[dts] [PATCH] test_plan/dynamic_queue: add dynamic queue test plan

[Lin, Xueqin]

Yes, queue stop is must, if not stop queue, will fail to setup this queue next. 
Use cmd to start testpmd, application first specifies how many Tx and Rx queues to include in the ethdev, such as "--txq=4 --rxq=4",
the application then sets up each 4 Tx and 4 Rx queues. Finally, once all the queues have been set up, the application may then start the device,
then traffic can flow. If device stops, this halts the flow of traffic on all queues in  the ethdev, not stop one queue.
So device stop is different from queue stop. 

Best regards,
Xueqin

> -----Original Message-----
> From: Liu, Yong
> Sent: Monday, August 20, 2018 5:24 PM
> To: Lin, Xueqin <xueqin.lin at intel.com>; dts at dpdk.org
> Subject: Re: [dts] [PATCH] test_plan/dynamic_queue: add dynamic queue
> test plan
> 
> 
> 
> On 08/07/2018 12:23 PM, Xueqin Lin wrote:
> > +Test Case: Rx queue configure at runtime
> > +========================================
> > +Stop some Rx queues on port 0::
> > +
> > +    testpmd> port 0 rxq <id> stop
> > +
> > +Set rxonly forward, start testpmd
> > +
> > +Send different src or dst IPv4 packets::
> > +
> > +    p=Ether()/IP(src="192.168.0.1", dst="192.168.0.1")/Raw("x"*20)
> > +
> > +Stop testpmd, find stopped queues can't receive packets, but other
> > +queues could receive packets
> Hi Xueqin,
> In the introduction chapter of this test plan, you have mentioned that no
> need to call dev_stop when configuring queue.
> Is here queue stop is must? If it is yes, please add some note for difference
> between device stop and queue stop.
> 
> Thanks,
> Marvin