)]}'
{
  "log": [
    {
      "commit": "e7cd2514ea506f06bd4f7b13a9b62afd60f9c73b",
      "tree": "66c46a54a2841091165a0b8083329c41f3504847",
      "parents": [
        "4726204200327c04a77b819e2c653c063f1bc6ab"
      ],
      "author": {
        "name": "Herbert Xu",
        "email": "herbert@gondor.apana.org.au",
        "time": "Fri Dec 14 22:28:14 2007 +0800"
      },
      "committer": {
        "name": "Herbert Xu",
        "email": "herbert@gondor.apana.org.au",
        "time": "Fri Jan 11 08:16:55 2008 +1100"
      },
      "message": "[CRYPTO] chainiv: Avoid lock spinning where possible\n\nThis patch makes chainiv avoid spinning by postponing requests on lock\ncontention if the user allows the use of asynchronous algorithms.  If\na synchronous algorithm is requested then we behave as before.\n\nThis should improve IPsec performance on SMP when two CPUs attempt to\ntransmit over the same SA.  Currently one of them will spin doing nothing\nwaiting for the other CPU to finish its encryption.  This patch makes it\npostpone the request and get on with other work.\n\nIf only one CPU is transmitting for a given SA, then we will process\nthe request synchronously as before.\n\nSigned-off-by: Herbert Xu \u003cherbert@gondor.apana.org.au\u003e\n"
    },
    {
      "commit": "7f47073911f0e4384d38a0827d28305a177c8816",
      "tree": "0732253520cdb0ff686166256923bc633ddceace",
      "parents": [
        "ecfc43292f68566c144afca966b46b371c26d56c"
      ],
      "author": {
        "name": "Herbert Xu",
        "email": "herbert@gondor.apana.org.au",
        "time": "Tue Nov 27 23:17:23 2007 +0800"
      },
      "committer": {
        "name": "Herbert Xu",
        "email": "herbert@gondor.apana.org.au",
        "time": "Fri Jan 11 08:16:44 2008 +1100"
      },
      "message": "[CRYPTO] chainiv: Add chain IV generator\n\nThe chain IV generator is the one we\u0027ve been using in the IPsec stack.\nIt simply starts out with a random IV, then uses the last block of each\nencrypted packet\u0027s cipher text as the IV for the next packet.\n\nIt can only be used by synchronous ciphers since we have to make sure\nthat we don\u0027t start the encryption of the next packet until the last\none has completed.\n\nIt does have the advantage of using very little CPU time since it doesn\u0027t\nhave to generate anything at all.\n\nSigned-off-by: Herbert Xu \u003cherbert@gondor.apana.org.au\u003e\n"
    }
  ]
}