Thanks. I agree this check makes no sense. I will remove it.
On 19.08.2024 20:58, Dan Carpenter wrote:
> Hello Jacek Lawrynowicz,
>
> This is a semi-automatic email about new static checker warnings.
>
> Commit 647371a6609d ("accel/ivpu: Add GEM buffer object management")
> from Jan 17, 2023, leads to the following Smatch complaint:
>
> ./drivers/accel/ivpu/ivpu_drv.c:119 ivpu_file_priv_put()
> warn: variable dereferenced before check 'file_priv' (see line 117)
>
> ./drivers/accel/ivpu/ivpu_drv.c
> 116 struct ivpu_file_priv *file_priv = *link;
> 117 struct ivpu_device *vdev = file_priv->vdev;
> ^^^^^^^^^^^^^^^
> Dereference
>
> 118
> 119 drm_WARN_ON(&vdev->drm, !file_priv);
> ^^^^^^^^^
> Pointless check. If there is a NULL dereference it already prints a stack
> trace. Also it's really unlikely that we're going to predict which pointer is
> going to lead to a NULL dereference ahead of time.
>
> 120
> 121 ivpu_dbg(vdev, KREF, "file_priv put: ctx %u refcount %u\n",
>
> regards,
> dan carpenter
