Re: [PATCH] staging: comedi: ni_mio_common.c: local functions shouldbe static

[Ian Abbott]

On 2012-06-06 17:22, H Hartley Sweeten wrote:
> On Wednesday, June 06, 2012 3:17 AM, Ian Abbott wrote:
> > On 2012-06-05 19:20, H Hartley Sweeten wrote:
> > > Local functions should be marked static to prevent them from
> > > being exposed globally.
> > >
> > > This quiets the following sparse warnings:
> > >
> > > warning: symbol 'ni_release_gpct_mite_channel' was not declared. Should it be static?
> > > warning: symbol 'ni_prime_channelgain_list' was not declared. Should it be static?
> > >
> > > Signed-off-by: H Hartley Sweeten<hsweeten@xxxxxxxxxxxxxxxxxxx>
> > > Cc: Ian Abbott<abbotti@xxxxxxxxx>
> > > Cc: Mori Hess<fmhess@xxxxxxxxxxxxxxxxxxxxx>
> > > Cc: Greg Kroah-Hartman<gregkh@xxxxxxxxxxxxxxxxxxx>
> > >
> > > ---
> > >
> > > diff --git a/drivers/staging/comedi/drivers/ni_mio_common.c b/drivers/staging/comedi/drivers/ni_mio_common.c
> > > index fd232bc..3b3a0b1 100644
> > > --- a/drivers/staging/comedi/drivers/ni_mio_common.c
> > > +++ b/drivers/staging/comedi/drivers/ni_mio_common.c
> > > @@ -644,7 +644,7 @@ static void ni_release_ao_mite_channel(struct comedi_device *dev)
> > >    #endif /*  PCIDMA */
> > >    }
> > >
> > > -void ni_release_gpct_mite_channel(struct comedi_device *dev,
> > > +static void ni_release_gpct_mite_channel(struct comedi_device *dev,
> > >    				  unsigned gpct_index)
> > >    {
> > >    #ifdef PCIDMA
> > > @@ -1880,7 +1880,7 @@ static int ni_ai_insn_read(struct comedi_device *dev,
> > >    	return insn->n;
> > >    }
> >
> > This results in a compiler warning about the unused function
> > ni_release_gpct_mite_channel when compiling ni_atmio.c.  It was unused
> > before of course, but now it is static, the compiler realizes it!
>
> Ugh... c code including c code... I didn't notice that until now.
>
> > It could be fixed for now by moving the '#ifdef PCIDMA' and matching
> > '#endif' to surround the whole function, though I guess we want to try
> > and get rid of the #ifdef's eventually.  It's tricky because
> > ni_mio_common.c isn't a standalone compilation unit, it's #include'd by
> > some other .c files.  We might need to have a go at splitting it up at
> > some point.
>
> The ni_mio_common.c file should just get turned into a library module.
>
> Right now it's a bit of a mess. The individual drivers define PCIDMA
> before including ni_mio_common.c. The only one that sets the define
> is ni_pcimio but it's also included by ni_atmio and ni_mio_cs.
>
> I'll look it over and see what might be done to fix it.

You have to be careful with module dependencies if doing that, as you 
don't want the ni_atmio module to depend on the mite and ni_tiocmd 
modules for example.

--
-=( Ian Abbott @ MEV Ltd.    E-mail: <abbotti@xxxxxxxxx>        )=-
-=( Tel: +44 (0)161 477 1898   FAX: +44 (0)161 718 3587         )=-
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/