[PATCH 1/2] drm/amdgpu: refine eeprom data check

Lazar, Lijo lijo.lazar at amd.com
Fri Jul 11 08:07:50 UTC 2025



On 7/11/2025 1:29 PM, ganglxie wrote:
> add eeprom data checksum check before driver unload. reset eeprom
> and save correct data to eeprom when check failed
> 
> Signed-off-by: ganglxie <ganglxie at amd.com>
> ---
>  drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c       |  1 +
>  .../gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c    | 28 +++++++++++++++++++
>  .../gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h    |  2 ++
>  3 files changed, 31 insertions(+)
> 
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> index 571b70da4562..1009b60f6ae4 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> @@ -2560,6 +2560,7 @@ amdgpu_pci_remove(struct pci_dev *pdev)
>  	struct drm_device *dev = pci_get_drvdata(pdev);
>  	struct amdgpu_device *adev = drm_to_adev(dev);
>  
> +	amdgpu_ras_eeprom_check_and_recover(adev);
>  	amdgpu_xcp_dev_unplug(adev);
>  	amdgpu_gmc_prepare_nps_mode_change(adev);
>  	drm_dev_unplug(dev);
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
> index 2af14c369bb9..df0e9b87d578 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
> @@ -1522,3 +1522,31 @@ int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control)
>  
>  	return res < 0 ? res : 0;
>  }
> +
> +void amdgpu_ras_eeprom_check_and_recover(struct amdgpu_device *adev)
> +{
> +	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
> +	struct amdgpu_ras_eeprom_control *control;
> +	int res = 0;

This initialization could be skipped. With that -

	Reviewed-by: Lijo Lazar <lijo.lazar at amd.com>

Thanks,
Lijo
> +
> +	if (!ras)
> +		return;
> +	control = &ras->eeprom_control;
> +	if (!control->is_eeprom_valid)
> +		return;
> +	res = __verify_ras_table_checksum(control);
> +	if (res) {
> +		dev_warn(adev->dev,
> +			"RAS table incorrect checksum or error:%d, try to recover\n",
> +			res);
> +		if (!amdgpu_ras_eeprom_reset_table(control))
> +			if (!amdgpu_ras_save_bad_pages(adev, NULL))
> +				if (!__verify_ras_table_checksum(control)) {
> +					dev_info(adev->dev, "RAS table recovery succeed\n");
> +					return;
> +				}
> +		dev_err(adev->dev, "RAS table recovery failed\n");
> +		control->is_eeprom_valid = false;
> +	}
> +	return;
> +}
> \ No newline at end of file
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
> index 35c69ac3dbeb..ebfca4cb5688 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
> @@ -161,6 +161,8 @@ void amdgpu_ras_debugfs_set_ret_size(struct amdgpu_ras_eeprom_control *control);
>  
>  int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control);
>  
> +void amdgpu_ras_eeprom_check_and_recover(struct amdgpu_device *adev);
> +
>  extern const struct file_operations amdgpu_ras_debugfs_eeprom_size_ops;
>  extern const struct file_operations amdgpu_ras_debugfs_eeprom_table_ops;
>  



More information about the amd-gfx mailing list