SAI testing and integration framework that is based on SONiC sairedis project.
git clone https://github.com/PLVision/sai-challenger.git
cd sai-challenger/
git submodule update --init --recursiveThe vslib SAI implementation is used as a virtual data-plane interface in SONiC Virtual Switch (SONiC VS). Though it does not configure the forwarding path but still process SAI CRUD calls in proper manner. This allows to use vslib for SAI testcases development without running traffic.
Build SAI Challenger Docker image with vslib SAI implementation:
docker build -f Dockerfile.saivs -t saivs-challenger .Run SAI Challenger testcases:
docker run --name sai-challenger-run \
-v $(pwd):/sai-challenger \
--cap-add=NET_ADMIN \
--device /dev/net/tun:/dev/net/tun \
-d saivs-challenger
docker exec -ti sai-challenger-run pytest -vIn client-server mode, SAI server - syncd linked with vslib - runs in one Docker container. Whereas the client - SAI Challenger - runs in the separate Docker container. These two Docker containers can also be running on the separate physical hosts.
Build Docker image with vslib SAI implementation:
docker build -f Dockerfile.saivs.server -t saivs-server .Build SAI Challenger Docker image with SAI tests:
docker build -f Dockerfile.client -t sai-challenger-client .Run SAI server:
docker run --name saivs -v $(pwd):/sai-challenger -d saivs-serverRun SAI Challenger testcases:
docker run --name sai-challenger \
-v $(pwd):/sai-challenger \
--cap-add=NET_ADMIN \
--device /dev/net/tun:/dev/net/tun \
-d sai-challenger-client
docker exec -ti sai-challenger pytest --sai-server=<saivs-server-ip> -v test_l2_basic.pyCopy Debian package with SAI library into sai-challenger/ folder.
Build SAI Challenger Docker image with vendor-specific SAI implementation:
docker build -f Dockerfile.sai -t sai-challenger .Run SAI Challenger testcases:
docker run --name sai-challenger-run \
-v $(pwd):/sai-challenger \
--cap-add=NET_ADMIN \
--device /dev/net/tun:/dev/net/tun \
-d sai-challenger
docker exec -ti sai-challenger-run pytest -vWhen Orchagent creates new SAI object, it actually performs two operations on Redis DB:
- Enqueue SAI object create operation through LPUSH (in fact, Redis linked list object is used as an queue to pass SAI object data between Orchagent and Syncd);
- Inform Syncd about new data in the linked list through PUBLISH operation;
/*
* KEYS[1] : tableName + "_KEY_VALUE_OP_QUEUE
* ARGV[1] : key
* ARGV[2] : value
* ARGV[3] : op
* KEYS[2] : tableName + "_CHANNEL"
* ARGV[4] : "G"
*/
string luaEnque =
"redis.call('LPUSH', KEYS[1], ARGV[1], ARGV[2], ARGV[3]);"
"redis.call('PUBLISH', KEYS[2], ARGV[4]);";Where 'op' string consists of two parts: DB operation ("S" - set, "D" - delete) and SAI operation ("create", "notify", "set", etc);
Both orchagent and syncd use VIDCOUNTER to generate new VID values.
/*
* Current VID format:
*
* bits 63..56 - switch index
* bits 55..48 - SAI object type
* bits 47..40 - global context
* bits 40..0 - object index
*/For more information, please refer to VirtualObjectIdManager::allocateNewObjectId() located in: sonic-sairedis/lib/src/VirtualObjectIdManager.cpp
LPUSH ASIC_STATE_KEY_VALUE_OP_QUEUE "SAI_OBJECT_TYPE_SWITCH:oid:0x21000000000000" '["SAI_SWITCH_ATTR_INIT_SWITCH","true","SAI_SWITCH_ATTR_SRC_MAC_ADDRESS","52:54:00:EE:BB:70"]' Screate
PUBLISH ASIC_STATE_CHANNEL Gredis-cli -n 1 LRANGE GETRESPONSE_KEY_VALUE_OP_QUEUE 0 -1
redis-cli -n 1 DEL GETRESPONSE_KEY_VALUE_OP_QUEUE1) "Sgetresponse"
2) "[]"
3) "SAI_STATUS_SUCCESS"LPUSH ASIC_STATE_KEY_VALUE_OP_QUEUE "SAI_OBJECT_TYPE_SWITCH:oid:0x21000000000000" '["SAI_SWITCH_ATTR_DEFAULT_VIRTUAL_ROUTER_ID","oid:0x0"]' Sget
PUBLISH ASIC_STATE_CHANNEL Gredis-cli -n 1 LRANGE GETRESPONSE_KEY_VALUE_OP_QUEUE 0 -1
redis-cli -n 1 DEL GETRESPONSE_KEY_VALUE_OP_QUEUE1) "Sgetresponse"
2) "[\"SAI_SWITCH_ATTR_DEFAULT_VIRTUAL_ROUTER_ID\",\"oid:0x3000000000022\"]"
3) "SAI_STATUS_SUCCESS"
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent