Re: [PATCH 1/2] tcg: add dup_const_tl wrapper

[Richard Henderson]

On 9/28/21 4:54 PM, Philipp Tomsich wrote:
> dup_const always generates a uint64_t, which may exceed the size of a
> target_long (generating warnings with recent-enough compilers).
>
> To ensure that we can use dup_const both for 64bit and 32bit targets,
> this adds dup_const_tl, which wraps dup_const and legalises the
> truncation to target_long by casting it to target_long.
>
> Signed-off-by: Philipp Tomsich <philipp.tomsich@vrull.eu>
>
> ---
>
>   include/tcg/tcg.h | 5 +++++
>   1 file changed, 5 insertions(+)
>
> diff --git a/include/tcg/tcg.h b/include/tcg/tcg.h
> index 44ccd86f3e..8f8a209600 100644
> --- a/include/tcg/tcg.h
> +++ b/include/tcg/tcg.h
> @@ -1272,6 +1272,11 @@ uint64_t dup_const(unsigned vece, uint64_t c);
>           : (qemu_build_not_reached_always(), 0))                    \
>        : dup_const(VECE, C))
>   
> +static inline target_long dup_const_tl(unsigned vece, uint64_t c)
> +{
> +    return (target_long)dup_const(vece, c);
> +}

While this works, it avoids the qemu_build_not_reached() sanity check within dup_const.  I 
think perhaps this should be like

#if TARGET_LONG_BITS == 64
# define dup_const_tl  dup_const
#else
# define dup_const_tl(VECE, C) \
   ... stuff, excluding the MO_64 case
   ... using 32-bit multiplier constants
   ... using (uint32_t)(dup_const)(VECE, C) at the end
#endif


r~