#include "monitor-private.h"
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
#include "meta-dbus-xrandr.h"
DisplayConfig: add the write side of the API Implement ApplyConfiguration in terms of XRandR calls. Error checking is done before actually committing the configuration. If mutter is using one of the other monitor config backends, an error is reported and nothing happens. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-22 11:31:10 +00:00
#ifndef HAVE_WAYLAND
enum wl_output_transform {
WL_OUTPUT_TRANSFORM_NORMAL,
WL_OUTPUT_TRANSFORM_90,
WL_OUTPUT_TRANSFORM_180,
WL_OUTPUT_TRANSFORM_270,
WL_OUTPUT_TRANSFORM_FLIPPED,
WL_OUTPUT_TRANSFORM_FLIPPED_90,
WL_OUTPUT_TRANSFORM_FLIPPED_180,
WL_OUTPUT_TRANSFORM_FLIPPED_270
};
#endif
typedef enum {
META_BACKEND_DUMMY,
META_BACKEND_XRANDR,
META_BACKEND_COGL
} MetaBackend;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
struct _MetaMonitorManager
{
GObject parent_instance;
DisplayConfig: add the write side of the API Implement ApplyConfiguration in terms of XRandR calls. Error checking is done before actually committing the configuration. If mutter is using one of the other monitor config backends, an error is reported and nothing happens. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-22 11:31:10 +00:00
MetaBackend backend;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
/* XXX: this structure is very badly
packed, but I like the logical organization
of fields */
unsigned int serial;
DisplayConfig: add the write side of the API Implement ApplyConfiguration in terms of XRandR calls. Error checking is done before actually committing the configuration. If mutter is using one of the other monitor config backends, an error is reported and nothing happens. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-22 11:31:10 +00:00
int max_screen_width;
int max_screen_height;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
/* Outputs refer to physical screens,
CRTCs refer to stuff that can drive outputs
(like encoders, but less tied to the HW),
while monitor_infos refer to logical ones.
See also the comment in monitor-private.h
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
*/
MetaOutput *outputs;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
unsigned int n_outputs;
MetaMonitorMode *modes;
unsigned int n_modes;
MetaCRTC *crtcs;
unsigned int n_crtcs;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
MetaMonitorInfo *monitor_infos;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
unsigned int n_monitor_infos;
int primary_monitor_index;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
#ifdef HAVE_RANDR
Display *xdisplay;
DisplayConfig: add the write side of the API Implement ApplyConfiguration in terms of XRandR calls. Error checking is done before actually committing the configuration. If mutter is using one of the other monitor config backends, an error is reported and nothing happens. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-22 11:31:10 +00:00
XRRScreenResources *resources;
int time;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
#endif
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
int dbus_name_id;
MetaDBusDisplayConfig *skeleton;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
};
struct _MetaMonitorManagerClass
{
GObjectClass parent_class;
};
enum {
MONITORS_CHANGED,
SIGNALS_LAST
};
static int signals[SIGNALS_LAST];
G_DEFINE_TYPE (MetaMonitorManager, meta_monitor_manager, G_TYPE_OBJECT);
static void
make_dummy_monitor_config (MetaMonitorManager *manager)
{
DisplayConfig: add the write side of the API Implement ApplyConfiguration in terms of XRandR calls. Error checking is done before actually committing the configuration. If mutter is using one of the other monitor config backends, an error is reported and nothing happens. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-22 11:31:10 +00:00
manager->backend = META_BACKEND_DUMMY;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
manager->modes = g_new0 (MetaMonitorMode, 1);
manager->n_modes = 1;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
manager->modes[0].mode_id = 1;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
if (manager->xdisplay)
{
Screen *screen = ScreenOfDisplay (manager->xdisplay,
DefaultScreen (manager->xdisplay));
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
manager->modes[0].width = WidthOfScreen (screen);
manager->modes[0].height = HeightOfScreen (screen);
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
}
else
{
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
manager->modes[0].width = 1024;
manager->modes[0].height = 768;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
}
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
manager->modes[0].refresh_rate = 60.0;
manager->crtcs = g_new0 (MetaCRTC, 1);
manager->n_crtcs = 1;
manager->crtcs[0].crtc_id = 2;
manager->crtcs[0].rect.x = 0;
manager->crtcs[0].rect.y = 0;
manager->crtcs[0].rect.width = manager->modes[0].width;
manager->crtcs[0].rect.height = manager->modes[0].height;
manager->crtcs[0].current_mode = &manager->modes[0];
DisplayConfig: add the write side of the API Implement ApplyConfiguration in terms of XRandR calls. Error checking is done before actually committing the configuration. If mutter is using one of the other monitor config backends, an error is reported and nothing happens. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-22 11:31:10 +00:00
manager->crtcs[0].is_dirty = FALSE;
manager->crtcs[0].logical_monitor = NULL;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
manager->outputs = g_new0 (MetaOutput, 1);
manager->n_outputs = 1;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
manager->outputs[0].crtc = &manager->crtcs[0];
manager->outputs[0].output_id = 3;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
manager->outputs[0].name = g_strdup ("LVDS");
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
manager->outputs[0].vendor = g_strdup ("unknown");
manager->outputs[0].product = g_strdup ("unknown");
manager->outputs[0].serial = g_strdup ("");
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
manager->outputs[0].width_mm = 222;
manager->outputs[0].height_mm = 125;
manager->outputs[0].subpixel_order = COGL_SUBPIXEL_ORDER_UNKNOWN;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
manager->outputs[0].preferred_mode = &manager->modes[0];
manager->outputs[0].n_modes = 1;
manager->outputs[0].modes = g_new0 (MetaMonitorMode *, 1);
manager->outputs[0].modes[0] = &manager->modes[0];
manager->outputs[0].n_possible_crtcs = 1;
manager->outputs[0].possible_crtcs = g_new0 (MetaCRTC *, 1);
manager->outputs[0].possible_crtcs[0] = &manager->crtcs[0];
Extend the DBus XRandR protocol to expose cloning restriction Turns out that even if two outputs say that they can be controlled by a given CRTC, you can't configure them in the same CRTC unless they are marked as "possible clones" one of the other. This can further restrict the configuration options, so we need to expose this limitation in the DBus API. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 16:47:01 +00:00
manager->outputs[0].n_possible_clones = 0;
manager->outputs[0].possible_clones = g_new0 (MetaOutput *, 0);
DisplayConfig: add the write side of the API Implement ApplyConfiguration in terms of XRandR calls. Error checking is done before actually committing the configuration. If mutter is using one of the other monitor config backends, an error is reported and nothing happens. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-22 11:31:10 +00:00
manager->max_screen_width = manager->modes[0].width;
manager->max_screen_height = manager->modes[0].height;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
}
#ifdef HAVE_RANDR
static void
read_monitor_infos_from_xrandr (MetaMonitorManager *manager)
{
XRRScreenResources *resources;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
RROutput primary_output;
unsigned int i, j, k;
unsigned int n_actual_outputs;
DisplayConfig: add the write side of the API Implement ApplyConfiguration in terms of XRandR calls. Error checking is done before actually committing the configuration. If mutter is using one of the other monitor config backends, an error is reported and nothing happens. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-22 11:31:10 +00:00
int min_width, min_height;
if (manager->resources)
XRRFreeScreenResources (manager->resources);
manager->resources = NULL;
XRRGetScreenSizeRange (manager->xdisplay, DefaultRootWindow (manager->xdisplay),
&min_width,
&min_height,
&manager->max_screen_width,
&manager->max_screen_height);
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
resources = XRRGetScreenResourcesCurrent (manager->xdisplay,
DefaultRootWindow (manager->xdisplay));
if (!resources)
return make_dummy_monitor_config (manager);
DisplayConfig: add the write side of the API Implement ApplyConfiguration in terms of XRandR calls. Error checking is done before actually committing the configuration. If mutter is using one of the other monitor config backends, an error is reported and nothing happens. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-22 11:31:10 +00:00
manager->backend = META_BACKEND_XRANDR;
manager->resources = resources;
manager->time = resources->configTimestamp;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
manager->n_outputs = resources->noutput;
manager->n_crtcs = resources->ncrtc;
manager->n_modes = resources->nmode;
manager->outputs = g_new0 (MetaOutput, manager->n_outputs);
manager->modes = g_new0 (MetaMonitorMode, manager->n_modes);
manager->crtcs = g_new0 (MetaCRTC, manager->n_crtcs);
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
for (i = 0; i < (unsigned)resources->nmode; i++)
{
XRRModeInfo *xmode = &resources->modes[i];
MetaMonitorMode *mode;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
mode = &manager->modes[i];
mode->mode_id = xmode->id;
mode->width = xmode->width;
mode->height = xmode->height;
mode->refresh_rate = (xmode->dotClock /
((float)xmode->hTotal * xmode->vTotal));
}
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
for (i = 0; i < (unsigned)resources->ncrtc; i++)
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
{
XRRCrtcInfo *crtc;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
MetaCRTC *meta_crtc;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
crtc = XRRGetCrtcInfo (manager->xdisplay, resources, resources->crtcs[i]);
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
meta_crtc = &manager->crtcs[i];
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
meta_crtc->crtc_id = resources->crtcs[i];
meta_crtc->rect.x = crtc->x;
meta_crtc->rect.y = crtc->y;
meta_crtc->rect.width = crtc->width;
meta_crtc->rect.height = crtc->height;
DisplayConfig: add the write side of the API Implement ApplyConfiguration in terms of XRandR calls. Error checking is done before actually committing the configuration. If mutter is using one of the other monitor config backends, an error is reported and nothing happens. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-22 11:31:10 +00:00
meta_crtc->is_dirty = FALSE;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
for (j = 0; j < (unsigned)resources->nmode; j++)
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
{
if (resources->modes[j].id == crtc->mode)
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
{
meta_crtc->current_mode = &manager->modes[j];
break;
}
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
}
XRRFreeCrtcInfo (crtc);
}
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
primary_output = XRRGetOutputPrimary (manager->xdisplay,
DefaultRootWindow (manager->xdisplay));
n_actual_outputs = 0;
for (i = 0; i < (unsigned)resources->noutput; i++)
{
XRROutputInfo *output;
MetaOutput *meta_output;
output = XRRGetOutputInfo (manager->xdisplay, resources, resources->outputs[i]);
meta_output = &manager->outputs[n_actual_outputs];
if (output->connection != RR_Disconnected)
{
meta_output->output_id = resources->outputs[i];
meta_output->name = g_strdup (output->name);
/* FIXME: to fill useful values for these we need an EDID parser */
meta_output->vendor = g_strdup ("unknown");
meta_output->product = g_strdup ("unknown");
meta_output->serial = g_strdup ("");
meta_output->width_mm = output->mm_width;
meta_output->height_mm = output->mm_height;
meta_output->subpixel_order = COGL_SUBPIXEL_ORDER_UNKNOWN;
meta_output->n_modes = output->nmode;
meta_output->modes = g_new0 (MetaMonitorMode *, meta_output->n_modes);
for (j = 0; j < meta_output->n_modes; j++)
{
for (k = 0; k < manager->n_modes; k++)
{
if (output->modes[j] == (XID)manager->modes[k].mode_id)
{
meta_output->modes[j] = &manager->modes[k];
break;
}
}
}
meta_output->preferred_mode = meta_output->modes[0];
meta_output->n_possible_crtcs = output->ncrtc;
meta_output->possible_crtcs = g_new0 (MetaCRTC *, meta_output->n_possible_crtcs);
for (j = 0; j < (unsigned)output->ncrtc; j++)
{
for (k = 0; k < manager->n_crtcs; k++)
{
if ((XID)manager->crtcs[k].crtc_id == output->crtcs[j])
{
meta_output->possible_crtcs[j] = &manager->crtcs[k];
break;
}
}
}
meta_output->crtc = NULL;
for (j = 0; j < manager->n_crtcs; j++)
{
if ((XID)manager->crtcs[j].crtc_id == output->crtc)
{
meta_output->crtc = &manager->crtcs[j];
break;
}
}
Extend the DBus XRandR protocol to expose cloning restriction Turns out that even if two outputs say that they can be controlled by a given CRTC, you can't configure them in the same CRTC unless they are marked as "possible clones" one of the other. This can further restrict the configuration options, so we need to expose this limitation in the DBus API. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 16:47:01 +00:00
meta_output->n_possible_clones = output->nclone;
meta_output->possible_clones = g_new0 (MetaOutput *, meta_output->n_possible_clones);
/* We can build the list of clones now, because we don't have the list of outputs
yet, so temporarily set the pointers to the bare XIDs, and then we'll fix them
in a second pass
*/
for (j = 0; j < (unsigned)output->nclone; j++)
{
meta_output->possible_clones = GINT_TO_POINTER (output->clones[j]);
}
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
meta_output->is_primary = ((XID)meta_output->output_id == primary_output);
meta_output->is_presentation = FALSE;
n_actual_outputs++;
}
XRRFreeOutputInfo (output);
}
manager->n_outputs = n_actual_outputs;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
Extend the DBus XRandR protocol to expose cloning restriction Turns out that even if two outputs say that they can be controlled by a given CRTC, you can't configure them in the same CRTC unless they are marked as "possible clones" one of the other. This can further restrict the configuration options, so we need to expose this limitation in the DBus API. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 16:47:01 +00:00
/* Now fix the clones */
for (i = 0; i < manager->n_outputs; i++)
{
MetaOutput *meta_output;
meta_output = &manager->outputs[i];
for (j = 0; j < meta_output->n_possible_clones; j++)
{
RROutput clone = GPOINTER_TO_INT (meta_output->possible_clones[j]);
for (k = 0; k < manager->n_outputs; k++)
{
if (clone == (XID)manager->outputs[k].output_id)
{
meta_output->possible_clones[j] = &manager->outputs[k];
break;
}
}
}
}
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
}
#endif
/*
* meta_has_dummy_output:
*
* Returns TRUE if the only available monitor is the dummy one
* backing the ClutterStage window.
*/
static gboolean
has_dummy_output (void)
{
return FALSE;
}
static void
meta_monitor_manager_init (MetaMonitorManager *manager)
{
}
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
static gboolean
make_debug_config (MetaMonitorManager *manager)
{
const char *env;
env = g_getenv ("META_DEBUG_MULTIMONITOR");
if (env == NULL)
return FALSE;
#ifdef HAVE_RANDR
if (strcmp (env, "xrandr") == 0)
read_monitor_infos_from_xrandr (manager);
else
#endif
make_dummy_monitor_config (manager);
return TRUE;
}
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
static void
read_current_config (MetaMonitorManager *manager)
{
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
if (make_debug_config (manager))
return;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
if (has_dummy_output ())
return make_dummy_monitor_config (manager);
#ifdef HAVE_RANDR
return read_monitor_infos_from_xrandr (manager);
#endif
}
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
/*
* make_logical_config:
*
* Turn outputs and CRTCs into logical MetaMonitorInfo,
* that will be used by the core and API layer (MetaScreen
* and friends)
*/
static void
make_logical_config (MetaMonitorManager *manager)
{
GArray *monitor_infos;
unsigned int i, j;
monitor_infos = g_array_sized_new (FALSE, TRUE, sizeof (MetaMonitorInfo),
manager->n_outputs);
/* Walk the list of MetaCRTCs, and build a MetaMonitorInfo
for each of them, unless they reference a rectangle that
is already there.
*/
for (i = 0; i < manager->n_crtcs; i++)
{
MetaCRTC *crtc = &manager->crtcs[i];
/* Ignore CRTCs not in use */
if (crtc->current_mode == NULL)
continue;
for (j = 0; j < monitor_infos->len; j++)
{
MetaMonitorInfo *info = &g_array_index (monitor_infos, MetaMonitorInfo, i);
if (meta_rectangle_equal (&crtc->rect,
&info->rect))
{
crtc->logical_monitor = info;
break;
}
}
if (crtc->logical_monitor == NULL)
{
MetaMonitorInfo info;
info.number = monitor_infos->len;
info.rect = crtc->rect;
info.is_primary = FALSE;
/* This starts true because we want
is_presentation only if all outputs are
marked as such (while for primary it's enough
that any is marked)
*/
info.is_presentation = TRUE;
info.in_fullscreen = -1;
info.output_id = 0;
g_array_append_val (monitor_infos, info);
crtc->logical_monitor = &g_array_index (monitor_infos, MetaMonitorInfo,
info.number);
}
}
/* Now walk the list of outputs applying extended properties (primary
and presentation)
*/
for (i = 0; i < manager->n_outputs; i++)
{
MetaOutput *output;
MetaMonitorInfo *info;
output = &manager->outputs[i];
/* Ignore outputs that are not active */
if (output->crtc == NULL)
continue;
/* We must have a logical monitor on every CRTC at this point */
g_assert (output->crtc->logical_monitor != NULL);
info = output->crtc->logical_monitor;
info->is_primary = info->is_primary || output->is_primary;
info->is_presentation = info->is_presentation && output->is_presentation;
if (output->is_primary || info->output_id == 0)
info->output_id = output->output_id;
if (info->is_primary)
manager->primary_monitor_index = info->number;
}
manager->n_monitor_infos = monitor_infos->len;
manager->monitor_infos = (void*)g_array_free (monitor_infos, FALSE);
}
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
static MetaMonitorManager *
meta_monitor_manager_new (Display *display)
{
MetaMonitorManager *manager;
manager = g_object_new (META_TYPE_MONITOR_MANAGER, NULL);
manager->xdisplay = display;
read_current_config (manager);
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
make_logical_config (manager);
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
return manager;
}
static void
free_output_array (MetaOutput *old_outputs,
int n_old_outputs)
{
int i;
for (i = 0; i < n_old_outputs; i++)
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
{
g_free (old_outputs[i].name);
g_free (old_outputs[i].vendor);
g_free (old_outputs[i].product);
g_free (old_outputs[i].serial);
g_free (old_outputs[i].modes);
g_free (old_outputs[i].possible_crtcs);
Extend the DBus XRandR protocol to expose cloning restriction Turns out that even if two outputs say that they can be controlled by a given CRTC, you can't configure them in the same CRTC unless they are marked as "possible clones" one of the other. This can further restrict the configuration options, so we need to expose this limitation in the DBus API. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 16:47:01 +00:00
g_free (old_outputs[i].possible_clones);
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
}
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
g_free (old_outputs);
}
static void
meta_monitor_manager_finalize (GObject *object)
{
MetaMonitorManager *manager = META_MONITOR_MANAGER (object);
free_output_array (manager->outputs, manager->n_outputs);
g_free (manager->monitor_infos);
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
g_free (manager->modes);
g_free (manager->crtcs);
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
G_OBJECT_CLASS (meta_monitor_manager_parent_class)->finalize (object);
}
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
static void
meta_monitor_manager_dispose (GObject *object)
{
MetaMonitorManager *manager = META_MONITOR_MANAGER (object);
if (manager->dbus_name_id != 0)
{
g_bus_unown_name (manager->dbus_name_id);
manager->dbus_name_id = 0;
}
g_clear_object (&manager->skeleton);
G_OBJECT_CLASS (meta_monitor_manager_parent_class)->dispose (object);
}
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
static void
meta_monitor_manager_class_init (MetaMonitorManagerClass *klass)
{
GObjectClass *object_class = G_OBJECT_CLASS (klass);
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
object_class->dispose = meta_monitor_manager_dispose;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
object_class->finalize = meta_monitor_manager_finalize;
signals[MONITORS_CHANGED] =
g_signal_new ("monitors-changed",
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
0,
NULL, NULL, NULL,
G_TYPE_NONE, 0);
}
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
static const double known_diagonals[] = {
12.1,
13.3,
15.6
};
static char *
diagonal_to_str (double d)
{
unsigned int i;
for (i = 0; i < G_N_ELEMENTS (known_diagonals); i++)
{
double delta;
delta = fabs(known_diagonals[i] - d);
if (delta < 0.1)
return g_strdup_printf ("%0.1lf\"", known_diagonals[i]);
}
return g_strdup_printf ("%d\"", (int) (d + 0.5));
}
static char *
make_display_name (MetaOutput *output)
{
if (output->width_mm != -1 && output->height_mm != -1)
{
double d = sqrt (output->width_mm * output->width_mm +
output->height_mm * output->height_mm);
char *inches = diagonal_to_str (d / 25.4);
char *ret;
ret = g_strdup_printf ("%s %s", output->vendor, inches);
g_free (inches);
return ret;
}
else
{
return g_strdup (output->vendor);
}
}
static gboolean
handle_get_resources (MetaDBusDisplayConfig *skeleton,
GDBusMethodInvocation *invocation,
MetaMonitorManager *manager)
{
GVariantBuilder crtc_builder, output_builder, mode_builder;
unsigned int i, j;
g_variant_builder_init (&crtc_builder, G_VARIANT_TYPE ("a(uxiiiiiuaua{sv})"));
Extend the DBus XRandR protocol to expose cloning restriction Turns out that even if two outputs say that they can be controlled by a given CRTC, you can't configure them in the same CRTC unless they are marked as "possible clones" one of the other. This can further restrict the configuration options, so we need to expose this limitation in the DBus API. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 16:47:01 +00:00
g_variant_builder_init (&output_builder, G_VARIANT_TYPE ("a(uxiausauaua{sv})"));
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
g_variant_builder_init (&mode_builder, G_VARIANT_TYPE ("a(uxuud)"));
for (i = 0; i < manager->n_crtcs; i++)
{
MetaCRTC *crtc = &manager->crtcs[i];
GVariantBuilder transforms;
g_variant_builder_init (&transforms, G_VARIANT_TYPE ("au"));
g_variant_builder_add (&transforms, "u", 0); /* 0 = WL_OUTPUT_TRANSFORM_NORMAL */
g_variant_builder_add (&crtc_builder, "(uxiiiiiuaua{sv})",
i, /* ID */
crtc->crtc_id,
(int)crtc->rect.x,
(int)crtc->rect.y,
(int)crtc->rect.width,
(int)crtc->rect.height,
(int)(crtc->current_mode ? crtc->current_mode - manager->modes : -1),
0, /* 0 = WL_OUTPUT_TRANSFORM_NORMAL */
&transforms,
NULL /* properties */);
}
for (i = 0; i < manager->n_outputs; i++)
{
MetaOutput *output = &manager->outputs[i];
Extend the DBus XRandR protocol to expose cloning restriction Turns out that even if two outputs say that they can be controlled by a given CRTC, you can't configure them in the same CRTC unless they are marked as "possible clones" one of the other. This can further restrict the configuration options, so we need to expose this limitation in the DBus API. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 16:47:01 +00:00
GVariantBuilder crtcs, modes, clones, properties;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
g_variant_builder_init (&crtcs, G_VARIANT_TYPE ("au"));
for (j = 0; j < output->n_possible_crtcs; j++)
g_variant_builder_add (&crtcs, "u",
(unsigned)(output->possible_crtcs[j] - manager->crtcs));
g_variant_builder_init (&modes, G_VARIANT_TYPE ("au"));
for (j = 0; j < output->n_modes; j++)
g_variant_builder_add (&modes, "u",
(unsigned)(output->modes[j] - manager->modes));
Extend the DBus XRandR protocol to expose cloning restriction Turns out that even if two outputs say that they can be controlled by a given CRTC, you can't configure them in the same CRTC unless they are marked as "possible clones" one of the other. This can further restrict the configuration options, so we need to expose this limitation in the DBus API. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 16:47:01 +00:00
g_variant_builder_init (&clones, G_VARIANT_TYPE ("au"));
for (j = 0; j < output->n_possible_clones; j++)
g_variant_builder_add (&clones, "u",
(unsigned)(output->possible_clones[j] - manager->outputs));
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
g_variant_builder_init (&properties, G_VARIANT_TYPE ("a{sv}"));
g_variant_builder_add (&properties, "{sv}", "vendor",
g_variant_new_string (output->vendor));
g_variant_builder_add (&properties, "{sv}", "product",
g_variant_new_string (output->product));
g_variant_builder_add (&properties, "{sv}", "serial",
g_variant_new_string (output->serial));
g_variant_builder_add (&properties, "{sv}", "display-name",
g_variant_new_take_string (make_display_name (output)));
g_variant_builder_add (&properties, "{sv}", "primary",
g_variant_new_boolean (output->is_primary));
g_variant_builder_add (&properties, "{sv}", "presentation",
g_variant_new_boolean (output->is_presentation));
Extend the DBus XRandR protocol to expose cloning restriction Turns out that even if two outputs say that they can be controlled by a given CRTC, you can't configure them in the same CRTC unless they are marked as "possible clones" one of the other. This can further restrict the configuration options, so we need to expose this limitation in the DBus API. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 16:47:01 +00:00
g_variant_builder_add (&output_builder, "(uxiausauaua{sv})",
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
i, /* ID */
output->output_id,
(int)(output->crtc ? output->crtc - manager->crtcs : -1),
&crtcs,
output->name,
&modes,
Extend the DBus XRandR protocol to expose cloning restriction Turns out that even if two outputs say that they can be controlled by a given CRTC, you can't configure them in the same CRTC unless they are marked as "possible clones" one of the other. This can further restrict the configuration options, so we need to expose this limitation in the DBus API. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 16:47:01 +00:00
&clones,
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
&properties);
}
for (i = 0; i < manager->n_modes; i++)
{
MetaMonitorMode *mode = &manager->modes[i];
g_variant_builder_add (&mode_builder, "(uxuud)",
i, /* ID */
mode->mode_id,
mode->width,
mode->height,
(double)mode->refresh_rate);
}
meta_dbus_display_config_complete_get_resources (skeleton,
invocation,
manager->serial,
g_variant_builder_end (&crtc_builder),
g_variant_builder_end (&output_builder),
DisplayConfig: add the write side of the API Implement ApplyConfiguration in terms of XRandR calls. Error checking is done before actually committing the configuration. If mutter is using one of the other monitor config backends, an error is reported and nothing happens. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-22 11:31:10 +00:00
g_variant_builder_end (&mode_builder),
manager->max_screen_width,
manager->max_screen_height);
return TRUE;
}
static gboolean
output_can_config (MetaOutput *output,
MetaCRTC *crtc,
MetaMonitorMode *mode)
{
unsigned int i;
gboolean ok = FALSE;
for (i = 0; i < output->n_possible_crtcs && !ok; i++)
ok = output->possible_crtcs[i] == crtc;
if (!ok)
return FALSE;
if (mode == NULL)
return TRUE;
ok = FALSE;
for (i = 0; i < output->n_modes && !ok; i++)
ok = output->modes[i] == mode;
return ok;
}
static gboolean
output_can_clone (MetaOutput *output,
MetaOutput *clone)
{
unsigned int i;
gboolean ok = FALSE;
for (i = 0; i < output->n_possible_clones && !ok; i++)
ok = output->possible_clones[i] == clone;
return ok;
}
#ifdef HAVE_RANDR
static Rotation
wl_transform_to_xrandr (enum wl_output_transform transform)
{
switch (transform)
{
case WL_OUTPUT_TRANSFORM_NORMAL:
return RR_Rotate_0;
case WL_OUTPUT_TRANSFORM_90:
return RR_Rotate_90;
case WL_OUTPUT_TRANSFORM_180:
return RR_Rotate_180;
case WL_OUTPUT_TRANSFORM_270:
return RR_Rotate_270;
case WL_OUTPUT_TRANSFORM_FLIPPED:
return RR_Reflect_X | RR_Rotate_0;
case WL_OUTPUT_TRANSFORM_FLIPPED_90:
return RR_Reflect_X | RR_Rotate_90;
case WL_OUTPUT_TRANSFORM_FLIPPED_180:
return RR_Reflect_X | RR_Rotate_180;
case WL_OUTPUT_TRANSFORM_FLIPPED_270:
return RR_Reflect_X | RR_Rotate_270;
}
g_assert_not_reached ();
}
static void
apply_config_xrandr (MetaMonitorManager *manager,
GVariantIter *crtcs,
GVariantIter *outputs)
{
GVariant *nested_outputs, *properties;
guint crtc_id, output_id, transform;
int new_mode, x, y;
unsigned i;
while (g_variant_iter_loop (crtcs, "(uiiiu@aua{sv})",
&crtc_id, &new_mode, &x, &y,
&transform, &nested_outputs, NULL))
{
MetaCRTC *crtc = &manager->crtcs[crtc_id];
crtc->is_dirty = TRUE;
if (new_mode == -1)
{
XRRSetCrtcConfig (manager->xdisplay,
manager->resources,
(XID)crtc->crtc_id,
manager->time,
0, 0,
None,
RR_Rotate_0,
NULL, 0);
}
else
{
MetaMonitorMode *mode;
XID *outputs;
int i, n_outputs;
guint output_id;
Status ok;
mode = &manager->modes[new_mode];
n_outputs = g_variant_n_children (nested_outputs);
outputs = g_new (XID, n_outputs);
for (i = 0; i < n_outputs; i++)
{
g_variant_get_child (nested_outputs, i, "u", &output_id);
outputs[i] = manager->outputs[output_id].output_id;
}
meta_error_trap_push (meta_get_display ());
ok = XRRSetCrtcConfig (manager->xdisplay,
manager->resources,
(XID)crtc->crtc_id,
manager->time,
x, y,
(XID)mode->mode_id,
wl_transform_to_xrandr (transform),
outputs, n_outputs);
meta_error_trap_pop (meta_get_display ());
if (ok != Success)
meta_warning ("Configuring CRTC %d with mode %d (%d x %d @ %f) at position %d, %d and transfrom %u failed\n",
(unsigned)(crtc->crtc_id), (unsigned)(mode->mode_id),
mode->width, mode->height, (float)mode->refresh_rate, x, y, transform);
g_free (outputs);
}
}
while (g_variant_iter_loop (outputs, "(u@a{sv})",
&output_id, &properties))
{
gboolean primary;
if (g_variant_lookup (properties, "primary", "b", &primary) && primary)
{
MetaOutput *output = &manager->outputs[output_id];
XRRSetOutputPrimary (manager->xdisplay,
DefaultRootWindow (manager->xdisplay),
(XID)output->output_id);
}
}
/* Disable CRTCs not mentioned in the list */
for (i = 0; i < manager->n_crtcs; i++)
{
MetaCRTC *crtc = &manager->crtcs[i];
if (crtc->is_dirty)
continue;
if (crtc->current_mode == NULL)
continue;
XRRSetCrtcConfig (manager->xdisplay,
manager->resources,
(XID)crtc->crtc_id,
manager->time,
0, 0,
None,
RR_Rotate_0,
NULL, 0);
}
}
#endif
static gboolean
handle_apply_configuration (MetaDBusDisplayConfig *skeleton,
GDBusMethodInvocation *invocation,
guint serial,
gboolean persistent,
GVariant *crtcs,
GVariant *outputs,
MetaMonitorManager *manager)
{
GVariantIter crtc_iter, output_iter, *nested_outputs;
guint crtc_id;
int new_mode, x, y;
guint transform;
guint output_id;
if (serial != manager->serial)
{
g_dbus_method_invocation_return_error (invocation, G_DBUS_ERROR,
G_DBUS_ERROR_ACCESS_DENIED,
"The requested configuration is based on stale information");
return TRUE;
}
if (persistent)
{
g_dbus_method_invocation_return_error (invocation, G_DBUS_ERROR,
G_DBUS_ERROR_NOT_SUPPORTED,
"Persistent configuration is not yet implemented");
return TRUE;
}
if (manager->backend != META_BACKEND_XRANDR)
{
g_dbus_method_invocation_return_error (invocation, G_DBUS_ERROR,
G_DBUS_ERROR_NOT_SUPPORTED,
"Changing configuration is not supported by the backend");
return TRUE;
}
/* Validate all arguments */
g_variant_iter_init (&crtc_iter, crtcs);
while (g_variant_iter_loop (&crtc_iter, "(uiiiuaua{sv})",
&crtc_id, &new_mode, &x, &y, &transform,
&nested_outputs, NULL))
{
MetaOutput *first_output;
MetaCRTC *crtc;
MetaMonitorMode *mode;
guint output_id;
if (crtc_id >= manager->n_crtcs)
{
g_dbus_method_invocation_return_error (invocation, G_DBUS_ERROR,
G_DBUS_ERROR_INVALID_ARGS,
"Invalid CRTC id");
return TRUE;
}
crtc = &manager->crtcs[crtc_id];
if (new_mode != -1 && (new_mode < 0 || (unsigned)new_mode >= manager->n_modes))
{
g_dbus_method_invocation_return_error (invocation, G_DBUS_ERROR,
G_DBUS_ERROR_INVALID_ARGS,
"Invalid mode id");
return TRUE;
}
mode = new_mode != -1 ? &manager->modes[new_mode] : NULL;
if (mode &&
(x < 0 ||
x + mode->width > manager->max_screen_width ||
y < 0 ||
y + mode->height > manager->max_screen_height))
{
g_dbus_method_invocation_return_error (invocation, G_DBUS_ERROR,
G_DBUS_ERROR_INVALID_ARGS,
"Invalid CRTC geometry");
return TRUE;
}
if (transform < WL_OUTPUT_TRANSFORM_NORMAL ||
transform > WL_OUTPUT_TRANSFORM_FLIPPED_270)
{
g_dbus_method_invocation_return_error (invocation, G_DBUS_ERROR,
G_DBUS_ERROR_INVALID_ARGS,
"Invalid transform");
return TRUE;
}
first_output = NULL;
while (g_variant_iter_loop (nested_outputs, "u", &output_id))
{
MetaOutput *output;
if (output_id >= manager->n_outputs)
{
g_dbus_method_invocation_return_error (invocation, G_DBUS_ERROR,
G_DBUS_ERROR_INVALID_ARGS,
"Invalid output id");
return TRUE;
}
output = &manager->outputs[output_id];
if (!output_can_config (output, crtc, mode))
{
g_dbus_method_invocation_return_error (invocation, G_DBUS_ERROR,
G_DBUS_ERROR_INVALID_ARGS,
"Output cannot be assigned to this CRTC or mode");
return TRUE;
}
if (first_output)
{
if (!output_can_clone (output, first_output))
{
g_dbus_method_invocation_return_error (invocation, G_DBUS_ERROR,
G_DBUS_ERROR_INVALID_ARGS,
"Outputs cannot be cloned");
return TRUE;
}
}
else
first_output = output;
}
if (!first_output && mode)
{
g_dbus_method_invocation_return_error (invocation, G_DBUS_ERROR,
G_DBUS_ERROR_INVALID_ARGS,
"Mode specified without outputs?");
return TRUE;
}
}
g_variant_iter_init (&output_iter, outputs);
while (g_variant_iter_loop (&output_iter, "(ua{sv})", &output_id, NULL))
{
if (output_id >= manager->n_outputs)
{
g_dbus_method_invocation_return_error (invocation, G_DBUS_ERROR,
G_DBUS_ERROR_INVALID_ARGS,
"Invalid output id");
return TRUE;
}
}
g_variant_iter_init (&crtc_iter, crtcs);
g_variant_iter_init (&output_iter, outputs);
apply_config_xrandr (manager, &crtc_iter, &output_iter);
meta_dbus_display_config_complete_apply_configuration (skeleton, invocation);
return TRUE;
}
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
static void
on_bus_acquired (GDBusConnection *connection,
const char *name,
gpointer user_data)
{
MetaMonitorManager *manager = user_data;
manager->skeleton = META_DBUS_DISPLAY_CONFIG (meta_dbus_display_config_skeleton_new ());
g_signal_connect_object (manager->skeleton, "handle-get-resources",
G_CALLBACK (handle_get_resources), manager, 0);
DisplayConfig: add the write side of the API Implement ApplyConfiguration in terms of XRandR calls. Error checking is done before actually committing the configuration. If mutter is using one of the other monitor config backends, an error is reported and nothing happens. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-22 11:31:10 +00:00
g_signal_connect_object (manager->skeleton, "handle-apply-configuration",
G_CALLBACK (handle_apply_configuration), manager, 0);
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
g_dbus_interface_skeleton_export (G_DBUS_INTERFACE_SKELETON (manager->skeleton),
connection,
"/org/gnome/Mutter/DisplayConfig",
NULL);
}
static void
on_name_acquired (GDBusConnection *connection,
const char *name,
gpointer user_data)
{
meta_topic (META_DEBUG_DBUS, "Acquired name %s\n", name);
}
static void
on_name_lost (GDBusConnection *connection,
const char *name,
gpointer user_data)
{
meta_topic (META_DEBUG_DBUS, "Lost or failed to acquire name %s\n", name);
}
static void
initialize_dbus_interface (MetaMonitorManager *manager)
{
manager->dbus_name_id = g_bus_own_name (G_BUS_TYPE_SESSION,
"org.gnome.Mutter.DisplayConfig",
G_BUS_NAME_OWNER_FLAGS_ALLOW_REPLACEMENT |
(meta_get_replace_current_wm () ?
G_BUS_NAME_OWNER_FLAGS_REPLACE : 0),
on_bus_acquired,
on_name_acquired,
on_name_lost,
g_object_ref (manager),
g_object_unref);
}
static MetaMonitorManager *global_monitor_manager;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
void
meta_monitor_manager_initialize (Display *display)
{
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
global_monitor_manager = meta_monitor_manager_new (display);
initialize_dbus_interface (global_monitor_manager);
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
}
MetaMonitorManager *
meta_monitor_manager_get (void)
{
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
g_assert (global_monitor_manager != NULL);
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
return global_monitor_manager;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
}
MetaMonitorInfo *
meta_monitor_manager_get_monitor_infos (MetaMonitorManager *manager,
int *n_infos)
{
*n_infos = manager->n_monitor_infos;
return manager->monitor_infos;
}
MetaOutput *
meta_monitor_manager_get_outputs (MetaMonitorManager *manager,
int *n_outputs)
{
*n_outputs = manager->n_outputs;
return manager->outputs;
}
int
meta_monitor_manager_get_primary_index (MetaMonitorManager *manager)
{
return manager->primary_monitor_index;
}
void
meta_monitor_manager_invalidate (MetaMonitorManager *manager)
{
MetaOutput *old_outputs;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
MetaCRTC *old_crtcs;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
MetaMonitorInfo *old_monitor_infos;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
MetaMonitorMode *old_modes;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
int n_old_outputs;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
/* Save the old structures, so they stay valid during the update */
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
old_outputs = manager->outputs;
n_old_outputs = manager->n_outputs;
old_monitor_infos = manager->monitor_infos;
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
old_modes = manager->modes;
old_crtcs = manager->crtcs;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
manager->serial ++;
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
read_current_config (manager);
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
make_logical_config (manager);
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
g_signal_emit (manager, signals[MONITORS_CHANGED], 0);
g_free (old_monitor_infos);
free_output_array (old_outputs, n_old_outputs);
Introduce a new DBus interface for display configuration This new interface will be used by the control center and possibly the settings daemon to configure the screens. It is designed to resemble a simplified XRandR, while still exposing all the quirks of the hardware, so that the panel can limit the user choices appropriately. To do so, MetaMonitorMode needs to track CRTCs, outputs and modes, so the low level objects have been decoupled from the high-level MetaMonitorInfo, which is used by core and API and offers a simplified view of HW, that hides away the details of what is cloned and how. This is still not efficient as it should be, because on every HW change we drop all data structures and rebuild them from scratch (which is not expensive because there aren't many of them, but at least in the XRandR path it involves a few sync X calls) https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-19 12:39:28 +00:00
g_free (old_modes);
g_free (old_crtcs);
Rework and consolidate monitor handling in MetaScreen Consolidate all places that deal with output configuration in MetaScreen, which gets it either from XRandR or from a dummy static configuration. We still need to read the Xinerama config, even when running xwayland, because we need the indices for _NET_WM_FULLSCREEN_MONITORS, but now we do it only when needed. https://bugzilla.gnome.org/show_bug.cgi?id=705670
2013-07-18 11:09:16 +00:00
}
Copy permalink