On Mon, Feb 01, 2016 at 09:24:12AM +0100, Ingo Molnar wrote:
> > So there are 37,000+ printk() calls in the kernel and 900+ ones in
> > arch/x86/
> > alone! Plus even after this patch there's 20 more printk()s left in
> > arch/x86/kernel/cpu/mcheck/...
> >
> > We don't want to create a churn of 10,000+ commits to convert them to
> > pr_*()
> > facilities...
> >
> > So we don't apply such 'conversion' patches unless it's done for a whole
> > subsystem
> > and done as part of a larger work with good reason, or if it's done as part
> > of
> > completely new facilities/drivers - and is done consistently.
> >
> > Even then it's dubious to convert: people keep re-adding printk()s (which
> > is a
> > perfectly fine facility), which generates never ending churn ...
Amen to that!
--
Regards/Gruss,
Boris.
ECO tip #101: Trim your mails when you reply.
