Re: [PATCH net-next] veth: clear rx queue hint in veth_xmit

On 10/13/26 10:14, Toshiaki Makita wrote:

On 2026/10/09 13:04, Jiayuan Chen wrote:

quoted

With more tx queues on one end of a veth pair than rx queues on the
other, and RPS or generic XDP enabled on the receiving end, we get:

     veth1 received packet on queue 2, but number of RX queues is 2
     WARNING: net/core/dev.c:5212 at get_rps_cpu+0x560/0x1360
     Call Trace:
      <TASK>
      netif_rx_internal+0x1af/0x4c0
      __netif_rx+0x99/0x350
      veth_xmit+0x713/0xca0
      dev_hard_start_xmit+0x166/0x5f0
      __dev_queue_xmit+0x1797/0x42d0
      ip_finish_output2+0xa34/0x1f40
      __ip_finish_output+0x510/0x7e0
      ip_finish_output+0x2f/0x320
      ip_output+0x17a/0x3f0
      ip_send_skb+0x1bc/0x220
      ......

Easy to hit with "ethtool -L veth0 tx 4", "ethtool -L veth1 rx 2",
rps_cpus set on veth1 and a few flows sent over the pair [1].

veth_xmit() uses skb->queue_mapping to pick the peer rq, but never
clears it before veth_forward_skb(), so the rx side still sees the tx
queue index. Everything on the rx side that goes through
skb_get_rx_queue(), like get_rps_cpu() and netif_get_rxqueue() for
generic XDP, takes it as a recorded rx queue and warns once it is out
of range.

Clear it before handing the skb to the peer:

- It is the tx queue index of this device, it says nothing about the
   rx queue of the peer.

- The two sides don't even agree on the encoding. The tx side stores
   the index as is, the rx side stores index + 1 so that 0 can mean
   "not recorded". So tx queue k is read back as rx queue k - 1, and
   tx queue 0 as "not recorded".

- Commit 710ad98c363a ("veth: Do not record rx queue hint in
   veth_xmit") already decided that veth should not pass any queue
   hint to the peer. There is no tx->rx queue mapping to preserve, so
   nothing is lost by clearing it.

On NETDEV_TX_BUSY the skb goes back to the qdisc, which looks up the
txq from skb->queue_mapping to decide when to retry, so restore it
there, next to the existing __skb_push().

[1]: https://lore.kernel.org/netdev/156834bb-8e40-496e-9443-9d515fa18eab@linux.dev/ (local)

Fixes: 710ad98c363a ("veth: Do not record rx queue hint in veth_xmit")

I guess the splat can happen before this commit.
I'm not very sure but probably it has been possible since after 76ff5cc91935
(with ip link add numtxqueues & numrxqueues, instead of ethtool -L)

I think the above is correct; I'll update the tag accordingly

For the same reason I agree this is net-next material.

/P