radv: disable implicit sync for radv allocated bos v3

Implicit sync kicks in when a buffer is used by two different amdgpu
contexts simultaneously. Jobs that use explicit synchronization
mechanisms end up needlessly waiting to be scheduled for long periods
of time in order to achieve serialized execution.

This patch disables implicit synchronization for all radv allocations
except for wsi bos. The only systems that require implicit
synchronization are DRI2/3 and PRIME.

v2: mark wsi bos as RADV_MEM_IMPLICIT_SYNC
v3: Add drm version check (Bas)

Signed-off-by: Andres Rodriguez <andresx7@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

src/amd/vulkan/radv_device.c

@@ -2115,6 +2115,9 @@ VkResult radv_alloc_memory(VkDevice                        _device,
 	if (pAllocateInfo->memoryTypeIndex == RADV_MEM_TYPE_GTT_WRITE_COMBINE)
 		flags |= RADEON_FLAG_GTT_WC;
 
+	if (mem_flags & RADV_MEM_IMPLICIT_SYNC)
+		flags |= RADEON_FLAG_IMPLICIT_SYNC;
+
 	mem->bo = device->ws->buffer_create(device->ws, alloc_size, device->physical_device->rad_info.max_alignment,
 					       domain, flags);
 

src/amd/vulkan/radv_radeon_winsys.h

@@ -53,6 +53,7 @@ enum radeon_bo_flag { /* bitfield */
 	RADEON_FLAG_NO_CPU_ACCESS = (1 << 2),
 	RADEON_FLAG_VIRTUAL =       (1 << 3),
 	RADEON_FLAG_VA_UNCACHED =   (1 << 4),
+	RADEON_FLAG_IMPLICIT_SYNC = (1 << 5),
 };
 
 enum radeon_bo_usage { /* bitfield */

src/amd/vulkan/radv_wsi.c

@@ -194,7 +194,7 @@ radv_wsi_image_create(VkDevice device_h,
 		.image = image_h
 	};
 
-	result = radv_AllocateMemory(device_h,
+	result = radv_alloc_memory(device_h,
 				     &(VkMemoryAllocateInfo) {
 					     .sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO,
 					     .pNext = &ded_alloc,
@@ -202,6 +202,7 @@ radv_wsi_image_create(VkDevice device_h,
 					     .memoryTypeIndex = linear ? 1 : 0,
 				     },
 				     NULL /* XXX: pAllocator */,
+				     RADV_MEM_IMPLICIT_SYNC,
 				     &memory_h);
 	if (result != VK_SUCCESS)
 		goto fail_create_image;

src/amd/vulkan/winsys/amdgpu/radv_amdgpu_bo.c

@@ -330,6 +330,8 @@ radv_amdgpu_winsys_bo_create(struct radeon_winsys *_ws,
 		request.flags |= AMDGPU_GEM_CREATE_NO_CPU_ACCESS;
 	if (flags & RADEON_FLAG_GTT_WC)
 		request.flags |= AMDGPU_GEM_CREATE_CPU_GTT_USWC;
+	if (!(flags & RADEON_FLAG_IMPLICIT_SYNC) && ws->info.drm_minor >= 22)
+		request.flags |= AMDGPU_GEM_CREATE_EXPLICIT_SYNC;
 
 	/* this won't do anything on pre 4.9 kernels */
 	if (ws->zero_all_vram_allocs && (initial_domain & RADEON_DOMAIN_VRAM))