[PATCH] drm/amd/pm: avoid potential UBSAN issue on legacy asics

Lazar, Lijo lijo.lazar at amd.com
Tue May 9 07:29:24 UTC 2023



On 5/9/2023 10:59 AM, Guchun Chen wrote:
> Prevent further dpm casting on legacy asics without od_enabled in
> amdgpu_dpm_is_overdrive_supported. This can avoid UBSAN complain
> in init sequence.
> 
> v2: add a macro to check legacy dpm instead of checking asic family/type
> 
> Suggested-by: Evan Quan <evan.quan at amd.com>
> Signed-off-by: Guchun Chen <guchun.chen at amd.com> > ---
>   drivers/gpu/drm/amd/pm/amdgpu_dpm.c | 9 +++++++--
>   1 file changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/amd/pm/amdgpu_dpm.c b/drivers/gpu/drm/amd/pm/amdgpu_dpm.c
> index 86246f69dbe1..4b28fd62ed7a 100644
> --- a/drivers/gpu/drm/amd/pm/amdgpu_dpm.c
> +++ b/drivers/gpu/drm/amd/pm/amdgpu_dpm.c
> @@ -36,6 +36,8 @@
>   #define amdgpu_dpm_enable_bapm(adev, e) \
>   		((adev)->powerplay.pp_funcs->enable_bapm((adev)->powerplay.pp_handle, (e)))
>   
> +#define is_legacy_dpm(adev) ((adev)->powerplay.pp_handle == (adev))
> +

For consistent naming, you may use amdgpu_dpm_is_legacy_dpm().

Regardless,

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

Thanks,
Lijo

>   int amdgpu_dpm_get_sclk(struct amdgpu_device *adev, bool low)
>   {
>   	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
> @@ -1467,8 +1469,11 @@ int amdgpu_dpm_is_overdrive_supported(struct amdgpu_device *adev)
>   	} else {
>   		struct pp_hwmgr *hwmgr;
>   
> -		/* SI asic does not carry od_enabled */
> -		if (adev->family == AMDGPU_FAMILY_SI)
> +		/*
> +		 * dpm on some legacy asics don't carry od_enabled member
> +		 * as its pp_handle is casted directly from adev.
> +		 */
> +		if (is_legacy_dpm(adev))
>   			return false;
>   
>   		hwmgr = (struct pp_hwmgr *)adev->powerplay.pp_handle;


More information about the amd-gfx mailing list