From: Yinan <yinan.wang@intel.com>
To: dts@dpdk.org
Cc: Wang Yinan <yinan.wang@intel.com>
Subject: [dts] [PATCH v1] test_plans: add packed ring vectorized case for loopback_virtio_user_server_mode test
Date: Thu, 9 Apr 2020 20:31:10 +0000 [thread overview]
Message-ID: <20200409203110.94053-1-yinan.wang@intel.com> (raw)
From: Wang Yinan <yinan.wang@intel.com>
Signed-off-by: Wang Yinan <yinan.wang@intel.com>
---
...back_virtio_user_server_mode_test_plan.rst | 138 ++++++++++++++----
1 file changed, 107 insertions(+), 31 deletions(-)
diff --git a/test_plans/loopback_virtio_user_server_mode_test_plan.rst b/test_plans/loopback_virtio_user_server_mode_test_plan.rst
index ebcbacb..2716c91 100644
--- a/test_plans/loopback_virtio_user_server_mode_test_plan.rst
+++ b/test_plans/loopback_virtio_user_server_mode_test_plan.rst
@@ -37,7 +37,11 @@ vhost/virtio-user loopback server mode test plan
Virtio-user server mode is a feature to enable virtio-user as the server, vhost as the client, thus after vhost-user is killed then relaunched,
virtio-user can reconnect back to vhost-user again; at another hand, virtio-user also can reconnect back to vhost-user after virtio-user is killed.
This feature test need cover different rx/tx paths with virtio 1.0 and virtio 1.1, includes split ring mergeable, non-mergeable, inorder mergeable,
-inorder non-mergeable, vector_rx path and packed ring mergeable, non-mergeable, inorder non-mergeable, inorder mergeable path.
+inorder non-mergeable, vector_rx path and packed ring mergeable, non-mergeable, inorder non-mergeable, inorder mergeable, vectorized path.
+Note: Packed virtqueue vectorized path need below three initial requirements:
+ 1. AVX512 is allowed in config file and supported by compiler
+ 2. Host cpu support AVX512F
+ 3. ring size is power of two
Test Case 1: Basic test for packed ring server mode
===================================================
@@ -442,7 +446,7 @@ Test Case 7: loopback reconnect test with split ring vector_rx path and server m
testpmd>stop
Test Case 8: loopback reconnect test with packed ring mergeable path and server mode
-====================================================================================
+===================================================================================
1. launch vhost as client mode with 2 queues::
@@ -503,18 +507,18 @@ Test Case 8: loopback reconnect test with packed ring mergeable path and server
10. Port restart at vhost side by below command and re-calculate the average throughput::
- testpmd>stop
- testpmd>port stop 0
- testpmd>port start 0
- testpmd>start tx_first 32
- testpmd>show port stats all
+ testpmd>stop
+ testpmd>port stop 0
+ testpmd>port start 0
+ testpmd>start tx_first 32
+ testpmd>show port stats all
11. Check each RX/TX queue has packets::
- testpmd>stop
+ testpmd>stop
Test Case 9: loopback reconnect test with packed ring non-mergeable path and server mode
-========================================================================================
+=======================================================================================
1. launch vhost as client mode with 2 queues::
@@ -575,18 +579,18 @@ Test Case 9: loopback reconnect test with packed ring non-mergeable path and ser
10. Port restart at vhost side by below command and re-calculate the average throughput::
- testpmd>stop
- testpmd>port stop 0
- testpmd>port start 0
- testpmd>start tx_first 32
- testpmd>show port stats all
+ testpmd>stop
+ testpmd>port stop 0
+ testpmd>port start 0
+ testpmd>start tx_first 32
+ testpmd>show port stats all
11. Check each RX/TX queue has packets::
- testpmd>stop
+ testpmd>stop
Test Case 10: loopback reconnect test with packed ring inorder mergeable path and server mode
-=============================================================================================
+===========================================================================================
1. launch vhost as client mode with 2 queues::
@@ -647,18 +651,18 @@ Test Case 10: loopback reconnect test with packed ring inorder mergeable path an
10. Port restart at vhost side by below command and re-calculate the average throughput::
- testpmd>stop
- testpmd>port stop 0
- testpmd>port start 0
- testpmd>start tx_first 32
- testpmd>show port stats all
+ testpmd>stop
+ testpmd>port stop 0
+ testpmd>port start 0
+ testpmd>start tx_first 32
+ testpmd>show port stats all
11. Check each RX/TX queue has packets::
- testpmd>stop
+ testpmd>stop
Test Case 11: loopback reconnect test with packed ring inorder non-mergeable path and server mode
-=================================================================================================
+===============================================================================================
1. launch vhost as client mode with 2 queues::
@@ -671,7 +675,79 @@ Test Case 11: loopback reconnect test with packed ring inorder non-mergeable pat
2. Launch virtio-user as server mode with 2 queues::
./testpmd -n 4 -l 5-7 --socket-mem 1024,1024 --legacy-mem --no-pci --file-prefix=virtio \
- --vdev=net_virtio_user0,mac=00:01:02:03:04:05,path=./vhost-net,server=1,queues=2,packed_vq=1,mrg_rxbuf=0,in_order=1 \
+ --vdev=net_virtio_user0,mac=00:01:02:03:04:05,path=./vhost-net,server=1,queues=2,packed_vq=1,mrg_rxbuf=0,in_order=1,packed_vec=1 \
+ -- -i --rx-offloads=0x10 --enable-hw-vlan-strip --rss-ip --nb-cores=2 --rxq=2 --txq=2
+ >set fwd mac
+ >start tx_first 32
+
+3. Quit vhost side testpmd, check the virtio-user side link status::
+
+ testpmd> show port info 0
+ #it should show "down"
+
+4. Relaunch vhost and send packets::
+
+ ./testpmd -c 0xe -n 4 --socket-mem 1024,1024 --legacy-mem --no-pci --file-prefix=vhost \
+ --vdev 'eth_vhost0,iface=vhost-net,client=1,queues=2' -- -i --nb-cores=2 --rxq=2 --txq=2
+ >set fwd mac
+ >start tx_first 32
+
+5. Check the virtio-user side link status and run below command to get throughput,verify the loopback throughput is not zero::
+
+ testpmd> show port info 0
+ #it should show up"
+ testpmd>show port stats all
+
+6. Check each RX/TX queue has packets::
+
+ testpmd>stop
+
+7. Quit virtio-user side testpmd, check the vhost side link status::
+
+ testpmd> show port info 0
+ #it should show "down"
+
+8. Relaunch virtio-user and send packets::
+
+ ./testpmd -n 4 -l 5-7 --socket-mem 1024,1024 --legacy-mem --no-pci --file-prefix=virtio \
+ --vdev=net_virtio_user0,mac=00:01:02:03:04:05,path=./vhost-net,server=1,queues=2,packed_vq=1,mrg_rxbuf=0,in_order=1,packed_vec=1 \
+ -- -i --rx-offloads=0x10 --enable-hw-vlan-strip --rss-ip --nb-cores=2 --rxq=2 --txq=2
+ >set fwd mac
+ >start tx_first 32
+
+9. Check the vhost side link status and run below command to get throughput, verify the loopback throughput is not zero::
+
+ testpmd> show port info 0
+ #it should show up"
+ testpmd>show port stats all
+
+10. Port restart at vhost side by below command and re-calculate the average throughput::
+
+ testpmd>stop
+ testpmd>port stop 0
+ testpmd>port start 0
+ testpmd>start tx_first 32
+ testpmd>show port stats all
+
+11. Check each RX/TX queue has packets::
+
+ testpmd>stop
+
+Test Case 12: loopback reconnect test with packed ring vectorized path and server mode
+=======================================================================================
+
+1. launch vhost as client mode with 2 queues::
+
+ rm -rf vhost-net*
+ ./testpmd -c 0xe -n 4 --log-level=pmd.net.vhost.driver,8 --no-pci --file-prefix=vhost \
+ --vdev 'eth_vhost0,iface=vhost-net,client=1,queues=2' -- -i --nb-cores=2 --rxq=2 --txq=2
+ >set fwd mac
+ >start
+
+2. Launch virtio-user as server mode with 2 queues::
+
+ ./testpmd -n 4 -l 5-7 --log-level=pmd.net.virtio.driver,8 --no-pci --file-prefix=virtio \
+ --vdev=net_virtio_user0,mac=00:01:02:03:04:05,path=./vhost-net,server=1,queues=2,packed_vq=1,mrg_rxbuf=0,in_order=1,packed_vec=1 \
-- -i --tx-offloads=0x0 --enable-hw-vlan-strip --rss-ip --nb-cores=2 --rxq=2 --txq=2
>set fwd mac
>start tx_first 32
@@ -706,7 +782,7 @@ Test Case 11: loopback reconnect test with packed ring inorder non-mergeable pat
8. Relaunch virtio-user and send packets::
./testpmd -n 4 -l 5-7 --socket-mem 1024,1024 --legacy-mem --no-pci --file-prefix=virtio \
- --vdev=net_virtio_user0,mac=00:01:02:03:04:05,path=./vhost-net,server=1,queues=2,packed_vq=1,mrg_rxbuf=0,in_order=1 \
+ --vdev=net_virtio_user0,mac=00:01:02:03:04:05,path=./vhost-net,server=1,queues=2,packed_vq=1,mrg_rxbuf=0,in_order=1,packed_vec=1 \
-- -i --tx-offloads=0x0 --enable-hw-vlan-strip --rss-ip --nb-cores=2 --rxq=2 --txq=2
>set fwd mac
>start tx_first 32
@@ -719,12 +795,12 @@ Test Case 11: loopback reconnect test with packed ring inorder non-mergeable pat
10. Port restart at vhost side by below command and re-calculate the average throughput::
- testpmd>stop
- testpmd>port stop 0
- testpmd>port start 0
- testpmd>start tx_first 32
- testpmd>show port stats all
+ testpmd>stop
+ testpmd>port stop 0
+ testpmd>port start 0
+ testpmd>start tx_first 32
+ testpmd>show port stats all
11. Check each RX/TX queue has packets::
- testpmd>stop
\ No newline at end of file
+ testpmd>stop
\ No newline at end of file
--
2.17.1
next reply other threads:[~2020-04-09 12:48 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-04-09 20:31 Yinan [this message]
2020-04-20 2:17 ` Tu, Lijuan
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200409203110.94053-1-yinan.wang@intel.com \
--to=yinan.wang@intel.com \
--cc=dts@dpdk.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).