Re: iMX ddrphy_utils difference with U-boot

[Ahmad Fatoum <a.fatoum@pengutronix.de>]

Hi,

On 6/15/26 10:49 AM, Sascha Hauer wrote:
> On 2026-06-08 19:26, Ahmad Fatoum wrote:
>> Hello Andrei,
>>
>> On 6/8/26 16:24, Andrei Lalaev wrote:
>>> Hi,
>>>
>>> I am moving an iMX8MP module from vendor U-Boot 2024.04 to Barebox 2025.02
>>
>> Sidenote: You'll probably want to use one of the still supported
>> v2026.04 or v2026.06 releases.
>>
>>> and found a strange difference in the DDR training code:
>>>
>>>   vim drivers/ddr/imx/ddrphy_utils.c +94
>>>
>>> And the corresponding line in U-Boot:
>>>
>>>   vim drivers/ddr/imx/phy/ddrphy_utils.c +101
>>>
>>> Is there any chance that somebody knows/remembers why "return -1" was replaced with "hang()"?
>>
>> I can't speak for Sascha, but having looked at the code, I see no reason
>> why not to propagate the error.
>>
>>> I couldn't find any explanation in the commits/mailing lists.
>>
>> My guess is that it wasn't anticipated that boards would handle
>> the error gracefully to fall back to a different DDR init.
> 
> That's my guess as well. Note the callers of
> imx8m_wait_ddrphy_training_complete() which is only a wrapper around
> calling wait_ddrphy_training_complete() do not check the error code, so
> when changing it back to return an error code we likely want to add
> error checking where missing.

I just stumbled upon this patch:
https://lore.kernel.org/all/20260819070758.51350-1-frieder@fris.de/

If we start allowing ddr_cfg_phy to propagate this error, you'll likely
want above patch as well.

Cheers,
Ahmad

> 
> Sascha
> 

-- 
Pengutronix e.K.                  |                             |
Steuerwalder Str. 21              | http://www.pengutronix.de/  |
31137 Hildesheim, Germany         | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686  | Fax:   +49-5121-206917-5555 |