cls_flower: Fix inability to match GRE/IPIP packets

[Bodong Wang]

* Explain the bug
    When a packet of a new flow arrives in openvswitch kernel module, it dissects
    the packet and passes the extracted flow key to ovs-vswtichd daemon. If hw-
    offload configuration is enabled, the daemon creates a new TC flower entry to
    bypass openvswitch kernel module for the flow (TC flower can also offload flows
    to NICs but this time that does not matter).

    In this processing flow, I found the following issue in cases of GRE/IPIP
    packets.

    When ovs_flow_key_extract() in openvswitch module parses a packet of a new
    GRE (or IPIP) flow received on non-tunneling vports, it extracts information
    of the outer IP header for ip_proto/src_ip/dst_ip match keys.

    This means ovs-vswitchd creates a TC flower entry with IP protocol/addresses
    match keys whose values are those of the outer IP header. OTOH, TC flower,
    which uses flow_dissector (different parser from openvswitch module), extracts
    information of the inner IP header.

* How to test
    The following flow is an example to describe the issue in more detail.

       <----------- Outer IP -----------------> <---------- Inner IP ---------->
      +----------+--------------+--------------+----------+----------+----------+
      | ip_proto | src_ip | dst_ip | ip_proto | src_ip | dst_ip |
      | 47 (GRE) | 192.168.10.1 | 192.168.10.2 | 6 (TCP) | 10.0.0.1 | 10.0.0.2 |
      +----------+--------------+--------------+----------+----------+----------+

    In this case, TC flower entry and extracted information are shown as below:

      - ovs-vswitchd creates TC flower entry with:
          - ip_proto: 47
          - src_ip: 192.168.10.1
          - dst_ip: 192.168.10.2

      - TC flower extracts below for IP header matches:
          - ip_proto: 6
          - src_ip: 10.0.0.1
          - dst_ip: 10.0.0.2

    Thus, GRE or IPIP packets never match the TC flower entry, as each
    dissector behaves differently.

    IMHO, the behavior of TC flower (flow dissector) does not look correct,
    as ip_proto/src_ip/dst_ip in TC flower match means the outermost IP
    header information except for GRE/IPIP cases. This patch adds a new
    flow_dissector flag FLOW_DISSECTOR_F_STOP_BEFORE_ENCAP which skips
    dissection of the encapsulated inner GRE/IPIP header in TC flower
    classifier.

* What it could break.
    N/A

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This bug is awaiting verification that the linux-bluefield/5.4.0-1037.40 kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-needed-focal' to 'verification-done-focal'. If the problem still exists, change the tag 'verification-needed-focal' to 'verification-failed-focal'.

If verification is not done by 5 working days from today, this fix will be dropped from the source code, and this bug will be closed.

See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you!

[Launchpad Janitor (janitor)]

This bug was fixed in the package linux-bluefield - 5.4.0-1040.44

---------------
linux-bluefield (5.4.0-1040.44) focal; urgency=medium

  * focal/linux-bluefield: 5.4.0-1040.44 -proposed tracker (LP: #1978639)

  * fix ref leak when switching zones (LP: #1979009)
    - net/sched: act_ct: fix ref leak when switching zones

  * Fix XFRM flags validity check (LP: #1978967)
    - SAUCE: net/xfrm: Fix XFRM flags validity check

  [ Ubuntu: 5.4.0-121.137 ]

  * focal/linux: 5.4.0-121.137 -proposed tracker (LP: #1978666)
  * Packaging resync (LP: #1786013)
    - debian/dkms-versions -- update from kernel-versions (main/2022.05.30)
  * CVE-2022-28388
    - can: usb_8dev: usb_8dev_start_xmit(): fix double dev_kfree_skb() in error
      path
  * test_vxlan_under_vrf.sh in net from ubuntu_kernel_selftests failed (Check VM
    connectivity through VXLAN (underlay in the default VRF) [FAIL])
    (LP: #1871015)
    - selftests: net: test_vxlan_under_vrf: fix HV connectivity test
  * [UBUNTU 20.04] CPU-MF: add extended counter set definitions for new IBM z16
    (LP: #1974433)
    - s390/cpumf: add new extended counter set for IBM z16
  * [UBUNTU 20.04] KVM nesting support leaks too much memory, might result in
    stalls during cleanup (LP: #1974017)
    - KVM: s390: vsie/gmap: reduce gmap_rmap overhead
  * [UBUNTU 20.04] Null Pointer issue in nfs code running Ubuntu on IBM Z
    (LP: #1968096)
    - NFS: Fix up nfs_ctx_key_to_expire()

  [ Ubuntu: 5.4.0-120.136 ]

  * CVE-2022-21123 // CVE-2022-21125 // CVE-2022-21166
    - cpu/speculation: Add prototype for cpu_show_srbds()
    - x86/cpu: Add Jasper Lake to Intel family
    - x86/cpu: Add Lakefield, Alder Lake and Rocket Lake models to the to Intel
      CPU family
    - x86/cpu: Add another Alder Lake CPU to the Intel family
    - Documentation: Add documentation for Processor MMIO Stale Data
    - x86/speculation/mmio: Enumerate Processor MMIO Stale Data bug
    - x86/speculation: Add a common function for MD_CLEAR mitigation update
    - x86/speculation/mmio: Add mitigation for Processor MMIO Stale Data
    - x86/bugs: Group MDS, TAA & Processor MMIO Stale Data mitigations
    - x86/speculation/mmio: Enable CPU Fill buffer clearing on idle
    - x86/speculation/mmio: Add sysfs reporting for Processor MMIO Stale Data
    - x86/speculation/srbds: Update SRBDS mitigation selection
    - x86/speculation/mmio: Reuse SRBDS mitigation for SBDS
    - KVM: x86/speculation: Disable Fill buffer clear within guests
    - x86/speculation/mmio: Print SMT warning

 -- Zachary Tahenakos <email address hidden> Tue, 21 Jun 2022 13:59:23 -0400