[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20210521143925.xjmwluaoyer322lg@pengutronix.de>
Date: Fri, 21 May 2021 16:39:25 +0200
From: Uwe Kleine-König <u.kleine-koenig@...gutronix.de>
To: Dan Carpenter <dan.carpenter@...cle.com>
Cc: Lee Jones <lee.jones@...aro.org>, Ian Abbott <abbotti@....co.uk>,
linux-kernel@...r.kernel.org,
H Hartley Sweeten <hsweeten@...ionengravers.com>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Thierry Reding <thierry.reding@...il.com>,
"David A. Schleef" <ds@...leef.org>,
Mori Hess <fmhess@...rs.sourceforge.net>,
Truxton Fulton <trux@...ode1.truxton.com>,
linux-staging@...ts.linux.dev, linux-pwm@...r.kernel.org
Subject: Re: [PATCH 5/6] comedi: drivers: ni_mio_common: Move
'range_ni_E_ao_ext' to where it is used
On Fri, May 21, 2021 at 02:54:31PM +0300, Dan Carpenter wrote:
> On Fri, May 21, 2021 at 08:26:35AM +0100, Lee Jones wrote:
> > On Thu, 20 May 2021, Ian Abbott wrote:
> >
> > > On 20/05/2021 13:25, Lee Jones wrote:
> > > > ... and mark it as __maybe_unused since not all users of the
> > > > header file reference it.
> > > >
> > > > Fixes the following W=1 kernel build warning(s):
> > > >
> > > > drivers/staging/comedi/drivers/ni_mio_common.c:163:35: warning: ‘range_ni_E_ao_ext’ defined but not used [-Wunused-const-variable=]
> > > >
> > > > Cc: Ian Abbott <abbotti@....co.uk>
> > > > Cc: H Hartley Sweeten <hsweeten@...ionengravers.com>
> > > > Cc: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
> > > > Cc: Thierry Reding <thierry.reding@...il.com>
> > > > Cc: "Uwe Kleine-König" <u.kleine-koenig@...gutronix.de>
> > > > Cc: Lee Jones <lee.jones@...aro.org>
> > > > Cc: "David A. Schleef" <ds@...leef.org>
> > > > Cc: Mori Hess <fmhess@...rs.sourceforge.net>
> > > > Cc: Truxton Fulton <trux@...xton.com>
> > > > Cc: linux-staging@...ts.linux.dev
> > > > Cc: linux-pwm@...r.kernel.org
> > > > Signed-off-by: Lee Jones <lee.jones@...aro.org>
> > > > ---
> > > > drivers/comedi/drivers/ni_mio_common.c | 9 ---------
> > > > drivers/comedi/drivers/ni_stc.h | 9 ++++++++-
> > > > 2 files changed, 8 insertions(+), 10 deletions(-)
> > > >
> > > > diff --git a/drivers/comedi/drivers/ni_mio_common.c b/drivers/comedi/drivers/ni_mio_common.c
> > > > index 4f80a4991f953..37615b4e2c10d 100644
> > > > --- a/drivers/comedi/drivers/ni_mio_common.c
> > > > +++ b/drivers/comedi/drivers/ni_mio_common.c
> > > > @@ -160,15 +160,6 @@ static const struct comedi_lrange range_ni_M_ai_628x = {
> > > > }
> > > > };
> > > > -static const struct comedi_lrange range_ni_E_ao_ext = {
> > > > - 4, {
> > > > - BIP_RANGE(10),
> > > > - UNI_RANGE(10),
> > > > - RANGE_ext(-1, 1),
> > > > - RANGE_ext(0, 1)
> > > > - }
> > > > -};
> > > > -
> > > > static const struct comedi_lrange *const ni_range_lkup[] = {
> > > > [ai_gain_16] = &range_ni_E_ai,
> > > > [ai_gain_8] = &range_ni_E_ai_limited,
> > > > diff --git a/drivers/comedi/drivers/ni_stc.h b/drivers/comedi/drivers/ni_stc.h
> > > > index fbc0b753a0f59..0822e65f709dd 100644
> > > > --- a/drivers/comedi/drivers/ni_stc.h
> > > > +++ b/drivers/comedi/drivers/ni_stc.h
> > > > @@ -1137,6 +1137,13 @@ struct ni_private {
> > > > u8 rgout0_usage;
> > > > };
> > > > -static const struct comedi_lrange range_ni_E_ao_ext;
> > > > +static const struct comedi_lrange __maybe_unused range_ni_E_ao_ext = {
> > > > + 4, {
> > > > + BIP_RANGE(10),
> > > > + UNI_RANGE(10),
> > > > + RANGE_ext(-1, 1),
> > > > + RANGE_ext(0, 1)
> > > > + }
> > > > +};
> > > > #endif /* _COMEDI_NI_STC_H */
> > > >
> > >
> > > The "ni_stc.h" header is also included by "ni_mio_cs.c" which doesn't need
> > > `range_ni_E_ao_ext` (admittedly, it was already pulling in a "tentative"
> > > definition of the variable).
> > >
> > > Thinking about it, I think it's probably better to move `range_ni_E_ao_ext`
> > > from "ni_mio_common.c" into *both* "ni_atmio.c" and "ni_pcimio.c" (I think
> > > we can live with the small amount of duplication), and to remove the
> > > tentative definition from "ni_stc.h".
> >
> > Happy to rework.
> >
> > Am I taking this or Uwe's suggestion?
>
> You should probably take Ian's suggestion because he is the maintainer
> and I really doubt Uwe's will build. :P But Uwe is right that
> including .c files is ugly.
I just tried
diff --git a/drivers/comedi/drivers/ni_mio_common.c b/drivers/comedi/drivers/ni_mio_common.c
index 4f80a4991f95..3b37c8b00ecd 100644
--- a/drivers/comedi/drivers/ni_mio_common.c
+++ b/drivers/comedi/drivers/ni_mio_common.c
@@ -160,7 +160,7 @@ static const struct comedi_lrange range_ni_M_ai_628x = {
}
};
-static const struct comedi_lrange range_ni_E_ao_ext = {
+const struct comedi_lrange range_ni_E_ao_ext = {
4, {
BIP_RANGE(10),
UNI_RANGE(10),
diff --git a/drivers/comedi/drivers/ni_stc.h b/drivers/comedi/drivers/ni_stc.h
index fbc0b753a0f5..026d215f2ef2 100644
--- a/drivers/comedi/drivers/ni_stc.h
+++ b/drivers/comedi/drivers/ni_stc.h
@@ -1137,6 +1137,6 @@ struct ni_private {
u8 rgout0_usage;
};
-static const struct comedi_lrange range_ni_E_ao_ext;
+extern const struct comedi_lrange range_ni_E_ao_ext;
#endif /* _COMEDI_NI_STC_H */
and it built for me.
Best regards
Uwe
--
Pengutronix e.K. | Uwe Kleine-König |
Industrial Linux Solutions | https://www.pengutronix.de/ |
Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)
Powered by blists - more mailing lists