Re: [PATCH 3/4] spi: cadence-quadspi: Fix cqspi_setup_flash()

[Santhosh Kumar K <s-k6@ti.com>]

Hello Theo,

On 04/09/25 21:02, Théo Lebrun wrote:
> Hello Santhosh,
> 
> On Thu Sep 4, 2025 at 3:31 PM CEST, Santhosh Kumar K wrote:
>> The 'max_cs' stores the largest chip select number. It should only
>> be updated when the current 'cs' is greater than existing 'max_cs'. So,
>> fix the condition accordingly.
> 
> Good catch. Current code can only work with one chip-select.
> 
> Reviewed-by: Théo Lebrun <theo.lebrun@bootlin.com>
> 
> Maybe we should error out if we don't enter the loop, ie if we have no
> flash declared?
>   - Before your patch, cqspi->num_chipselect was set to num-cs DT prop or
>     CQSPI_MAX_CHIPSELECT as fallback.
>   - After your patch, cqspi->num_chipselect is set to one.
> 
> In neither case do we get an error if no flash is defined in DT.
> 
> We could either return some error code or set cqspi->num_chipselect=0
> which will lead to spi_register_controller() to fail [0].

Yeah, we can use max_cs for this. Initiate max_cs with -1, check if it's 
still negative after the loop part - if yes, then it didn't enter the 
loop as there was no flash declared - return failure. I'll add this 
logic in v2. Thanks!

Regards,
Santhosh.

> 
> [0]: https://elixir.bootlin.com/linux/v6.16.4/source/drivers/spi/spi.c#L3322-L3329
> 
> Thanks,
> 
> --
> Théo Lebrun, Bootlin
> Embedded Linux and Kernel engineering
> https://bootlin.com
> 
>