[PATCH v2 1/3] drm/amdgpu: update XGMI physical node id and GMC configs on resume

[Christian König]

On 4/30/25 12:16, Samuel Zhang wrote:
> For virtual machine with vGPUs in SRIOV single device mode and XGMI
> is enabled, XGMI physical node ids may change when waking up from
> hiberation with different vGPU devices. So update XGMI physical node
> ids on resume.
> 
> Update GPU memory controller configuration on resume if XGMI physical
> node ids are changed.
> 
> Signed-off-by: Jiang Liu <gerry at linux.alibaba.com>
> Signed-off-by: Samuel Zhang <guoqing.zhang at amd.com>
> ---
>  drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 25 ++++++++++++++++++++++
>  drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c    |  3 +--
>  drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c      |  5 +++++
>  3 files changed, 31 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> index d477a901af84..af2c784a6ccd 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> @@ -5040,6 +5040,28 @@ int amdgpu_device_suspend(struct drm_device *dev, bool notify_clients)
>  	return 0;
>  }
>  
> +static int amdgpu_device_update_xgmi_nodes(struct amdgpu_device *adev)
> +{
> +	int r;
> +	unsigned int prev_physical_node_id;
> +
> +	/* Get xgmi info again for sriov to detect device changes */
> +	if (amdgpu_sriov_vf(adev) &&
> +	    !(adev->flags & AMD_IS_APU) &&
> +	    adev->gmc.xgmi.supported &&
> +	    !adev->gmc.xgmi.connected_to_cpu) {
> +		prev_physical_node_id = adev->gmc.xgmi.physical_node_id;
> +		r = adev->gfxhub.funcs->get_xgmi_info(adev);
> +		if (r)
> +			return r;
> +
> +		dev_info(adev->dev, "xgmi node, old id %d, new id %d\n",
> +			prev_physical_node_id, adev->gmc.xgmi.physical_node_id);
> +	}
> +	return 0;
> +}
> +
> +
>  /**
>   * amdgpu_device_resume - initiate device resume
>   *
> @@ -5059,6 +5081,9 @@ int amdgpu_device_resume(struct drm_device *dev, bool notify_clients)
>  		r = amdgpu_virt_request_full_gpu(adev, true);
>  		if (r)
>  			return r;
> +		r = amdgpu_device_update_xgmi_nodes(adev);
> +		if (r)
> +			return r;
>  	}
>  
>  	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
> index d1fa5e8e3937..a2abddf3c110 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
> @@ -1298,8 +1298,7 @@ int amdgpu_gmc_get_nps_memranges(struct amdgpu_device *adev,
>  	if (!mem_ranges || !exp_ranges)
>  		return -EINVAL;
>  
> -	refresh = (adev->init_lvl->level != AMDGPU_INIT_LEVEL_MINIMAL_XGMI) &&
> -		  (adev->gmc.reset_flags & AMDGPU_GMC_INIT_RESET_NPS);
> +	refresh = true;
>  	ret = amdgpu_discovery_get_nps_info(adev, &nps_type, &ranges,
>  					    &range_cnt, refresh);
>  
> diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
> index 59385da80185..3c950c75dea1 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
> @@ -2533,6 +2533,11 @@ static int gmc_v9_0_resume(struct amdgpu_ip_block *ip_block)
>  	struct amdgpu_device *adev = ip_block->adev;
>  	int r;
>  
> +	r = gmc_v9_0_mc_init(adev);
> +	if (r)
> +		return r;
> +	gmc_v9_0_init_sw_mem_ranges(adev, adev->gmc.mem_partitions);
> +

That is most likely a no-go. We should only adjust the base of the memory partitions and not re-initialize the number of partitions.

Regards,
Christian.

>  	/* If a reset is done for NPS mode switch, read the memory range
>  	 * information again.
>  	 */