[PATCH net-next] ibm emac: delete module references; phy.c only supported as built-in

[PATCH net-next] ibm emac: delete module references; phy.c only supported as built-in

[Paul Gortmaker]

The Makefile has it as "ibm_emac-y := mal.o core.o phy.o" so there is
no way this can be built modular, so remove all references to module
support.

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>

diff --git a/drivers/net/ethernet/ibm/emac/phy.c b/drivers/net/ethernet/ibm/emac/phy.c
index d3b9d10..fb96387 100644
--- a/drivers/net/ethernet/ibm/emac/phy.c
+++ b/drivers/net/ethernet/ibm/emac/phy.c
@@ -11,13 +11,15 @@
  * Copyright 2007 Benjamin Herrenschmidt, IBM Corp.
  *                <benh@kernel.crashing.org>
  *
+ * Originally listed as MODULE_LICENSE("GPL") in the code, but that
+ * was removed, since this code is only supported as non-modular.
+ *
  * Based on the arch/ppc version of the driver:
  *
  * (c) 2003, Benjamin Herrenscmidt (benh@kernel.crashing.org)
  * (c) 2004-2005, Eugene Surovegin <ebs@ebshome.net>
  *
  */
-#include <linux/module.h>
 #include <linux/kernel.h>
 #include <linux/types.h>
 #include <linux/netdevice.h>
@@ -537,5 +539,3 @@ int emac_mii_phy_probe(struct mii_phy *phy, int address)
 
 	return 0;
 }
-
-MODULE_LICENSE("GPL");
-- 
1.7.9.1

Re: [PATCH net-next] ibm emac: delete module references; phy.c only supported as built-in

[Ben Hutchings]

On Mon, 2012-02-27 at 14:47 -0500, Paul Gortmaker wrote:
> The Makefile has it as "ibm_emac-y := mal.o core.o phy.o" so there is
> no way this can be built modular, so remove all references to module
> support.

No, that's nonsense.  You need to look at whether ibm_emac.o can be
added to obj-m (which it can).

Ben.

> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
> 
> diff --git a/drivers/net/ethernet/ibm/emac/phy.c b/drivers/net/ethernet/ibm/emac/phy.c
> index d3b9d10..fb96387 100644
> --- a/drivers/net/ethernet/ibm/emac/phy.c
> +++ b/drivers/net/ethernet/ibm/emac/phy.c
> @@ -11,13 +11,15 @@
>   * Copyright 2007 Benjamin Herrenschmidt, IBM Corp.
>   *                <benh@kernel.crashing.org>
>   *
> + * Originally listed as MODULE_LICENSE("GPL") in the code, but that
> + * was removed, since this code is only supported as non-modular.
> + *
>   * Based on the arch/ppc version of the driver:
>   *
>   * (c) 2003, Benjamin Herrenscmidt (benh@kernel.crashing.org)
>   * (c) 2004-2005, Eugene Surovegin <ebs@ebshome.net>
>   *
>   */
> -#include <linux/module.h>
>  #include <linux/kernel.h>
>  #include <linux/types.h>
>  #include <linux/netdevice.h>
> @@ -537,5 +539,3 @@ int emac_mii_phy_probe(struct mii_phy *phy, int address)
>  
>  	return 0;
>  }
> -
> -MODULE_LICENSE("GPL");

-- 
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.

Re: [PATCH net-next] ibm emac: delete module references; phy.c only supported as built-in

[David Miller]

From: Paul Gortmaker <paul.gortmaker@windriver.com>
Date: Mon, 27 Feb 2012 14:47:04 -0500

> The Makefile has it as "ibm_emac-y := mal.o core.o phy.o" so there is
> no way this can be built modular, so remove all references to module
> support.

That doesn't mean it's only buildable statically.

"ibm_emacs-y :=" is merely a way to tell make what objects go into ibm_emac.{o,ko}

IBM_EMAC is tristate

Re: [PATCH net-next] ibm emac: delete module references; phy.c only supported as built-in

[Paul Gortmaker]

[Re: [PATCH net-next] ibm emac: delete module references; phy.c only supported as built-in] On 27/02/2012 (Mon 14:57) David Miller wrote:

> From: Paul Gortmaker <paul.gortmaker@windriver.com>
> Date: Mon, 27 Feb 2012 14:47:04 -0500
> 
> > The Makefile has it as "ibm_emac-y := mal.o core.o phy.o" so there is
> > no way this can be built modular, so remove all references to module
> > support.
> 
> That doesn't mean it's only buildable statically.
> 
> "ibm_emacs-y :=" is merely a way to tell make what objects go into ibm_emac.{o,ko}
> 
> IBM_EMAC is tristate

Argh, yes.  Sorry for the noise.  Since phy.c doesn't call any modular
functionality directly, the patch "worked" and I didn't catch the
error in my reasoning.

Paul.