)]}'
{
  "commit": "c68c5bb391c6c8063e38e99e3ecd57f5ed08cbab",
  "tree": "bf35420ca395ca110bfd17e26ea3f33a0449aec1",
  "parents": [
    "efa8447397c468a279db0b4b77befe62d9e2503b"
  ],
  "author": {
    "name": "Oliver Peralta",
    "email": "ocperalta@fuchsia.infra.roller.google.com",
    "time": "Wed Sep 30 10:08:10 2026 -0700"
  },
  "committer": {
    "name": "Copybara Service",
    "email": "copybara-worker@google.com",
    "time": "Wed Sep 30 10:08:10 2026 -0700"
  },
  "message": "[roll] Roll fuchsia [starnix] Handle concurrent exit in PidEntry::get_process\n\n`PidEntry::get_process()` reads `self.process` lock-free under RCU. If a\nreader loads a `Weak\u003cThreadGroup\u003e` right as the thread group exits and\ndrops its last strong reference, `upgrade()` returns `None`, causing an\n`.expect()` panic.\n\nTreat `None` as `Some(ProcessEntryRef::Zombie)`, matching the concurrent\ntransition to zombie state.\n\nAdd a unit test for `get_process()` with a released `ThreadGroup`, and\na syscall test racing `CLONE_PIDFD` exits against `pidfd_open()`.\n\nOriginal-Original-Bug: 565393899\nTest: fx test starnix_core_tests, pidfd\nOriginal-Original-Reviewed-on: https://fuchsia-review.googlesource.com/c/fuchsia/+/1844087\nOriginal-Original-Revision: 117f5f22f39bce9db184e736d68404aa8b4566da\nGitOrigin-RevId: 32dbbc9442ed4ba61d592d7cb716740d1f3ec67f\nChange-Id: I59c5a26365b0dd8ece9252a225c1d63e0867986e\nReviewed-on: https://fuchsia-review.googlesource.com/c/integration/+/1857336\nCr-Commit-Position: refs/heads/main@{#203679}\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "582490c3a0997096154153fd5d454910f4b8dc27",
      "old_mode": 33188,
      "old_path": "stem",
      "new_id": "eb5269b5b41018f6955df1a1d95aa0fad1e63f4c",
      "new_mode": 33188,
      "new_path": "stem"
    }
  ]
}