Re: [PATCH 1/3] mmc: renesas_sdhi: Deassert the reset signal on probe

[Claudiu Beznea <claudiu.beznea@tuxon.dev>]

Hi, Geert,

On 10/10/25 10:51, Geert Uytterhoeven wrote:
> Hi Claudiu,
> 
> On Fri, 10 Oct 2025 at 07:06, Claudiu <claudiu.beznea@tuxon.dev> wrote:
>> From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
>>
>> Deassert the reset signal of the SDHI controller during probe to avoid
>> relying on the previous bootloaders. Without deasserting the reset signal,
>> the SDHI controller will not function.
>>
>> Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
> 
> Thanks for your patch!
> 
>> --- a/drivers/mmc/host/renesas_sdhi_core.c
>> +++ b/drivers/mmc/host/renesas_sdhi_core.c
>> @@ -1103,7 +1103,7 @@ int renesas_sdhi_probe(struct platform_device *pdev,
>>         if (IS_ERR(priv->clk_cd))
>>                 return dev_err_probe(&pdev->dev, PTR_ERR(priv->clk_cd), "cannot get cd clock");
>>
>> -       priv->rstc = devm_reset_control_get_optional_exclusive(&pdev->dev, NULL);
>> +       priv->rstc = devm_reset_control_get_optional_exclusive_deasserted(&pdev->dev, NULL);
> 
> Note that this has the side effect of asserting reset again on probe
> failure or unbind.

I agree. I've tested unbind/bind on RZ/G3S and I saw no issue.

>  Also on SoCs that boot with reset already deasserted
> (e.g. R-Car).  I don't know if that would be a problem.

I think all SoCs boot with resets de-asserted as the driver does register
configuration before calling reset_control_reset() in renesas_sdhi_reset(),
e.g:
- one register configuration is just above reset_control_reset()
- one register configuration, through sd_ctrl_write32_as_16_and_16(), in
  renesas_sdhi_probe()
- register writes in tmio_mmc_reset() before host->reset() call

I did bonnie++ tests after boot on RZ/G2{H, M, N, L, UL}, RZ/V2L and saw no
issue.

Thank you for your review,
Claudiu

> 
>>         if (IS_ERR(priv->rstc))
>>                 return PTR_ERR(priv->rstc);
> 
> Gr{oetje,eeting}s,
> 
>                         Geert
>