Re: [PATCH] usb: dwc3: xilinx: add reset-controller support

[Thinh Nguyen <Thinh.Nguyen@synopsys.com>]

On Thu, Oct 05, 2023, Piyush Mehta wrote:
> Add a reset-controller for supporting Xilinx versal platforms. To reset
> the USB controller, get the reset ID from device-tree and using ID trigger
> the reset, with the assert and deassert reset controller APIs for USB
> controller initialization. Delay of microseconds is added in between assert
> and deassert to meet the setup and hold the time requirement of the reset.
> 
> Signed-off-by: Piyush Mehta <piyush.mehta@amd.com>
> ---
>  drivers/usb/dwc3/dwc3-xilinx.c | 17 ++++++++++-------
>  1 file changed, 10 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/usb/dwc3/dwc3-xilinx.c b/drivers/usb/dwc3/dwc3-xilinx.c
> index 19307d24f3a0..d0780bf231fb 100644
> --- a/drivers/usb/dwc3/dwc3-xilinx.c
> +++ b/drivers/usb/dwc3/dwc3-xilinx.c
> @@ -32,9 +32,6 @@
>  #define XLNX_USB_TRAFFIC_ROUTE_CONFIG		0x005C
>  #define XLNX_USB_TRAFFIC_ROUTE_FPD		0x1
>  
> -/* Versal USB Reset ID */
> -#define VERSAL_USB_RESET_ID			0xC104036
> -
>  #define XLNX_USB_FPD_PIPE_CLK			0x7c
>  #define PIPE_CLK_DESELECT			1
>  #define PIPE_CLK_SELECT				0
> @@ -72,20 +69,26 @@ static void dwc3_xlnx_mask_phy_rst(struct dwc3_xlnx *priv_data, bool mask)
>  static int dwc3_xlnx_init_versal(struct dwc3_xlnx *priv_data)
>  {
>  	struct device		*dev = priv_data->dev;
> +	struct reset_control	*crst;
>  	int			ret;
>  
> +	crst = devm_reset_control_get_exclusive(dev, NULL);
> +	if (IS_ERR(crst))
> +		return dev_err_probe(dev, PTR_ERR(crst), "failed to get reset signal\n");
> +
>  	dwc3_xlnx_mask_phy_rst(priv_data, false);
>  
>  	/* Assert and De-assert reset */
> -	ret = zynqmp_pm_reset_assert(VERSAL_USB_RESET_ID,
> -				     PM_RESET_ACTION_ASSERT);
> +	ret = reset_control_assert(crst);
>  	if (ret < 0) {
>  		dev_err_probe(dev, ret, "failed to assert Reset\n");
>  		return ret;
>  	}
>  
> -	ret = zynqmp_pm_reset_assert(VERSAL_USB_RESET_ID,
> -				     PM_RESET_ACTION_RELEASE);
> +	/* reset hold time */
> +	usleep_range(5, 10);

Was this delay needed before? If so, is this a fix patch?

> +
> +	ret = reset_control_deassert(crst);
>  	if (ret < 0) {
>  		dev_err_probe(dev, ret, "failed to De-assert Reset\n");
>  		return ret;
> -- 
> 2.17.1
> 

BR,
Thinh