Hi again,

The issue you were having is because I did a mistake, which is now
fixed in this new patch.

I've attached the patch to try. You need to try it without any other
patches or modifications, and to send the log to the mailing list.

Once you will have sent the log, we will most probably need to comment
the other failing functions in the same way.

Here are the details of this patch:
-----------------------------------
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> @@ -1713,8 +1713,7 @@ static int amdgpu_device_ip_early_init(struct
> amdgpu_device *adev) }
>  
>       r = amdgpu_device_parse_gpu_info_fw(adev);
> -     if (r)
> -             return r;
> +     /* Deblobbed */
Here this should take care of it because:
- amdgpu_device_parse_gpu_info_fw will return an error before any code
  that deference null pointers is ever executed.
- the error will then be ignored, and r is not reused before being
  asigned to a new value.

Sorry for the mistake.

Denis.
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index b5120ae41ce..9469db1ab8e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -1713,8 +1713,7 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev)
 	}
 
 	r = amdgpu_device_parse_gpu_info_fw(adev);
-	if (r)
-		return r;
+	/* Deblobbed */
 
 	if (amdgpu_discovery && adev->asic_type >= CHIP_NAVI10)
 		amdgpu_discovery_get_gfx_info(adev);

Attachment: pgpDgHeDEIZKF.pgp
Description: OpenPGP digital signature

_______________________________________________
linux-libre mailing list
[email protected]
http://www.fsfla.org/cgi-bin/mailman/listinfo/linux-libre

Reply via email to