[PATCH 4/8] spi: fsl-qspi: add a clock disable quirk

Alex Elder elder at riscstar.com
Tue Oct 21 21:34:23 PDT 2025


On 10/20/25 2:13 PM, Frank Li wrote:
> On Mon, Oct 20, 2025 at 11:51:47AM -0500, Alex Elder wrote:
>> The SpacemiT K1 SoC QSPI implementation needs to avoid shutting
>> off the clock when changing its rate.  Add a new quirk to indicate
>> the clock should not be disabled/enabled when changing its rate
>> for operations.
>>
>> Signed-off-by: Alex Elder <elder at riscstar.com>
>> ---
>>   drivers/spi/spi-fsl-qspi.c | 21 +++++++++++++++++----
>>   1 file changed, 17 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/spi/spi-fsl-qspi.c b/drivers/spi/spi-fsl-qspi.c
>> index 1e27647dd2a09..703a7df394c00 100644
>> --- a/drivers/spi/spi-fsl-qspi.c
>> +++ b/drivers/spi/spi-fsl-qspi.c
>> @@ -197,6 +197,11 @@
>>    */
>>   #define QUADSPI_QUIRK_USE_TDH_SETTING	BIT(5)
>>
>> +/*
>> + * Do not disable the "qspi" clock when changing its rate.
>> + */
>> +#define QUADSPI_QUIRK_NO_CLK_DISABLE	BIT(6)
> 
> NO_CLK_DISALBE look likes not clk disable capability. Maybe
> 
> QUADSPI_QUIRK_SKIP_CLK_DISABLE

OK, that's better.

> 
>> +
>>   struct fsl_qspi_devtype_data {
>>   	unsigned int rxfifo;
>>   	unsigned int txfifo;
>> @@ -306,6 +311,11 @@ static inline int needs_tdh_setting(struct fsl_qspi *q)
>>   	return q->devtype_data->quirks & QUADSPI_QUIRK_USE_TDH_SETTING;
>>   }
>>
>> +static inline int needs_clk_disable(struct fsl_qspi *q)
> 
> bool type?

Yes I agree with this suggestion.  However all of the other
needs_*() functions return int and are marked inline (neither
of which I would normally do).

You want me to add a patch to update the others too?  If I
do that it will look more like this:

static bool needs_swap_endian(struct fsl_qspi *q)
{
         return !!(q->devtype_data->quirks & QUADSPI_QUIRK_SWAP_ENDIAN);
}


					-Alex

> 
> Frank
> 
>> +{
>> +	return !(q->devtype_data->quirks & QUADSPI_QUIRK_NO_CLK_DISABLE);
>> +}
>> +
>>   /*
>>    * An IC bug makes it necessary to rearrange the 32-bit data.
>>    * Later chips, such as IMX6SLX, have fixed this bug.
>> @@ -536,15 +546,18 @@ static void fsl_qspi_select_mem(struct fsl_qspi *q, struct spi_device *spi,
>>   	if (needs_4x_clock(q))
>>   		rate *= 4;
>>
>> -	fsl_qspi_clk_disable_unprep(q);
>> +	if (needs_clk_disable(q))
>> +		fsl_qspi_clk_disable_unprep(q);
>>
>>   	ret = clk_set_rate(q->clk, rate);
>>   	if (ret)
>>   		return;
>>
>> -	ret = fsl_qspi_clk_prep_enable(q);
>> -	if (ret)
>> -		return;
>> +	if (needs_clk_disable(q)) {
>> +		ret = fsl_qspi_clk_prep_enable(q);
>> +		if (ret)
>> +			return;
>> +	}
>>
>>   	q->selected = spi_get_chipselect(spi, 0);
>>
>> --
>> 2.48.1
>>




More information about the linux-riscv mailing list