Re: [Xen-devel] [PATCH] x86/compat: fix compilation errors with clang 6

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 23.01.18 at 14:54, <roger.pau@xxxxxxxxxx> wrote:
> On Tue, Jan 23, 2018 at 06:34:44AM -0700, Jan Beulich wrote:
>> >>> On 23.01.18 at 13:38, <roger.pau@xxxxxxxxxx> wrote:
>> > Fix this by using pragma push/pop in order to store the current pragma
>> > value in the compiler stack and later restoring it.
>> 
>> For Linux this is fine all the way back to gcc 4.1 afaict, but what
>> about other OSes using gcc? In 4.1, only config/linux.h and a few
>> special config/i386/*.h define the necessary macro
>> (HANDLE_PRAGMA_PACK_PUSH_POP).
> 
> OK, then I have the following version:
> 
> diff --git a/xen/include/Makefile b/xen/include/Makefile
> index 1299b1962f..19066a33a0 100644
> --- a/xen/include/Makefile
> +++ b/xen/include/Makefile
> @@ -35,8 +35,13 @@ cppflags-y                := -include public/xen-compat.h 
> -DXEN_GENERATING_COMPA
>  cppflags-$(CONFIG_X86)    += -m32
>  
>  # 8-byte types are 4-byte aligned on x86_32 ...
> +ifeq ($(clang),y)
> +prefix-$(CONFIG_X86)      := \#pragma pack(push, 4)
> +suffix-$(CONFIG_X86)      := \#pragma pack(pop)
> +else
>  prefix-$(CONFIG_X86)      := \#pragma pack(4)
>  suffix-$(CONFIG_X86)      := \#pragma pack()
> +endif
>  
>  endif
>  
> 
> Since ATM that's a clang only issue I guess we can leave it like that.
> I'm not aware of anyway to know whether the compiler supports pragma
> pack with psuh/pop.
> 
> Let me know if that's fine and I will send out a formal v2.

Yes, that's fine indeed. Checking whether the construct is
supported would involve a test compile, which I think we should
bother doing only when we can't avoid this anymore.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel