SAI 9.x requires a SYNCD_SHM_SIZE specified otherwise it will default to 64mb which is insufficient for syncd.
E.G. of a few failures seen when insufficient shmem was set
ha_init: The file: warmboot_data_0 is of size=762[MB] and is beyond the directory: /dev/shm available storage of size=64[MB]#015
syncd.sh[26074]: Cannot get SYNCD_SHM_SIZE for chip: [869] in /usr/share/sonic/device/x86_64-broadcom_common/syncd_shm.ini. Skip set SYNCD_SHM_SIZE.
Syncd hangs here:
syncd#syncd: [none] SAI_API_SWITCH:_brcm_sai_shr_ha_section_resize:536 start=0x7f6e641b4000, end=0x7f6e645b4000, len=302276608, free=0x7f6e641b4000
Broadcom recommended using 1gb for DNX devices.
Since currently we don't use SAI9.x on master and 202305 this change won't fix anything until we upgrade the SAI on those branches.
These devices will not reliabily report the proper devid and vendorid
when reading it is read directly from the pci config space.
It can be read but shouldn't be compared against some fixed value like
the one stored in pcie.yaml.
Since this makes pcied unhappy, the simplest path forward is to just
remove this device from monitoring.
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
Add platform_asic file to each platform folder in sonic-device-data package. The file content will be used as the ground truth of mapping from PLATFORM_STRING to switch ASIC family.
One use case of the mapping is to prevent installing a wrong image, which targets for other ASIC platforms. For example, currently we have several ONIE images naming as sonic-*.bin, it's easy to mistakenly install the wrong image. With this mapping built into image, we could fetch the ONIE platform string, and figure out which ASIC it is using, and check we are installing the correct image.
After this PR merged, each platform vendor has to add one mandatory text file `device/PLATFORM_VENDOR/PLATFORM_STRING/platform_asic`, with the content of the platform's switch ASIC family.
I will update https://github.com/Azure/SONiC/wiki/Porting-Guide after this PR is merged.
You can get a list of the ASIC platforms by `ls -b platform | cat`. Currently the options are
```
barefoot
broadcom
cavium
centec
centec-arm64
generic
innovium
marvell
marvell-arm64
marvell-armhf
mellanox
nephos
p4
vs
```
Also support
```
broadcom-dnx
```
#### How I did it
#### How to verify it
Test one image on DUT. And check the folders under `/usr/share/sonic/device`
Introduce new sonic-buildimage images for Broadcom DNX ASIC family.
sonic-broadcom-dnx.bin
sonic-aboot-broadcom-dnx.swi
How I did it
NO CHANGE to existing make commands
make init; make configure PLATFORM=broadcom; make target/sonic-aboot-broadcom.swi; make target/sonic-broadcom.bin
The difference now is that it will result in new broadcom images for DNX asic family as well.
sonic-broadcom.bin, sonic-broadcom-dnx.bin
sonic-aboot-broadcom.swi, sonic-aboot-broadcom-dnx.swi
Note: This PR also adds support for Broadcom SAI 5.0 (based on 1.8 SAI ) for DNX based platform + changes in platform x86_64-arista_7280cr3_32p4 bcm config files and platform_env.conf files
Platform library changes
- Fix the use of /proc/modules during testing, fixes#7463
- Add `libsfp-eeprom.so` build to read/write xcvr eeproms in C
- Add some more reboot-cause information
- Write down temperature hw thresholds to the sensors
- Report software thresholds through platform api
- Writ `port_name sysfs` file of optoe`
- Tests enhancements
- Fix dependency issues for chassis provisioning
Platform configuration changes
- Add `pcie.yaml` configuration for a few platforms
- Mount `libsfp-eeprom.so` inside `pmon`
- Fix `Arista-7050SX3-48C8` and `Arista-7050SX3-48YC8' platform and hwsku
- Miscellaneous fixes
Co-authored-by: Boyang Yu <byu@arista.com>
Co-authored-by: Zhi Yuan Carl Zhao <zyzhao@arista.com>
- Add support for `DCS-7050SX3-48YC8` and `DCS-7050SX3-48C8` platform
- Add support for more variants of `DCS-7280CR3-32[PD]4`
- Add Supervisor to Linecard consutil support
- Complete Watchdog platform API support
- Fix some PSU behavior on `DCS-7050QX-32` and `DCS-7060CX-32S`
- Fix SEU management on `DCS-7060CX-32S`
- Allow kernel modules to build up to linux 5.10
- Rename led color `orange` to `amber`
- Miscellaneous fixes
Prevent system-healthd from service from failing at boot time due to missing configuration.
Also adds basic support for healthd.
The following caveat exists with this placeholder configuration:
- No PSU monitoring (sensors/fans)
- No ASIC temperature monitoring
platform.json is needed for sonic-mgmt testing. Also in the future it will be used as part of dynamic port breakout.
Also removed the folder symlink for BlackhawkDD because it has a different platform.json than BlackhawkO.
Co-authored-by: Zhi Yuan Carl Zhao <zyzhao@arista.com>
- Enable thermalctld support for our platforms
- Fix Chassis.get_num_sfp which had an off by one
- Implement read_eeprom and write_eeprom in SfpBase
- Refactor of Psus and PsuSlots. Psus they are now detected and metadata reported
- Improvements to modular support
Co-authored-by: Zhi Yuan Carl Zhao <zyzhao@arista.com>
However in SAI 3.7 default behaviout got changes to 128 Group and 128
Memeber each.
This change is to make sure we are using same ECMP Group/Memeber Per
Group for 3.7 also so that behaviour is consistent.
Signed-off-by: Abhishek Dosi <abdosi@microsoft.com>
thermalctld throwing error messages because it is not yet fully configured, disabling it for now on arista platforms.
Co-authored-by: Zhi Yuan Carl Zhao <zyzhao@arista.com>
- Sensor and Fan information added to primary platforms for thermal API.
- Refactors involving better abstractions, code reuse and dead code removal.
- Improvements to the diag capabilities
- Pylintrc added to improve code quality. Will become fatal at a later time.
Co-authored-by: Baptiste Covolato <baptiste@arista.com>
* Add boot0 support for the 7280CR3
* Add platform and plugins for 7280CR3
* Add port config for 7280CR3
* Add platform_reboot for 7280CR3
* Add support for 7280CR3-32D4 based on the 7280CR3-32P4
* Update arista driver submodules
- Introduce new 7280CR3-32P4
- Improve to the led plugin for OSFP
