[dpdk-dev] [PATCH v9 09/10] eal: replace rte_panic instances in init sequence

[Kevin Traynor]

On 04/26/2018 11:08 PM, Arnon Warshavsky wrote:
> Hi Kevin
> 
>     > +     if (rte_config_init() != 0)
>     > +             return -1;
>     > +
> 
>     I'm confused, is this deliberate? there is now two rte_config_init()
>     calls. 
> 
> 
> Please note that there are 2 eal.c files. One under linux and one under bsd
> There is a copy of rte_config_init() in each.
>  
> 
>     Aaron's comments on v4 about rte_eal_init_alert() are missed? and
>     it's missing the clear once and error number like the other returns.
> 
> 
> These comments were for code that I reverted to call panic again.
> As Aaron pointed out, the partial treatment of v4 did not only
> not-improve things, rather made them worst.
> This code along with other panic instances thrown from within threads
> will have a dedicated patchset.
> 

I was referencing the code above (but as it's a duplicate call, I don't
know if it was merge leftovers from a previous version). Other returns
from rte_eal_init() have something like:
		rte_eal_init_alert("Cannot init logging.");
		rte_errno = ENOMEM;
		rte_atomic32_clear(&run_once);
		return -1;

> Thanks
> /Arnon
>