Re: [PATCH] HSI: omap_ssi_core: move function declaration to header file

[Sebastian Reichel]

Hi,

On Sat, Sep 24, 2016 at 10:58:15AM +0200, Arnd Bergmann wrote:
> On Saturday, September 24, 2016 1:11:56 PM CEST Baoyou Xie wrote:
> > ---
> >  drivers/hsi/clients/ssi_protocol.c | 2 --
> >  include/linux/hsi/hsi.h            | 1 +
> >  2 files changed, 1 insertion(+), 2 deletions(-)
> > 
> > diff --git a/drivers/hsi/clients/ssi_protocol.c b/drivers/hsi/clients/ssi_protocol.c
> > index 6031cd1..fe86444 100644
> > --- a/drivers/hsi/clients/ssi_protocol.c
> > +++ b/drivers/hsi/clients/ssi_protocol.c
> > @@ -45,8 +45,6 @@
> >  #include <linux/hsi/hsi.h>
> >  #include <linux/hsi/ssi_protocol.h>
> >  
> > -void ssi_waketest(struct hsi_client *cl, unsigned int enable);
> > -
> >  #define SSIP_TXQUEUE_LEN       100
> >  #define SSIP_MAX_MTU           65535
> >  #define SSIP_DEFAULT_MTU       4000
> > 
> 
> Maybe the declaration should be moved to include/linux/hsi/ssi_protocol.h
> instead of include/linux/hsi/hsi.h?

ssi_protocol.h is not included by the ssi (controller) driver, so
that will not fix the warning.

> It's not completely clear as the exported function is a bit of a
> layering violation.

Yes it is and the code contains a^W lots of FIXME about this. I have
not yet tried to remove it as suggested in the FIXME (which predates
my maintenance, but it's probably not possible without modifying the
modem FW).

I'm fine with adding it the function prototype to hsi.h to fix the
warning, but please add a comment, that it's a layering violation
and should not be used.

-- Sebastian

Attachment: signature.asc
Description: PGP signature