)]}'
{
  "log": [
    {
      "commit": "f03c65993b98eeb909a4012ce7833c5857d74755",
      "tree": "a6dd5e353889b7fe4ab87c54170d09443d788fec",
      "parents": [
        "7b8a53fd815deb39542085897743fa0063f9fe06"
      ],
      "author": {
        "name": "Al Viro",
        "email": "viro@zeniv.linux.org.uk",
        "time": "Fri Jan 14 22:30:21 2011 -0500"
      },
      "committer": {
        "name": "Al Viro",
        "email": "viro@zeniv.linux.org.uk",
        "time": "Sun Jan 16 13:47:07 2011 -0500"
      },
      "message": "sanitize vfsmount refcounting changes\n\nInstead of splitting refcount between (per-cpu) mnt_count\nand (SMP-only) mnt_longrefs, make all references contribute\nto mnt_count again and keep track of how many are longterm\nones.\n\nAccounting rules for longterm count:\n\t* 1 for each fs_struct.root.mnt\n\t* 1 for each fs_struct.pwd.mnt\n\t* 1 for having non-NULL -\u003emnt_ns\n\t* decrement to 0 happens only under vfsmount lock exclusive\n\nThat allows nice common case for mntput() - since we can\u0027t drop the\nfinal reference until after mnt_longterm has reached 0 due to the rules\nabove, mntput() can grab vfsmount lock shared and check mnt_longterm.\nIf it turns out to be non-zero (which is the common case), we know\nthat this is not the final mntput() and can just blindly decrement\npercpu mnt_count.  Otherwise we grab vfsmount lock exclusive and\ndo usual decrement-and-check of percpu mnt_count.\n\nFor fs_struct.c we have mnt_make_longterm() and mnt_make_shortterm();\nnamespace.c uses the latter in places where we don\u0027t already hold\nvfsmount lock exclusive and opencodes a few remaining spots where\nwe need to manipulate mnt_longterm.\n\nNote that we mostly revert the code outside of fs/namespace.c back\nto what we used to have; in particular, normal code doesn\u0027t need\nto care about two kinds of references, etc.  And we get to keep\nthe optimization Nick\u0027s variant had bought us...\n\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n"
    },
    {
      "commit": "b3e19d924b6eaf2ca7d22cba99a517c5171007b6",
      "tree": "8c1fa4074114a883a4e2de2f7d12eb29ed91bdf1",
      "parents": [
        "c6653a838b1b2738561aff0b8c0f62a9b714bdd9"
      ],
      "author": {
        "name": "Nick Piggin",
        "email": "npiggin@kernel.dk",
        "time": "Fri Jan 07 17:50:11 2011 +1100"
      },
      "committer": {
        "name": "Nick Piggin",
        "email": "npiggin@kernel.dk",
        "time": "Fri Jan 07 17:50:33 2011 +1100"
      },
      "message": "fs: scale mntget/mntput\n\nThe problem that this patch aims to fix is vfsmount refcounting scalability.\nWe need to take a reference on the vfsmount for every successful path lookup,\nwhich often go to the same mount point.\n\nThe fundamental difficulty is that a \"simple\" reference count can never be made\nscalable, because any time a reference is dropped, we must check whether that\nwas the last reference. To do that requires communication with all other CPUs\nthat may have taken a reference count.\n\nWe can make refcounts more scalable in a couple of ways, involving keeping\ndistributed counters, and checking for the global-zero condition less\nfrequently.\n\n- check the global sum once every interval (this will delay zero detection\n  for some interval, so it\u0027s probably a showstopper for vfsmounts).\n\n- keep a local count and only taking the global sum when local reaches 0 (this\n  is difficult for vfsmounts, because we can\u0027t hold preempt off for the life of\n  a reference, so a counter would need to be per-thread or tied strongly to a\n  particular CPU which requires more locking).\n\n- keep a local difference of increments and decrements, which allows us to sum\n  the total difference and hence find the refcount when summing all CPUs. Then,\n  keep a single integer \"long\" refcount for slow and long lasting references,\n  and only take the global sum of local counters when the long refcount is 0.\n\nThis last scheme is what I implemented here. Attached mounts and process root\nand working directory references are \"long\" references, and everything else is\na short reference.\n\nThis allows scalable vfsmount references during path walking over mounted\nsubtrees and unattached (lazy umounted) mounts with processes still running\nin them.\n\nThis results in one fewer atomic op in the fastpath: mntget is now just a\nper-CPU inc, rather than an atomic inc; and mntput just requires a spinlock\nand non-atomic decrement in the common case. However code is otherwise bigger\nand heavier, so single threaded performance is basically a wash.\n\nSigned-off-by: Nick Piggin \u003cnpiggin@kernel.dk\u003e\n"
    },
    {
      "commit": "8df9d1a4142311c084ffeeacb67cd34d190eff74",
      "tree": "512e018114ea506659fac73d838bfca0fb97a5a4",
      "parents": [
        "ffd1f4ed5bddccf2277e3d8525bcedf1983319f8"
      ],
      "author": {
        "name": "Miklos Szeredi",
        "email": "mszeredi@suse.cz",
        "time": "Tue Aug 10 11:41:41 2010 +0200"
      },
      "committer": {
        "name": "Al Viro",
        "email": "viro@zeniv.linux.org.uk",
        "time": "Wed Aug 11 00:29:47 2010 -0400"
      },
      "message": "vfs: show unreachable paths in getcwd and proc\n\nPrepend \"(unreachable)\" to path strings if the path is not reachable\nfrom the current root.\n\nTwo places updated are\n - the return string from getcwd()\n - and symlinks under /proc/$PID.\n\nOther uses of d_path() are left unchanged (we know that some old\nsoftware crashes if /proc/mounts is changed).\n\nSigned-off-by: Miklos Szeredi \u003cmszeredi@suse.cz\u003e\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n"
    },
    {
      "commit": "5dd784d04924be5d8bc066aded0ec3274b20e612",
      "tree": "b92ba8ffd3cad1175a75df0356882e50b5999f98",
      "parents": [
        "09da5916baf6d3fb9ac16c125c801ae6ea151f97"
      ],
      "author": {
        "name": "Jan Blunck",
        "email": "jblunck@suse.de",
        "time": "Thu Feb 14 19:34:38 2008 -0800"
      },
      "committer": {
        "name": "Linus Torvalds",
        "email": "torvalds@woody.linux-foundation.org",
        "time": "Thu Feb 14 21:13:33 2008 -0800"
      },
      "message": "Introduce path_get()\n\nThis introduces the symmetric function to path_put() for getting a reference\nto the dentry and vfsmount of a struct path in the right order.\n\nSigned-off-by: Jan Blunck \u003cjblunck@suse.de\u003e\nSigned-off-by: Andreas Gruenbacher \u003cagruen@suse.de\u003e\nAcked-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n"
    },
    {
      "commit": "1d957f9bf87da74f420424d16ece005202bbebd3",
      "tree": "363d4770c0c74a536524c99ccd2762ce96ee9bbe",
      "parents": [
        "4ac9137858e08a19f29feac4e1f4df7c268b0ba5"
      ],
      "author": {
        "name": "Jan Blunck",
        "email": "jblunck@suse.de",
        "time": "Thu Feb 14 19:34:35 2008 -0800"
      },
      "committer": {
        "name": "Linus Torvalds",
        "email": "torvalds@woody.linux-foundation.org",
        "time": "Thu Feb 14 21:13:33 2008 -0800"
      },
      "message": "Introduce path_put()\n\n* Add path_put() functions for releasing a reference to the dentry and\n  vfsmount of a struct path in the right order\n\n* Switch from path_release(nd) to path_put(\u0026nd-\u003epath)\n\n* Rename dput_path() to path_put_conditional()\n\n[akpm@linux-foundation.org: fix cifs]\nSigned-off-by: Jan Blunck \u003cjblunck@suse.de\u003e\nSigned-off-by: Andreas Gruenbacher \u003cagruen@suse.de\u003e\nAcked-by: Christoph Hellwig \u003chch@lst.de\u003e\nCc: \u003clinux-fsdevel@vger.kernel.org\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nCc: Steven French \u003csfrench@us.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n"
    },
    {
      "commit": "c5e725f33b733a77de622e91b6ba5645fcf070be",
      "tree": "628b2cd365e85b5e6ae3bb8417790cdcd2acf516",
      "parents": [
        "429731b1553bacf9a331c260c317a28aaa878edb"
      ],
      "author": {
        "name": "Jan Blunck",
        "email": "jblunck@suse.de",
        "time": "Thu Feb 14 19:34:31 2008 -0800"
      },
      "committer": {
        "name": "Linus Torvalds",
        "email": "torvalds@woody.linux-foundation.org",
        "time": "Thu Feb 14 21:13:32 2008 -0800"
      },
      "message": "Move struct path into its own header\n\nMove the definition of struct path into its own header file for further\npatches.\n\nSigned-off-by: Jan Blunck \u003cjblunck@suse.de\u003e\nSigned-off-by: Andreas Gruenbacher \u003cagruen@suse.de\u003e\nAcked-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n"
    }
  ]
}