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

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

[Xueqin Lin]

From: "xueqin.lin" <xueqin.lin@intel.com>

Signed-off-by: Xueqin Lin <xueqin.lin@intel.com> 

---
 test_plans/dynamic_queue_test_plan.rst | 203 +++++++++++++++++++++++++++++++++
 1 file changed, 203 insertions(+)
 create mode 100644 test_plans/dynamic_queue_test_plan.rst

diff --git a/test_plans/dynamic_queue_test_plan.rst b/test_plans/dynamic_queue_test_plan.rst
new file mode 100644
index 0000000..780063d
--- /dev/null
+++ b/test_plans/dynamic_queue_test_plan.rst
@@ -0,0 +1,203 @@
+.. Copyright (c) <2018>, Intel Corporation
+   All rights reserved.
+
+   Redistribution and use in source and binary forms, with or without
+   modification, are permitted provided that the following conditions
+   are met:
+
+   - Redistributions of source code must retain the above copyright
+     notice, this list of conditions and the following disclaimer.
+
+   - Redistributions in binary form must reproduce the above copyright
+     notice, this list of conditions and the following disclaimer in
+     the documentation and/or other materials provided with the
+     distribution.
+
+   - Neither the name of Intel Corporation nor the names of its
+     contributors may be used to endorse or promote products derived
+     from this software without specific prior written permission.
+
+   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
+   FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
+   COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
+   INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+   (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+   SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+   HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+   STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+   ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+   OF THE POSSIBILITY OF SUCH DAMAGE.
+
+=============
+Dynamic queue 
+=============
+According to existing implementation, rte_eth_[rx|tx]_queue_setup will
+always return fail if device is already started(rte_eth_dev_start).
+This can't satisfy the usage when application wants to defer to setup
+part of the queues while keeping traffic running on those queues already
+be setup.
+Basically this is not a general hardware limitation, because for NIC
+like i40e, ixgbe, it is not necessary to stop the whole device before
+configure a fresh queue or reconfigure an existing queue with no traffic
+on it.
+Dynamic queue lets etherdev driver exposes the capability flag through
+rte_eth_dev_info_get when it supports deferred queue configuraiton,
+then base on this flag, rte_eth_[rx|tx]_queue_setup could decide to
+continue to setup the queue or just return fail when device already
+started.
+Allow ethdevs to setup/reconfigure/tear down queues at runtime without 
+stopping the device. Given an ethdev configuration with a specified 
+number of Tx and Rx queues, requirements as below:
+1.The application should be able to start the device with only some of the 
+queues set up.
+2.The application should be able to set up additional queues at runtime 
+without calling dev_stop().
+3.The application should be able to reconfigure existing queues at runtime 
+without calling dev_stop().
+4.This support should be implemented in such a way that it does not break 
+existing PMDs. 
+
+Prerequisites
+=============
+1. Host PF in DPDK driver::
+
+    ./tools/dpdk-devbind.py -b igb_uio 81:00.0
+
+2. Start testpmd on host, set chained port topology mode, add txq/rxq to 
+   enable multi-queues::
+   
+    ./testpmd -c 0xf -n 4  -- -i --port-topology=chained --txq=64 --rxq=64
+
+
+Test Case: Rx queue setup 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
+	
+Setup these stopped queues on the port::
+
+    testpmd> port 0 rxq <id> setup
+
+Start these stopped queues on the port, start testpmd::
+
+    testpmd> port 0 rxq <id> start
+
+Send different src or dst IPv4 packets
+
+Stop testpmd, check all the setup queues could receive packets
+
+
+Test Case: Tx queue setup at runtime
+====================================
+Check txq ring size is 256::
+                
+    testpmd> show txq info 0 <id>
+    Number of TXDs: 256
+
+Stop one Tx queue on port 0::
+
+    testpmd> port 0 txq <id> stop
+
+Set txonly forward, start testpmd
+
+Start testpmd, then stop, check this stopped queue only transmits 255 packets
+  
+Setup this stopped queue on the port::
+
+    testpmd> port 0 txq <id> setup
+
+Start this stopped queue on the port::
+
+    testpmd> port 0 txq <id> start
+
+Start then stop testpmd, check all queues could transmit lots of packets, 
+not only 255 packets
+
+Repeat above steps for 2 times
+
+
+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
+
+Check rxq ring size is 256::
+     
+    testpmd> show rxq info 0 <id>
+    Number of RXDs: 256
+
+Reconfigure ring size as 512 for the stopped queues on port 0::
+
+    testpmd> port config 0 rxq <id> ring_size 512
+
+Setup these stopped queues on the port::
+
+    testpmd> port 0 rxq <id> setup
+
+Check stopped rxq ring sizes have been changed to 512
+
+Start these stopped queues on the port, start testpmd::
+
+    testpmd> port 0 rxq <id> start
+
+Send different src or dst IPv4 packets
+
+Stop testpmd, check all the setup queues could receive packets
+
+
+Test Case: Tx queue configure at runtime
+========================================
+Check txq ring size is 256::
+
+    testpmd> show txq info 0 <id>
+    Number of TXDs: 256
+
+Stop one Tx queue on port 0::
+
+    testpmd> port 0 txq <id> stop
+
+Set txonly forward, start testpmd
+
+Start testpmd, then stop, check this stopped queue only transmits 255 packets
+
+Reconfigure ring size as 512 for the stopped queue on port 0::
+
+    testpmd> port config 0 txq <id> ring_size 512
+
+Setup this stopped queue on the port::
+
+    testpmd> port 0 txq <id> setup
+
+Check stopped txq ring size has been changed to 512
+
+Start this stopped queue on the port, start testpmd::
+
+    testpmd> port 0 txq <id> start
+
+Stop testpmd, check all queues could transmit lots of packets,
+not only 511 packets
+
+Repeat above steps for 2 times
+
-- 
2.7.5

Re: [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@intel.com>; dts@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

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

[Liu, Yong]

Thanks for clear it, xueqin. Please also add this information in test plan.

> -----Original Message-----
> From: Lin, Xueqin
> Sent: Monday, August 20, 2018 1:07 PM
> To: Liu, Yong <yong.liu@intel.com>; dts@dpdk.org
> Subject: RE: [dts] [PATCH] test_plan/dynamic_queue: add dynamic queue test
> plan
> 
> 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@intel.com>; dts@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

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

[Lin, Xueqin]

NP, v2 has added this information and sent it to dts. Thanks Marvin. 

Best regards,
Xueqin

-----Original Message-----
From: Liu, Yong 
Sent: Monday, August 20, 2018 2:39 PM
To: Lin, Xueqin <xueqin.lin@intel.com>; dts@dpdk.org
Subject: RE: [dts] [PATCH] test_plan/dynamic_queue: add dynamic queue test plan

Thanks for clear it, xueqin. Please also add this information in test plan.

> -----Original Message-----
> From: Lin, Xueqin
> Sent: Monday, August 20, 2018 1:07 PM
> To: Liu, Yong <yong.liu@intel.com>; dts@dpdk.org
> Subject: RE: [dts] [PATCH] test_plan/dynamic_queue: add dynamic queue 
> test plan
> 
> 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@intel.com>; dts@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

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

[Liu, Yong]

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