summaryrefslogtreecommitdiffstats
path: root/src/video_core
diff options
context:
space:
mode:
Diffstat (limited to 'src/video_core')
-rw-r--r--src/video_core/buffer_cache/buffer_cache.h20
-rw-r--r--src/video_core/buffer_cache/usage_tracker.h4
-rw-r--r--src/video_core/renderer_opengl/gl_buffer_cache.h4
-rw-r--r--src/video_core/renderer_opengl/gl_device.cpp18
-rw-r--r--src/video_core/renderer_opengl/gl_shader_cache.cpp2
-rw-r--r--src/video_core/renderer_vulkan/vk_buffer_cache.cpp4
-rw-r--r--src/video_core/renderer_vulkan/vk_buffer_cache.h2
-rw-r--r--src/video_core/renderer_vulkan/vk_pipeline_cache.cpp2
-rw-r--r--src/video_core/vulkan_common/vulkan_device.cpp34
9 files changed, 72 insertions, 18 deletions
diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h
index 90dbd352f..6d1fc3887 100644
--- a/src/video_core/buffer_cache/buffer_cache.h
+++ b/src/video_core/buffer_cache/buffer_cache.h
@@ -1753,15 +1753,25 @@ Binding BufferCache<P>::StorageBufferBinding(GPUVAddr ssbo_addr, u32 cbuf_index,
const u32 memory_layout_size = static_cast<u32>(gpu_memory->GetMemoryLayoutSize(gpu_addr));
return std::min(memory_layout_size, static_cast<u32>(8_MiB));
}();
- const std::optional<VAddr> cpu_addr = gpu_memory->GpuToCpuAddress(gpu_addr);
- if (!cpu_addr || size == 0) {
+ // Alignment only applies to the offset of the buffer
+ const u32 alignment = runtime.GetStorageBufferAlignment();
+ const GPUVAddr aligned_gpu_addr = Common::AlignDown(gpu_addr, alignment);
+ const u32 aligned_size = static_cast<u32>(gpu_addr - aligned_gpu_addr) + size;
+
+ const std::optional<VAddr> aligned_cpu_addr = gpu_memory->GpuToCpuAddress(aligned_gpu_addr);
+ if (!aligned_cpu_addr || size == 0) {
LOG_WARNING(HW_GPU, "Failed to find storage buffer for cbuf index {}", cbuf_index);
return NULL_BINDING;
}
- const VAddr cpu_end = Common::AlignUp(*cpu_addr + size, YUZU_PAGESIZE);
+ const std::optional<VAddr> cpu_addr = gpu_memory->GpuToCpuAddress(gpu_addr);
+ ASSERT_MSG(cpu_addr, "Unaligned storage buffer address not found for cbuf index {}",
+ cbuf_index);
+ // The end address used for size calculation does not need to be aligned
+ const VAddr cpu_end = Common::AlignUp(*cpu_addr + size, Core::Memory::YUZU_PAGESIZE);
+
const Binding binding{
- .cpu_addr = *cpu_addr,
- .size = is_written ? size : static_cast<u32>(cpu_end - *cpu_addr),
+ .cpu_addr = *aligned_cpu_addr,
+ .size = is_written ? aligned_size : static_cast<u32>(cpu_end - *aligned_cpu_addr),
.buffer_id = BufferId{},
};
return binding;
diff --git a/src/video_core/buffer_cache/usage_tracker.h b/src/video_core/buffer_cache/usage_tracker.h
index ab05fe415..5f8688d31 100644
--- a/src/video_core/buffer_cache/usage_tracker.h
+++ b/src/video_core/buffer_cache/usage_tracker.h
@@ -58,7 +58,7 @@ private:
void TrackPage(u64 page, u64 offset, u64 size) noexcept {
const size_t offset_in_page = offset % PAGE_BYTES;
const size_t first_bit = offset_in_page >> BYTES_PER_BIT_SHIFT;
- const size_t num_bits = std::min(size, PAGE_BYTES) >> BYTES_PER_BIT_SHIFT;
+ const size_t num_bits = std::min<size_t>(size, PAGE_BYTES) >> BYTES_PER_BIT_SHIFT;
const size_t mask = ~u64{0} >> (64 - num_bits);
pages[page] |= (~u64{0} & mask) << first_bit;
}
@@ -66,7 +66,7 @@ private:
bool IsPageUsed(u64 page, u64 offset, u64 size) const noexcept {
const size_t offset_in_page = offset % PAGE_BYTES;
const size_t first_bit = offset_in_page >> BYTES_PER_BIT_SHIFT;
- const size_t num_bits = std::min(size, PAGE_BYTES) >> BYTES_PER_BIT_SHIFT;
+ const size_t num_bits = std::min<size_t>(size, PAGE_BYTES) >> BYTES_PER_BIT_SHIFT;
const size_t mask = ~u64{0} >> (64 - num_bits);
const size_t mask2 = (~u64{0} & mask) << first_bit;
return (pages[page] & mask2) != 0;
diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.h b/src/video_core/renderer_opengl/gl_buffer_cache.h
index feccf06f9..000f29a82 100644
--- a/src/video_core/renderer_opengl/gl_buffer_cache.h
+++ b/src/video_core/renderer_opengl/gl_buffer_cache.h
@@ -191,6 +191,10 @@ public:
return device.CanReportMemoryUsage();
}
+ u32 GetStorageBufferAlignment() const {
+ return static_cast<u32>(device.GetShaderStorageBufferAlignment());
+ }
+
private:
static constexpr std::array PABO_LUT{
GL_VERTEX_PROGRAM_PARAMETER_BUFFER_NV, GL_TESS_CONTROL_PROGRAM_PARAMETER_BUFFER_NV,
diff --git a/src/video_core/renderer_opengl/gl_device.cpp b/src/video_core/renderer_opengl/gl_device.cpp
index 94258ccd0..993438a27 100644
--- a/src/video_core/renderer_opengl/gl_device.cpp
+++ b/src/video_core/renderer_opengl/gl_device.cpp
@@ -265,33 +265,33 @@ std::string Device::GetVendorName() const {
if (vendor_name == "Intel") {
// For Mesa, `Intel` is an overloaded vendor string that could mean crocus or iris.
// Simply return `INTEL` for those as well as the Windows driver.
- return "INTEL";
+ return "Intel";
}
if (vendor_name == "Intel Open Source Technology Center") {
- return "I965";
+ return "i965";
}
if (vendor_name == "Mesa Project") {
- return "I915";
+ return "i915";
}
if (vendor_name == "Mesa/X.org") {
// This vendor string is overloaded between llvmpipe, softpipe, and virgl, so just return
// MESA instead of one of those driver names.
- return "MESA";
+ return "Mesa";
}
if (vendor_name == "AMD") {
- return "RADEONSI";
+ return "RadeonSI";
}
if (vendor_name == "nouveau") {
- return "NOUVEAU";
+ return "Nouveau";
}
if (vendor_name == "X.Org") {
return "R600";
}
if (vendor_name == "Collabora Ltd") {
- return "ZINK";
+ return "Zink";
}
if (vendor_name == "Intel Corporation") {
- return "OPENSWR";
+ return "OpenSWR";
}
if (vendor_name == "Microsoft Corporation") {
return "D3D12";
@@ -300,7 +300,7 @@ std::string Device::GetVendorName() const {
// Mesa's tegra driver reports `NVIDIA`. Only present in this list because the default
// strategy would have returned `NVIDIA` here for this driver, the same result as the
// proprietary driver.
- return "TEGRA";
+ return "Tegra";
}
return vendor_name;
}
diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp
index 2888e0238..26f2d0ea7 100644
--- a/src/video_core/renderer_opengl/gl_shader_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp
@@ -232,6 +232,7 @@ ShaderCache::ShaderCache(RasterizerOpenGL& rasterizer_, Core::Frontend::EmuWindo
.has_gl_bool_ref_bug = device.HasBoolRefBug(),
.ignore_nan_fp_comparisons = true,
.gl_max_compute_smem_size = device.GetMaxComputeSharedMemorySize(),
+ .min_ssbo_alignment = device.GetShaderStorageBufferAlignment(),
},
host_info{
.support_float64 = true,
@@ -240,6 +241,7 @@ ShaderCache::ShaderCache(RasterizerOpenGL& rasterizer_, Core::Frontend::EmuWindo
.needs_demote_reorder = device.IsAmd(),
.support_snorm_render_buffer = false,
.support_viewport_index_layer = device.HasVertexViewportLayer(),
+ .min_ssbo_alignment = static_cast<u32>(device.GetShaderStorageBufferAlignment()),
.support_geometry_shader_passthrough = device.HasGeometryShaderPassthrough(),
.support_conditional_barrier = device.SupportsConditionalBarriers(),
} {
diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
index 7691cc2ba..5958f52f7 100644
--- a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
+++ b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
@@ -355,6 +355,10 @@ bool BufferCacheRuntime::CanReportMemoryUsage() const {
return device.CanReportMemoryUsage();
}
+u32 BufferCacheRuntime::GetStorageBufferAlignment() const {
+ return static_cast<u32>(device.GetStorageBufferAlignment());
+}
+
void BufferCacheRuntime::TickFrame(VideoCommon::SlotVector<Buffer>& slot_buffers) noexcept {
for (auto it = slot_buffers.begin(); it != slot_buffers.end(); it++) {
it->ResetUsageTracking();
diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.h b/src/video_core/renderer_vulkan/vk_buffer_cache.h
index 4416a902f..0b3fbd6d0 100644
--- a/src/video_core/renderer_vulkan/vk_buffer_cache.h
+++ b/src/video_core/renderer_vulkan/vk_buffer_cache.h
@@ -91,6 +91,8 @@ public:
bool CanReportMemoryUsage() const;
+ u32 GetStorageBufferAlignment() const;
+
[[nodiscard]] StagingBufferRef UploadStagingBuffer(size_t size);
[[nodiscard]] StagingBufferRef DownloadStagingBuffer(size_t size, bool deferred = false);
diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp
index 89b455bff..2a13b2a72 100644
--- a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp
+++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp
@@ -373,6 +373,7 @@ PipelineCache::PipelineCache(RasterizerVulkan& rasterizer_, const Device& device
driver_id == VK_DRIVER_ID_QUALCOMM_PROPRIETARY,
.has_broken_robust =
device.IsNvidia() && device.GetNvidiaArch() <= NvidiaArchitecture::Arch_Pascal,
+ .min_ssbo_alignment = device.GetStorageBufferAlignment(),
};
host_info = Shader::HostTranslateInfo{
@@ -383,6 +384,7 @@ PipelineCache::PipelineCache(RasterizerVulkan& rasterizer_, const Device& device
driver_id == VK_DRIVER_ID_AMD_PROPRIETARY || driver_id == VK_DRIVER_ID_AMD_OPEN_SOURCE,
.support_snorm_render_buffer = true,
.support_viewport_index_layer = device.IsExtShaderViewportIndexLayerSupported(),
+ .min_ssbo_alignment = static_cast<u32>(device.GetStorageBufferAlignment()),
.support_geometry_shader_passthrough = device.IsNvGeometryShaderPassthroughSupported(),
.support_conditional_barrier = device.SupportsConditionalBarriers(),
};
diff --git a/src/video_core/vulkan_common/vulkan_device.cpp b/src/video_core/vulkan_common/vulkan_device.cpp
index fde36a49c..188ceeed7 100644
--- a/src/video_core/vulkan_common/vulkan_device.cpp
+++ b/src/video_core/vulkan_common/vulkan_device.cpp
@@ -847,11 +847,41 @@ std::string Device::GetDriverName() const {
case VK_DRIVER_ID_NVIDIA_PROPRIETARY:
return "NVIDIA";
case VK_DRIVER_ID_INTEL_PROPRIETARY_WINDOWS:
- return "INTEL";
+ return "Intel";
case VK_DRIVER_ID_INTEL_OPEN_SOURCE_MESA:
return "ANV";
+ case VK_DRIVER_ID_IMAGINATION_PROPRIETARY:
+ return "PowerVR";
+ case VK_DRIVER_ID_QUALCOMM_PROPRIETARY:
+ return "Qualcomm";
+ case VK_DRIVER_ID_ARM_PROPRIETARY:
+ return "Mali";
+ case VK_DRIVER_ID_GOOGLE_SWIFTSHADER:
+ return "SwiftShader";
+ case VK_DRIVER_ID_BROADCOM_PROPRIETARY:
+ return "Broadcom";
case VK_DRIVER_ID_MESA_LLVMPIPE:
- return "LAVAPIPE";
+ return "Lavapipe";
+ case VK_DRIVER_ID_MOLTENVK:
+ return "MoltenVK";
+ case VK_DRIVER_ID_VERISILICON_PROPRIETARY:
+ return "Vivante";
+ case VK_DRIVER_ID_MESA_TURNIP:
+ return "Turnip";
+ case VK_DRIVER_ID_MESA_V3DV:
+ return "V3DV";
+ case VK_DRIVER_ID_MESA_PANVK:
+ return "PanVK";
+ case VK_DRIVER_ID_MESA_VENUS:
+ return "Venus";
+ case VK_DRIVER_ID_MESA_DOZEN:
+ return "Dozen";
+ case VK_DRIVER_ID_MESA_NVK:
+ return "NVK";
+ case VK_DRIVER_ID_IMAGINATION_OPEN_SOURCE_MESA:
+ return "PVR";
+ // case VK_DRIVER_ID_MESA_AGXV:
+ // return "Asahi";
default:
return properties.driver.driverName;
}
generated by cgit v1.2.3 (git 2.25.1) at 2024-07-19 21:55:12 +0200