[PATCH] Use -fno-unit-at-a-time if gcc supports it

[PATCH] Use -fno-unit-at-a-time if gcc supports it

[Andi Kleen]

Hallo,

gcc 3.4 current has switched to default -fno-unit-at-a-time mode for -O2. 
The 3.3-Hammer branch compiler used in some distributions also does this.

Unfortunately the kernel doesn't compile with unit-at-a-time currently,
it cannot tolerate the reordering of functions in relation to inline
assembly.

This patch just turns it off when gcc supports the option.

I only did it for i386 for now. The problem is actually not i386 specific
(other architectures break too), so it may make sense to move the check_gcc 
stuff into the main Makefile and do it for everybody.

-Andi

--- linux-2.6.0test4-work/arch/i386/Makefile-o	2003-08-23 13:03:08.000000000 +0200
+++ linux-2.6.0test4-work/arch/i386/Makefile	2003-09-05 02:14:07.000000000 +0200
@@ -26,6 +26,10 @@
 # prevent gcc from keeping the stack 16 byte aligned
 CFLAGS += $(call check_gcc,-mpreferred-stack-boundary=2,)
 
+# gcc 3.4/3.3-hammer support -funit-at-a-time mode, but the Kernel is not ready
+# for it yet
+CFLAGS += $(call check_gcc,-fno-unit-at-a-time,)
+
 align := $(subst -functions=0,,$(call check_gcc,-falign-functions=0,-malign-functions=0))
 
 cflags-$(CONFIG_M386)		+= -march=i386

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Aaron Lehmann]

On Fri, Sep 05, 2003 at 02:47:10AM +0200, Andi Kleen wrote:
> 
> Hallo,
> 
> gcc 3.4 current has switched to default -fno-unit-at-a-time mode for -O2. 
> The 3.3-Hammer branch compiler used in some distributions also does this.
> 
> Unfortunately the kernel doesn't compile with unit-at-a-time currently,

Did you mean -funit-at-a-time, rather than the converse?

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Andi Kleen]

On Thu, Sep 04, 2003 at 06:05:35PM -0700, Aaron Lehmann wrote:
> On Fri, Sep 05, 2003 at 02:47:10AM +0200, Andi Kleen wrote:
> > 
> > Hallo,
> > 
> > gcc 3.4 current has switched to default -fno-unit-at-a-time mode for -O2. 
> > The 3.3-Hammer branch compiler used in some distributions also does this.
> > 
> > Unfortunately the kernel doesn't compile with unit-at-a-time currently,
> 
> Did you mean -funit-at-a-time, rather than the converse?

Yep, sorry for the confusion.

It defaults to -funit-at-a-time now, but the kernel must use
 -fno-unit-at-a-time

-Andi

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Jan Hubicka]

> 
> Hallo,
> 
> gcc 3.4 current has switched to default -fno-unit-at-a-time mode for -O2. 
> The 3.3-Hammer branch compiler used in some distributions also does this.
> 
> Unfortunately the kernel doesn't compile with unit-at-a-time currently,
> it cannot tolerate the reordering of functions in relation to inline
> assembly.

How much work would be to fix kernel in this regard?
Are there some cases where this is esential?  Kernel would be nice
target to whole program optimization and GCC is not that far from it
right now.

Honza
> 
> This patch just turns it off when gcc supports the option.
> 
> I only did it for i386 for now. The problem is actually not i386 specific
> (other architectures break too), so it may make sense to move the check_gcc 
> stuff into the main Makefile and do it for everybody.
> 
> -Andi
> 
> --- linux-2.6.0test4-work/arch/i386/Makefile-o	2003-08-23 13:03:08.000000000 +0200
> +++ linux-2.6.0test4-work/arch/i386/Makefile	2003-09-05 02:14:07.000000000 +0200
> @@ -26,6 +26,10 @@
>  # prevent gcc from keeping the stack 16 byte aligned
>  CFLAGS += $(call check_gcc,-mpreferred-stack-boundary=2,)
>  
> +# gcc 3.4/3.3-hammer support -funit-at-a-time mode, but the Kernel is not ready
> +# for it yet
> +CFLAGS += $(call check_gcc,-fno-unit-at-a-time,)
> +
>  align := $(subst -functions=0,,$(call check_gcc,-falign-functions=0,-malign-functions=0))
>  
>  cflags-$(CONFIG_M386)		+= -march=i386

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Linus Torvalds]

On Fri, 5 Sep 2003, Andi Kleen wrote:
> 
> Unfortunately the kernel doesn't compile with unit-at-a-time currently,
> it cannot tolerate the reordering of functions in relation to inline
> assembly.

What is the problem exactly? Is it the exception table getting unordered?  
We _could_ just sort it at boot-time (or, even better, at build time after
the final link) instead...

		Linus

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Andreas Jaeger]

[-- Attachment #1: Type: text/plain, Size: 1347 bytes --]

Linus Torvalds <torvalds@osdl.org> writes:

> On Fri, 5 Sep 2003, Andi Kleen wrote:
>> 
>> Unfortunately the kernel doesn't compile with unit-at-a-time currently,
>> it cannot tolerate the reordering of functions in relation to inline
>> assembly.
>
> What is the problem exactly? Is it the exception table getting unordered?  
> We _could_ just sort it at boot-time (or, even better, at build time after
> the final link) instead...

The problem is that unit-at-a-time sees all functions used and finds
some static functions/variables that are not called anywhere and
therefore drops them, making a smaller binary.  Since GCC does not
look into inline assembler, anything referenced from inline assembler
only, will be treated as not used and therefore removed.

You have to options:
- use attribute ((used)) (implemented since GCC 3.2) to tell GCC that
  a function/variable should never be removed
- use -fno-unit-at-a-time.

Since unit-at-a-time has better inlining heuristics the better way is
to add the used attribute - but that takes some time.  The short-term
solution would be to add the compiler flag,

Andreas
-- 
 Andreas Jaeger, aj@suse.de, http://www.suse.de/~aj
  SuSE Linux AG, Deutschherrnstr. 15-19, 90429 Nürnberg, Germany
   GPG fingerprint = 93A3 365E CE47 B889 DF7F  FED1 389A 563C C272 A126

[-- Attachment #2: Type: application/pgp-signature, Size: 188 bytes --]

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Andreas Jaeger]

[-- Attachment #1: Type: text/plain, Size: 2110 bytes --]

Robert Love <rml@tech9.net> writes:

> On Fri, 2003-09-05 at 11:17, Andreas Jaeger wrote:
>
>
>> Since unit-at-a-time has better inlining heuristics the better way is
>> to add the used attribute - but that takes some time.  The short-term
>> solution would be to add the compiler flag,
>
> Won't we get a linker error if a static symbol is used but
> optimized-away?  It shouldn't be hard to fix the n linker errors that
> crop up.

Yes, we would get a linker error.

> And why are we using static symbols in inline assembly outside of the
> compilation scope?

Don't know.

> Anyhow, if it generates an error, this isn't hard to fix.

Just lots of places...

> Here is the start...
>
> 	Robert Love
>
>
> --- linux-rml/include/linux/compiler.h	Fri Sep  5 11:57:56 2003
> +++ linux/include/linux/compiler.h	Fri Sep  5 12:02:02 2003
> @@ -74,6 +74,19 @@
>  #define __attribute_pure__	/* unimplemented */
>  #endif
>  
> +/*
> + * As of gcc 3.2, we can mark a function as 'used' and gcc will assume that,
> + * even if it does not find a reference to it in any compilation unit.  We
> + * need this for gcc 3.4 and beyond, which can optimize on a program-wide
> + * scope, and not just one file at a time, to avoid static symbols being
> + * discarded.
> + */
> +#if (__GNUC__ == 3 && __GNUC_MINOR__ > 1) || __GNUC__ > 3
> +#define __attribute_used__	__attribute__((used))
> +#else
> +#define __attribute_used__	/* unimplemented */

In glibc we have for the else case:
# define __attribute_used__ __attribute__ ((__unused__))

This might reduce warnings about unused functions.  But this change is
not critical IMO, so your patch looks fine!

> +#endif
> +
>  /* This macro obfuscates arithmetic on a variable address so that gcc
>     shouldn't recognize the original var, and make assumptions about it */
>  #define RELOC_HIDE(ptr, off)					\

Andreas
-- 
 Andreas Jaeger, aj@suse.de, http://www.suse.de/~aj
  SuSE Linux AG, Deutschherrnstr. 15-19, 90429 Nürnberg, Germany
   GPG fingerprint = 93A3 365E CE47 B889 DF7F  FED1 389A 563C C272 A126

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Robert Love]

On Fri, 2003-09-05 at 11:17, Andreas Jaeger wrote:


> Since unit-at-a-time has better inlining heuristics the better way is
> to add the used attribute - but that takes some time.  The short-term
> solution would be to add the compiler flag,

Won't we get a linker error if a static symbol is used but
optimized-away?  It shouldn't be hard to fix the n linker errors that
crop up.

And why are we using static symbols in inline assembly outside of the
compilation scope?

Anyhow, if it generates an error, this isn't hard to fix.

Here is the start...

	Robert Love


--- linux-rml/include/linux/compiler.h	Fri Sep  5 11:57:56 2003
+++ linux/include/linux/compiler.h	Fri Sep  5 12:02:02 2003
@@ -74,6 +74,19 @@
 #define __attribute_pure__	/* unimplemented */
 #endif
 
+/*
+ * As of gcc 3.2, we can mark a function as 'used' and gcc will assume that,
+ * even if it does not find a reference to it in any compilation unit.  We
+ * need this for gcc 3.4 and beyond, which can optimize on a program-wide
+ * scope, and not just one file at a time, to avoid static symbols being
+ * discarded.
+ */
+#if (__GNUC__ == 3 && __GNUC_MINOR__ > 1) || __GNUC__ > 3
+#define __attribute_used__	__attribute__((used))
+#else
+#define __attribute_used__	/* unimplemented */
+#endif
+
 /* This macro obfuscates arithmetic on a variable address so that gcc
    shouldn't recognize the original var, and make assumptions about it */
 #define RELOC_HIDE(ptr, off)					\

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Jakub Jelinek]

On Fri, Sep 05, 2003 at 05:17:00PM +0200, Andreas Jaeger wrote:
> Linus Torvalds <torvalds@osdl.org> writes:
> 
> > On Fri, 5 Sep 2003, Andi Kleen wrote:
> >> 
> >> Unfortunately the kernel doesn't compile with unit-at-a-time currently,
> >> it cannot tolerate the reordering of functions in relation to inline
> >> assembly.
> >
> > What is the problem exactly? Is it the exception table getting unordered?  
> > We _could_ just sort it at boot-time (or, even better, at build time after
> > the final link) instead...
> 
> The problem is that unit-at-a-time sees all functions used and finds
> some static functions/variables that are not called anywhere and
> therefore drops them, making a smaller binary.  Since GCC does not
> look into inline assembler, anything referenced from inline assembler
> only, will be treated as not used and therefore removed.
> 
> You have to options:
> - use attribute ((used)) (implemented since GCC 3.2) to tell GCC that
>   a function/variable should never be removed

To be precise, implemented since GCC 3.2 for functions and since GCC 3.3
for variables.

	Jakub

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Andi Kleen]

> How much work would be to fix kernel in this regard?

The big problem is that -funit-at-a-time is not widely used yet,
so even if we fix the kernel at some point it would likely 
get broken again all the time by people who use older kernels
(= most kernel developers currently)

> Are there some cases where this is esential?  Kernel would be nice
> target to whole program optimization and GCC is not that far from it
> right now.

I'm not sure that is that good an idea. When I was still hacking 
TCP I especially moved some stuff out-of-line in the fast path to avoid 
register pressure. Otherwise gcc would inline rarely used sub functions 
and completely mess up the register allocation in the fast path.
Of course just a call alone messes up the registers somewhat because
of its clobbers, but a full inlining is usually worse.

That was a long time ago, of course the code has significantly changed by
then.

I suspect that is true for a lot of core kernel code - everything
that is worth inlining is already inlined and for the rest it doesn't matter.

On the other hand a lot of driver code seems to be written without
manual consideration for inline. For that it may be worth it. But then
I would consider core kernel code to be more important than driver
code.

Also I fear cross module inlining would expose a lot of latent bugs
(missing barriers etc.) when the optimizer becomes more aggressive. 
I'm not saying this would be a bad thing, just that it may be a lot 
of work to fix (both for compiler and kernel people)

-Andi

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Andi Kleen]

> You have to options:
> - use attribute ((used)) (implemented since GCC 3.2) to tell GCC that
>   a function/variable should never be removed
> - use -fno-unit-at-a-time.

Another problem is the way 32bit emulation is implemented in many 
64bit ports (all copying from sparc64) and now unified. This assumes 
an ordering between global functions and global assembly too. 

Not for i386 though. I think Andrew has already done some cleanups
in this area recently too, but it may still be dubious.

-Andi

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Jeff Garzik]

On Fri, Sep 05, 2003 at 07:27:15PM +0200, Andi Kleen wrote:
> I'm not sure that is that good an idea. When I was still hacking 
> TCP I especially moved some stuff out-of-line in the fast path to avoid 
> register pressure. Otherwise gcc would inline rarely used sub functions 
> and completely mess up the register allocation in the fast path.
> Of course just a call alone messes up the registers somewhat because
> of its clobbers, but a full inlining is usually worse.
[...]
> I suspect that is true for a lot of core kernel code - everything
> that is worth inlining is already inlined and for the rest it doesn't matter.

Definitely , agreed.  In fact, we are moving in the opposite direction:
looking into what we can un-inline...


> On the other hand a lot of driver code seems to be written without
> manual consideration for inline. For that it may be worth it. But then
> I would consider core kernel code to be more important than driver
> code.

Modern network drivers seem fairly aware of it ;-)

> Also I fear cross module inlining would expose a lot of latent bugs
> (missing barriers etc.) when the optimizer becomes more aggressive. 
> I'm not saying this would be a bad thing, just that it may be a lot 
> of work to fix (both for compiler and kernel people)

Agreed.

	Jeff

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Jan Hubicka]

> On Fri, 2003-09-05 at 11:17, Andreas Jaeger wrote:
> 
> 
> > Since unit-at-a-time has better inlining heuristics the better way is
> > to add the used attribute - but that takes some time.  The short-term
> > solution would be to add the compiler flag,
> 
> Won't we get a linker error if a static symbol is used but
> optimized-away?  It shouldn't be hard to fix the n linker errors that
> crop up.

Yes, you get linker error.
You may also run into misscompilation assuiming that function is static
and it is both called by hand in asm and by function call and there is
missing attribute used and asmlinkage definition.  In that case GCC
would conclude to change into register calling convention on i386
breaking asm code.

I would expect this to be rare as functions tends to be used either by
assembly or by normal code but not by both.
> 
> And why are we using static symbols in inline assembly outside of the
> compilation scope?

The toplevel asm statements are common source of this at least in glibc.
I didn't look much into the kernel sources.

I would be very happy if someone did look on that.  It may be well
possible that implementing tricks you do currently with toplevel asm
staements would need further extensions in GCC now and it would be nice
to know about that.

For instance it used to be possible to force function to go into given
section by changing the section by hand, but now you have to use section
attribute (that is cleaner anyway)
> 
> Anyhow, if it generates an error, this isn't hard to fix.
> 
> Here is the start...
> 
> 	Robert Love
> 
> 
> --- linux-rml/include/linux/compiler.h	Fri Sep  5 11:57:56 2003
> +++ linux/include/linux/compiler.h	Fri Sep  5 12:02:02 2003
> @@ -74,6 +74,19 @@
>  #define __attribute_pure__	/* unimplemented */
>  #endif
>  
> +/*
> + * As of gcc 3.2, we can mark a function as 'used' and gcc will assume that,
> + * even if it does not find a reference to it in any compilation unit.  We
> + * need this for gcc 3.4 and beyond, which can optimize on a program-wide
> + * scope, and not just one file at a time, to avoid static symbols being
> + * discarded.
> + */
> +#if (__GNUC__ == 3 && __GNUC_MINOR__ > 1) || __GNUC__ > 3
> +#define __attribute_used__	__attribute__((used))
> +#else
> +#define __attribute_used__	/* unimplemented */
> +#endif
> +
I believe there is little trick - attribute used works either for
variables or functions.  Functions can be marked as used only for GCC
3.4+ if I am right, so you may need __attribute_used_function__ and
__attribute_used_variable__ macros for that.

Honza
>  /* This macro obfuscates arithmetic on a variable address so that gcc
>     shouldn't recognize the original var, and make assumptions about it */
>  #define RELOC_HIDE(ptr, off)					\
> 
> 

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Jan Hubicka]

> > How much work would be to fix kernel in this regard?
> 
> The big problem is that -funit-at-a-time is not widely used yet,
> so even if we fix the kernel at some point it would likely 
> get broken again all the time by people who use older kernels
> (= most kernel developers currently)
> 
> > Are there some cases where this is esential?  Kernel would be nice
> > target to whole program optimization and GCC is not that far from it
> > right now.
> 
> I'm not sure that is that good an idea. When I was still hacking 
> TCP I especially moved some stuff out-of-line in the fast path to avoid 
> register pressure. Otherwise gcc would inline rarely used sub functions 
> and completely mess up the register allocation in the fast path.
> Of course just a call alone messes up the registers somewhat because
> of its clobbers, but a full inlining is usually worse.

You can use -O2 and rely on inline done by hand.  I can add option
-fno-inline-functions-called-once.  That should avoid such a problems.
Anyway it would be nice to mark functions that exist for this reason by
noinline attribute so compiler knows about it, but that is different
story.
> 
> Also I fear cross module inlining would expose a lot of latent bugs
> (missing barriers etc.) when the optimizer becomes more aggressive. 
> I'm not saying this would be a bad thing, just that it may be a lot 
> of work to fix (both for compiler and kernel people)

Some of this should be already tested by folks using Intel compiler I
would hope.

Honza
> 
> -Andi
> 

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[Helge Hafting]

Andreas Jaeger wrote:
[...]
> The problem is that unit-at-a-time sees all functions used and finds
> some static functions/variables that are not called anywhere and
> therefore drops them, making a smaller binary.  Since GCC does not
> look into inline assembler, anything referenced from inline assembler
> only, will be treated as not used and therefore removed.
> 
> You have to options:
> - use attribute ((used)) (implemented since GCC 3.2) to tell GCC that
>   a function/variable should never be removed
> - use -fno-unit-at-a-time.
> 
> Since unit-at-a-time has better inlining heuristics the better way is
> to add the used attribute - but that takes some time.  The short-term
> solution would be to add the compiler flag,
> 
Seems to me that a better solution is to mark the assembly code
in question so gcc knows that is somehow calls the function.
That still allows optimizing away the function whenever the
assembly itself is left out. (Module not compiled or similiar)
Marking the function "used" includes it anyway.

I realize this way probably isn' supported right now,
but people are talking about changing gcc so I
mentioned it as an ideal way.

Helge Hafting

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[David Mosberger]

Here is a minimal platform-independent patch which lets me build the
ia64 kernel with the current gcc-pre3.4 snapshot (which optimizes away
static variables which have no compiler-visible references).  All it
does is mark the init-call macros with "__attribute_unused__" (plus a
minor whitespace fix).

I did have to add an include of <linux/compiler.h> to <linux/init.h>.
That ought to be safe, because the only other include file used by
these files is <linux/config.h>.

Please apply, if there are no objections.

Thanks,

	--david

# This is a BitKeeper generated patch for the following project:
# Project Name: Linux kernel tree
# This patch format is intended for GNU patch command version 2.5 or higher.
# This patch includes the following deltas:
#	           ChangeSet	1.1305  -> 1.1306 
#	include/linux/init.h	1.26    -> 1.27   
#
# The following is the BitKeeper ChangeSet Log
# --------------------------------------------
# 03/09/08	davidm@tiger.hpl.hp.com	1.1306
# Mark initcall macros with __attribute_used__ so the definitions do not
# get optimized away by the compiler (such as the latest GCC pre-3.4).
# --------------------------------------------
#
diff -Nru a/include/linux/init.h b/include/linux/init.h
--- a/include/linux/init.h	Mon Sep  8 13:28:21 2003
+++ b/include/linux/init.h	Mon Sep  8 13:28:21 2003
@@ -2,6 +2,7 @@
 #define _LINUX_INIT_H
 
 #include <linux/config.h>
+#include <linux/compiler.h>
 
 /* These macros are used to mark some functions or 
  * initialized data (doesn't apply to uninitialized data)
@@ -71,15 +72,16 @@
 
 #ifndef __ASSEMBLY__
 
-/* initcalls are now grouped by functionality into separate 
+/* initcalls are now grouped by functionality into separate
  * subsections. Ordering inside the subsections is determined
- * by link order. 
- * For backwards compatibility, initcall() puts the call in 
+ * by link order.
+ * For backwards compatibility, initcall() puts the call in
  * the device init subsection.
  */
 
-#define __define_initcall(level,fn) \
-	static initcall_t __initcall_##fn __attribute__ ((unused,__section__ (".initcall" level ".init"))) = fn
+#define __define_initcall(level,fn)								\
+	static initcall_t __initcall_##fn							\
+	  __attribute_used__ __attribute__ ((__section__ (".initcall" level ".init"))) = fn
 
 #define core_initcall(fn)		__define_initcall("1",fn)
 #define postcore_initcall(fn)		__define_initcall("2",fn)
@@ -95,10 +97,12 @@
 	static exitcall_t __exitcall_##fn __exit_call = fn
 
 #define console_initcall(fn) \
-	static initcall_t __initcall_##fn __attribute__ ((unused,__section__ (".con_initcall.init")))=fn
+	static initcall_t __initcall_##fn \
+	  __attribute_used__ __attribute__ ((__section__ (".con_initcall.init"))) = fn
 
 #define security_initcall(fn) \
-	static initcall_t __initcall_##fn __attribute__ ((unused,__section__ (".security_initcall.init"))) = fn
+	static initcall_t __initcall_##fn \
+	  __attribute_used__ __attribute__ ((__section__ (".security_initcall.init"))) = fn
 
 struct obs_kernel_param {
 	const char *str;
@@ -106,10 +110,10 @@
 };
 
 /* OBSOLETE: see moduleparam.h for the right way. */
-#define __setup(str, fn)						\
-	static char __setup_str_##fn[] __initdata = str;		\
-	static struct obs_kernel_param __setup_##fn			\
-		 __attribute__((unused,__section__ (".init.setup")))	\
+#define __setup(str, fn)							\
+	static char __setup_str_##fn[] __initdata = str;			\
+	static struct obs_kernel_param __setup_##fn				\
+		__attribute_used__ __attribute__((__section__ (".init.setup")))	\
 		= { __setup_str_##fn, fn }
 
 #endif /* __ASSEMBLY__ */

Re: [PATCH] Use -fno-unit-at-a-time if gcc supports it

[David Mosberger-Tang]

For what it's worth, it was straight-forward to get the ia64 linux
kernel to compile & boot with the latest gcc snapshot.  I could make a
patch, if anyone cares, but it was mostly trival stuff: a small
cleanup in init_task.c and a few changes in init.h/compiler.h to use
"attribute ((used))".

Even the wait-channels seem to come out right, though I'm not sure
what's being done there is guaranteed to continue to work as the
compiler gets more aggressive...

	--david
--
Interested in learning more about IA-64 Linux?  Try http://www.lia64.org/book/

>>>>> On Sat, 06 Sep 2003 09:10:10 +0200, Jan Hubicka <jh@suse.cz> said:

  >> On Fri, 2003-09-05 at 11:17, Andreas Jaeger wrote:
  >> 
  >> 
  >> > Since unit-at-a-time has better inlining heuristics the better
  >> way is > to add the used attribute - but that takes some time.
  >> The short-term > solution would be to add the compiler flag,
  >> 
  >> Won't we get a linker error if a static symbol is used but
  >> optimized-away?  It shouldn't be hard to fix the n linker errors
  >> that crop up.

  Jan> Yes, you get linker error.  You may also run into
  Jan> misscompilation assuiming that function is static and it is
  Jan> both called by hand in asm and by function call and there is
  Jan> missing attribute used and asmlinkage definition.  In that case
  Jan> GCC would conclude to change into register calling convention
  Jan> on i386 breaking asm code.

  Jan> I would expect this to be rare as functions tends to be used
  Jan> either by assembly or by normal code but not by both.
  >>  And why are we using static symbols in inline assembly outside
  >> of the compilation scope?

  Jan> The toplevel asm statements are common source of this at least
  Jan> in glibc.  I didn't look much into the kernel sources.

  Jan> I would be very happy if someone did look on that.  It may be
  Jan> well possible that implementing tricks you do currently with
  Jan> toplevel asm staements would need further extensions in GCC now
  Jan> and it would be nice to know about that.

  Jan> For instance it used to be possible to force function to go
  Jan> into given section by changing the section by hand, but now you
  Jan> have to use section attribute (that is cleaner anyway)
  >>  Anyhow, if it generates an error, this isn't hard to fix.
  >> 
  >> Here is the start...
  >> 
  >> Robert Love
  >> 
  >> 
  >> --- linux-rml/include/linux/compiler.h Fri Sep 5 11:57:56 2003
  >> +++ linux/include/linux/compiler.h Fri Sep 5 12:02:02 2003 @@
  >> -74,6 +74,19 @@ #define __attribute_pure__ /* unimplemented */
  >> #endif
  >> 
  >> +/* + * As of gcc 3.2, we can mark a function as 'used' and gcc
  >> will assume that, + * even if it does not find a reference to it
  >> in any compilation unit.  We + * need this for gcc 3.4 and
  >> beyond, which can optimize on a program-wide + * scope, and not
  >> just one file at a time, to avoid static symbols being + *
  >> discarded.  + */ +#if (__GNUC__ == 3 && __GNUC_MINOR__ > 1) ||
  >> __GNUC__ > 3 +#define __attribute_used__ __attribute__((used))
  >> +#else +#define __attribute_used__ /* unimplemented */ +#endif +
  Jan> I believe there is little trick - attribute used works either
  Jan> for variables or functions.  Functions can be marked as used
  Jan> only for GCC 3.4+ if I am right, so you may need
  Jan> __attribute_used_function__ and __attribute_used_variable__
  Jan> macros for that.

  Jan> Honza
  >> /* This macro obfuscates arithmetic on a variable address so that
  >> gcc shouldn't recognize the original var, and make assumptions
  >> about it */ #define RELOC_HIDE(ptr, off) \
  >> 
  >> 
  Jan> - To unsubscribe from this list: send the line "unsubscribe
  Jan> linux-kernel" in the body of a message to
  Jan> majordomo@vger.kernel.org More majordomo info at
  Jan> http://vger.kernel.org/majordomo-info.html Please read the FAQ
  Jan> at http://www.tux.org/lkml/