)]}'
{
  "commit": "d7e58f412cf6c5426efda60558f0ccfbf709f646",
  "tree": "b149ccc8fa2ae8c69d88ffafcbdc7c3f8579ef4e",
  "parents": [
    "87bb9e953e6729920701dbc439a498586ae7e305"
  ],
  "author": {
    "name": "Fabiano Rosas",
    "email": "farosas@suse.de",
    "time": "Tue Aug 27 14:45:58 2024 -0300"
  },
  "committer": {
    "name": "Fabiano Rosas",
    "email": "farosas@suse.de",
    "time": "Tue Sep 03 16:24:35 2024 -0300"
  },
  "message": "migration/multifd: Don\u0027t send ram data during SYNC\n\nSkip saving and loading any ram data in the packet in the case of a\nSYNC. This fixes a shortcoming of the current code which requires a\nreset of the MultiFDPages_t fields right after the previous\npending_job finishes, otherwise the very next job might be a SYNC and\nmultifd_send_fill_packet() will put the stale values in the packet.\n\nBy not calling multifd_ram_fill_packet(), we can stop resetting\nMultiFDPages_t in the multifd core and leave that to the client code.\n\nActually moving the reset function is not yet done because\npages-\u003enum\u003d\u003d0 is used by the client code to determine whether the\nMultiFDPages_t needs to be flushed. The subsequent patches will\nreplace that with a generic flag that is not dependent on\nMultiFDPages_t.\n\nReviewed-by: Peter Xu \u003cpeterx@redhat.com\u003e\nSigned-off-by: Fabiano Rosas \u003cfarosas@suse.de\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "d64fcdf4ac6f619d93318aad45dda3aa21da1311",
      "old_mode": 33188,
      "old_path": "migration/multifd.c",
      "new_id": "3a164c124d70c63658c0a8b7a9c98c738639b1f3",
      "new_mode": 33188,
      "new_path": "migration/multifd.c"
    }
  ]
}