[PATCH] transitional config CONFIG_GENERIC_CLOCKEVENTS

[PATCH] transitional config CONFIG_GENERIC_CLOCKEVENTS

[Christoph Egger]

Hi all!

	As part of the VAMOS[0] research project at the University of
Erlangen we are looking at multiple integrity errors in linux'
configuration system.

	This was an aid for a transition finished 2007 so not
necessary for some time now, ready for removal.

	Please keep me informed of this patch getting confirmed /
merged so we can keep track of it.

Regards

	Christoph Egger

[0] http://vamos1.informatik.uni-erlangen.de/

----
>From c8092c26246db170d509dd3424e65e87e67c5a8a Mon Sep 17 00:00:00 2001
From: Christoph Egger <siccegge@cs.fau.de>
Date: Fri, 21 May 2010 17:24:01 +0200
Subject: [PATCH 1/2] Remove transitional, dead GENERIC_CLOCKEVENTS_MIGR

GENERIC_CLOCKEVENTS_MIGR was introduced to facilate transition to the
new codebase which was finished 2007 with
b8ce33590687888ebb900d09557b8807c4539022. However the config Item and
a tiny bit of code wasn't removed back then which is adressed by this
patch.

One might consider getting rid of GENERIC_CLOCKEVENTS_BUILD
alltogether however this patch is limited to changes to affecting the
actual build process in any way.

Signed-off-by: Christoph Egger <siccegge@cs.fau.de>
---
 include/linux/clockchips.h |    4 ----
 kernel/time/Kconfig        |    2 +-
 2 files changed, 1 insertions(+), 5 deletions(-)

diff --git a/include/linux/clockchips.h b/include/linux/clockchips.h
index fc53492..3ad8331 100644
--- a/include/linux/clockchips.h
+++ b/include/linux/clockchips.h
@@ -139,11 +139,7 @@ clockevents_calc_mult_shift(struct clock_event_device *ce, u32 freq, u32 minsec)
 				      freq, minsec);
 }
 
-#ifdef CONFIG_GENERIC_CLOCKEVENTS
 extern void clockevents_notify(unsigned long reason, void *arg);
-#else
-# define clockevents_notify(reason, arg) do { } while (0)
-#endif
 
 #else /* CONFIG_GENERIC_CLOCKEVENTS_BUILD */
 
diff --git a/kernel/time/Kconfig b/kernel/time/Kconfig
index 95ed429..18ed376 100644
--- a/kernel/time/Kconfig
+++ b/kernel/time/Kconfig
@@ -25,5 +25,5 @@ config HIGH_RES_TIMERS
 config GENERIC_CLOCKEVENTS_BUILD
 	bool
 	default y
-	depends on GENERIC_CLOCKEVENTS || GENERIC_CLOCKEVENTS_MIGR
+	depends on GENERIC_CLOCKEVENTS
 
-- 
1.6.3.3

Re: [PATCH] transitional config CONFIG_GENERIC_CLOCKEVENTS

[Christoph Hellwig]

On Wed, May 26, 2010 at 03:44:15PM +0200, Christoph Egger wrote:
> @@ -25,5 +25,5 @@ config HIGH_RES_TIMERS
>  config GENERIC_CLOCKEVENTS_BUILD
>  	bool
>  	default y
> -	depends on GENERIC_CLOCKEVENTS || GENERIC_CLOCKEVENTS_MIGR
> +	depends on GENERIC_CLOCKEVENTS

At this point you can also get rid of GENERIC_CLOCKEVENTS_BUILD and
replace all instances of it with GENERIC_CLOCKEVENTS.

Re: [PATCH] transitional config CONFIG_GENERIC_CLOCKEVENTS

[Thomas Gleixner]

On Wed, 26 May 2010, Christoph Egger wrote:

> Hi all!
> 
> 	As part of the VAMOS[0] research project at the University of
> Erlangen we are looking at multiple integrity errors in linux'
> configuration system.
> 
> 	This was an aid for a transition finished 2007 so not
> necessary for some time now, ready for removal.
> 
> 	Please keep me informed of this patch getting confirmed /
> merged so we can keep track of it.

It's still necessary/helpful for any arch which is not yet
converted. And there are a few unfortunately.
 
Thanks,

	tglx

Re: [PATCH] transitional config CONFIG_GENERIC_CLOCKEVENTS

[Christoph Egger]

On Wed, May 26, 2010 at 03:56:09PM +0200, Thomas Gleixner wrote:
> On Wed, 26 May 2010, Christoph Egger wrote:
> 
> > Hi all!
> > 
> > 	As part of the VAMOS[0] research project at the University of
> > Erlangen we are looking at multiple integrity errors in linux'
> > configuration system.
> > 
> > 	This was an aid for a transition finished 2007 so not
> > necessary for some time now, ready for removal.
> > 
> > 	Please keep me informed of this patch getting confirmed /
> > merged so we can keep track of it.
> 
> It's still necessary/helpful for any arch which is not yet
> converted. And there are a few unfortunately.

	Are you referring to some out-of-tree architectures? I've
verified the string GENERIC_CLOCKEVENTS_MIGR is only appearing at this
single spot in kconfig and nowhere else in code/Kconfig in my
up-to-date linux tree.

	If not maybe I was a bit unclear about the actual rational
behind the code patch. As GENERIC_CLOCKEVENTS_MIGR is not selectable
GENERIC_CLOCKEVENTS_BUILD implies GENERIC_CLOCKEVENTS so inside a
GENERIC_CLOCKEVENTS_BUILD block a check for GENERIC_CLOCKEVENTS is
non-necessary for everything inside the linus tree and the else branch
unavailable. In this case I'll prepare an more complete patch as
suggested by CHristoph Hellwig tomorrow or friday.

Regards

	Christoph

Re: [PATCH] transitional config CONFIG_GENERIC_CLOCKEVENTS

[Thomas Gleixner]

On Wed, 26 May 2010, Christoph Egger wrote:

> On Wed, May 26, 2010 at 03:56:09PM +0200, Thomas Gleixner wrote:
> > On Wed, 26 May 2010, Christoph Egger wrote:
> > 
> > > Hi all!
> > > 
> > > 	As part of the VAMOS[0] research project at the University of
> > > Erlangen we are looking at multiple integrity errors in linux'
> > > configuration system.
> > > 
> > > 	This was an aid for a transition finished 2007 so not
> > > necessary for some time now, ready for removal.
> > > 
> > > 	Please keep me informed of this patch getting confirmed /
> > > merged so we can keep track of it.
> > 
> > It's still necessary/helpful for any arch which is not yet
> > converted. And there are a few unfortunately.
> 
> 	Are you referring to some out-of-tree architectures? I've
> verified the string GENERIC_CLOCKEVENTS_MIGR is only appearing at this
> single spot in kconfig and nowhere else in code/Kconfig in my
> up-to-date linux tree.

No. I'm referring to in tree archs which do not use the clock events
at all yet. The MIGR symbol allows them to do a step by step
conversion which builds the generic clock events stuff first before
switching over to use them. That's why it was created in the first
place and it has been proven useful when doing such work. So I'd
rather see it stay until all archs actually use the generic framework.

But yes, we can apply your patch and just let them select
GENERIC_CLOCKEVENTS_BUILD instead.

Thanks,

	tglx

Re: [PATCH] transitional config CONFIG_GENERIC_CLOCKEVENTS

[Thomas Gleixner]

On Wed, 26 May 2010, Thomas Gleixner wrote:
> > 	Are you referring to some out-of-tree architectures? I've
> > verified the string GENERIC_CLOCKEVENTS_MIGR is only appearing at this
> > single spot in kconfig and nowhere else in code/Kconfig in my
> > up-to-date linux tree.
> 
> No. I'm referring to in tree archs which do not use the clock events
> at all yet. The MIGR symbol allows them to do a step by step
> conversion which builds the generic clock events stuff first before
> switching over to use them. That's why it was created in the first
> place and it has been proven useful when doing such work. So I'd
> rather see it stay until all archs actually use the generic framework.
> 
> But yes, we can apply your patch and just let them select
> GENERIC_CLOCKEVENTS_BUILD instead.

Gah, we can't.

     tglx