qf/quakeforge
0
0
Fork 0
mirror of https://git.code.sf.net/p/quake/quakeforge synced 2024-11-23 04:42:32 +00:00
Code Issues Projects Releases Packages Wiki Activity
quakeforge/libs/video/renderer/vulkan/vulkan_sprite.c

346 lines
10 KiB
C
Raw Normal View History

[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
/*
vulkan_sprite.c
Vulkan sprite 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: 2021/12/14
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
#ifdef HAVE_STRING_H
# include <string.h>
#endif
#ifdef HAVE_STRINGS_H
# include <strings.h>
#endif
#include <stdlib.h>
#include "qfalloca.h"
#include "QF/cvar.h"
#include "QF/darray.h"
#include "QF/image.h"
#include "QF/render.h"
#include "QF/skin.h"
#include "QF/sys.h"
#include "QF/va.h"
#include "QF/scene/entity.h"
#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] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
#include "QF/Vulkan/qf_sprite.h"
#include "QF/Vulkan/qf_texture.h"
#include "QF/Vulkan/buffer.h"
#include "QF/Vulkan/command.h"
#include "QF/Vulkan/debug.h"
#include "QF/Vulkan/descriptor.h"
#include "QF/Vulkan/device.h"
#include "QF/Vulkan/instance.h"
#include "r_internal.h"
#include "vid_vulkan.h"
static const char * __attribute__((used)) sprite_pass_names[] = {
"depth",
"g-buffer",
"translucent",
};
static QFV_Subpass subpass_map[] = {
QFV_passDepth, // QFV_spriteDepth
QFV_passGBuffer, // QFV_spriteGBuffer
[vulkan] Implement order implement transparency It's a bit flaky for particles, especially at higher frame rates, but that's due to supporting only 64 overlapping pixels. A reasonable solution is probably switching to a priority heap for the "sort" and upping the limit.
2022-11-30 18:00:47 +00:00
QFV_passTranslucentFrag,// QFV_spriteTranslucent
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
};
static void
emit_commands (VkCommandBuffer cmd, qfv_sprite_t *sprite,
int 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
qfv_renderframe_t *rFrame, entity_t ent)
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
{
vulkan_ctx_t *ctx = rFrame->vulkan_ctx;
qfv_device_t *device = ctx->device;
qfv_devfuncs_t *dfunc = device->funcs;
spritectx_t *sctx = ctx->sprite_context;
[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] Split out the push constants helper It proved to work well in the sprite pipeline, and should be useful for cleaning up the magic numbers in bsp and alias.
2021-12-17 05:54:02 +00:00
QFV_PushConstants (device, cmd, sctx->layout, numPC, constants);
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
VkDescriptorSet sets[] = {
sprite->descriptors,
};
dfunc->vkCmdBindDescriptorSets (cmd, VK_PIPELINE_BIND_POINT_GRAPHICS,
sctx->layout, 1, 1, sets, 0, 0);
dfunc->vkCmdDraw (cmd, 4, 1, 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] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +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_DrawSprite (entity_t ent, qfv_renderframe_t *rFrame)
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
{
vulkan_ctx_t *ctx = rFrame->vulkan_ctx;
spritectx_t *sctx = ctx->sprite_context;
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
spriteframe_t *sframe = &sctx->frames.a[ctx->curFrame];
[renderer] Tidy up entity component access Since entity_t has a pointer to the registry owning the entity, there's no need to access a global to get at the registry. Also move component getting closer to where it's used.
2022-10-25 03:53:30 +00:00
renderer_t *renderer = Ent_GetComponent (ent.id, scene_renderer, ent.reg);
[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] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
msprite_t *sprite = model->cache.data;
[vulkan] Clean up r_origin and modelorg Same deal as for GL and GLSL.
2022-03-08 09:41:09 +00:00
mat4f_t mat = {};
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
uint32_t frame;
qfv_push_constants_t push_constants[] = {
{ VK_SHADER_STAGE_VERTEX_BIT, 0, sizeof (mat), mat },
{ VK_SHADER_STAGE_VERTEX_BIT | VK_SHADER_STAGE_FRAGMENT_BIT,
64, sizeof (frame), &frame },
};
[renderer] Tidy up entity component access Since entity_t has a pointer to the registry owning the entity, there's no need to access a global to get at the registry. Also move component getting closer to where it's used.
2022-10-25 03:53:30 +00:00
animation_t *animation = Ent_GetComponent (ent.id, scene_animation,
ent.reg);
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
frame = (ptrdiff_t) R_GetSpriteFrame (sprite, animation);
[renderer] Tidy up entity component access Since entity_t has a pointer to the registry owning the entity, there's no need to access a global to get at the registry. Also move component getting closer to where it's used.
2022-10-25 03:53:30 +00:00
transform_t transform = Entity_Transform (ent);
[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
mat[3] = Transform_GetWorldPosition (transform);
[renderer] Clean up use of vup/vright/vpn This moves the common camera setup code out of the individual drivers, and completely removes vup/vright/vpn from the non-software renderers. This has highlighted the craziness around AngleVectors with it putting +X forward, -Y right and +Z up. The main issue with this is it requires a 90 degree pre-rotation about the Z axis to get the camera pointing in the right direction, and that's for the native sw renderer (vulkan needs a 90 degree pre-rotation about X, and gl and glsl need to invert an axis, too), though at least it's just a matrix swizzle and vector negation. However, it does mean the camera matrices can't be used directly. Also rename vpn to vfwd (still abbreviated, but fwd is much clearer in meaning (to me, at least) than pn (plane normal, I guess, but which way?)).
2022-03-14 00:34:24 +00:00
vec4f_t cameravec = r_refdef.frame.position - mat[3];
[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
R_BillboardFrame (transform, sprite->type, cameravec,
[renderer] Move a couple functions to using vec4f_t Makes for a few less FIXMEs and better consistency with vectors.
2022-03-30 14:50:12 +00:00
&mat[2], &mat[1], &mat[0]);
[vulkan] Clean up r_origin and modelorg Same deal as for GL and GLSL.
2022-03-08 09:41:09 +00:00
mat[0] = -mat[0];
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
emit_commands (sframe->cmdSet.a[QFV_spriteDepth],
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
(qfv_sprite_t *) ((byte *) sprite + sprite->data),
[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
2, push_constants, rFrame, ent);
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
emit_commands (sframe->cmdSet.a[QFV_spriteGBuffer],
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
(qfv_sprite_t *) ((byte *) sprite + sprite->data),
[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
2, push_constants, rFrame, ent);
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
}
static void
sprite_begin_subpass (QFV_SpriteSubpass subpass, VkPipeline pipeline,
qfv_renderframe_t *rFrame)
{
vulkan_ctx_t *ctx = rFrame->vulkan_ctx;
qfv_device_t *device = ctx->device;
qfv_devfuncs_t *dfunc = device->funcs;
spritectx_t *sctx = ctx->sprite_context;
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
spriteframe_t *sframe = &sctx->frames.a[ctx->curFrame];
VkCommandBuffer cmd = sframe->cmdSet.a[subpass];
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
dfunc->vkResetCommandBuffer (cmd, 0);
VkCommandBufferInheritanceInfo inherit = {
VK_STRUCTURE_TYPE_COMMAND_BUFFER_INHERITANCE_INFO, 0,
rFrame->renderpass->renderpass, subpass_map[subpass],
[vulkan] Get multiple frame buffers working Now each (high level) render pass can have its own frame buffer. The current goal is to get the final output render pass to just transfer the composed output to the swap chain image, potentially with scaling (my laptop might be able to cope).
2022-11-21 08:25:55 +00:00
rFrame->framebuffer,
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
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);
QFV_duCmdBeginLabel (device, cmd, va (ctx->va_ctx, "sprite:%s",
sprite_pass_names[subpass]),
{ 0.6, 0.5, 0, 1});
dfunc->vkCmdBindPipeline (cmd, VK_PIPELINE_BIND_POINT_GRAPHICS, pipeline);
VkDescriptorSet sets[] = {
Vulkan_Matrix_Descriptors (ctx, ctx->curFrame),
};
dfunc->vkCmdBindDescriptorSets (cmd, VK_PIPELINE_BIND_POINT_GRAPHICS,
sctx->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] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
//XXX glsl_Fog_GetColor (fog);
//XXX fog[3] = glsl_Fog_GetDensity () / 64.0;
}
static void
sprite_end_subpass (VkCommandBuffer cmd, vulkan_ctx_t *ctx)
{
qfv_device_t *device = ctx->device;
qfv_devfuncs_t *dfunc = device->funcs;
QFV_duCmdEndLabel (device, cmd);
dfunc->vkEndCommandBuffer (cmd);
}
void
Vulkan_SpriteBegin (qfv_renderframe_t *rFrame)
{
vulkan_ctx_t *ctx = rFrame->vulkan_ctx;
spritectx_t *sctx = ctx->sprite_context;
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
spriteframe_t *sframe = &sctx->frames.a[ctx->curFrame];
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
//XXX quat_t fog;
DARRAY_APPEND (&rFrame->subpassCmdSets[QFV_passDepth],
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
sframe->cmdSet.a[QFV_spriteDepth]);
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
DARRAY_APPEND (&rFrame->subpassCmdSets[QFV_passGBuffer],
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
sframe->cmdSet.a[QFV_spriteGBuffer]);
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
sprite_begin_subpass (QFV_spriteDepth, sctx->depth, rFrame);
sprite_begin_subpass (QFV_spriteGBuffer, sctx->gbuf, rFrame);
}
void
Vulkan_SpriteEnd (qfv_renderframe_t *rFrame)
{
vulkan_ctx_t *ctx = rFrame->vulkan_ctx;
spritectx_t *sctx = ctx->sprite_context;
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
spriteframe_t *sframe = &sctx->frames.a[ctx->curFrame];
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
sprite_end_subpass (sframe->cmdSet.a[QFV_spriteDepth], ctx);
sprite_end_subpass (sframe->cmdSet.a[QFV_spriteGBuffer], ctx);
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
}
static VkDescriptorBufferInfo base_buffer_info = {
.range = VK_WHOLE_SIZE,
};
static VkDescriptorImageInfo base_image_info = {
.imageLayout = VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL,
};
static VkWriteDescriptorSet base_buffer_write = {
.sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET,
.dstBinding = 0,
.descriptorCount = 1,
.descriptorType = VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER,
};
static VkWriteDescriptorSet base_image_write = {
.sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET,
.dstBinding = 1,
.descriptorCount = 1,
.descriptorType = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER,
};
void
Vulkan_Sprite_DescriptorSet (vulkan_ctx_t *ctx, qfv_sprite_t *sprite)
{
qfv_device_t *device = ctx->device;
qfv_devfuncs_t *dfunc = device->funcs;
spritectx_t *sctx = ctx->sprite_context;
//FIXME kinda dumb
__auto_type layouts = QFV_AllocDescriptorSetLayoutSet (1, alloca);
for (size_t i = 0; i < layouts->size; i++) {
layouts->a[i] = sctx->setLayout;
}
__auto_type sets = QFV_AllocateDescriptorSet (device, sctx->pool, layouts);
sprite->descriptors = sets->a[0];
free (sets);
VkDescriptorBufferInfo bufferInfo[1];
bufferInfo[0] = base_buffer_info;
bufferInfo[0].buffer = sprite->verts;
VkDescriptorImageInfo imageInfo[1];
imageInfo[0] = base_image_info;
imageInfo[0].sampler = sctx->sampler;
imageInfo[0].imageView = sprite->view;
VkWriteDescriptorSet write[2];
write[0] = base_buffer_write;
write[0].dstSet = sprite->descriptors;
write[0].pBufferInfo = bufferInfo;
write[1] = base_image_write;
write[1].dstSet = sprite->descriptors;
write[1].pImageInfo = imageInfo;
dfunc->vkUpdateDescriptorSets (device->dev, 2, write, 0, 0);
}
[vulkan] Plug some leaking sprite descriptors I'd forgotten to free them when unloading a model, thus the segfault after about 31 timedemo runs.
2022-05-10 16:39:33 +00:00
void
Vulkan_Sprint_FreeDescriptors (vulkan_ctx_t *ctx, qfv_sprite_t *sprite)
{
qfv_device_t *device = ctx->device;
qfv_devfuncs_t *dfunc = device->funcs;
spritectx_t *sctx = ctx->sprite_context;
dfunc->vkFreeDescriptorSets (device->dev, sctx->pool, 1,
&sprite->descriptors);
}
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
void
Vulkan_Sprite_Init (vulkan_ctx_t *ctx)
{
qfv_device_t *device = ctx->device;
qfvPushDebug (ctx, "sprite init");
spritectx_t *sctx = calloc (1, sizeof (spritectx_t));
ctx->sprite_context = sctx;
size_t frames = ctx->frames.size;
DARRAY_INIT (&sctx->frames, frames);
DARRAY_RESIZE (&sctx->frames, frames);
sctx->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
sctx->depth = Vulkan_CreateGraphicsPipeline (ctx, "sprite_depth");
sctx->gbuf = Vulkan_CreateGraphicsPipeline (ctx, "sprite_gbuf");
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
sctx->layout = Vulkan_CreatePipelineLayout (ctx, "sprite_layout");
sctx->sampler = Vulkan_CreateSampler (ctx, "sprite_sampler");
sctx->pool = Vulkan_CreateDescriptorPool (ctx, "sprite_pool");
sctx->setLayout = Vulkan_CreateDescriptorSetLayout (ctx, "sprite_set");
for (size_t i = 0; i < frames; i++) {
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
__auto_type sframe = &sctx->frames.a[i];
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
DARRAY_INIT (&sframe->cmdSet, QFV_spriteNumPasses);
DARRAY_RESIZE (&sframe->cmdSet, QFV_spriteNumPasses);
sframe->cmdSet.grow = 0;
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
QFV_AllocateCommandBuffers (device, ctx->cmdpool, 1, &sframe->cmdSet);
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
for (int j = 0; j < QFV_spriteNumPasses; j++) {
QFV_duSetObjectName (device, VK_OBJECT_TYPE_COMMAND_BUFFER,
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
sframe->cmdSet.a[j],
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
va (ctx->va_ctx, "cmd:sprite:%zd:%s", i,
sprite_pass_names[j]));
}
}
qfvPopDebug (ctx);
}
void
Vulkan_Sprite_Shutdown (vulkan_ctx_t *ctx)
{
qfv_device_t *device = ctx->device;
qfv_devfuncs_t *dfunc = device->funcs;
spritectx_t *sctx = ctx->sprite_context;
for (size_t i = 0; i < sctx->frames.size; i++) {
[vulkan] Clean up a pile of C&P fluff
2021-12-16 14:04:19 +00:00
__auto_type sframe = &sctx->frames.a[i];
free (sframe->cmdSet.a);
[vulkan] Implement the sprite render passes It turned out the bindless approach wouldn't work too well for my design of the sprite objects, but I don't think that's a big issue at this stage (and it seems bindless is causing problems for brush/alias rendering via renderdoc and on my versa pro). However, I have figured out how to make effective use of descriptor sets (finally :P). The actual normal still needs checking, but the sprites are currently unlit so not an issue at this stage.
2021-12-14 16:22:05 +00:00
}
dfunc->vkDestroyPipeline (device->dev, sctx->depth, 0);
dfunc->vkDestroyPipeline (device->dev, sctx->gbuf, 0);
free (sctx->frames.a);
free (sctx);
}
Reference in a new issue Copy permalink