[PATCH 0/4] Compile kernel with -fwhole-program --combine

[PATCH 0/4] Compile kernel with -fwhole-program --combine

[David Woodhouse]

`-combine'
     If you are compiling multiple source files, this option tells the
     driver to pass all the source files to the compiler at once (for
     those languages for which the compiler can handle this).  This
     will allow intermodule analysis (IMA) to be performed by the
     compiler.  Currently the only language for which this is supported
     is C.  If you pass source files for multiple languages to the
     driver, using this option, the driver will invoke the compiler(s)
     that support IMA once each, passing each compiler all the source
     files appropriate for it.  For those languages that do not support
     IMA this option will be ignored, and the compiler will be invoked
     once for each source file in that language.  If you use this
     option in conjunction with `-save-temps', the compiler will
     generate multiple pre-processed files (one for each source file),
     but only one (combined) `.o' or `.s' file.

`-fwhole-program'
     Assume that the current compilation unit represents whole program
     being compiled.  All public functions and variables with the
     exception of `main' and those merged by attribute
     `externally_visible' become static functions and in a affect gets
     more aggressively optimized by interprocedural optimizers.  While
     this option is equivalent to proper use of `static' keyword for
     programs consisting of single file, in combination with option
     `--combine' this flag can be used to compile most of smaller scale
     C programs since the functions and variables become local for the
     whole combined compilation unit, not for the single source file
     itself.

Using a combination of these two compiler options for building kernel
code leads to some useful optimisation -- especially with modules which
are made up of a bunch of incestuous C files, where none of the global
symbols actually _need_ to be visible outside the directory they reside
in. File systems are a prime example of this -- on PPC64 I see a
reduction in size of ext3.ko by 2.6%, jffs2.ko by 5%, cifs.ko by 8% and
befs.ko by a scary 14%. Strangely, udf.ko seems to have _grown_ by 6.6%
-- that'll probably be another optimisation bug like GCC PR28755.

The same benefits can be extended to the vmlinux too, although there are
caveats with making _everything_ static. However, it's relatively simple
to make EXPORT_SYMBOL() automatically set the 'externally_visible'
attribute on the symbol in question, and to introduce a new '__global'
tag which does the same for those symbols which aren't exported to
modules but which _are_ needed as a global symbol in vmlinux.

Size results from a test build on ppc64 are shown at
http://david.woodhou.se/combine/sizes.csv -- the format is
<old size>,<new size>,<delta>,<percentage * 100>,<object name>

The same file with objects where the size didn't change omitted, and
sorted on percentage is http://david.woodhou.se/combine/sizes-sorted.csv

There are a bunch of GCC bugs which make this interesting:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=27898
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=27889
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28706
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28712
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28744
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28755
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28779

Fixes (or workarounds) for some of these are at
http://david.woodhou.se/combine/gcc-patches/

(Actual patches will follow, to linux-kernel@vger.kernel.org only)

-- 
dwmw2

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[Jan Engelhardt]

>     If you are compiling multiple source files, this option tells the
>     driver to pass all the source files to the compiler at once (for
>     those languages for which the compiler can handle this).  This
>     will allow intermodule analysis (IMA) to be performed by the
>     compiler.  Currently the only language for which this is supported
>     is C.  If you pass source files for multiple languages to the
>     driver, using this option, the driver will invoke the compiler(s)
>     that support IMA once each, passing each compiler all the source
>     files appropriate for it. 

Compiling files on their own (`make drivers/foo/bar.o`) seems to make 
the optimization void. Sure, most people don't stop compiling in 
between. Just a note

> For those languages that do not support
>     IMA this option will be ignored, and the compiler will be invoked
>     once for each source file in that language.  If you use this
>     option in conjunction with `-save-temps', the compiler will
>     generate multiple pre-processed files (one for each source file),
>     but only one (combined) `.o' or `.s' file.

There should be an option (in the kernel's makefile system) to disable 
its use, just in case `gcc *.c` gobbles up a little more RAM 
than is present.

>Using a combination of these two compiler options for building kernel
>code leads to some useful optimisation -- especially with modules which
>are made up of a bunch of incestuous C files, where none of the global
>symbols actually _need_ to be visible outside the directory they reside
>in.

For modules, we have EXPORT_SYMBOL() and any other symbols that are 
'extern' but not exported are not visible to other modules.

>The same benefits can be extended to the vmlinux too, although there are
>caveats with making _everything_ static. However, it's relatively simple
>to make EXPORT_SYMBOL() automatically set the 'externally_visible'
>attribute on the symbol in question, and to introduce a new '__global'
>tag which does the same for those symbols which aren't exported to
>modules but which _are_ needed as a global symbol in vmlinux.

Does the kernel (at least for modules) really use ELF symbol visibility 
(read: __attribute__((visibility(xyz))) or -fvisibility=xyz) for 
lookups? I do not think so, since EXPORT_SYMBOL explicitly puts something 
into __ksymtab/__kstrtab.
If visibility supports had been in GCC a long time ago, I am sure we would 
not need EXPORT_SYMBOL today, or rather, would do it by use of 
__attribute__() rather than a macro that ksymtabs it. Or am I possibly 
misunderstanding something?


Jan Engelhardt
-- 

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[David Woodhouse]

On Thu, 2006-08-24 at 18:48 +0200, Jan Engelhardt wrote:
> >     If you are compiling multiple source files, this option tells the
> >     driver to pass all the source files to the compiler at once (for
> >     those languages for which the compiler can handle this).  This
> >     will allow intermodule analysis (IMA) to be performed by the
> >     compiler.  Currently the only language for which this is supported
> >     is C.  If you pass source files for multiple languages to the
> >     driver, using this option, the driver will invoke the compiler(s)
> >     that support IMA once each, passing each compiler all the source
> >     files appropriate for it. 
> 
> Compiling files on their own (`make drivers/foo/bar.o`) seems to make 
> the optimization void. Sure, most people don't stop compiling in 
> between. Just a note

Yeah, that's largely because my support for this in the makefiles is an
evil hack. I'm sure Sam can come up with something better :)

Actually I'm not entirely sure what you write is true. It'll _build_
fs/jffs2/read.o, for example, but it still won't then use it when I make
the kernel -- it'll just use fs/jffs2/jffs2.o which is built from all
the C files with --combine. So the optimisation isn't lost.

> There should be an option (in the kernel's makefile system) to disable 
> its use, just in case `gcc *.c` gobbles up a little more RAM 
> than is present.

Maybe, yeah. If you look in my 'hacks' patch you'll see I actually did
this for drivers/net/e1000.ko because of a known compiler bug. In an
ideal world it shouldn't be necessary though -- and there is always the
option of just turning off CONFIG_COMBINED_COMPILE if your compiler is
buggy or your system doesn't have the balls.

> >Using a combination of these two compiler options for building kernel
> >code leads to some useful optimisation -- especially with modules which
> >are made up of a bunch of incestuous C files, where none of the global
> >symbols actually _need_ to be visible outside the directory they reside
> >in.
> 
> For modules, we have EXPORT_SYMBOL() and any other symbols that are 
> 'extern' but not exported are not visible to other modules.

Indeed.

> >The same benefits can be extended to the vmlinux too, although there are
> >caveats with making _everything_ static. However, it's relatively simple
> >to make EXPORT_SYMBOL() automatically set the 'externally_visible'
> >attribute on the symbol in question, and to introduce a new '__global'
> >tag which does the same for those symbols which aren't exported to
> >modules but which _are_ needed as a global symbol in vmlinux.
> 
> Does the kernel (at least for modules) really use ELF symbol visibility 
> (read: __attribute__((visibility(xyz))) or -fvisibility=xyz) for 
> lookups? 

No, I think that's something entirely different. This is just about
whether something is static or not.

The -fwhole-program option makes _everything_ static -- not visible
outside the object file it's compiled into. For modules that's fine,
since they really don't need to provide _any_ global symbols.

For stuff compiled into the kernel, however, this can be a problem,
since sometimes we _need_ a symbol to actually be global, and accessible
from somewhere outside the directory it's provided by.

So to overcome this, we use GCC's __attribute__((externally_visible))
which, as documented, just makes it global again -- undoing the effect
of -fwhole-program just for this _one_ symbol.

For anything which is exported by EXPORT_SYMBOL or marked ASMLINKAGE, we
put the externally_visible attribute on it automatically, by modifying
the EXPORT_SYMBOL* and ASMLINKAGE macros to do so. That covers a large
number of the symbols which actually need to be global within vmlinux.
The new '__global' tag covers the rest.

-- 
dwmw2

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[Jan Engelhardt]

>> Compiling files on their own (`make drivers/foo/bar.o`) seems to make 
>> the optimization void. Sure, most people don't stop compiling in 
>> between. Just a note
>
>Actually I'm not entirely sure what you write is true. It'll _build_
>fs/jffs2/read.o, for example, but it still won't then use it when I make
>the kernel -- it'll just use fs/jffs2/jffs2.o which is built from all
>the C files with --combine. So the optimisation isn't lost.

Umm then it spends double the time in compilation, doing:

  read.o <- read.c
  foo.o <- foo.c
  bar.o <- bar.c
  built-in.o <- read.c foo.c bar.c

(cf. default current:
  built-in.o <- read.o foo.o bar.o)

>So to overcome this, we use GCC's __attribute__((externally_visible))
>which, as documented, just makes it global again -- undoing the effect
>of -fwhole-program just for this _one_ symbol.

Interesting. __attribute__((visibility("default"))) does the same?


Jan Engelhardt
-- 

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[Sam Ravnborg]

On Fri, Aug 25, 2006 at 08:01:27AM +0200, Jan Engelhardt wrote:
> >> Compiling files on their own (`make drivers/foo/bar.o`) seems to make 
> >> the optimization void. Sure, most people don't stop compiling in 
> >> between. Just a note
> >
> >Actually I'm not entirely sure what you write is true. It'll _build_
> >fs/jffs2/read.o, for example, but it still won't then use it when I make
> >the kernel -- it'll just use fs/jffs2/jffs2.o which is built from all
> >the C files with --combine. So the optimisation isn't lost.
> 
> Umm then it spends double the time in compilation, doing:
> 
>   read.o <- read.c
It will only do this if you ask for it.
The question was what happened when you did make read.o

>   foo.o <- foo.c
>   bar.o <- bar.c
>   built-in.o <- read.c foo.c bar.c
> 
> (cf. default current:
>   built-in.o <- read.o foo.o bar.o)

And this discussion is btw. mood. If the general opinion is that we shall
include the -combine support all the kbuild infrastructure will anyway
be redone.
There are several small things that are not addressed in todays
implementation and that will be fixed one way or the other.

	Sam

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[David Woodhouse]

On Fri, 2006-08-25 at 09:26 +0200, Sam Ravnborg wrote:
> And this discussion is btw. mood. If the general opinion is that we shall
> include the -combine support all the kbuild infrastructure will anyway
> be redone.
> There are several small things that are not addressed in todays
> implementation and that will be fixed one way or the other.

There's a few _big_ things that aren't addressed in my makefile hack
either. It's very much a proof of concept.

However, if we limit the use of --combine to files within the same
directory, the modifications to the kbuild infrastructure shouldn't be
too intrusive. I strongly suspect that files within the same directory
are where the majority of the benefit is anyway.

Whole-kernel optimisation is likely to be prohibitively expensive, and
can wait for LTO¹.

-- 
dwmw2
¹ http://www.gelato.org/pdf/apr2006/gelato_ICE06apr_lto_mitchell_codesourcery.pdf

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[David Woodhouse]

On Fri, 2006-08-25 at 08:01 +0200, Jan Engelhardt wrote:
> >> Compiling files on their own (`make drivers/foo/bar.o`) seems to make 
> >> the optimization void. Sure, most people don't stop compiling in 
> >> between. Just a note
> >
> >Actually I'm not entirely sure what you write is true. It'll _build_
> >fs/jffs2/read.o, for example, but it still won't then use it when I make
> >the kernel -- it'll just use fs/jffs2/jffs2.o which is built from all
> >the C files with --combine. So the optimisation isn't lost.
> 
> Umm then it spends double the time in compilation, doing:
> 
>   read.o <- read.c
>   foo.o <- foo.c
>   bar.o <- bar.c
>   built-in.o <- read.c foo.c bar.c

Only if you invoke make explicitly for read.o, foo.o and bar.o. If you
just type 'make' then it won't build those.

> >So to overcome this, we use GCC's __attribute__((externally_visible))
> >which, as documented, just makes it global again -- undoing the effect
> >of -fwhole-program just for this _one_ symbol.
> 
> Interesting. __attribute__((visibility("default"))) does the same?

That much you can test for yourself without a fixed compiler. No, it
doesn't.

-- 
dwmw2

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[Jan Engelhardt]

>> >> Compiling files on their own (`make drivers/foo/bar.o`) seems to make 
>> >> the optimization void. Sure, most people don't stop compiling in 
>> >> between. Just a note
>> >
>> >Actually I'm not entirely sure what you write is true. It'll _build_
>> >fs/jffs2/read.o, for example, but it still won't then use it when I make
>> >the kernel -- it'll just use fs/jffs2/jffs2.o which is built from all
>> >the C files with --combine. So the optimisation isn't lost.
>> 
>> Umm then it spends double the time in compilation, doing:
>> 
>>   read.o <- read.c
>>   foo.o <- foo.c
>>   bar.o <- bar.c
>>   built-in.o <- read.c foo.c bar.c
>
>Only if you invoke make explicitly for read.o, foo.o and bar.o. If you
>just type 'make' then it won't build those.

That's what I meant. Assume I explicitly built read.o foo.o and bar.o.
If I then run the regular make, it will rerun gcc for read.c foo.c and 
bar.c rather than using the already-created .o files for linking.


Jan Engelhardt
-- 

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[David Woodhouse]

On Fri, 2006-08-25 at 11:11 +0200, Jan Engelhardt wrote:
> That's what I meant. Assume I explicitly built read.o foo.o and bar.o.
> If I then run the regular make, it will rerun gcc for read.c foo.c and 
> bar.c rather than using the already-created .o files for linking. 

Yes. Just as if I run 'make fs/ntfs/inode.o' and then build my kernel,
my build of fs/ntfs/inode.o isn't used -- because I don't have
CONFIG_NTFS set. 

You built something manually that wasn't needed, and then it wasn't
used. Is there a problem here?

-- 
dwmw2

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[Jan Engelhardt]

>> That's what I meant. Assume I explicitly built read.o foo.o and bar.o.
>> If I then run the regular make, it will rerun gcc for read.c foo.c and 
>> bar.c rather than using the already-created .o files for linking. 
>
>You built something manually that wasn't needed, and then it wasn't
>used. Is there a problem here?

Although I cannot give you any logical reasons, there are possibly reasons 
why building some things manually might help. Oh well, I think in that case 
I just pass a make option CONFIG_NOCOMBINE, as you said.


Jan Engelhardt
-- 

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[David Woodhouse]

On Fri, 2006-08-25 at 11:51 +0200, Jan Engelhardt wrote:
> Although I cannot give you any logical reasons, there are possibly reasons 
> why building some things manually might help. Oh well, I think in that case 
> I just pass a make option CONFIG_NOCOMBINE, as you said. 

Of course. Sometimes when I'm debugging I make stuff manually with
EXTRA_CFLAGS='-O1 -g', and sometimes I'll probably want do the build
with CONFIG_COMBINED_COMPILE= as well.

That should certainly be possible -- and it _is_ possible with my
makefile hack but mostly by chance. It's a feature which we should
probably try to keep though.

-- 
dwmw2

Re: [OLPC-devel] Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[Arnd Bergmann]

On Thursday 24 August 2006 18:48, Jan Engelhardt wrote:
> If visibility supports had been in GCC a long time ago, I am sure we would 
> not need EXPORT_SYMBOL today, or rather, would do it by use of 
> __attribute__() rather than a macro that ksymtabs it. Or am I possibly 
> misunderstanding something?
> 
It's probably true, but the way it's done today gave us CONFIG_MODVERSIONS
and EXPORT_SYMBOL_GPL, which would break when turning EXPORT_SYMBOL into a
simple __attribute__().

	Arnd <><

Re: [OLPC-devel] Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[David Woodhouse]

On Thu, 2006-08-24 at 19:15 +0200, Arnd Bergmann wrote:
> It's probably true, but the way it's done today gave us CONFIG_MODVERSIONS
> and EXPORT_SYMBOL_GPL, which would break when turning EXPORT_SYMBOL into a
> simple __attribute__(). 

ELF can do symbol versioning too. But I don't really see a simple way to
make normal ELF linking work with a kernel like that. The vmlinux ELF
file is long gone by the time the kernel boots -- it needs its own
mechanism for managing its symbols anyway; its own internal dynamic
linker. I suppose you _could_ make it work from the standard ELF linkage
but I don't think it'd be fun.

ObMODVERSIONS: Forgot to mention that my makefile hacks for --combine
break that completely. It tries to preprocess all the files together
with 'gcc -E -D__GENKSYMS__ *.c' -- and you can't do that. 

-- 
dwmw2

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[Adrian Bunk]

On Thu, Aug 24, 2006 at 04:24:28PM +0100, David Woodhouse wrote:
>...
> Using a combination of these two compiler options for building kernel
> code leads to some useful optimisation -- especially with modules which
> are made up of a bunch of incestuous C files, where none of the global
> symbols actually _need_ to be visible outside the directory they reside
> in. File systems are a prime example of this -- on PPC64 I see a
> reduction in size of ext3.ko by 2.6%, jffs2.ko by 5%, cifs.ko by 8% and
> befs.ko by a scary 14%. Strangely, udf.ko seems to have _grown_ by 6.6%
> -- that'll probably be another optimisation bug like GCC PR28755.
>...

Looks good.  :-)

What kernel are your patches against?

2.6.18-rc4 gives me:

<--  snip  -->

...
  CHK     include/linux/version.h
  CHK     include/linux/utsrelease.h
  CC      arch/i386/kernel/asm-offsets.s
  GEN     include/asm-i386/asm-offsets.h
  HOSTCC  scripts/kallsyms
  HOSTCC  scripts/conmakehash
  HOSTCC  scripts/bin2c
make[1]: *** No rule to make target `/home/bunk/linux/linux-2.6.18-rc4/dummy.c', needed by `init/init.o'.  Stop.
make: *** [init] Error 2

<--  snip  -->
 
> dwmw2

cu
Adrian

-- 

    Gentoo kernels are 42 times more popular than SUSE kernels among
    KLive users  (a service by SUSE contractor Andrea Arcangeli that
    gathers data about kernels from many users worldwide).

       There are three kinds of lies: Lies, Damn Lies, and Statistics.
                                                    Benjamin Disraeli

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[David Woodhouse]

On Thu, 2006-08-24 at 23:49 +0200, Adrian Bunk wrote:
> 
> make[1]: *** No rule to make target
> `/home/bunk/linux/linux-2.6.18-rc4/dummy.c', needed by `init/init.o'.
> Stop. 

oops, sorry. touch dummy.c -- evil makefile hack ;)

-- 
dwmw2

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[Rob Landley]

On Thursday 24 August 2006 11:24 am, David Woodhouse wrote:
> Using a combination of these two compiler options for building kernel
> code leads to some useful optimisation

BusyBox has been doing this for months now: "build at once" is one of our 
config options.  I'd like to point out that gcc eats needs several hundred 
megabytes of ram to do this and you have no useful progress indicator between 
starting and ending.  But the result is definitely smaller.

We also tell the linker "--sort-common --gc-sections" which may or may not 
apply here...

Rob
-- 
Never bet against the cheap plastic solution.

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[David Woodhouse]

On Fri, 2006-08-25 at 16:11 -0400, Rob Landley wrote:
> BusyBox has been doing this for months now: "build at once" is one of our 
> config options.  I'd like to point out that gcc eats needs several hundred 
> megabytes of ram to do this and you have no useful progress indicator between 
> starting and ending.  But the result is definitely smaller.

It isn't that bad when you're only building a few files at a time -- I
wouldn't suggest doing it for the whole kernel.

And you get a nice progress indicator at the moment -- a new warning
about global register variables every time it eats a new file, due to
GCC PR27899 :)

> We also tell the linker "--sort-common --gc-sections" which may or may not 
> apply here...

We want --gc-sections (which requires -ffunction-sections
-fdata-sections to be useful) but that's a separate issue. We were doing
it a long time ago for the FR-V kernel; Marcelo has been looking at it
recently for other architectures.

-- 
dwmw2

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[Segher Boessenkool]

> And you get a nice progress indicator at the moment -- a new warning
> about global register variables every time it eats a new file, due to
> GCC PR27899 :)

Or pass -Q to the compiler, it'll print the name of every function
it compiles, and a nice timing report at the end.


Segher

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[Helge Hafting]

David Woodhouse wrote:
> On Fri, 2006-08-25 at 16:11 -0400, Rob Landley wrote:
>   
>> BusyBox has been doing this for months now: "build at once" is one of our 
>> config options.  I'd like to point out that gcc eats needs several hundred 
>> megabytes of ram to do this and you have no useful progress indicator between 
>> starting and ending.  But the result is definitely smaller.
>>     
>
> It isn't that bad when you're only building a few files at a time -- I
> wouldn't suggest doing it for the whole kernel.
>   
I suggest a new makefile target for this.

I.e. "make bzImage" as always for those who do development and
recompile after small changes/patches. 

And a "make optImage" (optimized image) when building a
kernel for production use, when you believe compiling every file
and spending lots of extra time is worth it.

Helge Hafting

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[Jan Engelhardt]

>> 
> I suggest a new makefile target for this.
>
> I.e. "make bzImage" as always for those who do development and
> recompile after small changes/patches. 
> And a "make optImage" (optimized image) when building a
> kernel for production use, when you believe compiling every file
> and spending lots of extra time is worth it.

Good idea. :)


Jan Engelhardt
-- 

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[David Woodhouse]

On Mon, 2006-08-28 at 12:52 +0200, Helge Hafting wrote:
> And a "make optImage" (optimized image) when building a
> kernel for production use, when you believe compiling every file
> and spending lots of extra time is worth it. 

If we revamp the entire kbuild infrastructure to allow building the
_whole_ kernel with --combine then I might be inclined to agree -- we
could do that instead of a CONFIG_COMBINED_COMPILE option.

But if, as I suggest, we're doing the simple option which combines only
the files which tend to get most benefit from it -- those which are in
the same directory -- then there's not a lot of point in the separate
target. It really doesn't take that much extra time.

-- 
dwmw2

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[Ian Stirling]

David Woodhouse wrote:
> On Mon, 2006-08-28 at 12:52 +0200, Helge Hafting wrote:
>> And a "make optImage" (optimized image) when building a
>> kernel for production use, when you believe compiling every file
>> and spending lots of extra time is worth it. 
<snip>
> But if, as I suggest, we're doing the simple option which combines only
> the files which tend to get most benefit from it -- those which are in
> the same directory -- then there's not a lot of point in the separate
> target. It really doesn't take that much extra time.

I thought that it used rather a lot more RAM. I still often(ish) compile 
a kernel on my PII/300/128M. It'd be moderately annoying if it got 
slower, and there was no way to turn it off.

Re: [PATCH 0/4] Compile kernel with -fwhole-program --combine

[David Woodhouse]

On Fri, 2006-09-01 at 20:35 +0100, Ian Stirling wrote:
> I thought that it used rather a lot more RAM. I still often(ish)
> compile a kernel on my PII/300/128M. It'd be moderately annoying if it
> got slower, and there was no way to turn it off.

There will definitely be a way to turn it off. Compilers with all the
relevant bugs fixed are not common yet -- in fact I think I may have the
_only_ existing builds with all the patches collected together.

-- 
dwmw2


-- 
VGER BF report: H 0

[PATCH 1/4] Inconsistent extern declarations.

[David Woodhouse]

When you compile multiple files together with --combine, the compiler
starts to _notice_ when you do things like this in one file:

 extern int ipxrtr_route_packet(struct sock *sk, struct sockaddr_ipx *usipx,
                                struct iovec *iov, int len, int noblock);

.. but the actual function looks like this:

 extern int ipxrtr_route_packet(struct sock *sk, struct sockaddr_ipx *usipx,
                                struct iovec *iov, size_t len, int noblock);

This fixes a bunch of those, which are mostly just a missing 'const' on
the extern declaration.

This patch is a bug-fix which is fairly much independent of the
--combine/-fwhole-program stuff, so can be applied today.

Signed-off-by: David Woodhouse <dwmw2@infradead.org>

diff --git a/drivers/isdn/hisax/config.c b/drivers/isdn/hisax/config.c
index e103503..f647ec3 100644
--- a/drivers/isdn/hisax/config.c
+++ b/drivers/isdn/hisax/config.c
@@ -369,11 +369,11 @@ #endif /* MODULE */
 
 int nrcards;
 
-extern char *l1_revision;
-extern char *l2_revision;
-extern char *l3_revision;
-extern char *lli_revision;
-extern char *tei_revision;
+extern const char *l1_revision;
+extern const char *l2_revision;
+extern const char *l3_revision;
+extern const char *lli_revision;
+extern const char *tei_revision;
 
 char *HiSax_getrev(const char *revision)
 {
diff --git a/drivers/net/skfp/smtinit.c b/drivers/net/skfp/smtinit.c
index 3c8964c..01bf76b 100644
--- a/drivers/net/skfp/smtinit.c
+++ b/drivers/net/skfp/smtinit.c
@@ -36,7 +36,7 @@ #endif
 
 #ifndef MULT_OEM
 #define OEMID(smc,i)	oem_id[i]
-	extern u_char	oem_id[] ;
+	extern const u_char	oem_id[] ;
 #else	/* MULT_OEM */
 #define OEMID(smc,i)	smc->hw.oem_id->oi_mark[i]
 	extern struct s_oem_ids	oem_ids[] ;
diff --git a/fs/debugfs/inode.c b/fs/debugfs/inode.c
index e8ae304..8a15f70 100644
--- a/fs/debugfs/inode.c
+++ b/fs/debugfs/inode.c
@@ -27,7 +27,7 @@ #include <linux/debugfs.h>
 #define DEBUGFS_MAGIC	0x64626720
 
 /* declared over in file.c */
-extern struct file_operations debugfs_file_operations;
+extern const struct file_operations debugfs_file_operations;
 
 static struct vfsmount *debugfs_mount;
 static int debugfs_mount_count;
diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c
index 924ecde..99f3120 100644
--- a/fs/hfsplus/inode.c
+++ b/fs/hfsplus/inode.c
@@ -269,7 +269,7 @@ static int hfsplus_file_release(struct i
 }
 
 extern struct inode_operations hfsplus_dir_inode_operations;
-extern struct file_operations hfsplus_dir_operations;
+extern const struct file_operations hfsplus_dir_operations;
 
 static struct inode_operations hfsplus_file_inode_operations = {
 	.lookup		= hfsplus_file_lookup,
diff --git a/fs/reiserfs/dir.c b/fs/reiserfs/dir.c
index 9aabcc0..68ab9c0 100644
--- a/fs/reiserfs/dir.c
+++ b/fs/reiserfs/dir.c
@@ -11,7 +11,7 @@ #include <linux/smp_lock.h>
 #include <linux/buffer_head.h>
 #include <asm/uaccess.h>
 
-extern struct reiserfs_key MIN_KEY;
+extern const struct reiserfs_key MIN_KEY;
 
 static int reiserfs_readdir(struct file *, void *, filldir_t);
 static int reiserfs_dir_fsync(struct file *filp, struct dentry *dentry,
diff --git a/net/core/sysctl_net_core.c b/net/core/sysctl_net_core.c
index 0253413..1316556 100644
--- a/net/core/sysctl_net_core.c
+++ b/net/core/sysctl_net_core.c
@@ -22,7 +22,7 @@ extern __u32 sysctl_rmem_max;
 extern int sysctl_core_destroy_delay;
 
 #ifdef CONFIG_NET_DIVERT
-extern char sysctl_divert_version[];
+extern const char sysctl_divert_version[];
 #endif /* CONFIG_NET_DIVERT */
 
 #ifdef CONFIG_XFRM
diff --git a/net/ipx/af_ipx.c b/net/ipx/af_ipx.c
index aa34ff4..94b3cb8 100644
--- a/net/ipx/af_ipx.c
+++ b/net/ipx/af_ipx.c
@@ -87,7 +87,7 @@ extern int ipxrtr_add_route(__u32 networ
 			    unsigned char *node);
 extern void ipxrtr_del_routes(struct ipx_interface *intrfc);
 extern int ipxrtr_route_packet(struct sock *sk, struct sockaddr_ipx *usipx,
-			       struct iovec *iov, int len, int noblock);
+			       struct iovec *iov, size_t len, int noblock);
 extern int ipxrtr_route_skb(struct sk_buff *skb);
 extern struct ipx_route *ipxrtr_lookup(__u32 net);
 extern int ipxrtr_ioctl(unsigned int cmd, void __user *arg);
diff --git a/net/irda/irsysctl.c b/net/irda/irsysctl.c
index 86805c3..7606ee7 100644
--- a/net/irda/irsysctl.c
+++ b/net/irda/irsysctl.c
@@ -44,9 +44,9 @@ extern int  sysctl_slot_timeout;
 extern int  sysctl_fast_poll_increase;
 extern char sysctl_devname[];
 extern int  sysctl_max_baud_rate;
-extern int  sysctl_min_tx_turn_time;
-extern int  sysctl_max_tx_data_size;
-extern int  sysctl_max_tx_window;
+extern unsigned  sysctl_min_tx_turn_time;
+extern unsigned  sysctl_max_tx_data_size;
+extern unsigned  sysctl_max_tx_window;
 extern int  sysctl_max_noreply_time;
 extern int  sysctl_warn_noreply_time;
 extern int  sysctl_lap_keepalive_time;
diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c
index 1ab03a2..2a3657b 100644
--- a/net/sctp/protocol.c
+++ b/net/sctp/protocol.c
@@ -83,11 +83,11 @@ kmem_cache_t *sctp_chunk_cachep __read_m
 kmem_cache_t *sctp_bucket_cachep __read_mostly;
 
 extern int sctp_snmp_proc_init(void);
-extern int sctp_snmp_proc_exit(void);
+extern void sctp_snmp_proc_exit(void);
 extern int sctp_eps_proc_init(void);
-extern int sctp_eps_proc_exit(void);
+extern void sctp_eps_proc_exit(void);
 extern int sctp_assocs_proc_init(void);
-extern int sctp_assocs_proc_exit(void);
+extern void sctp_assocs_proc_exit(void);
 
 /* Return the address of the control sock. */
 struct sock *sctp_get_ctl_sock(void)


-- 
dwmw2

Re: [PATCH 1/4] Inconsistent extern declarations.

[Alexey Dobriyan]

On Thu, Aug 24, 2006 at 04:25:18PM +0100, David Woodhouse wrote:
> When you compile multiple files together with --combine, the compiler
> starts to _notice_ when you do things like this in one file:
>
>  extern int ipxrtr_route_packet(struct sock *sk, struct sockaddr_ipx *usipx,
>                                 struct iovec *iov, int len, int noblock);
>
> .. but the actual function looks like this:
>
>  extern int ipxrtr_route_packet(struct sock *sk, struct sockaddr_ipx *usipx,
>                                 struct iovec *iov, size_t len, int noblock);

El-cheapo bug-finder, good. ;-)

> --- a/net/ipx/af_ipx.c
> +++ b/net/ipx/af_ipx.c
> @@ -87,7 +87,7 @@ extern int ipxrtr_add_route(__u32 networ
>  			    unsigned char *node);
>  extern void ipxrtr_del_routes(struct ipx_interface *intrfc);
>  extern int ipxrtr_route_packet(struct sock *sk, struct sockaddr_ipx *usipx,
> -			       struct iovec *iov, int len, int noblock);
> +			       struct iovec *iov, size_t len, int noblock);

Better move proto to include/net/ipx.h since net/ipx/af_ipx.c already
including it.

In general, this patch is wrong. Every external declaration and every
proto should be in only one place.

Re: [PATCH 1/4] Inconsistent extern declarations.

[David Woodhouse]

On Thu, 2006-08-24 at 20:13 +0400, Alexey Dobriyan wrote:
> Better move proto to include/net/ipx.h since net/ipx/af_ipx.c already
> including it.
> 
> In general, this patch is wrong. Every external declaration and every
> proto should be in only one place. 

Yes, there's a lot of merit in that argument.

-- 
dwmw2

Re: [PATCH 1/4] Inconsistent extern declarations.

[Adrian Bunk]

On Thu, Aug 24, 2006 at 04:25:18PM +0100, David Woodhouse wrote:

> When you compile multiple files together with --combine, the compiler
> starts to _notice_ when you do things like this in one file:
> 
>  extern int ipxrtr_route_packet(struct sock *sk, struct sockaddr_ipx *usipx,
>                                 struct iovec *iov, int len, int noblock);
> 
> .. but the actual function looks like this:
> 
>  extern int ipxrtr_route_packet(struct sock *sk, struct sockaddr_ipx *usipx,
>                                 struct iovec *iov, size_t len, int noblock);
> 
> This fixes a bunch of those, which are mostly just a missing 'const' on
> the extern declaration.
>...

Nice.

This is a subset of -Wmissing-prototypes warnings, and I'm working for 
some time to get the function prototypes into header files to avoid such 
bugs (that can in some cases lead to nasty stack corruptions).

But they should be fixed properly by moving the prototypes to header 
files.

> dwmw2

cu
Adrian

-- 

    Gentoo kernels are 42 times more popular than SUSE kernels among
    KLive users  (a service by SUSE contractor Andrea Arcangeli that
    gathers data about kernels from many users worldwide).

       There are three kinds of lies: Lies, Damn Lies, and Statistics.
                                                    Benjamin Disraeli

[PATCH 2/4] Core support for --combine -fwhole-program

[David Woodhouse]

This patch adds a config option for COMBINE_COMPILE, adds the __global
tag to compiler.h and makes EXPORT_SYMBOL automatically use it. It also
contains a crappy Makefile hack which uses -fwhole-program --combine for
all multi-obj .o files -- anything which was 
	obj-m := foo.o
	foo-objs := bar.o wibble.o 
... should now use gcc -o foo.o --combine -fwhole-program bar.c wibble.c

The makefile hack is known not to work properly for generated C files in
out-of-source-tree builds, and for multi-obj files where one of the
sources is _assembly_ instead of C. It's good enough for the proof of
concept, until someone more clueful can do it properly though. It would
be useful to make built-in.o build with --combine from _everything_
which uses standard CFLAGS, rather than doing just what I've done here.

For example only; do not apply as-is. 

diff --git a/include/asm-i386/linkage.h b/include/asm-i386/linkage.h
index f4a6eba..8b06e3e 100644
--- a/include/asm-i386/linkage.h
+++ b/include/asm-i386/linkage.h
@@ -1,7 +1,7 @@
 #ifndef __ASM_LINKAGE_H
 #define __ASM_LINKAGE_H
 
-#define asmlinkage CPP_ASMLINKAGE __attribute__((regparm(0)))
+#define asmlinkage CPP_ASMLINKAGE __attribute__((regparm(0))) __global
 #define FASTCALL(x)	x __attribute__((regparm(3)))
 #define fastcall	__attribute__((regparm(3)))
 
diff --git a/include/linux/compiler.h b/include/linux/compiler.h
index 9b4f110..ed4cf0c 100644
--- a/include/linux/compiler.h
+++ b/include/linux/compiler.h
@@ -156,4 +156,10 @@ #ifndef __attribute_const__
 # define __attribute_const__	/* unimplemented */
 #endif
 
+#ifdef CONFIG_COMBINED_COMPILE
+#define __global __attribute__((externally_visible,used))
+#else
+#define __global
+#endif
+
 #endif /* __LINUX_COMPILER_H */
diff --git a/include/linux/linkage.h b/include/linux/linkage.h
index 932021f..d78eec9 100644
--- a/include/linux/linkage.h
+++ b/include/linux/linkage.h
@@ -10,7 +10,7 @@ #define CPP_ASMLINKAGE
 #endif
 
 #ifndef asmlinkage
-#define asmlinkage CPP_ASMLINKAGE
+#define asmlinkage __global CPP_ASMLINKAGE
 #endif
 
 #ifndef prevent_tail_call
diff --git a/include/linux/module.h b/include/linux/module.h
index 0dfb794..725e5df 100644
--- a/include/linux/module.h
+++ b/include/linux/module.h
@@ -175,7 +175,7 @@ #ifdef CONFIG_MODVERSIONS
 #define __CRC_SYMBOL(sym, sec)					\
 	extern void *__crc_##sym __attribute__((weak));		\
 	static const unsigned long __kcrctab_##sym		\
-	__attribute_used__					\
+	__attribute_used__ __global				\
 	__attribute__((section("__kcrctab" sec), unused))	\
 	= (unsigned long) &__crc_##sym;
 #else
@@ -184,13 +184,15 @@ #endif
 
 /* For every exported symbol, place a struct in the __ksymtab section */
 #define __EXPORT_SYMBOL(sym, sec)				\
-	extern typeof(sym) sym;					\
+	extern typeof(sym) sym __global;		\
 	__CRC_SYMBOL(sym, sec)					\
 	static const char __kstrtab_##sym[]			\
+	asm("__kstrtab_" #sym)					\
 	__attribute__((section("__ksymtab_strings")))		\
 	= MODULE_SYMBOL_PREFIX #sym;                    	\
 	static const struct kernel_symbol __ksymtab_##sym	\
-	__attribute_used__					\
+	asm("__ksymtab_" #sym)					\
+	__attribute_used__ __global				\
 	__attribute__((section("__ksymtab" sec), unused))	\
 	= { (unsigned long)&sym, __kstrtab_##sym }
 
@@ -208,8 +210,8 @@ #ifdef CONFIG_UNUSED_SYMBOLS
 #define EXPORT_UNUSED_SYMBOL(sym) __EXPORT_SYMBOL(sym, "_unused")
 #define EXPORT_UNUSED_SYMBOL_GPL(sym) __EXPORT_SYMBOL(sym, "_unused_gpl")
 #else
-#define EXPORT_UNUSED_SYMBOL(sym)
-#define EXPORT_UNUSED_SYMBOL_GPL(sym)
+#define EXPORT_UNUSED_SYMBOL(sym) extern typeof(sym) sym  __global
+#define EXPORT_UNUSED_SYMBOL_GPL(sym) extern typeof(sym) sym  __global
 #endif
 
 #endif
@@ -470,11 +472,12 @@ void module_add_driver(struct module *, 
 void module_remove_driver(struct device_driver *);
 
 #else /* !CONFIG_MODULES... */
-#define EXPORT_SYMBOL(sym)
-#define EXPORT_SYMBOL_GPL(sym)
-#define EXPORT_SYMBOL_GPL_FUTURE(sym)
-#define EXPORT_UNUSED_SYMBOL(sym)
-#define EXPORT_UNUSED_SYMBOL_GPL(sym)
+
+#define EXPORT_SYMBOL(sym) extern typeof(sym) sym __global
+#define EXPORT_SYMBOL_GPL(sym) EXPORT_SYMBOL(sym)
+#define EXPORT_SYMBOL_GPL_FUTURE(sym) EXPORT_SYMBOL(sym)
+#define EXPORT_UNUSED_SYMBOL(sym) EXPORT_SYMBOL(sym)
+#define EXPORT_UNUSED_SYMBOL_GPL(sym) EXPORT_SYMBOL(sym)
 
 /* Given an address, look for it in the exception tables. */
 static inline const struct exception_table_entry *
diff --git a/init/Kconfig b/init/Kconfig
index a099fc6..b19cfd1 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -468,6 +468,16 @@ config MODULE_FORCE_UNLOAD
 	  rmmod).  This is mainly for kernel developers and desperate users.
 	  If unsure, say N.
 
+config COMBINED_COMPILE
+        bool "Compile multiple source files at once"
+	help
+	  GCC can perform optimisation between different C files by
+	  using its '--combine' option to compile them all at once. By
+	  also adding the '-fwhole-program' option to make global symbols
+	  static, further optimisations are made possible. Say N unless you
+	  have a version of GCC in which at least PR27898, PR28706, PR28712,
+	  PR28744 and PR28779 are fixed.
+	
 config MODVERSIONS
 	bool "Module versioning support"
 	depends on MODULES
diff --git a/scripts/Makefile.build b/scripts/Makefile.build
index 3cb445c..2f2075d 100644
--- a/scripts/Makefile.build
+++ b/scripts/Makefile.build
@@ -128,6 +128,12 @@ modname = $(basetarget)
 $(multi-objs-y:.o=.s)   : modname = $(modname-multi)
 $(multi-objs-y:.o=.lst) : modname = $(modname-multi)
 
+ifdef CONFIG_COMBINED_COMPILE
+$(multi-used-m)		: modkern_cflags += $(CFLAGS_MODULE) -fwhole-program --combine $(sort $(addprefix $(srctree)/$(obj)/,$($(subst $(obj)/,,$(@:.o=-y)):.o=.c) $($(subst $(obj)/,,$(@:.o=-objs)):.o=.c)))
+
+$(multi-used-y)		: CFLAGS += -fwhole-program --combine $(sort $(addprefix $(srctree)/$(obj)/,$($(subst $(obj)/,,$(@:.o=-y)):.o=.c) $($(subst $(obj)/,,$(@:.o=-objs)):.o=.c)))
+endif
+
 quiet_cmd_cc_s_c = CC $(quiet_modtag)  $@
 cmd_cc_s_c       = $(CC) $(c_flags) -fverbose-asm -S -o $@ $<
 
@@ -283,6 +289,23 @@ cmd_link_l_target = rm -f $@; $(AR) $(EX
 targets += $(lib-target)
 endif
 
+ifdef CONFIG_COMBINED_COMPILE
+# We would rather have a list of rules like
+# 	foo.o: $(foo-objs)
+# but that's not so easy, so we rather make all composite objects depend
+# on the set of all their parts
+
+$(multi-used-y) : %.o: $(srctree)/dummy.c $(multi-objs-y:.o=.c) FORCE
+	$(call cmd,force_checksrc)
+	$(call if_changed_rule,cc_o_c)
+
+$(multi-used-m) : %.o: $(srctree)/dummy.c $(multi-objs-m:.o=.c) FORCE
+	$(call cmd,force_checksrc)
+	$(call if_changed_rule,cc_o_c)
+	@{ echo $(@:.o=.ko); echo $(link_multi_deps); } > $(MODVERDIR)/$(@F:.o=.mod)
+
+targets += $(multi-used-y) $(multi-used-m)
+else
 #
 # Rule to link composite objects
 #
@@ -313,7 +336,7 @@ # on the set of all their parts
 	@{ echo $(@:.o=.ko); echo $(link_multi_deps); } > $(MODVERDIR)/$(@F:.o=.mod)
 
 targets += $(multi-used-y) $(multi-used-m)
-
+endif
 
 # Descending
 # ---------------------------------------------------------------------------

-- 
dwmw2

Re: [PATCH 2/4] Core support for --combine -fwhole-program

[Josh Triplett]

On Thu, 2006-08-24 at 16:26 +0100, David Woodhouse wrote:
> This patch adds a config option for COMBINE_COMPILE, adds the __global
> tag to compiler.h and makes EXPORT_SYMBOL automatically use it. It also
> contains a crappy Makefile hack which uses -fwhole-program --combine for
> all multi-obj .o files -- anything which was 
> 	obj-m := foo.o
> 	foo-objs := bar.o wibble.o 
> ... should now use gcc -o foo.o --combine -fwhole-program bar.c wibble.c
> 
> The makefile hack is known not to work properly for generated C files in
> out-of-source-tree builds, and for multi-obj files where one of the
> sources is _assembly_ instead of C. It's good enough for the proof of
> concept, until someone more clueful can do it properly though. It would
> be useful to make built-in.o build with --combine from _everything_
> which uses standard CFLAGS, rather than doing just what I've done here.

Would the generation and use of preprocessed source files ($x.c -> $x.i)
help solve the "standard CFLAGS" problem here, and allow a subsequent
single compilation of the entire kernel (sans modules) with --combine
-fwhole-program?

- Josh Triplett

Re: [PATCH 2/4] Core support for --combine -fwhole-program

[David Woodhouse]

On Thu, 2006-08-24 at 10:27 -0700, Josh Triplett wrote:
> Would the generation and use of preprocessed source files ($x.c -> $x.i)
> help solve the "standard CFLAGS" problem here, and allow a subsequent
> single compilation of the entire kernel (sans modules) with --combine
> -fwhole-program?

Not for stuff like CFLAGS_raid6altivec1.o := -maltivec -mabi-altivec

Stuff which is _purely_ #defines, such as...
	ncr53c8xx-flags-$(CONFIG_SCSI_ZALON) \
	                := -DCONFIG_NCR53C8XX_PREFETCH -DSCSI_NCR_BIG_ENDIAN \
	                        -DCONFIG_SCSI_NCR53C8XX_NO_WORD_TRANSFERS
	CFLAGS_ncr53c8xx.o      := $(ncr53c8xx-flags-y) $(ncr53c8xx-flags-m)

... can probably be done differently anyway.

It's not entirely unreasonable to treat files with their own CFLAGS just
as we have to treat .S files -- compile them separately, then link.

-- 
dwmw2

Re: [PATCH 2/4] Core support for --combine -fwhole-program

[Adrian Bunk]

On Thu, Aug 24, 2006 at 04:26:07PM +0100, David Woodhouse wrote:
>...
> index 0dfb794..725e5df 100644
> --- a/include/linux/module.h
> +++ b/include/linux/module.h
>...
> @@ -470,11 +472,12 @@ void module_add_driver(struct module *, 
>  void module_remove_driver(struct device_driver *);
>  
>  #else /* !CONFIG_MODULES... */
> -#define EXPORT_SYMBOL(sym)
> -#define EXPORT_SYMBOL_GPL(sym)
> -#define EXPORT_SYMBOL_GPL_FUTURE(sym)
> -#define EXPORT_UNUSED_SYMBOL(sym)
> -#define EXPORT_UNUSED_SYMBOL_GPL(sym)
> +
> +#define EXPORT_SYMBOL(sym) extern typeof(sym) sym __global
> +#define EXPORT_SYMBOL_GPL(sym) EXPORT_SYMBOL(sym)
> +#define EXPORT_SYMBOL_GPL_FUTURE(sym) EXPORT_SYMBOL(sym)
> +#define EXPORT_UNUSED_SYMBOL(sym) EXPORT_SYMBOL(sym)
> +#define EXPORT_UNUSED_SYMBOL_GPL(sym) EXPORT_SYMBOL(sym)
>...

If a "build everything except for assembler files at once" approach is 
possible, it should be possible to revert this and get even further 
savings.

> dwmw2

cu
Adrian

-- 

    Gentoo kernels are 42 times more popular than SUSE kernels among
    KLive users  (a service by SUSE contractor Andrea Arcangeli that
    gathers data about kernels from many users worldwide).

       There are three kinds of lies: Lies, Damn Lies, and Statistics.
                                                    Benjamin Disraeli

Re: [PATCH 2/4] Core support for --combine -fwhole-program

[David Woodhouse]

On Thu, 2006-08-24 at 23:33 +0200, Adrian Bunk wrote:
> If a "build everything except for assembler files at once" approach is 
> possible, it should be possible to revert this and get even further 
> savings.

Only if we build _everything_ at once, which may take an insane amount
of RAM. Doing it a directory at a time makes a certain amount of sense,
and tends to combine the most incestuous code -- although maybe
combinations like building arch/$ARCH/kernel/ with kernel/ (and likewise
mm) could be an interesting experiment.

I suspected that most of the 'further savings' to which you refer above
could be achieved more easily with -ffunction-sections -fdata-sections
--gc-sections

-- 
dwmw2

Re: [PATCH 2/4] Core support for --combine -fwhole-program

[Adrian Bunk]

On Fri, Aug 25, 2006 at 10:37:23AM +0100, David Woodhouse wrote:
> On Thu, 2006-08-24 at 23:33 +0200, Adrian Bunk wrote:
> > If a "build everything except for assembler files at once" approach is 
> > possible, it should be possible to revert this and get even further 
> > savings.
> 
> Only if we build _everything_ at once, which may take an insane amount
> of RAM. Doing it a directory at a time makes a certain amount of sense,
> and tends to combine the most incestuous code -- although maybe
> combinations like building arch/$ARCH/kernel/ with kernel/ (and likewise
> mm) could be an interesting experiment.

My hope is "insane" would be something like "1 GB of RAM" that is no 
longer insane on current computers. [1]

> I suspected that most of the 'further savings' to which you refer above
> could be achieved more easily with -ffunction-sections -fdata-sections
> --gc-sections

AFAIR -ffunction-sections/-fdata-sections cause some overhead in the 
resulting binary?

> dwmw2

cu
Adrian

[1] The interesting cases are embedded systems needing a small kernel
    that gets built on a much bigger system.
    Whether this should be the default compile mode for everyone is a
    different issue.

-- 

       "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: [PATCH 2/4] Core support for --combine -fwhole-program

[David Woodhouse]

On Fri, 2006-08-25 at 12:30 +0200, Adrian Bunk wrote:
> My hope is "insane" would be something like "1 GB of RAM" that is no 
> longer insane on current computers. [1]

Maybe -- but still, that's quite a steep requirement, and would also
require a fairly major redesign of the current kbuild architecture. 

I'm inclined to stick with per-directory builds for now, which is both
relatively simple to implement and where the _majority_ of the benefit
it likely to be seen.

Whole-kernel optimisation is something I'm inclined to leave until LTO
happens -- but don't let me stop you from investigating it.

> > I suspected that most of the 'further savings' to which you refer above
> > could be achieved more easily with -ffunction-sections -fdata-sections
> > --gc-sections
> 
> AFAIR -ffunction-sections/-fdata-sections cause some overhead in the 
> resulting binary?

Not that I'm aware of. There's overhead in the resulting ELF files, but
we combine sections in the vmlinux.lds so there isn't even that overhead
in the vmlinux. 

-- 
dwmw2

[PATCH 3/4] Add __global tag where needed.

[David Woodhouse]

This patch just adds the __global tag to functions and variables which
are used from outside the directory in which they're provided, to
prevent -fwhole-program from eating them.

If we just define __global as a dummy for now, then there's no harm in
applying this patch as-is -- but it's largely pointless without the
Makefile changes.

Signed-off-by: David Woodhouse <dwmw2@infradead.org>

diff --git a/arch/i386/mm/extable.c b/arch/i386/mm/extable.c
index de03c54..68a2409 100644
--- a/arch/i386/mm/extable.c
+++ b/arch/i386/mm/extable.c
@@ -6,7 +6,7 @@ #include <linux/module.h>
 #include <linux/spinlock.h>
 #include <asm/uaccess.h>
 
-int fixup_exception(struct pt_regs *regs)
+int __global fixup_exception(struct pt_regs *regs)
 {
 	const struct exception_table_entry *fixup;
 
diff --git a/arch/i386/mm/fault.c b/arch/i386/mm/fault.c
index f727946..f8f90b2 100644
--- a/arch/i386/mm/fault.c
+++ b/arch/i386/mm/fault.c
@@ -32,13 +32,13 @@ extern void die(const char *,struct pt_r
 
 #ifdef CONFIG_KPROBES
 ATOMIC_NOTIFIER_HEAD(notify_page_fault_chain);
-int register_page_fault_notifier(struct notifier_block *nb)
+int __global register_page_fault_notifier(struct notifier_block *nb)
 {
 	vmalloc_sync_all();
 	return atomic_notifier_chain_register(&notify_page_fault_chain, nb);
 }
 
-int unregister_page_fault_notifier(struct notifier_block *nb)
+int __global unregister_page_fault_notifier(struct notifier_block *nb)
 {
 	return atomic_notifier_chain_unregister(&notify_page_fault_chain, nb);
 }
@@ -325,8 +325,8 @@ static inline int vmalloc_fault(unsigned
  *	bit 3 == 1 means use of reserved bit detected
  *	bit 4 == 1 means fault was an instruction fetch
  */
-fastcall void __kprobes do_page_fault(struct pt_regs *regs,
-				      unsigned long error_code)
+fastcall void __kprobes __global do_page_fault(struct pt_regs *regs,
+					       unsigned long error_code)
 {
 	struct task_struct *tsk;
 	struct mm_struct *mm;
@@ -626,7 +626,7 @@ do_sigbus:
 }
 
 #ifndef CONFIG_X86_PAE
-void vmalloc_sync_all(void)
+void __global vmalloc_sync_all(void)
 {
 	/*
 	 * Note that races in the updates of insync and start aren't
diff --git a/arch/i386/mm/hugetlbpage.c b/arch/i386/mm/hugetlbpage.c
index 1719a81..2820b17 100644
--- a/arch/i386/mm/hugetlbpage.c
+++ b/arch/i386/mm/hugetlbpage.c
@@ -17,7 +17,7 @@ #include <asm/mman.h>
 #include <asm/tlb.h>
 #include <asm/tlbflush.h>
 
-pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr)
+pte_t * __global huge_pte_alloc(struct mm_struct *mm, unsigned long addr)
 {
 	pgd_t *pgd;
 	pud_t *pud;
@@ -32,7 +32,7 @@ pte_t *huge_pte_alloc(struct mm_struct *
 	return pte;
 }
 
-pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr)
+pte_t * __global huge_pte_offset(struct mm_struct *mm, unsigned long addr)
 {
 	pgd_t *pgd;
 	pud_t *pud;
@@ -48,7 +48,7 @@ pte_t *huge_pte_offset(struct mm_struct 
 }
 
 #if 0	/* This is just for testing */
-struct page *
+struct page * __global
 follow_huge_addr(struct mm_struct *mm, unsigned long address, int write)
 {
 	unsigned long start = address;
@@ -87,18 +87,18 @@ follow_huge_pmd(struct mm_struct *mm, un
 
 #else
 
-struct page *
+struct page * __global
 follow_huge_addr(struct mm_struct *mm, unsigned long address, int write)
 {
 	return ERR_PTR(-EINVAL);
 }
 
-int pmd_huge(pmd_t pmd)
+int __global pmd_huge(pmd_t pmd)
 {
 	return !!(pmd_val(pmd) & _PAGE_PSE);
 }
 
-struct page *
+struct page * __global
 follow_huge_pmd(struct mm_struct *mm, unsigned long address,
 		pmd_t *pmd, int write)
 {
@@ -245,7 +245,7 @@ fail:
 	return addr;
 }
 
-unsigned long
+unsigned long __global
 hugetlb_get_unmapped_area(struct file *file, unsigned long addr,
 		unsigned long len, unsigned long pgoff, unsigned long flags)
 {
diff --git a/arch/i386/mm/init.c b/arch/i386/mm/init.c
index 89e8486..89661e2 100644
--- a/arch/i386/mm/init.c
+++ b/arch/i386/mm/init.c
@@ -43,9 +43,9 @@ #include <asm/tlb.h>
 #include <asm/tlbflush.h>
 #include <asm/sections.h>
 
-unsigned int __VMALLOC_RESERVE = 128 << 20;
+unsigned int __global __VMALLOC_RESERVE = 128 << 20;
 
-DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);
+__global DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);
 unsigned long highstart_pfn, highend_pfn;
 
 static int noinline do_test_wp_bit(void);
@@ -329,7 +329,7 @@ #endif /* CONFIG_HIGHMEM */
 
 unsigned long long __PAGE_KERNEL = _PAGE_KERNEL;
 EXPORT_SYMBOL(__PAGE_KERNEL);
-unsigned long long __PAGE_KERNEL_EXEC = _PAGE_KERNEL_EXEC;
+unsigned long long __global __PAGE_KERNEL_EXEC = _PAGE_KERNEL_EXEC;
 
 #ifdef CONFIG_NUMA
 extern void __init remap_numa_kva(void);
@@ -390,7 +390,7 @@ #if defined(CONFIG_SOFTWARE_SUSPEND) || 
  * Swap suspend & friends need this for resume because things like the intel-agp
  * driver might have split up a kernel 4MB mapping.
  */
-char __nosavedata swsusp_pg_dir[PAGE_SIZE]
+char __nosavedata __global swsusp_pg_dir[PAGE_SIZE]
 	__attribute__ ((aligned (PAGE_SIZE)));
 
 static inline void save_pg_dir(void)
@@ -435,7 +435,7 @@ u64 __supported_pte_mask __read_mostly =
  * on      Enable
  * off     Disable
  */
-void __init noexec_setup(const char *str)
+void __init __global noexec_setup(const char *str)
 {
 	if (!strncmp(str, "on",2) && cpu_has_nx) {
 		__supported_pte_mask |= _PAGE_NX;
@@ -446,7 +446,7 @@ void __init noexec_setup(const char *str
 	}
 }
 
-int nx_enabled = 0;
+int __global nx_enabled = 0;
 #ifdef CONFIG_X86_PAE
 
 static void __init set_nx(void)
@@ -501,7 +501,7 @@ #endif
  * This routines also unmaps the page at virtual kernel address 0, so
  * that we can trap those pesky NULL-reference errors in the kernel.
  */
-void __init paging_init(void)
+void __init __global paging_init(void)
 {
 #ifdef CONFIG_X86_PAE
 	set_nx();
@@ -566,7 +566,7 @@ #endif
 
 static struct kcore_list kcore_mem, kcore_vmalloc; 
 
-void __init mem_init(void)
+void __init __global mem_init(void)
 {
 	extern int ppro_with_ram_bug(void);
 	int codesize, reservedpages, datasize, initsize;
@@ -674,7 +674,7 @@ #endif
 kmem_cache_t *pgd_cache;
 kmem_cache_t *pmd_cache;
 
-void __init pgtable_cache_init(void)
+void __init __global pgtable_cache_init(void)
 {
 	if (PTRS_PER_PMD > 1) {
 		pmd_cache = kmem_cache_create("pmd",
@@ -725,7 +725,7 @@ static int noinline do_test_wp_bit(void)
 
 #ifdef CONFIG_DEBUG_RODATA
 
-void mark_rodata_ro(void)
+void __global mark_rodata_ro(void)
 {
 	unsigned long addr = (unsigned long)__start_rodata;
 
@@ -745,7 +745,7 @@ void mark_rodata_ro(void)
 }
 #endif
 
-void free_init_pages(char *what, unsigned long begin, unsigned long end)
+void __global free_init_pages(char *what, unsigned long begin, unsigned long end)
 {
 	unsigned long addr;
 
@@ -759,7 +759,7 @@ void free_init_pages(char *what, unsigne
 	printk(KERN_INFO "Freeing %s: %ldk freed\n", what, (end - begin) >> 10);
 }
 
-void free_initmem(void)
+void __global free_initmem(void)
 {
 	free_init_pages("unused kernel memory",
 			(unsigned long)(&__init_begin),
@@ -767,7 +767,7 @@ void free_initmem(void)
 }
 
 #ifdef CONFIG_BLK_DEV_INITRD
-void free_initrd_mem(unsigned long start, unsigned long end)
+void __global free_initrd_mem(unsigned long start, unsigned long end)
 {
 	free_init_pages("initrd memory", start, end);
 }
diff --git a/arch/i386/mm/ioremap.c b/arch/i386/mm/ioremap.c
index 247fde7..5340869 100644
--- a/arch/i386/mm/ioremap.c
+++ b/arch/i386/mm/ioremap.c
@@ -280,7 +280,7 @@ void iounmap(volatile void __iomem *addr
 }
 EXPORT_SYMBOL(iounmap);
 
-void __init *bt_ioremap(unsigned long phys_addr, unsigned long size)
+void __init __global *bt_ioremap(unsigned long phys_addr, unsigned long size)
 {
 	unsigned long offset, last_addr;
 	unsigned int nrpages;
@@ -324,7 +324,7 @@ void __init *bt_ioremap(unsigned long ph
 	return (void*) (offset + fix_to_virt(FIX_BTMAP_BEGIN));
 }
 
-void __init bt_iounmap(void *addr, unsigned long size)
+void __init __global bt_iounmap(void *addr, unsigned long size)
 {
 	unsigned long virt_addr;
 	unsigned long offset;
diff --git a/arch/i386/mm/mmap.c b/arch/i386/mm/mmap.c
index e4730a1..ec8f27e 100644
--- a/arch/i386/mm/mmap.c
+++ b/arch/i386/mm/mmap.c
@@ -56,7 +56,7 @@ static inline unsigned long mmap_base(st
  * This function, called very early during the creation of a new
  * process VM image, sets up which VM layout function to use:
  */
-void arch_pick_mmap_layout(struct mm_struct *mm)
+void __global arch_pick_mmap_layout(struct mm_struct *mm)
 {
 	/*
 	 * Fall back to the standard layout if the personality
diff --git a/arch/i386/mm/pgtable.c b/arch/i386/mm/pgtable.c
index bd98768..6161dd9 100644
--- a/arch/i386/mm/pgtable.c
+++ b/arch/i386/mm/pgtable.c
@@ -21,7 +21,7 @@ #include <asm/e820.h>
 #include <asm/tlb.h>
 #include <asm/tlbflush.h>
 
-void show_mem(void)
+void __global show_mem(void)
 {
 	int total = 0, reserved = 0;
 	int shared = 0, cached = 0;
@@ -137,7 +137,7 @@ void set_pmd_pfn(unsigned long vaddr, un
 	__flush_tlb_one(vaddr);
 }
 
-void __set_fixmap (enum fixed_addresses idx, unsigned long phys, pgprot_t flags)
+void __global __set_fixmap (enum fixed_addresses idx, unsigned long phys, pgprot_t flags)
 {
 	unsigned long address = __fix_to_virt(idx);
 
@@ -148,12 +148,12 @@ void __set_fixmap (enum fixed_addresses 
 	set_pte_pfn(address, phys >> PAGE_SHIFT, flags);
 }
 
-pte_t *pte_alloc_one_kernel(struct mm_struct *mm, unsigned long address)
+pte_t * __global pte_alloc_one_kernel(struct mm_struct *mm, unsigned long address)
 {
 	return (pte_t *)__get_free_page(GFP_KERNEL|__GFP_REPEAT|__GFP_ZERO);
 }
 
-struct page *pte_alloc_one(struct mm_struct *mm, unsigned long address)
+struct page * __global pte_alloc_one(struct mm_struct *mm, unsigned long address)
 {
 	struct page *pte;
 
@@ -234,7 +234,7 @@ void pgd_dtor(void *pgd, kmem_cache_t *c
 	spin_unlock_irqrestore(&pgd_lock, flags);
 }
 
-pgd_t *pgd_alloc(struct mm_struct *mm)
+pgd_t * __global pgd_alloc(struct mm_struct *mm)
 {
 	int i;
 	pgd_t *pgd = kmem_cache_alloc(pgd_cache, GFP_KERNEL);
@@ -257,7 +257,7 @@ out_oom:
 	return NULL;
 }
 
-void pgd_free(pgd_t *pgd)
+void __global pgd_free(pgd_t *pgd)
 {
 	int i;
 
diff --git a/drivers/ide/ppc/pmac.c b/drivers/ide/ppc/pmac.c
index ebf961f..d6f5437 100644
--- a/drivers/ide/ppc/pmac.c
+++ b/drivers/ide/ppc/pmac.c
@@ -1056,7 +1056,7 @@ pmac_ide_get_irq(unsigned long base)
 
 static int ide_majors[] = { 3, 22, 33, 34, 56, 57 };
 
-dev_t __init
+dev_t __init __global
 pmac_find_ide_boot(char *bootdevice, int n)
 {
 	int i;
diff --git a/drivers/md/md.c b/drivers/md/md.c
index b6d1602..11cc13a 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -2887,7 +2887,7 @@ static struct kobj_type md_ktype = {
 	.default_attrs	= md_default_attrs,
 };
 
-int mdp_major = 0;
+int __global mdp_major = 0;
 
 static struct kobject *md_probe(dev_t dev, int *part, void *data)
 {
@@ -5578,13 +5578,12 @@ #ifndef MODULE
 static dev_t detected_devices[128];
 static int dev_cnt;
 
-void md_autodetect_dev(dev_t dev)
+void __global md_autodetect_dev(dev_t dev)
 {
 	if (dev_cnt >= 0 && dev_cnt < 127)
 		detected_devices[dev_cnt++] = dev;
 }
 
-
 static void autostart_arrays(int part)
 {
 	mdk_rdev_t *rdev;
diff --git a/drivers/video/fbcvt.c b/drivers/video/fbcvt.c
index b549899..baac324 100644
--- a/drivers/video/fbcvt.c
+++ b/drivers/video/fbcvt.c
@@ -301,7 +301,7 @@ static void fb_cvt_convert_to_mode(struc
  * DESCRIPTION:
  * Computes video timings using VESA(TM) Coordinated Video Timings
  */
-int fb_find_mode_cvt(struct fb_videomode *mode, int margins, int rb)
+int __global fb_find_mode_cvt(struct fb_videomode *mode, int margins, int rb)
 {
 	struct fb_cvt_data cvt;
 
diff --git a/fs/aio.c b/fs/aio.c
index 9506301..3290a73 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -42,8 +42,8 @@ #endif
 
 /*------ sysctl variables----*/
 static DEFINE_SPINLOCK(aio_nr_lock);
-unsigned long aio_nr;		/* current system wide number of aio requests */
-unsigned long aio_max_nr = 0x10000; /* system wide maximum number of aio requests */
+unsigned long __global aio_nr;		/* current system wide number of aio requests */
+unsigned long __global aio_max_nr = 0x10000; /* system wide maximum number of aio requests */
 /*----end sysctl variables---*/
 
 static kmem_cache_t	*kiocb_cachep;
@@ -332,7 +332,7 @@ ssize_t fastcall wait_on_sync_kiocb(stru
  * go away, they will call put_ioctx and release any pinned memory
  * associated with the request (held via struct page * references).
  */
-void fastcall exit_aio(struct mm_struct *mm)
+void fastcall __global exit_aio(struct mm_struct *mm)
 {
 	struct kioctx *ctx = mm->ioctx_list;
 	mm->ioctx_list = NULL;
diff --git a/fs/bio.c b/fs/bio.c
index 6a0b9ad..049afc4 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -745,9 +746,9 @@ struct bio *bio_map_user(request_queue_t
  *	Map the user space address into a bio suitable for io to a block
  *	device. Returns an error pointer in case of error.
  */
-struct bio *bio_map_user_iov(request_queue_t *q, struct block_device *bdev,
-			     struct sg_iovec *iov, int iov_count,
-			     int write_to_vm)
+struct bio * __global bio_map_user_iov(request_queue_t *q, struct block_device *bdev,
+				       struct sg_iovec *iov, int iov_count,
+				       int write_to_vm)
 {
 	struct bio *bio;
 	int len = 0, i;
@@ -917,7 +918,7 @@ struct bio *bio_map_kern(request_queue_t
 /*
  * bio_set_pages_dirty() will mark all the bio's pages as dirty.
  */
-void bio_set_pages_dirty(struct bio *bio)
+void __global bio_set_pages_dirty(struct bio *bio)
 {
 	struct bio_vec *bvec = bio->bi_io_vec;
 	int i;
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 3753457..de38fb4 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -388,7 +388,7 @@ struct block_device *bdget(dev_t dev)
 
 EXPORT_SYMBOL(bdget);
 
-long nr_blockdev_pages(void)
+long __global nr_blockdev_pages(void)
 {
 	struct list_head *p;
 	long ret = 0;
diff --git a/fs/buffer.c b/fs/buffer.c
index 71649ef..4073846 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -284,7 +284,7 @@ asmlinkage long sys_sync(void)
 	return 0;
 }
 
-void emergency_sync(void)
+void __global emergency_sync(void)
 {
 	pdflush_operation(do_sync, 0);
 }
@@ -318,7 +318,7 @@ int file_fsync(struct file *filp, struct
 	return ret;
 }
 
-long do_fsync(struct file *file, int datasync)
+long __global do_fsync(struct file *file, int datasync)
 {
 	int ret;
 	int err;
@@ -1630,7 +1630,7 @@ out:
 }
 EXPORT_SYMBOL(block_invalidatepage);
 
-void do_invalidatepage(struct page *page, unsigned long offset)
+void __global do_invalidatepage(struct page *page, unsigned long offset)
 {
 	void (*invalidatepage)(struct page *, unsigned long);
 	invalidatepage = page->mapping->a_ops->invalidatepage ? :
@@ -3061,7 +3061,7 @@ static kmem_cache_t *bh_cachep;
  */
 static int max_buffer_heads;
 
-int buffer_heads_over_limit;
+int __global buffer_heads_over_limit;
 
 struct bh_accounting {
 	int nr;			/* Number of live bh's */
@@ -3141,7 +3141,7 @@ static int buffer_cpu_notify(struct noti
 }
 #endif /* CONFIG_HOTPLUG_CPU */
 
-void __init buffer_init(void)
+void __init __global buffer_init(void)
 {
 	int nrpages;
 
diff --git a/fs/char_dev.c b/fs/char_dev.c
index 3483d3c..236a7d1 100644
--- a/fs/char_dev.c
+++ b/fs/char_dev.c
@@ -46,7 +46,7 @@ static inline int major_to_index(int maj
 
 #ifdef CONFIG_PROC_FS
 
-void chrdev_show(struct seq_file *f, off_t offset)
+void __global chrdev_show(struct seq_file *f, off_t offset)
 {
 	struct char_device_struct *cd;
 
diff --git a/fs/compat.c b/fs/compat.c
index e31e9cf..b635d8f 100644
--- a/fs/compat.c
+++ b/fs/compat.c
@@ -55,7 +55,7 @@ #include <asm/ioctls.h>
 
 extern void sigset_from_compat(sigset_t *set, compat_sigset_t *compat);
 
-int compat_log = 1;
+int __global compat_log = 1;
 
 int compat_printk(const char *fmt, ...)
 {
@@ -1503,7 +1503,7 @@ #endif /* CONFIG_MMU */
  * compat_do_execve() is mostly a copy of do_execve(), with the exception
  * that it processes 32 bit argv and envp pointers.
  */
-int compat_do_execve(char * filename,
+int __global compat_do_execve(char * filename,
 	compat_uptr_t __user *argv,
 	compat_uptr_t __user *envp,
 	struct pt_regs * regs)
diff --git a/fs/dcache.c b/fs/dcache.c
index 1b4a3a3..c9e20bb 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -63,7 +63,7 @@ static struct hlist_head *dentry_hashtab
 static LIST_HEAD(dentry_unused);
 
 /* Statistics gathering. */
-struct dentry_stat_t dentry_stat = {
+struct dentry_stat_t __global dentry_stat = {
 	.age_limit = 45,
 };
 
@@ -1120,7 +1120,7 @@ next:
  *
  * On hash failure or on lookup failure NULL is returned.
  */
-struct dentry *d_hash_and_lookup(struct dentry *dir, struct qstr *name)
+struct dentry * __global d_hash_and_lookup(struct dentry *dir, struct qstr *name)
 {
 	struct dentry *dentry = NULL;
 
@@ -1745,13 +1745,13 @@ EXPORT_SYMBOL(d_genocide);
 extern void bdev_cache_init(void);
 extern void chrdev_init(void);
 
-void __init vfs_caches_init_early(void)
+void __init __global vfs_caches_init_early(void)
 {
 	dcache_init_early();
 	inode_init_early();
 }
 
-void __init vfs_caches_init(unsigned long mempages)
+void __init __global vfs_caches_init(unsigned long mempages)
 {
 	unsigned long reserve;
 
diff --git a/fs/devpts/inode.c b/fs/devpts/inode.c
index f7aef5b..3b9f019 100644
--- a/fs/devpts/inode.c
+++ b/fs/devpts/inode.c
@@ -156,7 +156,7 @@ static struct dentry *get_node(int num)
 	return lookup_one_len(s, root, sprintf(s, "%d", num));
 }
 
-int devpts_pty_new(struct tty_struct *tty)
+int __global devpts_pty_new(struct tty_struct *tty)
 {
 	int number = tty->index;
 	struct tty_driver *driver = tty->driver;
@@ -188,7 +188,7 @@ int devpts_pty_new(struct tty_struct *tt
 	return 0;
 }
 
-struct tty_struct *devpts_get_tty(int number)
+struct tty_struct * __global devpts_get_tty(int number)
 {
 	struct dentry *dentry = get_node(number);
 	struct tty_struct *tty;
@@ -205,7 +205,7 @@ struct tty_struct *devpts_get_tty(int nu
 	return tty;
 }
 
-void devpts_pty_kill(int number)
+void __global devpts_pty_kill(int number)
 {
 	struct dentry *dentry = get_node(number);
 
diff --git a/fs/dnotify.c b/fs/dnotify.c
index f932591..c6c7ae3 100644
--- a/fs/dnotify.c
+++ b/fs/dnotify.c
@@ -21,7 +21,7 @@ #include <linux/init.h>
 #include <linux/spinlock.h>
 #include <linux/slab.h>
 
-int dir_notify_enable __read_mostly = 1;
+int dir_notify_enable __global __read_mostly = 1;
 
 static kmem_cache_t *dn_cache __read_mostly;
 
diff --git a/fs/drop_caches.c b/fs/drop_caches.c
index 4e47623..dfc591b 100644
--- a/fs/drop_caches.c
+++ b/fs/drop_caches.c
@@ -10,7 +10,7 @@ #include <linux/sysctl.h>
 #include <linux/gfp.h>
 
 /* A global variable is a bit ugly, but it keeps the code simple */
-int sysctl_drop_caches;
+int __global sysctl_drop_caches;
 
 static void drop_pagecache_sb(struct super_block *sb)
 {
@@ -54,7 +54,7 @@ void drop_slab(void)
 	} while (nr_objects > 10);
 }
 
-int drop_caches_sysctl_handler(ctl_table *table, int write,
+int __global drop_caches_sysctl_handler(ctl_table *table, int write,
 	struct file *file, void __user *buffer, size_t *length, loff_t *ppos)
 {
 	proc_dointvec_minmax(table, write, file, buffer, length, ppos);
diff --git a/fs/exec.c b/fs/exec.c
index 8344ba7..d14f122 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -57,8 +57,8 @@ #ifdef CONFIG_KMOD
 #include <linux/kmod.h>
 #endif
 
-int core_uses_pid;
-char core_pattern[65] = "core";
+int __global core_uses_pid;
+char __global core_pattern[65] = "core";
 int suid_dumpable = 0;
 
 EXPORT_SYMBOL(suid_dumpable);
@@ -805,7 +805,7 @@ static void flush_old_files(struct files
 	spin_unlock(&files->file_lock);
 }
 
-void get_task_comm(char *buf, struct task_struct *tsk)
+void __global get_task_comm(char *buf, struct task_struct *tsk)
 {
 	/* buf must be at least sizeof(tsk->comm) in size */
 	task_lock(tsk);
@@ -813,7 +813,7 @@ void get_task_comm(char *buf, struct tas
 	task_unlock(tsk);
 }
 
-void set_task_comm(struct task_struct *tsk, char *buf)
+void __global set_task_comm(struct task_struct *tsk, char *buf)
 {
 	task_lock(tsk);
 	strlcpy(tsk->comm, buf, sizeof(tsk->comm));
@@ -1131,7 +1131,7 @@ EXPORT_SYMBOL(search_binary_handler);
 /*
  * sys_execve() executes a new program.
  */
-int do_execve(char * filename,
+int __global do_execve(char * filename,
 	char __user *__user *argv,
 	char __user *__user *envp,
 	struct pt_regs * regs)
@@ -1465,7 +1465,7 @@ fail:
 	return core_waiters;
 }
 
-int do_coredump(long signr, int exit_code, struct pt_regs * regs)
+int __global do_coredump(long signr, int exit_code, struct pt_regs * regs)
 {
 	char corename[CORENAME_MAX_SIZE + 1];
 	struct mm_struct *mm = current->mm;
diff --git a/fs/fcntl.c b/fs/fcntl.c
index d35cbc6..5f1c0eb 100644
--- a/fs/fcntl.c
+++ b/fs/fcntl.c
@@ -23,7 +23,7 @@ #include <asm/poll.h>
 #include <asm/siginfo.h>
 #include <asm/uaccess.h>
 
-void fastcall set_close_on_exec(unsigned int fd, int flag)
+void fastcall __global set_close_on_exec(unsigned int fd, int flag)
 {
 	struct files_struct *files = current->files;
 	struct fdtable *fdt;
@@ -467,7 +467,7 @@ static void send_sigio_to_task(struct ta
 	}
 }
 
-void send_sigio(struct fown_struct *fown, int fd, int band)
+void __global send_sigio(struct fown_struct *fown, int fd, int band)
 {
 	struct task_struct *p;
 	int pid;
@@ -500,7 +500,7 @@ static void send_sigurg_to_task(struct t
 		group_send_sig_info(SIGURG, SEND_SIG_PRIV, p);
 }
 
-int send_sigurg(struct fown_struct *fown)
+int __global send_sigurg(struct fown_struct *fown)
 {
 	struct task_struct *p;
 	int pid, ret = 0;
diff --git a/fs/file.c b/fs/file.c
index b3c6b82..af0fff5 100644
--- a/fs/file.c
+++ b/fs/file.c
@@ -50,7 +50,7 @@ struct file ** alloc_fd_array(int num)
 	return new_fds;
 }
 
-void free_fd_array(struct file **array, int num)
+void __global free_fd_array(struct file **array, int num)
 {
 	int size = num * sizeof(struct file *);
 
@@ -154,7 +154,7 @@ static void free_fdtable_rcu(struct rcu_
 	}
 }
 
-void free_fdtable(struct fdtable *fdt)
+void __global free_fdtable(struct fdtable *fdt)
 {
 	if (fdt->free_files ||
 		fdt->max_fdset > EMBEDDED_FD_SET_SIZE ||
@@ -219,7 +219,7 @@ fd_set * alloc_fdset(int num)
 	return new_fdset;
 }
 
-void free_fdset(fd_set *array, int num)
+void __global free_fdset(fd_set *array, int num)
 {
 	if (num <= EMBEDDED_FD_SET_SIZE) /* Don't free an embedded fdset */
 		return;
@@ -336,7 +336,7 @@ out:
  * Return <0 on error; 0 nothing done; 1 files expanded, we may have blocked.
  * Should be called with the files->file_lock spinlock held for write.
  */
-int expand_files(struct files_struct *files, int nr)
+int __global expand_files(struct files_struct *files, int nr)
 {
 	int err, expand = 0;
 	struct fdtable *fdt;
diff --git a/fs/file_table.c b/fs/file_table.c
index 0131ba0..3e795f3 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -25,12 +25,12 @@ #include <linux/percpu_counter.h>
 #include <asm/atomic.h>
 
 /* sysctl tunables... */
-struct files_stat_struct files_stat = {
+struct files_stat_struct __global files_stat = {
 	.max_files = NR_FILE
 };
 
 /* public. Not pretty! */
-__cacheline_aligned_in_smp DEFINE_SPINLOCK(files_lock);
+__global __cacheline_aligned_in_smp DEFINE_SPINLOCK(files_lock);
 
 static struct percpu_counter nr_files __cacheline_aligned_in_smp;
 
@@ -67,14 +67,14 @@ EXPORT_SYMBOL_GPL(get_max_files);
  * Handle nr_files sysctl
  */
 #if defined(CONFIG_SYSCTL) && defined(CONFIG_PROC_FS)
-int proc_nr_files(ctl_table *table, int write, struct file *filp,
+int __global proc_nr_files(ctl_table *table, int write, struct file *filp,
                      void __user *buffer, size_t *lenp, loff_t *ppos)
 {
 	files_stat.nr_files = get_nr_files();
 	return proc_dointvec(table, write, filp, buffer, lenp, ppos);
 }
 #else
-int proc_nr_files(ctl_table *table, int write, struct file *filp,
+int __global proc_nr_files(ctl_table *table, int write, struct file *filp,
                      void __user *buffer, size_t *lenp, loff_t *ppos)
 {
 	return -ENOSYS;
@@ -210,7 +210,7 @@ EXPORT_SYMBOL(fget);
  * and a flag is returned to be passed to the corresponding fput_light().
  * There must not be a cloning between an fget_light/fput_light pair.
  */
-struct file fastcall *fget_light(unsigned int fd, int *fput_needed)
+struct file fastcall * __global fget_light(unsigned int fd, int *fput_needed)
 {
 	struct file *file;
 	struct files_struct *files = current->files;
@@ -235,7 +235,7 @@ struct file fastcall *fget_light(unsigne
 }
 
 
-void put_filp(struct file *file)
+void __global put_filp(struct file *file)
 {
 	if (atomic_dec_and_test(&file->f_count)) {
 		security_file_free(file);
@@ -244,7 +244,7 @@ void put_filp(struct file *file)
 	}
 }
 
-void file_move(struct file *file, struct list_head *list)
+void __global file_move(struct file *file, struct list_head *list)
 {
 	if (!list)
 		return;
@@ -253,7 +253,7 @@ void file_move(struct file *file, struct
 	file_list_unlock();
 }
 
-void file_kill(struct file *file)
+void __global file_kill(struct file *file)
 {
 	if (!list_empty(&file->f_u.fu_list)) {
 		file_list_lock();
diff --git a/fs/filesystems.c b/fs/filesystems.c
index 9f10728..2b338c6 100644
--- a/fs/filesystems.c
+++ b/fs/filesystems.c
@@ -198,7 +198,7 @@ asmlinkage long sys_sysfs(int option, un
 	return retval;
 }
 
-int get_filesystem_list(char * buf)
+int __global get_filesystem_list(char * buf)
 {
 	int len = 0;
 	struct file_system_type * tmp;
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index 892643d..c3a2e2b 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -409,7 +409,7 @@ sync_sb_inodes(struct super_block *sb, s
  * sync_sb_inodes will seekout the blockdev which matches `bdi'.  Maybe not
  * super-efficient but we're about to do a ton of I/O...
  */
-void
+void __global
 writeback_inodes(struct writeback_control *wbc)
 {
 	struct super_block *sb;
@@ -678,7 +678,7 @@ int writeback_acquire(struct backing_dev
  *
  * Determine whether there is writeback in progress against a backing device.
  */
-int writeback_in_progress(struct backing_dev_info *bdi)
+int __global writeback_in_progress(struct backing_dev_info *bdi)
 {
 	return test_bit(BDI_pdflush, &bdi->state);
 }
diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c
index c3920c9..5062ca1 100644
--- a/fs/hugetlbfs/inode.c
+++ b/fs/hugetlbfs/inode.c
@@ -44,7 +44,7 @@ static struct backing_dev_info hugetlbfs
 	.capabilities	= BDI_CAP_NO_ACCT_DIRTY | BDI_CAP_NO_WRITEBACK,
 };
 
-int sysctl_hugetlb_shm_group;
+int __global sysctl_hugetlb_shm_group;
 
 static void huge_pagevec_release(struct pagevec *pvec)
 {
@@ -562,7 +562,7 @@ static void init_once(void *foo, kmem_ca
 		inode_init_once(&ei->vfs_inode);
 }
 
-const struct file_operations hugetlbfs_file_operations = {
+const __global struct file_operations hugetlbfs_file_operations = {
 	.mmap			= hugetlbfs_file_mmap,
 	.fsync			= simple_sync_file,
 	.get_unmapped_area	= hugetlb_get_unmapped_area,
@@ -691,7 +691,7 @@ out_free:
 	return -ENOMEM;
 }
 
-int hugetlb_get_quota(struct address_space *mapping)
+int __global hugetlb_get_quota(struct address_space *mapping)
 {
 	int ret = 0;
 	struct hugetlbfs_sb_info *sbinfo = HUGETLBFS_SB(mapping->host->i_sb);
@@ -708,7 +708,7 @@ int hugetlb_get_quota(struct address_spa
 	return ret;
 }
 
-void hugetlb_put_quota(struct address_space *mapping)
+void __global hugetlb_put_quota(struct address_space *mapping)
 {
 	struct hugetlbfs_sb_info *sbinfo = HUGETLBFS_SB(mapping->host->i_sb);
 
@@ -740,7 +740,7 @@ static int can_do_hugetlb_shm(void)
 			can_do_mlock());
 }
 
-struct file *hugetlb_zero_setup(size_t size)
+struct file * __global hugetlb_zero_setup(size_t size)
 {
 	int error = -ENOMEM;
 	struct file *file;
diff --git a/fs/inode.c b/fs/inode.c
index 0bf9f04..0ed00e2 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -71,7 +71,7 @@ static unsigned int i_hash_shift __read_
  */
 
 LIST_HEAD(inode_in_use);
-LIST_HEAD(inode_unused);
+__global LIST_HEAD(inode_unused);
 static struct hlist_head *inode_hashtable __read_mostly;
 
 /*
@@ -80,7 +80,7 @@ static struct hlist_head *inode_hashtabl
  * NOTE! You also have to own the lock if you change
  * the i_state of an inode while it is in use..
  */
-DEFINE_SPINLOCK(inode_lock);
+__global DEFINE_SPINLOCK(inode_lock);
 
 /*
  * iprune_mutex provides exclusion between the kswapd or try_to_free_pages
@@ -95,7 +95,7 @@ static DEFINE_MUTEX(iprune_mutex);
 /*
  * Statistics gathering..
  */
-struct inodes_stat_t inodes_stat;
+struct inodes_stat_t __global inodes_stat;
 
 static kmem_cache_t * inode_cachep __read_mostly;
 
@@ -169,7 +169,7 @@ #endif
 	return inode;
 }
 
-void destroy_inode(struct inode *inode) 
+void __global destroy_inode(struct inode *inode) 
 {
 	BUG_ON(inode_has_buffers(inode));
 	security_inode_free(inode);
diff --git a/fs/inotify_user.c b/fs/inotify_user.c
index 017cb0f..a7997c2 100644
--- a/fs/inotify_user.c
+++ b/fs/inotify_user.c
@@ -112,7 +112,7 @@ #include <linux/sysctl.h>
 
 static int zero;
 
-ctl_table inotify_table[] = {
+ctl_table __global inotify_table[] = {
 	{
 		.ctl_name	= INOTIFY_MAX_USER_INSTANCES,
 		.procname	= "max_user_instances",
diff --git a/fs/jffs2/Makefile b/fs/jffs2/Makefile
diff --git a/fs/jffs2/read.c b/fs/jffs2/read.c
diff --git a/fs/locks.c b/fs/locks.c
index b0b41a6..e4ef939 100644
--- a/fs/locks.c
+++ b/fs/locks.c
@@ -133,8 +133,8 @@ #define IS_POSIX(fl)	(fl->fl_flags & FL_
 #define IS_FLOCK(fl)	(fl->fl_flags & FL_FLOCK)
 #define IS_LEASE(fl)	(fl->fl_flags & FL_LEASE)
 
-int leases_enable = 1;
-int lease_break_time = 45;
+int __global leases_enable = 1;
+int __global lease_break_time = 45;
 
 #define for_each_lock(inode, lockp) \
 	for (lockp = &inode->i_flock; *lockp != NULL; lockp = &(*lockp)->fl_next)
@@ -1072,7 +1072,7 @@ EXPORT_SYMBOL(posix_lock_file_wait);
  * Searches the inode's list of locks to find any POSIX locks which conflict.
  * This function is called from locks_verify_locked() only.
  */
-int locks_mandatory_locked(struct inode *inode)
+int __global locks_mandatory_locked(struct inode *inode)
 {
 	fl_owner_t owner = current->files;
 	struct file_lock *fl;
@@ -2110,7 +2110,7 @@ static void move_lock_status(char **p, o
  *	@length: how much to read
  */
 
-int get_locks_status(char *buffer, char **start, off_t offset, int length)
+int __global get_locks_status(char *buffer, char **start, off_t offset, int length)
 {
 	struct list_head *tmp;
 	char *q = buffer;
diff --git a/fs/namei.c b/fs/namei.c
index 55a1312..dee817e 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -322,7 +322,7 @@ int get_write_access(struct inode * inod
 	return 0;
 }
 
-int deny_write_access(struct file * file)
+int __global deny_write_access(struct file * file)
 {
 	struct inode *inode = file->f_dentry->d_inode;
 
@@ -1040,7 +1040,7 @@ static int __emul_lookup_dentry(const ch
 	return 1;
 }
 
-void set_fs_altroot(void)
+void __global set_fs_altroot(void)
 {
 	char *emul = __emul_prefix();
 	struct nameidata nd;
diff --git a/fs/namespace.c b/fs/namespace.c
index fa7ed6a..eed31ac 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -38,7 +38,7 @@ static inline int sysfs_init(void)
 #endif
 
 /* spinlock for vfsmount related operations, inplace of dcache_lock */
-__cacheline_aligned_in_smp DEFINE_SPINLOCK(vfsmount_lock);
+__global __cacheline_aligned_in_smp DEFINE_SPINLOCK(vfsmount_lock);
 
 static int event;
 
@@ -400,7 +400,7 @@ static int show_vfsmnt(struct seq_file *
 	return err;
 }
 
-struct seq_operations mounts_op = {
+struct seq_operations __global mounts_op = {
 	.start	= m_start,
 	.next	= m_next,
 	.stop	= m_stop,
@@ -438,7 +438,7 @@ static int show_vfsstat(struct seq_file 
 	return err;
 }
 
-struct seq_operations mountstats_op = {
+struct seq_operations __global mountstats_op = {
 	.start	= m_start,
 	.next	= m_next,
 	.stop	= m_stop,
@@ -1447,7 +1447,7 @@ dput_out:
  * Allocate a new namespace structure and populate it with contents
  * copied from the namespace of the passed in task structure.
  */
-struct namespace *dup_namespace(struct task_struct *tsk, struct fs_struct *fs)
+struct namespace * __global dup_namespace(struct task_struct *tsk, struct fs_struct *fs)
 {
 	struct namespace *namespace = tsk->namespace;
 	struct namespace *new_ns;
@@ -1514,7 +1514,7 @@ struct namespace *dup_namespace(struct t
 	return new_ns;
 }
 
-int copy_namespace(int flags, struct task_struct *tsk)
+int __global copy_namespace(int flags, struct task_struct *tsk)
 {
 	struct namespace *namespace = tsk->namespace;
 	struct namespace *new_ns;
@@ -1867,7 +1867,7 @@ void __init mnt_init(unsigned long mempa
 	init_mount_tree();
 }
 
-void __put_namespace(struct namespace *namespace)
+void __global __put_namespace(struct namespace *namespace)
 {
 	struct vfsmount *root = namespace->root;
 	LIST_HEAD(umount_list);
diff --git a/fs/pipe.c b/fs/pipe.c
index 2035257..eaa60a7 100644
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -822,7 +822,7 @@ struct pipe_inode_info * alloc_pipe_info
 	return pipe;
 }
 
-void __free_pipe_info(struct pipe_inode_info *pipe)
+void __global __free_pipe_info(struct pipe_inode_info *pipe)
 {
 	int i;
 
@@ -890,7 +890,7 @@ fail_inode:
 	return NULL;
 }
 
-int do_pipe(int *fd)
+int __global do_pipe(int *fd)
 {
 	struct qstr this;
 	char name[32];
diff --git a/fs/proc/base.c b/fs/proc/base.c
index fe8d55f..687c944 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -1975,7 +1975,7 @@ static struct inode_operations proc_self
  *       that no dcache entries will exist at process exit time it
  *       just makes it very unlikely that any will persist.
  */
-void proc_flush_task(struct task_struct *task)
+void __global proc_flush_task(struct task_struct *task)
 {
 	struct dentry *dentry, *leader, *dir;
 	char buf[PROC_NUMBUF];
diff --git a/fs/proc/generic.c b/fs/proc/generic.c
index 4ba0300..dbfdf4a 100644
--- a/fs/proc/generic.c
+++ b/fs/proc/generic.c
@@ -32,7 +32,7 @@ static loff_t proc_file_lseek(struct fil
 
 DEFINE_SPINLOCK(proc_subdir_lock);
 
-int proc_match(int len, const char *name, struct proc_dir_entry *de)
+int __global proc_match(int len, const char *name, struct proc_dir_entry *de)
 {
 	if (de->namelen != len)
 		return 0;
diff --git a/fs/proc/kcore.c b/fs/proc/kcore.c
index 6a984f6..7e4296c 100644
--- a/fs/proc/kcore.c
+++ b/fs/proc/kcore.c
@@ -54,7 +54,7 @@ struct memelfnote
 static struct kcore_list *kclist;
 static DEFINE_RWLOCK(kclist_lock);
 
-void
+void __global
 kclist_add(struct kcore_list *new, void *addr, size_t size)
 {
 	new->addr = (unsigned long)addr;
diff --git a/fs/proc/proc_devtree.c b/fs/proc/proc_devtree.c
index abdf068..cb53138 100644
--- a/fs/proc/proc_devtree.c
+++ b/fs/proc/proc_devtree.c
@@ -77,18 +77,18 @@ __proc_device_tree_add_prop(struct proc_
 }
 
 
-void proc_device_tree_add_prop(struct proc_dir_entry *pde, struct property *prop)
+void  __global proc_device_tree_add_prop(struct proc_dir_entry *pde, struct property *prop)
 {
 	__proc_device_tree_add_prop(pde, prop, prop->name);
 }
 
-void proc_device_tree_remove_prop(struct proc_dir_entry *pde,
+void __global proc_device_tree_remove_prop(struct proc_dir_entry *pde,
 				  struct property *prop)
 {
 	remove_proc_entry(prop->name, pde);
 }
 
-void proc_device_tree_update_prop(struct proc_dir_entry *pde,
+void __global proc_device_tree_update_prop(struct proc_dir_entry *pde,
 				  struct property *newprop,
 				  struct property *oldprop)
 {
@@ -172,8 +172,8 @@ retry:
 /*
  * Process a node, adding entries for its children and its properties.
  */
-void proc_device_tree_add_node(struct device_node *np,
-			       struct proc_dir_entry *de)
+void __global proc_device_tree_add_node(struct device_node *np,
+					struct proc_dir_entry *de)
 {
 	struct property *pp;
 	struct proc_dir_entry *ent;
diff --git a/fs/proc/proc_tty.c b/fs/proc/proc_tty.c
index 15c4455..c874f53 100644
--- a/fs/proc/proc_tty.c
+++ b/fs/proc/proc_tty.c
@@ -182,7 +182,7 @@ static int tty_ldiscs_read_proc(char *pa
  * This function is called by tty_register_driver() to handle
  * registering the driver's /proc handler into /proc/tty/driver/<foo>
  */
-void proc_tty_register_driver(struct tty_driver *driver)
+void __global proc_tty_register_driver(struct tty_driver *driver)
 {
 	struct proc_dir_entry *ent;
 		
@@ -205,7 +205,7 @@ void proc_tty_register_driver(struct tty
 /*
  * This function is called by tty_unregister_driver()
  */
-void proc_tty_unregister_driver(struct tty_driver *driver)
+void __global proc_tty_unregister_driver(struct tty_driver *driver)
 {
 	struct proc_dir_entry *ent;
 
diff --git a/fs/proc/root.c b/fs/proc/root.c
index 8901c65..7b99f88 100644
--- a/fs/proc/root.c
+++ b/fs/proc/root.c
@@ -22,7 +22,7 @@ #include "internal.h"
 struct proc_dir_entry *proc_net, *proc_net_stat, *proc_bus, *proc_root_fs, *proc_root_driver;
 
 #ifdef CONFIG_SYSCTL
-struct proc_dir_entry *proc_sys_root;
+struct proc_dir_entry * __global proc_sys_root;
 #endif
 
 static int proc_get_sb(struct file_system_type *fs_type,
@@ -37,7 +37,7 @@ static struct file_system_type proc_fs_t
 	.kill_sb	= kill_anon_super,
 };
 
-void __init proc_root_init(void)
+void __init __global proc_root_init(void)
 {
 	int err = proc_init_inodecache();
 	if (err)
diff --git a/fs/ramfs/inode.c b/fs/ramfs/inode.c
index b967733..9bd963d 100644
--- a/fs/ramfs/inode.c
+++ b/fs/ramfs/inode.c
@@ -222,7 +222,7 @@ static void __exit exit_ramfs_fs(void)
 module_init(init_ramfs_fs)
 module_exit(exit_ramfs_fs)
 
-int __init init_rootfs(void)
+int __init __global init_rootfs(void)
 {
 	return register_filesystem(&rootfs_fs_type);
 }
diff --git a/fs/splice.c b/fs/splice.c
index 684bca3..78500fc 100644
--- a/fs/splice.c
+++ b/fs/splice.c
@@ -707,7 +707,7 @@ out_nomem:
  * key here is the 'actor' worker passed in that actually moves the data
  * to the wanted destination. See pipe_to_file/pipe_to_sendpage above.
  */
-ssize_t splice_from_pipe(struct pipe_inode_info *pipe, struct file *out,
+ssize_t __global splice_from_pipe(struct pipe_inode_info *pipe, struct file *out,
 			 loff_t *ppos, size_t len, unsigned int flags,
 			 splice_actor *actor)
 {
diff --git a/fs/super.c b/fs/super.c
index 6d4e817..ffa2854 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -45,7 +45,7 @@ void put_filesystem(struct file_system_t
 struct file_system_type *get_fs_type(const char *name);
 
 LIST_HEAD(super_blocks);
-DEFINE_SPINLOCK(sb_lock);
+__global DEFINE_SPINLOCK(sb_lock);
 
 /**
  *	alloc_super	-	create new superblock
@@ -347,7 +347,7 @@ static inline void write_super(struct su
  * hold up the sync while mounting a device. (The newly
  * mounted device won't need syncing.)
  */
-void sync_supers(void)
+void __global sync_supers(void)
 {
 	struct super_block *sb;
 
@@ -592,7 +592,7 @@ static void do_emergency_remount(unsigne
 	printk("Emergency Remount complete\n");
 }
 
-void emergency_remount(void)
+void __global emergency_remount(void)
 {
 	pdflush_operation(do_emergency_remount, 0);
 }
@@ -646,7 +646,7 @@ void kill_anon_super(struct super_block 
 
 EXPORT_SYMBOL(kill_anon_super);
 
-void __init unnamed_dev_init(void)
+void __init __global unnamed_dev_init(void)
 {
 	idr_init(&unnamed_dev_idr);
 }
diff --git a/init/calibrate.c b/init/calibrate.c
index d206c75..938f200 100644
--- a/init/calibrate.c
+++ b/init/calibrate.c
@@ -112,7 +112,7 @@ #endif
  */
 #define LPS_PREC 8
 
-void __devinit calibrate_delay(void)
+void __devinit __global calibrate_delay(void)
 {
 	unsigned long ticks, loopbit;
 	int lps_precision = LPS_PREC;
diff --git a/init/do_mounts.c b/init/do_mounts.c
index 94aeec7..621bf15 100644
--- a/init/do_mounts.c
+++ b/init/do_mounts.c
@@ -17,13 +17,13 @@ #include "do_mounts.h"
 
 extern int get_filesystem_list(char * buf);
 
-int __initdata rd_doload;	/* 1 = load RAM disk, 0 = don't load */
+int __initdata __global rd_doload;	/* 1 = load RAM disk, 0 = don't load */
 
-int root_mountflags = MS_RDONLY | MS_SILENT;
+int __global root_mountflags = MS_RDONLY | MS_SILENT;
 char * __initdata root_device_name;
 static char __initdata saved_root_name[64];
 
-dev_t ROOT_DEV;
+dev_t __global ROOT_DEV;
 
 static int __init load_ramdisk(char *str)
 {
@@ -133,7 +133,7 @@ fail:
  *	is mounted on rootfs /sys.
  */
 
-dev_t name_to_dev_t(char *name)
+dev_t __global name_to_dev_t(char *name)
 {
 	char s[32];
 	char *p;
@@ -393,7 +393,7 @@ #endif
 /*
  * Prepare the namespace - decide what/where to mount, load ramdisks, etc.
  */
-void __init prepare_namespace(void)
+void __init __global prepare_namespace(void)
 {
 	int is_floppy;
 
@@ -430,4 +430,3 @@ out:
 	sys_chroot(".");
 	security_sb_post_mountroot();
 }
-
diff --git a/init/do_mounts_initrd.c b/init/do_mounts_initrd.c
index a06f037..6386990 100644
--- a/init/do_mounts_initrd.c
+++ b/init/do_mounts_initrd.c
@@ -10,9 +10,10 @@ #include <linux/sched.h>
 
 #include "do_mounts.h"
 
-unsigned long initrd_start, initrd_end;
-int initrd_below_start_ok;
-unsigned int real_root_dev;	/* do_proc_dointvec cannot handle kdev_t */
+unsigned long __global initrd_start;
+unsigned long __global initrd_end;
+int __global initrd_below_start_ok;
+unsigned int __global real_root_dev;	/* do_proc_dointvec cannot handle kdev_t */
 static int __initdata old_fd, root_fd;
 static int __initdata mount_initrd = 1;
 
diff --git a/init/do_mounts_rd.c b/init/do_mounts_rd.c
index ed652f4..4ef0705 100644
--- a/init/do_mounts_rd.c
+++ b/init/do_mounts_rd.c
@@ -12,7 +12,7 @@ #include "do_mounts.h"
 
 #define BUILD_CRAMDISK
 
-int __initdata rd_prompt = 1;/* 1 = prompt for RAM disk, 0 = don't prompt */
+int __initdata __global rd_prompt = 1;/* 1 = prompt for RAM disk, 0 = don't prompt */
 
 static int __init prompt_ramdisk(char *str)
 {
@@ -21,7 +21,7 @@ static int __init prompt_ramdisk(char *s
 }
 __setup("prompt_ramdisk=", prompt_ramdisk);
 
-int __initdata rd_image_start;		/* starting block # of image */
+int __initdata __global rd_image_start;		/* starting block # of image */
 
 static int __init ramdisk_start_setup(char *str)
 {
diff --git a/init/main.c b/init/main.c
index 8651a72..735b3f3 100644
--- a/init/main.c
+++ b/init/main.c
@@ -104,7 +104,7 @@ #ifdef CONFIG_TC
 extern void tc_init(void);
 #endif
 
-enum system_states system_state;
+enum system_states __global system_state;
 EXPORT_SYMBOL(system_state);
 
 /*
@@ -115,11 +115,11 @@ #define MAX_INIT_ENVS CONFIG_INIT_ENV_AR
 
 extern void time_init(void);
 /* Default late time init is NULL. archs can override this later. */
-void (*late_time_init)(void);
+void __global (*late_time_init)(void);
 extern void softirq_init(void);
 
 /* Untouched command line (eg. for /proc) saved by arch-specific code. */
-char saved_command_line[COMMAND_LINE_SIZE];
+char __global saved_command_line[COMMAND_LINE_SIZE];
 
 static char *execute_command;
 static char *ramdisk_execute_command;
@@ -188,7 +188,7 @@ static int __init obsolete_checksetup(ch
  * This should be approx 2 Bo*oMips to start (note initial shift), and will
  * still work even if initially too large, it will just take slightly longer
  */
-unsigned long loops_per_jiffy = (1<<12);
+unsigned long __global loops_per_jiffy = (1<<12);
 
 EXPORT_SYMBOL(loops_per_jiffy);
 
@@ -422,7 +422,7 @@ static int __init do_early_param(char *p
 }
 
 /* Arch code calls this early on, or if not, just before other parsing. */
-void __init parse_early_param(void)
+void __init __global parse_early_param(void)
 {
 	static __initdata int done = 0;
 	static __initdata char tmp_cmdline[COMMAND_LINE_SIZE];
@@ -453,7 +453,7 @@ void __init __attribute__((weak)) smp_se
 {
 }
 
-asmlinkage void __init start_kernel(void)
+asmlinkage void __init __global start_kernel(void)
 {
 	char * command_line;
 	extern struct kernel_param __start___param[], __stop___param[];
@@ -596,7 +596,7 @@ static int __init initcall_debug_setup(c
 }
 __setup("initcall_debug", initcall_debug_setup);
 
-struct task_struct *child_reaper = &init_task;
+struct task_struct * __global child_reaper = &init_task;
 
 extern initcall_t __initcall_start[], __initcall_end[];
 
diff --git a/ipc/compat.c b/ipc/compat.c
index 4d20cfd..3ff09e5 100644
--- a/ipc/compat.c
+++ b/ipc/compat.c
@@ -234,7 +234,7 @@ static inline int put_compat_semid_ds(st
 	return err;
 }
 
-long compat_sys_semctl(int first, int second, int third, void __user *uptr)
+long __global compat_sys_semctl(int first, int second, int third, void __user *uptr)
 {
 	union semun fourth;
 	u32 pad;
@@ -305,7 +305,7 @@ long compat_sys_semctl(int first, int se
 	return err;
 }
 
-long compat_sys_msgsnd(int first, int second, int third, void __user *uptr)
+long __global compat_sys_msgsnd(int first, int second, int third, void __user *uptr)
 {
 	struct msgbuf __user *p;
 	struct compat_msgbuf __user *up = uptr;
@@ -325,7 +325,7 @@ long compat_sys_msgsnd(int first, int se
 	return sys_msgsnd(first, p, second, third);
 }
 
-long compat_sys_msgrcv(int first, int second, int msgtyp, int third,
+long __global compat_sys_msgrcv(int first, int second, int msgtyp, int third,
 			   int version, void __user *uptr)
 {
 	struct msgbuf __user *p;
@@ -424,7 +424,7 @@ static inline int put_compat_msqid_ds(st
 	return err;
 }
 
-long compat_sys_msgctl(int first, int second, void __user *uptr)
+long __global compat_sys_msgctl(int first, int second, void __user *uptr)
 {
 	int err, err2;
 	struct msqid64_ds m64;
@@ -476,7 +476,7 @@ long compat_sys_msgctl(int first, int se
 	return err;
 }
 
-long compat_sys_shmat(int first, int second, compat_uptr_t third, int version,
+long __global compat_sys_shmat(int first, int second, compat_uptr_t third, int version,
 			void __user *uptr)
 {
 	int err;
@@ -592,7 +592,7 @@ static inline int put_compat_shm_info(st
 	return err;
 }
 
-long compat_sys_shmctl(int first, int second, void __user *uptr)
+long __global compat_sys_shmctl(int first, int second, void __user *uptr)
 {
 	void __user *p;
 	struct shmid64_ds s64;
@@ -671,7 +671,7 @@ long compat_sys_shmctl(int first, int se
 	return err;
 }
 
-long compat_sys_semtimedop(int semid, struct sembuf __user *tsems,
+long __global compat_sys_semtimedop(int semid, struct sembuf __user *tsems,
 		unsigned nsops, const struct compat_timespec __user *timeout)
 {
 	struct timespec __user *ts64 = NULL;
diff --git a/ipc/msg.c b/ipc/msg.c
index 2b4fccf..eeb2030 100644
--- a/ipc/msg.c
+++ b/ipc/msg.c
@@ -37,9 +37,9 @@ #include <asm/uaccess.h>
 #include "util.h"
 
 /* sysctl: */
-int msg_ctlmax = MSGMAX;
-int msg_ctlmnb = MSGMNB;
-int msg_ctlmni = MSGMNI;
+int __global msg_ctlmax = MSGMAX;
+int __global msg_ctlmnb = MSGMNB;
+int __global msg_ctlmni = MSGMNI;
 
 /*
  * one msg_receiver structure for each sleeping receiver:
diff --git a/ipc/sem.c b/ipc/sem.c
index 6013c75..6b744e1 100644
--- a/ipc/sem.c
+++ b/ipc/sem.c
@@ -110,7 +110,7 @@ #define SEMOPM_FAST	64  /* ~ 372 bytes o
  *	
  */
 
-int sem_ctls[4] = {SEMMSL, SEMMNS, SEMOPM, SEMMNI};
+int __global sem_ctls[4] = {SEMMSL, SEMMNS, SEMOPM, SEMMNI};
 #define sc_semmsl	(sem_ctls[0])
 #define sc_semmns	(sem_ctls[1])
 #define sc_semopm	(sem_ctls[2])
@@ -1234,7 +1234,7 @@ asmlinkage long sys_semop (int semid, st
  * because of the reasoning in the comment above unlock_semundo.
  */
 
-int copy_semundo(unsigned long clone_flags, struct task_struct *tsk)
+int __global copy_semundo(unsigned long clone_flags, struct task_struct *tsk)
 {
 	struct sem_undo_list *undo_list;
 	int error;
@@ -1263,7 +1263,7 @@ int copy_semundo(unsigned long clone_fla
  * The current implementation does not do so. The POSIX standard
  * and SVID should be consulted to determine what behavior is mandated.
  */
-void exit_sem(struct task_struct *tsk)
+void __global exit_sem(struct task_struct *tsk)
 {
 	struct sem_undo_list *undo_list;
 	struct sem_undo *u, **up;
diff --git a/ipc/shm.c b/ipc/shm.c
index 940b0c9..4d1dd3a 100644
--- a/ipc/shm.c
+++ b/ipc/shm.c
@@ -55,9 +55,9 @@ #ifdef CONFIG_PROC_FS
 static int sysvipc_shm_proc_show(struct seq_file *s, void *it);
 #endif
 
-size_t	shm_ctlmax = SHMMAX;
-size_t 	shm_ctlall = SHMALL;
-int 	shm_ctlmni = SHMMNI;
+size_t	__global shm_ctlmax = SHMMAX;
+size_t 	__global shm_ctlall = SHMALL;
+int 	__global shm_ctlmni = SHMMNI;
 
 static int shm_tot; /* total number of shared memory pages */
 
@@ -688,7 +688,7 @@ out:
  * "raddr" thing points to kernel space, and there has to be a wrapper around
  * this.
  */
-long do_shmat(int shmid, char __user *shmaddr, int shmflg, ulong *raddr)
+long __global do_shmat(int shmid, char __user *shmaddr, int shmflg, ulong *raddr)
 {
 	struct shmid_kernel *shp;
 	unsigned long addr;
diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c
index 6991bec..59a13da 100644
--- a/kernel/sys_ni.c
+++ b/kernel/sys_ni.c
@@ -1,4 +1,5 @@
 
+#include <linux/compiler.h>
 #include <linux/linkage.h>
 #include <linux/errno.h>
 
diff --git a/net/core/dev.c b/net/core/dev.c
index d95e262..1c08754 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -405,7 +405,7 @@ int netdev_boot_setup_check(struct net_d
  *	later in the device probing.
  *	Returns 0 if no settings found.
  */
-unsigned long netdev_boot_base(const char *prefix, int unit)
+unsigned long __global netdev_boot_base(const char *prefix, int unit)
 {
 	const struct netdev_boot_setup *s = dev_boot_setup;
 	char name[IFNAMSIZ];
@@ -429,7 +429,7 @@ unsigned long netdev_boot_base(const cha
 /*
  * Saves at boot time configured settings for any netdevice.
  */
-int __init netdev_boot_setup(char *str)
+int __init __global netdev_boot_setup(char *str)
 {
 	int ints[5];
 	struct ifmap map;
@@ -1330,7 +1330,7 @@ static int dev_gso_segment(struct sk_buf
 	return 0;
 }
 
-int dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
+int __global dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
 {
 	if (likely(!skb->next)) {
 		if (netdev_nit)
@@ -1538,7 +1538,7 @@ int netdev_max_backlog = 1000;
 int netdev_budget = 300;
 int weight_p = 64;            /* old backlog weight */
 
-DEFINE_PER_CPU(struct netif_rx_stats, netdev_rx_stat) = { 0, };
+__global DEFINE_PER_CPU(struct netif_rx_stats, netdev_rx_stat) = { 0, };
 
 /**
  *	netif_rx	-	post buffer to the network code
@@ -2646,7 +2645,7 @@ static int dev_ifsioc(struct ifreq *ifr,
  *	positive or a negative errno code on error.
  */
 
-int dev_ioctl(unsigned int cmd, void __user *arg)
+int __global dev_ioctl(unsigned int cmd, void __user *arg)
 {
 	struct ifreq ifr;
 	int ret;
diff --git a/net/core/dv.c b/net/core/dv.c
index 29ee77f..6569d40 100644
--- a/net/core/dv.c
+++ b/net/core/dv.c
@@ -186,7 +186,7 @@ #else
 #define	DVDBG(a)
 #endif
 
-int divert_ioctl(unsigned int cmd, struct divert_cf __user *arg)
+int __global divert_ioctl(unsigned int cmd, struct divert_cf __user *arg)
 {
 	struct divert_cf	div_cf;
 	struct divert_blk	*div_blk;
@@ -434,7 +434,6 @@ int divert_ioctl(unsigned int cmd, struc
 	return 0;
 }
 
-
 /*
  * Check if packet should have its dest mac address set to the box itself
  * for diversion
diff --git a/net/core/flow.c b/net/core/flow.c
index 2191af5..fe4240b 100644
--- a/net/core/flow.c
+++ b/net/core/flow.c
@@ -285,7 +285,7 @@ static void flow_cache_flush_per_cpu(voi
 	tasklet_schedule(tasklet);
 }
 
-void flow_cache_flush(void)
+void __global flow_cache_flush(void)
 {
 	struct flow_flush_info info;
 	static DEFINE_MUTEX(flow_flush_sem);
diff --git a/net/core/iovec.c b/net/core/iovec.c
index 65e4b56..e562e7d 100644
--- a/net/core/iovec.c
+++ b/net/core/iovec.c
@@ -37,7 +37,8 @@ #include <net/sock.h>
  *	in any case.
  */
 
-int verify_iovec(struct msghdr *m, struct iovec *iov, char *address, int mode)
+int __global verify_iovec(struct msghdr *m, struct iovec *iov, char *address,
+			  int mode)
 {
 	int size, err, ct;
 	
diff --git a/net/core/neighbour.c b/net/core/neighbour.c
index 5130d2e..1afcc7b 100644
--- a/net/core/neighbour.c
+++ b/net/core/neighbour.c
@@ -499,8 +499,8 @@ out:
 }
 
 
-int pneigh_delete(struct neigh_table *tbl, const void *pkey,
-		  struct net_device *dev)
+int __global pneigh_delete(struct neigh_table *tbl, const void *pkey,
+			   struct net_device *dev)
 {
 	struct pneigh_entry *n, **np;
 	int key_len = tbl->key_len;
diff --git a/net/core/request_sock.c b/net/core/request_sock.c
index 79ebd75..8a9c3aa 100644
--- a/net/core/request_sock.c
+++ b/net/core/request_sock.c
@@ -31,7 +31,7 @@ #include <net/request_sock.h>
  * (<=32Mb of memory) and to 1024 on normal or better ones (>=256Mb).
  * Further increasing requires to change hash table size.
  */
-int sysctl_max_syn_backlog = 256;
+int __global sysctl_max_syn_backlog = 256;
 
 int reqsk_queue_alloc(struct request_sock_queue *queue,
 		      const int nr_table_entries)
diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
index 20e5bb7..0c35f52 100644
--- a/net/core/rtnetlink.c
+++ b/net/core/rtnetlink.c
@@ -155,7 +155,7 @@ size_t rtattr_strlcpy(char *dest, const 
 	return ret;
 }
 
-int rtnetlink_send(struct sk_buff *skb, u32 pid, unsigned group, int echo)
+int __global rtnetlink_send(struct sk_buff *skb, u32 pid, unsigned group, int echo)
 {
 	int err = 0;
 
@@ -777,7 +777,7 @@ static struct notifier_block rtnetlink_d
 	.notifier_call	= rtnetlink_event,
 };
 
-void __init rtnetlink_init(void)
+void __init __global rtnetlink_init(void)
 {
 	int i;
 
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 022d889..7502897 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -2039,7 +2039,7 @@ err:
 
 EXPORT_SYMBOL_GPL(skb_segment);
 
-void __init skb_init(void)
+void __init __global skb_init(void)
 {
 	skbuff_head_cache = kmem_cache_create("skbuff_head_cache",
 					      sizeof(struct sk_buff),
diff --git a/net/core/sock.c b/net/core/sock.c
index 51fcfbc..96db5f9 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -636,8 +636,8 @@ #endif
 }
 
 
-int sock_getsockopt(struct socket *sock, int level, int optname,
-		    char __user *optval, int __user *optlen)
+int __global sock_getsockopt(struct socket *sock, int level, int optname,
+			     char __user *optval, int __user *optlen)
 {
 	struct sock *sk = sock->sk;
 	
@@ -983,7 +983,7 @@ out:
 
 EXPORT_SYMBOL_GPL(sk_clone);
 
-void __init sk_init(void)
+void __init __global sk_init(void)
 {
 	if (num_physpages <= 4096) {
 		sysctl_wmem_max = 32767;
@@ -1065,8 +1065,8 @@ struct sk_buff *sock_wmalloc(struct sock
 /*
  * Allocate a skb from the socket's receive buffer.
  */ 
-struct sk_buff *sock_rmalloc(struct sock *sk, unsigned long size, int force,
-			     gfp_t priority)
+struct sk_buff * __global sock_rmalloc(struct sock *sk, unsigned long size,
+				       int force, gfp_t priority)
 {
 	if (force || atomic_read(&sk->sk_rmem_alloc) < sk->sk_rcvbuf) {
 		struct sk_buff *skb = alloc_skb(size, priority);
diff --git a/net/core/sysctl_net_core.c b/net/core/sysctl_net_core.c
index 0253413..1316556 100644
--- a/net/core/sysctl_net_core.c
+++ b/net/core/sysctl_net_core.c
@@ -30,7 +30,7 @@ extern u32 sysctl_xfrm_aevent_etime;
 extern u32 sysctl_xfrm_aevent_rseqth;
 #endif
 
-ctl_table core_table[] = {
+ctl_table __global core_table[] = {
 #ifdef CONFIG_NET
 	{
 		.ctl_name	= NET_CORE_WMEM_MAX,
diff --git a/net/core/user_dma.c b/net/core/user_dma.c
index 248a6b6..676efe0 100644
--- a/net/core/user_dma.c
+++ b/net/core/user_dma.c
@@ -33,7 +33,7 @@ #include <net/netdma.h>
 
 #define NET_DMA_DEFAULT_COPYBREAK 4096
 
-int sysctl_tcp_dma_copybreak = NET_DMA_DEFAULT_COPYBREAK;
+int __global sysctl_tcp_dma_copybreak = NET_DMA_DEFAULT_COPYBREAK;
 
 /**
  *	dma_skb_copy_datagram_iovec - Copy a datagram to an iovec.
@@ -45,7 +45,7 @@ int sysctl_tcp_dma_copybreak = NET_DMA_D
  *
  *	Note: the iovec is modified during the copy.
  */
-int dma_skb_copy_datagram_iovec(struct dma_chan *chan,
+int __global dma_skb_copy_datagram_iovec(struct dma_chan *chan,
 			struct sk_buff *skb, int offset, struct iovec *to,
 			size_t len, struct dma_pinned_list *pinned_list)
 {
diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c
index 95fac55..944595e 100644
--- a/net/ipv4/inet_hashtables.c
+++ b/net/ipv4/inet_hashtables.c
@@ -27,7 +27,7 @@ #include <net/ip.h>
  * Allocate and initialize a new local port bind bucket.
  * The bindhash mutex for snum's hash chain must be held here.
  */
-struct inet_bind_bucket *inet_bind_bucket_create(kmem_cache_t *cachep,
+struct inet_bind_bucket * __global inet_bind_bucket_create(kmem_cache_t *cachep,
 						 struct inet_bind_hashbucket *head,
 						 const unsigned short snum)
 {
@@ -53,7 +53,7 @@ void inet_bind_bucket_destroy(kmem_cache
 	}
 }
 
-void inet_bind_hash(struct sock *sk, struct inet_bind_bucket *tb,
+void __global inet_bind_hash(struct sock *sk, struct inet_bind_bucket *tb,
 		    const unsigned short snum)
 {
 	inet_sk(sk)->num = snum;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 70cea9d..583acc3 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -129,7 +129,7 @@ static int sysctl_tcp_congestion_control
 }
 
 
-ctl_table ipv4_table[] = {
+ctl_table __global ipv4_table[] = {
         {
 		.ctl_name	= NET_IPV4_TCP_TIMESTAMPS,
 		.procname	= "tcp_timestamps",
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 7c1bde3..67af642 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -431,7 +431,7 @@ static void tcp_synack_timer(struct sock
 				   TCP_TIMEOUT_INIT, TCP_RTO_MAX);
 }
 
-void tcp_set_keepalive(struct sock *sk, int val)
+void __global tcp_set_keepalive(struct sock *sk, int val)
 {
 	if ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN))
 		return;
@@ -442,7 +442,6 @@ void tcp_set_keepalive(struct sock *sk, 
 		inet_csk_delete_keepalive_timer(sk);
 }
 
-
 static void tcp_keepalive_timer (unsigned long data)
 {
 	struct sock *sk = (struct sock *) data;
diff --git a/net/netfilter/core.c b/net/netfilter/core.c
index 5d29d5e..13af457 100644
--- a/net/netfilter/core.c
+++ b/net/netfilter/core.c
@@ -245,7 +245,7 @@ struct proc_dir_entry *proc_net_netfilte
 EXPORT_SYMBOL(proc_net_netfilter);
 #endif
 
-void __init netfilter_init(void)
+void __init __global netfilter_init(void)
 {
 	int i, h;
 	for (i = 0; i < NPROTO; i++) {
diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c
index 0834c2e..f897dc3 100644
--- a/net/sched/sch_generic.c
+++ b/net/sched/sch_generic.c
@@ -182,7 +182,7 @@ requeue:
 	return q->q.qlen;
 }
 
-void __qdisc_run(struct net_device *dev)
+void __global __qdisc_run(struct net_device *dev)
 {
 	if (unlikely(dev->qdisc == &noop_qdisc))
 		goto out;
@@ -535,7 +535,7 @@ void qdisc_destroy(struct Qdisc *qdisc)
 	call_rcu(&qdisc->q_rcu, __qdisc_destroy);
 }
 
-void dev_activate(struct net_device *dev)
+void __global dev_activate(struct net_device *dev)
 {
 	/* No queueing discipline is attached to device;
 	   create default one i.e. pfifo_fast for devices,
@@ -575,7 +575,7 @@ void dev_activate(struct net_device *dev
 	spin_unlock_bh(&dev->queue_lock);
 }
 
-void dev_deactivate(struct net_device *dev)
+void __global dev_deactivate(struct net_device *dev)
 {
 	struct Qdisc *qdisc;
 
@@ -602,7 +602,7 @@ void dev_deactivate(struct net_device *d
 	}
 }
 
-void dev_init_scheduler(struct net_device *dev)
+void __global dev_init_scheduler(struct net_device *dev)
 {
 	qdisc_lock_tree(dev);
 	dev->qdisc = &noop_qdisc;
@@ -613,7 +613,7 @@ void dev_init_scheduler(struct net_devic
 	dev_watchdog_init(dev);
 }
 
-void dev_shutdown(struct net_device *dev)
+void __global dev_shutdown(struct net_device *dev)
 {
 	struct Qdisc *qdisc;
 
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index f35bc67..e436c99 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -748,7 +748,7 @@ static struct xfrm_policy *clone_policy(
 	return newp;
 }
 
-int __xfrm_sk_clone_policy(struct sock *sk)
+int __global __xfrm_sk_clone_policy(struct sock *sk)
 {
 	struct xfrm_policy *p0 = sk->sk_policy[0],
 			   *p1 = sk->sk_policy[1];
@@ -1438,7 +1438,7 @@ static void __init xfrm_policy_init(void
 	register_netdevice_notifier(&xfrm_dev_notifier);
 }
 
-void __init xfrm_init(void)
+void __init __global xfrm_init(void)
 {
 	xfrm_state_init();
 	xfrm_policy_init();
diff --git a/security/selinux/avc.c b/security/selinux/avc.c
index a300702..1eaaa94 100644
--- a/security/selinux/avc.c
+++ b/security/selinux/avc.c
@@ -686,7 +686,7 @@ void avc_audit(u32 ssid, u32 tsid,
  * @perms based on @tclass.  Returns %0 on success or
  * -%ENOMEM if insufficient memory exists to add the callback.
  */
-int avc_add_callback(int (*callback)(u32 event, u32 ssid, u32 tsid,
+int __global avc_add_callback(int (*callback)(u32 event, u32 ssid, u32 tsid,
                                      u16 tclass, u32 perms,
                                      u32 *out_retained),
                      u32 events, u32 ssid, u32 tsid,
@@ -797,7 +797,7 @@ out:
  * avc_ss_reset - Flush the cache and revalidate migrated permissions.
  * @seqno: policy sequence number
  */
-int avc_ss_reset(u32 seqno)
+int __global avc_ss_reset(u32 seqno)
 {
 	struct avc_callback_node *c;
 	int i, rc = 0, tmprc;
diff --git a/security/selinux/exports.c b/security/selinux/exports.c
index 9d7737d..39ab9a5 100644
--- a/security/selinux/exports.c
+++ b/security/selinux/exports.c
@@ -21,7 +21,7 @@ #include <linux/ipc.h>
 #include "security.h"
 #include "objsec.h"
 
-void selinux_task_ctxid(struct task_struct *tsk, u32 *ctxid)
+void __global selinux_task_ctxid(struct task_struct *tsk, u32 *ctxid)
 {
 	struct task_security_struct *tsec = tsk->security;
 	if (selinux_enabled)
@@ -30,7 +30,7 @@ void selinux_task_ctxid(struct task_stru
 		*ctxid = 0;
 }
 
-int selinux_ctxid_to_string(u32 ctxid, char **ctx, u32 *ctxlen)
+int __global selinux_ctxid_to_string(u32 ctxid, char **ctx, u32 *ctxlen)
 {
 	if (selinux_enabled)
 		return security_sid_to_context(ctxid, ctx, ctxlen);
@@ -42,7 +42,7 @@ int selinux_ctxid_to_string(u32 ctxid, c
 	return 0;
 }
 
-void selinux_get_inode_sid(const struct inode *inode, u32 *sid)
+void __global selinux_get_inode_sid(const struct inode *inode, u32 *sid)
 {
 	if (selinux_enabled) {
 		struct inode_security_struct *isec = inode->i_security;
@@ -52,7 +52,7 @@ void selinux_get_inode_sid(const struct 
 	*sid = 0;
 }
 
-void selinux_get_ipc_sid(const struct kern_ipc_perm *ipcp, u32 *sid)
+void __global selinux_get_ipc_sid(const struct kern_ipc_perm *ipcp, u32 *sid)
 {
 	if (selinux_enabled) {
 		struct ipc_security_struct *isec = ipcp->security;
@@ -62,7 +62,7 @@ void selinux_get_ipc_sid(const struct ke
 	*sid = 0;
 }
 
-void selinux_get_task_sid(struct task_struct *tsk, u32 *sid)
+void __global selinux_get_task_sid(struct task_struct *tsk, u32 *sid)
 {
 	if (selinux_enabled) {
 		struct task_security_struct *tsec = tsk->security;
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
index 5d1b8c7..50e0269 100644
--- a/security/selinux/hooks.c
+++ b/security/selinux/hooks.c
@@ -83,7 +83,7 @@ extern int selinux_nlmsg_lookup(u16 scla
 extern int selinux_compat_net;
 
 #ifdef CONFIG_SECURITY_SELINUX_DEVELOP
-int selinux_enforcing = 0;
+int __global selinux_enforcing = 0;
 
 static int __init enforcing_setup(char *str)
 {
@@ -4683,7 +4683,7 @@ #endif
 	return 0;
 }
 
-void selinux_complete_init(void)
+void __global selinux_complete_init(void)
 {
 	printk(KERN_INFO "SELinux:  Completing initialization.\n");
 
diff --git a/security/selinux/netlink.c b/security/selinux/netlink.c
index e203883..dcb1ec4 100644
--- a/security/selinux/netlink.c
+++ b/security/selinux/netlink.c
@@ -97,7 +97,7 @@ void selnl_notify_setenforce(int val)
 	selnl_notify(SELNL_MSG_SETENFORCE, &val);
 }
 
-void selnl_notify_policyload(u32 seqno)
+void __global selnl_notify_policyload(u32 seqno)
 {
 	selnl_notify(SELNL_MSG_POLICYLOAD, &seqno);
 }
diff --git a/security/selinux/ss/policydb.c b/security/selinux/ss/policydb.c
index f03960e..3ed702f 100644
--- a/security/selinux/ss/policydb.c
+++ b/security/selinux/ss/policydb.c
@@ -45,7 +45,7 @@ static char *symtab_name[SYM_NUM] = {
 };
 #endif
 
-int selinux_mls_enabled = 0;
+int __global selinux_mls_enabled = 0;
 
 static unsigned int symtab_sizes[SYM_NUM] = {
 	2,
diff --git a/security/selinux/ss/services.c b/security/selinux/ss/services.c
index 85e4298..58bd308 100644
--- a/security/selinux/ss/services.c
+++ b/security/selinux/ss/services.c
@@ -42,7 +42,7 @@ #include "conditional.h"
 #include "mls.h"
 
 extern void selnl_notify_policyload(u32 seqno);
-unsigned int policydb_loaded_version;
+unsigned int __global policydb_loaded_version;
 
 static DEFINE_RWLOCK(policy_rwlock);
 #define POLICY_RDLOCK read_lock(&policy_rwlock)
@@ -56,7 +56,7 @@ #define LOAD_UNLOCK mutex_unlock(&load_m
 
 static struct sidtab sidtab;
 struct policydb policydb;
-int ss_initialized = 0;
+int __global ss_initialized = 0;
 
 /*
  * The largest sequence number that has been used when
@@ -397,7 +397,7 @@ out:
 	return -EPERM;
 }
 
-int security_validate_transition(u32 oldsid, u32 newsid, u32 tasksid,
+int __global security_validate_transition(u32 oldsid, u32 newsid, u32 tasksid,
                                  u16 tclass)
 {
 	struct context *ocontext;
@@ -484,7 +484,7 @@ out:
  * Return -%EINVAL if any of the parameters are invalid or %0
  * if the access vector decisions were computed successfully.
  */
-int security_compute_av(u32 ssid,
+int __global security_compute_av(u32 ssid,
 			u32 tsid,
 			u16 tclass,
 			u32 requested,
@@ -583,7 +583,7 @@ #include "initial_sid_to_string.h"
  * into a dynamically allocated string of the correct size.  Set @scontext
  * to point to this string and set @scontext_len to the length of the string.
  */
-int security_sid_to_context(u32 sid, char **scontext, u32 *scontext_len)
+int __global security_sid_to_context(u32 sid, char **scontext, u32 *scontext_len)
 {
 	struct context *context;
 	int rc = 0;
@@ -749,7 +749,7 @@ out:
  * Returns -%EINVAL if the context is invalid, -%ENOMEM if insufficient
  * memory is available, or 0 on success.
  */
-int security_context_to_sid(char *scontext, u32 scontext_len, u32 *sid)
+int __global security_context_to_sid(char *scontext, u32 scontext_len, u32 *sid)
 {
 	return security_context_to_sid_core(scontext, scontext_len,
 	                                    sid, SECSID_NULL);
@@ -772,7 +772,7 @@ int security_context_to_sid(char *sconte
  * Returns -%EINVAL if the context is invalid, -%ENOMEM if insufficient
  * memory is available, or 0 on success.
  */
-int security_context_to_sid_default(char *scontext, u32 scontext_len, u32 *sid, u32 def_sid)
+int __global security_context_to_sid_default(char *scontext, u32 scontext_len, u32 *sid, u32 def_sid)
 {
 	return security_context_to_sid_core(scontext, scontext_len,
 	                                    sid, def_sid);
@@ -959,7 +959,7 @@ out:
  * if insufficient memory is available, or %0 if the new SID was
  * computed successfully.
  */
-int security_transition_sid(u32 ssid,
+int __global security_transition_sid(u32 ssid,
 			    u32 tsid,
 			    u16 tclass,
 			    u32 *out_sid)
@@ -980,7 +980,7 @@ int security_transition_sid(u32 ssid,
  * if insufficient memory is available, or %0 if the SID was
  * computed successfully.
  */
-int security_member_sid(u32 ssid,
+int __global security_member_sid(u32 ssid,
 			u32 tsid,
 			u16 tclass,
 			u32 *out_sid)
@@ -1001,7 +1001,7 @@ int security_member_sid(u32 ssid,
  * if insufficient memory is available, or %0 if the SID was
  * computed successfully.
  */
-int security_change_sid(u32 ssid,
+int __global security_change_sid(u32 ssid,
 			u32 tsid,
 			u16 tclass,
 			u32 *out_sid)
@@ -1210,7 +1210,7 @@ extern void selinux_complete_init(void);
  * This function will flush the access vector cache after
  * loading the new policy.
  */
-int security_load_policy(void *data, size_t len)
+int __global security_load_policy(void *data, size_t len)
 {
 	struct policydb oldpolicydb, newpolicydb;
 	struct sidtab oldsidtab, newsidtab;
@@ -1314,7 +1314,7 @@ err:
  * @port: port number
  * @out_sid: security identifier
  */
-int security_port_sid(u16 domain,
+int __global security_port_sid(u16 domain,
 		      u16 type,
 		      u8 protocol,
 		      u16 port,
@@ -1358,7 +1358,7 @@ out:
  * @if_sid: interface SID
  * @msg_sid: default SID for received packets
  */
-int security_netif_sid(char *name,
+int __global security_netif_sid(char *name,
 		       u32 *if_sid,
 		       u32 *msg_sid)
 {
@@ -1419,7 +1419,7 @@ static int match_ipv6_addrmask(u32 *inpu
  * @addrlen: address length in bytes
  * @out_sid: security identifier
  */
-int security_node_sid(u16 domain,
+int __global security_node_sid(u16 domain,
 		      void *addrp,
 		      u32 addrlen,
 		      u32 *out_sid)
@@ -1502,7 +1502,7 @@ #define SIDS_NEL 25
  * number of elements in the array.
  */
 
-int security_get_user_sids(u32 fromsid,
+int __global security_get_user_sids(u32 fromsid,
 	                   char *username,
 			   u32 **sids,
 			   u32 *nel)
@@ -1605,7 +1605,7 @@ out:
  * cannot support xattr or use a fixed labeling behavior like
  * transition SIDs or task SIDs.
  */
-int security_genfs_sid(const char *fstype,
+int __global security_genfs_sid(const char *fstype,
 	               char *path,
 		       u16 sclass,
 		       u32 *sid)
@@ -1662,7 +1662,7 @@ out:
  * @behavior: labeling behavior
  * @sid: SID for filesystem (superblock)
  */
-int security_fs_use(
+int __global security_fs_use(
 	const char *fstype,
 	unsigned int *behavior,
 	u32 *sid)
@@ -1704,7 +1704,7 @@ out:
 	return rc;
 }
 
-int security_get_bools(int *len, char ***names, int **values)
+int __global security_get_bools(int *len, char ***names, int **values)
 {
 	int i, rc = -ENOMEM;
 
@@ -1750,7 +1750,7 @@ err:
 }
 
 
-int security_set_bools(int len, int *values)
+int __global security_set_bools(int len, int *values)
 {
 	int i, rc = 0;
 	int lenp, seqno = 0;
@@ -1798,7 +1798,7 @@ out:
 	return rc;
 }
 
-int security_get_bool_value(int bool)
+int __global security_get_bool_value(int bool)
 {
 	int rc = 0;
 	int len;
@@ -1822,7 +1822,7 @@ struct selinux_audit_rule {
 	struct context au_ctxt;
 };
 
-void selinux_audit_rule_free(struct selinux_audit_rule *rule)
+void __global selinux_audit_rule_free(struct selinux_audit_rule *rule)
 {
 	if (rule) {
 		context_destroy(&rule->au_ctxt);
@@ -1830,7 +1830,7 @@ void selinux_audit_rule_free(struct seli
 	}
 }
 
-int selinux_audit_rule_init(u32 field, u32 op, char *rulestr,
+int __global selinux_audit_rule_init(u32 field, u32 op, char *rulestr,
                             struct selinux_audit_rule **rule)
 {
 	struct selinux_audit_rule *tmprule;
@@ -1923,7 +1923,7 @@ int selinux_audit_rule_init(u32 field, u
 	return rc;
 }
 
-int selinux_audit_rule_match(u32 ctxid, u32 field, u32 op,
+int __global selinux_audit_rule_match(u32 ctxid, u32 field, u32 op,
                              struct selinux_audit_rule *rule,
                              struct audit_context *actx)
 {
@@ -2060,7 +2060,7 @@ static int __init aurule_init(void)
 }
 __initcall(aurule_init);
 
-void selinux_audit_set_callback(int (*callback)(void))
+void __global selinux_audit_set_callback(int (*callback)(void))
 {
 	aurule_callback = callback;
 }
diff --git a/sound/core/sound_oss.c b/sound/core/sound_oss.c

-- 
dwmw2

Re: [PATCH 3/4] Add __global tag where needed.

[Adrian Bunk]

On Thu, Aug 24, 2006 at 04:26:52PM +0100, David Woodhouse wrote:

> This patch just adds the __global tag to functions and variables which
> are used from outside the directory in which they're provided, to
> prevent -fwhole-program from eating them.
> 
> If we just define __global as a dummy for now, then there's no harm in
> applying this patch as-is -- but it's largely pointless without the
> Makefile changes.
>...

Applying this doesn't seem to make much sense until it's clear whether a
"build everything except for assembler files at once" approach (that 
needs less globals) or your current "compile only multi-obj at once" 
approach (that requires more globals).

> dwmw2

cu
Adrian

-- 

    Gentoo kernels are 42 times more popular than SUSE kernels among
    KLive users  (a service by SUSE contractor Andrea Arcangeli that
    gathers data about kernels from many users worldwide).

       There are three kinds of lies: Lies, Damn Lies, and Statistics.
                                                    Benjamin Disraeli

Re: [PATCH 3/4] Add __global tag where needed.

[David Woodhouse]

On Thu, 2006-08-24 at 23:30 +0200, Adrian Bunk wrote:
> Applying this doesn't seem to make much sense until it's clear whether a
> "build everything except for assembler files at once" approach (that 
> needs less globals) or your current "compile only multi-obj at once" 
> approach (that requires more globals). 

For the kernel itself, I think that building a directory at once is the
way forward. For modules, obviously the scope is more limited.

Either way, I'd like to prevent the unnecessary proliferation of
__global by instrument the link process somehow so that we get a
_warning_ during the final link if there are any global symbols which
aren't actually used. Having said that, --gc-sections will happily drop
them from the vmlinux anyway so I'm not _overly_ concerned by it.

In fact, I think binutils got patched recently so that it can _tell_ us
what sections got dropped by --gc-sections, which would do the job
nicely.

-- 
dwmw2

Re: [PATCH 3/4] Add __global tag where needed.

[Adrian Bunk]

On Fri, Aug 25, 2006 at 10:52:26AM +0100, David Woodhouse wrote:
> On Thu, 2006-08-24 at 23:30 +0200, Adrian Bunk wrote:
> > Applying this doesn't seem to make much sense until it's clear whether a
> > "build everything except for assembler files at once" approach (that 
> > needs less globals) or your current "compile only multi-obj at once" 
> > approach (that requires more globals). 
> 
> For the kernel itself, I think that building a directory at once is the
> way forward. For modules, obviously the scope is more limited.

For any desktop or server you buy today your patches are a nice 
improvement but not that important.

But projects like embedded systems or OLPC that really need want 
kernels should be the same projects that already avoid the
10% size penalty of CONFIG_MODULES=y.

> Either way, I'd like to prevent the unnecessary proliferation of
> __global by instrument the link process somehow so that we get a
> _warning_ during the final link if there are any global symbols which
> aren't actually used.
>...

That would give many false positives from files like fs/libfs.c .

> dwmw2

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: [PATCH 3/4] Add __global tag where needed.

[David Woodhouse]

On Fri, 2006-08-25 at 12:26 +0200, Adrian Bunk wrote:
> On Fri, Aug 25, 2006 at 10:52:26AM +0100, David Woodhouse wrote:
> > On Thu, 2006-08-24 at 23:30 +0200, Adrian Bunk wrote:
> > > Applying this doesn't seem to make much sense until it's clear whether a
> > > "build everything except for assembler files at once" approach (that 
> > > needs less globals) or your current "compile only multi-obj at once" 
> > > approach (that requires more globals). 
> > 
> > For the kernel itself, I think that building a directory at once is the
> > way forward. For modules, obviously the scope is more limited.
> 
> For any desktop or server you buy today your patches are a nice 
> improvement but not that important.

Yes, I agree -- although I haven't tested for performance yet; there
_may_ be something surprisingly improved in there but I suspect it's
unlikely.

> But projects like embedded systems or OLPC that really need want 
> kernels should be the same projects that already avoid the
> 10% size penalty of CONFIG_MODULES=y.

OLPC has USB ports and wants to be fairly flexible about being able to
connect stuff -- I don't think we can turn off CONFIG_MODULES in its
running kernel. However, its _boot_ kernel (LinuxBIOS) has
CONFIG_MODULES=n, and that's where we _really_ care about the space.

-- 
dwmw2

Re: [PATCH 3/4] Add __global tag where needed.

[Adrian Bunk]

On Fri, Aug 25, 2006 at 11:34:38AM +0100, David Woodhouse wrote:
> On Fri, 2006-08-25 at 12:26 +0200, Adrian Bunk wrote:
>...
> > But projects like embedded systems or OLPC that really need want 
> > kernels should be the same projects that already avoid the
> > 10% size penalty of CONFIG_MODULES=y.
> 
> OLPC has USB ports and wants to be fairly flexible about being able to
> connect stuff -- I don't think we can turn off CONFIG_MODULES in its
> running kernel.
>...

You can compile many USB modules statically into the kernel before 
using more additional space than what CONFIG_MODULES=n saves today.

> dwmw2

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

[PATCH 4/4] Some extra --combine hacks.

[David Woodhouse]

Do not apply. Mostly just Makefile hacks to make --combine get used by
using multi-obj compilation in a few places. The correct fix is to use
--combine properly for all C files in built-in.o

We do start to see interesting reductions in size in net/*/built-in.o
when we do this, though...

diff --git a/arch/i386/mm/Makefile b/arch/i386/mm/Makefile
index 80908b5..3bc5bc6 100644
--- a/arch/i386/mm/Makefile
+++ b/arch/i386/mm/Makefile
@@ -2,9 +2,12 @@ #
 # Makefile for the linux i386-specific parts of the memory manager.
 #
 
-obj-y	:= init.o pgtable.o fault.o ioremap.o extable.o pageattr.o mmap.o
+mm-y	:= init.o pgtable.o fault.o ioremap.o extable.o pageattr.o mmap.o
 
-obj-$(CONFIG_NUMA) += discontig.o
-obj-$(CONFIG_HUGETLB_PAGE) += hugetlbpage.o
-obj-$(CONFIG_HIGHMEM) += highmem.o
-obj-$(CONFIG_BOOT_IOREMAP) += boot_ioremap.o
+mm-$(CONFIG_NUMA) += discontig.o
+mm-$(CONFIG_HUGETLB_PAGE) += hugetlbpage.o
+mm-$(CONFIG_HIGHMEM) += highmem.o
+mm-$(CONFIG_BOOT_IOREMAP) += boot_ioremap.o
+
+obj-y := mm.o
+obj-m := $(mm-m)
diff --git a/drivers/md/Makefile b/drivers/md/Makefile
index 34957a6..48e925d 100644
--- a/drivers/md/Makefile
+++ b/drivers/md/Makefile
@@ -11,9 +11,9 @@ md-mod-objs     := md.o bitmap.o
 raid456-objs	:= raid5.o raid6algos.o raid6recov.o raid6tables.o \
 		   raid6int1.o raid6int2.o raid6int4.o \
 		   raid6int8.o raid6int16.o raid6int32.o \
-		   raid6altivec1.o raid6altivec2.o raid6altivec4.o \
-		   raid6altivec8.o \
 		   raid6mmx.o raid6sse1.o raid6sse2.o
+#		   raid6altivec1.o raid6altivec2.o raid6altivec4.o \
+#		   raid6altivec8.o \
 hostprogs-y	:= mktables
 
 # Note: link order is important.  All raid personalities
diff --git a/init/Makefile b/init/Makefile
index 633a268..6ca5c48 100644
--- a/init/Makefile
+++ b/init/Makefile
@@ -2,13 +2,12 @@ #
 # Makefile for the linux kernel.
 #
 
-obj-y				:= main.o version.o mounts.o initramfs.o
-obj-$(CONFIG_GENERIC_CALIBRATE_DELAY) += calibrate.o
-
-mounts-y			:= do_mounts.o
-mounts-$(CONFIG_BLK_DEV_RAM)	+= do_mounts_rd.o
-mounts-$(CONFIG_BLK_DEV_INITRD)	+= do_mounts_initrd.o
-mounts-$(CONFIG_BLK_DEV_MD)	+= do_mounts_md.o
+obj-y			:= init.o version.o
+init-y				:= main.o initramfs.o do_mounts.o
+init-$(CONFIG_GENERIC_CALIBRATE_DELAY) += calibrate.o
+init-$(CONFIG_BLK_DEV_RAM)	+= do_mounts_rd.o
+init-$(CONFIG_BLK_DEV_INITRD)	+= do_mounts_initrd.o
+init-$(CONFIG_BLK_DEV_MD)	+= do_mounts_md.o
 
 # files to be removed upon make clean
 clean-files := ../include/linux/compile.h
diff --git a/ipc/Makefile b/ipc/Makefile
index 0a6d626..7f23e4d 100644
--- a/ipc/Makefile
+++ b/ipc/Makefile
@@ -2,8 +2,9 @@ #
 # Makefile for the linux ipc.
 #
 
-obj-$(CONFIG_SYSVIPC_COMPAT) += compat.o
-obj-$(CONFIG_SYSVIPC) += util.o msgutil.o msg.o sem.o shm.o
+ipc-$(CONFIG_SYSVIPC_COMPAT) += compat.o
+ipc-$(CONFIG_SYSVIPC) += util.o msgutil.o msg.o sem.o shm.o
 obj_mq-$(CONFIG_COMPAT) += compat_mq.o
-obj-$(CONFIG_POSIX_MQUEUE) += mqueue.o msgutil.o $(obj_mq-y)
+ipc-$(CONFIG_POSIX_MQUEUE) += mqueue.o msgutil.o $(obj_mq-y)
 
+obj-y := ipc.o
diff --git a/net/core/Makefile b/net/core/Makefile
index 2645ba4..2dc2473 100644
--- a/net/core/Makefile
+++ b/net/core/Makefile
@@ -17,3 +17,6 @@ obj-$(CONFIG_NET_PKTGEN) += pktgen.o
 obj-$(CONFIG_WIRELESS_EXT) += wireless.o
 obj-$(CONFIG_NETPOLL) += netpoll.o
 obj-$(CONFIG_NET_DMA) += user_dma.o
+
+netcore-y := $(obj-y)
+obj-y := netcore.o
diff --git a/include/linux/kallsyms.h b/include/linux/kallsyms.h
index 849043c..9f91102 100644
--- a/include/linux/kallsyms.h
+++ b/include/linux/kallsyms.h
@@ -5,6 +5,7 @@
 #ifndef _LINUX_KALLSYMS_H
 #define _LINUX_KALLSYMS_H
 
+#include <linux/stddef.h>
 
 #define KSYM_NAME_LEN 127
 
diff --git a/net/ipv4/Makefile b/net/ipv4/Makefile
index 4878fc5..767ef0f 100644
--- a/net/ipv4/Makefile
+++ b/net/ipv4/Makefile
@@ -2,7 +2,9 @@ #
 # Makefile for the Linux TCP/IP (INET) layer.
 #
 
-obj-y     := route.o inetpeer.o protocol.o \
+obj-y	:= ipv4.o
+
+ipv4-y     := route.o inetpeer.o protocol.o \
 	     ip_input.o ip_fragment.o ip_forward.o ip_options.o \
 	     ip_output.o ip_sockglue.o inet_hashtables.o \
 	     inet_timewait_sock.o inet_connection_sock.o \
@@ -11,42 +13,45 @@ obj-y     := route.o inetpeer.o protocol
 	     datagram.o raw.o udp.o arp.o icmp.o devinet.o af_inet.o igmp.o \
 	     sysctl_net_ipv4.o fib_frontend.o fib_semantics.o
 
-obj-$(CONFIG_IP_FIB_HASH) += fib_hash.o
-obj-$(CONFIG_IP_FIB_TRIE) += fib_trie.o
-obj-$(CONFIG_PROC_FS) += proc.o
-obj-$(CONFIG_IP_MULTIPLE_TABLES) += fib_rules.o
-obj-$(CONFIG_IP_MROUTE) += ipmr.o
-obj-$(CONFIG_NET_IPIP) += ipip.o
-obj-$(CONFIG_NET_IPGRE) += ip_gre.o
-obj-$(CONFIG_SYN_COOKIES) += syncookies.o
-obj-$(CONFIG_INET_AH) += ah4.o
-obj-$(CONFIG_INET_ESP) += esp4.o
-obj-$(CONFIG_INET_IPCOMP) += ipcomp.o
-obj-$(CONFIG_INET_XFRM_TUNNEL) += xfrm4_tunnel.o
-obj-$(CONFIG_INET_TUNNEL) += tunnel4.o
-obj-$(CONFIG_INET_XFRM_MODE_TRANSPORT) += xfrm4_mode_transport.o
-obj-$(CONFIG_INET_XFRM_MODE_TUNNEL) += xfrm4_mode_tunnel.o
-obj-$(CONFIG_IP_PNP) += ipconfig.o
-obj-$(CONFIG_IP_ROUTE_MULTIPATH_RR) += multipath_rr.o
-obj-$(CONFIG_IP_ROUTE_MULTIPATH_RANDOM) += multipath_random.o
-obj-$(CONFIG_IP_ROUTE_MULTIPATH_WRANDOM) += multipath_wrandom.o
-obj-$(CONFIG_IP_ROUTE_MULTIPATH_DRR) += multipath_drr.o
-obj-$(CONFIG_NETFILTER)	+= netfilter.o netfilter/
+ipv4-$(CONFIG_IP_FIB_HASH) += fib_hash.o
+ipv4-$(CONFIG_IP_FIB_TRIE) += fib_trie.o
+ipv4-$(CONFIG_PROC_FS) += proc.o
+ipv4-$(CONFIG_IP_MULTIPLE_TABLES) += fib_rules.o
+ipv4-$(CONFIG_IP_MROUTE) += ipmr.o
+ipv4-$(CONFIG_NET_IPIP) += ipip.o
+ipv4-$(CONFIG_NET_IPGRE) += ip_gre.o
+ipv4-$(CONFIG_SYN_COOKIES) += syncookies.o
+ipv4-$(CONFIG_INET_AH) += ah4.o
+ipv4-$(CONFIG_INET_ESP) += esp4.o
+ipv4-$(CONFIG_INET_IPCOMP) += ipcomp.o
+ipv4-$(CONFIG_INET_XFRM_TUNNEL) += xfrm4_tunnel.o
+ipv4-$(CONFIG_INET_TUNNEL) += tunnel4.o
+ipv4-$(CONFIG_INET_XFRM_MODE_TRANSPORT) += xfrm4_mode_transport.o
+ipv4-$(CONFIG_INET_XFRM_MODE_TUNNEL) += xfrm4_mode_tunnel.o
+ipv4-$(CONFIG_IP_PNP) += ipconfig.o
+ipv4-$(CONFIG_IP_ROUTE_MULTIPATH_RR) += multipath_rr.o
+ipv4-$(CONFIG_IP_ROUTE_MULTIPATH_RANDOM) += multipath_random.o
+ipv4-$(CONFIG_IP_ROUTE_MULTIPATH_WRANDOM) += multipath_wrandom.o
+ipv4-$(CONFIG_IP_ROUTE_MULTIPATH_DRR) += multipath_drr.o
+ipv4-$(CONFIG_NETFILTER)	+= netfilter.o
+obj-$(CONFIG_NETFILTER)		+= netfilter/
 obj-$(CONFIG_IP_VS) += ipvs/
-obj-$(CONFIG_INET_DIAG) += inet_diag.o 
-obj-$(CONFIG_IP_ROUTE_MULTIPATH_CACHED) += multipath.o
-obj-$(CONFIG_INET_TCP_DIAG) += tcp_diag.o
-obj-$(CONFIG_NET_TCPPROBE) += tcp_probe.o
-obj-$(CONFIG_TCP_CONG_BIC) += tcp_bic.o
-obj-$(CONFIG_TCP_CONG_CUBIC) += tcp_cubic.o
-obj-$(CONFIG_TCP_CONG_WESTWOOD) += tcp_westwood.o
-obj-$(CONFIG_TCP_CONG_HSTCP) += tcp_highspeed.o
-obj-$(CONFIG_TCP_CONG_HYBLA) += tcp_hybla.o
-obj-$(CONFIG_TCP_CONG_HTCP) += tcp_htcp.o
-obj-$(CONFIG_TCP_CONG_VEGAS) += tcp_vegas.o
-obj-$(CONFIG_TCP_CONG_VENO) += tcp_veno.o
-obj-$(CONFIG_TCP_CONG_SCALABLE) += tcp_scalable.o
-obj-$(CONFIG_TCP_CONG_LP) += tcp_lp.o
+ipv4-$(CONFIG_INET_DIAG) += inet_diag.o 
+ipv4-$(CONFIG_IP_ROUTE_MULTIPATH_CACHED) += multipath.o
+ipv4-$(CONFIG_INET_TCP_DIAG) += tcp_diag.o
+ipv4-$(CONFIG_NET_TCPPROBE) += tcp_probe.o
+ipv4-$(CONFIG_TCP_CONG_BIC) += tcp_bic.o
+ipv4-$(CONFIG_TCP_CONG_CUBIC) += tcp_cubic.o
+ipv4-$(CONFIG_TCP_CONG_WESTWOOD) += tcp_westwood.o
+ipv4-$(CONFIG_TCP_CONG_HSTCP) += tcp_highspeed.o
+ipv4-$(CONFIG_TCP_CONG_HYBLA) += tcp_hybla.o
+ipv4-$(CONFIG_TCP_CONG_HTCP) += tcp_htcp.o
+ipv4-$(CONFIG_TCP_CONG_VEGAS) += tcp_vegas.o
+ipv4-$(CONFIG_TCP_CONG_VENO) += tcp_veno.o
+ipv4-$(CONFIG_TCP_CONG_SCALABLE) += tcp_scalable.o
+ipv4-$(CONFIG_TCP_CONG_LP) += tcp_lp.o
 
-obj-$(CONFIG_XFRM) += xfrm4_policy.o xfrm4_state.o xfrm4_input.o \
+ipv4-$(CONFIG_XFRM) += xfrm4_policy.o xfrm4_state.o xfrm4_input.o \
 		      xfrm4_output.o
+
+obj-m := $(ipv4-m)
diff --git a/net/sched/Makefile b/net/sched/Makefile
index 0f06aec..c605cc1 100644
--- a/net/sched/Makefile
+++ b/net/sched/Makefile
@@ -41,3 +41,6 @@ obj-$(CONFIG_NET_EMATCH_NBYTE)	+= em_nby
 obj-$(CONFIG_NET_EMATCH_U32)	+= em_u32.o
 obj-$(CONFIG_NET_EMATCH_META)	+= em_meta.o
 obj-$(CONFIG_NET_EMATCH_TEXT)	+= em_text.o
+
+netsched-y := $(obj-y)
+obj-y := netsched.o
diff --git a/net/xfrm/Makefile b/net/xfrm/Makefile
index 693aac1..3500f84 100644
--- a/net/xfrm/Makefile
+++ b/net/xfrm/Makefile
@@ -5,3 +5,5 @@ #
 obj-$(CONFIG_XFRM) := xfrm_policy.o xfrm_state.o xfrm_input.o xfrm_algo.o
 obj-$(CONFIG_XFRM_USER) += xfrm_user.o
 
+xfrm-y := $(obj-y)
+obj-y := xfrm.o
diff --git a/fs/Makefile b/fs/Makefile
index 8913542..faedce0 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -34,7 +34,6 @@ obj-$(CONFIG_BINFMT_FLAT)	+= binfmt_flat
 
 obj-$(CONFIG_FS_MBCACHE)	+= mbcache.o
 obj-$(CONFIG_FS_POSIX_ACL)	+= posix_acl.o xattr_acl.o
-obj-$(CONFIG_NFS_COMMON)	+= nfs_common/
 
 obj-$(CONFIG_QUOTA)		+= dquot.o
 obj-$(CONFIG_QFMT_V1)		+= quota_v1.o
@@ -42,6 +41,12 @@ obj-$(CONFIG_QFMT_V2)		+= quota_v2.o
 obj-$(CONFIG_QUOTACTL)		+= quota.o
 
 obj-$(CONFIG_DNOTIFY)		+= dnotify.o
+obj-$(CONFIG_PROFILING)		+= dcookies.o
+
+fs-y := $(obj-y)
+obj-y := fs.o
+
+obj-$(CONFIG_NFS_COMMON)	+= nfs_common/
 
 obj-$(CONFIG_PROC_FS)		+= proc/
 obj-y				+= partitions/
@@ -49,7 +54,6 @@ obj-$(CONFIG_SYSFS)		+= sysfs/
 obj-$(CONFIG_CONFIGFS_FS)	+= configfs/
 obj-y				+= devpts/
 
-obj-$(CONFIG_PROFILING)		+= dcookies.o
  
 # Do not add any filesystems before this line
 obj-$(CONFIG_REISERFS_FS)	+= reiserfs/
diff --git a/drivers/net/e1000/Makefile b/drivers/net/e1000/Makefile
index 5dea2b7..9bece1d 100644
--- a/drivers/net/e1000/Makefile
+++ b/drivers/net/e1000/Makefile
@@ -31,6 +31,10 @@ #
 # Makefile for the Intel(R) PRO/1000 ethernet driver
 #
 
+# Disable use of --combine for now. GCC PR28779
+CONFIG_COMBINED_COMPILE=
+
 obj-$(CONFIG_E1000) += e1000.o
 
 e1000-objs := e1000_main.o e1000_hw.o e1000_ethtool.o e1000_param.o
+

-- 
dwmw2