Re: [PATCH v4 2/5] mfd: mp2629: Add support for mps battery charger

[Jonathan Cameron <jic23@kernel.org>]

On Fri, 27 Mar 2020 10:22:21 +0000
Lee Jones <lee.jones@linaro.org> wrote:

> Saravanan, Jonathan,
> 
> On Fri, 27 Mar 2020, saravanan sekar wrote:
> > On 27/03/20 8:55 am, Lee Jones wrote:  
> > > On Sun, 22 Mar 2020, Saravanan Sekar wrote:
> > >   
> > > > mp2629 is a highly-integrated switching-mode battery charge management
> > > > device for single-cell Li-ion or Li-polymer battery.
> > > > 
> > > > Add MFD core enables chip access for ADC driver for battery readings,
> > > > and a power supply battery-charger driver
> > > > 
> > > > Signed-off-by: Saravanan Sekar <sravanhome@gmail.com>
> > > > ---
> > > >   drivers/mfd/Kconfig        |   9 +++
> > > >   drivers/mfd/Makefile       |   2 +
> > > >   drivers/mfd/mp2629.c       | 116 +++++++++++++++++++++++++++++++++++++
> > > >   include/linux/mfd/mp2629.h |  22 +++++++
> > > >   4 files changed, 149 insertions(+)
> > > >   create mode 100644 drivers/mfd/mp2629.c
> > > >   create mode 100644 include/linux/mfd/mp2629.h  
> 
> [...]
> 
> > > > +static int mp2629_probe(struct i2c_client *client)
> > > > +{
> > > > +	struct mp2629_info *info;  
> > > Call this ddata instead of info.  
> > Not sure the reason, I will do.  
> 
> Because this is device data.  Info is too loose of a definition.
> 
> > > > +	struct resource	*resources;
> > > > +	int ret;
> > > > +	int i;
> > > > +
> > > > +	info = devm_kzalloc(&client->dev, sizeof(*info), GFP_KERNEL);
> > > > +	if (!info)
> > > > +		return -ENOMEM;
> > > > +
> > > > +	info->dev = &client->dev;
> > > > +	i2c_set_clientdata(client, info);
> > > > +
> > > > +	info->regmap = devm_regmap_init_i2c(client, &mp2629_regmap_config);
> > > > +	if (IS_ERR(info->regmap)) {
> > > > +		dev_err(info->dev, "Failed to allocate regmap!\n");
> > > > +		return PTR_ERR(info->regmap);
> > > > +	}
> > > > +
> > > > +	for (i = 0; i < MP2629_MFD_MAX; i++) {
> > > > +		mp2629mfd[i].platform_data = &info->regmap;
> > > > +		mp2629mfd[i].pdata_size = sizeof(info->regmap);  
> > > You don't need to store this in platform data as well.
> > > 
> > > You already have it in device data (ddata [currently 'info']).  
> > 
> > "The IIO parts seems fine (minor comments inline) but I'm not keep on
> > directly accessing the internals of the mfd device info structure.
> > To my mind that should be opaque to the child drivers so as to provide
> > clear structure to any such accesses.
> > 
> > This mess in layering with the children directly using the parents
> > regmap is a little concerning. It means that the 3 drivers
> > really aren't very well separated and can't really be reviewed
> > independently (not a good thing)."
> > 
> > This is the review comments form Jonathan on V2, not to access parent data
> > structure directly.
> > Am I misunderstood his review comments? please suggest the better option to
> > follow as like in V2
> > or V2 + some improvements or V4 + improvements?  
> 
> I will take this up with Jonathan separately if necessary.
> 
> For your FYI (and Jonathan if he's Cc'ed), it's very common for a
> child of an MFD to acquire resources from their parent.  That is the
> point of a lot of MFDs, to obtain and register shared resources and
> pass them onto their offspring.  There are 10's of examples of this.
> 
> Things like Regmaps aren't platform data, they are device/driver data,
> which is usually passed though platform_set_drvdata().
> 

Fair enough.  It seemed a bit messy to have full visibility of the
parent driver structures just to access one element.

What I was actually meaning to suggest was a couple of wrapper functions,
not passing the regmap separately, but I guess that doesn't really make
any difference.  

So a read / write wrapper that just takes an abstract ddata pointer.

Fair enough if you think that's an unnecessary bit of abstraction.

Jonathan

> [...]
> 
> > > > + */
> > > > +
> > > > +#ifndef __MP2629_H__
> > > > +#define __MP2629_H__
> > > > +
> > > > +#include <linux/types.h>
> > > > +
> > > > +struct device;
> > > > +struct regmap;  
> > > Why not just add the includes?  
> > Some more shared enum added in ADC driver  
> 
> Sorry?
> 
> > > > +struct mp2629_info {
> > > > +	struct device *dev;
> > > > +	struct regmap *regmap;
> > > > +};
> > > > +
> > > > +#endif  
>