[PATCH] phy: stm32-usphyc: add mdelay(1) to fix timeout on some machines

[Fabrice Gasnier]

On 1/24/23 21:45, Michael Grzeschik wrote:
> An mdelay of 1 seems to be necessary on some machines, since

Hi Michael,

Could you precise on which board ?

> the monsel status does not seem to be accurate. On rare occasions just
> working with the phy after this pll check lead to no functional usb.

Could you elaborate on the symptoms (provide some error logs) ?

> With this short mdelay this issue was not reported again.
> 
> Signed-off-by: Michael Grzeschik <m.grzeschik at pengutronix.de>
> ---
>  drivers/phy/st/phy-stm32-usbphyc.c | 9 +++++++++
>  1 file changed, 9 insertions(+)
> 
> diff --git a/drivers/phy/st/phy-stm32-usbphyc.c b/drivers/phy/st/phy-stm32-usbphyc.c
> index 5bb9647b078f12..c452a0caceb9fa 100644
> --- a/drivers/phy/st/phy-stm32-usbphyc.c
> +++ b/drivers/phy/st/phy-stm32-usbphyc.c
> @@ -353,6 +353,15 @@ static int stm32_usbphyc_phy_init(struct phy *phy)
>  		goto pll_disable;
>  	}
>  
> +	/* This mdelay seems to be necessary on some machines, since the
> +	 * monsel status does not seem to be accurate. On rare occasions
> +	 * just working with the phy after this pll check the usb
> +	 * peripheral (e.g. on the dwc2) run into timeout issues and
> +	 * leading to no functional usb. With this short mdelay this
> +	 * issue was not reported again.
> +	 */
> +	mdelay(1);
> +

The USBPHYC provides two PHY interfaces:
- PHY port#1 is for USBH
- PHY port#2 can be assigned to USBH or DWC2.

Adding some delay here, we'll hit twice this penalty (e.g. for USBH +
DWC2 or dual USBH) on all MP1 (MP13 & MP15) platforms.

Could you try to narrow down the issue by adding some debug log in:
- stm32_usbphyc_phy_init
- stm32_usbphyc_clk48_prepare

Best Regards,
Fabrice

>  	usbphyc_phy->active = true;
>  
>  	return 0;