Cannot set rte_flow QUEUE action for VF representor

https://bugs.dpdk.org/show_bug.cgi?id=107

            Bug ID: 107
           Summary: Cannot set rte_flow QUEUE action for VF representor
           Product: DPDK
           Version: 18.11
          Hardware: x86
                OS: Linux
            Status: CONFIRMED
          Severity: normal
          Priority: Normal
         Component: ethdev
          Assignee: [email protected]
          Reporter: [email protected]

Target Milestone: ---

Hi all,

I'm trying to direct a given flow directly to a VF's specific queue.

Steps to reproduce:
sudo ./testpmd -w "pci:03:00.1,representor=0" -- -i

EAL: Detected 16 lcore(s)
EAL: Detected 2 NUMA nodes
EAL: Multi-process socket /var/run/dpdk/rte/mp_socket
EAL: Probing VFIO support...
EAL: VFIO support initialized
EAL: PCI device 0000:03:00.1 on NUMA socket 0
EAL: probe driver: 15b3:1017 net_mlx5
net_mlx5: port 1 ioctl(SIOCETHTOOL, ETHTOOL_GSET) failed: Operation not supported
Interactive-mode selected
testpmd: create a new mbuf pool <mbuf_pool_socket_1>: n=203456, size=2176, socket=1
testpmd: preferred mempool ops selected: ring_mp_mc
testpmd: create a new mbuf pool <mbuf_pool_socket_0>: n=203456, size=2176, socket=0
testpmd: preferred mempool ops selected: ring_mp_mc Configuring Port 0 (socket 0)
Port 0: 98:03:9B:33:FE:E3
Configuring Port 1 (socket 0)
Port 1: 62:74:FA:85:79:13
Checking link statuses...
Done

testpmd> flow create 0 transfer ingress pattern eth dst is 1e:00:b6:cb:66:e4 / end actions port_id id 1 / queue index 5 / end
Caught error type 16 (specific action): cause: 0x7ffce7c70428, action not
supported: Operation not supported

testpmd> flow create 0 transfer ingress pattern eth dst is 1e:00:b6:cb:66:e4 / end actions port_id id 1 / end
Flow rule #0 created

testpmd> flow create 1 ingress pattern eth dst is 1e:00:b6:cb:66:e4 / end actions queue index 5 / end
Caught error type 15 (action configuration): cause: 0x7ffce7c72368, queue index out of range: Invalid argument

testpmd> flow create 1 ingress pattern eth dst is 1e:00:b6:cb:66:e4 / end actions queue index 0 / end
Flow rule #0 created

testpmd>

Expected results:
- I can understand the first line, doing both VF and queue at the same time is
hard to support.
- The second line works and send packets to the VF. (Why can't I use the VF action by the way? And why do I need transfer for PF to VF?)
- The third line does not work because the VF representor has a single queue (the real VF has 8). Now at the end of the EAL lines I get 'net_mlx5: port 1 ioctl(SIOCETHTOOL, ETHTOOL_GSET) failed: Operation not supported.'. Is that a failing ethtool -L ?
- However if I use queue 0, it is accepted but does not work anyway (I get a random RSS choosen index).

I hope the answer is not "line 4" should fail, because I'd really like to be able to do this without having to send rules from the VF itself (if that works).

Tom

You are receiving this mail because:
You are the assignee for the bug.