[dpdk-dev] [PATCH 0/3] Rename field name for RX/TX queue start/stop

[Ouyang, Changchun]

Yes, Mark(Jing Chen) has mentioned some reasons as below,
Basically the comments/description for 2 fields(start_rx_per_q and start_tx_per_q) is not clear, so we need add more.
As for renaming, Mark has strong recommending to replace old one with new one.
Both Huawei and I agree with the new name, from my respective, both(old and new one) are readable enough.  

Thanks,
Changchun

-----Original Message-----
From: Chen, Jing D 
Sent: Tuesday, July 22, 2014 5:52 PM
To: Thomas Monjalon; Ouyang, Changchun
Cc: dev at dpdk.org
Subject: RE: [dpdk-dev] [PATCH 0/3] Rename field name for RX/TX queue start/stop

Hi Thomas,

> -----Original Message-----
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Thomas Monjalon
> Sent: Tuesday, July 22, 2014 5:38 PM
> To: Ouyang, Changchun
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH 0/3] Rename field name for RX/TX queue 
> start/stop
> 
> Hi,
> 
> 2014-07-22 15:47, Ouyang Changchun:
> > This patch series include 3 things:
> > 1) Rename the field name from start_rx_per_q to rx_enable_queue in 
> > struct rte_eth_rxconf, and do same thing for TX.
> > This patch also update description for field rx_enable_queue and
> tx_enable_queue.
> > 2) According to 1), update field name from start_rx_per_q to
> rx_enable_queue in struct igb_rx_queue
> >  in ixgbe PMD, do same thing for TX.
> > 3) Update its reference in sample vhost.
> 
> In order to be atomic (and do not break git bisect), you should submit 
> it in one patch.
> Title would be "ethdev: rename queue enabler field" or something like that.
> But the most important in such change is to explain why you make it.
> 
> Thanks
> --
> Thomas

The reason adding this patch is that "start_rx_per_q" and "start_tx_per_q" has requirement in NIC driver in some cases. The implication includes:
1. don't fill mbuf address in RX ring in later dev start function call.
2. don't try to switch this rx/tx queues on in later dev start function call.  Instead, application will call rte_eth_dev_rx/tx_queue_start/stop to control
    this queue.

If the NIC driver tried to support these 2 options, it will have to satisfy above 2 conditions. But the problem is that the 2 fields definition don't have a word to claim on their requirement. So, we needs this patch and add comments. 
As for renaming, it's not so important. Just for better understanding.