Re: [PATCH RFC] virtio-pci: new config layout: using memory BAR

[Paolo Bonzini <pbonzini@redhat.com>]

Il 29/05/2013 06:33, Rusty Russell ha scritto:
> Paolo Bonzini <pbonzini@redhat.com> writes:
>> Il 28/05/2013 19:32, Michael S. Tsirkin ha scritto:
>>>>>>> +
>>>>>>> +    switch (addr) {
>>>>>>> +    case offsetof(struct virtio_pci_common_cfg, device_feature_select):
>>>>>>> +        return proxy->device_feature_select;
>>>>>
>>>>> Oh dear no...  Please use defines like the rest of QEMU.
>>> Any good reason not to use offsetof?
>>
>> I'm not sure it's portable to use it in case labels.  IIRC, the
>> definition ((int)&(((T *)0)->field)) is not a valid C integer constant
>> expression.  Laszlo?
> 
> It's defined to yield an integer constant expression in the ISO standard
> (and I think ANSI too, though that's not at hand):

It's not in C89.  The oldest compiler QEMU cares about is GCC 4.2.  I
don't know if it has a builtin offsetof, probably it does.  But I'm not
sure about other users of virtio headers.

Paolo

> 
>     7.19, para 3:
> 
>         ...offsetof(type, member-designator)
>         which expands to an integer constant expression that has type
>         size_t, ...
> 
> The real question is whether compilers qemu cares about meet the
> standard (there's some evidence that older compilers fail this).  If
> not, we'll have to define them as raw offsets...
> 
> Cheers,
> Rusty.
>