Skip to contentSkip to conversation

ip_conntrack_helper_h323.c: EXPORT_SYMBOL'ed functions shouldn't be static

PATCH In review Adrian Bunk 3 participants 5 messages Mar 24, 2006

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ip_conntrack_helper_h323.c: EXPORT_SYMBOL'ed functions shouldn't be static

EXPORT_SYMBOL'ed functions shouldn't be static.

net/ipv4/netfilter/ip_conntrack_helper_h323.c renamed from net/ipv4/netfilter/ip_conntrack_helper_h323.c.old +2 -2
@@ -639,8 +639,8 @@
 }
 
 /****************************************************************************/
-static int get_h225_addr(unsigned char *data, TransportAddress * addr,
-			 u_int32_t * ip, u_int16_t * port)
+int get_h225_addr(unsigned char *data, TransportAddress * addr,
+		  u_int32_t * ip, u_int16_t * port)
 {
 	unsigned char *p;
 
maintainer
#1 lore raw

From: Adrian Bunk <[email protected]>
Date: Fri, 24 Mar 2006 01:08:01 +0100

Quoted 3 lines

EXPORT_SYMBOL'ed functions shouldn't be static.

Fixed in Linus's tree as of yesterday.

I actually have a patch from Patrick McHardy that will make this kind of error a build time failure instead of silently working in the modular case. I just need to test it out a bit before pushing.

#2 lore raw
Quoted 8 lines · David S. Miller

From: Adrian Bunk <[email protected]>
Date: Fri, 24 Mar 2006 01:08:01 +0100

Quoted 3 lines

EXPORT_SYMBOL'ed functions shouldn't be static.

Fixed in Linus's tree as of yesterday.

Sorry for missing this, this must have been after Andrew pulled Linus' tree for creating 2.6.16-mm1 (which is where I was looking at).

Quoted 4 lines

I actually have a patch from Patrick McHardy that will make this kind of error a build time failure instead of silently working in the modular case. I just need to test it out a bit before pushing.

Yes, it would be nice if it would also fail in CONFIG_MODULES=y builds.

cu
Adrian

Signature · 4 lines
       "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
#3 lore raw
In reply to David S. Miller
Quoted 9 lines · David S. Miller

From: Adrian Bunk <[email protected]>
Date: Fri, 24 Mar 2006 01:08:01 +0100

Quoted 3 lines

EXPORT_SYMBOL'ed functions shouldn't be static.

Quoted 6 lines

Fixed in Linus's tree as of yesterday.

I actually have a patch from Patrick McHardy that will make this kind of error a build time failure instead of silently working in the modular case. I just need to test it out a bit before pushing.

I guess I should send it to lkml anyway. It boots fine, I couldn't figure out how to compare checksums, since the time of compilation and a couple other dynamically generated strings end up in the binary.

[MODULES]: Don't allow statically declared exports

Add an extern declaration for exported symbols to make the compiler warn on symbols declared statically.

---
commit 8648236083e488ff4fc279b66d63b1187e22e558
tree cba9ee372f1056c8cf63cdc6a37a6a761fa490c9
parent 8b21e6d05d6ac0aeb44f5866ab611e2709c2f08e
author Patrick McHardy <[email protected]> Thu, 23 Mar 2006 05:07:39 +0100 committer Patrick McHardy <[email protected]> Thu, 23 Mar 2006 05:07:39 +0100

 include/linux/module.h |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)
include/linux/module.h +1 -0
index 70bd843..d956915 100644
@@ -183,6 +183,7 @@ void *__symbol_get_gpl(const char *symbo
 
 /* For every exported symbol, place a struct in the __ksymtab section */
 #define __EXPORT_SYMBOL(sym, sec)				\
+	extern typeof(sym) sym;					\
 	__CRC_SYMBOL(sym, sec)					\
 	static const char __kstrtab_##sym[]			\
 	__attribute__((section("__ksymtab_strings")))		\
maintainer
#4 lore raw

From: Patrick McHardy <[email protected]>
Date: Fri, 24 Mar 2006 02:27:44 +0100

Quoted 3 lines

I guess I should send it to lkml anyway. It boots fine, I couldn't figure out how to compare checksums, since the time of compilation and a couple other dynamically generated strings end up in the binary.

This looks fine, I'll push it in during my next round of networking updates.