Re: [PATCH v2 05/15] hwmon: mark const init data with __initconst instead of __initdata

[Guenter Roeck <guenter.roeck@ericsson.com>]

On Sat, Mar 31, 2012 at 10:09:24AM -0400, Uwe Kleine-König wrote:
> Hello Guenter,
> 
> On Fri, Mar 30, 2012 at 08:54:42PM -0700, Guenter Roeck wrote:
> > On Fri, Mar 30, 2012 at 04:04:55PM -0400, Uwe Kleine-König wrote:
> > > As long as there is no other non-const variable marked __initdata in the
> > > same compilation unit it doesn't hurt. If there were one however
> > > compilation would fail with
> > > 
> > > 	error: $variablename causes a section type conflict
> > > 
> > > because a section containing const variables is marked read only and so
> > > cannot contain non-const variables.
> > > 
> > > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
> > > Cc: Jean Delvare <khali@linux-fr.org>
> > > Cc: Guenter Roeck <guenter.roeck@ericsson.com>
> > > Cc: lm-sensors@lm-sensors.org
> > > ---
> > > changes since (implicit) v1:
> > >  - drop wrong changes to drivers/hwmon/w83627hf.c
> > > 
> > >  drivers/hwmon/w83627ehf.c |   18 +++++++++---------
> > >  1 file changed, 9 insertions(+), 9 deletions(-)
> > > 
> > > diff --git a/drivers/hwmon/w83627ehf.c b/drivers/hwmon/w83627ehf.c
> > > index a25350c..54922ed 100644
> > > --- a/drivers/hwmon/w83627ehf.c
> > > +++ b/drivers/hwmon/w83627ehf.c
> > > @@ -2619,15 +2619,15 @@ static struct platform_driver w83627ehf_driver = {
> > >  static int __init w83627ehf_find(int sioaddr, unsigned short *addr,
> > >  				 struct w83627ehf_sio_data *sio_data)
> > >  {
> > > -	static const char __initdata sio_name_W83627EHF[] = "W83627EHF";
> > > -	static const char __initdata sio_name_W83627EHG[] = "W83627EHG";
> > > -	static const char __initdata sio_name_W83627DHG[] = "W83627DHG";
> > > -	static const char __initdata sio_name_W83627DHG_P[] = "W83627DHG-P";
> > > -	static const char __initdata sio_name_W83627UHG[] = "W83627UHG";
> > > -	static const char __initdata sio_name_W83667HG[] = "W83667HG";
> > > -	static const char __initdata sio_name_W83667HG_B[] = "W83667HG-B";
> > > -	static const char __initdata sio_name_NCT6775[] = "NCT6775F";
> > > -	static const char __initdata sio_name_NCT6776[] = "NCT6776F";
> > > +	static const char sio_name_W83627EHF[] __initconst = "W83627EHF";
> > > +	static const char sio_name_W83627EHG[] __initconst = "W83627EHG";
> > > +	static const char sio_name_W83627DHG[] __initconst = "W83627DHG";
> > > +	static const char sio_name_W83627DHG_P[] __initconst = "W83627DHG-P";
> > > +	static const char sio_name_W83627UHG[] __initconst = "W83627UHG";
> > > +	static const char sio_name_W83667HG[] __initconst = "W83667HG";
> > > +	static const char sio_name_W83667HG_B[] __initconst = "W83667HG-B";
> > > +	static const char sio_name_NCT6775[] __initconst = "NCT6775F";
> > > +	static const char sio_name_NCT6776[] __initconst = "NCT6776F";
> > >  
> > Applied.
> > 
> > Just wondering: Why not the following ?
> > 
> > > +	static const char __initconst sio_name_NCT6776[] = "NCT6776F";
> > 
> > It does not make a difference in the generated code, and appears to be
> > less confusing, at least to me.
> hmm, I thought it does, maybe it's compiler dependant?! At least
> gcc-4.4.info tells:
> 
> 	An attribute specifier list may appear immediately before the
> 	comma, `=' or semicolon terminating the declaration of an
> 	identifier other than a function definition.
> 
I tried with gcc 4.4.3; it did not make a difference.

> and include/linux/init.h has:
> 
> 	You should insert __initdata between the variable name and equal
> 	sign followed by value [...].
> 
Seems to be a clear statement.

> I seem to remember that placing the attribute at the wrong place for a
> function made gcc ignore it (or apply it so something unintended).
> 
Apparently not for gcc 4.4.3, but possibly for others.

Thanks,
Guenter