Re: [Qemu-devel] [PULL 16/16] tcg/i386: Use MOVDQA for TCG_TYPE_V128 load/store

[Richard Henderson]

On 5/28/19 1:46 PM, David Hildenbrand wrote:
> FWIW, this seems to be the easiest way:
> 
> diff --git a/target/s390x/cpu.h b/target/s390x/cpu.h
> index f0d9a6a36d..d363ae0fb3 100644
> --- a/target/s390x/cpu.h
> +++ b/target/s390x/cpu.h
> @@ -66,7 +66,7 @@ struct CPUS390XState {
>       * The floating point registers are part of the vector registers.
>       * vregs[0][0] -> vregs[15][0] are 16 floating point registers
>       */
> -    CPU_DoubleU vregs[32][2];  /* vector registers */
> +    CPU_DoubleU vregs[32][2] QEMU_ALIGNED(16);  /* vector registers */
>      uint32_t aregs[16];    /* access registers */
>      uint8_t riccb[64];     /* runtime instrumentation control */
>      uint64_t gscb[4];      /* guarded storage control */
> 
> 
> Makes it work for me again.

That's the right fix, and exactly the bug that I was hoping to find with
11e2bfef7990 ("tcg/i386: Use MOVDQA for TCG_TYPE_V128 load/store").


r~