Re: [RFC v2 07/15] vt-d: Add API to update IRTE when VT-d PI is used

["Wu, Feng" <feng.wu@intel.com>]

> -----Original Message-----
> From: Jan Beulich [mailto:JBeulich@suse.com]
> Sent: Monday, June 15, 2015 5:35 PM
> To: Wu, Feng
> Cc: andrew.cooper3@citrix.com; george.dunlap@eu.citrix.com; Tian, Kevin;
> Zhang, Yang Z; xen-devel@lists.xen.org; keir@xen.org
> Subject: RE: [RFC v2 07/15] vt-d: Add API to update IRTE when VT-d PI is used
> 
> >>> On 15.06.15 at 11:20, <feng.wu@intel.com> wrote:
> 
> >
> >> -----Original Message-----
> >> From: Jan Beulich [mailto:JBeulich@suse.com]
> >> Sent: Friday, June 12, 2015 6:35 PM
> >> To: Wu, Feng
> >> Cc: andrew.cooper3@citrix.com; george.dunlap@eu.citrix.com; Tian, Kevin;
> >> Zhang, Yang Z; xen-devel@lists.xen.org; keir@xen.org
> >> Subject: RE: [RFC v2 07/15] vt-d: Add API to update IRTE when VT-d PI is used
> >>
> >> >>> On 12.06.15 at 11:40, <feng.wu@intel.com> wrote:
> >> >> -----Original Message-----
> >> >> From: Jan Beulich [mailto:JBeulich@suse.com]
> >> >> >>> On 08.05.15 at 11:07, <feng.wu@intel.com> wrote:
> >> >> > +static inline void setup_posted_irte(
> >> >> > +    struct iremap_entry *new_ire, struct pi_desc *pi_desc, uint8_t
> >> gvec)
> >> >> > +{
> >> >> > +    new_ire->post.urg = 0;
> >> >> > +    new_ire->post.vector = gvec;
> >> >> > +    new_ire->post.pda_l = (((u64)virt_to_maddr(pi_desc)) >>
> >> >> > +                           (32 - PDA_LOW_BIT)) &
> >> PDA_MASK(LOW);
> >> >> > +    new_ire->post.pda_h = (((u64)virt_to_maddr(pi_desc)) >> 32) &
> >> >> > +                           PDA_MASK(HIGH);
> >> >>
> >> >> Looking at this another time I can't see what the and-ing with
> >> >> PAD_MASK() is supposed to be good for.
> >> >
> >> > I cannot understand this well. Do you mean we don't need and
> PDA_MASK()
> >> > here?
> >>
> >> Correct - the bitfield width (where the data gets stored into) already
> >> enforces the intended truncation afaict.
> >
> > We may not need PDA_MASK(HIGH), but is PDA_MASK(LOW) really
> unnecessary
> > here?
> 
> I think so - iirc the bitfield is exactly the width you need (or if it
> wasn't, that would be a mistake). What you can't get rid of is the
> shifting.

You are right. I got the same result after removing the PDA_MASK.

Thanks,
Feng

> 
> Jan