Re: [Qemu-devel] [PATCH RFC] virtio: add virtqueue_fill_partial

[Michael S. Tsirkin]

On Mon, Apr 27, 2015 at 04:24:31PM +0200, Paolo Bonzini wrote:
> 
> 
> On 27/04/2015 16:18, Michael S. Tsirkin wrote:
> > +/*
> > + * Some devices dirty guest memory but don't want to tell guest about it. 
> > In
> > + * that case, use virtqueue_fill_partial: host_len is >= the amount of 
> > guest
> > + * memory actually written, guest_len is how much we guarantee to guest.
> > + * If you know exactly how much was written, use virtqueue_fill instead.
> 
> "If you never leave holes unwritten in the iov, use virtqueue_fill
> instead.

Is it just holes? Never data you are unsure about?

>  If the guest is not relying on iov boundaries, it should never
> be necessary to use this function."

You never know what guest does though, do you?

> The API looks okay though.
> 
> Paolo
> 
> > + */
> > +void virtqueue_fill_partial(VirtQueue *vq, const VirtQueueElement *elem,
> > +                            unsigned int host_len, unsigned int guest_len,
> > +                            unsigned int idx)
> >  {
> >      unsigned int offset;
> >      int i;
> >  
> > -    trace_virtqueue_fill(vq, elem, len, idx);
> > +    assert(host_len >= guest_len);
> > +
> > +    trace_virtqueue_fill(vq, elem, guest_len, idx);
> >  
> >      offset = 0;
> >      for (i = 0; i < elem->in_num; i++) {
> > -        size_t size = MIN(len - offset, elem->in_sg[i].iov_len);
> > +        size_t size = MIN(host_len - offset, elem->in_sg[i].iov_len);
> >  
> >          cpu_physical_memory_unmap(elem->in_sg[i].iov_base,
> >                                    elem->in_sg[i].iov_len,
> > @@ -269,7 +278,13 @@ void virtqueue_fill(VirtQueue *vq, const 
> > VirtQueueElement *elem,
> >  
> >      /* Get a pointer to the next entry in the used ring. */
> >      vring_used_ring_id(vq, idx, elem->index);
> > -    vring_used_ring_len(vq, idx, len);
> > +    vring_used_ring_len(vq, idx, guest_len);
> > +}
> > +