clang 15 build error

clang 15 build error

[Borislav Petkov]

Hey,

before I go dig into this, any ideas? The likelyhood that you know what the
issue is, is high: :-P

That's 32-bit allnoconfig of latest Linus master with Ubuntu clang version
15.0.7:

mm/maccess.c:41:3: error: invalid output size for constraint '=a'
                copy_from_kernel_nofault_loop(dst, src, size, u64, Efault);
                ^
mm/maccess.c:22:3: note: expanded from macro 'copy_from_kernel_nofault_loop'
                __get_kernel_nofault(dst, src, type, err_label);        \
                ^
./arch/x86/include/asm/uaccess.h:629:18: note: expanded from macro '__get_kernel_nofault'
        __get_user_size(*((type *)(dst)), (__force type __user *)(src), \
                        ^
mm/maccess.c:41:3: error: invalid output size for constraint '=a'
mm/maccess.c:22:3: note: expanded from macro 'copy_from_kernel_nofault_loop'
                __get_kernel_nofault(dst, src, type, err_label);        \
                ^
./arch/x86/include/asm/uaccess.h:629:18: note: expanded from macro '__get_kernel_nofault'
        __get_user_size(*((type *)(dst)), (__force type __user *)(src), \
                        ^
2 errors generated.
make[3]: *** [scripts/Makefile.build:287: mm/maccess.o] Error 1
make[3]: *** Waiting for unfinished jobs....
make[2]: *** [scripts/Makefile.build:556: mm] Error 2
make[2]: *** Waiting for unfinished jobs....
make[1]: *** [/home/amd/kernel/linux/Makefile:2010: .] Error 2
make: *** [Makefile:248: __sub-make] Error 2

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Re: clang 15 build error

[Nathan Chancellor]

Hey Boris,

On Wed, Oct 15, 2025 at 04:33:12PM +0200, Borislav Petkov wrote:
> before I go dig into this, any ideas? The likelyhood that you know what the
> issue is, is high: :-P
> 
> That's 32-bit allnoconfig of latest Linus master with Ubuntu clang version
> 15.0.7:
> 
> mm/maccess.c:41:3: error: invalid output size for constraint '=a'
>                 copy_from_kernel_nofault_loop(dst, src, size, u64, Efault);
>                 ^
> mm/maccess.c:22:3: note: expanded from macro 'copy_from_kernel_nofault_loop'
>                 __get_kernel_nofault(dst, src, type, err_label);        \
>                 ^
> ./arch/x86/include/asm/uaccess.h:629:18: note: expanded from macro '__get_kernel_nofault'
>         __get_user_size(*((type *)(dst)), (__force type __user *)(src), \
>                         ^
> mm/maccess.c:41:3: error: invalid output size for constraint '=a'
> mm/maccess.c:22:3: note: expanded from macro 'copy_from_kernel_nofault_loop'
>                 __get_kernel_nofault(dst, src, type, err_label);        \
>                 ^
> ./arch/x86/include/asm/uaccess.h:629:18: note: expanded from macro '__get_kernel_nofault'
>         __get_user_size(*((type *)(dst)), (__force type __user *)(src), \
>                         ^
> 2 errors generated.

Heh, I am aware of it but I have not had a chance to actually dig into
see what is going wrong here :/

https://github.com/ClangBuiltLinux/linux/issues/2121

I guess that LLVM is unhappy about a 64-bit variable being passed into
the constraint for eax on 32-bit?

https://github.com/llvm/llvm-project/blob/ce60a03573a4d9f28dd6de42f61f0a3c4b0f2be9/clang/test/CodeGen/X86/x86_32-inline-asm.c#L50
https://github.com/llvm/llvm-project/commit/974131ea8825041673a8c8dda7c42f9a9376eb17

Cheers,
Nathan