fix(metal): fix buffer to texture copy

src/video_core/renderer_metal/mtl_staging_buffer_pool.cpp

@@ -2,6 +2,7 @@
 // SPDX-License-Identifier: GPL-2.0-or-later
 
 #include <algorithm>
+#include <iostream>
 #include <utility>
 #include <vector>
 
@@ -77,9 +78,34 @@ StagingBufferRef StagingBufferPool::GetStreamBuffer(size_t size) {
 
 StagingBufferRef StagingBufferPool::GetStagingBuffer(size_t size, MemoryUsage usage,
                                                      bool deferred) {
+    if (const std::optional<StagingBufferRef> ref = TryGetReservedBuffer(size, usage, deferred)) {
+        return *ref;
+    }
+
     return CreateStagingBuffer(size, usage, deferred);
 }
 
+std::optional<StagingBufferRef> StagingBufferPool::TryGetReservedBuffer(size_t size,
+                                                                        MemoryUsage usage,
+                                                                        bool deferred) {
+    StagingBuffers& cache_level = GetCache(usage)[Common::Log2Ceil64(size)];
+
+    // TODO: don't always return true
+    const auto is_free = [](const StagingBuffer& entry) { return true; };
+    auto& entries = cache_level.entries;
+    const auto hint_it = entries.begin() + cache_level.iterate_index;
+    auto it = std::find_if(entries.begin() + cache_level.iterate_index, entries.end(), is_free);
+    if (it == entries.end()) {
+        it = std::find_if(entries.begin(), hint_it, is_free);
+        if (it == hint_it) {
+            return std::nullopt;
+        }
+    }
+    cache_level.iterate_index = std::distance(entries.begin(), it) + 1;
+
+    return it->Ref();
+}
+
 StagingBufferRef StagingBufferPool::CreateStagingBuffer(size_t size, MemoryUsage usage,
                                                         bool deferred) {
     const u32 log2 = Common::Log2Ceil64(size);

src/video_core/renderer_metal/mtl_staging_buffer_pool.h

@@ -71,6 +71,9 @@ private:
 
     StagingBufferRef GetStagingBuffer(size_t size, MemoryUsage usage, bool deferred = false);
 
+    std::optional<StagingBufferRef> TryGetReservedBuffer(size_t size, MemoryUsage usage,
+                                                         bool deferred);
+
     StagingBufferRef CreateStagingBuffer(size_t size, MemoryUsage usage, bool deferred);
 
     StagingBuffersCache& GetCache(MemoryUsage usage);

src/video_core/renderer_metal/mtl_texture_cache.cpp

@@ -5,6 +5,7 @@
 #include <array>
 #include <span>
 #include <vector>
+
 #include <boost/container/small_vector.hpp>
 
 #include "common/bit_cast.h"
@@ -59,7 +60,8 @@ Image::Image(TextureCacheRuntime& runtime_, const ImageInfo& info, GPUVAddr gpu_
     : VideoCommon::ImageBase(info, gpu_addr_, cpu_addr_), runtime{&runtime_} {
     MTL::TextureDescriptor* texture_descriptor = MTL::TextureDescriptor::alloc()->init();
     // TODO: don't hardcode the format
-    texture_descriptor->setPixelFormat(MTL::PixelFormatRGBA8Unorm);
+    texture_descriptor->setPixelFormat(info.size.width == 1600 ? MTL::PixelFormatRGBA8Unorm
+                                                               : MTL::PixelFormatASTC_4x4_sRGB);
     texture_descriptor->setWidth(info.size.width);
     texture_descriptor->setHeight(info.size.height);
 

src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp

@@ -49,8 +49,8 @@ size_t GetStreamBufferSize(const Device& device) {
 StagingBufferPool::StagingBufferPool(const Device& device_, MemoryAllocator& memory_allocator_,
                                      Scheduler& scheduler_)
     : device{device_}, memory_allocator{memory_allocator_}, scheduler{scheduler_},
-      stream_buffer_size{GetStreamBufferSize(device)}, region_size{stream_buffer_size /
-                                                                   StagingBufferPool::NUM_SYNCS} {
+      stream_buffer_size{GetStreamBufferSize(device)},
+      region_size{stream_buffer_size / StagingBufferPool::NUM_SYNCS} {
     VkBufferCreateInfo stream_ci = {
         .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO,
         .pNext = nullptr,