Re: [RFC] Circular include dependencies

Re: [RFC] Circular include dependencies

[Andrew Morton]

On Fri, 23 May 2008 14:20:34 +0100 Russell King <rmk+lkml@arm.linux.org.uk> wrote:

> Hi,
> 
> Having discovered some circular include dependencies in the ARM header
> files which were causing build issues, I created a script to walk ARM
> includes and report any similar issues found - which includes traversing
> any referenced linux/ includes.
> 
> It identified the following two in include/linux/:
> 
>   linux/mmzone.h <- linux/memory_hotplug.h <- linux/mmzone.h
>   linux/mmzone.h <- linux/topology.h <- linux/mmzone.h
> 
> Checking them by hand reveals that these are real.  Whether they're
> capable of causing a problem or not, I'm not going to comment on.
> However, they're not a good idea and someone should probably look at
> resolving the loops.

(cc's added).

Thanks.

I'm not sure who we could tap for the topology.h one.

A suitable (and often good) way of solving this is to identify the
things which a.h needs from b.h and hoist them out into a new c.h and
include that from both a.h and b.h.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [RFC] Circular include dependencies

[Yasunori Goto]

> On Fri, 23 May 2008 14:20:34 +0100 Russell King <rmk+lkml@arm.linux.org.uk> wrote:
> 
> > Hi,
> > 
> > Having discovered some circular include dependencies in the ARM header
> > files which were causing build issues, I created a script to walk ARM
> > includes and report any similar issues found - which includes traversing
> > any referenced linux/ includes.
> > 
> > It identified the following two in include/linux/:
> > 
> >   linux/mmzone.h <- linux/memory_hotplug.h <- linux/mmzone.h
> >   linux/mmzone.h <- linux/topology.h <- linux/mmzone.h
> > 
> > Checking them by hand reveals that these are real.  Whether they're
> > capable of causing a problem or not, I'm not going to comment on.
> > However, they're not a good idea and someone should probably look at
> > resolving the loops.
> 
> (cc's added).
> 
> Thanks.
> 
> I'm not sure who we could tap for the topology.h one.
> 
> A suitable (and often good) way of solving this is to identify the
> things which a.h needs from b.h and hoist them out into a new c.h and
> include that from both a.h and b.h.

Kame-san and I reviewed memory_hotplug.h.
We found its including was not necessary certainly.

This is the patch to fix it. I tested on IA64, and checked cross-compile
on powerpc. Kame-san tested this on x86-64.

Thanks for your report.

Bye.

----
Fix no need including of mmzone.h in memory_hotplug.h

Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>

---
 include/linux/memory_hotplug.h |    1 -
 1 file changed, 1 deletion(-)

Index: dptest/include/linux/memory_hotplug.h
===================================================================
--- dptest.orig/include/linux/memory_hotplug.h	2008-05-21 10:56:00.000000000 +0900
+++ dptest/include/linux/memory_hotplug.h	2008-05-26 20:32:06.000000000 +0900
@@ -1,7 +1,6 @@
 #ifndef __LINUX_MEMORY_HOTPLUG_H
 #define __LINUX_MEMORY_HOTPLUG_H
 
-#include <linux/mmzone.h>
 #include <linux/spinlock.h>
 #include <linux/notifier.h>
 
-- 
Yasunori Goto 


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>