Re: [tip:sched/urgent] sched: Restore printk sanity

[Ingo Molnar <mingo@elte.hu>]

* Pavel Machek <pavel@ucw.cz> wrote:

> On Sun 2009-12-20 18:00:38, tip-bot for Peter Zijlstra wrote:
> > Commit-ID:  6c9a74afc5d95321e33a727b3c563f3e7d6c4a5c
> > Gitweb:     http://git.kernel.org/tip/6c9a74afc5d95321e33a727b3c563f3e7d6c4a5c
> > Author:     Peter Zijlstra <peterz@infradead.org>
> > AuthorDate: Sun, 20 Dec 2009 14:23:57 +0100
> > Committer:  Ingo Molnar <mingo@elte.hu>
> > CommitDate: Sun, 20 Dec 2009 18:45:57 +0100
> > 
> > sched: Restore printk sanity
> > 
> > Revert the braindead pr_* crap. (Commit 663997d "sched: Use
> > pr_fmt() and pr_<level>()")
> > 
> > It's dumb and causes stupid "sched: " strings all over the place.
> 
> It seems rest of kernel is moving towards pr_*. [...]

Git stats do not agree with you:

 - for every pr_*() line in the kernel source code there's more than 6
   printk's:

     $ git grep '\<pr_' | wc -l
     10861
     $ git grep '\<printk' | wc -l
     61126

 - in the last stable kernel, v2.6.32, still more new printk()s were 
   introduced than pr_*() lines:

     $ git log -p v2.6.31..v2.6.32 | grep '^+.*\<pr_' | wc -l
     2016
     $ git log -p v2.6.31..v2.6.32 | grep '^+.*\<printk' | wc -l
     3531

An estimated completion of the 'conversion' to pr_*() to be in infinite number 
of years.

> [...] just removing the "sched:" would be better for consistency.
> 
> Or alternatively... deprecate pr_*?

Or alternatively, my favorite: let people who write the code use whichever 
variant they prefer.

	Ingo