mmotm on powerpc: walk_memory_resource?

mmotm on powerpc: walk_memory_resource?

[Hugh Dickins]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 797 bytes --]

Recent mmotm builds on powerpc have been failing with

arch/powerpc/mm/mem.c:169: error: ‘walk_memory_resource’ undeclared here (not in a function)
cc1: warnings being treated as errors
arch/powerpc/mm/mem.c:169: error: type defaults to ‘int’ in declaration of ‘walk_memory_resource’
make[1]: *** [arch/powerpc/mm/mem.o] Error 1

because walk-system-ram-range.patch renames walk_memory_resource()
but leaves the EXPORT_SYMBOL_GPL(walk_memory_resource).  I couldn't
tell from the patch description whether it's a matter of changing
the EXPORT_SYMBOL_GPL or removing the patch to arch/powerpc/mm/mem.c
or something else.  (I change the EXPORT to get my build working,
but there appears to be something in drivers/infiniband which still
wants a walk_memory_resource.)

Hugh

Re: mmotm on powerpc: walk_memory_resource?

[KAMEZAWA Hiroyuki]

On Tue, 15 Sep 2009 22:36:26 +0100 (BST)
Hugh Dickins <hugh.dickins@tiscali.co.uk> wrote:

> Recent mmotm builds on powerpc have been failing with
> 
> arch/powerpc/mm/mem.c:169: error: ‘walk_memory_resource’ undeclared here (not in a function)
> cc1: warnings being treated as errors
> arch/powerpc/mm/mem.c:169: error: type defaults to ‘int’ in declaration of ‘walk_memory_resource’
> make[1]: *** [arch/powerpc/mm/mem.o] Error 1
> 
> because walk-system-ram-range.patch renames walk_memory_resource()
> but leaves the EXPORT_SYMBOL_GPL(walk_memory_resource).  I couldn't
> tell from the patch description whether it's a matter of changing
> the EXPORT_SYMBOL_GPL or removing the patch to arch/powerpc/mm/mem.c
> or something else.  (I change the EXPORT to get my build working,
> but there appears to be something in drivers/infiniband which still
> wants a walk_memory_resource.)
> 

Very sorry..my mistake. Will fix it immedieately.

-Kame

> Hugh

[BUGFIX][mmotm] fix walk system ram range patch (Was Re: mmotm on powerpc: walk_memory_resource?

[KAMEZAWA Hiroyuki]

On Tue, 15 Sep 2009 22:36:26 +0100 (BST)
Hugh Dickins <hugh.dickins@tiscali.co.uk> wrote:

> Recent mmotm builds on powerpc have been failing with
> 
> arch/powerpc/mm/mem.c:169: error: ‘walk_memory_resource’ undeclared here (not in a function)
> cc1: warnings being treated as errors
> arch/powerpc/mm/mem.c:169: error: type defaults to ‘int’ in declaration of ‘walk_memory_resource’
> make[1]: *** [arch/powerpc/mm/mem.o] Error 1
> 
Thank you for pointing out.
-Kame
==


walk-system-ram-range.patch renames walk_memory_resource()
to be walk_system_ram_range().
But powerpc has its own one and it's exported.
(Because ehea driver (IBM's one) uses this, powerpc export this.)

> arch/powerpc/mm/mem.c:169: error: ‘walk_memory_resource’ undeclared here (not in a function)
> cc1: warnings being treated as errors
> arch/powerpc/mm/mem.c:169: error: type defaults to ‘int’ in declaration of ‘walk_memory_resource’
> make[1]: *** [arch/powerpc/mm/mem.o] Error 1
> 
The patch failed to update powerpc's EXPORT_SYMBOL, this patch fixes it.

Reported-by: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>

---
 arch/powerpc/mm/mem.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Index: mmotm-2.6.31-Sep14/arch/powerpc/mm/mem.c
===================================================================
--- mmotm-2.6.31-Sep14.orig/arch/powerpc/mm/mem.c
+++ mmotm-2.6.31-Sep14/arch/powerpc/mm/mem.c
@@ -166,7 +166,7 @@ walk_system_ram_range(unsigned long star
 	}
 	return ret;
 }
-EXPORT_SYMBOL_GPL(walk_memory_resource);
+EXPORT_SYMBOL_GPL(walk_system_ram_range);
 
 /*
  * Initialize the bootmem system and give it all the memory we

Re: [BUGFIX][mmotm] fix walk system ram range patch (Was Re: mmotm on powerpc: walk_memory_resource?

[David Rientjes]

On Wed, 16 Sep 2009, KAMEZAWA Hiroyuki wrote:

> Index: mmotm-2.6.31-Sep14/arch/powerpc/mm/mem.c
> ===================================================================
> --- mmotm-2.6.31-Sep14.orig/arch/powerpc/mm/mem.c
> +++ mmotm-2.6.31-Sep14/arch/powerpc/mm/mem.c
> @@ -166,7 +166,7 @@ walk_system_ram_range(unsigned long star
>  	}
>  	return ret;
>  }
> -EXPORT_SYMBOL_GPL(walk_memory_resource);
> +EXPORT_SYMBOL_GPL(walk_system_ram_range);
>  
>  /*
>   * Initialize the bootmem system and give it all the memory we
> 

Um, doesn't drivers/infiniband/hw/ehca/ehca_mrmw.c need to be fixed too?

[BUGFIX] fix walk systm ram range for infiniband (Was Re: [BUGFIX][mmotm] fix walk system ram range patch (Was Re: mmotm on powerpc: walk_memory_resource?

[KAMEZAWA Hiroyuki]

On Tue, 15 Sep 2009 23:16:47 -0700 (PDT)
David Rientjes <rientjes@google.com> wrote:

> On Wed, 16 Sep 2009, KAMEZAWA Hiroyuki wrote:
> 
> > Index: mmotm-2.6.31-Sep14/arch/powerpc/mm/mem.c
> > ===================================================================
> > --- mmotm-2.6.31-Sep14.orig/arch/powerpc/mm/mem.c
> > +++ mmotm-2.6.31-Sep14/arch/powerpc/mm/mem.c
> > @@ -166,7 +166,7 @@ walk_system_ram_range(unsigned long star
> >  	}
> >  	return ret;
> >  }
> > -EXPORT_SYMBOL_GPL(walk_memory_resource);
> > +EXPORT_SYMBOL_GPL(walk_system_ram_range);
> >  
> >  /*
> >   * Initialize the bootmem system and give it all the memory we
> > 
> 
> Um, doesn't drivers/infiniband/hw/ehca/ehca_mrmw.c need to be fixed too?
> 
Hmm?..yes. nice catch...
my GTAGS doesn't catch it...maybe need to be updated..
I'm sorry that I didn't imagine this is widely used because this was under
CONFIG_MEMORY_HOTPLUG....

How about this ?
==
walk-system-ram-range.patch failes to catch recent change in
inifiniband. This is a fix.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>

---
 drivers/infiniband/hw/ehca/ehca_mrmw.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Index: mmotm-2.6.31-Sep14/drivers/infiniband/hw/ehca/ehca_mrmw.c
===================================================================
--- mmotm-2.6.31-Sep14.orig/drivers/infiniband/hw/ehca/ehca_mrmw.c
+++ mmotm-2.6.31-Sep14/drivers/infiniband/hw/ehca/ehca_mrmw.c
@@ -2463,7 +2463,7 @@ int ehca_create_busmap(void)
 	int ret;
 
 	ehca_mr_len = 0;
-	ret = walk_memory_resource(0, 1ULL << MAX_PHYSMEM_BITS, NULL,
+	ret = walk_system_ram_range(0, 1ULL << MAX_PHYSMEM_BITS, NULL,
 				   ehca_create_busmap_callback);
 	return ret;
 }

Re: [BUGFIX][mmotm] fix walk system ram range patch (Was Re: mmotm on powerpc: walk_memory_resource?

[Américo Wang]

On Wed, Sep 16, 2009 at 2:16 PM, David Rientjes <rientjes@google.com> wrote:
> On Wed, 16 Sep 2009, KAMEZAWA Hiroyuki wrote:
>
>> Index: mmotm-2.6.31-Sep14/arch/powerpc/mm/mem.c
>> ===================================================================
>> --- mmotm-2.6.31-Sep14.orig/arch/powerpc/mm/mem.c
>> +++ mmotm-2.6.31-Sep14/arch/powerpc/mm/mem.c
>> @@ -166,7 +166,7 @@ walk_system_ram_range(unsigned long star
>>       }
>>       return ret;
>>  }
>> -EXPORT_SYMBOL_GPL(walk_memory_resource);
>> +EXPORT_SYMBOL_GPL(walk_system_ram_range);
>>
>>  /*
>>   * Initialize the bootmem system and give it all the memory we
>>
>
> Um, doesn't drivers/infiniband/hw/ehca/ehca_mrmw.c need to be fixed too?

Yes, all callers of walk_memory_resource() should be fixed.

Re: [BUGFIX][mmotm] fix walk system ram range patch (Was Re: mmotm on powerpc: walk_memory_resource?

[Américo Wang]

2009/9/16 KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>:
> On Tue, 15 Sep 2009 22:36:26 +0100 (BST)
> Hugh Dickins <hugh.dickins@tiscali.co.uk> wrote:
>
>> Recent mmotm builds on powerpc have been failing with
>>
>> arch/powerpc/mm/mem.c:169: error: ‘walk_memory_resource’ undeclared here (not in a function)
>> cc1: warnings being treated as errors
>> arch/powerpc/mm/mem.c:169: error: type defaults to ‘int’ in declaration of ‘walk_memory_resource’
>> make[1]: *** [arch/powerpc/mm/mem.o] Error 1
>>
> Thank you for pointing out.
> -Kame
> ==
>
>
> walk-system-ram-range.patch renames walk_memory_resource()
> to be walk_system_ram_range().
> But powerpc has its own one and it's exported.
> (Because ehea driver (IBM's one) uses this, powerpc export this.)
>
>> arch/powerpc/mm/mem.c:169: error: ‘walk_memory_resource’ undeclared here (not in a function)
>> cc1: warnings being treated as errors
>> arch/powerpc/mm/mem.c:169: error: type defaults to ‘int’ in declaration of ‘walk_memory_resource’
>> make[1]: *** [arch/powerpc/mm/mem.o] Error 1
>>
> The patch failed to update powerpc's EXPORT_SYMBOL, this patch fixes it.
>
> Reported-by: Hugh Dickins <hugh.dickins@tiscali.co.uk>
> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>


Acked-by: WANG Cong <xiyou.wangcong@gmail.com>

>
> ---
>  arch/powerpc/mm/mem.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> Index: mmotm-2.6.31-Sep14/arch/powerpc/mm/mem.c
> ===================================================================
> --- mmotm-2.6.31-Sep14.orig/arch/powerpc/mm/mem.c
> +++ mmotm-2.6.31-Sep14/arch/powerpc/mm/mem.c
> @@ -166,7 +166,7 @@ walk_system_ram_range(unsigned long star
>        }
>        return ret;
>  }
> -EXPORT_SYMBOL_GPL(walk_memory_resource);
> +EXPORT_SYMBOL_GPL(walk_system_ram_range);
>
>  /*
>  * Initialize the bootmem system and give it all the memory we
>