renderer_vulkan: Prefer depth stencil read-only layout when possible

* Persona reads a depth attachment while it is being attached with writes disabled. Now this works without spamming vk validation errors
This commit is contained in:
IndecisiveTurtle 2024-07-01 02:11:53 +03:00 committed by TheTurtle
parent 22b930ba5e
commit b4d24d8737
15 changed files with 106 additions and 65 deletions

View File

@ -26,7 +26,7 @@ Id EmitImageSampleImplicitLod(EmitContext& ctx, IR::Inst* inst, u32 handle, Id c
const Id sampled_image = ctx.OpSampledImage(texture.sampled_type, image, sampler);
ImageOperands operands;
if (Sirit::ValidId(offset)) {
operands.Add(spv::ImageOperandsMask::Offset, offset);
operands.Add(spv::ImageOperandsMask::ConstOffset, offset);
}
return ctx.OpImageSampleImplicitLod(ctx.F32[4], sampled_image, coords, operands.mask, operands.operands);
}

View File

@ -528,6 +528,14 @@ struct Liverpool {
BitField<0, 15, s32> bottom_right_x;
BitField<15, 15, s32> bottom_right_y;
};
u32 GetWidth() const {
return bottom_right_x - top_left_x;
}
u32 GetHeight() const {
return bottom_right_y - top_left_y;
}
};
struct ViewportDepth {

View File

@ -41,7 +41,8 @@ public:
Frame* PrepareFrame(const Libraries::VideoOut::BufferAttributeGroup& attribute,
VAddr cpu_address) {
const auto info = VideoCore::ImageInfo{attribute};
auto& image = texture_cache.FindImage(info, cpu_address);
const auto image_id = texture_cache.FindImage(info, cpu_address);
auto& image = texture_cache.GetImage(image_id);
return PrepareFrameInternal(image);
}
@ -54,7 +55,8 @@ public:
const Libraries::VideoOut::BufferAttributeGroup& attribute, VAddr cpu_address) {
vo_buffers_addr.emplace_back(cpu_address);
const auto info = VideoCore::ImageInfo{attribute};
return texture_cache.FindImage(info, cpu_address);
const auto image_id = texture_cache.FindImage(info, cpu_address);
return texture_cache.GetImage(image_id);
}
bool IsVideoOutSurface(const AmdGpu::Liverpool::ColorBuffer& color_buffer) {

View File

@ -125,17 +125,18 @@ bool ComputePipeline::BindResources(Core::MemoryManager* memory, StreamBuffer& s
}
}
for (const auto& image : info.images) {
const auto tsharp = info.ReadUd<AmdGpu::Image>(image.sgpr_base, image.dword_offset);
const auto& image_view = texture_cache.FindImageView(tsharp, image.is_storage, image.is_depth);
image_infos.emplace_back(VK_NULL_HANDLE, *image_view.image_view, vk::ImageLayout::eGeneral);
for (const auto& image_desc : info.images) {
const auto tsharp = info.ReadUd<AmdGpu::Image>(image_desc.sgpr_base, image_desc.dword_offset);
const auto& image_view = texture_cache.FindImageView(tsharp, image_desc.is_storage);
const auto& image = texture_cache.GetImage(image_view.image_id);
image_infos.emplace_back(VK_NULL_HANDLE, *image_view.image_view, image.layout);
set_writes.push_back({
.dstSet = VK_NULL_HANDLE,
.dstBinding = binding++,
.dstArrayElement = 0,
.descriptorCount = 1,
.descriptorType = image.is_storage ? vk::DescriptorType::eStorageImage
: vk::DescriptorType::eSampledImage,
.descriptorType = image_desc.is_storage ? vk::DescriptorType::eStorageImage
: vk::DescriptorType::eSampledImage,
.pImageInfo = &image_infos.back(),
});

View File

@ -348,19 +348,18 @@ void GraphicsPipeline::BindResources(Core::MemoryManager* memory, StreamBuffer&
}
}
for (const auto& image : stage.images) {
const auto tsharp = stage.ReadUd<AmdGpu::Image>(image.sgpr_base, image.dword_offset);
const auto& image_view = texture_cache.FindImageView(tsharp, image.is_storage, image.is_depth);
image_infos.emplace_back(VK_NULL_HANDLE, *image_view.image_view,
(image.is_storage || image.is_depth) ? vk::ImageLayout::eGeneral
: vk::ImageLayout::eShaderReadOnlyOptimal);
for (const auto& image_desc : stage.images) {
const auto tsharp = stage.ReadUd<AmdGpu::Image>(image_desc.sgpr_base, image_desc.dword_offset);
const auto& image_view = texture_cache.FindImageView(tsharp, image_desc.is_storage);
const auto& image = texture_cache.GetImage(image_view.image_id);
image_infos.emplace_back(VK_NULL_HANDLE, *image_view.image_view, image.layout);
set_writes.push_back({
.dstSet = VK_NULL_HANDLE,
.dstBinding = binding++,
.dstArrayElement = 0,
.descriptorCount = 1,
.descriptorType = image.is_storage ? vk::DescriptorType::eStorageImage
: vk::DescriptorType::eSampledImage,
.descriptorType = image_desc.is_storage ? vk::DescriptorType::eStorageImage
: vk::DescriptorType::eSampledImage,
.pImageInfo = &image_infos.back(),
});

View File

@ -203,6 +203,7 @@ bool Instance::CreateDevice() {
.independentBlend = true,
.geometryShader = features.geometryShader,
.logicOp = features.logicOp,
.multiViewport = true,
.samplerAnisotropy = features.samplerAnisotropy,
.fragmentStoresAndAtomics = features.fragmentStoresAndAtomics,
.shaderImageGatherExtended = true,

View File

@ -44,10 +44,11 @@ void Rasterizer::Draw(bool is_indexed, u32 index_offset) {
return;
}
UpdateDynamicState(*pipeline);
pipeline->BindResources(memory, vertex_index_buffer, texture_cache);
BeginRendering();
UpdateDynamicState(*pipeline);
cmdbuf.bindPipeline(vk::PipelineBindPoint::eGraphics, pipeline->Handle());
if (is_indexed) {
@ -113,12 +114,14 @@ void Rasterizer::BeginRendering() {
const bool is_clear = regs.depth_render_control.depth_clear_enable ||
texture_cache.IsMetaCleared(htile_address);
const auto& hint = liverpool->last_db_extent;
const auto& image_view = texture_cache.DepthTarget(regs.depth_buffer, htile_address, hint);
const auto& image_view = texture_cache.DepthTarget(regs.depth_buffer, htile_address, hint,
regs.depth_control.depth_write_enable);
const auto& image = texture_cache.GetImage(image_view.image_id);
state.width = std::min<u32>(state.width, hint.width);
state.height = std::min<u32>(state.height, hint.height);
state.depth_attachment = {
.imageView = *image_view.image_view,
.imageLayout = vk::ImageLayout::eGeneral,
.imageLayout = image.layout,
.loadOp = is_clear ? vk::AttachmentLoadOp::eClear : vk::AttachmentLoadOp::eLoad,
.storeOp = is_clear ? vk::AttachmentStoreOp::eNone : vk::AttachmentStoreOp::eStore,
.clearValue = vk::ClearValue{.depthStencil = {.depth = regs.depth_clear,
@ -192,23 +195,34 @@ void Rasterizer::UpdateDynamicState(const GraphicsPipeline& pipeline) {
void Rasterizer::UpdateViewportScissorState() {
auto& regs = liverpool->regs;
boost::container::static_vector<vk::Viewport, Liverpool::NumViewports> viewports;
boost::container::static_vector<vk::Rect2D, Liverpool::NumViewports> scissors;
const float reduce_z =
regs.clipper_control.clip_space == AmdGpu::Liverpool::ClipSpace::MinusWToW ? 1.0f : 0.0f;
for (u32 i = 0; i < Liverpool::NumViewports; i++) {
const auto& vp = regs.viewports[i];
const auto& vp_d = regs.viewport_depths[i];
if (vp.xscale == 0) {
continue;
}
viewports.push_back({
.x = vp.xoffset - vp.xscale,
.y = vp.yoffset - vp.yscale,
.width = vp.xscale * 2.0f,
.height = vp.yscale * 2.0f,
.minDepth = vp.zoffset - vp.zscale * reduce_z,
.maxDepth = vp.zscale + vp.zoffset,
});
}
const auto& sc = regs.screen_scissor;
scissors.push_back({
.offset = {sc.top_left_x, sc.top_left_y},
.extent = {sc.GetWidth(), sc.GetHeight()},
});
const auto cmdbuf = scheduler.CommandBuffer();
const vk::Viewport viewport{
.x = regs.viewports[0].xoffset - regs.viewports[0].xscale,
.y = regs.viewports[0].yoffset - regs.viewports[0].yscale,
.width = regs.viewports[0].xscale * 2.0f,
.height = regs.viewports[0].yscale * 2.0f,
.minDepth = regs.viewports[0].zoffset - regs.viewports[0].zscale * reduce_z,
.maxDepth = regs.viewports[0].zscale + regs.viewports[0].zoffset,
};
const vk::Rect2D scissor{
.offset = {regs.screen_scissor.top_left_x, regs.screen_scissor.top_left_y},
.extent = {regs.screen_scissor.GetWidth(), regs.screen_scissor.GetHeight()},
};
cmdbuf.setViewport(0, viewport);
cmdbuf.setScissor(0, scissor);
cmdbuf.setViewport(0, viewports);
cmdbuf.setScissor(0, scissors);
}
void Rasterizer::UpdateDepthStencilState() {

View File

@ -32,7 +32,7 @@ void Scheduler::BeginRendering(const RenderState& new_state) {
.extent = {render_state.width, render_state.height},
},
.layerCount = 1,
.colorAttachmentCount = static_cast<u32>(render_state.color_attachments.size()),
.colorAttachmentCount = render_state.num_color_attachments,
.pColorAttachments = render_state.color_attachments.data(),
.pDepthAttachment = render_state.num_depth_attachments ?
&render_state.depth_attachment : nullptr,

View File

@ -23,7 +23,7 @@ struct RenderState {
bool operator==(const RenderState& other) const noexcept {
return std::memcmp(this, &other, sizeof(RenderState)) == 0;
}
}
};
class Scheduler {
@ -46,6 +46,11 @@ public:
/// Ends current rendering scope.
void EndRendering();
/// Returns the current render state.
const RenderState& GetRenderState() const {
return render_state;
}
/// Returns the current command buffer.
vk::CommandBuffer CommandBuffer() const {
return current_cmdbuf;

View File

@ -226,7 +226,7 @@ void StreamBuffer::WaitPendingOperations(u64 requested_upper_bound) {
while (requested_upper_bound > wait_bound && wait_cursor < *invalidation_mark) {
auto& watch = previous_watches[wait_cursor];
wait_bound = watch.upper_bound;
scheduler.Wait(watch.tick);
//scheduler.Wait(watch.tick);
++wait_cursor;
}
}

View File

@ -267,7 +267,7 @@ Image::Image(const Vulkan::Instance& instance_, Vulkan::Scheduler& scheduler_,
if (info.is_tiled) {
ImageViewInfo view_info;
view_info.format = DemoteImageFormatForDetiling(info.pixel_format);
view_for_detiler.emplace(*instance, view_info, *this);
view_for_detiler.emplace(*instance, view_info, *this, ImageId{});
}
Transit(vk::ImageLayout::eGeneral, vk::AccessFlagBits::eNone);

View File

@ -71,8 +71,9 @@ ImageViewInfo::ImageViewInfo(const AmdGpu::Liverpool::ColorBuffer& col_buffer,
}
ImageView::ImageView(const Vulkan::Instance& instance, const ImageViewInfo& info_, Image& image,
ImageId image_id_,
std::optional<vk::ImageUsageFlags> usage_override /*= {}*/)
: info{info_} {
: info{info_}, image_id{image_id_} {
vk::ImageViewUsageCreateInfo usage_ci{};
if (usage_override) {
usage_ci.usage = usage_override.value();

View File

@ -36,6 +36,7 @@ struct Image;
struct ImageView {
explicit ImageView(const Vulkan::Instance& instance, const ImageViewInfo& info, Image& image,
ImageId image_id,
std::optional<vk::ImageUsageFlags> usage_override = {});
~ImageView();

View File

@ -93,7 +93,7 @@ TextureCache::TextureCache(const Vulkan::Instance& instance_, Vulkan::Scheduler&
ASSERT(null_id.index == 0);
ImageViewInfo view_info;
void(slot_image_views.insert(instance, view_info, slot_images[null_id]));
void(slot_image_views.insert(instance, view_info, slot_images[null_id], null_id));
}
TextureCache::~TextureCache() {
@ -112,7 +112,7 @@ void TextureCache::OnCpuWrite(VAddr address) {
});
}
Image& TextureCache::FindImage(const ImageInfo& info, VAddr cpu_address, bool refresh_on_create) {
ImageId TextureCache::FindImage(const ImageInfo& info, VAddr cpu_address, bool refresh_on_create) {
std::unique_lock lock{m_page_table};
boost::container::small_vector<ImageId, 2> image_ids;
ForEachImageInRegion(cpu_address, info.guest_size_bytes, [&](ImageId image_id, Image& image) {
@ -140,16 +140,16 @@ Image& TextureCache::FindImage(const ImageInfo& info, VAddr cpu_address, bool re
RegisterMeta(info, image_id);
Image& image = slot_images[image_id];
if (True(image.flags & ImageFlagBits::CpuModified) &&
(!image_ids.empty() || refresh_on_create)) {
if (True(image.flags & ImageFlagBits::CpuModified) && refresh_on_create) {
RefreshImage(image);
TrackImage(image, image_id);
}
return image;
return image_id;
}
ImageView& TextureCache::RegisterImageView(Image& image, const ImageViewInfo& view_info) {
ImageView& TextureCache::RegisterImageView(ImageId image_id, const ImageViewInfo& view_info) {
Image& image = slot_images[image_id];
if (const ImageViewId view_id = image.FindView(view_info); view_id) {
return slot_image_views[view_id];
}
@ -162,32 +162,37 @@ ImageView& TextureCache::RegisterImageView(Image& image, const ImageViewInfo& vi
usage_override = image.usage & ~vk::ImageUsageFlagBits::eStorage;
}
const ImageViewId view_id = slot_image_views.insert(instance, view_info, image, usage_override);
const ImageViewId view_id = slot_image_views.insert(instance, view_info, image, image_id, usage_override);
image.image_view_infos.emplace_back(view_info);
image.image_view_ids.emplace_back(view_id);
return slot_image_views[view_id];
}
ImageView& TextureCache::FindImageView(const AmdGpu::Image& desc, bool is_storage, bool is_depth) {
ImageView& TextureCache::FindImageView(const AmdGpu::Image& desc, bool is_storage) {
const ImageInfo info{desc};
Image& image = FindImage(info, desc.Address());
const ImageId image_id = FindImage(info, desc.Address());
Image& image = slot_images[image_id];
auto& usage = image.info.usage;
if (is_storage || is_depth) {
if (is_storage) {
image.Transit(vk::ImageLayout::eGeneral, vk::AccessFlagBits::eShaderWrite);
image.info.usage.storage = true;
usage.storage = true;
} else {
image.Transit(vk::ImageLayout::eShaderReadOnlyOptimal, vk::AccessFlagBits::eShaderRead);
image.info.usage.texture = true;
const auto new_layout = image.info.IsDepthStencil() ? vk::ImageLayout::eDepthStencilReadOnlyOptimal
: vk::ImageLayout::eShaderReadOnlyOptimal;
image.Transit(new_layout, vk::AccessFlagBits::eShaderRead);
usage.texture = true;
}
const ImageViewInfo view_info{desc, is_storage};
return RegisterImageView(image, view_info);
return RegisterImageView(image_id, view_info);
}
ImageView& TextureCache::RenderTarget(const AmdGpu::Liverpool::ColorBuffer& buffer,
const AmdGpu::Liverpool::CbDbExtent& hint) {
const ImageInfo info{buffer, hint};
auto& image = FindImage(info, buffer.Address(), false);
const ImageId image_id = FindImage(info, buffer.Address(), false);
Image& image = slot_images[image_id];
image.flags &= ~ImageFlagBits::CpuModified;
image.Transit(vk::ImageLayout::eColorAttachmentOptimal,
@ -197,17 +202,21 @@ ImageView& TextureCache::RenderTarget(const AmdGpu::Liverpool::ColorBuffer& buff
image.info.usage.render_target = true;
ImageViewInfo view_info{buffer, !!image.info.usage.vo_buffer};
return RegisterImageView(image, view_info);
return RegisterImageView(image_id, view_info);
}
ImageView& TextureCache::DepthTarget(const AmdGpu::Liverpool::DepthBuffer& buffer,
VAddr htile_address,
const AmdGpu::Liverpool::CbDbExtent& hint) {
const AmdGpu::Liverpool::CbDbExtent& hint,
bool write_enabled) {
const ImageInfo info{buffer, htile_address, hint};
auto& image = FindImage(info, buffer.Address(), false);
const ImageId image_id = FindImage(info, buffer.Address(), false);
Image& image = slot_images[image_id];
image.flags &= ~ImageFlagBits::CpuModified;
image.Transit(vk::ImageLayout::eGeneral,
const auto new_layout = write_enabled ? vk::ImageLayout::eDepthStencilAttachmentOptimal
: vk::ImageLayout::eDepthStencilReadOnlyOptimal;
image.Transit(new_layout,
vk::AccessFlagBits::eDepthStencilAttachmentWrite |
vk::AccessFlagBits::eDepthStencilAttachmentRead);
@ -215,7 +224,7 @@ ImageView& TextureCache::DepthTarget(const AmdGpu::Liverpool::DepthBuffer& buffe
ImageViewInfo view_info;
view_info.format = info.pixel_format;
return RegisterImageView(image, view_info);
return RegisterImageView(image_id, view_info);
}
void TextureCache::RefreshImage(Image& image) {

View File

@ -48,19 +48,19 @@ public:
void OnCpuWrite(VAddr address);
/// Retrieves the image handle of the image with the provided attributes and address.
[[nodiscard]] Image& FindImage(const ImageInfo& info, VAddr cpu_address,
bool refresh_on_create = true);
[[nodiscard]] ImageId FindImage(const ImageInfo& info, VAddr cpu_address,
bool refresh_on_create = true);
/// Retrieves an image view with the properties of the specified image descriptor.
[[nodiscard]] ImageView& FindImageView(const AmdGpu::Image& image, bool is_storage,
bool is_depth);
[[nodiscard]] ImageView& FindImageView(const AmdGpu::Image& image, bool is_storage);
/// Retrieves the render target with specified properties
[[nodiscard]] ImageView& RenderTarget(const AmdGpu::Liverpool::ColorBuffer& buffer,
const AmdGpu::Liverpool::CbDbExtent& hint);
[[nodiscard]] ImageView& DepthTarget(const AmdGpu::Liverpool::DepthBuffer& buffer,
VAddr htile_address,
const AmdGpu::Liverpool::CbDbExtent& hint);
const AmdGpu::Liverpool::CbDbExtent& hint,
bool write_enabled);
/// Reuploads image contents.
void RefreshImage(Image& image);
@ -95,7 +95,7 @@ public:
}
private:
ImageView& RegisterImageView(Image& image, const ImageViewInfo& view_info);
ImageView& RegisterImageView(ImageId image_id, const ImageViewInfo& view_info);
/// Iterate over all page indices in a range
template <typename Func>