Use SDMA rather than KIQ/MMIO. This is recommended by the hubs team and aligns with what other OSes do.
Signed-off-by: Alex Deucher <[email protected]> --- drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c index 3d545e0ebfd86..fa0fc4dc724c6 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c @@ -788,6 +788,7 @@ static int gmc_v10_0_hw_init(struct amdgpu_ip_block *ip_block) adev->gmc.flush_pasid_uses_kiq = !amdgpu_emu_mode; if (adev->gmc.flush_pasid_uses_kiq) adev->gmc.pasid_inv_method = AMDGPU_TLB_INV_METHOD_KIQ; + adev->gmc.gart_inv_method = AMDGPU_TLB_INV_METHOD_SDMA; /* The sequence of these two function calls matters.*/ gmc_v10_0_init_golden_registers(adev); -- 2.55.0
