mirror/godot
2
0
Fork 0
mirror of https://github.com/godotengine/godot.git synced 2024-12-09 10:09:20 +08:00
Code Issues Packages Projects Releases Wiki Activity
d6c0509b5a
godot/servers/physics_server_3d_wrap_mt.h

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

419 lines
15 KiB
C++
Raw Normal View History

Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
/**************************************************************************/
/* physics_server_3d_wrap_mt.h */
/**************************************************************************/
/* This file is part of: */
/* GODOT ENGINE */
/* https://godotengine.org */
/**************************************************************************/
/* Copyright (c) 2014-present Godot Engine contributors (see AUTHORS.md). */
/* Copyright (c) 2007-2014 Juan Linietsky, Ariel Manzur. */
/* */
/* Permission is hereby granted, free of charge, to any person obtaining */
/* a copy of this software and associated documentation files (the */
/* "Software"), to deal in the Software without restriction, including */
/* without limitation the rights to use, copy, modify, merge, publish, */
/* distribute, sublicense, and/or sell copies of the Software, and to */
/* permit persons to whom the Software is furnished to do so, subject to */
/* the following conditions: */
/* */
/* The above copyright notice and this permission notice shall be */
/* included in all copies or substantial portions of the Software. */
/* */
/* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, */
/* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF */
/* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. */
/* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY */
/* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, */
/* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE */
/* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */
/**************************************************************************/
Rename Godot Physics classes from *SW to Godot* Also moved MT physics server wrappers to the main servers folder, since they don't have to be implementation specific.
2021-10-19 03:24:30 +08:00
#ifndef PHYSICS_SERVER_3D_WRAP_MT_H
#define PHYSICS_SERVER_3D_WRAP_MT_H
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
#include "core/config/project_settings.h"
Use WorkerThreadPool for Server threads * Servers now use WorkerThreadPool for background computation. * This helps keep the number of threads used fixed at all times. * It also ensures everything works on HTML5 with threads. * And makes it easier to support disabling threads for also HTML5. CommandQueueMT now syncs with the servers via the WorkerThreadPool yielding mechanism, which makes its classic main sync semaphore superfluous. Also, some warnings about calls that kill performance when using threaded rendering are removed because there's a mechanism that warns about that in a more general fashion. Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
2024-04-04 16:18:08 +08:00
#include "core/object/worker_thread_pool.h"
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
#include "core/os/thread.h"
#include "core/templates/command_queue_mt.h"
#include "servers/physics_server_3d.h"
#ifdef DEBUG_SYNC
#define SYNC_DEBUG print_line("sync on: " + String(__FUNCTION__));
#else
#define SYNC_DEBUG
#endif
Use WorkerThreadPool for Server threads * Servers now use WorkerThreadPool for background computation. * This helps keep the number of threads used fixed at all times. * It also ensures everything works on HTML5 with threads. * And makes it easier to support disabling threads for also HTML5. CommandQueueMT now syncs with the servers via the WorkerThreadPool yielding mechanism, which makes its classic main sync semaphore superfluous. Also, some warnings about calls that kill performance when using threaded rendering are removed because there's a mechanism that warns about that in a more general fashion. Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
2024-04-04 16:18:08 +08:00
#ifdef DEBUG_ENABLED
Only print physics sync warning on dev builds to avoid warning spam from using move_and_slide The check should be restored once we implement move_and_slide in a way that doesn't sync the physics thread
2024-07-13 00:57:44 +08:00
#ifdef DEV_ENABLED
Use WorkerThreadPool for Server threads * Servers now use WorkerThreadPool for background computation. * This helps keep the number of threads used fixed at all times. * It also ensures everything works on HTML5 with threads. * And makes it easier to support disabling threads for also HTML5. CommandQueueMT now syncs with the servers via the WorkerThreadPool yielding mechanism, which makes its classic main sync semaphore superfluous. Also, some warnings about calls that kill performance when using threaded rendering are removed because there's a mechanism that warns about that in a more general fashion. Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
2024-04-04 16:18:08 +08:00
#define MAIN_THREAD_SYNC_WARN WARN_PRINT("Call to " + String(__FUNCTION__) + " causing PhysicsServer3D synchronizations on every frame. This significantly affects performance.");
Only print physics sync warning on dev builds to avoid warning spam from using move_and_slide The check should be restored once we implement move_and_slide in a way that doesn't sync the physics thread
2024-07-13 00:57:44 +08:00
#else
#define MAIN_THREAD_SYNC_WARN
#endif
Use WorkerThreadPool for Server threads * Servers now use WorkerThreadPool for background computation. * This helps keep the number of threads used fixed at all times. * It also ensures everything works on HTML5 with threads. * And makes it easier to support disabling threads for also HTML5. CommandQueueMT now syncs with the servers via the WorkerThreadPool yielding mechanism, which makes its classic main sync semaphore superfluous. Also, some warnings about calls that kill performance when using threaded rendering are removed because there's a mechanism that warns about that in a more general fashion. Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
2024-04-04 16:18:08 +08:00
#endif
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
class PhysicsServer3DWrapMT : public PhysicsServer3D {
Use WorkerThreadPool for Server threads * Servers now use WorkerThreadPool for background computation. * This helps keep the number of threads used fixed at all times. * It also ensures everything works on HTML5 with threads. * And makes it easier to support disabling threads for also HTML5. CommandQueueMT now syncs with the servers via the WorkerThreadPool yielding mechanism, which makes its classic main sync semaphore superfluous. Also, some warnings about calls that kill performance when using threaded rendering are removed because there's a mechanism that warns about that in a more general fashion. Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
2024-04-04 16:18:08 +08:00
mutable PhysicsServer3D *physics_server_3d = nullptr;
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
mutable CommandQueueMT command_queue;
Use WorkerThreadPool for Server threads * Servers now use WorkerThreadPool for background computation. * This helps keep the number of threads used fixed at all times. * It also ensures everything works on HTML5 with threads. * And makes it easier to support disabling threads for also HTML5. CommandQueueMT now syncs with the servers via the WorkerThreadPool yielding mechanism, which makes its classic main sync semaphore superfluous. Also, some warnings about calls that kill performance when using threaded rendering are removed because there's a mechanism that warns about that in a more general fashion. Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
2024-04-04 16:18:08 +08:00
Thread::ID server_thread = Thread::UNASSIGNED_ID;
WorkerThreadPool::TaskID server_task_id = WorkerThreadPool::INVALID_TASK_ID;
bool exit = false;
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
bool create_thread = false;
Apply additional fixes to servers' threading
2024-04-23 18:01:23 +08:00
void _assign_mt_ids(WorkerThreadPool::TaskID p_pump_task_id);
void _thread_exit();
void _thread_step(real_t p_delta);
void _thread_loop();
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
public:
#define ServerName PhysicsServer3D
#define ServerNameWrapMT PhysicsServer3DWrapMT
Rename variable names for some singletons
2022-04-27 11:52:53 +08:00
#define server_name physics_server_3d
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
#define WRITE_ACTION
#include "servers/server_wrap_mt_common.h"
//FUNC1RID(shape,ShapeType); todo fix
Rename WorldMarginShape to WorldBoundaryShape
2021-09-15 01:52:35 +08:00
FUNCRID(world_boundary_shape)
Rename RayShape to SeparationRayShape Makes it clearer that it's used for special cases when picking a collision shape.
2021-08-20 02:21:56 +08:00
FUNCRID(separation_ray_shape)
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNCRID(sphere_shape)
FUNCRID(box_shape)
FUNCRID(capsule_shape)
FUNCRID(cylinder_shape)
FUNCRID(convex_polygon_shape)
FUNCRID(concave_polygon_shape)
FUNCRID(heightmap_shape)
FUNCRID(custom_shape)
FUNC2(shape_set_data, RID, const Variant &);
FUNC2(shape_set_custom_solver_bias, RID, real_t);
FUNC2(shape_set_margin, RID, real_t)
FUNC1RC(real_t, shape_get_margin, RID)
FUNC1RC(ShapeType, shape_get_type, RID);
FUNC1RC(Variant, shape_get_data, RID);
FUNC1RC(real_t, shape_get_custom_solver_bias, RID);
#if 0
//these work well, but should be used from the main thread only
bool shape_collide(RID p_shape_A, const Transform &p_xform_A, const Vector3 &p_motion_A, RID p_shape_B, const Transform &p_xform_B, const Vector3 &p_motion_B, Vector3 *r_results, int p_result_max, int &r_result_count) {
Use WorkerThreadPool for Server threads * Servers now use WorkerThreadPool for background computation. * This helps keep the number of threads used fixed at all times. * It also ensures everything works on HTML5 with threads. * And makes it easier to support disabling threads for also HTML5. CommandQueueMT now syncs with the servers via the WorkerThreadPool yielding mechanism, which makes its classic main sync semaphore superfluous. Also, some warnings about calls that kill performance when using threaded rendering are removed because there's a mechanism that warns about that in a more general fashion. Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
2024-04-04 16:18:08 +08:00
ERR_FAIL_COND_V(!Thread::is_main_thread(), false);
Rename variable names for some singletons
2022-04-27 11:52:53 +08:00
return physics_server_3d->shape_collide(p_shape_A, p_xform_A, p_motion_A, p_shape_B, p_xform_B, p_motion_B, r_results, p_result_max, r_result_count);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
}
#endif
/* SPACE API */
FUNCRID(space);
FUNC2(space_set_active, RID, bool);
FUNC1RC(bool, space_is_active, RID);
FUNC3(space_set_param, RID, SpaceParameter, real_t);
FUNC2RC(real_t, space_get_param, RID, SpaceParameter);
// this function only works on physics process, errors and returns null otherwise
PhysicsDirectSpaceState3D *space_get_direct_state(RID p_space) override {
Use WorkerThreadPool for Server threads * Servers now use WorkerThreadPool for background computation. * This helps keep the number of threads used fixed at all times. * It also ensures everything works on HTML5 with threads. * And makes it easier to support disabling threads for also HTML5. CommandQueueMT now syncs with the servers via the WorkerThreadPool yielding mechanism, which makes its classic main sync semaphore superfluous. Also, some warnings about calls that kill performance when using threaded rendering are removed because there's a mechanism that warns about that in a more general fashion. Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
2024-04-04 16:18:08 +08:00
ERR_FAIL_COND_V(!Thread::is_main_thread(), nullptr);
Rename variable names for some singletons
2022-04-27 11:52:53 +08:00
return physics_server_3d->space_get_direct_state(p_space);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
}
FUNC2(space_set_debug_contacts, RID, int);
virtual Vector<Vector3> space_get_contacts(RID p_space) const override {
Use WorkerThreadPool for Server threads * Servers now use WorkerThreadPool for background computation. * This helps keep the number of threads used fixed at all times. * It also ensures everything works on HTML5 with threads. * And makes it easier to support disabling threads for also HTML5. CommandQueueMT now syncs with the servers via the WorkerThreadPool yielding mechanism, which makes its classic main sync semaphore superfluous. Also, some warnings about calls that kill performance when using threaded rendering are removed because there's a mechanism that warns about that in a more general fashion. Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
2024-04-04 16:18:08 +08:00
ERR_FAIL_COND_V(!Thread::is_main_thread(), Vector<Vector3>());
Rename variable names for some singletons
2022-04-27 11:52:53 +08:00
return physics_server_3d->space_get_contacts(p_space);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
}
virtual int space_get_contact_count(RID p_space) const override {
Use WorkerThreadPool for Server threads * Servers now use WorkerThreadPool for background computation. * This helps keep the number of threads used fixed at all times. * It also ensures everything works on HTML5 with threads. * And makes it easier to support disabling threads for also HTML5. CommandQueueMT now syncs with the servers via the WorkerThreadPool yielding mechanism, which makes its classic main sync semaphore superfluous. Also, some warnings about calls that kill performance when using threaded rendering are removed because there's a mechanism that warns about that in a more general fashion. Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
2024-04-04 16:18:08 +08:00
ERR_FAIL_COND_V(!Thread::is_main_thread(), 0);
Rename variable names for some singletons
2022-04-27 11:52:53 +08:00
return physics_server_3d->space_get_contact_count(p_space);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
}
/* AREA API */
//FUNC0RID(area);
FUNCRID(area);
FUNC2(area_set_space, RID, RID);
FUNC1RC(RID, area_get_space, RID);
Rename Transform to Transform3D in core
2020-10-17 13:08:21 +08:00
FUNC4(area_add_shape, RID, RID, const Transform3D &, bool);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC3(area_set_shape, RID, int, RID);
Rename Transform to Transform3D in core
2020-10-17 13:08:21 +08:00
FUNC3(area_set_shape_transform, RID, int, const Transform3D &);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC3(area_set_shape_disabled, RID, int, bool);
FUNC1RC(int, area_get_shape_count, RID);
FUNC2RC(RID, area_get_shape, RID, int);
Rename Transform to Transform3D in core
2020-10-17 13:08:21 +08:00
FUNC2RC(Transform3D, area_get_shape_transform, RID, int);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC2(area_remove_shape, RID, int);
FUNC1(area_clear_shapes, RID);
FUNC2(area_attach_object_instance_id, RID, ObjectID);
FUNC1RC(ObjectID, area_get_object_instance_id, RID);
FUNC3(area_set_param, RID, AreaParameter, const Variant &);
Rename Transform to Transform3D in core
2020-10-17 13:08:21 +08:00
FUNC2(area_set_transform, RID, const Transform3D &);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC2RC(Variant, area_get_param, RID, AreaParameter);
Rename Transform to Transform3D in core
2020-10-17 13:08:21 +08:00
FUNC1RC(Transform3D, area_get_transform, RID);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC2(area_set_collision_layer, RID, uint32_t);
Add `area_get_collision_layer` and `area_get_collision_mask`
2022-09-23 14:06:29 +08:00
FUNC1RC(uint32_t, area_get_collision_layer, RID);
FUNC2(area_set_collision_mask, RID, uint32_t);
FUNC1RC(uint32_t, area_get_collision_mask, RID);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC2(area_set_monitorable, RID, bool);
FUNC2(area_set_ray_pickable, RID, bool);
Use `Callable` in Area monitor callback
2021-11-01 05:45:58 +08:00
FUNC2(area_set_monitor_callback, RID, const Callable &);
FUNC2(area_set_area_monitor_callback, RID, const Callable &);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
/* BODY API */
//FUNC2RID(body,BodyMode,bool);
FUNCRID(body)
FUNC2(body_set_space, RID, RID);
FUNC1RC(RID, body_get_space, RID);
FUNC2(body_set_mode, RID, BodyMode);
FUNC1RC(BodyMode, body_get_mode, RID);
Rename Transform to Transform3D in core
2020-10-17 13:08:21 +08:00
FUNC4(body_add_shape, RID, RID, const Transform3D &, bool);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC3(body_set_shape, RID, int, RID);
Rename Transform to Transform3D in core
2020-10-17 13:08:21 +08:00
FUNC3(body_set_shape_transform, RID, int, const Transform3D &);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC1RC(int, body_get_shape_count, RID);
Rename Transform to Transform3D in core
2020-10-17 13:08:21 +08:00
FUNC2RC(Transform3D, body_get_shape_transform, RID, int);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC2RC(RID, body_get_shape, RID, int);
FUNC3(body_set_shape_disabled, RID, int, bool);
FUNC2(body_remove_shape, RID, int);
FUNC1(body_clear_shapes, RID);
FUNC2(body_attach_object_instance_id, RID, ObjectID);
FUNC1RC(ObjectID, body_get_object_instance_id, RID);
FUNC2(body_set_enable_continuous_collision_detection, RID, bool);
FUNC1RC(bool, body_is_continuous_collision_detection_enabled, RID);
FUNC2(body_set_collision_layer, RID, uint32_t);
FUNC1RC(uint32_t, body_get_collision_layer, RID);
FUNC2(body_set_collision_mask, RID, uint32_t);
FUNC1RC(uint32_t, body_get_collision_mask, RID);
Add collision weight to PhysicsBody for penetrations must be avoided Co-authored-by: Juan Linietsky <reduzio@gmail.com>
2022-08-11 00:45:36 +08:00
FUNC2(body_set_collision_priority, RID, real_t);
FUNC1RC(real_t, body_get_collision_priority, RID);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC2(body_set_user_flags, RID, uint32_t);
FUNC1RC(uint32_t, body_get_user_flags, RID);
Proper support for custom mass properties in 2D/3D physics bodies Changes: -Added support for custom inertia and center of mass in 3D -Added support for custom center of mass in 2D -Calculated center of mass from shapes in 2D (same as in 3D) -Fixed mass properties calculation with disabled shapes in 2D/3D -Removed first_integration which is not used in 2D and doesn't seem to make a lot of sense (prevents omit_force_integration to work during the first frame) -Support for custom inertia on different axes for RigidBody3D
2021-06-11 08:37:19 +08:00
FUNC3(body_set_param, RID, BodyParameter, const Variant &);
FUNC2RC(Variant, body_get_param, RID, BodyParameter);
FUNC1(body_reset_mass_properties, RID);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC3(body_set_state, RID, BodyState, const Variant &);
FUNC2RC(Variant, body_get_state, RID, BodyState);
FUNC2(body_apply_torque_impulse, RID, const Vector3 &);
FUNC2(body_apply_central_impulse, RID, const Vector3 &);
FUNC3(body_apply_impulse, RID, const Vector3 &, const Vector3 &);
Improve RigidDynamicBody force and torque API Makes the API for forces and impulses more flexible, easier to understand and harmonized between 2D and 3D. Rigid bodies now have 3 sets of methods for forces and impulses: -apply_impulse() for impulses (one-shot and time independent) -apply_force() for forces (time dependent) applied for the current step -add_constant_force() for forces that keeps being applied each step Also updated the documentation to clarify the different methods and parameters in rigid body nodes, body direct state and physics servers.
2021-12-08 09:09:54 +08:00
FUNC2(body_apply_central_force, RID, const Vector3 &);
FUNC3(body_apply_force, RID, const Vector3 &, const Vector3 &);
FUNC2(body_apply_torque, RID, const Vector3 &);
FUNC2(body_add_constant_central_force, RID, const Vector3 &);
FUNC3(body_add_constant_force, RID, const Vector3 &, const Vector3 &);
FUNC2(body_add_constant_torque, RID, const Vector3 &);
FUNC2(body_set_constant_force, RID, const Vector3 &);
FUNC1RC(Vector3, body_get_constant_force, RID);
FUNC2(body_set_constant_torque, RID, const Vector3 &);
FUNC1RC(Vector3, body_get_constant_torque, RID);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC2(body_set_axis_velocity, RID, const Vector3 &);
FUNC3(body_set_axis_lock, RID, BodyAxis, bool);
FUNC2RC(bool, body_is_axis_locked, RID, BodyAxis);
FUNC2(body_add_collision_exception, RID, RID);
FUNC2(body_remove_collision_exception, RID, RID);
FUNC2S(body_get_collision_exceptions, RID, List<RID> *);
FUNC2(body_set_max_contacts_reported, RID, int);
FUNC1RC(int, body_get_max_contacts_reported, RID);
FUNC2(body_set_contacts_reported_depth_threshold, RID, real_t);
FUNC1RC(real_t, body_get_contacts_reported_depth_threshold, RID);
FUNC2(body_set_omit_force_integration, RID, bool);
FUNC1RC(bool, body_is_omitting_force_integration, RID);
PhysicsServer2D and PhysicsServer3D: make body_set_state_sync_callback take a Callable Prefer Callable to a C-style callback. This is helpful for GDExtension.
2022-09-15 20:15:39 +08:00
FUNC2(body_set_state_sync_callback, RID, const Callable &);
Unexpose _direct_state_changed in PhysicsBody Removed _direct_state_changed bindings Affects 2D and 3D nodes Callbacks now use Callable Tests were changed accordingly
2021-03-30 14:22:23 +08:00
FUNC3(body_set_force_integration_callback, RID, const Callable &, const Variant &);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC2(body_set_ray_pickable, RID, bool);
Script interface improvements for test body motion -Physics servers test body motion use a class to hold parameters instead of multiple arguments to make it more readable and flexible since there are many options -Improved documentation for test body motion and kinematic collision -Removed read-only properties for body motion results (not handled in scripts, so they should be get_ methods only instead)
2021-10-01 02:28:57 +08:00
bool body_test_motion(RID p_body, const MotionParameters &p_parameters, MotionResult *r_result = nullptr) override {
Use WorkerThreadPool for Server threads * Servers now use WorkerThreadPool for background computation. * This helps keep the number of threads used fixed at all times. * It also ensures everything works on HTML5 with threads. * And makes it easier to support disabling threads for also HTML5. CommandQueueMT now syncs with the servers via the WorkerThreadPool yielding mechanism, which makes its classic main sync semaphore superfluous. Also, some warnings about calls that kill performance when using threaded rendering are removed because there's a mechanism that warns about that in a more general fashion. Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
2024-04-04 16:18:08 +08:00
ERR_FAIL_COND_V(!Thread::is_main_thread(), false);
Rename variable names for some singletons
2022-04-27 11:52:53 +08:00
return physics_server_3d->body_test_motion(p_body, p_parameters, r_result);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
}
// this function only works on physics process, errors and returns null otherwise
PhysicsDirectBodyState3D *body_get_direct_state(RID p_body) override {
Use WorkerThreadPool for Server threads * Servers now use WorkerThreadPool for background computation. * This helps keep the number of threads used fixed at all times. * It also ensures everything works on HTML5 with threads. * And makes it easier to support disabling threads for also HTML5. CommandQueueMT now syncs with the servers via the WorkerThreadPool yielding mechanism, which makes its classic main sync semaphore superfluous. Also, some warnings about calls that kill performance when using threaded rendering are removed because there's a mechanism that warns about that in a more general fashion. Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
2024-04-04 16:18:08 +08:00
ERR_FAIL_COND_V(!Thread::is_main_thread(), nullptr);
Rename variable names for some singletons
2022-04-27 11:52:53 +08:00
return physics_server_3d->body_get_direct_state(p_body);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
}
/* SOFT BODY API */
FUNCRID(soft_body)
Create GDExtension clases for PhysicsServer3D * Allows creating a GDExtension based 3D Physics Server (for Bullet, PhysX, etc. support) * Some changes on native struct binding for PhysicsServer This allows a 3D Physics server created entirely from GDExtension. Once it works, the idea is to port the 2D one to it.
2022-03-14 22:52:03 +08:00
FUNC2(soft_body_update_rendering_server, RID, PhysicsServer3DRenderingServerHandler *)
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC2(soft_body_set_space, RID, RID)
FUNC1RC(RID, soft_body_get_space, RID)
FUNC2(soft_body_set_ray_pickable, RID, bool);
FUNC2(soft_body_set_collision_layer, RID, uint32_t)
FUNC1RC(uint32_t, soft_body_get_collision_layer, RID)
FUNC2(soft_body_set_collision_mask, RID, uint32_t)
FUNC1RC(uint32_t, soft_body_get_collision_mask, RID)
FUNC2(soft_body_add_collision_exception, RID, RID)
FUNC2(soft_body_remove_collision_exception, RID, RID)
FUNC2S(soft_body_get_collision_exceptions, RID, List<RID> *)
FUNC3(soft_body_set_state, RID, BodyState, const Variant &);
FUNC2RC(Variant, soft_body_get_state, RID, BodyState);
Rename Transform to Transform3D in core
2020-10-17 13:08:21 +08:00
FUNC2(soft_body_set_transform, RID, const Transform3D &);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC2(soft_body_set_simulation_precision, RID, int);
FUNC1RC(int, soft_body_get_simulation_precision, RID);
FUNC2(soft_body_set_total_mass, RID, real_t);
FUNC1RC(real_t, soft_body_get_total_mass, RID);
FUNC2(soft_body_set_linear_stiffness, RID, real_t);
FUNC1RC(real_t, soft_body_get_linear_stiffness, RID);
FUNC2(soft_body_set_pressure_coefficient, RID, real_t);
FUNC1RC(real_t, soft_body_get_pressure_coefficient, RID);
FUNC2(soft_body_set_damping_coefficient, RID, real_t);
FUNC1RC(real_t, soft_body_get_damping_coefficient, RID);
FUNC2(soft_body_set_drag_coefficient, RID, real_t);
FUNC1RC(real_t, soft_body_get_drag_coefficient, RID);
Remove scene code in physics servers Replaced Mesh with mesh RID in Godot Physics 3D and Bullet.
2021-10-01 23:42:47 +08:00
FUNC2(soft_body_set_mesh, RID, RID);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
SoftBody support in GodotPhysics 3D - Fixed SoftBody surface update with new rendering system - Added GodotPhysics implementation for SoftBody - Added support to get SoftBody rid to interact with the physics server - Added support to get SoftBody bounds from the physics server - Removed support for unused get_vertex_position and get_point_offset from the physics server - Removed SoftBody properties that are unused in both Bullet and GodotPhysics (angular and volume stiffness, pose matching) - Added RenderingServerHandler interface to PhysicsServer3D so the physics servers don't need to reference the class from SoftBody node directly
2021-03-12 11:33:46 +08:00
FUNC1RC(AABB, soft_body_get_bounds, RID);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC3(soft_body_move_point, RID, int, const Vector3 &);
FUNC2RC(Vector3, soft_body_get_point_global_position, RID, int);
FUNC1(soft_body_remove_all_pinned_points, RID);
FUNC3(soft_body_pin_point, RID, int, bool);
FUNC2RC(bool, soft_body_is_point_pinned, RID, int);
/* JOINT API */
FUNCRID(joint)
FUNC1(joint_clear, RID)
FUNC5(joint_make_pin, RID, RID, const Vector3 &, RID, const Vector3 &)
FUNC3(pin_joint_set_param, RID, PinJointParam, real_t)
FUNC2RC(real_t, pin_joint_get_param, RID, PinJointParam)
FUNC2(pin_joint_set_local_a, RID, const Vector3 &)
FUNC1RC(Vector3, pin_joint_get_local_a, RID)
FUNC2(pin_joint_set_local_b, RID, const Vector3 &)
FUNC1RC(Vector3, pin_joint_get_local_b, RID)
Rename Transform to Transform3D in core
2020-10-17 13:08:21 +08:00
FUNC5(joint_make_hinge, RID, RID, const Transform3D &, RID, const Transform3D &)
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC7(joint_make_hinge_simple, RID, RID, const Vector3 &, const Vector3 &, RID, const Vector3 &, const Vector3 &)
FUNC3(hinge_joint_set_param, RID, HingeJointParam, real_t)
FUNC2RC(real_t, hinge_joint_get_param, RID, HingeJointParam)
FUNC3(hinge_joint_set_flag, RID, HingeJointFlag, bool)
FUNC2RC(bool, hinge_joint_get_flag, RID, HingeJointFlag)
Rename Transform to Transform3D in core
2020-10-17 13:08:21 +08:00
FUNC5(joint_make_slider, RID, RID, const Transform3D &, RID, const Transform3D &)
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC3(slider_joint_set_param, RID, SliderJointParam, real_t)
FUNC2RC(real_t, slider_joint_get_param, RID, SliderJointParam)
Rename Transform to Transform3D in core
2020-10-17 13:08:21 +08:00
FUNC5(joint_make_cone_twist, RID, RID, const Transform3D &, RID, const Transform3D &)
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC3(cone_twist_joint_set_param, RID, ConeTwistJointParam, real_t)
FUNC2RC(real_t, cone_twist_joint_get_param, RID, ConeTwistJointParam)
Rename Transform to Transform3D in core
2020-10-17 13:08:21 +08:00
FUNC5(joint_make_generic_6dof, RID, RID, const Transform3D &, RID, const Transform3D &)
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC4(generic_6dof_joint_set_param, RID, Vector3::Axis, G6DOFJointAxisParam, real_t)
FUNC3RC(real_t, generic_6dof_joint_get_param, RID, Vector3::Axis, G6DOFJointAxisParam)
FUNC4(generic_6dof_joint_set_flag, RID, Vector3::Axis, G6DOFJointAxisFlag, bool)
FUNC3RC(bool, generic_6dof_joint_get_flag, RID, Vector3::Axis, G6DOFJointAxisFlag)
FUNC1RC(JointType, joint_get_type, RID);
FUNC2(joint_set_solver_priority, RID, int);
FUNC1RC(int, joint_get_solver_priority, RID);
Create GDExtension clases for PhysicsServer3D * Allows creating a GDExtension based 3D Physics Server (for Bullet, PhysX, etc. support) * Some changes on native struct binding for PhysicsServer This allows a 3D Physics server created entirely from GDExtension. Once it works, the idea is to port the 2D one to it.
2022-03-14 22:52:03 +08:00
FUNC2(joint_disable_collisions_between_bodies, RID, bool);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
FUNC1RC(bool, joint_is_disabled_collisions_between_bodies, RID);
/* MISC */
FUNC1(free, RID);
FUNC1(set_active, bool);
virtual void init() override;
virtual void step(real_t p_step) override;
virtual void sync() override;
virtual void end_sync() override;
virtual void flush_queries() override;
virtual void finish() override;
virtual bool is_flushing_queries() const override {
Rename variable names for some singletons
2022-04-27 11:52:53 +08:00
return physics_server_3d->is_flushing_queries();
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
}
int get_process_info(ProcessInfo p_info) override {
Rename variable names for some singletons
2022-04-27 11:52:53 +08:00
return physics_server_3d->get_process_info(p_info);
Make Servers truly Thread Safe -Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread). -RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault. -Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory). -3D physics server changed to be made multithread friendly. -Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads. -Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 00:19:03 +08:00
}
PhysicsServer3DWrapMT(PhysicsServer3D *p_contained, bool p_create_thread);
~PhysicsServer3DWrapMT();
#undef ServerNameWrapMT
#undef ServerName
#undef server_name
#undef WRITE_ACTION
};
#ifdef DEBUG_SYNC
#undef DEBUG_SYNC
#endif
#undef SYNC_DEBUG
Use WorkerThreadPool for Server threads * Servers now use WorkerThreadPool for background computation. * This helps keep the number of threads used fixed at all times. * It also ensures everything works on HTML5 with threads. * And makes it easier to support disabling threads for also HTML5. CommandQueueMT now syncs with the servers via the WorkerThreadPool yielding mechanism, which makes its classic main sync semaphore superfluous. Also, some warnings about calls that kill performance when using threaded rendering are removed because there's a mechanism that warns about that in a more general fashion. Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
2024-04-04 16:18:08 +08:00
#ifdef DEBUG_ENABLED
#undef MAIN_THREAD_SYNC_WARN
#endif
Rename Godot Physics classes from *SW to Godot* Also moved MT physics server wrappers to the main servers folder, since they don't have to be implementation specific.
2021-10-19 03:24:30 +08:00
#endif // PHYSICS_SERVER_3D_WRAP_MT_H
Reference in New Issue Copy Permalink