Rusty Russell wrote:
> This patch modifies tun to allow a vringfd to specify the receive
> buffer. Because we can't copy to userspace in bh context, we queue
> like normal then use the "pull" hook to actually do the copy.
>
> More thought needs to be put into the possible races with ring
> registration and a simultaneous close, for example (see FIXME).
>
> We use struct virtio_net_hdr prepended to packets in the ring to allow
> userspace to receive GSO packets in future (at the moment, the tun
> driver doesn't tell the stack it can handle them, so these cases are
> never taken).
>
> Signed-off-by: Rusty Russell <[EMAIL PROTECTED]>
<snip>
> +
> +#ifdef CONFIG_VRINGFD
I think the rest of the code needs these for it to actually work without
VRINGFD.
> +static void unset_recv(void *_tun)
> +{
> + struct tun_struct *tun = _tun;
> +
> + tun->inring = NULL;
> +}
> +
> +/* Returns number of used buffers, or negative errno. */
> +static int pull_recv_skbs(void *_tun)
> +{
> + struct tun_struct *tun = _tun;
> + int err = 0, num_copied = 0;
> + struct sk_buff *skb;
> +
> + while ((skb = skb_dequeue(&tun->readq)) != NULL) {
> + struct iovec iov[1+MAX_SKB_FRAGS];
> + struct virtio_net_hdr gso = { 0 }; /* no info leak */
> + unsigned int iovnum = ARRAY_SIZE(iov);
> + unsigned long len;
> + int id;
> +
> + id = vring_get_buffer(tun->inring, iov, &iovnum, &len,
> + NULL, NULL, NULL);
> + if (id <= 0) {
> + err = id;
> + break;
> + }
Ah, I see now why you're passing something from the stack.
Regards,
Anthony Liguori
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/virtualization
