[RFC 3/3] include:linux: make ti_wilink_st like the rest

Savoy, Pavan pavan_savoy at ti.com
Thu Sep 16 21:24:31 UTC 2010


Greg,

> -----Original Message-----
> From: Greg KH [mailto:greg at kroah.com]
> Sent: Thursday, September 16, 2010 3:59 PM
> To: Savoy, Pavan
> Cc: gregkh at suse.de; alan at lxorguk.ukuu.org.uk; devel at driverdev.osuosl.org;
> linux-kernel at vger.kernel.org
> Subject: Re: [RFC 3/3] include:linux: make ti_wilink_st like the rest
> 
> On Fri, Sep 17, 2010 at 01:48:53AM +0530, Savoy, Pavan wrote:
> > > Also, why not just use 'u8' and friends instead?  That's the "proper"
> > > kernel types to be using, not the uint8_t mess that is not correct
> > > kernel types.
> >
> > Yes a mistake here too.
> > This should have been in the patch where I plan to combine all the headers
> into 1 header and have it in include/linux/
> > [the headers like st_core.h, st_kim.h, st_ll.h, bt_drv.h and fm and GPs
> headers]
> >
> > Yes, and the Bluetooth driver to go into drivers/Bluetooth/ and ST
> > driver (st_core.c/st_kim.c and st_ll.c) to go into drivers/char/ And
> > the upcoming FM V4L2 to go into the drivers/media/radio/
> >
> > Does this sound fine?
> 
> Yes.
> 
> > > > @@ -386,6 +392,7 @@ void st_ll_wakeup(struct st_data_s *);
> > > >
> > > >  /**
> > > >   * structures and declarations used by the st_core for FM packets
> > > > + * and GPS packets
> > > >   */
> > > >  struct fm_event_hdr {
> > > >  	unsigned char plen;
> > >
> > > Oh, is it?  That should go somewhere else.
> >
> > Plan is to have just 1 large header to be included by all protocol drivers
> (BT, FM and GPS)...
> > Is it a bad idea?
> 
> No, as long as all drivers need it.
> 
> > > >  #endif /* ST_H */
> > >
> > > I don't think this file is called "ST_H" here anymore.
> >
> > Yes, need your input here. This file will be made use by the rest of
> > the protocol drivers I plan to push patches for.
> > So does ti_wilink_st.h sound fine? And in include/linux? So that rest
> > of the drivers can be in their individual directories.
> 
> Yes, that name is fine, but it needs to be the same as the #ifdef
> guard as well.


I've re-worked on the patch;
I hope now it's close to ready to be in include/linux/?

> thanks,
> 
> greg k-h



More information about the devel mailing list