Re: [PATCH v3 1/2] phy: Add new PHY attribute max_link_rate and APIs to get/set PHY attributes

[Kishon Vijay Abraham I]

On 7/13/2020 4:41 PM, Vinod Koul wrote:
> On 13-07-20, 11:38, Swapnil Jakhade wrote:
>> Add new PHY attribute max_link_rate to struct phy_attrs.
>> Add a pair of PHY APIs to get/set all the PHY attributes.
>> Use phy_set_attrs() to set attribute values in the PHY provider driver.
>> Use phy_get_attrs() to get attribute values in the controller driver.
>>
>> Signed-off-by: Yuti Amonkar <yamonkar@xxxxxxxxxxx>
>> Signed-off-by: Swapnil Jakhade <sjakhade@xxxxxxxxxxx>
>> ---
>>  include/linux/phy/phy.h | 22 ++++++++++++++++++++++
>>  1 file changed, 22 insertions(+)
>>
>> diff --git a/include/linux/phy/phy.h b/include/linux/phy/phy.h
>> index bcee8eba62b3..7fb59359ab7b 100644
>> --- a/include/linux/phy/phy.h
>> +++ b/include/linux/phy/phy.h
>> @@ -115,10 +115,12 @@ struct phy_ops {
>>  /**
>>   * struct phy_attrs - represents phy attributes
>>   * @bus_width: Data path width implemented by PHY
>> + * @max_link_rate: Maximum link rate supported by PHY (in Mbps)
>>   * @mode: PHY mode
>>   */
>>  struct phy_attrs {
>>  	u32			bus_width;
>> +	u32			max_link_rate;
>>  	enum phy_mode		mode;
>>  };
>>  
>> @@ -231,6 +233,16 @@ static inline void phy_set_bus_width(struct phy *phy, int bus_width)
>>  {
>>  	phy->attrs.bus_width = bus_width;
>>  }
>> +
>> +static inline void phy_get_attrs(struct phy *phy, struct phy_attrs *attrs)
>> +{
>> +	memcpy(attrs, &phy->attrs, sizeof(struct phy_attrs));
>> +}
>> +
>> +static inline void phy_set_attrs(struct phy *phy, struct phy_attrs attrs)
>> +{
>> +	memcpy(&phy->attrs, &attrs, sizeof(struct phy_attrs));
>> +}
> 
> we already have APIs for mode and bus_width so why not add one for
> link_rate and call them?
> 
> Also I see you are using phy_set_attrs() in second patch, why add
> phy_get_attrs() when we have no user?

One of the factors to consider is who will set the attributes; it could either
be phy provider (like 2/2 of this series) or phy consumer (factors like PCIe
speed, lane are usually negotiated with the phy consumer).

Now if phy provider is setting/getting the attributes, then
phy_set_attrs/phy_get_attrs should be protected by mutex. We don't want to be
updating attributes when phy consumer is doing some phy ops.

If phy_consumer is updating attributes, it could directly access the phy
attributes if it's updating within one of those phy ops. Don't really see a
need for using an API to update the attributes then.

However if it's updating outside the phy_ops, then it would still make sense to
use the APIs to update attributes with all those mutex protection.

Regards
Kishon
> 
>>  struct phy *phy_get(struct device *dev, const char *string);
>>  struct phy *phy_optional_get(struct device *dev, const char *string);
>>  struct phy *devm_phy_get(struct device *dev, const char *string);
>> @@ -389,6 +401,16 @@ static inline void phy_set_bus_width(struct phy *phy, int bus_width)
>>  	return;
>>  }
>>  
>> +static inline void phy_get_attrs(struct phy *phy, struct phy_attrs *attrs)
>> +{
>> +	return;
>> +}
>> +
>> +static inline void phy_set_attrs(struct phy *phy, struct phy_attrs attrs)
>> +{
>> +	return;
>> +}
>> +
>>  static inline struct phy *phy_get(struct device *dev, const char *string)
>>  {
>>  	return ERR_PTR(-ENOSYS);
>> -- 
>> 2.26.1
>