[PATCH v2] arm: Replace ASSEMBLY with ASSEMBLER in uapi headers

[PATCH v2] arm: Replace ASSEMBLY with ASSEMBLER in uapi headers

[Nick Huang]

While the transition from __ASSEMBLY__ to __ASSEMBLER__ is not a regression,
the use of the modern __ASSEMBLER__ macro requires a sufficiently recent
compiler (GCC 3.0+).

In some environments, userland might still be using ancient tools like
GCC 2.95.x, which does not natively define __ASSEMBLER__. To provide a
complete fix and prevent silent build issues or header ambiguity, add
an #error guard to catch unsupported toolchains.

Changes in v2:
- Added explicit #error check for GCC versions earlier than 3.0.
- Updated commit message to clarify the rationale for legacy toolchain support.
- Link to v1: https://lore.kernel.org/all/20260218133952.5356-1-sef1548@gmail.com/

Signed-off-by: Nick Huang <sef1548@gmail.com>
---
 arch/arm/include/uapi/asm/ptrace.h | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/arch/arm/include/uapi/asm/ptrace.h b/arch/arm/include/uapi/asm/ptrace.h
index 8896c23cc..51cfefeae 100644
--- a/arch/arm/include/uapi/asm/ptrace.h
+++ b/arch/arm/include/uapi/asm/ptrace.h
@@ -119,7 +119,12 @@
 #define PT_DATA_ADDR		0x10004
 #define PT_TEXT_END_ADDR	0x10008
 
-#ifndef __ASSEMBLY__
+#if defined(__GNUC__) && (__GNUC__ <  3)
+# error "GCC 3.0+ is required for proper __ASSEMBLER__ support. \
+Your compiler is too old to safely handle modern kernel assembly headers."
+#endif
+
+#ifndef __ASSEMBLER__ 
 
 /*
  * This struct defines the way the registers are stored on the
@@ -158,6 +163,6 @@ struct pt_regs {
 #define ARM_VFPREGS_SIZE ( 32 * 8 /*fpregs*/ + 4 /*fpscr*/ )
 
 
-#endif /* __ASSEMBLY__ */
+#endif /* __ASSEMBLER__  */
 
 #endif /* _UAPI__ASM_ARM_PTRACE_H */
-- 
2.43.0

Re: [PATCH v2] arm: Replace ASSEMBLY with ASSEMBLER in uapi headers

[Thomas Weißschuh]

On 2026-02-28 12:13:11+0000, Nick Huang wrote:
> While the transition from __ASSEMBLY__ to __ASSEMBLER__ is not a regression,
> the use of the modern __ASSEMBLER__ macro requires a sufficiently recent
> compiler (GCC 3.0+).
> 
> In some environments, userland might still be using ancient tools like
> GCC 2.95.x, which does not natively define __ASSEMBLER__. To provide a
> complete fix and prevent silent build issues or header ambiguity, add
> an #error guard to catch unsupported toolchains.
> 
> Changes in v2:
> - Added explicit #error check for GCC versions earlier than 3.0.
> - Updated commit message to clarify the rationale for legacy toolchain support.
> - Link to v1: https://lore.kernel.org/all/20260218133952.5356-1-sef1548@gmail.com/
> 
> Signed-off-by: Nick Huang <sef1548@gmail.com>
> ---
>  arch/arm/include/uapi/asm/ptrace.h | 9 +++++++--
>  1 file changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/arm/include/uapi/asm/ptrace.h b/arch/arm/include/uapi/asm/ptrace.h
> index 8896c23cc..51cfefeae 100644
> --- a/arch/arm/include/uapi/asm/ptrace.h
> +++ b/arch/arm/include/uapi/asm/ptrace.h
> @@ -119,7 +119,12 @@
>  #define PT_DATA_ADDR		0x10004
>  #define PT_TEXT_END_ADDR	0x10008
>  
> -#ifndef __ASSEMBLY__
> +#if defined(__GNUC__) && (__GNUC__ <  3)
> +# error "GCC 3.0+ is required for proper __ASSEMBLER__ support. \
> +Your compiler is too old to safely handle modern kernel assembly headers."
> +#endif

Having this check in a random, architecture-specific header does not
make much sense. It should go into a generic header if we want to have it.
Or we can have some generic postprocessing in 'headers_install'.
But as most other architecture UAPI headers already have switched to
__ASSEMBLER__, it seems nobody will be affected by this, so this check
should be unnecessary.

> +
> +#ifndef __ASSEMBLER__ 
>  
>  /*
>   * This struct defines the way the registers are stored on the
> @@ -158,6 +163,6 @@ struct pt_regs {
>  #define ARM_VFPREGS_SIZE ( 32 * 8 /*fpregs*/ + 4 /*fpscr*/ )
>  
>  
> -#endif /* __ASSEMBLY__ */
> +#endif /* __ASSEMBLER__  */
>  
>  #endif /* _UAPI__ASM_ARM_PTRACE_H */
> -- 
> 2.43.0
> 

Re: [PATCH v2] arm: Replace ASSEMBLY with ASSEMBLER in uapi headers

[Nick Huang]

Thomas Weißschuh <linux@weissschuh.net> 於 2026年2月28日週六 下午8:20寫道：

> > diff --git a/arch/arm/include/uapi/asm/ptrace.h b/arch/arm/include/uapi/asm/ptrace.h
> > index 8896c23cc..51cfefeae 100644
> > --- a/arch/arm/include/uapi/asm/ptrace.h
> > +++ b/arch/arm/include/uapi/asm/ptrace.h
> > @@ -119,7 +119,12 @@
> >  #define PT_DATA_ADDR         0x10004
> >  #define PT_TEXT_END_ADDR     0x10008
> >
> > -#ifndef __ASSEMBLY__
> > +#if defined(__GNUC__) && (__GNUC__ <  3)
> > +# error "GCC 3.0+ is required for proper __ASSEMBLER__ support. \
> > +Your compiler is too old to safely handle modern kernel assembly headers."
> > +#endif
>
> Having this check in a random, architecture-specific header does not
> make much sense. It should go into a generic header if we want to have it.
> Or we can have some generic postprocessing in 'headers_install'.
> But as most other architecture UAPI headers already have switched to
> __ASSEMBLER__, it seems nobody will be affected by this, so this check
> should be unnecessary.
>
Hi Thomas Weißschu
I'm aware that other architectures have transitioned to __ASSEMBLER__.
However, I implemented this check here because Maciej W. Rozycki
specifically suggested this approach during our discussion.

-- 
Regards,
Nick Huang

Re: [PATCH v2] arm: Replace ASSEMBLY with ASSEMBLER in uapi headers

[Maciej W. Rozycki]

On Sat, 28 Feb 2026, Nick Huang wrote:

> > But as most other architecture UAPI headers already have switched to
> > __ASSEMBLER__, it seems nobody will be affected by this, so this check
> > should be unnecessary.

 It seems to whom?  Have you got a reference to a discussion with libc 
maintainers where the consensus has been that support for older compilers 
can be dropped?  Why does it remain in the GNU C library?

> I'm aware that other architectures have transitioned to __ASSEMBLER__.
> However, I implemented this check here because Maciej W. Rozycki
> specifically suggested this approach during our discussion.

 I asked a question whether this is relevant.  The outcome could well be 
that it is not, but with data available at hand such a conclusion cannot 
be made in my opinion.

  Maciej

Re: [PATCH v2] arm: Replace ASSEMBLY with ASSEMBLER in uapi headers

[Thomas Weißschuh]

On 2026-03-01 20:12:11+0000, Maciej W. Rozycki wrote:
> On Sat, 28 Feb 2026, Nick Huang wrote:
> 
> > > But as most other architecture UAPI headers already have switched to
> > > __ASSEMBLER__, it seems nobody will be affected by this, so this check
> > > should be unnecessary.
> 
>  It seems to whom?  Have you got a reference to a discussion with libc 
> maintainers where the consensus has been that support for older compilers 
> can be dropped?  Why does it remain in the GNU C library?

To me, based on the fact that nobody has complained so far about the
other architectures. But I don't really want to advocate for dropping
compatibility.

> > I'm aware that other architectures have transitioned to __ASSEMBLER__.
> > However, I implemented this check here because Maciej W. Rozycki
> > specifically suggested this approach during our discussion.
> 
>  I asked a question whether this is relevant.  The outcome could well be 
> that it is not, but with data available at hand such a conclusion cannot 
> be made in my opinion.

Fair enough. I have no issues with retaining the compatibility.
But the approach in this patch is counterproductive to compatibility.
For regular C sources, which I assume are most users of the UAPI
headers, the __ASSEMBLER__ or __ASSEMBLY__ are wholly irrelevant.
But with this patch they will be hard broken. Also UAPI consumers which
might be perfectly happy to manually define __ASSEMBLY__ will now run
into a hard error. So this check is strictly *less* compatible with
exactly those old compilers.

Given that there is a general switch to __ASSEMBLY__ for regular kernel
headers, the chances are high that adding these manual version checks to
the UAPI headers will be forgotten frequently and will require a lot of
fixups. If we want to stay compatible, we should do the following:

diff --git a/scripts/headers_install.sh b/scripts/headers_install.sh
index 9c15e748761c..2f1d1767ca26 100755
--- a/scripts/headers_install.sh
+++ b/scripts/headers_install.sh
@@ -36,6 +36,7 @@ sed -E -e '
        s/(^|[^a-zA-Z0-9])__packed([^a-zA-Z0-9_]|$)/\1__attribute__((packed))\2/g
        s/(^|[[:space:](])(inline|asm|volatile)([[:space:](]|$)/\1__\2__\3/g
        s@#(ifndef|define|endif[[:space:]]*/[*])[[:space:]]*_UAPI@#\1 @
+       s/__ASSEMBLY__/__ASSEMBLER__/g
 ' $INFILE > $TMPFILE || exit 1

 scripts/unifdef -U__KERNEL__ -D__EXPORTED_HEADERS__ $TMPFILE > $OUTFILE


Then we can convert all UAPI headers to the in-kernel style using
__ASSEMBLY__ and drop -DASSEMBLER from kbuild. Userspace will still
work as before and if we come to the point to drop compatibility,
we can do so far all UAPI headers at the same time, without any
inconsistencies.

Right now I think the consistency argument is enough that this should
have been done from the beginning. Your thoughts?


Thomas

Re: [PATCH v2] arm: Replace ASSEMBLY with ASSEMBLER in uapi headers

[Nick Huang]

On Sun, 1 Mar 2026, Maciej W. Rozycki wrote:

> It seems to whom?  Have you got a reference to a discussion with libc
> maintainers where the consensus has been that support for older compilers
> can be dropped?  Why does it remain in the GNU C library?

I have now consulted the glibc maintainers directly on this question.
The thread is available here:

  https://inbox.sourceware.org/libc-alpha/mvm342hx8tf.fsf@suse.de/

Andreas Schwab (SUSE Labs) responded:

  "Since the compiler will define __ASSEMBLER__ automatically when
   preprocessing assembler-with-cpp files, there is no need to change
   anything on the glibc side.  The places that define __ASSEMBLY__ for
   that purpose will just become redundant."

Best regards,
Nick Huang <sef1548@gmail.com>

Re: [PATCH v2] arm: Replace ASSEMBLY with ASSEMBLER in uapi

[Nick Huang]

Add a compatibility shim for __ASSEMBLER__ to support older GCC versions.
This bridge ensures that toolchains defining only __ASSEMBLY__ can correctly identify assembly stages.
This block can be removed once consensus is reached to drop support for these legacy compilers. If this version is acceptable, I will send it out in the v3 patch.

diff --git a/arch/arm/include/uapi/asm/ptrace.h b/arch/arm/include/uapi/asm/ptrace.h
index 8896c23cc..427eb70bd 100644
--- a/arch/arm/include/uapi/asm/ptrace.h
+++ b/arch/arm/include/uapi/asm/ptrace.h
@@ -119,7 +119,11 @@
 #define PT_DATA_ADDR		0x10004
 #define PT_TEXT_END_ADDR	0x10008
 
-#ifndef __ASSEMBLY__
+#if !defined(__ASSEMBLER__) && defined(__ASSEMBLY__)
+#define __ASSEMBLER__ __ASSEMBLY__
+#endif
+
+#ifndef __ASSEMBLER__ 
 
 /*
  * This struct defines the way the registers are stored on the
@@ -158,6 +162,6 @@ struct pt_regs {
 #define ARM_VFPREGS_SIZE ( 32 * 8 /*fpregs*/ + 4 /*fpscr*/ )
 
 
-#endif /* __ASSEMBLY__ */
+#endif /* __ASSEMBLER__  */
 
 #endif /* _UAPI__ASM_ARM_PTRACE_H */
-- 
2.43.0

Re: [PATCH v2] arm: Replace ASSEMBLY with ASSEMBLER in uapi

[Thomas Weißschuh]

Hi Nick,

thanks for the update. After our last discussion I prepared a
supplemental patch for this, but of course forgot to actually
send it out.

On 2026-03-09 15:24:23+0000, Nick Huang wrote:
> Add a compatibility shim for __ASSEMBLER__ to support older GCC versions.
> This bridge ensures that toolchains defining only __ASSEMBLY__ can correctly identify assembly stages.
> This block can be removed once consensus is reached to drop support for these legacy compilers. If this version is acceptable, I will send it out in the v3 patch.
> 
> diff --git a/arch/arm/include/uapi/asm/ptrace.h b/arch/arm/include/uapi/asm/ptrace.h
> index 8896c23cc..427eb70bd 100644
> --- a/arch/arm/include/uapi/asm/ptrace.h
> +++ b/arch/arm/include/uapi/asm/ptrace.h
> @@ -119,7 +119,11 @@
>  #define PT_DATA_ADDR		0x10004
>  #define PT_TEXT_END_ADDR	0x10008
 
> -#ifndef __ASSEMBLY__
> +#if !defined(__ASSEMBLER__) && defined(__ASSEMBLY__)
> +#define __ASSEMBLER__ __ASSEMBLY__
> +#endif

With [0] this hunk should also be unnecesarry. It would free all
UAPI authors from worrying about the problem at all.

[0] https://lore.kernel.org/lkml/20260309-uapi-assembly-v1-1-a7ebfbf14309@weissschuh.net/T/#u

> +#ifndef __ASSEMBLER__ 
>  
>  /*
>   * This struct defines the way the registers are stored on the
> @@ -158,6 +162,6 @@ struct pt_regs {
>  #define ARM_VFPREGS_SIZE ( 32 * 8 /*fpregs*/ + 4 /*fpscr*/ )
>  
>  
> -#endif /* __ASSEMBLY__ */
> +#endif /* __ASSEMBLER__  */
>  
>  #endif /* _UAPI__ASM_ARM_PTRACE_H */
> -- 
> 2.43.0
> 

Re: [PATCH v2] arm: Replace ASSEMBLY with ASSEMBLER in uapi

[Maciej W. Rozycki]

On Mon, 9 Mar 2026, Thomas Weißschuh wrote:

> thanks for the update. After our last discussion I prepared a
> supplemental patch for this, but of course forgot to actually
> send it out.

 I have now dug a bit deeper and realised that GCC used to have explicit 
-D__ASSEMBLER__ in the specs for `@assembler-with-cpp' before switching to 
the internal macro definition.  So I take my concern back, the original 
change is good to go in.  Apologies for the confusion.

  Maciej