Re: [PATCH] spi: dw: assert reset before deasserting reset

[Serge Semin <fancer.lancer@xxxxxxxxx>]

Hello Niklas

On Tue, Mar 01, 2022 at 11:17:20AM +0000, Niklas Cassel wrote:
> From: Niklas Cassel <niklas.cassel@xxxxxxx>
> 
> Simply deasserting reset just ensures that the hardware is taken out of
> reset, if it was booted with the hardware reset asserted.
> 
> In order actually reset the SPI controller, we need to assert reset before
> deasserting.
> 
> By doing this, we ensure that the hardware is not in some bad state, and we
> ensure that the hardware registers get set to their reset default, clearing
> any setting set by e.g. a bootloader.
> 
> Signed-off-by: Niklas Cassel <niklas.cassel@xxxxxxx>
> ---
>  drivers/spi/spi-dw-mmio.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/spi/spi-dw-mmio.c b/drivers/spi/spi-dw-mmio.c
> index 5101c4c6017b..eb1dacb45ca2 100644
> --- a/drivers/spi/spi-dw-mmio.c
> +++ b/drivers/spi/spi-dw-mmio.c
> @@ -289,6 +289,8 @@ static int dw_spi_mmio_probe(struct platform_device *pdev)
>  		ret = PTR_ERR(dwsmmio->rstc);
>  		goto out_clk;
>  	}

> +	reset_control_assert(dwsmmio->rstc);
> +	udelay(2);

Do we really need this? dw_spi_add_host() is doing a sort of soft reset
anyway by calling the dw_spi_hw_init() method. Do you have a real
platform, which requires to do a full hard-reset?

What about the self-reset based controllers?

-Sergey

>  	reset_control_deassert(dwsmmio->rstc);
>  
>  	dws->bus_num = pdev->id;
> -- 
> 2.35.1