vk: Improve driver compatibility

- Add workarounds to disable compression on mesa drivers
- Separate intel proprietary vs mesa into two enumerants

rpcs3/Emu/RSX/VK/VKCompute.cpp

@@ -75,6 +75,7 @@ namespace vk
 			{
 			case vk::driver_vendor::unknown:
 			case vk::driver_vendor::INTEL:
+			case vk::driver_vendor::ANV:
 				// Intel hw has 8 threads, but LDS allocation behavior makes optimal group size between 64 and 256
 				// Based on intel's own OpenCL recommended settings
 				unroll_loops = true;

rpcs3/Emu/RSX/VK/VKGSRender.cpp

@@ -583,7 +583,7 @@ VKGSRender::VKGSRender() : GSRender()
 		case vk::driver_vendor::AMD:
 #if !defined(__linux__)
 			// Intel chipsets would fail on BSD in most cases and DRM_IOCTL_i915_GEM_USERPTR unimplemented
-		case vk::driver_vendor::INTEL:
+		case vk::driver_vendor::ANV:
 #endif
 			if (backend_config.supports_passthrough_dma)
 			{

rpcs3/Emu/RSX/VK/VKHelpers.cpp

@@ -121,6 +121,9 @@ namespace vk
 			g_drv_sanitize_fp_values = true;
 			break;
 		case driver_vendor::INTEL:
+		case driver_vendor::ANV:
+			// INTEL vulkan drivers are mostly OK, workarounds are applied when creating the device
+			break;
 		default:
 			rsx_log.warning("Unsupported device: %s", gpu_name);
 		}

rpcs3/Emu/RSX/VK/VKRenderTargets.h

@@ -89,6 +89,35 @@ namespace vk
 		using download_buffer_object = void*;
 		using barrier_descriptor_t = rsx::deferred_clipped_region<vk::render_target*>;
 
+		static VkFlags get_attachment_compression_usage_flags()
+		{
+			if (g_cfg.video.strict_rendering_mode)
+			{
+				return 0;
+			}
+
+			switch (vk::get_driver_vendor())
+			{
+			case driver_vendor::NVIDIA:
+			case driver_vendor::INTEL:
+			case driver_vendor::AMD: // TODO
+				return 0;
+
+			// Workaround to force transition to GENERAL to decompress.
+			// Fixes corruption in FBO loops for ANV and RADV.
+			case driver_vendor::ANV:
+				return VK_IMAGE_USAGE_STORAGE_BIT;
+			case driver_vendor::RADV:
+				// Only needed for GFX10+
+				return (vk::get_chip_family() >= chip_class::AMD_navi1x) ?
+					VK_IMAGE_USAGE_STORAGE_BIT : 0;
+
+			default:
+				rsx_log.error("Unknown driver vendor!");
+				return 0;
+			}
+		}
+
 		static std::unique_ptr<vk::render_target> create_new_surface(
 			u32 address,
 			rsx::surface_color_format format,
@@ -115,7 +144,7 @@ namespace vk
 			VkImageUsageFlags usage_flags = VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT | VK_IMAGE_USAGE_TRANSFER_DST_BIT;
 			if (samples == 1) [[likely]]
 			{
-				usage_flags |= VK_IMAGE_USAGE_TRANSFER_SRC_BIT | VK_IMAGE_USAGE_SAMPLED_BIT;
+				usage_flags |= get_attachment_compression_usage_flags() | VK_IMAGE_USAGE_TRANSFER_SRC_BIT | VK_IMAGE_USAGE_SAMPLED_BIT;
 			}
 			else
 			{
@@ -180,7 +209,7 @@ namespace vk
 
 			if (samples == 1) [[likely]]
 			{
-				usage_flags |= VK_IMAGE_USAGE_TRANSFER_SRC_BIT;
+				usage_flags |= get_attachment_compression_usage_flags() | VK_IMAGE_USAGE_TRANSFER_SRC_BIT;
 			}
 
 			std::unique_ptr<vk::render_target> ds;

rpcs3/Emu/RSX/VK/vkutils/chip_class.h

@@ -29,7 +29,8 @@ namespace vk
 		AMD,
 		NVIDIA,
 		RADV,
-		INTEL
+		INTEL,
+		ANV
 	};
 
 	driver_vendor get_driver_vendor();

rpcs3/Emu/RSX/VK/vkutils/device.cpp

@@ -128,7 +128,11 @@ namespace vk
 
 			if (gpu_name.find("Intel") != umax)
 			{
+#ifdef _WIN32
 				return driver_vendor::INTEL;
+#else
+				return driver_vendor::ANV;
+#endif
 			}
 
 			return driver_vendor::unknown;
@@ -145,10 +149,11 @@ namespace vk
 			case VK_DRIVER_ID_NVIDIA_PROPRIETARY_KHR:
 				return driver_vendor::NVIDIA;
 			case VK_DRIVER_ID_INTEL_PROPRIETARY_WINDOWS_KHR:
-			case VK_DRIVER_ID_INTEL_OPEN_SOURCE_MESA_KHR:
 				return driver_vendor::INTEL;
+			case VK_DRIVER_ID_INTEL_OPEN_SOURCE_MESA_KHR:
+				return driver_vendor::ANV;
 			default:
-				// Mobile
+				// Mobile?
 				return driver_vendor::unknown;
 			}
 		}

rpcs3/Emu/RSX/VK/vkutils/swapchain.hpp

@@ -520,11 +520,9 @@ namespace vk
 				switch (gpu.get_driver_vendor())
 				{
 				case driver_vendor::AMD:
-					break;
 				case driver_vendor::INTEL:
-	#ifdef _WIN32
 					break;
-	#endif
+				case driver_vendor::ANV:
 				case driver_vendor::NVIDIA:
 				case driver_vendor::RADV:
 					m_wm_reports_flag = true;