)]}'
{
  "commit": "170364619ac21c2b14869571eeaf767ae825f96c",
  "tree": "db7a32b63816e0523aef421588aedd8b99402180",
  "parents": [
    "e6c033f79a0a1e9ca850575dcfa51bb583b592fa"
  ],
  "author": {
    "name": "Christoph Hellwig",
    "email": "hch@lst.de",
    "time": "Mon Oct 05 09:31:06 2015 +0200"
  },
  "committer": {
    "name": "NeilBrown",
    "email": "neilb@suse.com",
    "time": "Sun Nov 01 13:48:27 2015 +1100"
  },
  "message": "raid5-cache: free I/O units earlier\n\nThere is no good reason to keep the I/O unit structures around after the\nstripe has been written back to the RAID array.  The only information\nwe need is the log sequence number, and the checkpoint offset of the\nhighest successfull writeback.  Store those in the log structure, and\nfree the IO units from __r5l_stripe_write_finished.\n\nBesides simplifying the code this also avoid having to keep the allocation\nfor the I/O unit around for a potentially long time as superblock updates\nthat checkpoint the log do not happen very often.\n\nThis also fixes the previously incorrect calculation of \u0027free\u0027 in\nr5l_do_reclaim as a side effect: previous if took the last unit which\nisn\u0027t checkpointed into account.\n\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Shaohua Li \u003cshli@fb.com\u003e\nSigned-off-by: NeilBrown \u003cneilb@suse.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "289ca3f5d4b32e57284f6f05b83691ff38164f1b",
      "old_mode": 33188,
      "old_path": "drivers/md/raid5-cache.c",
      "new_id": "604c64505232def69873195194d3d3d3d51b85bf",
      "new_mode": 33188,
      "new_path": "drivers/md/raid5-cache.c"
    }
  ]
}