[PATCH v2 2/2] ARM: boards: i.MX93: Add NXP FRDM i.MX93 support

[Maud Spierings]

On 10/21/25 13:43, Ahmad Fatoum wrote:
> Hello Fabian,
> 
> On 10/21/25 1:37 PM, Fabian Pflug wrote:
>> Hello Ahmad,
>>
>> On Tue, 2025-10-21 at 13:11 +0200, Ahmad Fatoum wrote:
>>> Hello Fabian,
>>>
>>>> +static noinline void frdm_imx93_continue(void)
>>>> +{
>>>> +	void __iomem *base = IOMEM(MX9_UART1_BASE_ADDR);
>>>> +	void __iomem *muxbase = IOMEM(MX9_IOMUXC_BASE_ADDR);
>>>> +
>>>> +	/* configure LPUART1 TX pin */
>>>> +	writel(0x0, muxbase + 0x184);
>>>> +
>>>> +	imx9_uart_setup(base);
>>>> +	pbl_set_putc(lpuart32_putc, base + 0x10);
>>>> +	putchar('>');
>>>
>>> Do you see the > character on bootup if CONFIG_PBL_CONSOLE is enabled?
>>> Maud (Cc'd) mentioned on the IRC that the 0x10 offset shouldn't be
>>> necessary.
>>>
>>
>> I did see the > character with (working with a 2025.07.0 barebox here):
> 
> Ok, good to know.
> 
> 
>> But there seems to be another Problem. If I enable
>>
>> CONFIG_DEBUG_LL=y
>> CONFIG_DEBUG_IMX_UART=y
>> CONFIG_DEBUG_IMX8M_UART=y
>> CONFIG_DEBUG_IMX_UART_PORT=1
>> CONFIG_DEBUG_PBL=y
> 
> Given that you are using the i.MX9's LPUART, I think it's expected that
> DEBUG_LL misbehaves. PBL output works, because the PBL console has
> precedence over DEBUG_LL.
> 
> Cheers,
> Ahmad
> 

I think I may have confused something.
Could've sworn I saw the base + 0x10 in the _uart_setup() call, but now 
that I am looking at it again I don't see it.

I may have been a bit too tired yesterday and made a premature 
conclusion, based on incorrectly reading the code.

Kind regards,
Maud