Re: [BUG] vhost_net: livelock in handle_rx() when GRO packet exceeds virtqueue capacity

["Michael S. Tsirkin" <mst@redhat.com>]

On Sun, Mar 01, 2026 at 07:10:06PM -0500, Michael S. Tsirkin wrote:
> On Sun, Mar 01, 2026 at 10:36:39PM +0000, ShuangYu wrote:
> > Hi,
> > 
> > We have hit a severe livelock in vhost_net on 6.18.x. The vhost
> > kernel thread spins at 100% CPU indefinitely in handle_rx(), and
> > QEMU becomes unkillable (stuck in D state).
> > [This is a text/plain messages]
> > 
> > Environment
> > -----------
> >   Kernel:  6.18.10-1.el8.elrepo.x86_64
> >   QEMU:    7.2.19
> >   Virtio:  VIRTIO_F_IN_ORDER is negotiated
> >   Backend: vhost (kernel)
> > 
> > Symptoms
> > --------
> >   - vhost-<pid> kernel thread at 100% CPU (R state, never yields)
> >   - QEMU stuck in D state at vhost_dev_flush() after receiving SIGTERM
> >   - kill -9 has no effect on the QEMU process
> >   - libvirt management plane deadlocks ("cannot acquire state change lock")
> > 
> > Root Cause
> > ----------
> > The livelock is triggered when a GRO-merged packet on the host TAP
> > interface (e.g., ~60KB) exceeds the remaining free capacity of the
> > guest's RX virtqueue (e.g., ~40KB of available buffers).
> > 
> > The loop in handle_rx() (drivers/vhost/net.c) proceeds as follows:
> > 
> >   1. get_rx_bufs() calls vhost_get_vq_desc_n() to fetch descriptors.
> >     It advances vq->last_avail_idx and vq->next_avail_head as it
> >     consumes buffers, but runs out before satisfying datalen.
> > 
> >   2. get_rx_bufs() jumps to err: and calls
> >     vhost_discard_vq_desc(vq, headcount, n), which rolls back
> >     vq->last_avail_idx and vq->next_avail_head.
> > 
> >     Critically, vq->avail_idx (the cached copy of the guest's
> >     avail->idx) is NOT rolled back. This is correct behavior in
> >     isolation, but creates a persistent mismatch:
> > 
> >       vq->avail_idx      = 108  (cached, unchanged)
> >       vq->last_avail_idx = 104  (rolled back)
> > 
> >   3. handle_rx() sees headcount == 0 and calls vhost_enable_notify().
> >     Inside, vhost_get_avail_idx() finds:
> > 
> >       vq->avail_idx (108) != vq->last_avail_idx (104)
> > 
> >     It returns 1 (true), indicating "new buffers available."
> >     But these are the SAME buffers that were just discarded.
> > 
> >   4. handle_rx() hits `continue`, restarting the loop.
> > 
> >   5. In the next iteration, vhost_get_vq_desc_n() checks:
> > 
> >       if (vq->avail_idx == vq->last_avail_idx)
> > 
> >     This is FALSE (108 != 104), so it skips re-reading the guest's
> >     actual avail->idx and directly fetches the same descriptors.
> > 
> >   6. The exact same sequence repeats: fetch -> too small -> discard
> >     -> rollback -> "new buffers!" -> continue. Indefinitely.
> > 
> > This appears to be a regression introduced by the VIRTIO_F_IN_ORDER
> > support, which added vhost_get_vq_desc_n() with the cached avail_idx
> > short-circuit check, and the two-argument vhost_discard_vq_desc()
> > with next_avail_head rollback. The mismatch between the rollback
> > scope (last_avail_idx, next_avail_head) and the check scope
> > (avail_idx vs last_avail_idx) was not present before this change.
> > 
> > bpftrace Evidence
> > -----------------
> > During the 100% CPU lockup, we traced:
> > 
> >   @get_rx_ret[0]:      4468052   // get_rx_bufs() returns 0 every time
> >   @peek_ret[60366]:    4385533   // same 60KB packet seen every iteration
> >   @sock_err[recvmsg]:        0   // tun_recvmsg() is never reached
> > 
> > vhost_get_vq_desc_n() was observed iterating over the exact same 11
> > descriptor addresses millions of times per second.
> > 
> > Workaround
> > ----------
> > Either of the following avoids the livelock:
> > 
> >   - Disable GRO/GSO on the TAP interface:
> >      ethtool -K <tap> gro off gso off
> > 
> >   - Switch from kernel vhost to userspace QEMU backend:
> >      <driver name='qemu'/> in libvirt XML
> > 
> > Bisect
> > ------
> > We have not yet completed a full git bisect, but the issue does not
> > occur on 6.17.x kernels which lack the VIRTIO_F_IN_ORDER vhost
> > support. We will follow up with a Fixes: tag if we can identify the
> > exact commit.
> > 
> > Suggested Fix Direction
> > -----------------------
> > In handle_rx(), when get_rx_bufs() returns 0 (headcount == 0) due to
> > insufficient buffers (not because the queue is truly empty), the code
> > should break out of the loop rather than relying on
> > vhost_enable_notify() to make that determination. For example, when
> > get_rx_bufs() returns r == 0 with datalen still > 0, this indicates a
> > "packet too large" condition, not a "queue empty" condition, and
> > should be handled differently.
> > 
> > Thanks,
> > ShuangYu
> 
> Hmm. On a hunch, does the following help? completely untested,
> it is night here, sorry.
> 
> 
> diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c
> index 2f2c45d20883..aafae15d5156 100644
> --- a/drivers/vhost/vhost.c
> +++ b/drivers/vhost/vhost.c
> @@ -1522,6 +1522,7 @@ static void vhost_dev_unlock_vqs(struct vhost_dev *d)
>  static inline int vhost_get_avail_idx(struct vhost_virtqueue *vq)
>  {
>  	__virtio16 idx;
> +	u16 avail_idx;
>  	int r;
>  
>  	r = vhost_get_avail(vq, idx, &vq->avail->idx);
> @@ -1532,17 +1533,19 @@ static inline int vhost_get_avail_idx(struct vhost_virtqueue *vq)
>  	}
>  
>  	/* Check it isn't doing very strange thing with available indexes */
> -	vq->avail_idx = vhost16_to_cpu(vq, idx);
> -	if (unlikely((u16)(vq->avail_idx - vq->last_avail_idx) > vq->num)) {
> +	avail_idx = vhost16_to_cpu(vq, idx);
> +	if (unlikely((u16)(avail_idx - vq->last_avail_idx) > vq->num)) {
>  		vq_err(vq, "Invalid available index change from %u to %u",
>  		       vq->last_avail_idx, vq->avail_idx);
>  		return -EINVAL;
>  	}
>  
>  	/* We're done if there is nothing new */
> -	if (vq->avail_idx == vq->last_avail_idx)
> +	if (avail_idx == vq->avail_idx)
>  		return 0;
>  
> +	vq->avail_idx == avail_idx;
> +

meaning 
	vq->avail_idx = avail_idx; 
of course

>  	/*
>  	 * We updated vq->avail_idx so we need a memory barrier between
>  	 * the index read above and the caller reading avail ring entries.