)]}'
{
  "commit": "c03c520d508ba8b3a384f9849700987df8e4c328",
  "tree": "d03407b7d8e749143af57a6e4cdc1eb538f3c43c",
  "parents": [
    "2dddbc2123681f0cc37a891fa61d97a88d5e641c"
  ],
  "author": {
    "name": "Andreas Färber",
    "email": "afaerber@suse.de",
    "time": "Fri Jan 25 16:11:42 2013 +0100"
  },
  "committer": {
    "name": "Andreas Färber",
    "email": "afaerber@suse.de",
    "time": "Mon Jan 28 16:57:56 2013 +0100"
  },
  "message": "cpu: Unconditionalize CPUState fields\n\nCommits fc8c5b8c41ee5ba69d7a2be63b02a08c7b0b155b (Makefile.user: Define\nCONFIG_USER_ONLY for libuser/) and\ndd83b06ae61cfa2dc4381ab49f365bd0995fc930 (qom: Introduce CPU class)\nspecifically prepared the qom/cpu.c file to be compiled differently for\nsoftmmu and *-user. This broke as part of build system refactorings\nwhile CPU patches were in flight, adding conditional fields\nkvm_fd (8737c51c0444f832c4e97d7eb7540eae457e08e4) and\nkvm_vcpu_dirty (20d695a9254c1b086a456d3b79a3c311236643ba) for softmmu.\n\nlinux-user and bsd-user would therefore get a CPUState type with\ninstance_size ~8 bytes longer than expected.\nFix this by unconditionally having the fields in CPUState.\n\nIn practice, target-specific CPU types\u0027 instance_size would compensate\nthis, and upstream qom/cpu.c does not yet touch any affected field.\n\nSigned-off-by: Andreas Färber \u003cafaerber@suse.de\u003e\nReviewed-by: Paolo Bonzini \u003cpbonzini@redhat.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "8097692bdea258103f37d9bf7c0155cd0ef0f97e",
      "old_mode": 33188,
      "old_path": "include/qom/cpu.h",
      "new_id": "46f2247274a82eaed90cfaf90d40c0d91b927eff",
      "new_mode": 33188,
      "new_path": "include/qom/cpu.h"
    }
  ]
}