- Why I did it
Support DSCP remapping in dual ToR topo on T0 switch for SKU Mellanox-SN4600c-C64, Mellanox-SN4600c-D48C40, Mellanox-SN2700, Mellanox-SN2700-D48C8.
- How I did it
Regarding buffer settings, originally, there are two lossless PGs and queues 3, 4. In dual ToR scenario, the lossless traffic from the leaf switch to the uplink of the ToR switch can be bounced back.
To avoid PFC deadlock, we need to map the bounce-back lossless traffic to different PGs and queues. Therefore, 2 additional lossless PGs and queues are allocated on uplink ports on ToR switches.
On uplink ports, map DSCP 2/6 to TC 2/6 respectively
On downlink ports, both DSCP 2/6 are still mapped to TC 1
Buffer adjusted according to the ports information:
Mellanox-SN4600c-C64:
56 downlinks 50G + 8 uplinks 100G
Mellanox-SN4600c-D48C40, Mellanox-SN2700, Mellanox-SN2700-D48C8:
24 downlinks 50G + 8 uplinks 100G
- How to verify it
Unit test.
Signed-off-by: Stephen Sun <stephens@nvidia.com>
Why I did it
Default pcieutil uses one configuration for all models of platform
How I did it
Take the configuration file as base for all models of concrete platform where model-specific devices are not listed in this configuration
How to verify it
Run pmon#pcied and verify that there is no error/warning logs on initialization step
Why I did it
fix DX010 fan drawer and watchdog platform test case issues
How I did it
1. Add fan_drawer get_maximum_consumed_power support
2. Adjust maximum watchdog timeout value check
How to verify it
Run test_fan_drawer and test_watchdog test cases.
Why I did it
By specifying 'status_led' 'controllable' to false for psu section, it means the platform is not yet supporting psu status led
How I did it
specify 'status_led' 'controllable' to false for psu section
How to verify it
by running test in pdb, manually add {'status_led' : {'controllable' : False}} in dictionary
this flag will be able to get False and skip testing:
ce290c735d/tests/platform_tests/api/test_psu.py (L337)
After upgrade to brcmsai 8.1, the sdk running environment (container) recommended with mininum memory size as below
TH4/TD4(ltsw) uses 512MB
TH3 used 300MB
Helix4/TD2/TD3/TH/TH 256 MB
Base on this requirement, adjust the default syncd share memory size and set the memory size for special ACISs in platform_env.conf file for different types of Broadcom ASICs.
How I did it
Add the platform_env.conf file if none of it for broadcom platform (base on platform_asic file)
Add the 'SYNCD_SHM_SIZE' and set the value
for ltsw(TD4/TH4) devices set to 512M at least (update the platform_env.conf)
for Td2/TH2/TH devices set to 256M
for TH3 set to 300M
verify
How to verify it
verify the image with code fix
Check with UT
Check on lab devices
On a problematic device which cannot start successfully
Run with the command
$ cat /proc/linux-kernel-bde
Broadcom Device Enumerator (linux-kernel-bde)
Module parameters:
maxpayload=128
usemsi=0
dmasize=32M
himem=(null)
himemaddr=(null)
DMA Memory (kernel): 33554432 bytes, 0 used, 33554432 free, local mmap
No devices found
$ docker rm -f syncd
syncd
$ sudo /usr/bin/syncd.sh start
Cannot get Broadcom Chip Id. Skip set SYNCD_SHM_SIZE.
Creating new syncd container with HWSKU Force10-S6000
a4862129a7fea04f00ed71a88715eac65a41cdae51c3158f9cdd7de3ccc3dd31
$ docker inspect syncd | grep -i shm
"ShmSize": 67108864,
"Tag": "fix_8.1_shm_issue.67873427-9f7ca60a0e",
On Normal device
$ docker inspect syncd | grep -i shm
"ShmSize": 268435456,
"Tag": "fix_8.1_shm_issue.67873427-9f7ca60a0e"
change the config syncd_shm.ini to b85=128m
$ docker rm -f syncd
syncd
$ sudo /usr/bin/syncd.sh start
Creating new syncd container with HWSKU Force10-S6000
3209ffc1e5a7224b99640eb9a286c4c7aa66a2e6a322be32fb7fe2113bb9524c
$ docker inspect syncd | grep -i shm
"ShmSize": 134217728,
"Tag": "fix_8.1_shm_issue.67873427-9f7ca60a0e",
change the config under
/usr/share/sonic/device/x86_64-dell_s6000_s1220-r0/Force10-S6000/platform_env.conf
and run command
$ cat /usr/share/sonic/device/x86_64-dell_s6000_s1220-r0/platform_env.conf
SYNCD_SHM_SIZE=300m
$ sudo /usr/bin/syncd.sh start
Creating new syncd container with HWSKU Force10-S6000
897f6fcde1f669ad2caab7da4326079abd7e811bf73f018c6dacc24cf24bfda5
$ docker inspect syncd | grep -i shm
"ShmSize": 314572800,
"Tag": "fix_8.1_shm_issue.67873427-9f7ca60a0e",
Signed-off-by: richardyu-ms <richard.yu@microsoft.com>
Eariler the SDK stat polling was erroneously set to once every msec
which is far more frequent than required by SWSS. The new setting, which
is consistent with other vendor SKUs, is once a second. The net result
is reduced CPU MHz by syncd.
Why I did it
Fix the following issues for Seastone platform:
- system-health issue: show system-health detail will not complete #9530, Celestica Seastone DX010-C32: show system-health detail fails with 'Chassis' object has no attribute 'initizalize_system_led' #11322
- show platform firmware updates issue: Celestica Seastone DX010-C32: show platform firmware updates #11317
- other platform optimization
How I did it
Modify and optimize the platform implememtation.
How to verify it
Manual run the test commands described in these issues.
Why I did it
Ragile adapter ra-b6510-32c ra-b6510-48v8c ra-b6910-64c ra-b6920-4s to kernel 5.x
Signed-off-by: “pettershao” pettershao@ragilenetworks.com
Fix code issue when SonicV2Connector.get() return None.
#### Why I did it
When database key does not exist, SonicV2Connector.get() will return None.
Code will break if not check return value.
#### How I did it
Check SonicV2Connector.get() return value before use it.
#### How to verify it
Pass all E2E test case.
Why I did it
Add two platform that support s3IP framework
How I did it
Add two platforms supporting S3IP SYSFS (TCS8400, TCS9400)
How to verify it
Manual test
Adding platform support for FS s5800-48t4s and s5800-48t8s-mars8p.
Both s5800-48t4s and s5800-48t8s-mars8p have 48 * 10/100/1000 Base-T ports, 4 * 10GE SFP+ Ports on Centec TsingMa.
s5800-48t4s is different from s5800-48t8s-mars8p in that:
The phy chip used by s5800-48t4s is Marvell 88e1680;
The phy chip used by s5800-48t4s-mars8p is Centec ctc21108;
Why I did it
This is to fix test_forward_ip_packet_with_0xffff_chksum_tolerant test failure on 720DT-48S. IP packets with checksum set to 0xffff will be forwarded with the same checksum on this platform, instead of updating to the correct value.
How I did it
Add bcm config sai_verify_incoming_chksum=0 so that checksum is updated instead of being left unchanged when checksum is 0xffff. Note that packets with invalid checksum are still dropped with this config.
Why I did it
On DNX (J2/J2c/J2c+) platforms, Single Path Nexthops and ECMp Nexthop resources(FECs) are shared. BRCM SAI do not have partition of this resource, and hence more single path Nexthop entries, causes ECMP programming to fail in scaled setup.
How I did it
Broadcom provided SAI changes to reserve resources for single path nexthop entries(More details in CSP: https://brcmsemiconductor-csm.wolkenservicedesk.com/wolken-support/allcases/request-details?requestId=CS00012251649).
Along with SAI changes, they provided configurable Macro/flag to reserve NON_ECMP entries.
This PR is to add that flag in various sai.profile files wherever applicable.
PS: We are reserving 3072 single path Nexthop entries on each Linecard. Calculation is as follows.
Max Slots per chassis: 8
Max No of Ports(each LC): 64
MyIP/Subnet Entries per port: 4(v4/v6)
Nbr Entries Per port: 2(v4/v6)
Total Non_ECMP Count: 8x64x(4+2) = 3072
How to verify it
Without this change, the ECMP group count will be shown as Max_count in 'crm show resources all' command, and with this change the ECMP group count will be decreased by 24(3072/128).
Port indexes of front panel ports are not contiguous in multi-asic because we didn't distiguish between
front panel and internal ports, e.g., recycle ports. Fix this by assigning index to front panel port first
and then internal ports.
Add missing aggregate port_config.ini needed by sonic-mgmt
Concatenate the ASIC specific port_config.ini from device/arista/x86_64-arista_7800r3a_36d2_lc/Arista-7800R3A-36D2-C36/[01] to create the aggregate file.
Signed-off-by: maipbui <maibui@microsoft.com>
Dependency: [PR (#12065)](https://github.com/sonic-net/sonic-buildimage/pull/12065) needs to merge first.
#### Why I did it
1. `eval()` - not secure against maliciously constructed input, can be dangerous if used to evaluate dynamic content. This may be a code injection vulnerability.
2. `subprocess()` - when using with `shell=True` is dangerous. Using subprocess function without a static string can lead to command injection.
3. `os` - not secure against maliciously constructed input and dangerous if used to evaluate dynamic content.
4. `is` operator - string comparison should not be used with reference equality.
5. `globals()` - extremely dangerous because it may allow an attacker to execute arbitrary code on the system
#### How I did it
1. `eval()` - use `literal_eval()`
2. `subprocess()` - use `shell=False` instead. use an array string. Ref: [https://semgrep.dev/docs/cheat-sheets/python-command-injection/#mitigation](https://semgrep.dev/docs/cheat-sheets/python-command-injection/#mitigation)
3. `os` - use with `subprocess`
4. `is` - replace by `==` operator for value equality
5. `globals()` - avoid the use of globals()
This change is to disable the pcie firmware check done by Broadcom SAI. The change is needed for the Arista platform x86_64-arista_7050cx3_32s; otherwise, the check will fail, blocking the initialization.
There was a pcie firmware check added in brcm SDK and certain Arista hardwares do not compliant with the check, so we added the disable_pcie_firmware_check originally for x86_64-arista_7060dx4_32. For x86_64-arista_7050cx3_32s, it was able to pass the check but some firmware change done in August made it fail.
Why I did it:
Fix multiple seastone platform issues caused by sonic kernel upgrade.
How I did it:
Get gpio base id with new label path in gpio sys fs.
How to verify it:
After the change, show platform fan/psustatus/temperature works well.
Signed-off-by: maipbui <maibui@microsoft.com>
Dependency: [https://github.com/sonic-net/sonic-buildimage/pull/12065](https://github.com/sonic-net/sonic-buildimage/pull/12065)
#### Why I did it
`subprocess.Popen()` and `subprocess.run()` is used with `shell=True`, which is very dangerous for shell injection.
`os` - not secure against maliciously constructed input and dangerous if used to evaluate dynamic content
`getstatusoutput` is dangerous because it contains `shell=True` in the implementation
#### How I did it
Replace `os` by `subprocess`, use with `shell=False`
Remove unused functions
#### Why I did it
`os` and `commands` modules are not secure against maliciously constructed input
`getstatusoutput` is detected without a static string, uses `shell=True`
#### How I did it
Eliminate the use of `os` and `commands`
Use `subprocess` instead
Signed-off-by: Neetha John <nejo@microsoft.com>
Why I did it
ECN parameters need to be updated for storage backend
How I did it
Included the check for storage backend devices to update qos configs
How to verify it
Verified that the new ecn settings are applied on storage backend device.
Verified that the old ecn settings are applied for storage frontend, non storage frontend/backend devices
Signed-off-by: maipbui <maibui@microsoft.com>
#### Why I did it
The [xml.etree.ElementTree](https://docs.python.org/3/library/xml.etree.elementtree.html#module-xml.etree.ElementTree) module is not secure against maliciously constructed data.
`os` - not secure against maliciously constructed input and dangerous if used to evaluate dynamic content
`subprocess.getstatusoutput` is dangerous because include shell=True in the implementation
#### How I did it
Remove xml. Use [lxml](https://pypi.org/project/lxml/) XML parsers package that prevent potentially malicious operation.
Replace `os` by `subprocess`
Use command as an array instead of string
Use `getstatusoutput_noshell` in `sonic_py_common` lib
Why I did it
Added to allow test_crm_route to pass; the test tries to add a /126 ipv6 route and this change is required in order for the count of available routes to be updated correctly.
Why I did it
TX FIR tuning should be done based on the type of inserted transceiver
How I did it
Add media_settings.json which contains the tuning data for 100G optic and 400G optic.
How to verify it
Tested against x86_64-arista_7800r3a_36d2_lc
Why I did it
The PR is to apply separated DSCP_TO_TC_MAP and TC_TO_QUEUE_MAP to uplink ports on dualtor.
The traffic with DSCP 2 and DSCP 6 from T1 is treated as lossless traffic.
DSCP TC Queue
2 2 2
6 6 6
Traffic with DSCP 2 or DSCP 6 from downlink is still treated as lossy traffic as before.
How I did it
Define DSCP_TO_TC_MAP|AZURE_UPLINK and TC_TO_QUEUE_MAP|AZURE_UPLINK.
How to verify it
Verified by UT
Verified by coping the new template to a testbed, and rendering a config_db.json
Why I did it
Some sonic-mgmt platform_tests/api were failing on the 7060CX-32S
How I did it
Added the missing metadata in platform.json and platform_components.json
This is purely test data and does not impact our API implementation.
How to verify it
Run platform_tests / api and expect 100% pass rate.
Why I did it
Some sonic-mgmt platform_tests/api were failing on the 7260CX3-64
How I did it
Added the missing metadata in platform.json and platform_components.json
This is purely test data and does not impact our API implementation.
How to verify it
Run platform_tests/api and expect 100% passrate.
Signed-off-by: Neetha John <nejo@microsoft.com>
Why I did it
There is a need to have separate profiles on compute and storage and this infra update will help achieve that
How I did it
Moved buffer pool/profile and qos definitions on TD2 to a common folder and all TD2 hwsku's will reference that folder
