Re: [PATCH] drm/amdgpu: Enable translate_further to extend UTCL2 reach

[Christian König]

I can't 100% judge if the IP version check is sufficient or not, so keep 
an eye open for regression reports when this goes upstream.

Apart from that Acked-by: Christian König <christian.koe...@amd.com> as 
well.

Christian.

Am 08.08.22 um 15:12 schrieb Russell, Kent:
[AMD Official Use Only - General]

You can throw an Acked-by Kent Russell <kent.russ...@amd.com> as well . Not 
sure if there's someone who wants to provide a more concrete RB.

  Kent

-----Original Message-----
From: Kuehling, Felix <felix.kuehl...@amd.com>
Sent: Thursday, August 4, 2022 2:54 PM
To: Greathouse, Joseph <joseph.greatho...@amd.com>; amd-
g...@lists.freedesktop.org
Cc: Russell, Kent <kent.russ...@amd.com>; Koenig, Christian
<christian.koe...@amd.com>
Subject: Re: [PATCH] drm/amdgpu: Enable translate_further to extend UTCL2
reach

On 2022-08-04 12:01, Joseph Greathouse wrote:
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.greatho...@amd.com>
Acked-by: Felix Kuehling <felix.kuehl...@amd.com>


---
   drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 3 +++
   1 file changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 22761a3bb818..ab89d91975ab 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -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;