[Xen-devel] [PATCH V2] libxc, libxenstore: make the headers C++-friendlier

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 23.01.13 at 10:19, Razvan Cojocaru <rzvncj@xxxxxxxxx> wrote:
>  static inline int _hvm_hw_fix_cpu(void *h) {
> -    struct hvm_hw_cpu *new=h;
> -    struct hvm_hw_cpu_compat *old=h;
> +    struct hvm_hw_cpu *newcpu = (struct hvm_hw_cpu *)h;
> +    struct hvm_hw_cpu_compat *old = (struct hvm_hw_cpu_compat *)h;
>  
>      /* If we copy from the end backwards, we should
>       * be able to do the modification in-place */
> -    new->error_code=old->error_code;
> -    new->pending_event=old->pending_event;
> -    new->tsc=old->tsc;
> -    new->msr_tsc_aux=0;
> +    newcpu->error_code = old->error_code;
> +    newcpu->pending_event = old->pending_event;
> +    newcpu->tsc = old->tsc;
> +    newcpu->msr_tsc_aux = 0;
>  
>      return 0;
>  }

Actually I think this needs to be done differently anyway, as the
way it currently is it isn't even correct. In particular, I don't
think there's anything keeping the compiler from re-ordering
the reads and writes (gcc: without -fno-strict-aliasing), thus
invalidating the purposeful backward processing.

By means of using

union hvm_hw_cpu_union {
    struct hvm_hw_cpu nat;
    struct hvm_hw_cpu_compat cmp;
};

this can be avoided in a compatible manner (and the renaming
can be done basically as a side effect).

Sorry for not having noticed on the first run through.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel