[PATCH] __devexit_p macro

[PATCH] __devexit_p macro

[Felipe W Damasio]

	Hi,

	This patch defines  __devexit_p when CONFIG_HOTPLUG || MODULE, instead of
when just CONFIG_HOTPLUG is defined.

	This is needed for some net drivers (at least) that use "remove_one"
(which use unregister_netdev), allowing the driver to be re-installed.

	This is the same behaviour that 2.4.

	Patch against 2.5.30

	Please consider pulling it from:

http://cscience.org/~coqueiro/linux/patches-fwd/2.5/init.h-__devexit_p.patch

Felipe

--- ./include/linux/init.h.orig	Fri Aug  2 09:15:44 2002
+++ ./include/linux/init.h	Fri Aug  2 09:06:39 2002
@@ -177,12 +177,16 @@
 #define __devinitdata
 #define __devexit
 #define __devexitdata
-#define __devexit_p(x)  &(x)
 #else
 #define __devinit __init
 #define __devinitdata __initdata
 #define __devexit __exit
 #define __devexitdata __exitdata
+#endif
+
+#ifdef MODULE || CONFIG_HOTPLUG
+#define __devexit_p(x)  &(x)
+#else
 #define __devexit_p(x)  0
 #endif
 

Re: [PATCH] __devexit_p macro

[Dave Jones]

On Fri, Aug 02, 2002 at 09:24:56AM +0000, Felipe W Damasio wrote:
 > --- ./include/linux/init.h.orig	Fri Aug  2 09:15:44 2002
 > +++ ./include/linux/init.h	Fri Aug  2 09:06:39 2002
 > @@ -177,12 +177,16 @@
 >  #define __devinitdata
 >  #define __devexit
 >  #define __devexitdata
 > -#define __devexit_p(x)  &(x)
 >  #else
 >  #define __devinit __init
 >  #define __devinitdata __initdata
 >  #define __devexit __exit
 >  #define __devexitdata __exitdata
 > +#endif
 > +
 > +#ifdef MODULE || CONFIG_HOTPLUG
 > +#define __devexit_p(x)  &(x)
 > +#else
 >  #define __devexit_p(x)  0
 >  #endif

Instead of making this a maze of #if/else's, you can acheive
the same effect with the following patch that has been in my
tree for a few months.. (hand pasted, may not apply cleanly)



@@ -167,12 +167,18 @@ typedef void (*__cleanup_module_func_t)(
 #define device_initcall(fn)        module_init(fn)
 #define late_initcall(fn)      module_init(fn)

-#endif
+#endif /* !MODULE */

 /* Data marked not to be saved by software_suspend() */
 #define __nosavedata __attribute__ ((__section__ (".data.nosave")))

-#ifdef CONFIG_HOTPLUG
+/* Functions marked as __devexit may be discarded at kernel link time, depending
+   on config options.  Newer versions of binutils detect references from
+   retained sections to discarded sections and flag an error.  Pointers to
+   __devexit functions must use __devexit_p(function_name), the wrapper will
+   insert either the function_name or NULL, depending on the config options.
+ */
+#if defined(MODULE) || defined(CONFIG_HOTPLUG)
 #define __devinit
 #define __devinitdata
 #define __devexit

-- 
| Dave Jones.        http://www.codemonkey.org.uk
| SuSE Labs

Re: [PATCH] __devexit_p macro

[Keith Owens]

On Fri, 2 Aug 2002 14:33:49 +0200, 
Dave Jones <davej@suse.de> wrote:
>On Fri, Aug 02, 2002 at 09:24:56AM +0000, Felipe W Damasio wrote:
> > +#ifdef MODULE || CONFIG_HOTPLUG
> > +#define __devexit_p(x)  &(x)
> > +#else
> >  #define __devexit_p(x)  0
> >  #endif
>
>Instead of making this a maze of #if/else's, you can acheive
>the same effect with the following patch that has been in my
>tree for a few months.. (hand pasted, may not apply cleanly)

Better still, copy the end of 2.5.19-rc5/include/linux/init.h to
2.5.30, from #ifdef CONFIG_HOTPLUG onwards.  There is no point in
having slight differences between the 2.4 and 2.5 versions, ATM they
should be the same.

Re: [PATCH] __devexit_p macro

[Keith Owens]

On Fri, 02 Aug 2002 23:46:39 +1000, 
Keith Owens <kaos@ocs.com.au> wrote:
>On Fri, 2 Aug 2002 14:33:49 +0200, 
>Dave Jones <davej@suse.de> wrote:
>>On Fri, Aug 02, 2002 at 09:24:56AM +0000, Felipe W Damasio wrote:
>> > +#ifdef MODULE || CONFIG_HOTPLUG
>> > +#define __devexit_p(x)  &(x)
>> > +#else
>> >  #define __devexit_p(x)  0
>> >  #endif
>>
>>Instead of making this a maze of #if/else's, you can acheive
>>the same effect with the following patch that has been in my
>>tree for a few months.. (hand pasted, may not apply cleanly)
>
>Better still, copy the end of 2.5.19-rc5/include/linux/init.h to

Duh, that should be 2.4.19-rc5 of course.

>2.5.30, from #ifdef CONFIG_HOTPLUG onwards.  There is no point in
>having slight differences between the 2.4 and 2.5 versions, ATM they
>should be the same.