On (Tue) 26 Jun 2012 [08:01:20], Anthony Liguori wrote:
> >>>+/* Send data from a char device over to the guest */
> >>>+static void chr_read(void *opaque, const void *buf, size_t size)
> >>>+{
> >>>+ VirtIORNG *vrng = opaque;
> >>>+ size_t len;
> >>>+ int offset;
> >>>+
> >>>+ if (!is_guest_ready(vrng)) {
> >>>+ return;
> >>>+ }
> >>>+
> >>>+ offset = 0;
> >>>+ while (offset< size) {
> >>>+ if (!pop_an_elem(vrng)) {
> >>>+ break;
> >>>+ }
> >>>+ len = iov_from_buf(vrng->elem.in_sg, vrng->elem.in_num,
> >>>+ buf + offset, 0, size - offset);
> >>>+ offset += len;
> >>>+
> >>>+ virtqueue_push(vrng->vq,&vrng->elem, len);
> >>>+ vrng->popped = false;
> >>>+ }
> >>>+ virtio_notify(&vrng->vdev, vrng->vq);
> >>>+
> >>>+ /*
> >>>+ * Lastly, if we had multiple elems queued by the guest, and we
> >>>+ * didn't have enough data to fill them all, indicate we want more
> >>>+ * data.
> >>>+ */
> >>>+ len = pop_an_elem(vrng);
> >>>+ if (len) {
> >>>+ rng_backend_request_entropy(vrng->rng, size, chr_read, vrng);
> >>>+ }
> >>
> >>Because of this above while() loop, you won't see entropy requests
> >>for every request that comes from the guest depending on how data
> >>gets buffered in the socket.
> >
> >So the issue is we currently can't get the iov_size without popping
> >the elem from the vq.
>
> I think we could split out some of the logic in virtqueue_pop to
> implement a virtqueue_peek().
I remember vaguely I looked at it before, and can't recollect why I
didn't follow through with it. Will re-look and note it.
> >>>+
> >>>+static void virtio_rng_save(QEMUFile *f, void *opaque)
> >>>+{
> >>>+ VirtIORNG *vrng = opaque;
> >>>+
> >>>+ virtio_save(&vrng->vdev, f);
> >>>+
> >>>+ qemu_put_byte(f, vrng->popped);
> >>>+ if (vrng->popped) {
> >>>+ qemu_put_buffer(f, (unsigned char *)&vrng->elem,
> >>>+ sizeof(vrng->elem));
> >>>+ }
> >>
> >>Okay, new rule: if you copy a struct verbatim to savevm, your next 5
> >>patches will be automatically nacked.
> >>
> >>Seriously, this is an awful thing to do. I don't care if we do it
> >>in other places in the code. It's never the right thing to do.]
>
> Err, this came in with virtio-scsi too apparently.
>
> Paolo/Stefan, please fix this in virtio-scsi before the 1.2 release.
> This is really not something we want to have to maintain long term.
>
> >>This is an unpacked unaligned structure with non-fixed sized types
> >>in it. that are greater than a byte. This breaks across endianness,
> >>compiler versions, etc.
> >
> >Any ideas how to get it done? (CC'ed Juan).
>
> Just save the individual fields of the structure using the
> appropriate accessors. Then you don't have to worry about alignment
> and endianness.
I think we need an accessor for the whole for VirtQueueElem. That
keeps individual devices agnostic from struct changes.
Then, we would need a way to deal with the individual elements in that
struct, which is tricky. There are pointers there.
> We can't change the existing devices (sans virtio-scsi) because it's
> now part of the ABI but we can avoid making the same mistakes again.
Yup.
Amit
