Re: [PATCH v2] accel/rocket: Report per-core engine usage through fdinfo

[Sidong Yang <sidong.yang@furiosa.ai>]

On Tue, Nov 03, 2026 at 07:14:14PM +0900, Gyeyoung Baek wrote:
> Hi Sidong,
> 
> > +
> >  static int
> >  rocket_open(struct drm_device *dev, struct drm_file *file)
> >  {
> > @@ -86,10 +107,19 @@ rocket_open(struct drm_device *dev, struct drm_file *file)
> >         }
> >
> >         rocket_priv->rdev = rdev;
> > +
> > +       rocket_priv->engine_usage = kzalloc_flex(*rocket_priv->engine_usage,
> > +                                                elapsed_ns, rdev->max_cores);
> > +       if (!rocket_priv->engine_usage) {
> > +               ret = -ENOMEM;
> > +               goto err_free;
> > +       }
> > +       kref_init(&rocket_priv->engine_usage->kref);
> 
> How about moving these lines into a rocket_engine_usage_create() helper, to pair
> with rocket_engine_usage_destroy()?

Sure, I'll make it for v3.

Thanks,
Sidong
> 
> Otherwise LGTM. Tested on RK3588 (ROCK 5B) with KASAN: the OOB from v1 no
> longer reproduces, and drm-engine-npu* times are accounted as expected
> while running benchmark_model with teflon.
> 
> Reviewed-by: Gyeyoung Baek <gye976@gmail.com>
> Tested-by: Gyeyoung Baek <gye976@gmail.com>
> 
> -- 
> Thanks,
> Gyeyoung