EXPORT_SYMBOL_NOVERS question

EXPORT_SYMBOL_NOVERS question

[vlobanov]

Hi,

I was looking over the /include/linux/module.h file, and the
EXPORT_SYMBOL_NOVERS macro caught my eye. To quote the source:

  /* We don't mangle the actual symbol anymore, so no need for
   * special casing EXPORT_SYMBOL_NOVERS.  FIXME: Deprecated */
  #define EXPORT_SYMBOL_NOVERS(sym) EXPORT_SYMBOL(sym)

A quick grep through the tree brought up no usage cases for this macro.
Is there any reason to keep it around, instead of cutting it out, as the
FIXME comment seems to suggest?

-Vadim Lobanov

Re: EXPORT_SYMBOL_NOVERS question

[Rusty Russell]

On Wed, 2004-11-03 at 00:10 -0800, vlobanov wrote:
> Hi,
> 
> I was looking over the /include/linux/module.h file, and the
> EXPORT_SYMBOL_NOVERS macro caught my eye. To quote the source:
> 
>   /* We don't mangle the actual symbol anymore, so no need for
>    * special casing EXPORT_SYMBOL_NOVERS.  FIXME: Deprecated */
>   #define EXPORT_SYMBOL_NOVERS(sym) EXPORT_SYMBOL(sym)
> 
> A quick grep through the tree brought up no usage cases for this macro.
> Is there any reason to keep it around, instead of cutting it out, as the
> FIXME comment seems to suggest?

Yep, it's time.

Rusty.

Name: Remove EXPORT_SYMBOL_NOVERS
Status: Trivial
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Vadim Lobanov points out that EXPORT_SYMBOL_NOVERS is no longer used;
in fact, SH still uses it, but once we fix that, the kernel is clean.
Remove it.

Index: linux-2.6.10-rc2-bk11-Module/arch/sh/kernel/sh_ksyms.c
===================================================================
--- linux-2.6.10-rc2-bk11-Module.orig/arch/sh/kernel/sh_ksyms.c	2004-11-16 15:29:12.000000000 +1100
+++ linux-2.6.10-rc2-bk11-Module/arch/sh/kernel/sh_ksyms.c	2004-11-29 09:38:48.007889896 +1100
@@ -89,7 +89,7 @@
 
 EXPORT_SYMBOL(__div64_32);
 
-#define DECLARE_EXPORT(name) extern void name(void);EXPORT_SYMBOL_NOVERS(name)
+#define DECLARE_EXPORT(name) extern void name(void);EXPORT_SYMBOL(name)
 
 /* These symbols are generated by the compiler itself */
 DECLARE_EXPORT(__udivsi3);
@@ -100,7 +100,7 @@
 DECLARE_EXPORT(__lshrdi3);
 DECLARE_EXPORT(__movstr);
 
-EXPORT_SYMBOL_NOVERS(strcpy);
+EXPORT_SYMBOL(strcpy);
 
 #ifdef CONFIG_CPU_SH4
 DECLARE_EXPORT(__movstr_i4_even);
Index: linux-2.6.10-rc2-bk11-Module/Makefile
===================================================================
--- linux-2.6.10-rc2-bk11-Module.orig/Makefile	2004-11-29 07:48:40.000000000 +1100
+++ linux-2.6.10-rc2-bk11-Module/Makefile	2004-11-29 09:39:06.146132464 +1100
@@ -1166,7 +1166,7 @@
 quiet_cmd_tags = MAKE   $@
 define cmd_tags
 	rm -f $@; \
-	CTAGSF=`ctags --version | grep -i exuberant >/dev/null && echo "-I __initdata,__exitdata,EXPORT_SYMBOL,EXPORT_SYMBOL_NOVERS"`; \
+	CTAGSF=`ctags --version | grep -i exuberant >/dev/null && echo "-I __initdata,__exitdata,EXPORT_SYMBOL,EXPORT_SYMBOL_GPL"`; \
 	$(all-sources) | xargs ctags $$CTAGSF -a
 endef
 
Index: linux-2.6.10-rc2-bk11-Module/include/linux/module.h
===================================================================
--- linux-2.6.10-rc2-bk11-Module.orig/include/linux/module.h	2004-11-29 07:48:57.000000000 +1100
+++ linux-2.6.10-rc2-bk11-Module/include/linux/module.h	2004-11-29 09:39:31.970206608 +1100
@@ -206,10 +206,6 @@
 
 #endif
 
-/* We don't mangle the actual symbol anymore, so no need for
- * special casing EXPORT_SYMBOL_NOVERS.  FIXME: Deprecated */
-#define EXPORT_SYMBOL_NOVERS(sym) EXPORT_SYMBOL(sym)
-
 struct module_ref
 {
 	local_t count;
@@ -449,7 +445,6 @@
 #else /* !CONFIG_MODULES... */
 #define EXPORT_SYMBOL(sym)
 #define EXPORT_SYMBOL_GPL(sym)
-#define EXPORT_SYMBOL_NOVERS(sym)
 
 /* Given an address, look for it in the exception tables. */
 static inline const struct exception_table_entry *

-- 
A bad analogy is like a leaky screwdriver -- Richard Braakman

Re: EXPORT_SYMBOL_NOVERS question

[Adrian Bunk]

On Mon, Nov 29, 2004 at 09:42:20AM +1100, Rusty Russell wrote:
> On Wed, 2004-11-03 at 00:10 -0800, vlobanov wrote:
> > Hi,
> > 
> > I was looking over the /include/linux/module.h file, and the
> > EXPORT_SYMBOL_NOVERS macro caught my eye. To quote the source:
> > 
> >   /* We don't mangle the actual symbol anymore, so no need for
> >    * special casing EXPORT_SYMBOL_NOVERS.  FIXME: Deprecated */
> >   #define EXPORT_SYMBOL_NOVERS(sym) EXPORT_SYMBOL(sym)
> > 
> > A quick grep through the tree brought up no usage cases for this macro.
> > Is there any reason to keep it around, instead of cutting it out, as the
> > FIXME comment seems to suggest?
> 
> Yep, it's time.
> 
> Rusty.
> 
> Name: Remove EXPORT_SYMBOL_NOVERS
> Status: Trivial
> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
> 
> Vadim Lobanov points out that EXPORT_SYMBOL_NOVERS is no longer used;
> in fact, SH still uses it, but once we fix that, the kernel is clean.
> Remove it.
>...

Thaat's true for Linus' tree, but not for -mm.

@David:
arch/frv/kernel/frv_ksyms.c has to be changed.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: EXPORT_SYMBOL_NOVERS question

[David Howells]

> arch/frv/kernel/frv_ksyms.c has to be changed.

Okay.

David

Re: EXPORT_SYMBOL_NOVERS question

[Paul Mundt]

[-- Attachment #1: Type: text/plain, Size: 248 bytes --]

On Mon, Nov 29, 2004 at 09:42:20AM +1100, Rusty Russell wrote:
> Vadim Lobanov points out that EXPORT_SYMBOL_NOVERS is no longer used;
> in fact, SH still uses it, but once we fix that, the kernel is clean.
> Remove it.
> 
Looks good to me..

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]