)]}'
{
  "commit": "fa26b342fe85a7b5c7bfade1fe1a7f015adfe51d",
  "tree": "2b048eb8e2bc095dc46351a535c3672b95af6ba4",
  "parents": [
    "2734b812763991117d7f79ecd6d07b6ad5faa5c0"
  ],
  "author": {
    "name": "Ayush Kumar Thakur",
    "email": "ayushkumart@nvidia.com",
    "time": "Mon Jun 30 20:05:55 2025 +0530"
  },
  "committer": {
    "name": "Akshat Jain",
    "email": "akshatzen@google.com",
    "time": "Tue Sep 02 16:48:51 2025 -0700"
  },
  "message": "Refactor Static Inventory, MCTP Discovery and Rediscovery Flow\n\n\u0027\u0027\u0027\nwhy:\n1. NsmDevice was created based on static inventory provided by EM config file and it was mapped to the MCTP EID in sensor polling loop.\n   This was not an efficient and cleaner method to remap device and there was dependency on sensor manager to map EID from Static inventory.\n2. The MCTP rediscovery flow was not efficient, leading to many incoming bugs in field. There was multiple in between operations involved in\n   MCTP rediscovery signal and the corresponding nsmDevice going online and offline.\n3. Due to issue 2, it was creating hindrance in making discovery flow atomic.\n\nwhat:\n1. Restructuring of static inventory object in EM config file, it will help in removing instance mapping file from platform configs.\n2. MCTP discovery/Rediscovery flow and Static inventory mapping is restructured and made more modular.\n\u0027\u0027\u0027\n\nFixes Jira https://jirasw.nvidia.com/browse/DGXOPENBMC-17745\nsigned-off-by: \u003cayushkumart@nvidia.com\u003e\n\nRefactor Init, Mctp Rediscovery and Synchronize Post/Patch\n\nFixes Jira https://jirasw.nvidia.com/browse/DGXOPENBMC-17745\nsigned-off-by: \u003cayushkumart@nvidia.com\u003e\n\nFix for state maintenance during rediscovery and state management\n\n\u0027\u0027\u0027\n- Implement sequential processing of rediscovery signals\n- Add queue management for MCTP info in device manager\n- Update NsmDevice state based on queued MCTP info\n- Ensure state consistency with last received signal\n\u0027\u0027\u0027\n\nFixes Jira https://jirasw.nvidia.com/browse/DGXOPENBMC-17745\nsigned-off-by \u003cayushkumart@nvidia.com\u003e\n\nAdd Coroutine Semaphore for locking nsm device update\n\n\u0027\u0027\u0027\nSemaphore is required for synchronizing update nsm device during init if mctp discovery signal is also received.\n\u0027\u0027\u0027\n\nFixes Jira https://jirasw.nvidia.com/browse/DGXOPENBMC-17745\nsigned-off-by: \u003cayushkumart@nvidia.com\u003e\n\nRefactor: Extract capability sensor refresh into separate function\n\n\u0027\u0027\u0027\nCapability sensors must be refreshed after MCTP rediscovery events and driver reloads. Previously, this functionality was embedded within a monolithic function that also handled command code matrix refresh and FRU updates. This commit extracts the capability sensor refresh logic into a dedicated `refreshCapabilitySensor()` function to improve code modularity and maintainability.\n\nThe new function is called from:\n- MCTP rediscovery event handlers\n- Driver reload scenarios\n- Device state transitions\n\u0027\u0027\u0027\n\nFixes Jira https://jirasw.nvidia.com/browse/DGXOPENBMC-17745\nsigned-off-by: \u003cayushkumart@nvidia.com\u003e\n\nDefer Rediscovery Event processing while MCTP rediscovery is in progress\n\nFixes Jira https://jirasw.nvidia.com/browse/DGXOPENBMC-17745\nsigned-off-by: \u003cayushkumart@nvidia.com\u003e\n\nRebase after upstream sync\n\nFixes Jira https://jirasw.nvidia.com/browse/DGXOPENBMC-17745\nsigned-off-by: \u003cayushkumart@nvidia.com\u003e\n\nChange-Id: I2d87efe8a0b49ce5d3ee324aea89e9a50fc53e58\n",
  "tree_diff": [
    {
      "type": "add",
      "old_id": "0000000000000000000000000000000000000000",
      "old_mode": 0,
      "old_path": "/dev/null",
      "new_id": "974828348c62798411bdca841b327f6e442d504f",
      "new_mode": 33188,
      "new_path": "nsmd/nsmDumpCollection/nsmDeviceDiagnostics.cpp"
    }
  ]
}
