)]}'
{
  "commit": "33304ec9fa484e765c6249673e09e1b7d49c5b85",
  "tree": "9e469f03b22e55233921b39cb9c6ec765dceb293",
  "parents": [
    "16f0587e0a5da5b1ad76cb7c3739491bc042201c"
  ],
  "author": {
    "name": "Hu Tao",
    "email": "hutao@cn.fujitsu.com",
    "time": "Sun Jan 26 11:12:38 2014 +0800"
  },
  "committer": {
    "name": "Kevin Wolf",
    "email": "kwolf@redhat.com",
    "time": "Sun Feb 09 09:12:39 2014 +0100"
  },
  "message": "qcow2: fix offset overflow in qcow2_alloc_clusters_at()\n\nWhen cluster size is big enough it can lead to an offset overflow\nin qcow2_alloc_clusters_at(). This patch fixes it.\n\nThe allocation is stopped each time at L2 table boundary\n(see handle_alloc()), so the possible maximum bytes could be\n\n  2^(cluster_bits - 3 + cluster_bits)\n\ncluster_bits - 3 is used to compute the number of entry by L2\nand the additional cluster_bits is to take into account each\nclusters referenced by the L2 entries.\n\nso int is safe for cluster_bits\u003c\u003d17, unsafe otherwise.\n\nSigned-off-by: Hu Tao \u003chutao@cn.fujitsu.com\u003e\nReviewed-by: Max Reitz \u003cmreitz@redhat.com\u003e\nReviewed-by: Benoit Canet \u003cbenoit@irqsave.net\u003e\nSigned-off-by: Kevin Wolf \u003ckwolf@redhat.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "c974abe7951093ca5a085fea4ad7bd2f45de65d5",
      "old_mode": 33188,
      "old_path": "block/qcow2-refcount.c",
      "new_id": "8712d8bd549d70df987e54052cabab5d4734b857",
      "new_mode": 33188,
      "new_path": "block/qcow2-refcount.c"
    }
  ]
}