RE: [PATCH v4 3/3] i2c: Extend i2c_device_get_match_data() to support i2c sysfs

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Andy Shevchenko,

Thanks for the feedback.

> Subject: Re: [PATCH v4 3/3] i2c: Extend i2c_device_get_match_data() to
> support i2c sysfs
> 
> On Wed, Aug 02, 2023 at 12:23:17PM +0100, Biju Das wrote:
> > Extend i2c_device_get_match_data() to i2c sysfs interface by changing
> > i2c_of_match_device_sysfs() to non-static and call this function as a
> > fallback for i2c_get_match_data_helper().
> 
> ...
> 
> >  	/* TODO: use i2c_verify_client() when it accepts const pointer */
> >  	const struct i2c_client *client = (dev->type ==
> &i2c_client_type) ?
> >  					  to_i2c_client(dev) : NULL;
> > +	const void *data;
> 
> Oh, yeah, definitely, please follow advice for previous patch I have
> given.

OK.

> 
> >  	if (!client || !dev->driver)
> >  		return NULL;
> >
> > -	return i2c_get_match_data_helper(client);
> > +	data = i2c_get_match_data_helper(client);
> > +	if (data)
> > +		return data;
> > +
> > +	if (dev->driver->of_match_table) {
> > +		const struct of_device_id *match;
> > +
> > +		match = i2c_of_match_device_sysfs(dev->driver-
> >of_match_table,
> > +						  client);
> 
> So, with this it makes sense to have
> 
> 	strucct device_driver *drv = dev->driver;
> 	...
> 	if (!drv)
> 		return NULL;
> 
> The above for the first patch even. And here

Agreed.

> 
> 		match = i2c_of_match_device_sysfs(drv->of_match_table,
> client);
> 
> > +		if (match)
> > +			return match->data;
> > +	}
> 
> > +	return data;
> 
> This actually should be
> 
> 	return NULL;
> 
> to avoid (potentially) stale pointers.

OK.

> 
> >  }
> 
> ...
> 
> > -static const struct of_device_id*
> > +const struct of_device_id *
> >  i2c_of_match_device_sysfs(const struct of_device_id *matches,
> > -				  struct i2c_client *client)
> > +			  const struct i2c_client *client)
> >  {
> >  	const char *name;
> >
> > @@ -141,6 +141,7 @@ i2c_of_match_device_sysfs(const struct
> > of_device_id *matches,
> >
> >  	return NULL;
> >  }
> > +EXPORT_SYMBOL_GPL(i2c_of_match_device_sysfs);
> >
> >  const struct of_device_id
> >  *i2c_of_match_device(const struct of_device_id *matches, diff --git
> > a/drivers/i2c/i2c-core.h b/drivers/i2c/i2c-core.h index
> > 1247e6e6e975..e4d397b67989 100644
> > --- a/drivers/i2c/i2c-core.h
> > +++ b/drivers/i2c/i2c-core.h
> > @@ -82,8 +82,17 @@ static inline void
> > i2c_acpi_remove_space_handler(struct i2c_adapter *adapter) {
> >
> >  #ifdef CONFIG_OF
> >  void of_i2c_register_devices(struct i2c_adapter *adap);
> > +const struct of_device_id *
> > +i2c_of_match_device_sysfs(const struct of_device_id *matches,
> > +			  const struct i2c_client *client);
> >  #else
> >  static inline void of_i2c_register_devices(struct i2c_adapter *adap)
> > { }
> > +static inline const struct of_device_id *
> > +i2c_of_match_device_sysfs(const struct of_device_id *matches,
> > +			  const struct i2c_client *client) {
> > +	return NULL;
> > +}
> >  #endif
> >  extern struct notifier_block i2c_of_notifier;
> 
> That's what above I meant to see as a separate change.

OK will create a separate patch for this change.

Cheers,
Biju




[Index of Archives]     [Linux GPIO]     [Linux SPI]     [Linux Hardward Monitoring]     [LM Sensors]     [Linux USB Devel]     [Linux Media]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux