linux-next.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* linux-next: manual merge of the signal tree with the modules tree
@ 2013-03-14  6:27 Stephen Rothwell
  2013-03-14 11:24 ` James Hogan
  2013-03-15  4:41 ` Rusty Russell
  0 siblings, 2 replies; 6+ messages in thread
From: Stephen Rothwell @ 2013-03-14  6:27 UTC (permalink / raw)
  To: Al Viro; +Cc: linux-next, linux-kernel, Rusty Russell

[-- Attachment #1: Type: text/plain, Size: 2278 bytes --]

Hi Al,

Today's linux-next merge of the signal tree got a conflict in
include/asm-generic/unistd.h between commit 837718bfd28b
("CONFIG_SYMBOL_PREFIX: cleanup") from the modules tree and commit
e1b5bb6d1236 ("consolidate cond_syscall and SYSCALL_ALIAS declarations")
from the signal tree.

The latter moved the cond_syscall stuff to linkage.h, so I applied the
following patch as a merge fixup and can carry the fix as necessary (no
action is required).  I am not sure if this is completely correct or all
that is needed.

From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Thu, 14 Mar 2013 17:14:41 +1100
Subject: [PATCH] cond_syscall and SYSCALL_ALIAS merge fixup

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
 include/linux/linkage.h | 20 ++++++--------------
 1 file changed, 6 insertions(+), 14 deletions(-)

diff --git a/include/linux/linkage.h b/include/linux/linkage.h
index 829d66c..bedcddf 100644
--- a/include/linux/linkage.h
+++ b/include/linux/linkage.h
@@ -2,7 +2,7 @@
 #define _LINUX_LINKAGE_H
 
 #include <linux/compiler.h>
-#include <linux/stringify.h>
+#include <linux/export.h>
 #include <asm/linkage.h>
 
 #ifdef __cplusplus
@@ -15,24 +15,16 @@
 #define asmlinkage CPP_ASMLINKAGE
 #endif
 
-#ifndef SYMBOL_NAME
-#ifdef CONFIG_SYMBOL_PREFIX
-#define SYMBOL_NAME(x) CONFIG_SYMBOL_PREFIX ## x
-#else
-#define SYMBOL_NAME(x) x
-#endif
-#endif
-#define __SYMBOL_NAME(x) __stringify(SYMBOL_NAME(x))
-
 #ifndef cond_syscall
-#define cond_syscall(x) asm(".weak\t" __SYMBOL_NAME(x) \
-	"\n\t.set\t" __SYMBOL_NAME(x) "," __SYMBOL_NAME(sys_ni_syscall));
+#define cond_syscall(x) asm(".weak\t" VMLINUX_SYMBOL_STR(x) "\n\t"	\
+			    ".set\t" VMLINUX_SYMBOL_STR(x) ","	\
+			    VMLINUX_SYMBOL_STR(sys_ni_syscall))
 #endif
 
 #ifndef SYSCALL_ALIAS
 #define SYSCALL_ALIAS(alias, name)				\
-	asm ("\t.globl " __SYMBOL_NAME(alias)			\
-	"\n\t.set\t" __SYMBOL_NAME(alias) "," __SYMBOL_NAME(name))
+	asm ("\t.globl " VMLINUX_SYMBOL_STR(alias)			\
+	"\n\t.set\t" VMLINUX_SYMBOL_STR(alias) "," VMLINUX_SYMBOL_STR(name))
 #endif
 
 #define __page_aligned_data	__section(.data..page_aligned) __aligned(PAGE_SIZE)
-- 
1.8.1

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

^ permalink raw reply related	[flat|nested] 6+ messages in thread
* Re: linux-next: manual merge of the signal tree with the modules tree
  2013-03-14  6:27 linux-next: manual merge of the signal tree with the modules tree Stephen Rothwell
@ 2013-03-14 11:24 ` James Hogan
  2013-03-15  4:41 ` Rusty Russell
  1 sibling, 0 replies; 6+ messages in thread
From: James Hogan @ 2013-03-14 11:24 UTC (permalink / raw)
  To: Stephen Rothwell, Al Viro; +Cc: linux-next, linux-kernel, Rusty Russell

[-- Attachment #1: Type: text/plain, Size: 1010 bytes --]

On 14/03/13 06:27, Stephen Rothwell wrote:
> Hi Al,
> 
> Today's linux-next merge of the signal tree got a conflict in
> include/asm-generic/unistd.h between commit 837718bfd28b
> ("CONFIG_SYMBOL_PREFIX: cleanup") from the modules tree and commit
> e1b5bb6d1236 ("consolidate cond_syscall and SYSCALL_ALIAS declarations")
> from the signal tree.
> 
> The latter moved the cond_syscall stuff to linkage.h, so I applied the
> following patch as a merge fixup and can carry the fix as necessary (no
> action is required).  I am not sure if this is completely correct or all
> that is needed.

FYI, I've tested that symbol prefix stuff works on linux-next for metag
(after fixes applied from Rusty's latest symbol prefix patch).

Al: your signal tree still breaks symbol prefixed arches due to stuff
like this in kernel/sys_ni.i:

asm(".weak\t" "CONFIG_SYMBOL_PREFIXsys_quotactl" "\n\t.set\t"
"CONFIG_SYMBOL_PREFIXsys_quotactl" ","
"CONFIG_SYMBOL_PREFIXsys_ni_syscall");;

Cheers
James


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

^ permalink raw reply	[flat|nested] 6+ messages in thread
* Re: linux-next: manual merge of the signal tree with the modules tree
  2013-03-14  6:27 linux-next: manual merge of the signal tree with the modules tree Stephen Rothwell
  2013-03-14 11:24 ` James Hogan
@ 2013-03-15  4:41 ` Rusty Russell
  2013-03-15 10:21   ` Sedat Dilek
  1 sibling, 1 reply; 6+ messages in thread
From: Rusty Russell @ 2013-03-15  4:41 UTC (permalink / raw)
  To: Stephen Rothwell, Al Viro; +Cc: linux-next, linux-kernel

Stephen Rothwell <sfr@canb.auug.org.au> writes:
> Hi Al,
>
> Today's linux-next merge of the signal tree got a conflict in
> include/asm-generic/unistd.h between commit 837718bfd28b
> ("CONFIG_SYMBOL_PREFIX: cleanup") from the modules tree and commit
> e1b5bb6d1236 ("consolidate cond_syscall and SYSCALL_ALIAS declarations")
> from the signal tree.
>
> The latter moved the cond_syscall stuff to linkage.h, so I applied the
> following patch as a merge fixup and can carry the fix as necessary (no
> action is required).  I am not sure if this is completely correct or all
> that is needed.

Your fix looks correct, thanks.

I've been forced to update that patch after another round of
improvements, so you may need to re-do the merge.

Cheers,
Rusty.

^ permalink raw reply	[flat|nested] 6+ messages in thread
* Re: linux-next: manual merge of the signal tree with the modules tree
  2013-03-15  4:41 ` Rusty Russell
@ 2013-03-15 10:21   ` Sedat Dilek
  2013-03-15 11:28     ` Sedat Dilek
  2013-03-18  0:56     ` Rusty Russell
  0 siblings, 2 replies; 6+ messages in thread
From: Sedat Dilek @ 2013-03-15 10:21 UTC (permalink / raw)
  To: Rusty Russell; +Cc: Stephen Rothwell, Al Viro, linux-next, linux-kernel

[-- Attachment #1: Type: text/plain, Size: 2502 bytes --]

On Fri, Mar 15, 2013 at 5:41 AM, Rusty Russell <rusty@rustcorp.com.au> wrote:
> Stephen Rothwell <sfr@canb.auug.org.au> writes:
>> Hi Al,
>>
>> Today's linux-next merge of the signal tree got a conflict in
>> include/asm-generic/unistd.h between commit 837718bfd28b
>> ("CONFIG_SYMBOL_PREFIX: cleanup") from the modules tree and commit
>> e1b5bb6d1236 ("consolidate cond_syscall and SYSCALL_ALIAS declarations")
>> from the signal tree.
>>
>> The latter moved the cond_syscall stuff to linkage.h, so I applied the
>> following patch as a merge fixup and can carry the fix as necessary (no
>> action is required).  I am not sure if this is completely correct or all
>> that is needed.
>
> Your fix looks correct, thanks.
>
> I've been forced to update that patch after another round of
> improvements, so you may need to re-do the merge.
>

Hi,

I just looked into modules-next...
The improved version is in [1]...
...and contains a file called "kernel/modsign_certificate.S" which is
NOT in the latest Linux-Next tree [2].
So, I thought about reverting the one in -next and apply the new one
from modules-next.
This is not possible!

$ git log --oneline -2
d98f244 Merge tag 'next-20130315' of
git://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next into
Linux-Next-v20130315
1765898 Add linux-next specific files for 20130315

$ LC_ALL=C ls -l kernel/mod*
-rw-r--r-- 1 wearefam wearefam   458 Mar 15 11:13 kernel/module-internal.h
-rw-r--r-- 1 wearefam wearefam 99000 Mar 15 11:13 kernel/module.c
-rw-r--r-- 1 wearefam wearefam  6063 Mar 15 11:13 kernel/module_signing.c

Is this a generated file?
I have attached both patches extracted from Linux-Next [3] and modules-next [4].
Both patches list changes to "kernel/modsign_certificate.S".
/me ---> confused!

Can you please enlighten me?


Regards,
- Sedat -

[1] http://git.kernel.org/cgit/linux/kernel/git/rusty/linux.git/commit/?h=modules-next&id=b92021b09df70c1609e3547f3d6128dd560be97f
[2] http://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/tree/kernel?id=next-20130315
[3] http://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/patch/?id=837718bfd28ba5f474c1182ef2639b6949d854fc
[4] http://git.kernel.org/cgit/linux/kernel/git/rusty/linux.git/patch/?id=b92021b09df70c1609e3547f3d6128dd560be97f

> Cheers,
> Rusty.
> --
> To unsubscribe from this list: send the line "unsubscribe linux-next" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

[-- Attachment #2: CONFIG_SYMBOL_PREFIX-cleanup-from-modules-next.patch --]
[-- Type: application/octet-stream, Size: 14634 bytes --]

From b92021b09df70c1609e3547f3d6128dd560be97f Mon Sep 17 00:00:00 2001
From: Rusty Russell <rusty@rustcorp.com.au>
Date: Fri, 15 Mar 2013 04:34:17 +0000
Subject: CONFIG_SYMBOL_PREFIX: cleanup.

We have CONFIG_SYMBOL_PREFIX, which three archs define to the string
"_".  But Al Viro broke this in "consolidate cond_syscall and
SYSCALL_ALIAS declarations" (in linux-next), and he's not the first to
do so.

Using CONFIG_SYMBOL_PREFIX is awkward, since we usually just want to
prefix it so something.  So various places define helpers which are
defined to nothing if CONFIG_SYMBOL_PREFIX isn't set:

1) include/asm-generic/unistd.h defines __SYMBOL_PREFIX.
2) include/asm-generic/vmlinux.lds.h defines VMLINUX_SYMBOL(sym)
3) include/linux/export.h defines MODULE_SYMBOL_PREFIX.
4) include/linux/kernel.h defines SYMBOL_PREFIX (which differs from #7)
5) kernel/modsign_certificate.S defines ASM_SYMBOL(sym)
6) scripts/modpost.c defines MODULE_SYMBOL_PREFIX
7) scripts/Makefile.lib defines SYMBOL_PREFIX on the commandline if
   CONFIG_SYMBOL_PREFIX is set, so that we have a non-string version
   for pasting.

(arch/h8300/include/asm/linkage.h defines SYMBOL_NAME(), too).

Let's solve this properly:
1) No more generic prefix, just CONFIG_HAVE_UNDERSCORE_SYMBOL_PREFIX.
2) Make linux/export.h usable from asm.
3) Define VMLINUX_SYMBOL() and VMLINUX_SYMBOL_STR().
4) Make everyone use them.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Reviewed-by: James Hogan <james.hogan@imgtec.com>
Tested-by: James Hogan <james.hogan@imgtec.com> (metag)
---
diff --git a/Makefile b/Makefile
index a05ea42..0b09ba5 100644
--- a/Makefile
+++ b/Makefile
@@ -1398,7 +1398,7 @@ quiet_cmd_rmfiles = $(if $(wildcard $(rm-files)),CLEAN   $(wildcard $(rm-files))
 # Run depmod only if we have System.map and depmod is executable
 quiet_cmd_depmod = DEPMOD  $(KERNELRELEASE)
       cmd_depmod = $(CONFIG_SHELL) $(srctree)/scripts/depmod.sh $(DEPMOD) \
-                   $(KERNELRELEASE) "$(patsubst "%",%,$(CONFIG_SYMBOL_PREFIX))"
+                   $(KERNELRELEASE) "$(patsubst y,_,$(CONFIG_HAVE_UNDERSCORE_SYMBOL_PREFIX))"
 
 # Create temporary dir for module support files
 # clean it up only when building all modules
diff --git a/arch/Kconfig b/arch/Kconfig
index 1455579..7b433a4 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -384,6 +384,12 @@ config MODULES_USE_ELF_REL
 	  Modules only use ELF REL relocations.  Modules with ELF RELA
 	  relocations will give an error.
 
+config HAVE_UNDERSCORE_SYMBOL_PREFIX
+	bool
+	help
+	  Some architectures generate an _ in front of C symbols; things like
+	  module loading and assembly files need to know about this.
+
 #
 # ABI hall of shame
 #
diff --git a/arch/blackfin/Kconfig b/arch/blackfin/Kconfig
index c3f2e0b..453ebe4 100644
--- a/arch/blackfin/Kconfig
+++ b/arch/blackfin/Kconfig
@@ -1,7 +1,3 @@
-config SYMBOL_PREFIX
-	string
-	default "_"
-
 config MMU
 	def_bool n
 
@@ -33,6 +29,7 @@ config BLACKFIN
 	select ARCH_HAVE_CUSTOM_GPIO_H
 	select ARCH_WANT_OPTIONAL_GPIOLIB
 	select HAVE_UID16
+	select HAVE_UNDERSCORE_SYMBOL_PREFIX
 	select VIRT_TO_BUS
 	select ARCH_WANT_IPC_PARSE_VERSION
 	select HAVE_GENERIC_HARDIRQS
diff --git a/arch/h8300/Kconfig b/arch/h8300/Kconfig
index 79250de..303e4f9 100644
--- a/arch/h8300/Kconfig
+++ b/arch/h8300/Kconfig
@@ -12,10 +12,7 @@ config H8300
 	select MODULES_USE_ELF_RELA
 	select OLD_SIGSUSPEND3
 	select OLD_SIGACTION
-
-config SYMBOL_PREFIX
-	string
-	default "_"
+	select HAVE_UNDERSCORE_SYMBOL_PREFIX
 
 config MMU
 	bool
diff --git a/arch/metag/Kconfig b/arch/metag/Kconfig
index afc8973..2099617 100644
--- a/arch/metag/Kconfig
+++ b/arch/metag/Kconfig
@@ -1,7 +1,3 @@
-config SYMBOL_PREFIX
-	string
-	default "_"
-
 config METAG
 	def_bool y
 	select EMBEDDED
@@ -27,6 +23,7 @@ config METAG
 	select HAVE_MOD_ARCH_SPECIFIC
 	select HAVE_PERF_EVENTS
 	select HAVE_SYSCALL_TRACEPOINTS
+	select HAVE_UNDERSCORE_SYMBOL_PREFIX
 	select IRQ_DOMAIN
 	select MODULES_USE_ELF_RELA
 	select OF
diff --git a/drivers/mtd/chips/gen_probe.c b/drivers/mtd/chips/gen_probe.c
index 3b9a284..74dbb6b 100644
--- a/drivers/mtd/chips/gen_probe.c
+++ b/drivers/mtd/chips/gen_probe.c
@@ -204,14 +204,16 @@ static inline struct mtd_info *cfi_cmdset_unknown(struct map_info *map,
 	struct cfi_private *cfi = map->fldrv_priv;
 	__u16 type = primary?cfi->cfiq->P_ID:cfi->cfiq->A_ID;
 #ifdef CONFIG_MODULES
-	char probename[16+sizeof(MODULE_SYMBOL_PREFIX)];
+	char probename[sizeof(VMLINUX_SYMBOL_STR(cfi_cmdset_%4.4X))];
 	cfi_cmdset_fn_t *probe_function;
 
-	sprintf(probename, MODULE_SYMBOL_PREFIX "cfi_cmdset_%4.4X", type);
+	sprintf(probename, VMLINUX_SYMBOL_STR(cfi_cmdset_%4.4X), type);
 
 	probe_function = __symbol_get(probename);
 	if (!probe_function) {
-		request_module(probename + sizeof(MODULE_SYMBOL_PREFIX) - 1);
+		char modname[sizeof("cfi_cmdset_%4.4X")];
+		sprintf(modname, "cfi_cmdset_%4.4X", type);
+		request_module(modname);
 		probe_function = __symbol_get(probename);
 	}
 
diff --git a/include/asm-generic/unistd.h b/include/asm-generic/unistd.h
index 4077b5d..15c0598 100644
--- a/include/asm-generic/unistd.h
+++ b/include/asm-generic/unistd.h
@@ -1,4 +1,5 @@
 #include <uapi/asm-generic/unistd.h>
+#include <linux/export.h>
 
 /*
  * These are required system calls, we should
@@ -17,12 +18,7 @@
  * but it doesn't work on all toolchains, so we just do it by hand
  */
 #ifndef cond_syscall
-#ifdef CONFIG_SYMBOL_PREFIX
-#define __SYMBOL_PREFIX CONFIG_SYMBOL_PREFIX
-#else
-#define __SYMBOL_PREFIX
-#endif
-#define cond_syscall(x) asm(".weak\t" __SYMBOL_PREFIX #x "\n\t" \
-			    ".set\t" __SYMBOL_PREFIX #x "," \
-			    __SYMBOL_PREFIX "sys_ni_syscall")
+#define cond_syscall(x) asm(".weak\t" VMLINUX_SYMBOL_STR(x) "\n\t"	\
+			    ".set\t" VMLINUX_SYMBOL_STR(x) ","	\
+			    VMLINUX_SYMBOL_STR(sys_ni_syscall))
 #endif
diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
index afa12c7..eb58d2d 100644
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@ -52,13 +52,7 @@
 #define LOAD_OFFSET 0
 #endif
 
-#ifndef SYMBOL_PREFIX
-#define VMLINUX_SYMBOL(sym) sym
-#else
-#define PASTE2(x,y) x##y
-#define PASTE(x,y) PASTE2(x,y)
-#define VMLINUX_SYMBOL(sym) PASTE(SYMBOL_PREFIX, sym)
-#endif
+#include <linux/export.h>
 
 /* Align . to a 8 byte boundary equals to maximum function alignment. */
 #define ALIGN_FUNCTION()  . = ALIGN(8)
diff --git a/include/linux/export.h b/include/linux/export.h
index 696c0f4..412cd50 100644
--- a/include/linux/export.h
+++ b/include/linux/export.h
@@ -5,17 +5,24 @@
  * to reduce the amount of pointless cruft we feed to gcc when only
  * exporting a simple symbol or two.
  *
- * If you feel the need to add #include <linux/foo.h> to this file
- * then you are doing something wrong and should go away silently.
+ * Try not to add #includes here.  It slows compilation and makes kernel
+ * hackers place grumpy comments in header files.
  */
 
 /* Some toolchains use a `_' prefix for all user symbols. */
-#ifdef CONFIG_SYMBOL_PREFIX
-#define MODULE_SYMBOL_PREFIX CONFIG_SYMBOL_PREFIX
+#ifdef CONFIG_HAVE_UNDERSCORE_SYMBOL_PREFIX
+#define __VMLINUX_SYMBOL(x) _##x
+#define __VMLINUX_SYMBOL_STR(x) "_" #x
 #else
-#define MODULE_SYMBOL_PREFIX ""
+#define __VMLINUX_SYMBOL(x) x
+#define __VMLINUX_SYMBOL_STR(x) #x
 #endif
 
+/* Indirect, so macros are expanded before pasting. */
+#define VMLINUX_SYMBOL(x) __VMLINUX_SYMBOL(x)
+#define VMLINUX_SYMBOL_STR(x) __VMLINUX_SYMBOL_STR(x)
+
+#ifndef __ASSEMBLY__
 struct kernel_symbol
 {
 	unsigned long value;
@@ -51,7 +58,7 @@ extern struct module __this_module;
 	__CRC_SYMBOL(sym, sec)					\
 	static const char __kstrtab_##sym[]			\
 	__attribute__((section("__ksymtab_strings"), aligned(1))) \
-	= MODULE_SYMBOL_PREFIX #sym;				\
+	= VMLINUX_SYMBOL_STR(sym);				\
 	static const struct kernel_symbol __ksymtab_##sym	\
 	__used							\
 	__attribute__((section("___ksymtab" sec "+" #sym), unused))	\
@@ -85,5 +92,6 @@ extern struct module __this_module;
 #define EXPORT_UNUSED_SYMBOL_GPL(sym)
 
 #endif /* CONFIG_MODULES */
+#endif /* !__ASSEMBLY__ */
 
 #endif /* _LINUX_EXPORT_H */
diff --git a/include/linux/kernel.h b/include/linux/kernel.h
index 80d3687..e13e992 100644
--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -723,13 +723,6 @@ static inline void ftrace_dump(enum ftrace_dump_mode oops_dump_mode) { }
 /* Trap pasters of __FUNCTION__ at compile-time */
 #define __FUNCTION__ (__func__)
 
-/* This helps us to avoid #ifdef CONFIG_SYMBOL_PREFIX */
-#ifdef CONFIG_SYMBOL_PREFIX
-#define SYMBOL_PREFIX CONFIG_SYMBOL_PREFIX
-#else
-#define SYMBOL_PREFIX ""
-#endif
-
 /* Rebuild everything on CONFIG_FTRACE_MCOUNT_RECORD */
 #ifdef CONFIG_FTRACE_MCOUNT_RECORD
 # define REBUILD_DUE_TO_FTRACE_MCOUNT_RECORD
diff --git a/include/linux/module.h b/include/linux/module.h
index ead1b57..46f1ea0 100644
--- a/include/linux/module.h
+++ b/include/linux/module.h
@@ -190,7 +190,7 @@ extern int modules_disabled; /* for sysctl */
 /* Get/put a kernel symbol (calls must be symmetric) */
 void *__symbol_get(const char *symbol);
 void *__symbol_get_gpl(const char *symbol);
-#define symbol_get(x) ((typeof(&x))(__symbol_get(MODULE_SYMBOL_PREFIX #x)))
+#define symbol_get(x) ((typeof(&x))(__symbol_get(VMLINUX_SYMBOL_STR(x))))
 
 /* modules using other modules: kdb wants to see this. */
 struct module_use {
@@ -453,7 +453,7 @@ extern void __module_put_and_exit(struct module *mod, long code)
 #ifdef CONFIG_MODULE_UNLOAD
 unsigned long module_refcount(struct module *mod);
 void __symbol_put(const char *symbol);
-#define symbol_put(x) __symbol_put(MODULE_SYMBOL_PREFIX #x)
+#define symbol_put(x) __symbol_put(VMLINUX_SYMBOL_STR(x))
 void symbol_put_addr(void *addr);
 
 /* Sometimes we know we already have a refcount, and it's easier not
diff --git a/kernel/modsign_certificate.S b/kernel/modsign_certificate.S
index 246b4c6..4a9a86d 100644
--- a/kernel/modsign_certificate.S
+++ b/kernel/modsign_certificate.S
@@ -1,15 +1,8 @@
-/* SYMBOL_PREFIX defined on commandline from CONFIG_SYMBOL_PREFIX */
-#ifndef SYMBOL_PREFIX
-#define ASM_SYMBOL(sym) sym
-#else
-#define PASTE2(x,y) x##y
-#define PASTE(x,y) PASTE2(x,y)
-#define ASM_SYMBOL(sym) PASTE(SYMBOL_PREFIX, sym)
-#endif
+#include <linux/export.h>
 
 #define GLOBAL(name)	\
-	.globl ASM_SYMBOL(name);	\
-	ASM_SYMBOL(name):
+	.globl VMLINUX_SYMBOL(name);	\
+	VMLINUX_SYMBOL(name):
 
 	.section ".init.data","aw"
 
diff --git a/kernel/module.c b/kernel/module.c
index 0925c9a..cfd4a3f 100644
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -1209,7 +1209,7 @@ static inline int check_modstruct_version(Elf_Shdr *sechdrs,
 
 	/* Since this should be found in kernel (which can't be removed),
 	 * no locking is necessary. */
-	if (!find_symbol(MODULE_SYMBOL_PREFIX "module_layout", NULL,
+	if (!find_symbol(VMLINUX_SYMBOL_STR(module_layout), NULL,
 			 &crc, true, false))
 		BUG();
 	return check_version(sechdrs, versindex, "module_layout", mod, crc,
diff --git a/scripts/Makefile.lib b/scripts/Makefile.lib
index 07125e6..a373a1f 100644
--- a/scripts/Makefile.lib
+++ b/scripts/Makefile.lib
@@ -119,13 +119,6 @@ _c_flags += $(if $(patsubst n%,, \
 		$(CFLAGS_GCOV))
 endif
 
-ifdef CONFIG_SYMBOL_PREFIX
-_sym_flags = -DSYMBOL_PREFIX=$(patsubst "%",%,$(CONFIG_SYMBOL_PREFIX))
-_cpp_flags += $(_sym_flags)
-_a_flags += $(_sym_flags)
-endif
-
-
 # If building the kernel in a separate objtree expand all occurrences
 # of -Idir to -I$(srctree)/dir except for absolute paths (starting with '/').
 
diff --git a/scripts/link-vmlinux.sh b/scripts/link-vmlinux.sh
index 3d569d6..0149949 100644
--- a/scripts/link-vmlinux.sh
+++ b/scripts/link-vmlinux.sh
@@ -74,9 +74,8 @@ kallsyms()
 	info KSYM ${2}
 	local kallsymopt;
 
-	if [ -n "${CONFIG_SYMBOL_PREFIX}" ]; then
-		kallsymopt="${kallsymopt} \
-			    --symbol-prefix=${CONFIG_SYMBOL_PREFIX}"
+	if [ -n "${CONFIG_HAVE_UNDERSCORE_SYMBOL_PREFIX}" ]; then
+		kallsymopt="${kallsymopt} --symbol-prefix=_"
 	fi
 
 	if [ -n "${CONFIG_KALLSYMS_ALL}" ]; then
diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c
index 78b30c1..282decf 100644
--- a/scripts/mod/modpost.c
+++ b/scripts/mod/modpost.c
@@ -18,14 +18,7 @@
 #include "modpost.h"
 #include "../../include/generated/autoconf.h"
 #include "../../include/linux/license.h"
-
-/* Some toolchains use a `_' prefix for all user symbols. */
-#ifdef CONFIG_SYMBOL_PREFIX
-#define MODULE_SYMBOL_PREFIX CONFIG_SYMBOL_PREFIX
-#else
-#define MODULE_SYMBOL_PREFIX ""
-#endif
-
+#include "../../include/linux/export.h"
 
 /* Are we using CONFIG_MODVERSIONS? */
 int modversions = 0;
@@ -562,7 +555,7 @@ static void parse_elf_finish(struct elf_info *info)
 static int ignore_undef_symbol(struct elf_info *info, const char *symname)
 {
 	/* ignore __this_module, it will be resolved shortly */
-	if (strcmp(symname, MODULE_SYMBOL_PREFIX "__this_module") == 0)
+	if (strcmp(symname, VMLINUX_SYMBOL_STR(__this_module)) == 0)
 		return 1;
 	/* ignore global offset table */
 	if (strcmp(symname, "_GLOBAL_OFFSET_TABLE_") == 0)
@@ -583,8 +576,8 @@ static int ignore_undef_symbol(struct elf_info *info, const char *symname)
 	return 0;
 }
 
-#define CRC_PFX     MODULE_SYMBOL_PREFIX "__crc_"
-#define KSYMTAB_PFX MODULE_SYMBOL_PREFIX "__ksymtab_"
+#define CRC_PFX     VMLINUX_SYMBOL_STR(__crc_)
+#define KSYMTAB_PFX VMLINUX_SYMBOL_STR(__ksymtab_)
 
 static void handle_modversions(struct module *mod, struct elf_info *info,
 			       Elf_Sym *sym, const char *symname)
@@ -637,14 +630,15 @@ static void handle_modversions(struct module *mod, struct elf_info *info,
 		}
 #endif
 
-		if (memcmp(symname, MODULE_SYMBOL_PREFIX,
-			   strlen(MODULE_SYMBOL_PREFIX)) == 0) {
-			mod->unres =
-			  alloc_symbol(symname +
-			               strlen(MODULE_SYMBOL_PREFIX),
-			               ELF_ST_BIND(sym->st_info) == STB_WEAK,
-			               mod->unres);
-		}
+#ifdef CONFIG_HAVE_UNDERSCORE_SYMBOL_PREFIX
+		if (symname[0] != '_')
+			break;
+		else
+			symname++;
+#endif
+		mod->unres = alloc_symbol(symname,
+					  ELF_ST_BIND(sym->st_info) == STB_WEAK,
+					  mod->unres);
 		break;
 	default:
 		/* All exported symbols */
@@ -652,9 +646,9 @@ static void handle_modversions(struct module *mod, struct elf_info *info,
 			sym_add_exported(symname + strlen(KSYMTAB_PFX), mod,
 					export);
 		}
-		if (strcmp(symname, MODULE_SYMBOL_PREFIX "init_module") == 0)
+		if (strcmp(symname, VMLINUX_SYMBOL_STR(init_module)) == 0)
 			mod->has_init = 1;
-		if (strcmp(symname, MODULE_SYMBOL_PREFIX "cleanup_module") == 0)
+		if (strcmp(symname, VMLINUX_SYMBOL_STR(cleanup_module)) == 0)
 			mod->has_cleanup = 1;
 		break;
 	}
--
cgit v0.9.1

[-- Attachment #3: CONFIG_SYMBOL_PREFIX-cleanup-from-next-20130315.patch --]
[-- Type: application/octet-stream, Size: 14461 bytes --]

From 837718bfd28ba5f474c1182ef2639b6949d854fc Mon Sep 17 00:00:00 2001
From: Rusty Russell <rusty@rustcorp.com.au>
Date: Wed, 13 Mar 2013 02:03:45 +0000
Subject: CONFIG_SYMBOL_PREFIX: cleanup.

We have CONFIG_SYMBOL_PREFIX, which three archs define to the string
"_".  But Al Viro broke this in "consolidate cond_syscall and
SYSCALL_ALIAS declarations" (in linux-next), and he's not the first to
do so.

Using CONFIG_SYMBOL_PREFIX is awkward, since we usually just want to
prefix it so something.  So various places define helpers which are
defined to nothing if CONFIG_SYMBOL_PREFIX isn't set:

1) include/asm-generic/unistd.h defines __SYMBOL_PREFIX.
2) include/asm-generic/vmlinux.lds.h defines VMLINUX_SYMBOL(sym)
3) include/linux/export.h defines MODULE_SYMBOL_PREFIX.
4) include/linux/kernel.h defines SYMBOL_PREFIX (which differs from #7)
5) kernel/modsign_certificate.S defines ASM_SYMBOL(sym)
6) scripts/modpost.c defines MODULE_SYMBOL_PREFIX
7) scripts/Makefile.lib defines SYMBOL_PREFIX on the commandline if
   CONFIG_SYMBOL_PREFIX is set, so that we have a non-string version
   for pasting.

(arch/h8300/include/asm/linkage.h defines SYMBOL_NAME(), too).

Let's solve this properly:
1) No more generic prefix, just CONFIG_HAVE_UNDERSCORE_SYMBOL_PREFIX.
2) Make linux/export.h usable from asm.
3) Define VMLINUX_SYMBOL, VMLINUX_SYMBOL_STR and VMLINUX_SYMBOL_PREFIX_STR.
4) Make everyone use them.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Reviewed-by: James Hogan <james.hogan@imgtec.com>
Tested-by: James Hogan <james.hogan@imgtec.com> (on metag)
---
diff --git a/Makefile b/Makefile
index a05ea42..10fb6c7 100644
--- a/Makefile
+++ b/Makefile
@@ -1398,7 +1398,7 @@ quiet_cmd_rmfiles = $(if $(wildcard $(rm-files)),CLEAN   $(wildcard $(rm-files))
 # Run depmod only if we have System.map and depmod is executable
 quiet_cmd_depmod = DEPMOD  $(KERNELRELEASE)
       cmd_depmod = $(CONFIG_SHELL) $(srctree)/scripts/depmod.sh $(DEPMOD) \
-                   $(KERNELRELEASE) "$(patsubst "%",%,$(CONFIG_SYMBOL_PREFIX))"
+                   $(KERNELRELEASE) "$(patsubst y,_,$(CONFIG_HAVE_SYMBOL_PREFIX_UNDERSCORE))"
 
 # Create temporary dir for module support files
 # clean it up only when building all modules
diff --git a/arch/Kconfig b/arch/Kconfig
index 1455579..7b433a4 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -384,6 +384,12 @@ config MODULES_USE_ELF_REL
 	  Modules only use ELF REL relocations.  Modules with ELF RELA
 	  relocations will give an error.
 
+config HAVE_UNDERSCORE_SYMBOL_PREFIX
+	bool
+	help
+	  Some architectures generate an _ in front of C symbols; things like
+	  module loading and assembly files need to know about this.
+
 #
 # ABI hall of shame
 #
diff --git a/arch/blackfin/Kconfig b/arch/blackfin/Kconfig
index c3f2e0b..453ebe4 100644
--- a/arch/blackfin/Kconfig
+++ b/arch/blackfin/Kconfig
@@ -1,7 +1,3 @@
-config SYMBOL_PREFIX
-	string
-	default "_"
-
 config MMU
 	def_bool n
 
@@ -33,6 +29,7 @@ config BLACKFIN
 	select ARCH_HAVE_CUSTOM_GPIO_H
 	select ARCH_WANT_OPTIONAL_GPIOLIB
 	select HAVE_UID16
+	select HAVE_UNDERSCORE_SYMBOL_PREFIX
 	select VIRT_TO_BUS
 	select ARCH_WANT_IPC_PARSE_VERSION
 	select HAVE_GENERIC_HARDIRQS
diff --git a/arch/h8300/Kconfig b/arch/h8300/Kconfig
index 79250de..303e4f9 100644
--- a/arch/h8300/Kconfig
+++ b/arch/h8300/Kconfig
@@ -12,10 +12,7 @@ config H8300
 	select MODULES_USE_ELF_RELA
 	select OLD_SIGSUSPEND3
 	select OLD_SIGACTION
-
-config SYMBOL_PREFIX
-	string
-	default "_"
+	select HAVE_UNDERSCORE_SYMBOL_PREFIX
 
 config MMU
 	bool
diff --git a/arch/metag/Kconfig b/arch/metag/Kconfig
index afc8973..2099617 100644
--- a/arch/metag/Kconfig
+++ b/arch/metag/Kconfig
@@ -1,7 +1,3 @@
-config SYMBOL_PREFIX
-	string
-	default "_"
-
 config METAG
 	def_bool y
 	select EMBEDDED
@@ -27,6 +23,7 @@ config METAG
 	select HAVE_MOD_ARCH_SPECIFIC
 	select HAVE_PERF_EVENTS
 	select HAVE_SYSCALL_TRACEPOINTS
+	select HAVE_UNDERSCORE_SYMBOL_PREFIX
 	select IRQ_DOMAIN
 	select MODULES_USE_ELF_RELA
 	select OF
diff --git a/drivers/mtd/chips/gen_probe.c b/drivers/mtd/chips/gen_probe.c
index 3b9a284..00d20ad 100644
--- a/drivers/mtd/chips/gen_probe.c
+++ b/drivers/mtd/chips/gen_probe.c
@@ -204,14 +204,14 @@ static inline struct mtd_info *cfi_cmdset_unknown(struct map_info *map,
 	struct cfi_private *cfi = map->fldrv_priv;
 	__u16 type = primary?cfi->cfiq->P_ID:cfi->cfiq->A_ID;
 #ifdef CONFIG_MODULES
-	char probename[16+sizeof(MODULE_SYMBOL_PREFIX)];
+	char probename[16+sizeof(VMLINUX_SYMBOL_PREFIX_STR)];
 	cfi_cmdset_fn_t *probe_function;
 
-	sprintf(probename, MODULE_SYMBOL_PREFIX "cfi_cmdset_%4.4X", type);
+	sprintf(probename, VMLINUX_SYMBOL_PREFIX_STR "cfi_cmdset_%4.4X", type);
 
 	probe_function = __symbol_get(probename);
 	if (!probe_function) {
-		request_module(probename + sizeof(MODULE_SYMBOL_PREFIX) - 1);
+		request_module(probename + sizeof(VMLINUX_SYMBOL_PREFIX_STR)-1);
 		probe_function = __symbol_get(probename);
 	}
 
diff --git a/include/asm-generic/unistd.h b/include/asm-generic/unistd.h
index 4077b5d..15c0598 100644
--- a/include/asm-generic/unistd.h
+++ b/include/asm-generic/unistd.h
@@ -1,4 +1,5 @@
 #include <uapi/asm-generic/unistd.h>
+#include <linux/export.h>
 
 /*
  * These are required system calls, we should
@@ -17,12 +18,7 @@
  * but it doesn't work on all toolchains, so we just do it by hand
  */
 #ifndef cond_syscall
-#ifdef CONFIG_SYMBOL_PREFIX
-#define __SYMBOL_PREFIX CONFIG_SYMBOL_PREFIX
-#else
-#define __SYMBOL_PREFIX
-#endif
-#define cond_syscall(x) asm(".weak\t" __SYMBOL_PREFIX #x "\n\t" \
-			    ".set\t" __SYMBOL_PREFIX #x "," \
-			    __SYMBOL_PREFIX "sys_ni_syscall")
+#define cond_syscall(x) asm(".weak\t" VMLINUX_SYMBOL_STR(x) "\n\t"	\
+			    ".set\t" VMLINUX_SYMBOL_STR(x) ","	\
+			    VMLINUX_SYMBOL_STR(sys_ni_syscall))
 #endif
diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
index afa12c7..eb58d2d 100644
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@ -52,13 +52,7 @@
 #define LOAD_OFFSET 0
 #endif
 
-#ifndef SYMBOL_PREFIX
-#define VMLINUX_SYMBOL(sym) sym
-#else
-#define PASTE2(x,y) x##y
-#define PASTE(x,y) PASTE2(x,y)
-#define VMLINUX_SYMBOL(sym) PASTE(SYMBOL_PREFIX, sym)
-#endif
+#include <linux/export.h>
 
 /* Align . to a 8 byte boundary equals to maximum function alignment. */
 #define ALIGN_FUNCTION()  . = ALIGN(8)
diff --git a/include/linux/export.h b/include/linux/export.h
index 696c0f4..fc83b2a 100644
--- a/include/linux/export.h
+++ b/include/linux/export.h
@@ -7,15 +7,27 @@
  *
  * If you feel the need to add #include <linux/foo.h> to this file
  * then you are doing something wrong and should go away silently.
+ *
+ * If you think the above arrogance just encourages more people to add
+ * random crap to this file, you're not alone.
  */
 
 /* Some toolchains use a `_' prefix for all user symbols. */
-#ifdef CONFIG_SYMBOL_PREFIX
-#define MODULE_SYMBOL_PREFIX CONFIG_SYMBOL_PREFIX
+#ifdef CONFIG_HAVE_UNDERSCORE_SYMBOL_PREFIX
+#define __VMLINUX_SYMBOL(x) _##x
+#define __VMLINUX_SYMBOL_STR(x) "_" #x
+#define VMLINUX_SYMBOL_PREFIX_STR "_"
 #else
-#define MODULE_SYMBOL_PREFIX ""
+#define __VMLINUX_SYMBOL(x) x
+#define __VMLINUX_SYMBOL_STR(x) #x
+#define VMLINUX_SYMBOL_PREFIX_STR ""
 #endif
 
+/* Indirect, so macros are expanded before pasting. */
+#define VMLINUX_SYMBOL(x) __VMLINUX_SYMBOL(x)
+#define VMLINUX_SYMBOL_STR(x) __VMLINUX_SYMBOL_STR(x)
+
+#ifndef __ASSEMBLY__
 struct kernel_symbol
 {
 	unsigned long value;
@@ -51,7 +63,7 @@ extern struct module __this_module;
 	__CRC_SYMBOL(sym, sec)					\
 	static const char __kstrtab_##sym[]			\
 	__attribute__((section("__ksymtab_strings"), aligned(1))) \
-	= MODULE_SYMBOL_PREFIX #sym;				\
+	= VMLINUX_SYMBOL_STR(sym);				\
 	static const struct kernel_symbol __ksymtab_##sym	\
 	__used							\
 	__attribute__((section("___ksymtab" sec "+" #sym), unused))	\
@@ -85,5 +97,6 @@ extern struct module __this_module;
 #define EXPORT_UNUSED_SYMBOL_GPL(sym)
 
 #endif /* CONFIG_MODULES */
+#endif /* !__ASSEMBLY__ */
 
 #endif /* _LINUX_EXPORT_H */
diff --git a/include/linux/kernel.h b/include/linux/kernel.h
index 80d3687..e13e992 100644
--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -723,13 +723,6 @@ static inline void ftrace_dump(enum ftrace_dump_mode oops_dump_mode) { }
 /* Trap pasters of __FUNCTION__ at compile-time */
 #define __FUNCTION__ (__func__)
 
-/* This helps us to avoid #ifdef CONFIG_SYMBOL_PREFIX */
-#ifdef CONFIG_SYMBOL_PREFIX
-#define SYMBOL_PREFIX CONFIG_SYMBOL_PREFIX
-#else
-#define SYMBOL_PREFIX ""
-#endif
-
 /* Rebuild everything on CONFIG_FTRACE_MCOUNT_RECORD */
 #ifdef CONFIG_FTRACE_MCOUNT_RECORD
 # define REBUILD_DUE_TO_FTRACE_MCOUNT_RECORD
diff --git a/include/linux/module.h b/include/linux/module.h
index ead1b57..46f1ea0 100644
--- a/include/linux/module.h
+++ b/include/linux/module.h
@@ -190,7 +190,7 @@ extern int modules_disabled; /* for sysctl */
 /* Get/put a kernel symbol (calls must be symmetric) */
 void *__symbol_get(const char *symbol);
 void *__symbol_get_gpl(const char *symbol);
-#define symbol_get(x) ((typeof(&x))(__symbol_get(MODULE_SYMBOL_PREFIX #x)))
+#define symbol_get(x) ((typeof(&x))(__symbol_get(VMLINUX_SYMBOL_STR(x))))
 
 /* modules using other modules: kdb wants to see this. */
 struct module_use {
@@ -453,7 +453,7 @@ extern void __module_put_and_exit(struct module *mod, long code)
 #ifdef CONFIG_MODULE_UNLOAD
 unsigned long module_refcount(struct module *mod);
 void __symbol_put(const char *symbol);
-#define symbol_put(x) __symbol_put(MODULE_SYMBOL_PREFIX #x)
+#define symbol_put(x) __symbol_put(VMLINUX_SYMBOL_STR(x))
 void symbol_put_addr(void *addr);
 
 /* Sometimes we know we already have a refcount, and it's easier not
diff --git a/kernel/modsign_certificate.S b/kernel/modsign_certificate.S
index 246b4c6..4a9a86d 100644
--- a/kernel/modsign_certificate.S
+++ b/kernel/modsign_certificate.S
@@ -1,15 +1,8 @@
-/* SYMBOL_PREFIX defined on commandline from CONFIG_SYMBOL_PREFIX */
-#ifndef SYMBOL_PREFIX
-#define ASM_SYMBOL(sym) sym
-#else
-#define PASTE2(x,y) x##y
-#define PASTE(x,y) PASTE2(x,y)
-#define ASM_SYMBOL(sym) PASTE(SYMBOL_PREFIX, sym)
-#endif
+#include <linux/export.h>
 
 #define GLOBAL(name)	\
-	.globl ASM_SYMBOL(name);	\
-	ASM_SYMBOL(name):
+	.globl VMLINUX_SYMBOL(name);	\
+	VMLINUX_SYMBOL(name):
 
 	.section ".init.data","aw"
 
diff --git a/kernel/module.c b/kernel/module.c
index 0925c9a..cfd4a3f 100644
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -1209,7 +1209,7 @@ static inline int check_modstruct_version(Elf_Shdr *sechdrs,
 
 	/* Since this should be found in kernel (which can't be removed),
 	 * no locking is necessary. */
-	if (!find_symbol(MODULE_SYMBOL_PREFIX "module_layout", NULL,
+	if (!find_symbol(VMLINUX_SYMBOL_STR(module_layout), NULL,
 			 &crc, true, false))
 		BUG();
 	return check_version(sechdrs, versindex, "module_layout", mod, crc,
diff --git a/scripts/Makefile.lib b/scripts/Makefile.lib
index 07125e6..a373a1f 100644
--- a/scripts/Makefile.lib
+++ b/scripts/Makefile.lib
@@ -119,13 +119,6 @@ _c_flags += $(if $(patsubst n%,, \
 		$(CFLAGS_GCOV))
 endif
 
-ifdef CONFIG_SYMBOL_PREFIX
-_sym_flags = -DSYMBOL_PREFIX=$(patsubst "%",%,$(CONFIG_SYMBOL_PREFIX))
-_cpp_flags += $(_sym_flags)
-_a_flags += $(_sym_flags)
-endif
-
-
 # If building the kernel in a separate objtree expand all occurrences
 # of -Idir to -I$(srctree)/dir except for absolute paths (starting with '/').
 
diff --git a/scripts/link-vmlinux.sh b/scripts/link-vmlinux.sh
index 3d569d6..0149949 100644
--- a/scripts/link-vmlinux.sh
+++ b/scripts/link-vmlinux.sh
@@ -74,9 +74,8 @@ kallsyms()
 	info KSYM ${2}
 	local kallsymopt;
 
-	if [ -n "${CONFIG_SYMBOL_PREFIX}" ]; then
-		kallsymopt="${kallsymopt} \
-			    --symbol-prefix=${CONFIG_SYMBOL_PREFIX}"
+	if [ -n "${CONFIG_HAVE_UNDERSCORE_SYMBOL_PREFIX}" ]; then
+		kallsymopt="${kallsymopt} --symbol-prefix=_"
 	fi
 
 	if [ -n "${CONFIG_KALLSYMS_ALL}" ]; then
diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c
index 78b30c1..6985021 100644
--- a/scripts/mod/modpost.c
+++ b/scripts/mod/modpost.c
@@ -18,14 +18,7 @@
 #include "modpost.h"
 #include "../../include/generated/autoconf.h"
 #include "../../include/linux/license.h"
-
-/* Some toolchains use a `_' prefix for all user symbols. */
-#ifdef CONFIG_SYMBOL_PREFIX
-#define MODULE_SYMBOL_PREFIX CONFIG_SYMBOL_PREFIX
-#else
-#define MODULE_SYMBOL_PREFIX ""
-#endif
-
+#include "../../include/linux/export.h"
 
 /* Are we using CONFIG_MODVERSIONS? */
 int modversions = 0;
@@ -562,7 +555,7 @@ static void parse_elf_finish(struct elf_info *info)
 static int ignore_undef_symbol(struct elf_info *info, const char *symname)
 {
 	/* ignore __this_module, it will be resolved shortly */
-	if (strcmp(symname, MODULE_SYMBOL_PREFIX "__this_module") == 0)
+	if (strcmp(symname, VMLINUX_SYMBOL_STR(__this_module)) == 0)
 		return 1;
 	/* ignore global offset table */
 	if (strcmp(symname, "_GLOBAL_OFFSET_TABLE_") == 0)
@@ -583,8 +576,8 @@ static int ignore_undef_symbol(struct elf_info *info, const char *symname)
 	return 0;
 }
 
-#define CRC_PFX     MODULE_SYMBOL_PREFIX "__crc_"
-#define KSYMTAB_PFX MODULE_SYMBOL_PREFIX "__ksymtab_"
+#define CRC_PFX     VMLINUX_SYMBOL_STR(__crc_)
+#define KSYMTAB_PFX VMLINUX_SYMBOL_STR(__ksymtab_)
 
 static void handle_modversions(struct module *mod, struct elf_info *info,
 			       Elf_Sym *sym, const char *symname)
@@ -637,11 +630,11 @@ static void handle_modversions(struct module *mod, struct elf_info *info,
 		}
 #endif
 
-		if (memcmp(symname, MODULE_SYMBOL_PREFIX,
-			   strlen(MODULE_SYMBOL_PREFIX)) == 0) {
+		if (memcmp(symname, VMLINUX_SYMBOL_PREFIX_STR,
+			   strlen(VMLINUX_SYMBOL_PREFIX_STR)) == 0) {
 			mod->unres =
 			  alloc_symbol(symname +
-			               strlen(MODULE_SYMBOL_PREFIX),
+			               strlen(VMLINUX_SYMBOL_PREFIX_STR),
 			               ELF_ST_BIND(sym->st_info) == STB_WEAK,
 			               mod->unres);
 		}
@@ -652,9 +645,9 @@ static void handle_modversions(struct module *mod, struct elf_info *info,
 			sym_add_exported(symname + strlen(KSYMTAB_PFX), mod,
 					export);
 		}
-		if (strcmp(symname, MODULE_SYMBOL_PREFIX "init_module") == 0)
+		if (strcmp(symname, VMLINUX_SYMBOL_STR(init_module)) == 0)
 			mod->has_init = 1;
-		if (strcmp(symname, MODULE_SYMBOL_PREFIX "cleanup_module") == 0)
+		if (strcmp(symname, VMLINUX_SYMBOL_STR(cleanup_module)) == 0)
 			mod->has_cleanup = 1;
 		break;
 	}
--
cgit v0.9.1

^ permalink raw reply related	[flat|nested] 6+ messages in thread
* Re: linux-next: manual merge of the signal tree with the modules tree
  2013-03-15 10:21   ` Sedat Dilek
@ 2013-03-15 11:28     ` Sedat Dilek
  2013-03-18  0:56     ` Rusty Russell
  1 sibling, 0 replies; 6+ messages in thread
From: Sedat Dilek @ 2013-03-15 11:28 UTC (permalink / raw)
  To: Rusty Russell
  Cc: Stephen Rothwell, Al Viro, linux-next, linux-kernel, James Hogan

[-- Attachment #1: Type: text/plain, Size: 2926 bytes --]

On Fri, Mar 15, 2013 at 11:21 AM, Sedat Dilek <sedat.dilek@gmail.com> wrote:
> On Fri, Mar 15, 2013 at 5:41 AM, Rusty Russell <rusty@rustcorp.com.au> wrote:
>> Stephen Rothwell <sfr@canb.auug.org.au> writes:
>>> Hi Al,
>>>
>>> Today's linux-next merge of the signal tree got a conflict in
>>> include/asm-generic/unistd.h between commit 837718bfd28b
>>> ("CONFIG_SYMBOL_PREFIX: cleanup") from the modules tree and commit
>>> e1b5bb6d1236 ("consolidate cond_syscall and SYSCALL_ALIAS declarations")
>>> from the signal tree.
>>>
>>> The latter moved the cond_syscall stuff to linkage.h, so I applied the
>>> following patch as a merge fixup and can carry the fix as necessary (no
>>> action is required).  I am not sure if this is completely correct or all
>>> that is needed.
>>
>> Your fix looks correct, thanks.
>>
>> I've been forced to update that patch after another round of
>> improvements, so you may need to re-do the merge.
>>
>
> Hi,
>
> I just looked into modules-next...
> The improved version is in [1]...
> ...and contains a file called "kernel/modsign_certificate.S" which is
> NOT in the latest Linux-Next tree [2].
> So, I thought about reverting the one in -next and apply the new one
> from modules-next.
> This is not possible!
>
> $ git log --oneline -2
> d98f244 Merge tag 'next-20130315' of
> git://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next into
> Linux-Next-v20130315
> 1765898 Add linux-next specific files for 20130315
>
> $ LC_ALL=C ls -l kernel/mod*
> -rw-r--r-- 1 wearefam wearefam   458 Mar 15 11:13 kernel/module-internal.h
> -rw-r--r-- 1 wearefam wearefam 99000 Mar 15 11:13 kernel/module.c
> -rw-r--r-- 1 wearefam wearefam  6063 Mar 15 11:13 kernel/module_signing.c
>

OK, renamed: kernel/modsign_certificate.S -> kernel/system_certificates.S... see

commit ebe2e946f60e0012c02a27845bdab70e34cc4202
KEYS: Separate the kernel signature checking keyring from module signing

> Is this a generated file?
> I have attached both patches extracted from Linux-Next [3] and modules-next [4].
> Both patches list changes to "kernel/modsign_certificate.S".
> /me ---> confused!
>
> Can you please enlighten me?
>

Is this attached follow-up correct?

- Sedat -

>
> Regards,
> - Sedat -
>
> [1] http://git.kernel.org/cgit/linux/kernel/git/rusty/linux.git/commit/?h=modules-next&id=b92021b09df70c1609e3547f3d6128dd560be97f
> [2] http://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/tree/kernel?id=next-20130315
> [3] http://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/patch/?id=837718bfd28ba5f474c1182ef2639b6949d854fc
> [4] http://git.kernel.org/cgit/linux/kernel/git/rusty/linux.git/patch/?id=b92021b09df70c1609e3547f3d6128dd560be97f
>
>> Cheers,
>> Rusty.
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-next" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html

[-- Attachment #2: 0001-CONFIG_SYMBOL_PREFIX-cleanup-improved-version.patch --]
[-- Type: application/octet-stream, Size: 4261 bytes --]

From 7c921951784cbc978a5e0193b91f0c3f8a464810 Mon Sep 17 00:00:00 2001
From: Sedat Dilek <sedat.dilek@gmail.com>
Date: Fri, 15 Mar 2013 12:22:26 +0100
Subject: [PATCH] CONFIG_SYMBOL_PREFIX: cleanup (improved version)

Import missing bits from improved version pending in modules-next.

This is a follow up to commit 837718bfd28b ("CONFIG_SYMBOL_PREFIX: cleanup.")
in Linux-Next (next-20130315).

Signed-off-by: Sedat Dilek <sedat.dilek@gmail.com>
---
 Makefile                      |  2 +-
 drivers/mtd/chips/gen_probe.c |  8 +++++---
 include/linux/export.h        |  9 ++-------
 scripts/mod/modpost.c         | 17 +++++++++--------
 4 files changed, 17 insertions(+), 19 deletions(-)

diff --git a/Makefile b/Makefile
index 10fb6c7..0b09ba5 100644
--- a/Makefile
+++ b/Makefile
@@ -1398,7 +1398,7 @@ quiet_cmd_rmfiles = $(if $(wildcard $(rm-files)),CLEAN   $(wildcard $(rm-files))
 # Run depmod only if we have System.map and depmod is executable
 quiet_cmd_depmod = DEPMOD  $(KERNELRELEASE)
       cmd_depmod = $(CONFIG_SHELL) $(srctree)/scripts/depmod.sh $(DEPMOD) \
-                   $(KERNELRELEASE) "$(patsubst y,_,$(CONFIG_HAVE_SYMBOL_PREFIX_UNDERSCORE))"
+                   $(KERNELRELEASE) "$(patsubst y,_,$(CONFIG_HAVE_UNDERSCORE_SYMBOL_PREFIX))"
 
 # Create temporary dir for module support files
 # clean it up only when building all modules
diff --git a/drivers/mtd/chips/gen_probe.c b/drivers/mtd/chips/gen_probe.c
index 00d20ad..74dbb6b 100644
--- a/drivers/mtd/chips/gen_probe.c
+++ b/drivers/mtd/chips/gen_probe.c
@@ -204,14 +204,16 @@ static inline struct mtd_info *cfi_cmdset_unknown(struct map_info *map,
 	struct cfi_private *cfi = map->fldrv_priv;
 	__u16 type = primary?cfi->cfiq->P_ID:cfi->cfiq->A_ID;
 #ifdef CONFIG_MODULES
-	char probename[16+sizeof(VMLINUX_SYMBOL_PREFIX_STR)];
+	char probename[sizeof(VMLINUX_SYMBOL_STR(cfi_cmdset_%4.4X))];
 	cfi_cmdset_fn_t *probe_function;
 
-	sprintf(probename, VMLINUX_SYMBOL_PREFIX_STR "cfi_cmdset_%4.4X", type);
+	sprintf(probename, VMLINUX_SYMBOL_STR(cfi_cmdset_%4.4X), type);
 
 	probe_function = __symbol_get(probename);
 	if (!probe_function) {
-		request_module(probename + sizeof(VMLINUX_SYMBOL_PREFIX_STR)-1);
+		char modname[sizeof("cfi_cmdset_%4.4X")];
+		sprintf(modname, "cfi_cmdset_%4.4X", type);
+		request_module(modname);
 		probe_function = __symbol_get(probename);
 	}
 
diff --git a/include/linux/export.h b/include/linux/export.h
index fc83b2a..412cd50 100644
--- a/include/linux/export.h
+++ b/include/linux/export.h
@@ -5,22 +5,17 @@
  * to reduce the amount of pointless cruft we feed to gcc when only
  * exporting a simple symbol or two.
  *
- * If you feel the need to add #include <linux/foo.h> to this file
- * then you are doing something wrong and should go away silently.
- *
- * If you think the above arrogance just encourages more people to add
- * random crap to this file, you're not alone.
+ * Try not to add #includes here.  It slows compilation and makes kernel
+ * hackers place grumpy comments in header files.
  */
 
 /* Some toolchains use a `_' prefix for all user symbols. */
 #ifdef CONFIG_HAVE_UNDERSCORE_SYMBOL_PREFIX
 #define __VMLINUX_SYMBOL(x) _##x
 #define __VMLINUX_SYMBOL_STR(x) "_" #x
-#define VMLINUX_SYMBOL_PREFIX_STR "_"
 #else
 #define __VMLINUX_SYMBOL(x) x
 #define __VMLINUX_SYMBOL_STR(x) #x
-#define VMLINUX_SYMBOL_PREFIX_STR ""
 #endif
 
 /* Indirect, so macros are expanded before pasting. */
diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c
index 6985021..282decf 100644
--- a/scripts/mod/modpost.c
+++ b/scripts/mod/modpost.c
@@ -630,14 +630,15 @@ static void handle_modversions(struct module *mod, struct elf_info *info,
 		}
 #endif
 
-		if (memcmp(symname, VMLINUX_SYMBOL_PREFIX_STR,
-			   strlen(VMLINUX_SYMBOL_PREFIX_STR)) == 0) {
-			mod->unres =
-			  alloc_symbol(symname +
-			               strlen(VMLINUX_SYMBOL_PREFIX_STR),
-			               ELF_ST_BIND(sym->st_info) == STB_WEAK,
-			               mod->unres);
-		}
+#ifdef CONFIG_HAVE_UNDERSCORE_SYMBOL_PREFIX
+		if (symname[0] != '_')
+			break;
+		else
+			symname++;
+#endif
+		mod->unres = alloc_symbol(symname,
+					  ELF_ST_BIND(sym->st_info) == STB_WEAK,
+					  mod->unres);
 		break;
 	default:
 		/* All exported symbols */
-- 
1.8.2


^ permalink raw reply related	[flat|nested] 6+ messages in thread
* Re: linux-next: manual merge of the signal tree with the modules tree
  2013-03-15 10:21   ` Sedat Dilek
  2013-03-15 11:28     ` Sedat Dilek
@ 2013-03-18  0:56     ` Rusty Russell
  1 sibling, 0 replies; 6+ messages in thread
From: Rusty Russell @ 2013-03-18  0:56 UTC (permalink / raw)
  To: sedat.dilek; +Cc: Stephen Rothwell, Al Viro, linux-next, linux-kernel

Sedat Dilek <sedat.dilek@gmail.com> writes:
> On Fri, Mar 15, 2013 at 5:41 AM, Rusty Russell <rusty@rustcorp.com.au> wrote:
>> Stephen Rothwell <sfr@canb.auug.org.au> writes:
>>> Hi Al,
>>>
>>> Today's linux-next merge of the signal tree got a conflict in
>>> include/asm-generic/unistd.h between commit 837718bfd28b
>>> ("CONFIG_SYMBOL_PREFIX: cleanup") from the modules tree and commit
>>> e1b5bb6d1236 ("consolidate cond_syscall and SYSCALL_ALIAS declarations")
>>> from the signal tree.
>>>
>>> The latter moved the cond_syscall stuff to linkage.h, so I applied the
>>> following patch as a merge fixup and can carry the fix as necessary (no
>>> action is required).  I am not sure if this is completely correct or all
>>> that is needed.
>>
>> Your fix looks correct, thanks.
>>
>> I've been forced to update that patch after another round of
>> improvements, so you may need to re-do the merge.
>>
>
> Hi,
>
> I just looked into modules-next...
> The improved version is in [1]...
> ...and contains a file called "kernel/modsign_certificate.S" which is
> NOT in the latest Linux-Next tree [2].
> So, I thought about reverting the one in -next and apply the new one
> from modules-next.
> This is not possible!

I'd wait until Stephen has done the new merge, which should happen
within 12 hours from now.

Thanks,
Rusty.

^ permalink raw reply	[flat|nested] 6+ messages in thread
end of thread, other threads:[~2013-03-18  2:45 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-03-14  6:27 linux-next: manual merge of the signal tree with the modules tree Stephen Rothwell
2013-03-14 11:24 ` James Hogan
2013-03-15  4:41 ` Rusty Russell
2013-03-15 10:21   ` Sedat Dilek
2013-03-15 11:28     ` Sedat Dilek
2013-03-18  0:56     ` Rusty Russell

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).