)]}'
{
  "commit": "0d31c7c200b3dca2aeeaa6f74ff3fd539aad803a",
  "tree": "e8b2a49fe7133c5c81387cbcc1406518789e6b2a",
  "parents": [
    "2d68e22e94e8bc5a0d32a38b53c592c320db8261"
  ],
  "author": {
    "name": "Denis V. Lunev",
    "email": "den@openvz.org",
    "time": "Tue Apr 28 10:46:57 2015 +0300"
  },
  "committer": {
    "name": "Stefan Hajnoczi",
    "email": "stefanha@redhat.com",
    "time": "Fri May 22 09:37:32 2015 +0100"
  },
  "message": "block/parallels: delay writing to BAT till bdrv_co_flush_to_os\n\nThe idea is that we do not need to immediately sync BAT to the image as\nfrom the guest point of view there is a possibility that IO is lost\neven in the physical controller until flush command was finished.\nbdrv_co_flush_to_os is exactly the right place for this purpose.\n\nTechnically the patch uses loaded BAT data as a cache and performs\nactual on-disk metadata updates in parallels_co_flush_to_os callback.\n\nThis patch speed ups\n  qemu-img create -f parallels -o cluster_size\u003d64k ./1.hds 64G\n  qemu-io -f parallels -c \"write -P 0x11 0 1024k\" 1.hds\nwriting from 50-60 Mb/sec to 80-90 Mb/sec on rotational media and\nfrom 160 Mb/sec to 190 Mb/sec on SSD disk.\n\nSigned-off-by: Denis V. Lunev \u003cden@openvz.org\u003e\nReviewed-by: Roman Kagan \u003crkagan@parallels.com\u003e\nReviewed-by: Stefan Hajnoczi \u003cstefanha@redhat.com\u003e\nSigned-off-by: Roman Kagan \u003crkagan@parallels.com\u003e\nMessage-id: 1430207220-24458-25-git-send-email-den@openvz.org\nCC: Kevin Wolf \u003ckwolf@redhat.com\u003e\nSigned-off-by: Stefan Hajnoczi \u003cstefanha@redhat.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "4d8a0d48aad1a1f4c5bc9cbcee6e037748ad36f4",
      "old_mode": 33188,
      "old_path": "block/parallels.c",
      "new_id": "05fe030f055bfc5bd6b4a0b50c2ed0632ddc4cc8",
      "new_mode": 33188,
      "new_path": "block/parallels.c"
    }
  ]
}