Re: [PATCH v5 11/17] vt-d: Add API to update IRTE when VT-d PI is used

[Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>]

On Thu, Aug 13, 2015 at 02:33:03AM -0600, Jan Beulich wrote:
> >>> On 12.08.15 at 18:23, <konrad.wilk@oracle.com> wrote:
> > On Wed, Aug 12, 2015 at 10:35:32AM +0800, Feng Wu wrote:
> >> +    GET_IREMAP_ENTRY(ir_ctrl->iremap_maddr, remap_index, iremap_entries, p);
> >> +
> >> +    old_ire = new_ire = *p;
> >> +
> >> +    /* Setup/Update interrupt remapping table entry. */
> >> +    setup_posted_irte(&new_ire, pi_desc, gvec);
> >> +    ret = cmpxchg16b(p, &old_ire, &new_ire);
> >> +
> >> +    ASSERT(ret == *(__uint128_t *)&old_ire);
> >> +
> >> +    iommu_flush_cache_entry(p, sizeof(*p));
> > 
> > The other use sites of iommu_flush_cache_entry are sizeof(struct 
> > iremap_entry).
> > I think if you are doing this modification - then you should also have an
> > patch to modify the other code to be in sync.
> > 
> > Or just use the sizeof(struct ...).
> 
> I asked for this (to avoid proliferation of the bad style); I
> certainly wouldn't mind another cleanup patch to deal with the
> other use sites, but I certainly don't want to see this go back
> to sizeof(struct ...).

Aah, I missed that!

Thank you.
> 
> Jan
>