Re: [PATCH 13/62] gpio: gpio-f7188x: Use new GPIO_LINE_DIRECTION

[Vaittinen, Matti]

Hello Geert,

On Wed, 2019-11-06 at 13:38 +0100, Geert Uytterhoeven wrote:
> Hi Matti,
> 
> On Wed, Nov 6, 2019 at 7:45 AM Vaittinen, Matti
> <Matti.Vaittinen@xxxxxxxxxxxxxxxxx> wrote:
> > On Wed, 2019-11-06 at 06:34 +0100, Simon Guinot wrote:
> > > On Tue, Nov 05, 2019 at 12:16:03PM +0200, Matti Vaittinen wrote:
> > > > It's hard for occasional GPIO code reader/writer to know if
> > > > values
> > > > 0/1
> > > > equal to IN or OUT. Use defined GPIO_LINE_DIRECTION_IN and
> > > > GPIO_LINE_DIRECTION_OUT to help them out.
> > > > 
> > > > Signed-off-by: Matti Vaittinen <
> > > > matti.vaittinen@xxxxxxxxxxxxxxxxx>
> > > > ---
> > > >  drivers/gpio/gpio-f7188x.c | 5 ++++-
> > > >  1 file changed, 4 insertions(+), 1 deletion(-)
> > > > 
> > > > diff --git a/drivers/gpio/gpio-f7188x.c b/drivers/gpio/gpio-
> > > > f7188x.c
> > > > index fdc639f856f1..cadd02993539 100644
> > > > --- a/drivers/gpio/gpio-f7188x.c
> > > > +++ b/drivers/gpio/gpio-f7188x.c
> > > > @@ -250,7 +250,10 @@ static int
> > > > f7188x_gpio_get_direction(struct
> > > > gpio_chip *chip, unsigned offset)
> > > > 
> > > >     superio_exit(sio->addr);
> > > > 
> > > > -   return !(dir & 1 << offset);
> > > > +   if (dir & 1 << offset)
> > > > +           return GPIO_LINE_DIRECTION_OUT;
> > > > +
> > > > +   return GPIO_LINE_DIRECTION_IN
> > > 
> > > Hi Matti,
> > > 
> > > I am probably missing something but I can't find
> > > GPIO_LINE_DIRECTION_IN
> > > and GPIO_LINE_DIRECTION_OUT defined anywhere.
> > 
> > Sorry. I accidentally sent the patch 01/62 to limited audience -
> > and
> > also messed up the message-ID from the series so threading messages
> > is
> > probably not working :( I did resend the patch adding defines to
> > all
> > reviewers yesterday - title should be "[RESEND PATCH 01/62] gpio:
> > Add
> > definition for GPIO direction".
> > 
> > > Besides I am an occasional code reader/writer and I find the
> > > original
> > > code easy to understand.
> > 
> > Glad to hear that. When I read code:
> > 
> > return !(dir & 1 << offset);
> > 
> > It's impossible for me to tell if dir having bit at offset 'offset'
> > set
> > means IN or OUT - I know the meaning of code, it checks this bit
> > for
> > in/out - but which dir value is IN and which is OUT?
> > 
> > When this is written as:
> > 
> >         if (dir & 1 << offset)
> >                 return GPIO_LINE_DIRECTION_OUT;
> > 
> >         return GPIO_LINE_DIRECTION_IN
> > 
> > it get's quite obvious even for me that having the matching bit set
> > means direction to be OUT.
> 
> "suggest parentheses around... " warning?

I don't think I saw that. Also, simple test:

#include <stdio.h>

int main(int argc, char *argv[])
{

	int offset = argc;

	if (4 & 1 << offset)
		printf("foo\n");

	printf("%d", !(4 & 1 << offset));
	return !(4 & 1 << offset);

}

and building with gcc -Wall does not show any warnings.

nor did compilation for the driver:
[mvaittin@localhost linux]$ make ARCH=arm CROSS_COMPILE=${CC}
LOADADDR=0x80008000 
  CALL    scripts/checksyscalls.sh
  CALL    scripts/atomic/check-atomics.sh
  CHK     include/generated/compile.h
  CC [M]  drivers/gpio/gpio-f7188x.o
  Kernel: arch/arm/boot/Image is ready
  Kernel: arch/arm/boot/zImage is ready
  Building modules, stage 2.
  MODPOST 1282 modules
  LD [M]  drivers/gpio/gpio-f7188x.ko
[mvaittin@localhost linux]$ 


In any case, my intention was to keep the logic exactly same - except
for the documented change for cases where I changed bit-position
specific return value to 1.

Br,
	Matti Vaittinen