[dpdk-dev] [PATCH] ethdev: force offloading API rules

[Ferruh Yigit]

On 6/27/2018 9:31 AM, Ido Goshen wrote:
> I guess the error below relates to f52f1a6 ethdev: force offloading API rules

Hi Ido,

Yes it is related.

This specific error is related to the requested DEV_RX_OFFLOAD_CRC_STRIP, and
following patch should fix it because it adds to CRC strip capability to virtual
drivers:
https://patches.dpdk.org/patch/41359/

But perhaps we should also add some checks in examples as done in RSS check:
https://patches.dpdk.org/patch/41584/

There are a few pieces are moving, making it hard to manage, all these errors
should be detected and fixed by end of this week. Changes:
- We are removing old offload API and structus
- We are making ethdev more strict about what application request about offloads
and rss_hf
- We are in the process of deprecating CRC_STRIP flag


> 
>  
> 
> cgs at ubuntu:~/dpdk-next-net$ sudo examples/l2fwd/build/l2fwd -c 3 -n1 --no-huge
> --vdev=eth_pcap0,iface=dummy0 --vdev=eth_pcap1,iface=dummy1  -- -p 3 -T 1
> 
> EAL: Detected 4 lcore(s)
> 
> EAL: Detected 1 NUMA nodes
> 
> EAL: Multi-process socket /var/run/dpdk/rte/mp_socket
> 
> EAL: Probing VFIO support...
> 
> EAL: Started without hugepages support, physical addresses not available
> 
> EAL: PCI device 0000:02:01.0 on NUMA socket -1
> 
> EAL:   Invalid NUMA socket, default to 0
> 
> EAL:   probe driver: 8086:100f net_e1000_em
> 
> EAL: PCI device 0000:02:06.0 on NUMA socket -1
> 
> EAL:   Invalid NUMA socket, default to 0
> 
> EAL:   probe driver: 8086:100f net_e1000_em
> 
> MAC updating enabled
> 
> Lcore 0: RX port 0
> 
> Lcore 1: RX port 1
> 
> Initializing port 0... ethdev port_id=0 requested Rx offloads 0x1000 doesn't
> match Rx offloads capabilities 0x0 in rte_eth_dev_configure()
> 
>  
> 
> EAL: Error - exiting with code: 1
> 
>   Cause: Cannot configure device: err=-22, port=0
> 
>  
> 
> Please advice
>