Re: [PATCH] hwrng: rockchip: propagate reset errors

[Nicolas Frattaroli]

On Sunday, 30 August 2026 15:36:52 Central European Summer Time Pengpeng Hou wrote:
> rk_rng_probe() pulses the optional or required reset array before
> registering
> the hardware RNG, but discards errors from both reset operations.
> 
> Return reset failures before publishing the RNG device.
> 
> Fixes: 8eff8eb83fc0 ("hwrng: rockchip - add support for rk3588's standalone TRNG")
> Signed-off-by: Pengpeng Hou <pengpeng@xxxxxxxxxxx>
> ---
>  drivers/char/hw_random/rockchip-rng.c | 9 +++++++--
>  1 file changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/char/hw_random/rockchip-rng.c b/drivers/char/hw_random/rockchip-rng.c
> index 6e3ed4b856051..ff98fcdc68911 100644
> --- a/drivers/char/hw_random/rockchip-rng.c
> +++ b/drivers/char/hw_random/rockchip-rng.c
> @@ -412,9 +412,14 @@ static int rk_rng_probe(struct platform_device *pdev)
>  		if (IS_ERR(rst))
>  			return dev_err_probe(dev, PTR_ERR(rst), "Failed to get reset property\n");
>  
> -		reset_control_assert(rst);
> +		ret = reset_control_assert(rst);
> +		if (ret)
> +			return ret;
> +
>  		udelay(2);
> -		reset_control_deassert(rst);
> +		ret = reset_control_deassert(rst);
> +		if (ret)
> +			return ret;
>  	}
>  
>  	platform_set_drvdata(pdev, rk_rng);
> 
> base-commit: 08dbfad3f5040f5bdb6c529da20d6d4e81fefd72
> 

The reset controller on rockchip SoCs does not return anything other
than 0 on the assert/deassert ops, and even if it did, the probe
function should use dev_err_probe to report those failures.

In other words, this patch only adds (non-ideal) dead code.

Kind regards,
Nicolas Frattaroli