)]}'
{
  "commit": "29c361f18bafeb4afa3b6d3a9ea180fcec0b81d7",
  "tree": "b2cdbd312524e10e511f2cbe48c826cf3bc030d6",
  "parents": [
    "ea1013c1539270e372fc99854bc6e4d94eaeff66"
  ],
  "author": {
    "name": "Jens Axboe",
    "email": "axboe@kernel.dk",
    "time": "Wed Dec 17 19:14:27 2025 -0700"
  },
  "committer": {
    "name": "Jens Axboe",
    "email": "axboe@kernel.dk",
    "time": "Thu Dec 18 14:29:13 2025 -0700"
  },
  "message": "af_unix: don\u0027t post cmsg for SO_INQ unless explicitly asked for\n\nA previous commit added SO_INQ support for AF_UNIX (SOCK_STREAM), but it\nposts a SCM_INQ cmsg even if just msg-\u003emsg_get_inq is set. This is\nincorrect, as -\u003emsg_get_inq is just the caller asking for the remainder\nto be passed back in msg-\u003emsg_inq, it has nothing to do with cmsg. The\noriginal commit states that this is done to make sockets\nio_uring-friendly\", but it\u0027s actually incorrect as io_uring doesn\u0027t use\ncmsg headers internally at all, and it\u0027s actively wrong as this means\nthat cmsg\u0027s are always posted if someone does recvmsg via io_uring.\n\nFix that up by only posting a cmsg if u-\u003erecvmsg_inq is set.\n\nAdditionally, mirror how TCP handles inquiry handling in that it should\nonly be done for a successful return. This makes the logic for the two\nidentical.\n\nCc: stable@vger.kernel.org\nFixes: df30285b3670 (\"af_unix: Introduce SO_INQ.\")\nReported-by: Julian Orth \u003cju.orth@gmail.com\u003e\nLink: https://github.com/axboe/liburing/issues/1509\nSigned-off-by: Jens Axboe \u003caxboe@kernel.dk\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "55cdebfa0da029857d99dce62eec80973675c971",
      "old_mode": 33188,
      "old_path": "net/unix/af_unix.c",
      "new_id": "a7ca74653d94646b2308df5236d2a12a05dfbe01",
      "new_mode": 33188,
      "new_path": "net/unix/af_unix.c"
    }
  ]
}