Re: [PATCH v19 008/130] x86/tdx: Warning with 32bit build shift-count-overflow

[Huang, Kai]

On 26/02/2024 9:25 pm, isaku.yamahata@xxxxxxxxx wrote:
> From: Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
>
> This patch fixes the following warnings.
>
>     In file included from arch/x86/kernel/asm-offsets.c:22:
>     arch/x86/include/asm/tdx.h:92:87: warning: shift count >= width of type [-Wshift-count-overflow]
>     arch/x86/include/asm/tdx.h:20:21: note: expanded from macro 'TDX_ERROR'
>     #define TDX_ERROR                       _BITUL(63)
>
>                                             ^~~~~~~~~~
>
> Also consistently use ULL for TDX_SEAMCALL_VMFAILINVALID.
>
> Fixes: 527a534c7326 ("x86/tdx: Provide common base for SEAMCALL and TDCALL C wrappers")

+Kirill.

This kinda fix should be sent out as a separate patch.

> Signed-off-by: Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
> ---
>   arch/x86/include/asm/tdx.h | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/arch/x86/include/asm/tdx.h b/arch/x86/include/asm/tdx.h
> index 16be3a1e4916..1e9dcdf9912b 100644
> --- a/arch/x86/include/asm/tdx.h
> +++ b/arch/x86/include/asm/tdx.h
> @@ -17,9 +17,9 @@
>    * Bits 47:40 == 0xFF indicate Reserved status code class that never used by
>    * TDX module.
>    */
> -#define TDX_ERROR			_BITUL(63)
> +#define TDX_ERROR			_BITULL(63)
>   #define TDX_SW_ERROR			(TDX_ERROR | GENMASK_ULL(47, 40))
> -#define TDX_SEAMCALL_VMFAILINVALID	(TDX_SW_ERROR | _UL(0xFFFF0000))
> +#define TDX_SEAMCALL_VMFAILINVALID	(TDX_SW_ERROR | _ULL(0xFFFF0000))
>   
>   #define TDX_SEAMCALL_GP			(TDX_SW_ERROR | X86_TRAP_GP)
>   #define TDX_SEAMCALL_UD			(TDX_SW_ERROR | X86_TRAP_UD)

Both TDX guest and TDX host code depends on X86_64 in the Kconfig.  This 
issue seems due to asm-offsets.c includes <asm/tdx.h> unconditionally.

It doesn't make sense to generate any TDX related code in asm-offsets.h 
so I am wondering whether it is better to just make the inclusion of 
<asm/tdx.h> conditionally or move it the asm-offsets_64.c?


Kirill what's your opinion?

Btw after quick try seems I cannot reproduce this (w/o this KVM TDX 
patchset).  Isaku, could you share your .config?