)]}'
{
  "commit": "85784f9395987a422fa04263e7c0fb13da11eb5c",
  "tree": "f8f998acb306c25992468d0a52e9082616b37ea5",
  "parents": [
    "3eb2ce825ea1ad89d20f7a3b5780df850e4be274"
  ],
  "author": {
    "name": "Yan, Zheng",
    "email": "zyan@redhat.com",
    "time": "Fri Mar 16 11:22:29 2018 +0800"
  },
  "committer": {
    "name": "Ilya Dryomov",
    "email": "idryomov@gmail.com",
    "time": "Fri Mar 30 11:17:48 2018 +0200"
  },
  "message": "ceph: only dirty ITER_IOVEC pages for direct read\n\nIf a page is already locked, attempting to dirty it leads to a deadlock\nin lock_page().  This is what currently happens to ITER_BVEC pages when\na dio-enabled loop device is backed by ceph:\n\n  $ losetup --direct-io /dev/loop0 /mnt/cephfs/img\n  $ xfs_io -c \u0027pread 0 4k\u0027 /dev/loop0\n\nFollow other file systems and only dirty ITER_IOVEC pages.\n\nCc: stable@kernel.org\nSigned-off-by: \"Yan, Zheng\" \u003czyan@redhat.com\u003e\nReviewed-by: Ilya Dryomov \u003cidryomov@gmail.com\u003e\nSigned-off-by: Ilya Dryomov \u003cidryomov@gmail.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "6639926eed4e8461d0c50b6a686df236e1c96338",
      "old_mode": 33188,
      "old_path": "fs/ceph/file.c",
      "new_id": "b67eec3532a125c0c8ca0c97b7b0f22c650f7966",
      "new_mode": 33188,
      "new_path": "fs/ceph/file.c"
    }
  ]
}