Mark Brown writes:
On Fri, Jun 19, 2020 at 01:31:18PM +0200, Lars Povlsen wrote:
>> +/*
>> + * The Designware SPI controller (referred to as master in the
>> + * documentation) automatically deasserts chip select when the tx fifo
>> + * is empty. The chip selects then needs to be driven by a CS override
>> + * register. enable is an active low signal.
>> + */
>> +static void dw_spi_sparx5_set_cs(struct spi_device *spi, bool nEnable)
>
>The value that is passed in here is the value that should be driven on
>the output pin, the driver should not be interpreting the value in any
>way here. Documenting it as nEnable adds a layer of confusion, and it
>may not be an active high signal depending on the system.
Ok, I will make the CS function more like the others.
>
>> + if (!nEnable) {
>> + /* Ensure CS toggles, so start off all disabled */
>> + regmap_write(dwsmscc->syscon, SPARX5_FORCE_VAL, ~0);
>> + /* CS override drive enable */
>> + regmap_write(dwsmscc->syscon, SPARX5_FORCE_ENA, 1);
>
>This should just be setting the value to whatever the core asked for it
>to be set to, the driver adding extra toggles is likely to disrupt
>things.
I will have a look at this again. But it was added for a reason. The
issue is that we have two different busses in front of the controller,
so we might need more settle time when switching interface.
Thank you for you comments,
Cheers
--
Lars Povlsen,
Microchip
