Re: [PATCH 1/3] intel_pstate: Use pr_fmt

From: Joe Perches
Date: Wed Apr 06 2016 - 11:13:07 EST


On Wed, 2016-04-06 at 07:51 -0700, Doug Smythies wrote:
> On 2016.04.05 02:44 Srinivas Pandruvada wrote:
> >
> > On Tue, 2016-04-05 at 13:28 -0700, Joe Perches wrote:
> > >
> > > Prefix the output using the more common kernel style.
> > >
> > > Signed-off-by: Joe Perches <joe@xxxxxxxxxxx>
> > Acked-by: Srinivas Pandruvada <srinivas.pandruvada@xxxxxxxxxxxxxxx>
> > >
> > > ---
> > >  drivers/cpufreq/intel_pstate.c | 18 ++++++++++--------
> > >  1 file changed, 10 insertions(+), 8 deletions(-)
> ...[cut, example left]...
>
> >
> > - pr_warn("intel_pstate: Turbo disabled by BIOS or
> > unavailable on processor\n");
> > + pr_warn("Turbo disabled by BIOS or unavailable on
> > processor\n");
> I do  not understand.
> The common and unique string "intel_pstate" was added on purpose
> so as to provide a way to easily extract the related message from
> an otherwise huge log file.
>

The more common kernel mechanism to prefix messages
is using a pr_fmt define like:

#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

This style is used ~1000 times in the kernel tree.

All of the pr_<level> macros are defined like:

#define pr_info(fmt, ...) \
printk(KERN_INFO pr_fmt(fmt), ##__VA_ARGS__)

so this prefixes all messages and means that any
new message added later will also be prefixed without
copy/paste defects or omission.