[EXTERNAL] Re: [PATCH v4 4/5] soc: ti: pruss: Add helper functions to set GPI mode, MII_RT_event and XFR

Roger Quadros rogerq at kernel.org
Thu Mar 16 04:36:47 PDT 2023


Hi,

On 16/03/2023 13:05, Md Danish Anwar wrote:
> Hi Roger,
> 
> On 15/03/23 17:52, Roger Quadros wrote:
>>
>>
>> On 13/03/2023 13:11, MD Danish Anwar wrote:
>>> From: Suman Anna <s-anna at ti.com>
>>>
>>> The PRUSS CFG module is represented as a syscon node and is currently
>>> managed by the PRUSS platform driver. Add easy accessor functions to set
>>> GPI mode, MII_RT event enable/disable and XFR (XIN XOUT) enable/disable
>>> to enable the PRUSS Ethernet usecase. These functions reuse the generic
>>> pruss_cfg_update() API function.
>>>
>>> Signed-off-by: Suman Anna <s-anna at ti.com>
>>> Co-developed-by: Grzegorz Jaszczyk <grzegorz.jaszczyk at linaro.org>
>>> Signed-off-by: Grzegorz Jaszczyk <grzegorz.jaszczyk at linaro.org>
>>> Signed-off-by: Puranjay Mohan <p-mohan at ti.com>
>>> Signed-off-by: MD Danish Anwar <danishanwar at ti.com>
>>> ---
>>>  drivers/soc/ti/pruss.c           | 60 ++++++++++++++++++++++++++++++++
>>>  include/linux/remoteproc/pruss.h | 22 ++++++++++++
>>>  2 files changed, 82 insertions(+)
>>>
>>> diff --git a/drivers/soc/ti/pruss.c b/drivers/soc/ti/pruss.c
>>> index 26d8129b515c..2f04b7922ddb 100644
>>> --- a/drivers/soc/ti/pruss.c
>>> +++ b/drivers/soc/ti/pruss.c
>>> @@ -203,6 +203,66 @@ static int pruss_cfg_update(struct pruss *pruss, unsigned int reg,
>>>  	return regmap_update_bits(pruss->cfg_regmap, reg, mask, val);
>>>  }
>>>  
>>> +/**
>>> + * pruss_cfg_gpimode() - set the GPI mode of the PRU
>>> + * @pruss: the pruss instance handle
>>> + * @pru_id: id of the PRU core within the PRUSS
>>> + * @mode: GPI mode to set
>>> + *
>>> + * Sets the GPI mode for a given PRU by programming the
>>> + * corresponding PRUSS_CFG_GPCFGx register
>>> + *
>>> + * Return: 0 on success, or an error code otherwise
>>> + */
>>> +int pruss_cfg_gpimode(struct pruss *pruss, enum pruss_pru_id pru_id,
>>> +		      enum pruss_gpi_mode mode)
>>> +{
>>> +	if (pru_id < 0 || pru_id >= PRUSS_NUM_PRUS)
>>> +		return -EINVAL;
>>> +
>>> +	if (mode < 0 || mode > PRUSS_GPI_MODE_MAX)
>>> +		return -EINVAL;
>>> +
>>> +	return pruss_cfg_update(pruss, PRUSS_CFG_GPCFG(pru_id),
>>> +				PRUSS_GPCFG_PRU_GPI_MODE_MASK,
>>> +				mode << PRUSS_GPCFG_PRU_GPI_MODE_SHIFT);
>>> +}
>>> +EXPORT_SYMBOL_GPL(pruss_cfg_gpimode);
>>> +
>>> +/**
>>> + * pruss_cfg_miirt_enable() - Enable/disable MII RT Events
>>> + * @pruss: the pruss instance
>>> + * @enable: enable/disable
>>> + *
>>> + * Enable/disable the MII RT Events for the PRUSS.
>>> + *
>>> + * Return: 0 on success, or an error code otherwise
>>> + */
>>> +int pruss_cfg_miirt_enable(struct pruss *pruss, bool enable)
>>> +{
>>> +	u32 set = enable ? PRUSS_MII_RT_EVENT_EN : 0;
>>> +
>>> +	return pruss_cfg_update(pruss, PRUSS_CFG_MII_RT,
>>> +				PRUSS_MII_RT_EVENT_EN, set);
>>> +}
>>> +EXPORT_SYMBOL_GPL(pruss_cfg_miirt_enable);
>>> +
>>> +/**
>>> + * pruss_cfg_xfr_enable() - Enable/disable XIN XOUT shift functionality
>>> + * @pruss: the pruss instance
>>> + * @enable: enable/disable
>>> + * @mask: Mask for PRU / RTU
>>
>> You should not expect the user to provide the mask but only
>> the core type e.g. 
>>
>> enum pru_type {
>>         PRU_TYPE_PRU = 0,
>>         PRU_TYPE_RTU,
>>         PRU_TYPE_TX_PRU,
>>         PRU_TYPE_MAX,
>> };
>>
>> Then you figure out the mask in the function.
>> Also check for invalid pru_type and return error if so.
>>
> 
> Sure Roger, I will create a enum and take it as parameter in API. Based on
> these enum I will calculate mask and do XFR shifting inside the API
> pruss_cfg_xfr_enable().
> 
> There are two registers for XFR shift.
> 
> #define PRUSS_SPP_XFER_SHIFT_EN                 BIT(1)
> #define PRUSS_SPP_RTU_XFR_SHIFT_EN              BIT(3)
> 
> For PRU XFR shifting, the mask should be PRUSS_SPP_XFER_SHIFT_EN,
> for RTU shifting mask should be PRUSS_SPP_RTU_XFR_SHIFT_EN and for PRU and RTU
> shifting mask should be (PRUSS_SPP_XFER_SHIFT_EN | PRUSS_SPP_RTU_XFR_SHIFT_EN)
> 
> So the enum would be something like this.
> 
> /**
>  * enum xfr_shift_type - XFR shift type
>  * @XFR_SHIFT_PRU: Enables XFR shift for PRU
>  * @XFR_SHIFT_RTU: Enables XFR shift for RTU
>  * @XFR_SHIFT_PRU_RTU: Enables XFR shift for both PRU and RTU

This is not required. User can call the API twice. once for PRU and once for RTU.

>  * @XFR_SHIFT_MAX: Total number of XFR shift types available.
>  *
>  */
> 
> enum xfr_shift_type {
>         XFR_SHIFT_PRU = 0,
>         XFR_SHIFT_RTU,
>         XFR_SHIFT_PRU_RTU,
>         XFR_SHIFT_MAX,
> };

Why do you need this new enum definition?
We already have pru_type defined somewhere. You can move it to a public header
if not there yet.

enum pru_type {
         PRU_TYPE_PRU = 0,
         PRU_TYPE_RTU,
         PRU_TYPE_TX_PRU,
         PRU_TYPE_MAX,
};


> 
> In pruss_cfg_xfr_enable() API, I will use switch case, and for first three
> enums, I will calculate the mask.
> 
> If input is anything other than first three, I will retun -EINVAL. This will
> serve as check for valid xfr_shift_type.
> 
> The API will look like this.
> 
> int pruss_cfg_xfr_enable(struct pruss *pruss, enum xfr_shift_type xfr_type,
> 			 bool enable);
> {
> 	u32 mask;
> 
> 	switch (xfr_type) {
> 	case XFR_SHIFT_PRU:
> 		mask = PRUSS_SPP_XFER_SHIFT_EN;
> 		break;
> 	case XFR_SHIFT_RTU:
> 		mask = PRUSS_SPP_RTU_XFR_SHIFT_EN;
> 		break;
> 	case XFR_SHIFT_PRU_RTU:
> 		mask = PRUSS_SPP_XFER_SHIFT_EN | PRUSS_SPP_RTU_XFR_SHIFT_EN;
> 		break;
> 	default:
> 		return -EINVAL;
> 	}
> 
> 	u32 set = enable ? mask : 0;
> 
> 	return pruss_cfg_update(pruss, PRUSS_CFG_SPP, mask, set);
> }
> 
> This entire change I will keep as part of this patch only.
> 
> Please let me know if this looks OK to you.
> 
> 

cheers,
-roger



More information about the linux-arm-kernel mailing list