[SmartSwitch] Extend reboot script for rebooting SmartSwitch

[vvolam]

What I did

Extended reboot script for SmartSwitch cases to reboot entire SmartSwitch or a specific DPU

How I did it

Implemented changes according to https://github.com/sonic-net/SONiC/blob/605c3a56ac2717dbbb638433e7bb13054fc05a31/doc/smart-switch/reboot/reboot-hld.md

How to verify it

• Verified the script on non-smart switch and didn't find any regressions. Also, script throws errors if any new smart switch related parameters are given by user.
• Verified on NVIDIA smart switch.

Previous command output (if the output of a command-line utility has changed)

New command output (if the output of a command-line utility has changed)

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[gpunathilell]

2025-01-28 21:11:44 - Error: Failed to send reboot status command to DPU dpu1
/usr/local/bin/reboot_smartswitch_helper: line 126: [: null: integer expression expected

Error seen with latest changes

[gpunathilell]

Also there are two issues which need to be addressed, the provisioning for the DPU and the switch GNMI container configuration (as these configurations are required to make sure that the gnoi_client command can be executed from the switch)
GNMI command execution fails with EOF error (due to GNMI container being shut down during pre shutdown) This issue has to be handled as per HLD (make sure GNMI container is not shutdown on DPU)

[KrisNey-MSFT]

Dependent upon PMON via Cisco, and a few Issues filed.

[gpunathilell]

For the pcie related changes which are being done, we need to have differentiation in the way DPU reboot is handled in single DPU reboot and system reboot scenarios (ModuleBase.MODULE_REBOOT_DPU vs ModuleBase.MODULE_REBOOT_SMARTSWITCH) for MODULE_REBOOT_SMARTSWITCH we just schedule the DPU reboot and exit from the reboot call i.e. pcie is not back up after the reboot is executed, so we can not remove the STATE_DB entry in case of ModuleBase.MODULE_REBOOT_SMARTSWITCH (or perform rescan) so the entry deletion has to be skippped in case of system reboot

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[vvolam]

2025-01-28 21:11:44 - Error: Failed to send reboot status command to DPU dpu1
/usr/local/bin/reboot_smartswitch_helper: line 126: [: null: integer expression expected

Error seen with latest changes

This is because of dpu_halt_services_timeout not present in platform.json. I have updated script to have a default timeout in such cse. Thank you!

[vvolam]

not back up after the reboot is executed, so we can not remove the STATE_DB entry in case of ModuleBase.MODULE_R

This is done. Thank you!

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).