)]}'
{
  "commit": "b5f862180d7011d9575d0499fa37f0f25b423b12",
  "tree": "1148c40af363908755859fc2179e1890821f27aa",
  "parents": [
    "39a4b86f0de4ce5024985a56fc39b16194b04313"
  ],
  "author": {
    "name": "Hangbin Liu",
    "email": "liuhangbin@gmail.com",
    "time": "Mon Nov 06 09:01:57 2017 +0800"
  },
  "committer": {
    "name": "David S. Miller",
    "email": "davem@davemloft.net",
    "time": "Wed Nov 08 11:28:43 2017 +0900"
  },
  "message": "bonding: discard lowest hash bit for 802.3ad layer3+4\n\nAfter commit 07f4c90062f8 (\"tcp/dccp: try to not exhaust ip_local_port_range\nin connect()\"), we will try to use even ports for connect(). Then if an\napplication (seen clearly with iperf) opens multiple streams to the same\ndestination IP and port, each stream will be given an even source port.\n\nSo the bonding driver\u0027s simple xmit_hash_policy based on layer3+4 addressing\nwill always hash all these streams to the same interface. And the total\nthroughput will limited to a single slave.\n\nChange the tcp code will impact the whole tcp behavior, only for bonding\nusage. Paolo Abeni suggested fix this by changing the bonding code only,\nwhich should be more reasonable, and less impact.\n\nFix this by discarding the lowest hash bit because it contains little entropy.\nAfter the fix we can re-balance between slaves.\n\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\nSigned-off-by: Hangbin Liu \u003cliuhangbin@gmail.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "c99dc59d729b30dd3ffc4a673deae38630fb0a84",
      "old_mode": 33188,
      "old_path": "drivers/net/bonding/bond_main.c",
      "new_id": "76e8054bfc4eb371f03a5d1f18584c842bbb6d03",
      "new_mode": 33188,
      "new_path": "drivers/net/bonding/bond_main.c"
    }
  ]
}