)]}'
{
  "commit": "0b0a0806b0d8635e046bf533225a25903b1cddce",
  "tree": "245ac39eb1b7c6008a669ab2d68ed174d66bd7bc",
  "parents": [
    "c15d8a6499d04e5d2cac07f8120f207bb275f60f"
  ],
  "author": {
    "name": "Hugh Dickins",
    "email": "hugh@veritas.com",
    "time": "Tue Feb 24 20:51:52 2009 +0000"
  },
  "committer": {
    "name": "Linus Torvalds",
    "email": "torvalds@linux-foundation.org",
    "time": "Wed Feb 25 12:21:42 2009 -0800"
  },
  "message": "shmem: fix shared anonymous accounting\n\nEach time I exit Firefox, /proc/meminfo\u0027s Committed_AS goes down almost\n400 kB: OVERCOMMIT_NEVER would be allowing overcommits it should\nprohibit.\n\nCommit fc8744adc870a8d4366908221508bb113d8b72ee \"Stop playing silly\ngames with the VM_ACCOUNT flag\" changed shmem_file_setup() to set the\nshmem file\u0027s VM_ACCOUNT flag according to VM_NORESERVE not being set in\nthe vma flags; but did so only _after_ the shmem_acct_size(flags, size)\ncall which is expected to pre-account a shared anonymous object.\n\nIt\u0027s all clearer if we switch shmem.c over to use VM_NORESERVE\nthroughout in place of !VM_ACCOUNT.\n\nBut I very nearly sent in a patch which mistakenly removed the\naccounting from tmpfs files: shmem_get_inode()\u0027s memset was good for not\nsetting VM_ACCOUNT, but now it needs to set VM_NORESERVE.\n\nRather than setting that by default, then perhaps clearing it again in\nshmem_file_setup(), let\u0027s pass it as a flag to shmem_get_inode(): that\nallows us to remove the #ifdef CONFIG_SHMEM from shmem_file_setup().\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "19d566ccdeeab8c6a1f0431602b3bd3dd5a76026",
      "old_mode": 33188,
      "old_path": "mm/shmem.c",
      "new_id": "4103a239ce843326faca51c61c5b0b105797b24b",
      "new_mode": 33188,
      "new_path": "mm/shmem.c"
    }
  ]
}