qf/quakeforge
0
0
Fork 0
mirror of https://git.code.sf.net/p/quake/quakeforge synced 2024-11-27 14:42:07 +00:00
Code Issues Projects Releases Packages Wiki Activity
quakeforge/libs/video/renderer/vulkan/vulkan_alias.c

355 lines
10 KiB
C
Raw Normal View History

[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
/*
vulkan_alias.c
Vulkan alias model pipeline
Copyright (C) 2012 Bill Currie <bill@taniwha.org>
Copyright (C) 2021 Bill Currie <bill@taniwha.org>
Author: Bill Currie <bill@taniwha.org>
Date: 2012/1/7
Date: 2021/1/26
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:
Free Software Foundation, Inc.
59 Temple Place - Suite 330
Boston, MA 02111-1307, USA
*/
#ifdef HAVE_CONFIG_H
# include "config.h"
#endif
[vulkan] Clean up a lot of unnecessary includes Too much copying of base files.
2022-03-30 06:54:07 +00:00
#include <string.h>
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
#include "QF/cvar.h"
#include "QF/va.h"
[scene] Make entity_t just an entity id for ECS This puts the hierarchy (transform) reference, animation, visibility, renderer, active, and old_origin data in separate components. There are a few bugs (crashes on grenade explosions in gl/glsl/vulkan, immediately in sw, reasons known, missing brush models in vulkan). While quake doesn't really need an ECS, the direction I want to take QF does, and it does seem to have improved memory bandwidth a little (uncertain). However, there's a lot more work to go (especially fixing the above bugs), but this seems to be a good start.
2022-10-23 01:32:09 +00:00
#include "QF/scene/component.h"
[scene] Rename libQFentity to libQFscene And start working on scene management.
2021-07-24 05:19:52 +00:00
#include "QF/scene/entity.h"
[scene] Make entity_t just an entity id for ECS This puts the hierarchy (transform) reference, animation, visibility, renderer, active, and old_origin data in separate components. There are a few bugs (crashes on grenade explosions in gl/glsl/vulkan, immediately in sw, reasons known, missing brush models in vulkan). While quake doesn't really need an ECS, the direction I want to take QF does, and it does seem to have improved memory bandwidth a little (uncertain). However, there's a lot more work to go (especially fixing the above bugs), but this seems to be a good start.
2022-10-23 01:32:09 +00:00
#include "QF/scene/scene.h"
[scene] Rename libQFentity to libQFscene And start working on scene management.
2021-07-24 05:19:52 +00:00
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
#include "QF/Vulkan/qf_alias.h"
[vulkan] Switch alias pipeline to multi-texture This gets the alias pipeline in line with the bsp pipeline, and thus everything is about as functional as it was before the rework (minus dealing with large texture sets).
2021-12-08 15:25:50 +00:00
#include "QF/Vulkan/qf_matrices.h"
[vulkan] Rename the renderpass files I plan on moving the current hard-coded render pass creation from vulkan_vid_common and this maintains the (possibly silly) naming convention).
2022-05-30 03:10:00 +00:00
#include "QF/Vulkan/qf_renderpass.h"
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
#include "QF/Vulkan/qf_texture.h"
[vulkan] Name most resources This makes debugging with renderdoc and validation messages much easier (no more "what buffer is that?").
2021-01-31 10:58:55 +00:00
#include "QF/Vulkan/debug.h"
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
#include "QF/Vulkan/device.h"
[vulkan] Add some more debug stack contexts
2021-12-02 13:48:50 +00:00
#include "QF/Vulkan/instance.h"
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
#include "r_internal.h"
#include "vid_vulkan.h"
[vulkan] Clean up alias push-constants
2021-12-17 07:46:54 +00:00
typedef struct {
mat4f_t mat;
float blend;
byte colorA[4];
byte colorB[4];
vec4f_t base_color;
vec4f_t fog;
} alias_push_constants_t;
[vulkan] Fix some 32-bit compatibility issues mostly dealing with vulkan's handles always being 64-bit, but pointers being either 32 or 64.
2021-04-01 23:48:11 +00:00
static const char * __attribute__((used)) alias_pass_names[] = {
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
"depth",
"g-buffer",
"translucent",
};
[vulkan] Fix sky depth issues There were actually several problems: translucency wasn't using or depending on the depth buffer, and the depth buffer wasn't marked as read-only in the g-buffer pass. Getting that correct seems to have given bigass1 a 0.5% boost (hard to say, could be the usual noise).
2021-03-23 02:26:24 +00:00
static QFV_Subpass subpass_map[] = {
QFV_passDepth, // QFV_aliasDepth
QFV_passGBuffer, // QFV_aliasGBuffer
QFV_passTranslucent, // QFV_aliasTranslucent
};
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
static void
emit_commands (VkCommandBuffer cmd, int pose1, int pose2,
qfv_alias_skin_t *skin,
[vulkan] Clean up alias push-constants
2021-12-17 07:46:54 +00:00
uint32_t numPC, qfv_push_constants_t *constants,
[scene] Make entity_t just an entity id for ECS This puts the hierarchy (transform) reference, animation, visibility, renderer, active, and old_origin data in separate components. There are a few bugs (crashes on grenade explosions in gl/glsl/vulkan, immediately in sw, reasons known, missing brush models in vulkan). While quake doesn't really need an ECS, the direction I want to take QF does, and it does seem to have improved memory bandwidth a little (uncertain). However, there's a lot more work to go (especially fixing the above bugs), but this seems to be a good start.
2022-10-23 01:32:09 +00:00
aliashdr_t *hdr, qfv_renderframe_t *rFrame, entity_t ent)
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
{
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
vulkan_ctx_t *ctx = rFrame->vulkan_ctx;
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
qfv_device_t *device = ctx->device;
qfv_devfuncs_t *dfunc = device->funcs;
aliasctx_t *actx = ctx->alias_context;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
__auto_type mesh = (qfv_alias_mesh_t *) ((byte *) hdr + hdr->commands);
VkDeviceSize offsets[] = {
pose1 * hdr->poseverts * sizeof (aliasvrt_t),
pose2 * hdr->poseverts * sizeof (aliasvrt_t),
0,
};
VkBuffer buffers[] = {
mesh->vertex_buffer,
mesh->vertex_buffer,
mesh->uv_buffer,
};
int bindingCount = skin ? 3 : 2;
[vulkan] Add entity labels to command queues Brush models looked a little too tricky due to the very different style of command queue, so that's left for now, but alias, iqm and sprite entities are now labeled. The labels are made up of the lower 5 hex digits of the entity address, the position, and colored by the normalized position vector. Not sure that's the best choice as it does mean the color changes as the entity moves, and can be quite subtle between nearby entities, but it still helps identify the entities in the command buffer. And, as I suspected, I've got multiple draw calls for the one ogre. Now to find out why.
2022-05-07 05:12:02 +00:00
Vulkan_BeginEntityLabel (ctx, cmd, ent);
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
dfunc->vkCmdBindVertexBuffers (cmd, 0, bindingCount, buffers, offsets);
dfunc->vkCmdBindIndexBuffer (cmd, mesh->index_buffer, 0,
VK_INDEX_TYPE_UINT32);
[vulkan] Clean up alias push-constants
2021-12-17 07:46:54 +00:00
QFV_PushConstants (device, cmd, actx->layout, numPC, constants);
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
if (skin) {
[vulkan] Abandon bindless rendering At least for now, it is more trouble than it is worth as it (or my implementation of it) breaks lavapipe and renderdoc.
2021-12-17 03:20:32 +00:00
VkDescriptorSet sets[] = {
skin->descriptor,
};
dfunc->vkCmdBindDescriptorSets (cmd, VK_PIPELINE_BIND_POINT_GRAPHICS,
actx->layout, 1, 1, sets, 0, 0);
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
}
dfunc->vkCmdDrawIndexed (cmd, 3 * hdr->mdl.numtris, 1, 0, 0, 0);
[vulkan] Add entity labels to command queues Brush models looked a little too tricky due to the very different style of command queue, so that's left for now, but alias, iqm and sprite entities are now labeled. The labels are made up of the lower 5 hex digits of the entity address, the position, and colored by the normalized position vector. Not sure that's the best choice as it does mean the color changes as the entity moves, and can be quite subtle between nearby entities, but it still helps identify the entities in the command buffer. And, as I suspected, I've got multiple draw calls for the one ogre. Now to find out why.
2022-05-07 05:12:02 +00:00
QFV_CmdEndLabel (device, cmd);
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
}
void
[scene] Make entity_t just an entity id for ECS This puts the hierarchy (transform) reference, animation, visibility, renderer, active, and old_origin data in separate components. There are a few bugs (crashes on grenade explosions in gl/glsl/vulkan, immediately in sw, reasons known, missing brush models in vulkan). While quake doesn't really need an ECS, the direction I want to take QF does, and it does seem to have improved memory bandwidth a little (uncertain). However, there's a lot more work to go (especially fixing the above bugs), but this seems to be a good start.
2022-10-23 01:32:09 +00:00
Vulkan_DrawAlias (entity_t ent, qfv_renderframe_t *rFrame)
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
{
[scene] Make entity_t just an entity id for ECS This puts the hierarchy (transform) reference, animation, visibility, renderer, active, and old_origin data in separate components. There are a few bugs (crashes on grenade explosions in gl/glsl/vulkan, immediately in sw, reasons known, missing brush models in vulkan). While quake doesn't really need an ECS, the direction I want to take QF does, and it does seem to have improved memory bandwidth a little (uncertain). However, there's a lot more work to go (especially fixing the above bugs), but this seems to be a good start.
2022-10-23 01:32:09 +00:00
transform_t transform = Entity_Transform (ent);
renderer_t *renderer = Ent_GetComponent (ent.id, scene_renderer,
r_refdef.scene->reg);
animation_t *animation = Ent_GetComponent (ent.id, scene_animation,
r_refdef.scene->reg);
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
vulkan_ctx_t *ctx = rFrame->vulkan_ctx;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
aliasctx_t *actx = ctx->alias_context;
aliasframe_t *aframe = &actx->frames.a[ctx->curFrame];
[scene] Make entity_t just an entity id for ECS This puts the hierarchy (transform) reference, animation, visibility, renderer, active, and old_origin data in separate components. There are a few bugs (crashes on grenade explosions in gl/glsl/vulkan, immediately in sw, reasons known, missing brush models in vulkan). While quake doesn't really need an ECS, the direction I want to take QF does, and it does seem to have improved memory bandwidth a little (uncertain). However, there's a lot more work to go (especially fixing the above bugs), but this seems to be a good start.
2022-10-23 01:32:09 +00:00
model_t *model = renderer->model;
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
aliashdr_t *hdr;
[vulkan] Implement team colors It's not quite as expected, but that may be due to one of msaa, the 0-15 range in the palette not being all the way to white, the color gradients being not quite linear (haven't checked yet) or some combination of the above. However, it's that what should be yellow is more green. At least the zombies are no longer white and the ogres don't look like they're wearing skeleton suits.
2021-02-02 10:53:36 +00:00
qfv_alias_skin_t *skin;
[vulkan] Clean up alias push-constants
2021-12-17 07:46:54 +00:00
alias_push_constants_t constants = {};
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
if (!(hdr = model->aliashdr)) {
hdr = Cache_Get (&model->cache);
}
[vulkan] Clean up alias push-constants
2021-12-17 07:46:54 +00:00
constants.blend = R_AliasGetLerpedFrames (animation, hdr);
qfv_push_constants_t push_constants[] = {
{ VK_SHADER_STAGE_VERTEX_BIT,
field_offset (alias_push_constants_t, mat),
[scene] Make entity_t just an entity id for ECS This puts the hierarchy (transform) reference, animation, visibility, renderer, active, and old_origin data in separate components. There are a few bugs (crashes on grenade explosions in gl/glsl/vulkan, immediately in sw, reasons known, missing brush models in vulkan). While quake doesn't really need an ECS, the direction I want to take QF does, and it does seem to have improved memory bandwidth a little (uncertain). However, there's a lot more work to go (especially fixing the above bugs), but this seems to be a good start.
2022-10-23 01:32:09 +00:00
sizeof (mat4f_t), Transform_GetWorldMatrixPtr (transform) },
[vulkan] Clean up alias push-constants
2021-12-17 07:46:54 +00:00
{ VK_SHADER_STAGE_VERTEX_BIT,
field_offset (alias_push_constants_t, blend),
sizeof (float), &constants.blend },
{ VK_SHADER_STAGE_FRAGMENT_BIT,
field_offset (alias_push_constants_t, colorA),
sizeof (constants.colorA), constants.colorA },
{ VK_SHADER_STAGE_FRAGMENT_BIT,
field_offset (alias_push_constants_t, colorB),
sizeof (constants.colorB), constants.colorB },
{ VK_SHADER_STAGE_FRAGMENT_BIT,
field_offset (alias_push_constants_t, base_color),
sizeof (constants.base_color), &constants.base_color },
{ VK_SHADER_STAGE_FRAGMENT_BIT,
field_offset (alias_push_constants_t, fog),
sizeof (constants.fog), &constants.fog },
};
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
if (0/*XXX ent->skin && ent->skin->tex*/) {
//skin = ent->skin->tex;
} else {
maliasskindesc_t *skindesc;
[scene] Make entity_t just an entity id for ECS This puts the hierarchy (transform) reference, animation, visibility, renderer, active, and old_origin data in separate components. There are a few bugs (crashes on grenade explosions in gl/glsl/vulkan, immediately in sw, reasons known, missing brush models in vulkan). While quake doesn't really need an ECS, the direction I want to take QF does, and it does seem to have improved memory bandwidth a little (uncertain). However, there's a lot more work to go (especially fixing the above bugs), but this seems to be a good start.
2022-10-23 01:32:09 +00:00
skindesc = R_AliasGetSkindesc (animation, renderer->skinnum, hdr);
[vulkan] Implement team colors It's not quite as expected, but that may be due to one of msaa, the 0-15 range in the palette not being all the way to white, the color gradients being not quite linear (haven't checked yet) or some combination of the above. However, it's that what should be yellow is more green. At least the zombies are no longer white and the ogres don't look like they're wearing skeleton suits.
2021-02-02 10:53:36 +00:00
skin = (qfv_alias_skin_t *) ((byte *) hdr + skindesc->skin);
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
}
[scene] Make entity_t just an entity id for ECS This puts the hierarchy (transform) reference, animation, visibility, renderer, active, and old_origin data in separate components. There are a few bugs (crashes on grenade explosions in gl/glsl/vulkan, immediately in sw, reasons known, missing brush models in vulkan). While quake doesn't really need an ECS, the direction I want to take QF does, and it does seem to have improved memory bandwidth a little (uncertain). However, there's a lot more work to go (especially fixing the above bugs), but this seems to be a good start.
2022-10-23 01:32:09 +00:00
QuatCopy (renderer->colormod, constants.base_color);
[vulkan] Clean up alias push-constants
2021-12-17 07:46:54 +00:00
QuatCopy (skin->colora, constants.colorA);
QuatCopy (skin->colorb, constants.colorB);
QuatZero (constants.fog);
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
[renderer] Clean up entity_t to a certain extent This is the first step towards component-based entities. There's still some transform-related stuff in the struct that needs to be moved, but it's all entirely client related (rather than renderer) and will probably go into a "client" component. Also, the current components are directly included structs rather than references as I didn't want to deal with the object management at this stage. As part of the process (because transforms use simd) this also starts the process of moving QF to using simd for vectors and matrices. There's now a mess of simd and sisd code mixed together, but it works surprisingly well together.
2021-03-09 14:52:40 +00:00
emit_commands (aframe->cmdSet.a[QFV_aliasDepth],
[scene] Make entity_t just an entity id for ECS This puts the hierarchy (transform) reference, animation, visibility, renderer, active, and old_origin data in separate components. There are a few bugs (crashes on grenade explosions in gl/glsl/vulkan, immediately in sw, reasons known, missing brush models in vulkan). While quake doesn't really need an ECS, the direction I want to take QF does, and it does seem to have improved memory bandwidth a little (uncertain). However, there's a lot more work to go (especially fixing the above bugs), but this seems to be a good start.
2022-10-23 01:32:09 +00:00
animation->pose1, animation->pose2,
[vulkan] Add entity labels to command queues Brush models looked a little too tricky due to the very different style of command queue, so that's left for now, but alias, iqm and sprite entities are now labeled. The labels are made up of the lower 5 hex digits of the entity address, the position, and colored by the normalized position vector. Not sure that's the best choice as it does mean the color changes as the entity moves, and can be quite subtle between nearby entities, but it still helps identify the entities in the command buffer. And, as I suspected, I've got multiple draw calls for the one ogre. Now to find out why.
2022-05-07 05:12:02 +00:00
0, 2, push_constants, hdr, rFrame, ent);
[renderer] Clean up entity_t to a certain extent This is the first step towards component-based entities. There's still some transform-related stuff in the struct that needs to be moved, but it's all entirely client related (rather than renderer) and will probably go into a "client" component. Also, the current components are directly included structs rather than references as I didn't want to deal with the object management at this stage. As part of the process (because transforms use simd) this also starts the process of moving QF to using simd for vectors and matrices. There's now a mess of simd and sisd code mixed together, but it works surprisingly well together.
2021-03-09 14:52:40 +00:00
emit_commands (aframe->cmdSet.a[QFV_aliasGBuffer],
[scene] Make entity_t just an entity id for ECS This puts the hierarchy (transform) reference, animation, visibility, renderer, active, and old_origin data in separate components. There are a few bugs (crashes on grenade explosions in gl/glsl/vulkan, immediately in sw, reasons known, missing brush models in vulkan). While quake doesn't really need an ECS, the direction I want to take QF does, and it does seem to have improved memory bandwidth a little (uncertain). However, there's a lot more work to go (especially fixing the above bugs), but this seems to be a good start.
2022-10-23 01:32:09 +00:00
animation->pose1, animation->pose2,
[vulkan] Add entity labels to command queues Brush models looked a little too tricky due to the very different style of command queue, so that's left for now, but alias, iqm and sprite entities are now labeled. The labels are made up of the lower 5 hex digits of the entity address, the position, and colored by the normalized position vector. Not sure that's the best choice as it does mean the color changes as the entity moves, and can be quite subtle between nearby entities, but it still helps identify the entities in the command buffer. And, as I suspected, I've got multiple draw calls for the one ogre. Now to find out why.
2022-05-07 05:12:02 +00:00
skin, 6, push_constants, hdr, rFrame, ent);
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
}
[vulkan] Sort of implement ambient lighting It works, but as it uses the camera position rather than the lit entity's position, it is incorrect. I have some interim ideas to handle it, though.
2021-01-28 10:17:40 +00:00
static void
[vulkan] Convert bsp and draw to deferred This has bsp and draw passing muster with the validation layers.
2021-02-19 02:14:54 +00:00
alias_begin_subpass (QFV_AliasSubpass subpass, VkPipeline pipeline,
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
qfv_renderframe_t *rFrame)
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
{
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
vulkan_ctx_t *ctx = rFrame->vulkan_ctx;
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
qfv_device_t *device = ctx->device;
qfv_devfuncs_t *dfunc = device->funcs;
aliasctx_t *actx = ctx->alias_context;
[vulkan] Rename vulkan_framebuffer_t It turns out I had conflated frame buffers with frames and wound up making a minor mess when separating the number of frames the renderer could have in flight from the number of swap-chain images. This is the first step towards correcting that mistake.
2021-02-05 01:22:32 +00:00
__auto_type cframe = &ctx->frames.a[ctx->curFrame];
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
aliasframe_t *aframe = &actx->frames.a[ctx->curFrame];
[vulkan] Convert bsp and draw to deferred This has bsp and draw passing muster with the validation layers.
2021-02-19 02:14:54 +00:00
VkCommandBuffer cmd = aframe->cmdSet.a[subpass];
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
dfunc->vkResetCommandBuffer (cmd, 0);
VkCommandBufferInheritanceInfo inherit = {
VK_STRUCTURE_TYPE_COMMAND_BUFFER_INHERITANCE_INFO, 0,
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
rFrame->renderpass->renderpass, subpass_map[subpass],
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
cframe->framebuffer,
0, 0, 0,
};
VkCommandBufferBeginInfo beginInfo = {
VK_STRUCTURE_TYPE_COMMAND_BUFFER_BEGIN_INFO, 0,
VK_COMMAND_BUFFER_USAGE_ONE_TIME_SUBMIT_BIT
| VK_COMMAND_BUFFER_USAGE_RENDER_PASS_CONTINUE_BIT, &inherit,
};
dfunc->vkBeginCommandBuffer (cmd, &beginInfo);
[vulkan] Label the drawing command buffers
2021-04-18 11:40:43 +00:00
QFV_duCmdBeginLabel (device, cmd, va (ctx->va_ctx, "alias:%s",
alias_pass_names[subpass]),
{ 0.6, 0.5, 0, 1});
[vulkan] Implement lighting and compose passes Lighting doesn't actually do lights yet, but it's producing pixels. Translucent seems to be working (2d draw uses it), and compose seems to be working.
2021-02-24 10:58:31 +00:00
dfunc->vkCmdBindPipeline (cmd, VK_PIPELINE_BIND_POINT_GRAPHICS, pipeline);
[vulkan] Switch alias pipeline to multi-texture This gets the alias pipeline in line with the bsp pipeline, and thus everything is about as functional as it was before the rework (minus dealing with large texture sets).
2021-12-08 15:25:50 +00:00
VkDescriptorSet sets[] = {
[vulkan] Fix a typo and gcc purity warning
2021-12-13 01:39:01 +00:00
Vulkan_Matrix_Descriptors (ctx, ctx->curFrame),
[vulkan] Switch alias pipeline to multi-texture This gets the alias pipeline in line with the bsp pipeline, and thus everything is about as functional as it was before the rework (minus dealing with large texture sets).
2021-12-08 15:25:50 +00:00
};
dfunc->vkCmdBindDescriptorSets (cmd, VK_PIPELINE_BIND_POINT_GRAPHICS,
[vulkan] Abandon bindless rendering At least for now, it is more trouble than it is worth as it (or my implementation of it) breaks lavapipe and renderdoc.
2021-12-17 03:20:32 +00:00
actx->layout, 0, 1, sets, 0, 0);
[vulkan] Move viewport and scissor into qfv_renderpass_t This makes much more sense as they are intimately tied to the frame buffer on which a render pass is working. Now, just the window width and height are stored in vulkan_ctx_t. As a side benefit, QFV_CreateSwapchain no long references viddef (now just palette and conview in vulkan_draw.c to go).
2022-04-01 11:34:41 +00:00
dfunc->vkCmdSetViewport (cmd, 0, 1, &rFrame->renderpass->viewport);
dfunc->vkCmdSetScissor (cmd, 0, 1, &rFrame->renderpass->scissor);
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
//XXX glsl_Fog_GetColor (fog);
//XXX fog[3] = glsl_Fog_GetDensity () / 64.0;
}
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
static void
alias_end_subpass (VkCommandBuffer cmd, vulkan_ctx_t *ctx)
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
{
qfv_device_t *device = ctx->device;
qfv_devfuncs_t *dfunc = device->funcs;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
[vulkan] Label the drawing command buffers
2021-04-18 11:40:43 +00:00
QFV_duCmdEndLabel (device, cmd);
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
dfunc->vkEndCommandBuffer (cmd);
}
void
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
Vulkan_AliasBegin (qfv_renderframe_t *rFrame)
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
{
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
vulkan_ctx_t *ctx = rFrame->vulkan_ctx;
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
aliasctx_t *actx = ctx->alias_context;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
aliasframe_t *aframe = &actx->frames.a[ctx->curFrame];
//XXX quat_t fog;
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
DARRAY_APPEND (&rFrame->subpassCmdSets[QFV_passDepth],
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
aframe->cmdSet.a[QFV_aliasDepth]);
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
DARRAY_APPEND (&rFrame->subpassCmdSets[QFV_passGBuffer],
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
aframe->cmdSet.a[QFV_aliasGBuffer]);
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
alias_begin_subpass (QFV_aliasDepth, actx->depth, rFrame);
alias_begin_subpass (QFV_aliasGBuffer, actx->gbuf, rFrame);
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
}
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
void
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
Vulkan_AliasEnd (qfv_renderframe_t *rFrame)
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
{
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
vulkan_ctx_t *ctx = rFrame->vulkan_ctx;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
aliasctx_t *actx = ctx->alias_context;
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
aliasframe_t *aframe = &actx->frames.a[ctx->curFrame];
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
alias_end_subpass (aframe->cmdSet.a[QFV_aliasDepth], ctx);
alias_end_subpass (aframe->cmdSet.a[QFV_aliasGBuffer], ctx);
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
}
[vulkan] Implement view model rendering Quake just looked wrong without the view model. I can't say I like the way the depth range is hacked, but it was necessary because the view model needs to be processed along with the rest of the alias models (didn't feel like adding more command buffers, which I imagine would be expensive with the pipeline switching).
2021-07-25 05:03:25 +00:00
void
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
Vulkan_AliasDepthRange (qfv_renderframe_t *rFrame,
float minDepth, float maxDepth)
[vulkan] Implement view model rendering Quake just looked wrong without the view model. I can't say I like the way the depth range is hacked, but it was necessary because the view model needs to be processed along with the rest of the alias models (didn't feel like adding more command buffers, which I imagine would be expensive with the pipeline switching).
2021-07-25 05:03:25 +00:00
{
[vulkan] Support multiple render passes Multiple render passes are needed for supporting shadow mapping, and this is a huge step towards breaking the Vulkan render free of Quake, and hopefully will lead the way for breaking the GL renderers free as well.
2021-12-02 12:58:29 +00:00
vulkan_ctx_t *ctx = rFrame->vulkan_ctx;
[vulkan] Implement view model rendering Quake just looked wrong without the view model. I can't say I like the way the depth range is hacked, but it was necessary because the view model needs to be processed along with the rest of the alias models (didn't feel like adding more command buffers, which I imagine would be expensive with the pipeline switching).
2021-07-25 05:03:25 +00:00
qfv_device_t *device = ctx->device;
qfv_devfuncs_t *dfunc = device->funcs;
aliasctx_t *actx = ctx->alias_context;
aliasframe_t *aframe = &actx->frames.a[ctx->curFrame];
[vulkan] Move viewport and scissor into qfv_renderpass_t This makes much more sense as they are intimately tied to the frame buffer on which a render pass is working. Now, just the window width and height are stored in vulkan_ctx_t. As a side benefit, QFV_CreateSwapchain no long references viddef (now just palette and conview in vulkan_draw.c to go).
2022-04-01 11:34:41 +00:00
VkViewport viewport = rFrame->renderpass->viewport;
[vulkan] Implement view model rendering Quake just looked wrong without the view model. I can't say I like the way the depth range is hacked, but it was necessary because the view model needs to be processed along with the rest of the alias models (didn't feel like adding more command buffers, which I imagine would be expensive with the pipeline switching).
2021-07-25 05:03:25 +00:00
viewport.minDepth = minDepth;
viewport.maxDepth = maxDepth;
dfunc->vkCmdSetViewport (aframe->cmdSet.a[QFV_aliasDepth], 0, 1,
&viewport);
dfunc->vkCmdSetViewport (aframe->cmdSet.a[QFV_aliasGBuffer], 0, 1,
&viewport);
}
[vulkan] Switch alias pipeline to multi-texture This gets the alias pipeline in line with the bsp pipeline, and thus everything is about as functional as it was before the rework (minus dealing with large texture sets).
2021-12-08 15:25:50 +00:00
void
Vulkan_AliasAddSkin (vulkan_ctx_t *ctx, qfv_alias_skin_t *skin)
{
aliasctx_t *actx = ctx->alias_context;
[vulkan] Abandon bindless rendering At least for now, it is more trouble than it is worth as it (or my implementation of it) breaks lavapipe and renderdoc.
2021-12-17 03:20:32 +00:00
skin->descriptor = Vulkan_CreateCombinedImageSampler (ctx, skin->view,
actx->sampler);
[vulkan] Switch alias pipeline to multi-texture This gets the alias pipeline in line with the bsp pipeline, and thus everything is about as functional as it was before the rework (minus dealing with large texture sets).
2021-12-08 15:25:50 +00:00
}
void
Vulkan_AliasRemoveSkin (vulkan_ctx_t *ctx, qfv_alias_skin_t *skin)
{
[vulkan] Abandon bindless rendering At least for now, it is more trouble than it is worth as it (or my implementation of it) breaks lavapipe and renderdoc.
2021-12-17 03:20:32 +00:00
Vulkan_FreeTexture (ctx, skin->descriptor);
skin->descriptor = 0;
[vulkan] Switch alias pipeline to multi-texture This gets the alias pipeline in line with the bsp pipeline, and thus everything is about as functional as it was before the rework (minus dealing with large texture sets).
2021-12-08 15:25:50 +00:00
}
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
void
Vulkan_Alias_Init (vulkan_ctx_t *ctx)
{
qfv_device_t *device = ctx->device;
[vulkan] Add some more debug stack contexts
2021-12-02 13:48:50 +00:00
qfvPushDebug (ctx, "alias init");
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
aliasctx_t *actx = calloc (1, sizeof (aliasctx_t));
ctx->alias_context = actx;
[vulkan] Rename vulkan_framebuffer_t It turns out I had conflated frame buffers with frames and wound up making a minor mess when separating the number of frames the renderer could have in flight from the number of swap-chain images. This is the first step towards correcting that mistake.
2021-02-05 01:22:32 +00:00
size_t frames = ctx->frames.size;
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
DARRAY_INIT (&actx->frames, frames);
DARRAY_RESIZE (&actx->frames, frames);
actx->frames.grow = 0;
[vulkan] Add support for compute shaders This needed changing Vulkan_CreatePipeline to Vulkan_CreateGraphicsPipeline for consistency (and parsing the difference from a plist seemed... not worth thinking about).
2021-12-16 14:00:17 +00:00
actx->depth = Vulkan_CreateGraphicsPipeline (ctx, "alias_depth");
actx->gbuf = Vulkan_CreateGraphicsPipeline (ctx, "alias_gbuf");
[vulkan] Rework pipeline parsing for better reuse It's not entirely there yet, but the basics are working. Work is still needed for avoiding duplication of objects (different threads will have different contexts and thus different tables, so necessary per-thread duplication should not become a problem) and general access to arbitrary fields (mostly just parsing the strings)
2021-02-04 08:03:49 +00:00
actx->layout = Vulkan_CreatePipelineLayout (ctx, "alias_layout");
actx->sampler = Vulkan_CreateSampler (ctx, "alias_sampler");
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
for (size_t i = 0; i < frames; i++) {
__auto_type aframe = &actx->frames.a[i];
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
DARRAY_INIT (&aframe->cmdSet, QFV_aliasNumPasses);
DARRAY_RESIZE (&aframe->cmdSet, QFV_aliasNumPasses);
aframe->cmdSet.grow = 0;
QFV_AllocateCommandBuffers (device, ctx->cmdpool, 1, &aframe->cmdSet);
for (int j = 0; j < QFV_aliasNumPasses; j++) {
QFV_duSetObjectName (device, VK_OBJECT_TYPE_COMMAND_BUFFER,
aframe->cmdSet.a[j],
va (ctx->va_ctx, "cmd:alias:%zd:%s", i,
alias_pass_names[j]));
}
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
}
[vulkan] Add some more debug stack contexts
2021-12-02 13:48:50 +00:00
qfvPopDebug (ctx);
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
}
void
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
Vulkan_Alias_Shutdown (vulkan_ctx_t *ctx)
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
{
qfv_device_t *device = ctx->device;
qfv_devfuncs_t *dfunc = device->funcs;
aliasctx_t *actx = ctx->alias_context;
[vulkan] Clean up alias light buffer Still wedging, but it seems to be something to do with the matrix buffer.
2021-01-27 08:59:09 +00:00
for (size_t i = 0; i < actx->frames.size; i++) {
__auto_type aframe = &actx->frames.a[i];
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
free (aframe->cmdSet.a);
[vulkan] Clean up alias light buffer Still wedging, but it seems to be something to do with the matrix buffer.
2021-01-27 08:59:09 +00:00
}
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
dfunc->vkDestroyPipeline (device->dev, actx->depth, 0);
dfunc->vkDestroyPipeline (device->dev, actx->gbuf, 0);
free (actx->frames.a);
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
free (actx);
}
Reference in a new issue Copy permalink