Re: [PATCH 6/6] drm/imagination: Warn or error on unsupported hardware

[Geert Uytterhoeven]

Hi Matt,

On Tue, 13 Jan 2026 at 11:21, Matt Coster <matt.coster@xxxxxxxxxx> wrote:
> Gate the use of unsupported hardware behind a new module parameter
> (exp_hw_support).
>
> Signed-off-by: Matt Coster <matt.coster@xxxxxxxxxx>

Thanks for your patch, which is now commit 1c21f240fbc1e47b
("drm/imagination: Warn or error on unsupported hardware")
in v7.0-rc1.

> --- a/drivers/gpu/drm/imagination/pvr_device.c
> +++ b/drivers/gpu/drm/imagination/pvr_device.c
> @@ -525,6 +525,77 @@ pvr_gpuid_decode_string(const struct pvr_device *pvr_dev,
>  }
>  EXPORT_SYMBOL_IF_KUNIT(pvr_gpuid_decode_string);
>
> +static bool pvr_exp_hw_support;
> +module_param_named(exp_hw_support, pvr_exp_hw_support, bool, 0600);
> +MODULE_PARM_DESC(exp_hw_support, "Bypass runtime checks for fully supported GPU cores. WARNING: enabling this option may result in a buggy, insecure, or otherwise unusable driver.");
> +
> +/**
> + * enum pvr_gpu_support_level - The level of support for a gpu_id in the current
> + * version of the driver.
> + *
> + * @PVR_GPU_UNKNOWN: Cores that are unknown to the driver. These may not even exist.
> + * @PVR_GPU_EXPERIMENTAL: Cores that have experimental support.
> + * @PVR_GPU_SUPPORTED: Cores that are supported and maintained.
> + */
> +enum pvr_gpu_support_level {
> +       PVR_GPU_UNKNOWN,
> +       PVR_GPU_EXPERIMENTAL,
> +       PVR_GPU_SUPPORTED,
> +};
> +
> +static enum pvr_gpu_support_level
> +pvr_gpu_support_level(const struct pvr_gpu_id *gpu_id)
> +{
> +       switch (pvr_gpu_id_to_packed_bvnc(gpu_id)) {
> +       case PVR_PACKED_BVNC(33, 15, 11, 3):
> +       case PVR_PACKED_BVNC(36, 53, 104, 796):
> +               return PVR_GPU_SUPPORTED;
> +
> +       case PVR_PACKED_BVNC(36, 52, 104, 182):
> +               return PVR_GPU_EXPERIMENTAL;
> +
> +       default:
> +               return PVR_GPU_UNKNOWN;
> +       }
> +}
> +
> +static int
> +pvr_check_gpu_supported(struct pvr_device *pvr_dev,
> +                       const struct pvr_gpu_id *gpu_id)
> +{
> +       struct drm_device *drm_dev = from_pvr_device(pvr_dev);
> +
> +       switch (pvr_gpu_support_level(gpu_id)) {
> +       case PVR_GPU_SUPPORTED:
> +               if (pvr_exp_hw_support)
> +                       drm_info(drm_dev, "Module parameter 'exp_hw_support' was set, but this hardware is fully supported by the current driver.");
> +
> +               break;
> +
> +       case PVR_GPU_EXPERIMENTAL:
> +               if (!pvr_exp_hw_support) {
> +                       drm_err(drm_dev, "Unsupported GPU! Set 'exp_hw_support' to bypass this check.");
> +                       return -ENODEV;
> +               }
> +
> +               drm_warn(drm_dev, "Running on unsupported hardware; you may encounter bugs!");
> +               break;
> +
> +       /* NOTE: This code path may indicate misbehaving hardware. */
> +       case PVR_GPU_UNKNOWN:
> +       default:
> +               if (!pvr_exp_hw_support) {
> +                       drm_err(drm_dev, "Unknown GPU! Set 'exp_hw_support' to bypass this check.");

After marking the gpu device node enabled in DTS for Renesas Salvator-X
with R-Car M3-W (as you are aware, it is still disabled upstream
because the system may crash when the firmware is not found), this
error message is triggered:

    powervr fd000000.gpu: [drm] *ERROR* Unknown GPU! Set
'exp_hw_support' to bypass this check.

and the driver fails to probe. Using the exp_hw_support flag makes
the driver probe (when the firmware is available):

    # echo 1 > /sys/module/powervr/parameters/exp_hw_support
    # echo fd000000.gpu > /sys/bus/platform/drivers/powervr/bind
    powervr fd000000.gpu: [drm] Running on unknown hardware; expect issues.
    powervr fd000000.gpu: [drm] loaded firmware powervr/rogue_4.45.2.58_v1.fw
    powervr fd000000.gpu: [drm] FW version v1.0 (build 6513336 OS)
    [drm] Initialized powervr 1.0.0 for fd000000.gpu on minor 1

I am not sure if this counts as a regression.  Given firmware is
available for this device, I'd expect it to fall (at least) in the the
PVR_GPU_EXPERIMENTAL case?

Thanks!

> +                       return -ENODEV;
> +               }
> +
> +               drm_warn(drm_dev, "Running on unknown hardware; expect issues.");
> +               break;
> +       }
> +
> +       return 0;
> +}
> +
>  static char *pvr_gpuid_override;
>  module_param_named(gpuid, pvr_gpuid_override, charp, 0400);
>  MODULE_PARM_DESC(gpuid, "GPU ID (BVNC) to be used instead of the value read from hardware.");
> @@ -555,7 +626,7 @@ pvr_load_gpu_id(struct pvr_device *pvr_dev)
>                         return err;
>         }
>
> -       return 0;
> +       return pvr_check_gpu_supported(pvr_dev, gpu_id);
>  }
>
>  /**

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds