Bug 1964137

Description of problem:

pyverbs-tests fail with the following error when tested on rdma-virt-02 or rdma-virt-03 as an RDMA client. When run as RDMA server, no such issue exists.

test_tx_packet_reformat (tests.test_mlx5_flow.Mlx5MatcherTest) ... ERROR


Version-Release number of selected component (if applicable):


DISTRO=RHEL-8.5.0-20210521.n.1

Red Hat Enterprise Linux release 8.5 Beta (Ootpa)

Linux rdma-virt-03.lab.bos.redhat.com 4.18.0-305.8.el8.x86_64 #1 SMP Mon May 17 14:15:59 EDT 2021 x86_64 x86_64 x86_64 GNU/Linux

BOOT_IMAGE=(hd0,msdos1)/vmlinuz-4.18.0-305.8.el8.x86_64 root=/dev/mapper/rhel_rdma--virt--03-root ro intel_idle.max_cstate=0 processor.max_cstate=0 intel_iommu=on iommu=on console=tty0 rd_NO_PLYMOUTH crashkernel=auto resume=/dev/mapper/rhel_rdma--virt--03-swap rd.lvm.lv=rhel_rdma-virt-03/root rd.lvm.lv=rhel_rdma-virt-03/swap console=ttyS1,115200n81

rdma-core-35.0-1.el8.x86_64

linux-firmware-20201218-102.git05789708.el8.noarch

==> /sys/class/infiniband/mlx5_0/fw_ver <==
12.25.1020

==> /sys/class/infiniband/mlx5_1/fw_ver <==
12.25.1020

==> /sys/class/infiniband/mlx5_bond_0/fw_ver <==
14.27.1016
02:00.0 Ethernet controller: Broadcom Inc. and subsidiaries NetXtreme BCM5720 2-port Gigabit Ethernet PCIe
02:00.1 Ethernet controller: Broadcom Inc. and subsidiaries NetXtreme BCM5720 2-port Gigabit Ethernet PCIe
03:00.0 Ethernet controller: Broadcom Inc. and subsidiaries NetXtreme BCM5720 2-port Gigabit Ethernet PCIe
03:00.1 Ethernet controller: Broadcom Inc. and subsidiaries NetXtreme BCM5720 2-port Gigabit Ethernet PCIe
04:00.0 Infiniband controller: Mellanox Technologies MT27700 Family [ConnectX-4]
04:00.1 Infiniband controller: Mellanox Technologies MT27700 Family [ConnectX-4]
05:00.0 Ethernet controller: Mellanox Technologies MT27710 Family [ConnectX-4 Lx]
05:00.1 Ethernet controller: Mellanox Technologies MT27710 Family [ConnectX-4 Lx]


RDMA hosts tested:

Clients: rdma-virt-03
Servers: rdma-virt-02


Installed:
  python3-pyverbs-35.0-1.el8.x86_64    

Refer to the following beaker test job:

  https://beaker.engineering.redhat.com/jobs/5404400

  

How reproducible:

100%

Steps to Reproduce:
1. With the above RHEL8.5 build, install the following packages on both server and client hosts

2. execute the pyverbs tests

 ./run_tests.py -v --dev $HCA_ID

    <HCA_ID: mlx5_0>


3.

Actual results:

1. No issue on the RDMA server
2. Error on the RDMA client, as the following:

======================================================================
ERROR: test_tx_packet_reformat (tests.test_mlx5_flow.Mlx5MatcherTest)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/tmp/tmp.WYNlJJKsTn/rdma-core/tests/test_mlx5_flow.py", line 45, in func_wrapper
    MLX5_CMD_OP_QUERY_HCA_CAP_OUT_LEN)
  File "mlx5dv.pyx", line 223, in pyverbs.providers.mlx5.mlx5dv.Mlx5Context.devx_general_cmd
pyverbs.pyverbs_error.PyverbsRDMAError: DevX general command failed. Errno: 93, Protocol not supported

----------------------------------------------------------------------
Ran 183 tests in 5.379s

FAILED (errors=1, skipped=75)
---
- TEST RESULT FOR rdma-core
-   Test:   Run pyverbs tests
-   Result: FAIL
-   Return: 1
---
/mnt/tests/kernel/infiniband/pyverbs-tests
---
- TEST RESULT FOR pyverbs-tests
-   Test:   Remove temp directory
-   Result: PASS
-   Return: 0
---
      FAIL |      1 | Run pyverbs tests
** /kernel/infiniband/pyverbs-tests/standalone FAIL Score:0
Uploading resultoutputfile.log 


Expected results:

As shown on the RDMA server hosts:

test_tx_packet_reformat (tests.test_mlx5_flow.Mlx5MatcherTest) ... skipped 'NIC flow table does not support reformat'



Additional info:

Only client host of mlx5 MT27700 CX-4 IB0 incurs the above issue