)]}'
{
  "commit": "f4ac7b5eb79ef15819c966b1f6b84bf443949123",
  "tree": "cfc81caaa2d1d50f0986c99cf59764060993eb60",
  "parents": [
    "f13f4ca8036516ca1b99a41f95f7dea7e4dce104"
  ],
  "author": {
    "name": "Benjamin Herrenschmidt",
    "email": "benh@ozlabs.org",
    "time": "Wed Apr 09 17:21:36 2008 +1000"
  },
  "committer": {
    "name": "Paul Mackerras",
    "email": "paulus@samba.org",
    "time": "Fri Apr 18 15:37:16 2008 +1000"
  },
  "message": "[POWERPC] Fix device-tree locking vs. interrupts\n\nLockdep found out that we can occasionally take the device-tree\nlock for reading from softirq time (from rtas_token called\nby the rtas real time clock code called by the NTP code),\nwhile we take it occasionally for writing without masking\ninterrupts. The combination of those two can thus deadlock.\n\nWhile some of those cases of interrupt read lock could be fixed\n(such as caching the RTAS tokens) I figured that taking the\nlock for writing is so rare (device-tree modification) that we\nmay as well penalize that case and allow reading from interrupts.\n\nThus, this turns all the writers to take the lock with irqs\nmasked to avoid the situation.\n\nSigned-off-by: Benjamin Herrenschmidt \u003cbenh@kernel.crashing.org\u003e\nSigned-off-by: Paul Mackerras \u003cpaulus@samba.org\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "8a9359ae471853688e509d42bd2999b78d1a87ca",
      "old_mode": 33188,
      "old_path": "arch/powerpc/kernel/prom.c",
      "new_id": "3bfe7837e8207978062b0d8426a0f50b04b6c448",
      "new_mode": 33188,
      "new_path": "arch/powerpc/kernel/prom.c"
    }
  ]
}