Ethernet Products
Determine ramifications of Intel® Ethernet products and technologies
4810 Discussions

RSS Support in IXGBEVF

MTang5
Beginner
3,228 Views

I am using 3.3.2 ixgbevf in guest and 5.0.4 ixgbe on host. I noticed that even though there are two RX queues for each VF, but pretty much all the packets go to the first RX queue. The RSS related features are also not supported. How to turn on the RSS support on ixgbevf? If it does not support RSS, what's the point having multiple queues in ixgbevf?

root@ubuntu14:~# ethtool -n eth1

2 RX rings available

rxclass: Cannot get RX class rule count: Operation not supported

RX classification rule retrieval failed

root@ubuntu14:~# ethtool -x eth1

Cannot get RX flow hash indirection table: Operation not permitted

root@ubuntu14:~# ethtool -g eth1

Ring parameters for eth1:

Pre-set maximums:

RX: 4096

RX Mini: 0

RX Jumbo: 0

TX: 4096

Current hardware settings:

RX: 512

RX Mini: 0

RX Jumbo: 0

TX: 1024

root@ubuntu14:~# ethtool -k eth1

Features for eth1:

rx-checksumming: on

tx-checksumming: on

tx-checksum-ipv4: on

tx-checksum-ip-generic: off [fixed]

tx-checksum-ipv6: on

tx-checksum-fcoe-crc: off [fixed]

tx-checksum-sctp: off [fixed]

scatter-gather: on

tx-scatter-gather: on

tx-scatter-gather-fraglist: off [fixed]

tcp-segmentation-offload: on

tx-tcp-segmentation: on

tx-tcp-ecn-segmentation: off [fixed]

tx-tcp6-segmentation: on

udp-fragmentation-offload: off [fixed]

generic-segmentation-offload: on

generic-receive-offload: on

large-receive-offload: off [fixed]

rx-vlan-offload: on

tx-vlan-offload: on

ntuple-filters: off [fixed]

receive-hashing: off [fixed]

highdma: on [fixed]

rx-vlan-filter: on

vlan-challenged: off [fixed]

tx-lockless: off [fixed]

netns-local: off [fixed]

tx-gso-robust: off [fixed]

tx-fcoe-segmentation: off [fixed]

tx-gre-segmentation: off [fixed]

tx-ipip-segmentation: off [fixed]

tx-sit-segmentation: off [fixed]

tx-udp_tnl-segmentation: off [fixed]

fcoe-mtu: off [fixed]

tx-nocache-copy: off

loopback: off [fixed]

rx-fcs: off [fixed]

rx-all: off [fixed]

tx-vlan-stag-hw-insert: off [fixed]

rx-vlan-stag-hw-parse: off [fixed]

rx-vlan-stag-filter: off [fixed]

l2-fwd-offload: off [fixed]

busy-poll: on [fixed]

0 Kudos
9 Replies
idata
Employee
1,868 Views

Hi Mintang,

 

 

Thank you for posting at Wired Communities. Please share what is the exact model of your network adapter and what is the system board(brand and model).

 

 

thanks,

 

wb

 

0 Kudos
MTang5
Beginner
1,868 Views

Intel Corporation 82599ES 10-Gigabit SFI/SFP+ Network Connection (rev 01).

0 Kudos
idata
Employee
1,868 Views

Hi Mintang,

 

 

Thank you for the information. Is this an onboard NIC? and what is the brand and model of your server?

 

 

Thanks,

 

wb

 

0 Kudos
idata
Employee
1,868 Views

Hi Mintang,

 

 

RSS is not supported on ixgbevf driver as ixgbevf driver is intended for virtual operating system. RSS is used on host operating system

 

http://www.intel.com/content/www/us/en/support/network-and-i-o/ethernet-products/000006703.html

 

 

. If you want to have multiple queue in ixgbevf (this is driver for virtual function) you can use vmdq features, please refer to this website for reference:

 

https://www.intel.com/content/www/us/en/ethernet-products/converged-network-adapters/io-acceleration-technology-vmdq.html

 

 

 

It is recommended to check with Ubuntu if vmdq is supported.

 

 

Thanks,

 

wb
0 Kudos
MTang5
Beginner
1,868 Views

VMDQ is not an option for us because we have to do SR-IOV. According to thehttp://dpdk.org/doc/guides/nics/intel_vf.html DPDK document, RSS is supported in VF but it has to set on host:

"On host, to enable VF RSS functionality, rx mq mode should be set as ETH_MQ_RX_VMDQ_RSS or ETH_MQ_RX_RSS mode, and SRIOV mode should be activated (max_vfs >= 1). It also needs config VF RSS information like hash function, RSS key, RSS key length."

"Note

The limitation for VF RSS on Intel® 82599 10 Gigabit Ethernet Controller is: The hash and key are shared among PF and all VF, the RETA table with 128 entries is also shared among PF and all VF; So it could not to provide a method to query the hash and reta content per VF on guest, while, if possible, please query them on host for the shared RETA information."

It is for the DPDK driver, but not the regular linux kernel driver. It looks to me that hardware supports it, just the linux kernel driver on host does not expose the interface to do it.

0 Kudos
idata
Employee
1,868 Views

Hi Mintang,

 

 

Thank you for the clarification. Are you using the DPDK? If yes, then all Intel® DPDK questions and technical problems including those regarding the Ethernet* Controllers for the Poll Mode Driver should be reported through the Intel® Premier Support site

 

http://premier.intel.com/premier http://premier.intel.com/premier or access your IBL account and click

 

the Intel® Premier Support link to enter issues under the Product Name

 

"Data Plane Development Kit (DPDK)", which are then routed to our support team

 

 

Thanks,

 

wb

 

 

 

0 Kudos
MTang5
Beginner
1,868 Views

I am not using DPDK. I just used the DPDK doc to show that RSS is supported on VF. Actually the http://elixir.free-electrons.com/linux/v4.1/source/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c# L109 ixgbe driver in the linux kernel also suggests that RSS is supported on VF. For 82599, the indirection table and RSS hash key cannot be changed on VF, but RSS should work on VF.

for (i = 0; i < adapter->num_vfs; i++) {

/* enable spoof checking for all VFs */

adapter->vfinfo[i].spoofchk_enabled = true;

/* We support VF RSS querying only for 82599 and x540

* devices at the moment. These devices share RSS

* indirection table and RSS hash key with PF therefore

* we want to disable the querying by default.

*/

adapter->vfinfo[i].rss_query_enabled = 0;

}

0 Kudos
idata
Employee
1,868 Views

Hi Mintang,

 

 

Thank you for the clarification.

 

 

regards,

 

wb

 

0 Kudos
idata
Employee
1,868 Views

Hi Mintang,

 

 

Further checking, please use DPDK option to configure.

 

 

Thanks,

 

sharon

 

0 Kudos
Reply