Re: [PATCH v3 2/5] platform/chrome: cros_ec_lpc: MEC access can use an AML mutex

[Tzung-Bi Shih <tzungbi@kernel.org>]

On Wed, Jun 05, 2024 at 07:33:48AM +0100, Ben Walsh wrote:
> @@ -93,7 +146,9 @@ int cros_ec_lpc_io_bytes_mec(enum cros_ec_lpc_mec_io_type io_type,
>  	else
>  		access = ACCESS_TYPE_LONG_AUTO_INCREMENT;
>  
> -	mutex_lock(&io_mutex);
> +	ret = cros_ec_lpc_mec_lock();
> +	if (ret)
> +		return ret;
>  
>  	/* Initialize I/O at desired address */
>  	cros_ec_lpc_mec_emi_write_address(offset, access);
> @@ -135,7 +190,9 @@ int cros_ec_lpc_io_bytes_mec(enum cros_ec_lpc_mec_io_type io_type,
>  	}
>  
>  done:
> -	mutex_unlock(&io_mutex);
> +	ret = cros_ec_lpc_mec_unlock();
> +	if (ret < 0)
> +		return ret;

A nit haven't caught in previous round review:

- `if (ret)` for cros_ec_lpc_mec_lock().
- `if (ret < 0)` for cros_ec_lpc_mec_unlock().

If I see the inconsistency when reading the code, it would take me further
cycles for seeking why.  We should fix it.

No need to resend for this; I will fix it for you.