[dpdk-users] DPDK 17.05 unable to access Mellanox VF device

["Sinha, Anirban (Nokia - IN/Bangalore)" <anirban.sinha@nokia.com>]

Hi :

We are using DPDK version 17.05.1 and have installed 
MLNX_OFED_LINUX-4.5-1.0.1.0-rhel7.6-x86_64 on a centos 7.6 box. I have 
followed the instructions here in order to install OFED:

https://mails.dpdk.org/archives/users/2019-April/004076.html

and installed using:

./mlnxofedinstall --dpdk --upstream-libs --force
Then:
/etc/init.d/openibd force-restart

DPDK is compiled with the relevant PMDs as follows:

diff --git a/config/common_linuxapp b/config/common_linuxapp
index 5330215..5f2fc2e 100644
--- a/config/common_linuxapp
+++ b/config/common_linuxapp
@@ -55,3 +55,8 @@ CONFIG_RTE_LIBRTE_AVP_PMD=y
CONFIG_RTE_LIBRTE_NFP_PMD=y
CONFIG_RTE_LIBRTE_POWER=y
CONFIG_RTE_VIRTIO_USER=y
+
+CONFIG_RTE_LIBRTE_MLX5_PMD=y
+CONFIG_RTE_LIBRTE_MLX5_DEBUG=n
+CONFIG_RTE_LIBRTE_MLX5_TX_MP_CACHE=8


I have bound one of the VFs to dpdk as follows:

# dpdk-devbind.py --status-dev net

Network devices using DPDK-compatible driver
============================================

a8ba:00:02.0 'MT27710 Family [ConnectX-4 Lx Virtual Function] 1016' 
drv=igb_uio unused=

Network devices using kernel driver
===================================
8abb:00:02.0 'MT27710 Family [ConnectX-4 Lx Virtual Function] 1016' 
if=eth0 drv=mlx5_core unused=igb_uio Other Network devices
=====================

<none>

O/P from idv_devinfo:

# ibv_devinfo
hca_id: mlx5_0
transport: InfiniBand (0)
fw_ver: 14.25.8362
node_guid: 6045:bdff:fe72:c782
sys_image_guid: 0000:0000:0000:0000
vendor_id: 0x02c9
vendor_part_id: 4118
hw_ver: 0x80
board_id: MSF0010110035
phys_port_cnt: 1
port: 1
state: PORT_ACTIVE (4)
max_mtu: 4096 (5)
active_mtu: 1024 (3)
sm_lid: 0
port_lid: 0
port_lmc: 0x00

link_layer: Ethernet


Now, when I run testpmd application, I get this:


# ./testpmd -w a8ba:00:02.0 --vdev="net_vdev_netvsc0,iface=eth2" -- -i 
--port-topology=chained
EAL: Detected 2 lcore(s)
EAL: Probing VFIO support...
EAL: WARNING: cpu flags constant_tsc=yes nonstop_tsc=no -> using 
unreliable clock cycles !
EAL: Master lcore 0 is ready (tid=37756900;cpuset=[0])
EAL: PCI device a8ba:00:02.0 on NUMA socket -1
EAL: probe driver: 15b3:1016 net_mlx5
PMD: net_mlx5: cannot access device, is mlx5_ib loaded?
EAL: Requested device a8ba:00:02.0 cannot be used
EAL: failed to initialize net_vdev_netvsc0 device
EAL: Bus (virtual) probe failed.
EAL: FATAL: Cannot probe devicesEAL: Cannot probe devicesPANIC in main():
Cannot init EAL
5: [./testpmd() [0x45c520]]
4: [/lib64/libc.so.6(__libc_start_main+0xf5) [0x7fea35fb6555]]
3: [./testpmd(main+0x5e9) [0x4586d9]]
2: [./testpmd(__rte_panic+0xb0) [0x4509d2]]
1: [./testpmd(rte_dump_stack+0x1a) [0x4c12ca]]
Aborted (core dumped)


The kernel modules loaded are:

# cat /proc/modules | grep mlx
mlx5_fpga_tools 14392 0 - Live 0xffffffffc02f0000 (OE)
mlx5_ib 339996 0 - Live 0xffffffffc0623000 (OE)
ib_uverbs 121776 3 rdma_ucm,ib_ucm,mlx5_ib, Live 0xffffffffc05aa000 (OE)
mlx5_core 915439 2 mlx5_fpga_tools,mlx5_ib, Live 0xffffffffc04c9000 (OE)
mlxfw 18227 1 mlx5_core, Live 0xffffffffc02fa000 (OE)
mlx4_en 146509 0 - Live 0xffffffffc04a4000 (OE)
mlx4_ib 211832 0 - Live 0xffffffffc0432000 (OE)
ib_core 294554 10 
rdma_ucm,ib_ucm,rdma_cm,iw_cm,ib_ipoib,ib_cm,ib_umad,mlx5_ib,ib_uverbs,mlx4_ib, 
Live 0xffffffffc03e9000 (OE)
mlx4_core 360644 2 mlx4_en,mlx4_ib, Live 0xffffffffc0378000 (OE)
mlx_compat 28730 15 
rdma_ucm,ib_ucm,rdma_cm,iw_cm,ib_ipoib,ib_cm,ib_umad,mlx5_fpga_tools,mlx5_ib,ib_uverbs,mlx5_core,mlx4_en,mlx4_ib,ib_core,mlx4_core, 
Live 0xffffffffc02e7000 (OE)
devlink 48345 4 mlx5_core,mlx4_en,mlx4_ib,mlx4_core, Live 0xffffffffc02a3000

ptp 19231 3 mlx5_core,mlx4_en,hv_utils, Live 0xffffffffc0116000

$ lsmod | grep mlx
mlx5_fpga_tools        14392  0
mlx5_ib               339996  0
ib_uverbs             121776  3 mlx5_ib,ib_ucm,rdma_ucm
mlx5_core             915439  2 mlx5_ib,mlx5_fpga_tools
mlxfw                  18227  1 mlx5_core
mlx4_en               146509  0
mlx4_ib               211832  0
ib_core               294554  10 
rdma_cm,ib_cm,iw_cm,mlx4_ib,mlx5_ib,ib_ucm,ib_umad,ib_uverbs,rdma_ucm,ib_ipoib
mlx4_core             360644  2 mlx4_en,mlx4_ib
mlx_compat             28730  15 
rdma_cm,ib_cm,iw_cm,mlx4_en,mlx4_ib,mlx5_ib,mlx5_fpga_tools,ib_ucm,ib_core,ib_umad,ib_uverbs,mlx4_core,mlx5_core,rdma_ucm,ib_ipoib
devlink                48345  4 mlx4_en,mlx4_ib,mlx4_core,mlx5_core
ptp                    19231  3 hv_utils,mlx4_en,mlx5_core


Is there anything I am missing here? Are we missing any kernel drivers 
for this to work?

$ uname -r
3.10.0-957.27.2.el7.nsgv7.x86_64

Thanks for the help,
Ani