[PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Arnd Bergmann]

This adds an asm/asm-prototypes.h header for ARM to fix the broken symbol
versioning for symbols exported from assembler files.

I couldn't find the correct prototypes for the compiler builtins,
so I went with the fake 'void f(void)' prototypes that we had
before, restoring the state before they were moved.

Originally I assumed that the problem was just a harmless warning
in unusual configurations, but as Uwe found, we actually need this
to load most modules when symbol versioning is enabled, as it is
in many distro kernels.

Cc: Uwe Kleine-König <uwe@kleine-koenig.org>
Fixes: 4dd1837d7589 ("arm: move exports to definitions")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
Compared to the earlier version, I dropped the changes to the
csumpartial files, which now get handled correctly by Kbuild
even when the export comes from a macro, and I also dropped the
changes to the bitops files, which were already fixed in a
patch from Nico.

The patch applies cleanly on top of the rmk/fixes tree but has
no effect there, as it also needs 4efca4ed05cb ("kbuild: modversions
for EXPORT_SYMBOL() for asm") and cc6acc11cad1 ("kbuild: be more
careful about matching preprocessed asm ___EXPORT_SYMBOL").

With the combination of rmk/fixes, torvalds/master and these two
patches, symbol versioning works again on ARM. As it is still
broken on almost all other architectures (powerpc is fixed,
x86 has a patch), I wonder if we should make CONFIG_MODVERSIONS
as broken for everything else.
---
 arch/arm/include/asm/asm-prototypes.h | 34 ++++++++++++++++++++++++++++++++++
 1 file changed, 34 insertions(+)
 create mode 100644 arch/arm/include/asm/asm-prototypes.h

diff --git a/arch/arm/include/asm/asm-prototypes.h b/arch/arm/include/asm/asm-prototypes.h
new file mode 100644
index 000000000000..04e5616a7b15
--- /dev/null
+++ b/arch/arm/include/asm/asm-prototypes.h
@@ -0,0 +1,34 @@
+#include <linux/arm-smccc.h>
+#include <linux/bitops.h>
+#include <linux/ftrace.h>
+#include <linux/io.h>
+#include <linux/platform_data/asoc-imx-ssi.h>
+#include <linux/string.h>
+#include <linux/uaccess.h>
+
+#include <asm/checksum.h>
+#include <asm/div64.h>
+#include <asm/memory.h>
+
+extern void __aeabi_idivmod(void);
+extern void __aeabi_idiv(void);
+extern void __aeabi_lasr(void);
+extern void __aeabi_llsl(void);
+extern void __aeabi_llsr(void);
+extern void __aeabi_lmul(void);
+extern void __aeabi_uidivmod(void);
+extern void __aeabi_uidiv(void);
+extern void __aeabi_ulcmp(void);
+
+extern void __ashldi3(void);
+extern void __ashrdi3(void);
+extern void __bswapdi2(void);
+extern void __bswapsi2(void);
+extern void __divsi3(void);
+extern void __do_div64(void);
+extern void __lshrdi3(void);
+extern void __modsi3(void);
+extern void __muldi3(void);
+extern void __ucmpdi2(void);
+extern void __udivsi3(void);
+extern void __umodsi3(void);
-- 
2.9.0

[PATCH 2/2] ARM: move mmiocpy/mmioset exports to io.c

[Arnd Bergmann]

The prototypes for mmioset/mmiocpy are intentionally hidden
inside of inline functions, which breaks the EXPORT_SYMBOL
statements when symbol versioning is enabled.

This moves the two exports from the files that implement the
code into the kernel/io.c file, adding another local declaration
there.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
 arch/arm/kernel/io.c  | 7 +++++++
 arch/arm/lib/memcpy.S | 1 -
 arch/arm/lib/memset.S | 1 -
 3 files changed, 7 insertions(+), 2 deletions(-)

diff --git a/arch/arm/kernel/io.c b/arch/arm/kernel/io.c
index eedefe050022..c74746997626 100644
--- a/arch/arm/kernel/io.c
+++ b/arch/arm/kernel/io.c
@@ -82,3 +82,10 @@ void _memset_io(volatile void __iomem *dst, int c, size_t count)
 	}
 }
 EXPORT_SYMBOL(_memset_io);
+
+/* can't export them from memcpy.S/memset.S because of hidden declaration */
+void mmioset(void __iomem *addr, unsigned int c, size_t n);
+EXPORT_SYMBOL(mmioset);
+
+void mmiocpy(void *dest, const void __iomem *src, size_t n);
+EXPORT_SYMBOL(mmiocpy);
diff --git a/arch/arm/lib/memcpy.S b/arch/arm/lib/memcpy.S
index 1be5b6ddf37c..1f822fc52400 100644
--- a/arch/arm/lib/memcpy.S
+++ b/arch/arm/lib/memcpy.S
@@ -70,4 +70,3 @@ ENTRY(memcpy)
 ENDPROC(memcpy)
 ENDPROC(mmiocpy)
 EXPORT_SYMBOL(memcpy)
-EXPORT_SYMBOL(mmiocpy)
diff --git a/arch/arm/lib/memset.S b/arch/arm/lib/memset.S
index 7b72044cba62..6f075ca09abc 100644
--- a/arch/arm/lib/memset.S
+++ b/arch/arm/lib/memset.S
@@ -137,4 +137,3 @@ UNWIND( .fnend   )
 ENDPROC(memset)
 ENDPROC(mmioset)
 EXPORT_SYMBOL(memset)
-EXPORT_SYMBOL(mmioset)
-- 
2.9.0

Re: [PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Nicolas Pitre]

[-- Attachment #1: Type: text/plain, Size: 2371 bytes --]

On Tue, 22 Nov 2016, Arnd Bergmann wrote:

> This adds an asm/asm-prototypes.h header for ARM to fix the broken symbol
> versioning for symbols exported from assembler files.
> 
> I couldn't find the correct prototypes for the compiler builtins,
> so I went with the fake 'void f(void)' prototypes that we had
> before, restoring the state before they were moved.
> 
> Originally I assumed that the problem was just a harmless warning
> in unusual configurations, but as Uwe found, we actually need this
> to load most modules when symbol versioning is enabled, as it is
> in many distro kernels.
> 
> Cc: Uwe Kleine-König <uwe@kleine-koenig.org>
> Fixes: 4dd1837d7589 ("arm: move exports to definitions")
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---
> Compared to the earlier version, I dropped the changes to the
> csumpartial files, which now get handled correctly by Kbuild
> even when the export comes from a macro, and I also dropped the
> changes to the bitops files, which were already fixed in a
> patch from Nico.
> 
> The patch applies cleanly on top of the rmk/fixes tree but has
> no effect there, as it also needs 4efca4ed05cb ("kbuild: modversions
> for EXPORT_SYMBOL() for asm") and cc6acc11cad1 ("kbuild: be more
> careful about matching preprocessed asm ___EXPORT_SYMBOL").
> 
> With the combination of rmk/fixes, torvalds/master and these two
> patches, symbol versioning works again on ARM. As it is still
> broken on almost all other architectures (powerpc is fixed,
> x86 has a patch), I wonder if we should make CONFIG_MODVERSIONS
> as broken for everything else.

I'm not sure I like this at all.

The goal for moving EXPORT_SYMBOL() to assembly code where symbols were 
defined is to make things close together and avoid those centralized 
list of symbols that you can easily miss when modifying the actual code.

This series is therefore bringing back a centralized list of symbols in 
a slightly different form, nullifying the advantages from having moved 
EXPORT_SYMBOL() to asm code.  To me this looks like a big step backward.

Why not simply extending the original idea of keeping exports close to 
the actual code by _also_ having a macro that provides the function 
prototype alongside the EXPORT_SYMBOL() instance?  That could even be 
expressed with some EXPORT_SYMBOL_PROTO(ret, sym, arg...) macro that 
does it all.


Nicolas

Re: [PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Russell King - ARM Linux]

On Tue, Nov 22, 2016 at 11:34:48AM -0500, Nicolas Pitre wrote:
> On Tue, 22 Nov 2016, Arnd Bergmann wrote:
> > This adds an asm/asm-prototypes.h header for ARM to fix the broken symbol
> > versioning for symbols exported from assembler files.
> > 
> > I couldn't find the correct prototypes for the compiler builtins,
> > so I went with the fake 'void f(void)' prototypes that we had
> > before, restoring the state before they were moved.
> > 
> > Originally I assumed that the problem was just a harmless warning
> > in unusual configurations, but as Uwe found, we actually need this
> > to load most modules when symbol versioning is enabled, as it is
> > in many distro kernels.
> > 
> > Cc: Uwe Kleine-König <uwe@kleine-koenig.org>
> > Fixes: 4dd1837d7589 ("arm: move exports to definitions")
> > Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> > ---
> > Compared to the earlier version, I dropped the changes to the
> > csumpartial files, which now get handled correctly by Kbuild
> > even when the export comes from a macro, and I also dropped the
> > changes to the bitops files, which were already fixed in a
> > patch from Nico.
> > 
> > The patch applies cleanly on top of the rmk/fixes tree but has
> > no effect there, as it also needs 4efca4ed05cb ("kbuild: modversions
> > for EXPORT_SYMBOL() for asm") and cc6acc11cad1 ("kbuild: be more
> > careful about matching preprocessed asm ___EXPORT_SYMBOL").
> > 
> > With the combination of rmk/fixes, torvalds/master and these two
> > patches, symbol versioning works again on ARM. As it is still
> > broken on almost all other architectures (powerpc is fixed,
> > x86 has a patch), I wonder if we should make CONFIG_MODVERSIONS
> > as broken for everything else.
> 
> I'm not sure I like this at all.
> 
> The goal for moving EXPORT_SYMBOL() to assembly code where symbols were 
> defined is to make things close together and avoid those centralized 
> list of symbols that you can easily miss when modifying the actual code.
> 
> This series is therefore bringing back a centralized list of symbols in 
> a slightly different form, nullifying the advantages from having moved 
> EXPORT_SYMBOL() to asm code.  To me this looks like a big step backward.
> 
> Why not simply extending the original idea of keeping exports close to 
> the actual code by _also_ having a macro that provides the function 
> prototype alongside the EXPORT_SYMBOL() instance?  That could even be 
> expressed with some EXPORT_SYMBOL_PROTO(ret, sym, arg...) macro that 
> does it all.

What you're saying is that you don't like the solution that's taken
weeks to get merged up to this point, so where do we go from here
now?  This crap has been broken since 4.9-rc1, and is a regression.

I think at this point, we just declare that modversions are broken
on ARM, and those who created this mess get to explain to people
why the fsck they broke the kernel.

4.9 is the next LTS kernel?  ROTFL!

I agree with Nicolas - it seems that the whole EXPORT_SYMBOL() crap
has just been a pointless exercise in churn, resulting only in
something "different" because it looks "cool" to do it some other
way.  There's no real benefit here at all, only harm.

Just revert the damned patches that created this breakage in the
first place please.  It's now way too late to be trying to fix it
any other way.

-- 
RMK's Patch system: http://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line: currently at 9.6Mbps down 400kbps up
according to speedtest.net.

Re: [PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Nicholas Piggin]

On Wed, 23 Nov 2016 00:41:07 +0000
Russell King - ARM Linux <linux@armlinux.org.uk> wrote:

> On Tue, Nov 22, 2016 at 11:34:48AM -0500, Nicolas Pitre wrote:
> > On Tue, 22 Nov 2016, Arnd Bergmann wrote:  
> > > This adds an asm/asm-prototypes.h header for ARM to fix the broken symbol
> > > versioning for symbols exported from assembler files.
> > > 
> > > I couldn't find the correct prototypes for the compiler builtins,
> > > so I went with the fake 'void f(void)' prototypes that we had
> > > before, restoring the state before they were moved.
> > > 
> > > Originally I assumed that the problem was just a harmless warning
> > > in unusual configurations, but as Uwe found, we actually need this
> > > to load most modules when symbol versioning is enabled, as it is
> > > in many distro kernels.
> > > 
> > > Cc: Uwe Kleine-König <uwe@kleine-koenig.org>
> > > Fixes: 4dd1837d7589 ("arm: move exports to definitions")
> > > Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> > > ---
> > > Compared to the earlier version, I dropped the changes to the
> > > csumpartial files, which now get handled correctly by Kbuild
> > > even when the export comes from a macro, and I also dropped the
> > > changes to the bitops files, which were already fixed in a
> > > patch from Nico.
> > > 
> > > The patch applies cleanly on top of the rmk/fixes tree but has
> > > no effect there, as it also needs 4efca4ed05cb ("kbuild: modversions
> > > for EXPORT_SYMBOL() for asm") and cc6acc11cad1 ("kbuild: be more
> > > careful about matching preprocessed asm ___EXPORT_SYMBOL").
> > > 
> > > With the combination of rmk/fixes, torvalds/master and these two
> > > patches, symbol versioning works again on ARM. As it is still
> > > broken on almost all other architectures (powerpc is fixed,
> > > x86 has a patch), I wonder if we should make CONFIG_MODVERSIONS
> > > as broken for everything else.  
> > 
> > I'm not sure I like this at all.
> > 
> > The goal for moving EXPORT_SYMBOL() to assembly code where symbols were 
> > defined is to make things close together and avoid those centralized 
> > list of symbols that you can easily miss when modifying the actual code.
> > 
> > This series is therefore bringing back a centralized list of symbols in 
> > a slightly different form, nullifying the advantages from having moved 
> > EXPORT_SYMBOL() to asm code.  To me this looks like a big step backward.
> > 
> > Why not simply extending the original idea of keeping exports close to 
> > the actual code by _also_ having a macro that provides the function 
> > prototype alongside the EXPORT_SYMBOL() instance?  That could even be 
> > expressed with some EXPORT_SYMBOL_PROTO(ret, sym, arg...) macro that 
> > does it all.  
> 
> What you're saying is that you don't like the solution that's taken
> weeks to get merged up to this point, so where do we go from here
> now?  This crap has been broken since 4.9-rc1, and is a regression.
> 
> I think at this point, we just declare that modversions are broken
> on ARM, and those who created this mess get to explain to people
> why the fsck they broke the kernel.

Let's fix it instead :)

Why not merge Arnd's 2 patches? I think he mostly addressed your concerns
of them. Or...

> 
> 4.9 is the next LTS kernel?  ROTFL!
> 
> I agree with Nicolas - it seems that the whole EXPORT_SYMBOL() crap
> has just been a pointless exercise in churn, resulting only in
> something "different" because it looks "cool" to do it some other
> way.  There's no real benefit here at all, only harm.
> 
> Just revert the damned patches that created this breakage in the
> first place please.  It's now way too late to be trying to fix it
> any other way.
> 

4dd1837d7589 diffstat is entirely in arch/arm. I think reverting that
would fix it (I haven't tested it myself so I would advise testing
before committing). So the ball is in your court.

As for process concerns, you have made valid points. Sometimes a mistake
is made or we make an incorrect assumption about how another person works
or what mailing lists they have read, or do not anticipate the fallout
from some change.

Everyone does it sometimes, so it's never a bad time to reflect on how
we work with others and try to do better.

Thanks,
Nick

Re: [PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Nicholas Piggin]

On Tue, 22 Nov 2016 11:34:48 -0500 (EST)
Nicolas Pitre <nicolas.pitre@linaro.org> wrote:

> On Tue, 22 Nov 2016, Arnd Bergmann wrote:
> 
> > This adds an asm/asm-prototypes.h header for ARM to fix the broken symbol
> > versioning for symbols exported from assembler files.
> > 
> > I couldn't find the correct prototypes for the compiler builtins,
> > so I went with the fake 'void f(void)' prototypes that we had
> > before, restoring the state before they were moved.
> > 
> > Originally I assumed that the problem was just a harmless warning
> > in unusual configurations, but as Uwe found, we actually need this
> > to load most modules when symbol versioning is enabled, as it is
> > in many distro kernels.
> > 
> > Cc: Uwe Kleine-König <uwe@kleine-koenig.org>
> > Fixes: 4dd1837d7589 ("arm: move exports to definitions")
> > Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> > ---
> > Compared to the earlier version, I dropped the changes to the
> > csumpartial files, which now get handled correctly by Kbuild
> > even when the export comes from a macro, and I also dropped the
> > changes to the bitops files, which were already fixed in a
> > patch from Nico.
> > 
> > The patch applies cleanly on top of the rmk/fixes tree but has
> > no effect there, as it also needs 4efca4ed05cb ("kbuild: modversions
> > for EXPORT_SYMBOL() for asm") and cc6acc11cad1 ("kbuild: be more
> > careful about matching preprocessed asm ___EXPORT_SYMBOL").
> > 
> > With the combination of rmk/fixes, torvalds/master and these two
> > patches, symbol versioning works again on ARM. As it is still
> > broken on almost all other architectures (powerpc is fixed,
> > x86 has a patch), I wonder if we should make CONFIG_MODVERSIONS
> > as broken for everything else.  
> 
> I'm not sure I like this at all.
> 
> The goal for moving EXPORT_SYMBOL() to assembly code where symbols were 
> defined is to make things close together and avoid those centralized 
> list of symbols that you can easily miss when modifying the actual code.

Right.

> 
> This series is therefore bringing back a centralized list of symbols in 
> a slightly different form, nullifying the advantages from having moved 
> EXPORT_SYMBOL() to asm code.  To me this looks like a big step backward.

Exported symbols have C declarations in headers already. For the most
part, anyway -- these ones Arnd adds are for compiler runtime which is
why some architectures haven't had the prototypes.

> Why not simply extending the original idea of keeping exports close to 
> the actual code by _also_ having a macro that provides the function 
> prototype alongside the EXPORT_SYMBOL() instance?  That could even be 
> expressed with some EXPORT_SYMBOL_PROTO(ret, sym, arg...) macro that 
> does it all.

Well, the reason is to get 4.9 working, I never thought asm-prototypes.h
was a beautiful solution or it should not be changed if we can find ways
to improve it.

EXPORT_SYMBOL_PROTO() for asm code seems possibly a good idea for 4.10.
Of course your exported symbols will still have their prototypes in
various headers -- that redundancy is inherent.

Thanks,
Nick

Re: [PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Nicolas Pitre]

[-- Attachment #1: Type: text/plain, Size: 3082 bytes --]

On Wed, 23 Nov 2016, Nicholas Piggin wrote:

> On Tue, 22 Nov 2016 11:34:48 -0500 (EST)
> Nicolas Pitre <nicolas.pitre@linaro.org> wrote:
> 
> > On Tue, 22 Nov 2016, Arnd Bergmann wrote:
> > 
> > > This adds an asm/asm-prototypes.h header for ARM to fix the broken symbol
> > > versioning for symbols exported from assembler files.
> > > 
> > > I couldn't find the correct prototypes for the compiler builtins,
> > > so I went with the fake 'void f(void)' prototypes that we had
> > > before, restoring the state before they were moved.
> > > 
> > > Originally I assumed that the problem was just a harmless warning
> > > in unusual configurations, but as Uwe found, we actually need this
> > > to load most modules when symbol versioning is enabled, as it is
> > > in many distro kernels.
> > > 
> > > Cc: Uwe Kleine-König <uwe@kleine-koenig.org>
> > > Fixes: 4dd1837d7589 ("arm: move exports to definitions")
> > > Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> > > ---
> > > Compared to the earlier version, I dropped the changes to the
> > > csumpartial files, which now get handled correctly by Kbuild
> > > even when the export comes from a macro, and I also dropped the
> > > changes to the bitops files, which were already fixed in a
> > > patch from Nico.
> > > 
> > > The patch applies cleanly on top of the rmk/fixes tree but has
> > > no effect there, as it also needs 4efca4ed05cb ("kbuild: modversions
> > > for EXPORT_SYMBOL() for asm") and cc6acc11cad1 ("kbuild: be more
> > > careful about matching preprocessed asm ___EXPORT_SYMBOL").
> > > 
> > > With the combination of rmk/fixes, torvalds/master and these two
> > > patches, symbol versioning works again on ARM. As it is still
> > > broken on almost all other architectures (powerpc is fixed,
> > > x86 has a patch), I wonder if we should make CONFIG_MODVERSIONS
> > > as broken for everything else.  
> > 
> > I'm not sure I like this at all.
> > 
> > The goal for moving EXPORT_SYMBOL() to assembly code where symbols were 
> > defined is to make things close together and avoid those centralized 
> > list of symbols that you can easily miss when modifying the actual code.
> 
> Right.
> 
> > 
> > This series is therefore bringing back a centralized list of symbols in 
> > a slightly different form, nullifying the advantages from having moved 
> > EXPORT_SYMBOL() to asm code.  To me this looks like a big step backward.
> 
> Exported symbols have C declarations in headers already. For the most
> part, anyway -- these ones Arnd adds are for compiler runtime which is
> why some architectures haven't had the prototypes.

Hmmm. That's right.  That makes it much more justifiable.
My main objection is withdrawn.

However there is a bunch of includes added to asm-prototypes.h:

+#include <linux/arm-smccc.h>
+#include <linux/bitops.h>
+#include <linux/ftrace.h>
+#include <linux/io.h>
+#include <linux/platform_data/asoc-imx-ssi.h>
+#include <linux/string.h>
+#include <linux/uaccess.h>
+
+#include <asm/checksum.h>
+#include <asm/div64.h>
+#include <asm/memory.h>

Are those necessary at all?


Nicolas

Re: [PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Russell King - ARM Linux]

On Tue, Nov 22, 2016 at 08:35:48PM -0500, Nicolas Pitre wrote:
> On Wed, 23 Nov 2016, Nicholas Piggin wrote:
> 
> > On Tue, 22 Nov 2016 11:34:48 -0500 (EST)
> > Nicolas Pitre <nicolas.pitre@linaro.org> wrote:
> > 
> > > On Tue, 22 Nov 2016, Arnd Bergmann wrote:
> > > 
> > > > This adds an asm/asm-prototypes.h header for ARM to fix the broken symbol
> > > > versioning for symbols exported from assembler files.
> > > > 
> > > > I couldn't find the correct prototypes for the compiler builtins,
> > > > so I went with the fake 'void f(void)' prototypes that we had
> > > > before, restoring the state before they were moved.
> > > > 
> > > > Originally I assumed that the problem was just a harmless warning
> > > > in unusual configurations, but as Uwe found, we actually need this
> > > > to load most modules when symbol versioning is enabled, as it is
> > > > in many distro kernels.
> > > > 
> > > > Cc: Uwe Kleine-König <uwe@kleine-koenig.org>
> > > > Fixes: 4dd1837d7589 ("arm: move exports to definitions")
> > > > Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> > > > ---
> > > > Compared to the earlier version, I dropped the changes to the
> > > > csumpartial files, which now get handled correctly by Kbuild
> > > > even when the export comes from a macro, and I also dropped the
> > > > changes to the bitops files, which were already fixed in a
> > > > patch from Nico.
> > > > 
> > > > The patch applies cleanly on top of the rmk/fixes tree but has
> > > > no effect there, as it also needs 4efca4ed05cb ("kbuild: modversions
> > > > for EXPORT_SYMBOL() for asm") and cc6acc11cad1 ("kbuild: be more
> > > > careful about matching preprocessed asm ___EXPORT_SYMBOL").
> > > > 
> > > > With the combination of rmk/fixes, torvalds/master and these two
> > > > patches, symbol versioning works again on ARM. As it is still
> > > > broken on almost all other architectures (powerpc is fixed,
> > > > x86 has a patch), I wonder if we should make CONFIG_MODVERSIONS
> > > > as broken for everything else.  
> > > 
> > > I'm not sure I like this at all.
> > > 
> > > The goal for moving EXPORT_SYMBOL() to assembly code where symbols were 
> > > defined is to make things close together and avoid those centralized 
> > > list of symbols that you can easily miss when modifying the actual code.
> > 
> > Right.
> > 
> > > 
> > > This series is therefore bringing back a centralized list of symbols in 
> > > a slightly different form, nullifying the advantages from having moved 
> > > EXPORT_SYMBOL() to asm code.  To me this looks like a big step backward.
> > 
> > Exported symbols have C declarations in headers already. For the most
> > part, anyway -- these ones Arnd adds are for compiler runtime which is
> > why some architectures haven't had the prototypes.
> 
> Hmmm. That's right.  That makes it much more justifiable.
> My main objection is withdrawn.

I don't see it makes any difference - the armksyms.c originally had
the same:

-#include <linux/export.h>
-#include <linux/sched.h>
-#include <linux/string.h>
-#include <linux/cryptohash.h>
-#include <linux/delay.h>
-#include <linux/in6.h>
-#include <linux/syscalls.h>
-#include <linux/uaccess.h>
-#include <linux/io.h>
-#include <linux/arm-smccc.h>
-
-#include <asm/checksum.h>
-#include <asm/ftrace.h>

followed by prototypes for the GCC internal functions, and:

-extern void fpundefinstr(void);
-
-void mmioset(void *, unsigned int, size_t);
-void mmiocpy(void *, const void *, size_t);

So, the asm-prototypes.h approach is just the same, only that we now
have a bunch of prototypes in a header file, and the EXPORT_SYMBOL()s
in the assembly files.

As the C prototypes are remote from the definitions, it means that
the C prototypes are going to get forgotten about in exactly the same
way that armksyms.c would've been forgotten about too.

It _is_ worse than that though - with the armksyms.c approach, if the
assembly code for it is removed, you get a build error reminding you
to remove the export (and prototype).  With this approach, you get no
reminder to touch asm-prototypes.h.

It's also error prone for another reason - adding a new assembly level
export, if you forget to add it to asm-prototypes.h, we're back into
the problem we have right now with MODVERSIONS breaking.

So, I still think the whole approach is wrong - it's added extra
fragility that wasn't there with the armksyms.c approach.

-- 
RMK's Patch system: http://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line: currently at 9.6Mbps down 400kbps up
according to speedtest.net.

Re: [PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Russell King - ARM Linux]

On Wed, Nov 23, 2016 at 09:33:32AM +0000, Russell King - ARM Linux wrote:
> I don't see it makes any difference - the armksyms.c originally had
> the same:
> 
> -#include <linux/export.h>
> -#include <linux/sched.h>
> -#include <linux/string.h>
> -#include <linux/cryptohash.h>
> -#include <linux/delay.h>
> -#include <linux/in6.h>
> -#include <linux/syscalls.h>
> -#include <linux/uaccess.h>
> -#include <linux/io.h>
> -#include <linux/arm-smccc.h>
> -
> -#include <asm/checksum.h>
> -#include <asm/ftrace.h>
> 
> followed by prototypes for the GCC internal functions, and:
> 
> -extern void fpundefinstr(void);
> -
> -void mmioset(void *, unsigned int, size_t);
> -void mmiocpy(void *, const void *, size_t);
> 
> So, the asm-prototypes.h approach is just the same, only that we now
> have a bunch of prototypes in a header file, and the EXPORT_SYMBOL()s
> in the assembly files.
> 
> As the C prototypes are remote from the definitions, it means that
> the C prototypes are going to get forgotten about in exactly the same
> way that armksyms.c would've been forgotten about too.
> 
> It _is_ worse than that though - with the armksyms.c approach, if the
> assembly code for it is removed, you get a build error reminding you
> to remove the export (and prototype).  With this approach, you get no
> reminder to touch asm-prototypes.h.
> 
> It's also error prone for another reason - adding a new assembly level
> export, if you forget to add it to asm-prototypes.h, we're back into
> the problem we have right now with MODVERSIONS breaking.
> 
> So, I still think the whole approach is wrong - it's added extra
> fragility that wasn't there with the armksyms.c approach.

Here's what the diffstat and patch looks like when you combine the
original commit and the three fixes.  The LoC delta of 25 lines can
be accounted for as deleted commentry.  So, I think (as I've detailed
above) that the _technical_ benefit of the approach is very low.

If we want to move the exports into assembly files, I've no problem
with that, provided we can do it better than this - and by better I
mean not creating the fragile asm-prototypes.h which divorses the
prototypes from everything else.

Looking at _this_ patch, there's also an issue with the __raw_*
functions which are conditionally exported, which isn't taken account
of in these fixes.  We currently build most of the io-*.S files whether
we use them or not, and rely on the linker's archive processing, along
with the ifdefs in armksyms to omit them from the kernel image.  This
won't happen with this new approach.

So, I'm reverting the original commit today, because there's no clear
benefit, it's fragile, and I think there's still a few corner cases
that need to be fixed.  We can revisit when we have more time to
properly review and test these changes, rather than at what is now the
11th hour.

 arch/arm/include/asm/Kbuild           |   1 +
 arch/arm/include/asm/asm-prototypes.h |  34 +++++++
 arch/arm/kernel/Makefile              |   2 +-
 arch/arm/kernel/armksyms.c            | 183 ----------------------------------
 arch/arm/kernel/entry-ftrace.S        |   3 +
 arch/arm/kernel/head.S                |   3 +
 arch/arm/kernel/io.c                  |   7 ++
 arch/arm/kernel/smccc-call.S          |   3 +
 arch/arm/lib/ashldi3.S                |   3 +
 arch/arm/lib/ashrdi3.S                |   3 +
 arch/arm/lib/bswapsdi2.S              |   3 +
 arch/arm/lib/changebit.S              |   3 +
 arch/arm/lib/clear_user.S             |   4 +
 arch/arm/lib/clearbit.S               |   3 +
 arch/arm/lib/copy_from_user.S         |   2 +
 arch/arm/lib/copy_page.S              |   2 +
 arch/arm/lib/copy_to_user.S           |   4 +
 arch/arm/lib/csumipv6.S               |   3 +-
 arch/arm/lib/csumpartial.S            |   2 +
 arch/arm/lib/csumpartialcopy.S        |   1 +
 arch/arm/lib/csumpartialcopygeneric.S |   2 +
 arch/arm/lib/csumpartialcopyuser.S    |   1 +
 arch/arm/lib/delay.c                  |   2 +
 arch/arm/lib/div64.S                  |   2 +
 arch/arm/lib/findbit.S                |   9 ++
 arch/arm/lib/getuser.S                |   9 ++
 arch/arm/lib/io-readsb.S              |   2 +
 arch/arm/lib/io-readsl.S              |   2 +
 arch/arm/lib/io-readsw-armv3.S        |   3 +-
 arch/arm/lib/io-readsw-armv4.S        |   2 +
 arch/arm/lib/io-writesb.S             |   2 +
 arch/arm/lib/io-writesl.S             |   2 +
 arch/arm/lib/io-writesw-armv3.S       |   2 +
 arch/arm/lib/io-writesw-armv4.S       |   2 +
 arch/arm/lib/lib1funcs.S              |   9 ++
 arch/arm/lib/lshrdi3.S                |   3 +
 arch/arm/lib/memchr.S                 |   2 +
 arch/arm/lib/memcpy.S                 |   2 +
 arch/arm/lib/memmove.S                |   2 +
 arch/arm/lib/memset.S                 |   2 +
 arch/arm/lib/memzero.S                |   2 +
 arch/arm/lib/muldi3.S                 |   3 +
 arch/arm/lib/putuser.S                |   5 +
 arch/arm/lib/setbit.S                 |   3 +
 arch/arm/lib/strchr.S                 |   2 +
 arch/arm/lib/strrchr.S                |   2 +
 arch/arm/lib/testchangebit.S          |   3 +
 arch/arm/lib/testclearbit.S           |   3 +
 arch/arm/lib/testsetbit.S             |   3 +
 arch/arm/lib/uaccess_with_memcpy.c    |   3 +
 arch/arm/lib/ucmpdi2.S                |   3 +
 arch/arm/mach-imx/Makefile            |   1 -
 arch/arm/mach-imx/ssi-fiq-ksym.c      |  20 ----
 arch/arm/mach-imx/ssi-fiq.S           |   7 +-
 54 files changed, 183 insertions(+), 208 deletions(-)

diff --git a/arch/arm/include/asm/Kbuild b/arch/arm/include/asm/Kbuild
index 55e0e3ea9cb6..0745538b26d3 100644
--- a/arch/arm/include/asm/Kbuild
+++ b/arch/arm/include/asm/Kbuild
@@ -8,6 +8,7 @@ generic-y += early_ioremap.h
 generic-y += emergency-restart.h
 generic-y += errno.h
 generic-y += exec.h
+generic-y += export.h
 generic-y += ioctl.h
 generic-y += ipcbuf.h
 generic-y += irq_regs.h
diff --git a/arch/arm/include/asm/asm-prototypes.h b/arch/arm/include/asm/asm-prototypes.h
new file mode 100644
index 000000000000..04e5616a7b15
--- /dev/null
+++ b/arch/arm/include/asm/asm-prototypes.h
@@ -0,0 +1,34 @@
+#include <linux/arm-smccc.h>
+#include <linux/bitops.h>
+#include <linux/ftrace.h>
+#include <linux/io.h>
+#include <linux/platform_data/asoc-imx-ssi.h>
+#include <linux/string.h>
+#include <linux/uaccess.h>
+
+#include <asm/checksum.h>
+#include <asm/div64.h>
+#include <asm/memory.h>
+
+extern void __aeabi_idivmod(void);
+extern void __aeabi_idiv(void);
+extern void __aeabi_lasr(void);
+extern void __aeabi_llsl(void);
+extern void __aeabi_llsr(void);
+extern void __aeabi_lmul(void);
+extern void __aeabi_uidivmod(void);
+extern void __aeabi_uidiv(void);
+extern void __aeabi_ulcmp(void);
+
+extern void __ashldi3(void);
+extern void __ashrdi3(void);
+extern void __bswapdi2(void);
+extern void __bswapsi2(void);
+extern void __divsi3(void);
+extern void __do_div64(void);
+extern void __lshrdi3(void);
+extern void __modsi3(void);
+extern void __muldi3(void);
+extern void __ucmpdi2(void);
+extern void __udivsi3(void);
+extern void __umodsi3(void);
diff --git a/arch/arm/kernel/Makefile b/arch/arm/kernel/Makefile
index ad325a8c7e1e..68c2c097cffe 100644
--- a/arch/arm/kernel/Makefile
+++ b/arch/arm/kernel/Makefile
@@ -33,7 +33,7 @@ endif
 obj-$(CONFIG_CPU_IDLE)		+= cpuidle.o
 obj-$(CONFIG_ISA_DMA_API)	+= dma.o
 obj-$(CONFIG_FIQ)		+= fiq.o fiqasm.o
-obj-$(CONFIG_MODULES)		+= armksyms.o module.o
+obj-$(CONFIG_MODULES)		+= module.o
 obj-$(CONFIG_ARM_MODULE_PLTS)	+= module-plts.o
 obj-$(CONFIG_ISA_DMA)		+= dma-isa.o
 obj-$(CONFIG_PCI)		+= bios32.o isa.o
diff --git a/arch/arm/kernel/armksyms.c b/arch/arm/kernel/armksyms.c
deleted file mode 100644
index 7e45f69a0ddc..000000000000
--- a/arch/arm/kernel/armksyms.c
+++ /dev/null
@@ -1,183 +0,0 @@
-/*
- *  linux/arch/arm/kernel/armksyms.c
- *
- *  Copyright (C) 2000 Russell King
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- */
-#include <linux/export.h>
-#include <linux/sched.h>
-#include <linux/string.h>
-#include <linux/cryptohash.h>
-#include <linux/delay.h>
-#include <linux/in6.h>
-#include <linux/syscalls.h>
-#include <linux/uaccess.h>
-#include <linux/io.h>
-#include <linux/arm-smccc.h>
-
-#include <asm/checksum.h>
-#include <asm/ftrace.h>
-
-/*
- * libgcc functions - functions that are used internally by the
- * compiler...  (prototypes are not correct though, but that
- * doesn't really matter since they're not versioned).
- */
-extern void __ashldi3(void);
-extern void __ashrdi3(void);
-extern void __divsi3(void);
-extern void __lshrdi3(void);
-extern void __modsi3(void);
-extern void __muldi3(void);
-extern void __ucmpdi2(void);
-extern void __udivsi3(void);
-extern void __umodsi3(void);
-extern void __do_div64(void);
-extern void __bswapsi2(void);
-extern void __bswapdi2(void);
-
-extern void __aeabi_idiv(void);
-extern void __aeabi_idivmod(void);
-extern void __aeabi_lasr(void);
-extern void __aeabi_llsl(void);
-extern void __aeabi_llsr(void);
-extern void __aeabi_lmul(void);
-extern void __aeabi_uidiv(void);
-extern void __aeabi_uidivmod(void);
-extern void __aeabi_ulcmp(void);
-
-extern void fpundefinstr(void);
-
-void mmioset(void *, unsigned int, size_t);
-void mmiocpy(void *, const void *, size_t);
-
-	/* platform dependent support */
-EXPORT_SYMBOL(arm_delay_ops);
-
-	/* networking */
-EXPORT_SYMBOL(csum_partial);
-EXPORT_SYMBOL(csum_partial_copy_from_user);
-EXPORT_SYMBOL(csum_partial_copy_nocheck);
-EXPORT_SYMBOL(__csum_ipv6_magic);
-
-	/* io */
-#ifndef __raw_readsb
-EXPORT_SYMBOL(__raw_readsb);
-#endif
-#ifndef __raw_readsw
-EXPORT_SYMBOL(__raw_readsw);
-#endif
-#ifndef __raw_readsl
-EXPORT_SYMBOL(__raw_readsl);
-#endif
-#ifndef __raw_writesb
-EXPORT_SYMBOL(__raw_writesb);
-#endif
-#ifndef __raw_writesw
-EXPORT_SYMBOL(__raw_writesw);
-#endif
-#ifndef __raw_writesl
-EXPORT_SYMBOL(__raw_writesl);
-#endif
-
-	/* string / mem functions */
-EXPORT_SYMBOL(strchr);
-EXPORT_SYMBOL(strrchr);
-EXPORT_SYMBOL(memset);
-EXPORT_SYMBOL(memcpy);
-EXPORT_SYMBOL(memmove);
-EXPORT_SYMBOL(memchr);
-EXPORT_SYMBOL(__memzero);
-
-EXPORT_SYMBOL(mmioset);
-EXPORT_SYMBOL(mmiocpy);
-
-#ifdef CONFIG_MMU
-EXPORT_SYMBOL(copy_page);
-
-EXPORT_SYMBOL(arm_copy_from_user);
-EXPORT_SYMBOL(arm_copy_to_user);
-EXPORT_SYMBOL(arm_clear_user);
-
-EXPORT_SYMBOL(__get_user_1);
-EXPORT_SYMBOL(__get_user_2);
-EXPORT_SYMBOL(__get_user_4);
-EXPORT_SYMBOL(__get_user_8);
-
-#ifdef __ARMEB__
-EXPORT_SYMBOL(__get_user_64t_1);
-EXPORT_SYMBOL(__get_user_64t_2);
-EXPORT_SYMBOL(__get_user_64t_4);
-EXPORT_SYMBOL(__get_user_32t_8);
-#endif
-
-EXPORT_SYMBOL(__put_user_1);
-EXPORT_SYMBOL(__put_user_2);
-EXPORT_SYMBOL(__put_user_4);
-EXPORT_SYMBOL(__put_user_8);
-#endif
-
-	/* gcc lib functions */
-EXPORT_SYMBOL(__ashldi3);
-EXPORT_SYMBOL(__ashrdi3);
-EXPORT_SYMBOL(__divsi3);
-EXPORT_SYMBOL(__lshrdi3);
-EXPORT_SYMBOL(__modsi3);
-EXPORT_SYMBOL(__muldi3);
-EXPORT_SYMBOL(__ucmpdi2);
-EXPORT_SYMBOL(__udivsi3);
-EXPORT_SYMBOL(__umodsi3);
-EXPORT_SYMBOL(__do_div64);
-EXPORT_SYMBOL(__bswapsi2);
-EXPORT_SYMBOL(__bswapdi2);
-
-#ifdef CONFIG_AEABI
-EXPORT_SYMBOL(__aeabi_idiv);
-EXPORT_SYMBOL(__aeabi_idivmod);
-EXPORT_SYMBOL(__aeabi_lasr);
-EXPORT_SYMBOL(__aeabi_llsl);
-EXPORT_SYMBOL(__aeabi_llsr);
-EXPORT_SYMBOL(__aeabi_lmul);
-EXPORT_SYMBOL(__aeabi_uidiv);
-EXPORT_SYMBOL(__aeabi_uidivmod);
-EXPORT_SYMBOL(__aeabi_ulcmp);
-#endif
-
-	/* bitops */
-EXPORT_SYMBOL(_set_bit);
-EXPORT_SYMBOL(_test_and_set_bit);
-EXPORT_SYMBOL(_clear_bit);
-EXPORT_SYMBOL(_test_and_clear_bit);
-EXPORT_SYMBOL(_change_bit);
-EXPORT_SYMBOL(_test_and_change_bit);
-EXPORT_SYMBOL(_find_first_zero_bit_le);
-EXPORT_SYMBOL(_find_next_zero_bit_le);
-EXPORT_SYMBOL(_find_first_bit_le);
-EXPORT_SYMBOL(_find_next_bit_le);
-
-#ifdef __ARMEB__
-EXPORT_SYMBOL(_find_first_zero_bit_be);
-EXPORT_SYMBOL(_find_next_zero_bit_be);
-EXPORT_SYMBOL(_find_first_bit_be);
-EXPORT_SYMBOL(_find_next_bit_be);
-#endif
-
-#ifdef CONFIG_FUNCTION_TRACER
-#ifdef CONFIG_OLD_MCOUNT
-EXPORT_SYMBOL(mcount);
-#endif
-EXPORT_SYMBOL(__gnu_mcount_nc);
-#endif
-
-#ifdef CONFIG_ARM_PATCH_PHYS_VIRT
-EXPORT_SYMBOL(__pv_phys_pfn_offset);
-EXPORT_SYMBOL(__pv_offset);
-#endif
-
-#ifdef CONFIG_HAVE_ARM_SMCCC
-EXPORT_SYMBOL(arm_smccc_smc);
-EXPORT_SYMBOL(arm_smccc_hvc);
-#endif
diff --git a/arch/arm/kernel/entry-ftrace.S b/arch/arm/kernel/entry-ftrace.S
index c73c4030ca5d..b629d3f11c3d 100644
--- a/arch/arm/kernel/entry-ftrace.S
+++ b/arch/arm/kernel/entry-ftrace.S
@@ -7,6 +7,7 @@
 #include <asm/assembler.h>
 #include <asm/ftrace.h>
 #include <asm/unwind.h>
+#include <asm/export.h>
 
 #include "entry-header.S"
 
@@ -153,6 +154,7 @@ ENTRY(mcount)
 	__mcount _old
 #endif
 ENDPROC(mcount)
+EXPORT_SYMBOL(mcount)
 
 #ifdef CONFIG_DYNAMIC_FTRACE
 ENTRY(ftrace_caller_old)
@@ -205,6 +207,7 @@ UNWIND(.fnstart)
 #endif
 UNWIND(.fnend)
 ENDPROC(__gnu_mcount_nc)
+EXPORT_SYMBOL(__gnu_mcount_nc)
 
 #ifdef CONFIG_DYNAMIC_FTRACE
 ENTRY(ftrace_caller)
diff --git a/arch/arm/kernel/head.S b/arch/arm/kernel/head.S
index 04286fd9e09c..f41cee4c5746 100644
--- a/arch/arm/kernel/head.S
+++ b/arch/arm/kernel/head.S
@@ -22,6 +22,7 @@
 #include <asm/memory.h>
 #include <asm/thread_info.h>
 #include <asm/pgtable.h>
+#include <asm/export.h>
 
 #if defined(CONFIG_DEBUG_LL) && !defined(CONFIG_DEBUG_SEMIHOSTING)
 #include CONFIG_DEBUG_LL_INCLUDE
@@ -727,6 +728,8 @@ ENDPROC(fixup_pv_table)
 __pv_offset:
 	.quad	0
 	.size	__pv_offset, . -__pv_offset
+EXPORT_SYMBOL(__pv_phys_pfn_offset)
+EXPORT_SYMBOL(__pv_offset)
 #endif
 
 #include "head-common.S"
diff --git a/arch/arm/kernel/io.c b/arch/arm/kernel/io.c
index eedefe050022..c74746997626 100644
--- a/arch/arm/kernel/io.c
+++ b/arch/arm/kernel/io.c
@@ -82,3 +82,10 @@ void _memset_io(volatile void __iomem *dst, int c, size_t count)
 	}
 }
 EXPORT_SYMBOL(_memset_io);
+
+/* can't export them from memcpy.S/memset.S because of hidden declaration */
+void mmioset(void __iomem *addr, unsigned int c, size_t n);
+EXPORT_SYMBOL(mmioset);
+
+void mmiocpy(void *dest, const void __iomem *src, size_t n);
+EXPORT_SYMBOL(mmiocpy);
diff --git a/arch/arm/kernel/smccc-call.S b/arch/arm/kernel/smccc-call.S
index 2e48b674aab1..37669e7e13af 100644
--- a/arch/arm/kernel/smccc-call.S
+++ b/arch/arm/kernel/smccc-call.S
@@ -16,6 +16,7 @@
 #include <asm/opcodes-sec.h>
 #include <asm/opcodes-virt.h>
 #include <asm/unwind.h>
+#include <asm/export.h>
 
 	/*
 	 * Wrap c macros in asm macros to delay expansion until after the
@@ -51,6 +52,7 @@ UNWIND(	.fnend)
 ENTRY(arm_smccc_smc)
 	SMCCC SMCCC_SMC
 ENDPROC(arm_smccc_smc)
+EXPORT_SYMBOL(arm_smccc_smc)
 
 /*
  * void smccc_hvc(unsigned long a0, unsigned long a1, unsigned long a2,
@@ -60,3 +62,4 @@ ENDPROC(arm_smccc_smc)
 ENTRY(arm_smccc_hvc)
 	SMCCC SMCCC_HVC
 ENDPROC(arm_smccc_hvc)
+EXPORT_SYMBOL(arm_smccc_hvc)
diff --git a/arch/arm/lib/ashldi3.S b/arch/arm/lib/ashldi3.S
index b05e95840651..a7e7de89bd75 100644
--- a/arch/arm/lib/ashldi3.S
+++ b/arch/arm/lib/ashldi3.S
@@ -28,6 +28,7 @@ Boston, MA 02110-1301, USA.  */
 
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 #ifdef __ARMEB__
 #define al r1
@@ -52,3 +53,5 @@ ENTRY(__aeabi_llsl)
 
 ENDPROC(__ashldi3)
 ENDPROC(__aeabi_llsl)
+EXPORT_SYMBOL(__ashldi3)
+EXPORT_SYMBOL(__aeabi_llsl)
diff --git a/arch/arm/lib/ashrdi3.S b/arch/arm/lib/ashrdi3.S
index 275d7d2341a4..490336e42518 100644
--- a/arch/arm/lib/ashrdi3.S
+++ b/arch/arm/lib/ashrdi3.S
@@ -28,6 +28,7 @@ Boston, MA 02110-1301, USA.  */
 
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 #ifdef __ARMEB__
 #define al r1
@@ -52,3 +53,5 @@ ENTRY(__aeabi_lasr)
 
 ENDPROC(__ashrdi3)
 ENDPROC(__aeabi_lasr)
+EXPORT_SYMBOL(__ashrdi3)
+EXPORT_SYMBOL(__aeabi_lasr)
diff --git a/arch/arm/lib/bswapsdi2.S b/arch/arm/lib/bswapsdi2.S
index 07cda737bb11..f05f78247304 100644
--- a/arch/arm/lib/bswapsdi2.S
+++ b/arch/arm/lib/bswapsdi2.S
@@ -1,5 +1,6 @@
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 #if __LINUX_ARM_ARCH__ >= 6
 ENTRY(__bswapsi2)
@@ -35,3 +36,5 @@ ENTRY(__bswapdi2)
 	ret lr
 ENDPROC(__bswapdi2)
 #endif
+EXPORT_SYMBOL(__bswapsi2)
+EXPORT_SYMBOL(__bswapdi2)
diff --git a/arch/arm/lib/changebit.S b/arch/arm/lib/changebit.S
index f4027862172f..1cfdb138d2d9 100644
--- a/arch/arm/lib/changebit.S
+++ b/arch/arm/lib/changebit.S
@@ -9,7 +9,10 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 #include "bitops.h"
                 .text
 
 bitop	_change_bit, eor
+
+EXPORT_SYMBOL(_change_bit)
diff --git a/arch/arm/lib/clear_user.S b/arch/arm/lib/clear_user.S
index e936352ccb00..b566154f5cf4 100644
--- a/arch/arm/lib/clear_user.S
+++ b/arch/arm/lib/clear_user.S
@@ -10,6 +10,7 @@
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 #include <asm/unwind.h>
+#include <asm/export.h>
 
 		.text
 
@@ -50,6 +51,9 @@ USER(		strnebt	r2, [r0])
 UNWIND(.fnend)
 ENDPROC(arm_clear_user)
 ENDPROC(__clear_user_std)
+#ifndef CONFIG_UACCESS_WITH_MEMCPY
+EXPORT_SYMBOL(arm_clear_user)
+#endif
 
 		.pushsection .text.fixup,"ax"
 		.align	0
diff --git a/arch/arm/lib/clearbit.S b/arch/arm/lib/clearbit.S
index f6b75fb64d30..e901ca5af0df 100644
--- a/arch/arm/lib/clearbit.S
+++ b/arch/arm/lib/clearbit.S
@@ -9,7 +9,10 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 #include "bitops.h"
                 .text
 
 bitop	_clear_bit, bic
+
+EXPORT_SYMBOL(_clear_bit)
diff --git a/arch/arm/lib/copy_from_user.S b/arch/arm/lib/copy_from_user.S
index 1512bebfbf1b..f549c57ea435 100644
--- a/arch/arm/lib/copy_from_user.S
+++ b/arch/arm/lib/copy_from_user.S
@@ -13,6 +13,7 @@
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 #include <asm/unwind.h>
+#include <asm/export.h>
 
 /*
  * Prototype:
@@ -94,6 +95,7 @@ ENTRY(arm_copy_from_user)
 #include "copy_template.S"
 
 ENDPROC(arm_copy_from_user)
+EXPORT_SYMBOL(arm_copy_from_user)
 
 	.pushsection .fixup,"ax"
 	.align 0
diff --git a/arch/arm/lib/copy_page.S b/arch/arm/lib/copy_page.S
index 6ee2f6706f86..d97851d4af7a 100644
--- a/arch/arm/lib/copy_page.S
+++ b/arch/arm/lib/copy_page.S
@@ -13,6 +13,7 @@
 #include <asm/assembler.h>
 #include <asm/asm-offsets.h>
 #include <asm/cache.h>
+#include <asm/export.h>
 
 #define COPY_COUNT (PAGE_SZ / (2 * L1_CACHE_BYTES) PLD( -1 ))
 
@@ -45,3 +46,4 @@ ENTRY(copy_page)
 	PLD(	beq	2b			)
 		ldmfd	sp!, {r4, pc}			@	3
 ENDPROC(copy_page)
+EXPORT_SYMBOL(copy_page)
diff --git a/arch/arm/lib/copy_to_user.S b/arch/arm/lib/copy_to_user.S
index caf5019d8161..592c179112d1 100644
--- a/arch/arm/lib/copy_to_user.S
+++ b/arch/arm/lib/copy_to_user.S
@@ -13,6 +13,7 @@
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 #include <asm/unwind.h>
+#include <asm/export.h>
 
 /*
  * Prototype:
@@ -99,6 +100,9 @@ WEAK(arm_copy_to_user)
 
 ENDPROC(arm_copy_to_user)
 ENDPROC(__copy_to_user_std)
+#ifndef CONFIG_UACCESS_WITH_MEMCPY
+EXPORT_SYMBOL(arm_copy_to_user)
+#endif
 
 	.pushsection .text.fixup,"ax"
 	.align 0
diff --git a/arch/arm/lib/csumipv6.S b/arch/arm/lib/csumipv6.S
index 3ac6ef01bc43..68603b5ee537 100644
--- a/arch/arm/lib/csumipv6.S
+++ b/arch/arm/lib/csumipv6.S
@@ -9,6 +9,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 		.text
 
@@ -30,4 +31,4 @@ ENTRY(__csum_ipv6_magic)
 		adcs	r0, r0, #0
 		ldmfd	sp!, {pc}
 ENDPROC(__csum_ipv6_magic)
-
+EXPORT_SYMBOL(__csum_ipv6_magic)
diff --git a/arch/arm/lib/csumpartial.S b/arch/arm/lib/csumpartial.S
index 984e0f29d548..830b20e81c37 100644
--- a/arch/arm/lib/csumpartial.S
+++ b/arch/arm/lib/csumpartial.S
@@ -9,6 +9,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 		.text
 
@@ -140,3 +141,4 @@ ENTRY(csum_partial)
 		bne	4b
 		b	.Lless4
 ENDPROC(csum_partial)
+EXPORT_SYMBOL(csum_partial)
diff --git a/arch/arm/lib/csumpartialcopy.S b/arch/arm/lib/csumpartialcopy.S
index d03fc71fc88c..9c3383fed129 100644
--- a/arch/arm/lib/csumpartialcopy.S
+++ b/arch/arm/lib/csumpartialcopy.S
@@ -49,5 +49,6 @@
 
 #define FN_ENTRY	ENTRY(csum_partial_copy_nocheck)
 #define FN_EXIT		ENDPROC(csum_partial_copy_nocheck)
+#define FN_EXPORT	EXPORT_SYMBOL(csum_partial_copy_nocheck)
 
 #include "csumpartialcopygeneric.S"
diff --git a/arch/arm/lib/csumpartialcopygeneric.S b/arch/arm/lib/csumpartialcopygeneric.S
index 10b45909610c..8b94d20e51d1 100644
--- a/arch/arm/lib/csumpartialcopygeneric.S
+++ b/arch/arm/lib/csumpartialcopygeneric.S
@@ -8,6 +8,7 @@
  * published by the Free Software Foundation.
  */
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 /*
  * unsigned int
@@ -331,3 +332,4 @@ FN_ENTRY
 		mov	r5, r4, get_byte_1
 		b	.Lexit
 FN_EXIT
+FN_EXPORT
diff --git a/arch/arm/lib/csumpartialcopyuser.S b/arch/arm/lib/csumpartialcopyuser.S
index 1712f132b80d..5d495edf3d83 100644
--- a/arch/arm/lib/csumpartialcopyuser.S
+++ b/arch/arm/lib/csumpartialcopyuser.S
@@ -73,6 +73,7 @@
 
 #define FN_ENTRY	ENTRY(csum_partial_copy_from_user)
 #define FN_EXIT		ENDPROC(csum_partial_copy_from_user)
+#define FN_EXPORT	EXPORT_SYMBOL(csum_partial_copy_from_user)
 
 #include "csumpartialcopygeneric.S"
 
diff --git a/arch/arm/lib/delay.c b/arch/arm/lib/delay.c
index 8044591dca72..e60ce1549759 100644
--- a/arch/arm/lib/delay.c
+++ b/arch/arm/lib/delay.c
@@ -24,6 +24,7 @@
 #include <linux/init.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
+#include <linux/export.h>
 #include <linux/timex.h>
 
 /*
@@ -34,6 +35,7 @@ struct arm_delay_ops arm_delay_ops = {
 	.const_udelay	= __loop_const_udelay,
 	.udelay		= __loop_udelay,
 };
+EXPORT_SYMBOL(arm_delay_ops);
 
 static const struct delay_timer *delay_timer;
 static bool delay_calibrated;
diff --git a/arch/arm/lib/div64.S b/arch/arm/lib/div64.S
index a9eafe4981eb..0c9e1c18fc9e 100644
--- a/arch/arm/lib/div64.S
+++ b/arch/arm/lib/div64.S
@@ -15,6 +15,7 @@
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 #include <asm/unwind.h>
+#include <asm/export.h>
 
 #ifdef __ARMEB__
 #define xh r0
@@ -210,3 +211,4 @@ UNWIND(.save {lr})
 
 UNWIND(.fnend)
 ENDPROC(__do_div64)
+EXPORT_SYMBOL(__do_div64)
diff --git a/arch/arm/lib/findbit.S b/arch/arm/lib/findbit.S
index 7848780e8834..26302b8cd38f 100644
--- a/arch/arm/lib/findbit.S
+++ b/arch/arm/lib/findbit.S
@@ -15,6 +15,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
                 .text
 
 /*
@@ -37,6 +38,7 @@ ENTRY(_find_first_zero_bit_le)
 3:		mov	r0, r1			@ no free bits
 		ret	lr
 ENDPROC(_find_first_zero_bit_le)
+EXPORT_SYMBOL(_find_first_zero_bit_le)
 
 /*
  * Purpose  : Find next 'zero' bit
@@ -57,6 +59,7 @@ ENTRY(_find_next_zero_bit_le)
 		add	r2, r2, #1		@ align bit pointer
 		b	2b			@ loop for next bit
 ENDPROC(_find_next_zero_bit_le)
+EXPORT_SYMBOL(_find_next_zero_bit_le)
 
 /*
  * Purpose  : Find a 'one' bit
@@ -78,6 +81,7 @@ ENTRY(_find_first_bit_le)
 3:		mov	r0, r1			@ no free bits
 		ret	lr
 ENDPROC(_find_first_bit_le)
+EXPORT_SYMBOL(_find_first_bit_le)
 
 /*
  * Purpose  : Find next 'one' bit
@@ -97,6 +101,7 @@ ENTRY(_find_next_bit_le)
 		add	r2, r2, #1		@ align bit pointer
 		b	2b			@ loop for next bit
 ENDPROC(_find_next_bit_le)
+EXPORT_SYMBOL(_find_next_bit_le)
 
 #ifdef __ARMEB__
 
@@ -116,6 +121,7 @@ ENTRY(_find_first_zero_bit_be)
 3:		mov	r0, r1			@ no free bits
 		ret	lr
 ENDPROC(_find_first_zero_bit_be)
+EXPORT_SYMBOL(_find_first_zero_bit_be)
 
 ENTRY(_find_next_zero_bit_be)
 		teq	r1, #0
@@ -133,6 +139,7 @@ ENTRY(_find_next_zero_bit_be)
 		add	r2, r2, #1		@ align bit pointer
 		b	2b			@ loop for next bit
 ENDPROC(_find_next_zero_bit_be)
+EXPORT_SYMBOL(_find_next_zero_bit_be)
 
 ENTRY(_find_first_bit_be)
 		teq	r1, #0
@@ -150,6 +157,7 @@ ENTRY(_find_first_bit_be)
 3:		mov	r0, r1			@ no free bits
 		ret	lr
 ENDPROC(_find_first_bit_be)
+EXPORT_SYMBOL(_find_first_bit_be)
 
 ENTRY(_find_next_bit_be)
 		teq	r1, #0
@@ -166,6 +174,7 @@ ENTRY(_find_next_bit_be)
 		add	r2, r2, #1		@ align bit pointer
 		b	2b			@ loop for next bit
 ENDPROC(_find_next_bit_be)
+EXPORT_SYMBOL(_find_next_bit_be)
 
 #endif
 
diff --git a/arch/arm/lib/getuser.S b/arch/arm/lib/getuser.S
index 8ecfd15c3a02..9d09a38e73af 100644
--- a/arch/arm/lib/getuser.S
+++ b/arch/arm/lib/getuser.S
@@ -31,6 +31,7 @@
 #include <asm/assembler.h>
 #include <asm/errno.h>
 #include <asm/domain.h>
+#include <asm/export.h>
 
 ENTRY(__get_user_1)
 	check_uaccess r0, 1, r1, r2, __get_user_bad
@@ -38,6 +39,7 @@ ENTRY(__get_user_1)
 	mov	r0, #0
 	ret	lr
 ENDPROC(__get_user_1)
+EXPORT_SYMBOL(__get_user_1)
 
 ENTRY(__get_user_2)
 	check_uaccess r0, 2, r1, r2, __get_user_bad
@@ -58,6 +60,7 @@ rb	.req	r0
 	mov	r0, #0
 	ret	lr
 ENDPROC(__get_user_2)
+EXPORT_SYMBOL(__get_user_2)
 
 ENTRY(__get_user_4)
 	check_uaccess r0, 4, r1, r2, __get_user_bad
@@ -65,6 +68,7 @@ ENTRY(__get_user_4)
 	mov	r0, #0
 	ret	lr
 ENDPROC(__get_user_4)
+EXPORT_SYMBOL(__get_user_4)
 
 ENTRY(__get_user_8)
 	check_uaccess r0, 8, r1, r2, __get_user_bad
@@ -78,6 +82,7 @@ ENTRY(__get_user_8)
 	mov	r0, #0
 	ret	lr
 ENDPROC(__get_user_8)
+EXPORT_SYMBOL(__get_user_8)
 
 #ifdef __ARMEB__
 ENTRY(__get_user_32t_8)
@@ -91,6 +96,7 @@ ENTRY(__get_user_32t_8)
 	mov	r0, #0
 	ret	lr
 ENDPROC(__get_user_32t_8)
+EXPORT_SYMBOL(__get_user_32t_8)
 
 ENTRY(__get_user_64t_1)
 	check_uaccess r0, 1, r1, r2, __get_user_bad8
@@ -98,6 +104,7 @@ ENTRY(__get_user_64t_1)
 	mov	r0, #0
 	ret	lr
 ENDPROC(__get_user_64t_1)
+EXPORT_SYMBOL(__get_user_64t_1)
 
 ENTRY(__get_user_64t_2)
 	check_uaccess r0, 2, r1, r2, __get_user_bad8
@@ -114,6 +121,7 @@ rb	.req	r0
 	mov	r0, #0
 	ret	lr
 ENDPROC(__get_user_64t_2)
+EXPORT_SYMBOL(__get_user_64t_2)
 
 ENTRY(__get_user_64t_4)
 	check_uaccess r0, 4, r1, r2, __get_user_bad8
@@ -121,6 +129,7 @@ ENTRY(__get_user_64t_4)
 	mov	r0, #0
 	ret	lr
 ENDPROC(__get_user_64t_4)
+EXPORT_SYMBOL(__get_user_64t_4)
 #endif
 
 __get_user_bad8:
diff --git a/arch/arm/lib/io-readsb.S b/arch/arm/lib/io-readsb.S
index c31b2f3153f1..3dff7a3a2aef 100644
--- a/arch/arm/lib/io-readsb.S
+++ b/arch/arm/lib/io-readsb.S
@@ -9,6 +9,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 .Linsb_align:	rsb	ip, ip, #4
 		cmp	ip, r2
@@ -121,3 +122,4 @@ ENTRY(__raw_readsb)
 
 		ldmfd	sp!, {r4 - r6, pc}
 ENDPROC(__raw_readsb)
+EXPORT_SYMBOL(__raw_readsb)
diff --git a/arch/arm/lib/io-readsl.S b/arch/arm/lib/io-readsl.S
index 2ed86fa5465f..bfd39682325b 100644
--- a/arch/arm/lib/io-readsl.S
+++ b/arch/arm/lib/io-readsl.S
@@ -9,6 +9,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 ENTRY(__raw_readsl)
 		teq	r2, #0		@ do we have to check for the zero len?
@@ -77,3 +78,4 @@ ENTRY(__raw_readsl)
 		strb	r3, [r1, #0]
 		ret	lr
 ENDPROC(__raw_readsl)
+EXPORT_SYMBOL(__raw_readsl)
diff --git a/arch/arm/lib/io-readsw-armv3.S b/arch/arm/lib/io-readsw-armv3.S
index 413da9914529..b3af3db6caac 100644
--- a/arch/arm/lib/io-readsw-armv3.S
+++ b/arch/arm/lib/io-readsw-armv3.S
@@ -9,6 +9,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 .Linsw_bad_alignment:
 		adr	r0, .Linsw_bad_align_msg
@@ -103,4 +104,4 @@ ENTRY(__raw_readsw)
 
 		ldmfd	sp!, {r4, r5, r6, pc}
 
-
+EXPORT_SYMBOL(__raw_readsw)
diff --git a/arch/arm/lib/io-readsw-armv4.S b/arch/arm/lib/io-readsw-armv4.S
index d9a45e9692ae..3c7a7a40b33e 100644
--- a/arch/arm/lib/io-readsw-armv4.S
+++ b/arch/arm/lib/io-readsw-armv4.S
@@ -9,6 +9,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 		.macro	pack, rd, hw1, hw2
 #ifndef __ARMEB__
@@ -129,3 +130,4 @@ ENTRY(__raw_readsw)
 		strneb	ip, [r1]
 		ldmfd	sp!, {r4, pc}
 ENDPROC(__raw_readsw)
+EXPORT_SYMBOL(__raw_readsw)
diff --git a/arch/arm/lib/io-writesb.S b/arch/arm/lib/io-writesb.S
index a46bbc9b168b..fa3633594415 100644
--- a/arch/arm/lib/io-writesb.S
+++ b/arch/arm/lib/io-writesb.S
@@ -9,6 +9,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 		.macro	outword, rd
 #ifndef __ARMEB__
@@ -92,3 +93,4 @@ ENTRY(__raw_writesb)
 
 		ldmfd	sp!, {r4, r5, pc}
 ENDPROC(__raw_writesb)
+EXPORT_SYMBOL(__raw_writesb)
diff --git a/arch/arm/lib/io-writesl.S b/arch/arm/lib/io-writesl.S
index 4ea2435988c1..98ed6aec0b47 100644
--- a/arch/arm/lib/io-writesl.S
+++ b/arch/arm/lib/io-writesl.S
@@ -9,6 +9,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 ENTRY(__raw_writesl)
 		teq	r2, #0		@ do we have to check for the zero len?
@@ -65,3 +66,4 @@ ENTRY(__raw_writesl)
 		bne	6b
 		ret	lr
 ENDPROC(__raw_writesl)
+EXPORT_SYMBOL(__raw_writesl)
diff --git a/arch/arm/lib/io-writesw-armv3.S b/arch/arm/lib/io-writesw-armv3.S
index 121789eb6802..577184c082bb 100644
--- a/arch/arm/lib/io-writesw-armv3.S
+++ b/arch/arm/lib/io-writesw-armv3.S
@@ -9,6 +9,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 .Loutsw_bad_alignment:
 		adr	r0, .Loutsw_bad_align_msg
@@ -124,3 +125,4 @@ ENTRY(__raw_writesw)
 		strne	ip, [r0]
 
 		ldmfd	sp!, {r4, r5, r6, pc}
+EXPORT_SYMBOL(__raw_writesw)
diff --git a/arch/arm/lib/io-writesw-armv4.S b/arch/arm/lib/io-writesw-armv4.S
index 269f90c51ad2..e335f489d1fc 100644
--- a/arch/arm/lib/io-writesw-armv4.S
+++ b/arch/arm/lib/io-writesw-armv4.S
@@ -9,6 +9,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 		.macro	outword, rd
 #ifndef __ARMEB__
@@ -98,3 +99,4 @@ ENTRY(__raw_writesw)
 		strneh	ip, [r0]
 		ret	lr
 ENDPROC(__raw_writesw)
+EXPORT_SYMBOL(__raw_writesw)
diff --git a/arch/arm/lib/lib1funcs.S b/arch/arm/lib/lib1funcs.S
index 9397b2e532af..f541bc013bff 100644
--- a/arch/arm/lib/lib1funcs.S
+++ b/arch/arm/lib/lib1funcs.S
@@ -36,6 +36,7 @@ Boston, MA 02111-1307, USA.  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 #include <asm/unwind.h>
+#include <asm/export.h>
 
 .macro ARM_DIV_BODY dividend, divisor, result, curbit
 
@@ -238,6 +239,8 @@ UNWIND(.fnstart)
 UNWIND(.fnend)
 ENDPROC(__udivsi3)
 ENDPROC(__aeabi_uidiv)
+EXPORT_SYMBOL(__udivsi3)
+EXPORT_SYMBOL(__aeabi_uidiv)
 
 ENTRY(__umodsi3)
 UNWIND(.fnstart)
@@ -256,6 +259,7 @@ UNWIND(.fnstart)
 
 UNWIND(.fnend)
 ENDPROC(__umodsi3)
+EXPORT_SYMBOL(__umodsi3)
 
 #ifdef CONFIG_ARM_PATCH_IDIV
 	.align 3
@@ -303,6 +307,8 @@ UNWIND(.fnstart)
 UNWIND(.fnend)
 ENDPROC(__divsi3)
 ENDPROC(__aeabi_idiv)
+EXPORT_SYMBOL(__divsi3)
+EXPORT_SYMBOL(__aeabi_idiv)
 
 ENTRY(__modsi3)
 UNWIND(.fnstart)
@@ -327,6 +333,7 @@ UNWIND(.fnstart)
 
 UNWIND(.fnend)
 ENDPROC(__modsi3)
+EXPORT_SYMBOL(__modsi3)
 
 #ifdef CONFIG_AEABI
 
@@ -343,6 +350,7 @@ UNWIND(.save {r0, r1, ip, lr}	)
 
 UNWIND(.fnend)
 ENDPROC(__aeabi_uidivmod)
+EXPORT_SYMBOL(__aeabi_uidivmod)
 
 ENTRY(__aeabi_idivmod)
 UNWIND(.fnstart)
@@ -356,6 +364,7 @@ UNWIND(.save {r0, r1, ip, lr}	)
 
 UNWIND(.fnend)
 ENDPROC(__aeabi_idivmod)
+EXPORT_SYMBOL(__aeabi_idivmod)
 
 #endif
 
diff --git a/arch/arm/lib/lshrdi3.S b/arch/arm/lib/lshrdi3.S
index 922dcd88b02b..e40833981417 100644
--- a/arch/arm/lib/lshrdi3.S
+++ b/arch/arm/lib/lshrdi3.S
@@ -28,6 +28,7 @@ Boston, MA 02110-1301, USA.  */
 
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 #ifdef __ARMEB__
 #define al r1
@@ -52,3 +53,5 @@ ENTRY(__aeabi_llsr)
 
 ENDPROC(__lshrdi3)
 ENDPROC(__aeabi_llsr)
+EXPORT_SYMBOL(__lshrdi3)
+EXPORT_SYMBOL(__aeabi_llsr)
diff --git a/arch/arm/lib/memchr.S b/arch/arm/lib/memchr.S
index 74a5bed6d999..44182bf686a5 100644
--- a/arch/arm/lib/memchr.S
+++ b/arch/arm/lib/memchr.S
@@ -11,6 +11,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 	.text
 	.align	5
@@ -24,3 +25,4 @@ ENTRY(memchr)
 2:	movne	r0, #0
 	ret	lr
 ENDPROC(memchr)
+EXPORT_SYMBOL(memchr)
diff --git a/arch/arm/lib/memcpy.S b/arch/arm/lib/memcpy.S
index 64111bd4440b..1f822fc52400 100644
--- a/arch/arm/lib/memcpy.S
+++ b/arch/arm/lib/memcpy.S
@@ -13,6 +13,7 @@
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 #include <asm/unwind.h>
+#include <asm/export.h>
 
 #define LDR1W_SHIFT	0
 #define STR1W_SHIFT	0
@@ -68,3 +69,4 @@ ENTRY(memcpy)
 
 ENDPROC(memcpy)
 ENDPROC(mmiocpy)
+EXPORT_SYMBOL(memcpy)
diff --git a/arch/arm/lib/memmove.S b/arch/arm/lib/memmove.S
index 69a9d47fc5ab..71dcc5400d02 100644
--- a/arch/arm/lib/memmove.S
+++ b/arch/arm/lib/memmove.S
@@ -13,6 +13,7 @@
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 #include <asm/unwind.h>
+#include <asm/export.h>
 
 		.text
 
@@ -225,3 +226,4 @@ ENTRY(memmove)
 18:		backward_copy_shift	push=24	pull=8
 
 ENDPROC(memmove)
+EXPORT_SYMBOL(memmove)
diff --git a/arch/arm/lib/memset.S b/arch/arm/lib/memset.S
index 3c65e3bd790f..6f075ca09abc 100644
--- a/arch/arm/lib/memset.S
+++ b/arch/arm/lib/memset.S
@@ -12,6 +12,7 @@
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 #include <asm/unwind.h>
+#include <asm/export.h>
 
 	.text
 	.align	5
@@ -135,3 +136,4 @@ UNWIND( .fnstart            )
 UNWIND( .fnend   )
 ENDPROC(memset)
 ENDPROC(mmioset)
+EXPORT_SYMBOL(memset)
diff --git a/arch/arm/lib/memzero.S b/arch/arm/lib/memzero.S
index 0eded952e089..6dec26ed5bcc 100644
--- a/arch/arm/lib/memzero.S
+++ b/arch/arm/lib/memzero.S
@@ -10,6 +10,7 @@
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 #include <asm/unwind.h>
+#include <asm/export.h>
 
 	.text
 	.align	5
@@ -135,3 +136,4 @@ UNWIND(	.fnstart			)
 	ret	lr			@ 1
 UNWIND(	.fnend				)
 ENDPROC(__memzero)
+EXPORT_SYMBOL(__memzero)
diff --git a/arch/arm/lib/muldi3.S b/arch/arm/lib/muldi3.S
index 204305956925..b8f12388ccac 100644
--- a/arch/arm/lib/muldi3.S
+++ b/arch/arm/lib/muldi3.S
@@ -12,6 +12,7 @@
 
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 #ifdef __ARMEB__
 #define xh r0
@@ -46,3 +47,5 @@ ENTRY(__aeabi_lmul)
 
 ENDPROC(__muldi3)
 ENDPROC(__aeabi_lmul)
+EXPORT_SYMBOL(__muldi3)
+EXPORT_SYMBOL(__aeabi_lmul)
diff --git a/arch/arm/lib/putuser.S b/arch/arm/lib/putuser.S
index 38d660d3705f..11de126e2ed6 100644
--- a/arch/arm/lib/putuser.S
+++ b/arch/arm/lib/putuser.S
@@ -31,6 +31,7 @@
 #include <asm/assembler.h>
 #include <asm/errno.h>
 #include <asm/domain.h>
+#include <asm/export.h>
 
 ENTRY(__put_user_1)
 	check_uaccess r0, 1, r1, ip, __put_user_bad
@@ -38,6 +39,7 @@ ENTRY(__put_user_1)
 	mov	r0, #0
 	ret	lr
 ENDPROC(__put_user_1)
+EXPORT_SYMBOL(__put_user_1)
 
 ENTRY(__put_user_2)
 	check_uaccess r0, 2, r1, ip, __put_user_bad
@@ -62,6 +64,7 @@ ENTRY(__put_user_2)
 	mov	r0, #0
 	ret	lr
 ENDPROC(__put_user_2)
+EXPORT_SYMBOL(__put_user_2)
 
 ENTRY(__put_user_4)
 	check_uaccess r0, 4, r1, ip, __put_user_bad
@@ -69,6 +72,7 @@ ENTRY(__put_user_4)
 	mov	r0, #0
 	ret	lr
 ENDPROC(__put_user_4)
+EXPORT_SYMBOL(__put_user_4)
 
 ENTRY(__put_user_8)
 	check_uaccess r0, 8, r1, ip, __put_user_bad
@@ -82,6 +86,7 @@ ENTRY(__put_user_8)
 	mov	r0, #0
 	ret	lr
 ENDPROC(__put_user_8)
+EXPORT_SYMBOL(__put_user_8)
 
 __put_user_bad:
 	mov	r0, #-EFAULT
diff --git a/arch/arm/lib/setbit.S b/arch/arm/lib/setbit.S
index 618fedae4b37..3c8b11240fca 100644
--- a/arch/arm/lib/setbit.S
+++ b/arch/arm/lib/setbit.S
@@ -9,7 +9,10 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 #include "bitops.h"
 		.text
 
 bitop	_set_bit, orr
+
+EXPORT_SYMBOL(_set_bit)
diff --git a/arch/arm/lib/strchr.S b/arch/arm/lib/strchr.S
index 013d64c71e8d..7301f6e6046c 100644
--- a/arch/arm/lib/strchr.S
+++ b/arch/arm/lib/strchr.S
@@ -11,6 +11,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 		.text
 		.align	5
@@ -25,3 +26,4 @@ ENTRY(strchr)
 		subeq	r0, r0, #1
 		ret	lr
 ENDPROC(strchr)
+EXPORT_SYMBOL(strchr)
diff --git a/arch/arm/lib/strrchr.S b/arch/arm/lib/strrchr.S
index 3cec1c7482c4..aaf9fd98b754 100644
--- a/arch/arm/lib/strrchr.S
+++ b/arch/arm/lib/strrchr.S
@@ -11,6 +11,7 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 		.text
 		.align	5
@@ -24,3 +25,4 @@ ENTRY(strrchr)
 		mov	r0, r3
 		ret	lr
 ENDPROC(strrchr)
+EXPORT_SYMBOL(strrchr)
diff --git a/arch/arm/lib/testchangebit.S b/arch/arm/lib/testchangebit.S
index 4becdc3a59cb..e3d19b87fbe0 100644
--- a/arch/arm/lib/testchangebit.S
+++ b/arch/arm/lib/testchangebit.S
@@ -9,7 +9,10 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 #include "bitops.h"
                 .text
 
 testop	_test_and_change_bit, eor, str
+
+EXPORT_SYMBOL(_test_and_change_bit)
diff --git a/arch/arm/lib/testclearbit.S b/arch/arm/lib/testclearbit.S
index 918841dcce7a..d247e6f70fe6 100644
--- a/arch/arm/lib/testclearbit.S
+++ b/arch/arm/lib/testclearbit.S
@@ -9,7 +9,10 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 #include "bitops.h"
                 .text
 
 testop	_test_and_clear_bit, bicne, strne
+
+EXPORT_SYMBOL(_test_and_clear_bit)
diff --git a/arch/arm/lib/testsetbit.S b/arch/arm/lib/testsetbit.S
index 8d1b2fe9e487..76800ff601ff 100644
--- a/arch/arm/lib/testsetbit.S
+++ b/arch/arm/lib/testsetbit.S
@@ -9,7 +9,10 @@
  */
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 #include "bitops.h"
                 .text
 
 testop	_test_and_set_bit, orreq, streq
+
+EXPORT_SYMBOL(_test_and_set_bit)
diff --git a/arch/arm/lib/uaccess_with_memcpy.c b/arch/arm/lib/uaccess_with_memcpy.c
index 6bd1089b07e0..1626e3a551a1 100644
--- a/arch/arm/lib/uaccess_with_memcpy.c
+++ b/arch/arm/lib/uaccess_with_memcpy.c
@@ -19,6 +19,7 @@
 #include <linux/gfp.h>
 #include <linux/highmem.h>
 #include <linux/hugetlb.h>
+#include <linux/export.h>
 #include <asm/current.h>
 #include <asm/page.h>
 
@@ -156,6 +157,7 @@ arm_copy_to_user(void __user *to, const void *from, unsigned long n)
 	}
 	return n;
 }
+EXPORT_SYMBOL(arm_copy_to_user);
 	
 static unsigned long noinline
 __clear_user_memset(void __user *addr, unsigned long n)
@@ -213,6 +215,7 @@ unsigned long arm_clear_user(void __user *addr, unsigned long n)
 	}
 	return n;
 }
+EXPORT_SYMBOL(arm_clear_user);
 
 #if 0
 
diff --git a/arch/arm/lib/ucmpdi2.S b/arch/arm/lib/ucmpdi2.S
index ad4a6309141a..127a91af46f3 100644
--- a/arch/arm/lib/ucmpdi2.S
+++ b/arch/arm/lib/ucmpdi2.S
@@ -12,6 +12,7 @@
 
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 #ifdef __ARMEB__
 #define xh r0
@@ -35,6 +36,7 @@ ENTRY(__ucmpdi2)
 	ret	lr
 
 ENDPROC(__ucmpdi2)
+EXPORT_SYMBOL(__ucmpdi2)
 
 #ifdef CONFIG_AEABI
 
@@ -48,6 +50,7 @@ ENTRY(__aeabi_ulcmp)
 	ret	lr
 
 ENDPROC(__aeabi_ulcmp)
+EXPORT_SYMBOL(__aeabi_ulcmp)
 
 #endif
 
diff --git a/arch/arm/mach-imx/Makefile b/arch/arm/mach-imx/Makefile
index 9f5fffd62702..8ed8ab56bb78 100644
--- a/arch/arm/mach-imx/Makefile
+++ b/arch/arm/mach-imx/Makefile
@@ -32,7 +32,6 @@ endif
 
 ifdef CONFIG_SND_IMX_SOC
 obj-y += ssi-fiq.o
-obj-y += ssi-fiq-ksym.o
 endif
 
 # i.MX1 based machines
diff --git a/arch/arm/mach-imx/ssi-fiq-ksym.c b/arch/arm/mach-imx/ssi-fiq-ksym.c
deleted file mode 100644
index 792090f9a032..000000000000
--- a/arch/arm/mach-imx/ssi-fiq-ksym.c
+++ /dev/null
@@ -1,20 +0,0 @@
-/*
- * Exported ksyms for the SSI FIQ handler
- *
- * Copyright (C) 2009, Sascha Hauer <s.hauer@pengutronix.de>
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- */
-
-#include <linux/module.h>
-
-#include <linux/platform_data/asoc-imx-ssi.h>
-
-EXPORT_SYMBOL(imx_ssi_fiq_tx_buffer);
-EXPORT_SYMBOL(imx_ssi_fiq_rx_buffer);
-EXPORT_SYMBOL(imx_ssi_fiq_start);
-EXPORT_SYMBOL(imx_ssi_fiq_end);
-EXPORT_SYMBOL(imx_ssi_fiq_base);
-
diff --git a/arch/arm/mach-imx/ssi-fiq.S b/arch/arm/mach-imx/ssi-fiq.S
index a8b93c5f29b5..fd7917f1c204 100644
--- a/arch/arm/mach-imx/ssi-fiq.S
+++ b/arch/arm/mach-imx/ssi-fiq.S
@@ -8,6 +8,7 @@
 
 #include <linux/linkage.h>
 #include <asm/assembler.h>
+#include <asm/export.h>
 
 /*
  * r8  = bit 0-15: tx offset, bit 16-31: tx buffer size
@@ -144,4 +145,8 @@
 		.word 0x0
 .L_imx_ssi_fiq_end:
 imx_ssi_fiq_end:
-
+EXPORT_SYMBOL(imx_ssi_fiq_tx_buffer)
+EXPORT_SYMBOL(imx_ssi_fiq_rx_buffer)
+EXPORT_SYMBOL(imx_ssi_fiq_start)
+EXPORT_SYMBOL(imx_ssi_fiq_end)
+EXPORT_SYMBOL(imx_ssi_fiq_base)

-- 
RMK's Patch system: http://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line: currently at 9.6Mbps down 400kbps up
according to speedtest.net.

Re: [PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Nicolas Pitre]

On Wed, 23 Nov 2016, Russell King - ARM Linux wrote:

> On Wed, Nov 23, 2016 at 09:33:32AM +0000, Russell King - ARM Linux wrote:
> > So, I still think the whole approach is wrong - it's added extra
> > fragility that wasn't there with the armksyms.c approach.
> 
> Here's what the diffstat and patch looks like when you combine the
> original commit and the three fixes.  The LoC delta of 25 lines can
> be accounted for as deleted commentry.  So, I think (as I've detailed
> above) that the _technical_ benefit of the approach is very low.
> 
> If we want to move the exports into assembly files, I've no problem
> with that, provided we can do it better than this - and by better I
> mean not creating the fragile asm-prototypes.h which divorses the
> prototypes from everything else.

I think we agree on that.

I see that the revert made it to Linus' tree.  That's probably the best 
outcome for now.

I was working on that better solution I alluded to previously. And it 
isn't really complicated either, with much fewer lines of code.  But 
this can wait for the next merge window.


Nicolas

Re: [GIT PULL] kbuild changes for v4.9-rc1

[Adam Borowski]

On Mon, Oct 17, 2016 at 02:57:09PM +1100, Nicholas Piggin wrote:
> On Sat, 15 Oct 2016 17:22:05 -0700 Omar Sandoval <osandov@osandov.com> wrote:
> > On Fri, Oct 14, 2016 at 10:12:46PM +0200, Michal Marek wrote:
> > > please pull these kbuild changes for v4.9-rc1:
> > > 
> > > - EXPORT_SYMBOL for asm source by Al Viro. This does bring a regression,
> > >   because genksyms no longer generates checksums for these symbols
> > >   (CONFIG_MODVERSIONS). Nick Piggin is working on a patch to fix this.
> > >   Plus, we are talking about functions like strcpy(), which rarely
> > >   change prototypes.  
> > 
> > So this has broken all module loading for me. I get the following dmesg
> > spew:
> > ...
> > [    4.586914] scsi_mod: no symbol version for memset
> > [    4.587920] scsi_mod: Unknown symbol memset (err -22)
> > [    4.588443] scsi_mod: no symbol version for ___preempt_schedule
> > [    4.589026] scsi_mod: Unknown symbol ___preempt_schedule (err -22)
> > ...
> > 
> > Reverting 784d5699eddc ("x86: move exports to actual definitions") fixes
> > it for me. This is with GCC 6.2.1, binutils 2.27, attached config.
> 
> Thanks for the report. Could you try this patch and see if it helps?
[patch snipped]

Omar probably won't wake up in quite a while, so I've tested the patch.
Alas, doesn't help.  Similar spew (for the few modules I don't have =y),
while reverting 784d5699eddc fixes it for me too.

Debian sid toolchain: gcc 6.2.0-6, binutils 2.27-8, config at
https://angband.pl/tmp/config-4.9.0-rc1-debug+.gz


Meow!
-- 
A MAP07 (Dead Simple) raspberry tincture recipe: 0.5l 95% alcohol, 1kg
raspberries, 0.4kg sugar; put into a big jar for 1 month.  Filter out and
throw away the fruits (can dump them into a cake, etc), let the drink age
at least 3-6 months.

Re: [PATCH] kbuild: provide include/asm/asm-prototypes.h for ARM

[Nicholas Piggin]

On Wed, 19 Oct 2016 16:32:00 +0100
Russell King - ARM Linux <linux@armlinux.org.uk> wrote:

> On Wed, Oct 19, 2016 at 05:02:55PM +0200, Arnd Bergmann wrote:
> > On Wednesday, October 19, 2016 4:52:06 PM CEST Michal Marek wrote:  
> > > Dne 17.10.2016 v 14:26 Arnd Bergmann napsal(a):  
> > > > This adds an asm/asm-prototypes.h header for ARM to fix the
> > > > broken symbol versioning for symbols exported from assembler
> > > > files.
> > > > 
> > > > In addition to the header, we have to do these other small
> > > > changes:
> > > > 
> > > > - move the 'extern' declarations out of memset_io/memcpy_io
> > > >   to make them visible to the symbol version generator
> > > > - move the exports from bitops.h to {change,clear,set,...}bit.S
> > > > - move the exports from csumpartialgeneric.S into the files
> > > >   including it
> > > > 
> > > > I couldn't find the correct prototypes for the compiler builtins,
> > > > so I went with the fake 'void f(void)' prototypes that we had
> > > > before.
> > > > 
> > > > Signed-off-by: Arnd Bergmann <arnd@arndb.de>  
> > > 
> > > Hi Arnd,
> > > 
> > > just to make sure I'm looking at the right code - is this based on the
> > > patch by Nick here: https://patchwork.kernel.org/patch/9377783/?
> > >   
> > 
> > (adding Russell to Cc, I missed him during my earlier mail, which
> > is now archived at https://lkml.org/lkml/2016/10/17/356)  
> 
> I'm not in favour of this.
> 
>    +extern void mmioset(void *, unsigned int, size_t);
>    +extern void mmiocpy(void *, const void *, size_t);
>    +
>     #ifndef __ARMBE__
>     static inline void memset_io(volatile void __iomem *dst, unsigned c,
>            size_t count)
>     {
>    -       extern void mmioset(void *, unsigned int, size_t);
>            mmioset((void __force *)dst, c, count);
>     }
> 
> The reason they're declared _within_ memset_io() is to prevent people
> from using them by hiding their declaration.  Moving them outside is
> an open invitation to stupid people starting to use them as an "oh it
> must be an official API".
> 
> We know this happens, there's been a long history of this kind of stupid
> in the ARM community, not only with cache flushing APIs, but also the
> DMA APIs.
> 
> The way the existing code is written is a completely valid way to hide
> declarations from outside the intended caller's scope.
> 
> We've been here many times, we've had many people doing this crap, so
> I'm now at the point of NAKing changes which result in an increased
> visibility to the rest of the kernel of symbols that should not be
> used by stupid driver authors.
> 
> Now, why do we have these extra functions when they're just aliased to
> memset()/memcpy() - to avoid GCC optimising them because it thinks that
> they're standard memset()/memcpy().
> 
> So overall this gets a NAK from me.

Fair point, what about leaving those as they are, and also adding
them to asm-prototypes.h protected with GENKSYMS ifdef? It's not
beautiful, but still better than armksyms.c before Al's patches (or
at least no worse).


> Now, it would have _ALSO_ been nice to have been at least COPIED on the
> original set of changes that caused the need for this change.  I wasn't.
> So I want to see the original set of changes reverted, because they're
> clearly causing breakage.  Let's revert them and then go through the
> proper process of maintainer review, rather than bypassing maintainers
> and screwing up architectures in the process.  There really is no
> excuse for this crap.

You may have a point about improvement of the process. I wasn't
involved in the original patches, but we did cc linux-arch when the
.S CRC issue became known.

However let's work on the assumption that they won't be reverted at this
stage, and try to come up with something to fix it that you're happy with.

Thanks,
Nick

Re: [PATCH] kbuild: provide include/asm/asm-prototypes.h for ARM

[Arnd Bergmann]

On Thursday, October 20, 2016 3:08:14 PM CEST Nicholas Piggin wrote:
> On Wed, 19 Oct 2016 16:32:00 +0100 Russell King - ARM Linux <linux@armlinux.org.uk> wrote:
> > I'm not in favour of this.
> > 
> >    +extern void mmioset(void *, unsigned int, size_t);
> >    +extern void mmiocpy(void *, const void *, size_t);
> >    +
> >     #ifndef __ARMBE__
> >     static inline void memset_io(volatile void __iomem *dst, unsigned c,
> >            size_t count)
> >     {
> >    -       extern void mmioset(void *, unsigned int, size_t);
> >            mmioset((void __force *)dst, c, count);
> >     }
> > 
> > The reason they're declared _within_ memset_io() is to prevent people
> > from using them by hiding their declaration.  Moving them outside is
> > an open invitation to stupid people starting to use them as an "oh it
> > must be an official API".
> > 

I've split out that change from the other ones now, and will follow
up with the patch to address all the other ones first.

> Fair point, what about leaving those as they are, and also adding
> them to asm-prototypes.h protected with GENKSYMS ifdef? It's not
> beautiful, but still better than armksyms.c before Al's patches (or
> at least no worse).

I'm trying this one, and an alternative patch that moves the
export into arch/arm/kernel/io.h. Let's see if we can agree
on one of these.

	Arnd

[PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Arnd Bergmann]

This adds an asm/asm-prototypes.h header for ARM to fix the
broken symbol versioning for symbols exported from assembler
files.

In addition to the header, we have to do these other small
changes:

- move the exports from bitops.h to {change,clear,set,...}bit.S
- move the exports from csumpartialgeneric.S into the files
  including it

I couldn't find the correct prototypes for the compiler builtins,
so I went with the fake 'void f(void)' prototypes that we had
before.

This leaves the mmioset/mmiocpy function for now, as it's not
obvious how to best handle them.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>

diff --git a/arch/arm/include/asm/asm-prototypes.h b/arch/arm/include/asm/asm-prototypes.h
new file mode 100644
index 000000000000..04e5616a7b15
--- /dev/null
+++ b/arch/arm/include/asm/asm-prototypes.h
@@ -0,0 +1,34 @@
+#include <linux/arm-smccc.h>
+#include <linux/bitops.h>
+#include <linux/ftrace.h>
+#include <linux/io.h>
+#include <linux/platform_data/asoc-imx-ssi.h>
+#include <linux/string.h>
+#include <linux/uaccess.h>
+
+#include <asm/checksum.h>
+#include <asm/div64.h>
+#include <asm/memory.h>
+
+extern void __aeabi_idivmod(void);
+extern void __aeabi_idiv(void);
+extern void __aeabi_lasr(void);
+extern void __aeabi_llsl(void);
+extern void __aeabi_llsr(void);
+extern void __aeabi_lmul(void);
+extern void __aeabi_uidivmod(void);
+extern void __aeabi_uidiv(void);
+extern void __aeabi_ulcmp(void);
+
+extern void __ashldi3(void);
+extern void __ashrdi3(void);
+extern void __bswapdi2(void);
+extern void __bswapsi2(void);
+extern void __divsi3(void);
+extern void __do_div64(void);
+extern void __lshrdi3(void);
+extern void __modsi3(void);
+extern void __muldi3(void);
+extern void __ucmpdi2(void);
+extern void __udivsi3(void);
+extern void __umodsi3(void);
diff --git a/arch/arm/lib/bitops.h b/arch/arm/lib/bitops.h
index df06638b327c..afaef2a7faec 100644
--- a/arch/arm/lib/bitops.h
+++ b/arch/arm/lib/bitops.h
@@ -26,7 +26,6 @@ UNWIND(	.fnstart	)
 	bx	lr
 UNWIND(	.fnend		)
 ENDPROC(\name		)
-EXPORT_SYMBOL(\name	)
 	.endm
 
 	.macro	testop, name, instr, store
@@ -57,7 +56,6 @@ UNWIND(	.fnstart	)
 2:	bx	lr
 UNWIND(	.fnend		)
 ENDPROC(\name		)
-EXPORT_SYMBOL(\name	)
 	.endm
 #else
 	.macro	bitop, name, instr
@@ -77,7 +75,6 @@ UNWIND(	.fnstart	)
 	ret	lr
 UNWIND(	.fnend		)
 ENDPROC(\name		)
-EXPORT_SYMBOL(\name	)
 	.endm
 
 /**
@@ -106,6 +103,5 @@ UNWIND(	.fnstart	)
 	ret	lr
 UNWIND(	.fnend		)
 ENDPROC(\name		)
-EXPORT_SYMBOL(\name	)
 	.endm
 #endif
diff --git a/arch/arm/lib/changebit.S b/arch/arm/lib/changebit.S
index f4027862172f..005fdd18c509 100644
--- a/arch/arm/lib/changebit.S
+++ b/arch/arm/lib/changebit.S
@@ -13,3 +13,4 @@
                 .text
 
 bitop	_change_bit, eor
+EXPORT_SYMBOL(_change_bit)
diff --git a/arch/arm/lib/clearbit.S b/arch/arm/lib/clearbit.S
index f6b75fb64d30..501eff09968d 100644
--- a/arch/arm/lib/clearbit.S
+++ b/arch/arm/lib/clearbit.S
@@ -13,3 +13,4 @@
                 .text
 
 bitop	_clear_bit, bic
+EXPORT_SYMBOL(_clear_bit)
diff --git a/arch/arm/lib/csumpartialcopy.S b/arch/arm/lib/csumpartialcopy.S
index 9c3383fed129..bdcc2eea4e5c 100644
--- a/arch/arm/lib/csumpartialcopy.S
+++ b/arch/arm/lib/csumpartialcopy.S
@@ -49,6 +49,7 @@
 
 #define FN_ENTRY	ENTRY(csum_partial_copy_nocheck)
 #define FN_EXIT		ENDPROC(csum_partial_copy_nocheck)
-#define FN_EXPORT	EXPORT_SYMBOL(csum_partial_copy_nocheck)
 
 #include "csumpartialcopygeneric.S"
+
+EXPORT_SYMBOL(csum_partial_copy_nocheck)
diff --git a/arch/arm/lib/csumpartialcopygeneric.S b/arch/arm/lib/csumpartialcopygeneric.S
index 8b94d20e51d1..06825566c0f7 100644
--- a/arch/arm/lib/csumpartialcopygeneric.S
+++ b/arch/arm/lib/csumpartialcopygeneric.S
@@ -332,4 +332,3 @@ FN_ENTRY
 		mov	r5, r4, get_byte_1
 		b	.Lexit
 FN_EXIT
-FN_EXPORT
diff --git a/arch/arm/lib/csumpartialcopyuser.S b/arch/arm/lib/csumpartialcopyuser.S
index 5d495edf3d83..d5522c94f58c 100644
--- a/arch/arm/lib/csumpartialcopyuser.S
+++ b/arch/arm/lib/csumpartialcopyuser.S
@@ -73,9 +73,9 @@
 
 #define FN_ENTRY	ENTRY(csum_partial_copy_from_user)
 #define FN_EXIT		ENDPROC(csum_partial_copy_from_user)
-#define FN_EXPORT	EXPORT_SYMBOL(csum_partial_copy_from_user)
 
 #include "csumpartialcopygeneric.S"
+EXPORT_SYMBOL(csum_partial_copy_from_user)
 
 /*
  * FIXME: minor buglet here
diff --git a/arch/arm/lib/setbit.S b/arch/arm/lib/setbit.S
index 618fedae4b37..d748b8d1326f 100644
--- a/arch/arm/lib/setbit.S
+++ b/arch/arm/lib/setbit.S
@@ -13,3 +13,4 @@
 		.text
 
 bitop	_set_bit, orr
+EXPORT_SYMBOL(_set_bit)
diff --git a/arch/arm/lib/testchangebit.S b/arch/arm/lib/testchangebit.S
index 4becdc3a59cb..4d2dafa9b787 100644
--- a/arch/arm/lib/testchangebit.S
+++ b/arch/arm/lib/testchangebit.S
@@ -13,3 +13,4 @@
                 .text
 
 testop	_test_and_change_bit, eor, str
+EXPORT_SYMBOL(_test_and_change_bit)
diff --git a/arch/arm/lib/testclearbit.S b/arch/arm/lib/testclearbit.S
index 918841dcce7a..fe5cae2e480a 100644
--- a/arch/arm/lib/testclearbit.S
+++ b/arch/arm/lib/testclearbit.S
@@ -13,3 +13,4 @@
                 .text
 
 testop	_test_and_clear_bit, bicne, strne
+EXPORT_SYMBOL(_test_and_clear_bit)
diff --git a/arch/arm/lib/testsetbit.S b/arch/arm/lib/testsetbit.S
index 8d1b2fe9e487..25fed837edb3 100644
--- a/arch/arm/lib/testsetbit.S
+++ b/arch/arm/lib/testsetbit.S
@@ -13,3 +13,4 @@
                 .text
 
 testop	_test_and_set_bit, orreq, streq
+EXPORT_SYMBOL(_test_and_set_bit)

Re: [PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Nicholas Piggin]

On Mon, 24 Oct 2016 17:05:26 +0200
Arnd Bergmann <arnd@arndb.de> wrote:

> This adds an asm/asm-prototypes.h header for ARM to fix the
> broken symbol versioning for symbols exported from assembler
> files.
> 
> In addition to the header, we have to do these other small
> changes:
> 
> - move the exports from bitops.h to {change,clear,set,...}bit.S
> - move the exports from csumpartialgeneric.S into the files
>   including it
> 
> I couldn't find the correct prototypes for the compiler builtins,
> so I went with the fake 'void f(void)' prototypes that we had
> before.
> 
> This leaves the mmioset/mmiocpy function for now, as it's not
> obvious how to best handle them.


This looks nicer. I like variant B because it keeps the GENKSYMS cruft to
a single location, but either one isn't too bad.

I'd like to get moving on this, so let's at least get the generic kbuild
change merged. In the end, the kbuild code does not prevent a maintainer
from putting their EXPORT_SYMBOL in whatever location they like, so there
is no reason not to merge it (certainly there will be archs that do use
it).

Michal, what's your thoughts? If you merge my patch 2/2 and skip 1/2, it
should not give any new build warnings or errors, so then arch patches can
go via arch trees. 1/2 could go in after everyone is up to date.

Thanks,
Nick

Re: [PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Russell King - ARM Linux]

On Tue, Oct 25, 2016 at 07:32:00PM +1100, Nicholas Piggin wrote:
> On Mon, 24 Oct 2016 17:05:26 +0200
> Arnd Bergmann <arnd@arndb.de> wrote:
> 
> > This adds an asm/asm-prototypes.h header for ARM to fix the
> > broken symbol versioning for symbols exported from assembler
> > files.
> > 
> > In addition to the header, we have to do these other small
> > changes:
> > 
> > - move the exports from bitops.h to {change,clear,set,...}bit.S
> > - move the exports from csumpartialgeneric.S into the files
> >   including it
> > 
> > I couldn't find the correct prototypes for the compiler builtins,
> > so I went with the fake 'void f(void)' prototypes that we had
> > before.
> > 
> > This leaves the mmioset/mmiocpy function for now, as it's not
> > obvious how to best handle them.
> 
> 
> This looks nicer. I like variant B because it keeps the GENKSYMS cruft to
> a single location, but either one isn't too bad.
> 
> I'd like to get moving on this, so let's at least get the generic kbuild
> change merged. In the end, the kbuild code does not prevent a maintainer
> from putting their EXPORT_SYMBOL in whatever location they like, so there
> is no reason not to merge it (certainly there will be archs that do use
> it).
> 
> Michal, what's your thoughts? If you merge my patch 2/2 and skip 1/2, it
> should not give any new build warnings or errors, so then arch patches can
> go via arch trees. 1/2 could go in after everyone is up to date.

So what's the conclusion on this?  I've just had a failure due to
CONFIG_TRIM_UNUSED_KSYMS reported on ARM, and it looks like (at
least some of) patch 1 could resolve it.

Do we need to split patch 1?

Has any of these patches been committed yet?

-- 
RMK's Patch system: http://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line: currently at 9.6Mbps down 400kbps up
according to speedtest.net.

Re: [PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Linus Torvalds]

On Sun, Nov 20, 2016 at 5:21 AM, Russell King - ARM Linux
<linux@armlinux.org.uk> wrote:
> On Tue, Oct 25, 2016 at 07:32:00PM +1100, Nicholas Piggin wrote:
>>
>> Michal, what's your thoughts? If you merge my patch 2/2 and skip 1/2, it
>> should not give any new build warnings or errors, so then arch patches can
>> go via arch trees. 1/2 could go in after everyone is up to date.
>
> So what's the conclusion on this?  I've just had a failure due to
> CONFIG_TRIM_UNUSED_KSYMS reported on ARM, and it looks like (at
> least some of) patch 1 could resolve it.

Hmm. I've got

  cc6acc11cad1 kbuild: be more careful about matching preprocessed asm
___EXPORT_SYMBOL
  4efca4ed05cb kbuild: modversions for EXPORT_SYMBOL() for asm

in my tree. Is that sufficient, or do we still have issues?

             Linus

Re: [PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Russell King - ARM Linux]

On Sun, Nov 20, 2016 at 10:32:50AM -0800, Linus Torvalds wrote:
> On Sun, Nov 20, 2016 at 5:21 AM, Russell King - ARM Linux
> <linux@armlinux.org.uk> wrote:
> > On Tue, Oct 25, 2016 at 07:32:00PM +1100, Nicholas Piggin wrote:
> >>
> >> Michal, what's your thoughts? If you merge my patch 2/2 and skip 1/2, it
> >> should not give any new build warnings or errors, so then arch patches can
> >> go via arch trees. 1/2 could go in after everyone is up to date.
> >
> > So what's the conclusion on this?  I've just had a failure due to
> > CONFIG_TRIM_UNUSED_KSYMS reported on ARM, and it looks like (at
> > least some of) patch 1 could resolve it.
> 
> Hmm. I've got
> 
>   cc6acc11cad1 kbuild: be more careful about matching preprocessed asm
> ___EXPORT_SYMBOL
>   4efca4ed05cb kbuild: modversions for EXPORT_SYMBOL() for asm
> 
> in my tree. Is that sufficient, or do we still have issues?

Hmm, those seem to have gone in during the last week, so I haven't
tested it yet (build running, but it'll take a while).  However, I
don't think they'll solve _this_ problem.

Some of the issue here is that we use a mixture of assembly macros
and preprocessor for the ARM bitops - the ARM bitops are created
with an assembly macro which contains some pre-processor expanded
macros (eg, EXPORT_SYMBOL()).

This means that the actual symbol being exported is not known to
the preprocessor, so doing the "__is_defined(__KSYM_##sym)" inside
"EXPORT_SYMBOL(\name)" becomes "__is_defined(__KSYM_\name)" to the
preprocessor.  As "__KSYM_\name" is never defined, it always comes
out as zero, hence we always use __cond_export_sym_0, which omits
the symbol export from the assembly macro definition:

 .macro bitop, name, instr
.globl \name ; .align 0 ; \name:

...

.type \name, %function; .size \name, .-\name

 .endm

In other words, using preprocessor macros inside an assembly macro
may not work as expected, and now leads to config-specific failures.

-- 
RMK's Patch system: http://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line: currently at 9.6Mbps down 400kbps up
according to speedtest.net.

Re: [PATCH 1/2] kbuild: provide include/asm/asm-prototypes.h for ARM

[Nicholas Piggin]

On Sun, 20 Nov 2016 19:12:57 +0000
Russell King - ARM Linux <linux@armlinux.org.uk> wrote:

> On Sun, Nov 20, 2016 at 10:32:50AM -0800, Linus Torvalds wrote:
> > On Sun, Nov 20, 2016 at 5:21 AM, Russell King - ARM Linux
> > <linux@armlinux.org.uk> wrote:  
> > > On Tue, Oct 25, 2016 at 07:32:00PM +1100, Nicholas Piggin wrote:  
> > >>
> > >> Michal, what's your thoughts? If you merge my patch 2/2 and skip 1/2, it
> > >> should not give any new build warnings or errors, so then arch patches can
> > >> go via arch trees. 1/2 could go in after everyone is up to date.  
> > >
> > > So what's the conclusion on this?  I've just had a failure due to
> > > CONFIG_TRIM_UNUSED_KSYMS reported on ARM, and it looks like (at
> > > least some of) patch 1 could resolve it.  
> > 
> > Hmm. I've got
> > 
> >   cc6acc11cad1 kbuild: be more careful about matching preprocessed asm
> > ___EXPORT_SYMBOL
> >   4efca4ed05cb kbuild: modversions for EXPORT_SYMBOL() for asm
> > 
> > in my tree. Is that sufficient, or do we still have issues?  
> 
> Hmm, those seem to have gone in during the last week, so I haven't
> tested it yet (build running, but it'll take a while).  However, I
> don't think they'll solve _this_ problem.
> 
> Some of the issue here is that we use a mixture of assembly macros
> and preprocessor for the ARM bitops - the ARM bitops are created
> with an assembly macro which contains some pre-processor expanded
> macros (eg, EXPORT_SYMBOL()).
> 
> This means that the actual symbol being exported is not known to
> the preprocessor, so doing the "__is_defined(__KSYM_##sym)" inside
> "EXPORT_SYMBOL(\name)" becomes "__is_defined(__KSYM_\name)" to the
> preprocessor.  As "__KSYM_\name" is never defined, it always comes
> out as zero, hence we always use __cond_export_sym_0, which omits
> the symbol export from the assembly macro definition:
> 
>  .macro bitop, name, instr
> .globl \name ; .align 0 ; \name:
> 
> ...
> 
> .type \name, %function; .size \name, .-\name
> 
>  .endm
> 
> In other words, using preprocessor macros inside an assembly macro
> may not work as expected, and now leads to config-specific failures.
> 

Yes, that's a limitation. cpp expansion we can handle, but not gas macros.
You will need Arnd's patches for ARM.

http://marc.info/?l=linux-kbuild&m=147732160529499&w=2

If that doesn't fix it for you, send me your .config offline and I'll set
up a cross compile to work on it.

Again, any arch always has the option of going back to doing asm exports
in the old style of putting them into a .c file, but hopefully you'll find
Arnd's reworked patches to be something you're willing to merge.

Thanks,
Nick