Re: [Qemu-devel] [PATCH maybe-for-3.1?] linux-user: Fix compilation with clang 3.4

[Laurent Vivier]

On 30/11/2018 11:33, Thomas Huth wrote:
> Clang version 3.4.2 does not know the -Wpragmas option yet and bails
> out with an error when we try to disable it in linux-user/qemu.h.
> Fortunately, clang has a __has_warning() macro which allows us to add
> an explicit check for this option to fix this issue.
> 
> Fixes: 850d5e330a9c68cc998cecc02caf8a3c8d1ee8a3
> Signed-off-by: Thomas Huth <address@hidden>
> ---
>  Today, I noticed that building the linux-user targets is currently also
>  not working with Clang 3.4 anymore ... so in case we will roll out 3.1-rc4,
>  I think this patch should be considered, too. Otherwise, this certainly
>  does not justify another rc, so it can also be added later via stable.
> 
>  include/qemu/compiler.h | 4 ++++
>  linux-user/qemu.h       | 2 +-
>  2 files changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/include/qemu/compiler.h b/include/qemu/compiler.h
> index 6b92710..ca9bc85 100644
> --- a/include/qemu/compiler.h
> +++ b/include/qemu/compiler.h
> @@ -119,6 +119,10 @@
>  #define GCC_FMT_ATTR(n, m)
>  #endif
>  
> +#ifndef __has_warning
> +#define __has_warning(x) 0 /* compatibility with non-clang compilers */
> +#endif
> +
>  #ifndef __has_feature
>  #define __has_feature(x) 0 /* compatibility with non-clang compilers */
>  #endif
> diff --git a/linux-user/qemu.h b/linux-user/qemu.h
> index dd5771c..c7b3c37 100644
> --- a/linux-user/qemu.h
> +++ b/linux-user/qemu.h
> @@ -481,7 +481,7 @@ static inline int access_ok(int type, abi_ulong addr, 
> abi_ulong size)
>   *   https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83256 -- so we only
>   *   include the warning-suppression pragmas for clang
>   */
> -#ifdef __clang__
> +#if defined(__clang__) && __has_warning("-Wpragmas")

Perhaps we could remove the "-Wpragmas" and only use
"-Waddress-of-packed-member" with
__has_warning("-Waddress-of-packed-member"),
as -Wpragmas warning is there to avoid a complaint about an unknown
warning "-Waddress-of-packed-member"?

Thanks,
Laurent