Marc
Thanks for the comments
On 5/8/19 9:35 AM, Marc Kleine-Budde wrote:
> On 3/19/19 6:26 PM, Dan Murphy wrote:
>> Create a m_can platform framework that peripheral
>> devices can register to and use common code and register sets.
>> The peripheral devices may provide read/write and configuration
>> support of the IP.
>>
>> Acked-by: Wolfgang Grandegger <wg@xxxxxxxxxxxxxx>
>> Signed-off-by: Dan Murphy <dmurphy@xxxxxx>
>
> [...]
>
>> -/* m_can private data structure */
>> -struct m_can_priv {
>> - struct can_priv can; /* must be the first member */
>> - struct napi_struct napi;
>> - struct net_device *dev;
>> - struct device *device;
>> - struct clk *hclk;
>> - struct clk *cclk;
>> - void __iomem *base;
>> - u32 irqstatus;
>> - int version;
>> -
>> - /* message ram configuration */
>> - void __iomem *mram_base;
>> - struct mram_cfg mcfg[MRAM_CFG_NUM];
>> -};
>> +static u32 m_can_read(struct m_can_priv *priv, enum m_can_reg reg)
>> +{
>> + if (priv->ops->read_reg)
>> + return priv->ops->read_reg(priv, reg);
>> + else
>> + return -EINVAL;
>> +}
>
> How do you plan to check the return value here?
> What's the difference between a register value of 0xffffffe9 and
> returning -EINVAL?
>
Good point. I could just inline this and return whatever is sent from the callback
and as you said allow a backtrace to happen if read_reg is invalid.
>>
>> -static inline u32 m_can_read(const struct m_can_priv *priv, enum m_can_reg reg)
>> +static int m_can_write(struct m_can_priv *priv, enum m_can_reg reg, u32 val)
>> {
>> - return readl(priv->base + reg);
>> + if (priv->ops->write_reg)
>> + return priv->ops->write_reg(priv, reg, val);
>> + else
>> + return -EINVAL;
>> }
>
> I don't see anyone checking the return value. Better just dereference
> the pointer and the kernel will produce a nice backtrace.
>
> Same should be done for all read and write variants.
>
I will need to go through this and see if there is any caller checking the return. But
I think you are correct. If thats true I will just change this to a void, inline the function
and allow a backtrace if the callback is null
Dan
> Marc
>
