Re: [qemu-s390x] [PATCH v6] s390x/pci: add common function measurement block

[Pierre Morel]

On 07/01/2019 10:11, Thomas Huth wrote:
> On 2019-01-07 10:08, Pierre Morel wrote:
> > On 04/01/2019 15:39, Cornelia Huck wrote:
> > > On Thu,  3 Jan 2019 11:17:00 +0100
> > > Pierre Morel <address@hidden> wrote:
> > >
> > > > +static void fmb_update(void *opaque)
> > > > +{
> > > > +    S390PCIBusDevice *pbdev = opaque;
> > > > +    int64_t t = qemu_clock_get_ms(QEMU_CLOCK_VIRTUAL);
> > > > +    int i;
> > > > +
> > > > +    /* Update U bit */
> > > > +    pbdev->fmb.last_update *= 2;
> > > > +    pbdev->fmb.last_update |= UPDATE_U_BIT;
> > > > +    if (fmb_do_update(pbdev, offsetof(ZpciFmb, last_update),
> > > > +                      pbdev->fmb.last_update,
> > > > +                      sizeof(((ZpciFmb *)0)->last_update))) {
> > >
> > > I really don't want to be a pain, but... this looks weird. Why not
> > > simply sizeof(pbdev->fmb.last_update)?
> >
> > Isn't it preferable to use the size of the destination?
>
> But "(ZpciFmb *)0" in the context of QEMU isn't the destination either,
> it's an invalid pointer. So I think I agree with Cornelia,
> "sizeof(pbdev->fmb.last_update)" looks much easier to read and
> understand here.
>
>   Thomas

sizeof() returns the field length, the pointer value is not relevant.
But I will do as you wish if you both prefer to use this.
Anyway, this should be the same.

regards,
Pierre

--
Pierre Morel
Linux/KVM/QEMU in Böblingen - Germany