qf
/
quakeforge
mirror of https://git.code.sf.net/p/quake/quakeforge
0
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
quakeforge/include/vid_vulkan.h

123 lines
3.5 KiB
C
Raw Normal View History

Separate render and presentation initialization This paves the way for clean initialization of the Vulkan renderer, and very much cleans up the older renderer initialization code as gl and sw are no longer intertwined.
2019-07-08 16:00:47 +00:00
#ifndef __vid_vulkan_h
#define __vid_vulkan_h
#ifndef VK_NO_PROTOTYPES
#define VK_NO_PROTOTYPES
#endif
#include <vulkan/vulkan.h>
Create and destroy render pass and frame buffers
2020-02-18 08:18:37 +00:00
#include "QF/darray.h"
[vulkan] Move projection matrix code to its own file And move more stuff over to simd.
2021-04-25 06:47:50 +00:00
#include "QF/simd/types.h"
Create command pool and frame timing fences Nothing is actually done yet, so the reported fps is around 172k (yes, k), but startup and shutdown seems to be clean (yay validation layers).
2020-02-16 13:45:27 +00:00
[vulkan] Break render pass parsing away from swapchain This allows a single render pass description to be used for both on-screen and off-screen targets. While Vulkan does allow a VkRenderPass to be used with any compatible frame buffer, and vkparse caches a VkRenderPass created from the same description, this allows the same description to be used for a compatible off-screen target without any dependence on the swapchain. However, there is a problem in the caching when it comes to targeting outputs with different formats.
2022-04-02 02:42:36 +00:00
//FIXME name
typedef struct qfv_output_s {
VkExtent2D extent;
VkImageView view;
VkFormat format;
} qfv_output_t;
[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
typedef struct vulkan_frame_s {
Create and destroy render pass and frame buffers
2020-02-18 08:18:37 +00:00
VkFramebuffer framebuffer;
VkFence fence;
VkSemaphore imageAvailableSemaphore;
VkSemaphore renderDoneSemaphore;
VkCommandBuffer cmdBuffer;
[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
} vulkan_frame_t;
Create and destroy render pass and frame buffers
2020-02-18 08:18:37 +00:00
[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
typedef struct vulkan_frameset_s
DARRAY_TYPE (vulkan_frame_t) vulkan_frameset_t;
Create and destroy render pass and frame buffers
2020-02-18 08:18:37 +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
typedef struct qfv_renderpassset_s
DARRAY_TYPE (struct qfv_renderpass_s *) qfv_renderpassset_t;
[vulkan] Parse clear values And get the render pass working in general. Note that this is only the render pass and frame buffers: actual commands still have problems.
2021-02-23 05:37:48 +00:00
Separate render and presentation initialization This paves the way for clean initialization of the Vulkan renderer, and very much cleans up the older renderer initialization code as gl and sw are no longer intertwined.
2019-07-08 16:00:47 +00:00
typedef struct vulkan_ctx_s {
void (*load_vulkan) (struct vulkan_ctx_s *ctx);
void (*unload_vulkan) (struct vulkan_ctx_s *ctx);
Remove global vulkan_ctx While I can't say that I'm happy with the details of vulkan_ctx_t, I am pretty sure I don't want to be limited to having only one.
2019-07-09 02:54:23 +00:00
const char **required_extensions;
Separate render and presentation initialization This paves the way for clean initialization of the Vulkan renderer, and very much cleans up the older renderer initialization code as gl and sw are no longer intertwined.
2019-07-08 16:00:47 +00:00
struct vulkan_presentation_s *presentation;
int (*get_presentation_support) (struct vulkan_ctx_s *ctx,
VkPhysicalDevice physicalDevice,
uint32_t queueFamilyIndex);
Create a window and a surface for vulkan Yay, segfaults in R_Init :)
2019-07-09 07:33:44 +00:00
void (*choose_visual) (struct vulkan_ctx_s *ctx);
void (*create_window) (struct vulkan_ctx_s *ctx);
Separate render and presentation initialization This paves the way for clean initialization of the Vulkan renderer, and very much cleans up the older renderer initialization code as gl and sw are no longer intertwined.
2019-07-08 16:00:47 +00:00
VkSurfaceKHR (*create_surface) (struct vulkan_ctx_s *ctx);
Remove global vulkan_ctx While I can't say that I'm happy with the details of vulkan_ctx_t, I am pretty sure I don't want to be limited to having only one.
2019-07-09 02:54:23 +00:00
[util] Make va thread-safe It now takes a context pointer (opaque data) that holds the buffers it uses for the temporary strings. If the context pointer is null, a static context is used (making those uses of va NOT thread-safe). Most calls to va use the static context, but all such calls have been formatted consistently so they are easy to find when it comes time to do a full audit.
2021-01-31 07:01:20 +00:00
struct va_ctx_s *va_ctx;
Completely rework the vulkan related api Things don't work yet, but this feels much cleaner.
2019-07-12 04:15:25 +00:00
struct qfv_instance_s *instance;
struct qfv_device_s *device;
struct qfv_swapchain_s *swapchain;
[vulkan] Ensure msaa settings are consistent I had forgotten that msaa samples was governed by the driver (as a max) and the renderpass setup code simply took the max. Thus why 1 vs 8 caused the display to render incorrectly.
2020-12-21 11:17:27 +00:00
VkSampleCountFlagBits msaaSamples; // FIXME not here?
[vulkan] Hook up the expression parser The pipeline parser still isn't hooked up yet as something isn't quite right, but it seems all the parsing works.
2020-12-21 09:38:31 +00:00
struct hashlink_s *hashlinks; //FIXME want per thread
Completely rework the vulkan related api Things don't work yet, but this feels much cleaner.
2019-07-12 04:15:25 +00:00
VkSurfaceKHR surface; //FIXME surface = window, so "contains" swapchain
[vulkan] Remove renderpass from generated code The creation of a render pass seems to need special handling.
2021-01-05 08:43:26 +00:00
struct plitem_s *pipelineDef;
[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
[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
uint32_t swapImageIndex;
[vulkan] Generate code for resource nodes This makes it easy to add resource nodes defining contextually named resources. It is already used for shaders, set layouts, and pipeline layouts.
2021-01-05 05:15:35 +00:00
struct hashtab_s *shaderModules;
[vulkan] Implement pipeline layout creation It seems to work when parsing the layouts. They can be created in-line (in theory) or in a "setLayouts" node and then referenced by name.
2021-01-04 08:36:11 +00:00
struct hashtab_s *setLayouts;
[vulkan] Generate code for vulkan handles The handles can be created in place or in resource blocks and referred to by name (resource block code gen next).
2021-01-04 23:38:35 +00:00
struct hashtab_s *pipelineLayouts;
[vulkan] Parse descriptor pools
2021-01-09 16:53:15 +00:00
struct hashtab_s *descriptorPools;
[vulkan] Make a start on the 2D pipeline Short wrappers for Draw functins are in vid_render_vulkan.c so the vulkan context can be passed on to the actual functions. The 2D shaders are set up similar to those in glsl, but with full 32-bit color (rgba) support instead of paletted. However, the textures are not loaded yet, nor is anything bound.
2021-01-10 06:52:27 +00:00
struct hashtab_s *samplers;
[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
struct hashtab_s *images;
struct hashtab_s *imageViews;
[vulkan] Parse clear values And get the render pass working in general. Note that this is only the render pass and frame buffers: actual commands still have problems.
2021-02-23 05:37:48 +00:00
struct hashtab_s *renderpasses;
Create command pool and frame timing fences Nothing is actually done yet, so the reported fps is around 172k (yes, k), but startup and shutdown seems to be clean (yay validation layers).
2020-02-16 13:45:27 +00:00
[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
struct texturectx_s *texture_context;
[vulkan] Move matrices into per-frame buffers This should fix the horrid frame rate dependent behavior of the view model. They are also in their own descriptor set so they can be easily shared between pipelines. This has been verified to work for Draw.
2021-12-08 06:28:21 +00:00
struct matrixctx_s *matrix_context;
[vulkan] Start work on alias model rendering
2021-01-26 11:58:24 +00:00
struct aliasctx_s *alias_context;
[vulkan] Make a lot of progress for brush models Light maps are maybe updating, but as nothing is actually rendered yet, it's hard to tell.
2021-01-19 16:25:54 +00:00
struct bspctx_s *bsp_context;
[vulkan] Add support for IQM models Despite the base IQM specification not supporting blend-shapes, I think IQM will become the basis for QF's generic model representation (at least for the more advanced renderers). After my experience with .mu models (KSP) and unity mesh objects (both normal and skinned), and reviewing the IQM spec, it looks like with the addition of support for blend-shapes, IQM is actually pretty good. This is just the preliminary work to get standard IQM models loading in vulkan (seems to work, along with unloading), and they very basics into the renderer (most likely not working: not tested yet). The rest of the renderer seems to be unaffected, though, which is good.
2022-05-04 05:07:27 +00:00
struct iqmctx_s *iqm_context;
[vulkan] Start work on particles for Vulkan This gets the pipelines loaded (and unloaded on shutdown). Probably the easy part :P. Still need to sort out the command buffers, synchronization, and particle generation (and probably a bunch else that's not coming to mind).
2021-12-16 14:17:01 +00:00
struct particlectx_s *particle_context;
[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
struct spritectx_s *sprite_context;
[vulkan] Rework draw to use a context struct Cleans up global space and makes it usable in multiple contexts. Also, max quads dropped to 32k as each frame now has its own vertex buffer to avoid issues with vertex overwrites (which I have seen). However, all vertex buffers are in the one memory/buffer object (using offsets) and the index buffer has been moved into a device-local memory object.
2021-01-16 05:42:25 +00:00
struct drawctx_s *draw_context;
[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
struct lightingctx_s *lighting_context;
struct composectx_s *compose_context;
Rework much of the Vulkan array handling So much for all those little wrappers hiding the device. Some survived, but mostly just the bigger things like device, swapchain, etc.
2020-02-17 11:29:35 +00:00
VkCommandPool cmdpool;
Create and destroy color/depth resources
2020-02-18 05:28:28 +00:00
VkCommandBuffer cmdbuffer;
VkFence fence; // for ctx->cmdbuffer only
[vulkan] Use only one global staging buffer I think I did two as a bit of a ring buffer, but the new ring buffer system used inside a staging buffer makes it less necessary. Also, the staging buffer is now a fair bit bigger (4M is probably not really enough)
2021-01-16 05:37:16 +00:00
struct qfv_stagebuf_s *staging;
Create and destroy render pass and frame buffers
2020-02-18 08:18:37 +00:00
size_t curFrame;
[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
vulkan_frameset_t frames;
[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_renderpassset_t renderPasses;
Create command pool and frame timing fences Nothing is actually done yet, so the reported fps is around 172k (yes, k), but startup and shutdown seems to be clean (yay validation layers).
2020-02-16 13:45:27 +00:00
[vulkan] Implement screenshot Finally, I can brag about my progress on irc :)
2021-03-24 10:20:53 +00:00
struct qfv_capture_s *capture;
void (*capture_callback) (const byte *data, int width, int height);
[renderer] Make screen capture support asynchronous operation This fixes (*ahem*) the vulkan renderer segfaulting when attempting to take a screenshot. However, the image is upside down. Also, remote snapshots and demo capture are broken for the moment.
2022-03-31 10:07:15 +00:00
// make a queue?
void *capture_complete;// really capfunc_t
void *capture_complete_data;
[vulkan] Implement screenshot Finally, I can brag about my progress on irc :)
2021-03-24 10:20:53 +00:00
[vulkan] Create some default textures Black, white and magenta. This gets glow (full bright) maps working.
2021-01-23 06:40:31 +00:00
struct qfv_tex_s *default_black;
struct qfv_tex_s *default_white;
struct qfv_tex_s *default_magenta;
[vulkan] Add an array view for default_magenta The validation layers get very picky about image view types (rightfully so, I imagine).
2021-12-08 12:08:37 +00:00
struct qfv_tex_s *default_magenta_array;
[vulkan] Create some default textures Black, white and magenta. This gets glow (full bright) maps working.
2021-01-23 06:40:31 +00:00
[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
// size of window
int window_width;
int window_height;
[vulkan] Add some matrix buffers It seems they could all be in the one buffer: there are indications that uniform binding can be fairly fine-grained. I need to investigate that.
2021-01-12 02:26:20 +00:00
[vulkan] Break render pass parsing away from swapchain This allows a single render pass description to be used for both on-screen and off-screen targets. While Vulkan does allow a VkRenderPass to be used with any compatible frame buffer, and vkparse caches a VkRenderPass created from the same description, this allows the same description to be used for a compatible off-screen target without any dependence on the swapchain. However, there is a problem in the caching when it comes to targeting outputs with different formats.
2022-04-02 02:42:36 +00:00
//FIXME not sure I like it being here (also, type name)
qfv_output_t output;
Separate render and presentation initialization This paves the way for clean initialization of the Vulkan renderer, and very much cleans up the older renderer initialization code as gl and sw are no longer intertwined.
2019-07-08 16:00:47 +00:00
#define EXPORTED_VULKAN_FUNCTION(fname) PFN_##fname fname;
#define GLOBAL_LEVEL_VULKAN_FUNCTION(fname) PFN_##fname fname;
#include "QF/Vulkan/funclist.h"
} vulkan_ctx_t;
[vulkan] Add a debug stack to the vulkan instance The stack is arbitrary strings that the validation layer debug callback prints in reverse order after each message. This makes it easy to work out what nodes in a pipeline/render pass plist are causing validation errors. Still have to narrow down the actual line, but the messages seem to help with that. Putting qfvPushDebug/qfvPopDebug around other calls to vulkan should help out a lot, tool. As a bonus, the stack is printed before debug_breakpoint is called, so it's immediately visible in gdb.
2021-11-30 03:24:45 +00:00
#define qfvPushDebug(ctx, x) \
do { \
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
if (developer & SYS_vulkan) { \
[vulkan] Add a debug stack to the vulkan instance The stack is arbitrary strings that the validation layer debug callback prints in reverse order after each message. This makes it easy to work out what nodes in a pipeline/render pass plist are causing validation errors. Still have to narrow down the actual line, but the messages seem to help with that. Putting qfvPushDebug/qfvPopDebug around other calls to vulkan should help out a lot, tool. As a bonus, the stack is printed before debug_breakpoint is called, so it's immediately visible in gdb.
2021-11-30 03:24:45 +00:00
DARRAY_APPEND(&(ctx)->instance->debug_stack, (x)); \
} \
} while (0)
#define qfvPopDebug(ctx) \
do { \
[cvar] Make cvars properly typed This is an extremely extensive patch as it hits every cvar, and every usage of the cvars. Cvars no longer store the value they control, instead, they use a cexpr value object to reference the value and specify the value's type (currently, a null type is used for strings). Non-string cvars are passed through cexpr, allowing expressions in the cvars' settings. Also, cvars have returned to an enhanced version of the original (id quake) registration scheme. As a minor benefit, relevant code having direct access to the cvar-controlled variables is probably a slight optimization as it removed a pointer dereference, and the variables can be located for data locality. The static cvar descriptors are made private as an additional safety layer, though there's nothing stopping external modification via Cvar_FindVar (which is needed for adding listeners). While not used yet (partly due to working out the design), cvars can have a validation function. Registering a cvar allows a primary listener (and its data) to be specified: it will always be called first when the cvar is modified. The combination of proper listeners and direct access to the controlled variable greatly simplifies the more complex cvar interactions as much less null checking is required, and there's no need for one cvar's callback to call another's. nq-x11 is known to work at least well enough for the demos. More testing will come.
2022-04-23 03:22:45 +00:00
if (developer & SYS_vulkan) { \
[vulkan] Add a debug stack to the vulkan instance The stack is arbitrary strings that the validation layer debug callback prints in reverse order after each message. This makes it easy to work out what nodes in a pipeline/render pass plist are causing validation errors. Still have to narrow down the actual line, but the messages seem to help with that. Putting qfvPushDebug/qfvPopDebug around other calls to vulkan should help out a lot, tool. As a bonus, the stack is printed before debug_breakpoint is called, so it's immediately visible in gdb.
2021-11-30 03:24:45 +00:00
__auto_type ds = &(ctx)->instance->debug_stack; \
DARRAY_REMOVE_AT(ds, ds->size - 1); \
} \
} while (0)
Separate render and presentation initialization This paves the way for clean initialization of the Vulkan renderer, and very much cleans up the older renderer initialization code as gl and sw are no longer intertwined.
2019-07-08 16:00:47 +00:00
#endif//__vid_vulkan_h