Re: [PATCH] usb: dwc2: Fix call location of dwc2_check_core_endianness

[Minas Harutyunyan]

On 8/28/2018 1:37 AM, Bruno Meirelles Herrera wrote:
> Some SoC/IP as STM32F469, the snpsid can only be read after clock is
> enabled, otherwise it will read as 0, and the dwc2_check_core_endianness
> will assume the core and AHB have opposite endianness, leading to the
> following error:
> 
> [    1.976339] dwc2 50000000.usb: 50000000.usb supply vusb_d not found, using dummy regulator
> [    1.986124] dwc2 50000000.usb: Linked as a consumer to regulator.0
> [    1.992711] dwc2 50000000.usb: 50000000.usb supply vusb_a not found, using dummy regulator
> [    2.003672] dwc2 50000000.usb: dwc2_core_reset: HANG! AHB Idle timeout GRSTCTL GRSTCTL_AHBIDLE
> [    2.015176] dwc2: probe of 50000000.usb failed with error -16
> 
> The proposed patch changes the location where dwc2_check_core_endianness
> is called, allowing the clock peripheral to be enabled first.
> 
> Signed-off-by: Bruno Meirelles Herrera <bmh@xxxxxxxxxxxx>

Acked-by: Minas Harutyunyan <hminas@xxxxxxxxxxxx>

> ---
>   drivers/usb/dwc2/platform.c | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/usb/dwc2/platform.c b/drivers/usb/dwc2/platform.c
> index 9a53a58e676e..577642895b57 100644
> --- a/drivers/usb/dwc2/platform.c
> +++ b/drivers/usb/dwc2/platform.c
> @@ -412,8 +412,6 @@ static int dwc2_driver_probe(struct platform_device *dev)
>   	dev_dbg(&dev->dev, "mapped PA %08lx to VA %p\n",
>   		(unsigned long)res->start, hsotg->regs);
>   
> -	hsotg->needs_byte_swap = dwc2_check_core_endianness(hsotg);
> -
>   	retval = dwc2_lowlevel_hw_init(hsotg);
>   	if (retval)
>   		return retval;
> @@ -438,6 +436,8 @@ static int dwc2_driver_probe(struct platform_device *dev)
>   	if (retval)
>   		return retval;
>   
> +	hsotg->needs_byte_swap = dwc2_check_core_endianness(hsotg);
> +
>   	retval = dwc2_get_dr_mode(hsotg);
>   	if (retval)
>   		goto error;
>