)]}'
{
  "commit": "15d33c070ddde99f7368e6b17b71d22c866f97d9",
  "tree": "00e779b01fb3c48e61b3ccc4de8f082e1b929cc4",
  "parents": [
    "699784b7614ce61b16a075445b8e5b6c379c5086"
  ],
  "author": {
    "name": "David S. Miller",
    "email": "davem@sunset.davemloft.net",
    "time": "Mon Apr 09 13:23:14 2007 -0700"
  },
  "committer": {
    "name": "David S. Miller",
    "email": "davem@sunset.davemloft.net",
    "time": "Mon Apr 09 13:31:15 2007 -0700"
  },
  "message": "[TCP]: slow_start_after_idle should influence cwnd validation too\n\nFor the cases that slow_start_after_idle are meant to deal\nwith, it is almost a certainty that the congestion window\ntests will think the connection is application limited and\nwe\u0027ll thus decrease the cwnd there too.  This defeats the\nwhole point of setting slow_start_after_idle to zero.\n\nSo test it there too.\n\nWe do not cancel out the entire tcp_cwnd_validate() function\nso that if the sysctl is changed we still have the validation\nstate maintained.\n\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "688b95594f2e9de0f2724889855e470ce6aab376",
      "old_mode": 33188,
      "old_path": "net/ipv4/tcp_output.c",
      "new_id": "3c24881f2a6598941fbc5663d8bc6c2263e0ee21",
      "new_mode": 33188,
      "new_path": "net/ipv4/tcp_output.c"
    }
  ]
}