)]}'
{
  "commit": "dd4d4ea0add97df078d571b48192adaf7c4b0d87",
  "tree": "a4dbaadf652c32716f3e7583d260cca6c045ffed",
  "parents": [
    "63b66a5147069bda815989732442cb07790609db"
  ],
  "author": {
    "name": "Jordan Niethe",
    "email": "jniethe5@gmail.com",
    "time": "Tue Aug 29 10:12:01 2023 +1000"
  },
  "committer": {
    "name": "Alexey Kardashevskiy",
    "email": "aik@ozlabs.ru",
    "time": "Mon Sep 18 18:20:45 2023 +1000"
  },
  "message": "virtio-serial: Do not close stdout on quiesce\n\nCommit 76fee95 (\"slof: Only close stdout for virtio-serial devices\")\nsays that commit cf28264 (\"virtio-serial: Rework shutdown sequence\")\nfixed a hang. The problem was believed to be that it was necessary to\nclose stdout to shutdown the underlying virtio device.\n\nCommit cf28264 (\"virtio-serial: Rework shutdown sequence\") closed stdout\non quiesce. This meant when prom_init() called write on stdout after\nquiesce, there is a use after free so this is unreliable, and can also\nhang (especially after reboots).\n\nQuiescing is intended to put hardware into a safe state for the client\nto take over. It is incorrect for SLOF to close ihandles that the client\ncould still be using, even after a quiesce.\n\nRather than closing the stdout device, all that needs to happen is to\nensure virtio-serial-shutdown gets called. On quiesce, close the virtio\ndevice, but leave the stdout device itself open.\n\nCommit 8174acd (\"virtio-serial: Close device completely\") handles reads\nand writes as no-ops if the underlying virtio device is closed so there\nis no problem with the client calling \"write\" on stdout after this, but\nno output will be displayed.\n\nFixes: cf28264 (\"virtio-serial: Rework shutdown sequence\")\nDebugged-by: Kautuk Consul \u003ckconsul@linux.vnet.ibm.com\u003e\nCo-developed-by: Kautuk Consul \u003ckconsul@linux.vnet.ibm.com\u003e\nSigned-off-by: Kautuk Consul \u003ckconsul@linux.vnet.ibm.com\u003e\nSigned-off-by: Jordan Niethe \u003cjniethe5@gmail.com\u003e\nReviewed-by: Thomas Huth \u003cthuth@redhat.com\u003e\nSigned-off-by: Alexey Kardashevskiy \u003caik@ozlabs.ru\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "41e2e0400b686e14bfbe81948da5984757737a00",
      "old_mode": 33188,
      "old_path": "board-qemu/slof/virtio-serial.fs",
      "new_id": "de42cc723ee344f0733636e20abd2e18e173dfdd",
      "new_mode": 33188,
      "new_path": "board-qemu/slof/virtio-serial.fs"
    }
  ]
}