Dear Alex Waterman,
> Adds a short busy loop wait to the atmel_usart.c serial_init()
> function.
> 
> Signed-off-by: Alex Waterman <awater...@dawning.com>
> ---
>  drivers/serial/atmel_usart.c |    6 ++++++
>  1 files changed, 6 insertions(+), 0 deletions(-)
> 
> diff --git a/drivers/serial/atmel_usart.c b/drivers/serial/atmel_usart.c
> index e326b2b..e355706 100644
> --- a/drivers/serial/atmel_usart.c
> +++ b/drivers/serial/atmel_usart.c
> @@ -47,6 +47,8 @@ void serial_setbrg(void)
>  
>  int serial_init(void)
>  {
> +
> +     volatile int i = 0;
>       atmel_usart3_t *usart = (atmel_usart3_t *)CONFIG_USART_BASE;
>  
>       writel(USART3_BIT(RSTRX) | USART3_BIT(RSTTX), &usart->cr);
> @@ -61,6 +63,10 @@ int serial_init(void)
>                          | USART3_BF(NBSTOP, USART3_NBSTOP_1)),
>                          &usart->mr);
>  
> +     /* Short wait to let the serial port init. */
> +     for (; i < 10000; i++)
> +             ;
> +
>       return 0;
>  }

1. Why is this delay needed in the first place?
Apparently it has not been required before.

2. Such a delay loop is highly dependant on toolchain,
cache and page situation and processor speed. It can not be
accepted as a solution. If a delay is really required here
and cannot be substituting by reading the UARTs status
register, udelay() must be used.

Best Regards,
Reinhard
_______________________________________________
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot

Reply via email to