drm/amdgpu: Enable translate_further to extend UTCL2 reach
authorJoseph Greathouse <Joseph.Greathouse@amd.com>
Thu, 4 Aug 2022 14:54:57 +0000 (09:54 -0500)
committerAlex Deucher <alexander.deucher@amd.com>
Wed, 10 Aug 2022 19:41:23 +0000 (15:41 -0400)
Enable translate_further on Arcturus and Aldebaran server chips
in order to increase the UTCL2 reach from 8 GiB to 64 GiB,
which is more in line with the amount of framebuffer DRAM in
the devices.

Signed-off-by: Joseph Greathouse <Joseph.Greathouse@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Kent Russell <kent.russell@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c

index 22761a3..ab89d91 100644 (file)
@@ -1624,12 +1624,15 @@ static int gmc_v9_0_sw_init(void *handle)
                        amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 47);
                else
                        amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
+               if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2))
+                       adev->gmc.translate_further = adev->vm_manager.num_level > 1;
                break;
        case IP_VERSION(9, 4, 1):
                adev->num_vmhubs = 3;
 
                /* Keep the vm size same with Vega20 */
                amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
+               adev->gmc.translate_further = adev->vm_manager.num_level > 1;
                break;
        default:
                break;