Hi Tom,
On 7 December 2014 at 19:01, Masahiro Yamada yamada.m@jp.panasonic.com wrote:
Hi Simon,
On Sun, 7 Dec 2014 14:44:21 -0700 Simon Glass sjg@chromium.org wrote:
#endif diff --git a/arch/x86/include/asm/linkage.h b/arch/x86/include/asm/linkage.h new file mode 100644 index 0000000..bdca72e --- /dev/null +++ b/arch/x86/include/asm/linkage.h @@ -0,0 +1,6 @@ +#ifndef _ASM_X86_LINKAGE_H +#define _ASM_X86_LINKAGE_H
+#define asmlinkage CPP_ASMLINKAGE __attribute__((regparm(0)))
Why CPP_ASMLINKAGE here?
The intention of the generic asmlinkage (defined in <linux/linkage.h>) is to add 'extern "C"' if __cplusplus is defined. The x86-specific asmlinkage should be supposed to add "__attribute__((regparm(0)))" onto that rather than replacing it.
Tested on chromebook_link:
Tested-by: Simon Glass sjg@chromium.org
With the above question answered, I'd like to apply this as it is a clean-up. Is it OK to so this independently of the ARM patch?
It must be accompanied with the ARM patch, otherwise the latter will get a conflict.
Will you apply both to u-boot-x86? I think it is OK because 2/2 is trivial enough.
Do you agree with this?
Regards, Simon