)]}'
{
  "commit": "2c8fc867602e385fd2abe76da0b6bda8ed907547",
  "tree": "21507c7b823953e8632ae4a1853d1196b6ff02a3",
  "parents": [
    "9b13776977d45505469edc6decc93e9e3799afe2"
  ],
  "author": {
    "name": "Borislav Petkov",
    "email": "bp@alien8.de",
    "time": "Mon Oct 03 14:28:18 2011 -0400"
  },
  "committer": {
    "name": "David S. Miller",
    "email": "davem@davemloft.net",
    "time": "Mon Oct 03 14:28:18 2011 -0400"
  },
  "message": "ide-disk: Fix request requeuing\n\nSimon Kirby reported that on his RAID setup with idedisk underneath\nthe box OOMs after a couple of days of runtime. Running with\nCONFIG_DEBUG_KMEMLEAK pointed to idedisk_prep_fn() which unconditionally\nallocates an ide_cmd struct. However, ide_requeue_and_plug() can be\ncalled more than once per request, either from the request issue or the\nIRQ handler path and do blk_peek_request() ends up in idedisk_prep_fn()\nrepeatedly, allocating a struct ide_cmd everytime and \"forgetting\" the\nprevious pointer.\n\nMake sure the code reuses the old allocated chunk.\n\nReported-and-tested-by: Simon Kirby \u003csim@hostway.ca\u003e\nCc: \u003cstable@kernel.org\u003e [ 39.x, 3.0.x ]\nLink: http://marc.info/?l\u003dlinux-kernel\u0026m\u003d131667641517919\nLink: http://lkml.kernel.org/r/20110922072643.GA27232@hostway.ca\nSigned-off-by: Borislav Petkov \u003cbp@alien8.de\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "274798068a54648c0b504045982a02fda95c0892",
      "old_mode": 33188,
      "old_path": "drivers/ide/ide-disk.c",
      "new_id": "16f69be820c7bb13da94f253b5e7bb33c831fea5",
      "new_mode": 33188,
      "new_path": "drivers/ide/ide-disk.c"
    }
  ]
}