Re: [PATCH] spi: mediatek: Attempt to address style issues in spi-mt7621.c

[Stefan Roese]

Hi Armando,

On 14.03.19 12:13, Armando Miraglia wrote:
> My answers are in-line below. BTW bare with me as this is my attempt to get my
> feet wet in how to contribute to the linux kernel for my own pleasure and
> interest :)
>
> On Wed, Mar 13, 2019 at 03:34:54PM +0300, Dan Carpenter wrote:
> > On Wed, Mar 13, 2019 at 01:24:04PM +0100, Armando Miraglia wrote:
> > > Running Lindent on the mt7621-spi.c file in drivers/staging I noticed that the
> > > file contained style issues. This change attempts to address such style
> > > problems.
> >
> > Don't run lindent.  I think checkpatch.pl has a --fix option that might
> > be better, but once the code is merged then our standard become much
> > higher for follow up patches.
> >
> > > Signed-off-by: Armando Miraglia <armax@xxxxxxxxxx>
> > > ---
> > > NOTE: resend this patch to include all mainteners listed by get_mantainers.pl.
> > >   drivers/staging/mt7621-spi/spi-mt7621.c | 27 +++++++++++++------------
> > >   1 file changed, 14 insertions(+), 13 deletions(-)
> > >
> > > diff --git a/drivers/staging/mt7621-spi/spi-mt7621.c b/drivers/staging/mt7621-spi/spi-mt7621.c
> > > index b509f9fe3346..03d53845f8c5 100644
> > > --- a/drivers/staging/mt7621-spi/spi-mt7621.c
> > > +++ b/drivers/staging/mt7621-spi/spi-mt7621.c
> > > @@ -52,14 +52,14 @@
> > >   #define MT7621_LSB_FIRST	BIT(3)
> > >   
> > >   struct mt7621_spi {
> > > -	struct spi_master	*master;
> > > -	void __iomem		*base;
> > > -	unsigned int		sys_freq;
> > > -	unsigned int		speed;
> > > -	struct clk		*clk;
> > > -	int			pending_write;
> > > -
> > > -	struct mt7621_spi_ops	*ops;
> > > +	struct spi_master *master;
> > > +	void __iomem *base;
> > > +	unsigned int sys_freq;
> > > +	unsigned int speed;
> > > +	struct clk *clk;
> > > +	int pending_write;
> > > +
> > > +	struct mt7621_spi_ops *ops;
> >
> > The original is fine.  I don't encourage people to do fancy indenting
> > with their local variable declarations inside functions but for a struct
> > the declarations aren't going to change a lot so people can get fancy
> > if they want.
> Is there an explicit intent to deprecate Lindent in favor of checkpatch.pl
> --fix? If one would like to contribute to fixing the tooling for linting which
> of the two would be the right target for such an effort?
>
> > The problem with a local is if you need to add a new variable then you
> > have to re-indent a bunch of unrelated lines or have one out of
> > alignment line.  Most people know this intuitively so they don't get
> > fancy.
> >
> > >   };
> > >   
> > >   static inline struct mt7621_spi *spidev_to_mt7621_spi(struct spi_device *spi)
> > > @@ -303,7 +303,7 @@ static int mt7621_spi_setup(struct spi_device *spi)
> > >   	struct mt7621_spi *rs = spidev_to_mt7621_spi(spi);
> > >   
> > >   	if ((spi->max_speed_hz == 0) ||
> > > -		(spi->max_speed_hz > (rs->sys_freq / 2)))
> > > +	    (spi->max_speed_hz > (rs->sys_freq / 2)))
> >
> > Yeah.  Lindent is correct here.
>
> Funny enough, this is something I adjusted manually :)
>
> > >   		spi->max_speed_hz = (rs->sys_freq / 2);
> > >   
> > >   	if (spi->max_speed_hz < (rs->sys_freq / 4097)) {
> > > @@ -316,9 +316,10 @@ static int mt7621_spi_setup(struct spi_device *spi)
> > >   }
> > >   
> > >   static const struct of_device_id mt7621_spi_match[] = {
> > > -	{ .compatible = "ralink,mt7621-spi" },
> > > +	{.compatible = "ralink,mt7621-spi"},
> >
> > The original was better.
> >
> > >   	{},
> > >   };
> > > +
> > >   MODULE_DEVICE_TABLE(of, mt7621_spi_match);
> >
> > No need for a blank.  These are closely related.
>
> Ack.
>
> > >   
> > >   static int mt7621_spi_probe(struct platform_device *pdev)
> > > @@ -408,9 +409,9 @@ MODULE_ALIAS("platform:" DRIVER_NAME);
> > >   
> > >   static struct platform_driver mt7621_spi_driver = {
> > >   	.driver = {
> > > -		.name = DRIVER_NAME,
> > > -		.of_match_table = mt7621_spi_match,
> > > -	},
> > > +		   .name = DRIVER_NAME,
> > > +		   .of_match_table = mt7621_spi_match,
> > > +		   },
> >
> > The new indenting is very wrong.
>
> Ack. In fact, I was thinking this could be one target to fix the logic in
> Lindent to do this appropriately.
>
> I have a process question here: to post a change for the only accepted change I
> have in this patch should I send out a new patch?

Would it be possible for you to wait a bit with this minor cleanup?
As I'm preparing a patch to move this driver out of staging right
now. You can definitely follow-up with your cleanup, once this move
is done. Otherwise the move might be delayed even more.

Thanks,
Stefan