)]}'
{
  "commit": "dc49e3445aa703eb7fd33c7ddb7e4a7bbcf06d30",
  "tree": "eb164ad61b92f2df2ffab9628adffe06e635c8e2",
  "parents": [
    "be0d03f1c3d3612fe2b6aa451ae87a89382c9231"
  ],
  "author": {
    "name": "Satoshi Oshima",
    "email": "soshima@redhat.com",
    "time": "Sat May 20 15:00:21 2006 -0700"
  },
  "committer": {
    "name": "Linus Torvalds",
    "email": "torvalds@g5.osdl.org",
    "time": "Sun May 21 12:59:21 2006 -0700"
  },
  "message": "[PATCH] kprobes: bad manipulation of 2 byte opcode on x86_64\n\nProblem:\n\nIf we put a probe onto a callq instruction and the probe is executed,\nkernel panic of Bad RIP value occurs.\n\nRoot cause:\n\nIf resume_execution() found 0xff at first byte of p-\u003eainsn.insn, it must\ncheck the _second_ byte.  But current resume_execution check _first_ byte\nagain.\n\nI changed it checks second byte of p-\u003eainsn.insn.\n\nKprobes on i386 don\u0027t have this problem, because the implementation is a\nlittle bit different from x86_64.\n\nCc: Andi Kleen \u003cak@muc.de\u003e\nSigned-off-by: Satoshi Oshima \u003csoshima@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "1eaa5dae6174ff57b514a56828f07e6f8ed0a149",
      "old_mode": 33188,
      "old_path": "arch/x86_64/kernel/kprobes.c",
      "new_id": "fa1d19ca700ae82a23df77f068bee660e47f80f2",
      "new_mode": 33188,
      "new_path": "arch/x86_64/kernel/kprobes.c"
    }
  ]
}