1# SPDX-License-Identifier: MIT 2# 3# Heterogeneous system architecture configuration 4# 5 6config HSA_AMD 7 bool "HSA kernel driver for AMD GPU devices" 8 depends on DRM_AMDGPU && (X86_64 || ARM64 || PPC64) 9 imply AMD_IOMMU_V2 if X86_64 10 select HMM_MIRROR 11 select MMU_NOTIFIER 12 select DRM_AMDGPU_USERPTR 13 help 14 Enable this if you want to use HSA features on AMD GPU devices. 15 16config HSA_AMD_SVM 17 bool "Enable HMM-based shared virtual memory manager" 18 depends on HSA_AMD && DEVICE_PRIVATE 19 default y 20 select HMM_MIRROR 21 select MMU_NOTIFIER 22 help 23 Enable this to use unified memory and managed memory in HIP. This 24 memory manager supports two modes of operation. One based on 25 preemptions and one based on page faults. To enable page fault 26 based memory management on most GFXv9 GPUs, set the module 27 parameter amdgpu.noretry=0. 28 29config HSA_AMD_P2P 30 bool "HSA kernel driver support for peer-to-peer for AMD GPU devices" 31 depends on HSA_AMD && PCI_P2PDMA && DMABUF_MOVE_NOTIFY 32 help 33 Enable peer-to-peer (P2P) communication between AMD GPUs over 34 the PCIe bus. This can improve performance of multi-GPU compute 35 applications and libraries by enabling GPUs to access data directly 36 in peer GPUs' memory without intermediate copies in system memory. 37 38 This P2P feature is only enabled on compatible chipsets, and between 39 GPUs with large memory BARs that expose the entire VRAM in PCIe bus 40 address space within the physical address limits of the GPUs. 41 42