This reply is helpful to me,thanks a lot.




------------------ 原始邮件 ------------------
发件人: "Ferruh Yigit"<ferruh.yigit@intel.com>; 
发送时间: 2016年6月28日(星期二) 凌晨2:04
收件人: "成桦者别"<373755701@qq.com>; "users"<users@dpdk.org>; "Damjan Marion (damarion)"<damarion@cisco.com>; 
主题: Re: [dpdk-users] my problem aboutixgbe_rx_vec_dev_conf_condition_check is called by ixgbe_set_rx_function inixgbe



On 5/27/2016 2:32 PM, 成桦者别 wrote:
> hello,
> 
> 
> firstly ,gcc version is 3.4.3.
> 
> 
> "PMD: ixgbe_set_rx_function(): Port[0] doesn't meet Vector Rx preconditions or RTE_IXGBE_INC_VECTOR is not enabled" is outputed while running l2fwd  and l3fwd ,i GDB the l2fwd and l3fwd.
> i found ixgbe_rx_vec_dev_conf_condition_check has two definition ,one is:
> int __attribute__((weak))
> ixgbe_rx_vec_dev_conf_condition_check(struct rte_eth_dev __rte_unused *dev)
> {
> 	return -1;
> }
> 
> 
> the other is:
> int __attribute__((cold))
> ixgbe_rx_vec_dev_conf_condition_check(struct rte_eth_dev *dev)
> {
> ....
> }
> when i GDB the EXE, port_conf.rxmode.hw_ip_checksum was set to 0 .
> ixgbe_set_rx_function call the former at every turn ,i want to know why .
> 
> 
> any help would be appreciated.
> 

I don't know if this is from same person/team, but if not please check
following mail thread for same issue:

http://dpdk.org/ml/archives/dev/2016-June/042323.html

If issue is still valid for you:
Briefly, weak function is defined for the case RTE_IXGBE_INC_VECTOR
config option not enabled.

Which one to link against will be chosen on build time, if app linked
against first one, I would guess RTE_IXGBE_INC_VECTOR is not enabled,
can you please share the output of:
objdump -x librte_pmd_i40e.a| grep i40e_rx_vec_dev_conf_condition_check

Thanks,
ferruh