On Mon, 2018-11-05 at 06:13 +0000, Tony Chuang wrote:
> > -----Original Message-----
> > From: Joe Perches [mailto:[email protected]]
> > Sent: Sunday, November 04, 2018 12:29 PM
> > To: Tony Chuang; [email protected]; [email protected]
> > Cc: [email protected]; [email protected]; Pkshih; Andy
> > Huang; [email protected]
> > Subject: Re: [RFC v5 08/13] rtw88: debug files
> >
> > On Wed, 2018-10-31 at 18:12 +0800, [email protected] wrote
> > > From: Yan-Hsuan Chuang <[email protected]>
> > > debug files for Realtek 802.11ac wireless network chips
> > []
> > > diff --git a/drivers/net/wireless/realtek/rtw88/debug.h
> > b/drivers/net/wireless/realtek/rtw88/debug.h
> > []
> > > +#ifdef CONFIG_RTW88_DEBUG
> > []
> > > +#else
> > > +
> > > +static inline void rtw_dbg(struct rtw_dev *rtwdev, const char *fmt, ...)
> > > {}
> > > +static inline void rtw_info(struct rtw_dev *rtwdev, const char *fmt,
> > > ...) {}
> > > +static inline void rtw_warn(struct rtw_dev *rtwdev, const char *fmt,
> > > ...) {}
> > > +static inline void rtw_err(struct rtw_dev *rtwdev, const char *fmt, ...)
> > > {}
> >
> > I still think it's _very, very bad_, and extremely
> > unusual to turn off _all_ logging, even for error
> > conditions, based on a debug config flag.
> >
[]
> So I think we should move rtw_[err/warn/info] out of the flag.
> Should we need to send it in RFC v6 or in the next patch series?
Up to you.
I think you should look at all the rtw_err/rtw_warn/rtw_info uses
as many seem to be more like they should use rtw_dbg instead.
Also, I think the rtw_err/rtw_warn/rtw_info function calls aren't
particularly useful and these should be simple macros or static
inlines and not use %pV at all.
#define rtw_err(rtw_dev, fmt, ...) \
dev_err((rtw_dev)->dev, fmt, ##__VA_ARGS__)
etc
