)]}'
{
  "commit": "4a44bac1f98223ed77e47bf3b42fcfd10cddd85f",
  "tree": "4a31217a28d701d9ae6733ad236163d4c093c977",
  "parents": [
    "8144737def6abc49457124424887436531bd6a50"
  ],
  "author": {
    "name": "Ingo Molnar",
    "email": "mingo@elte.hu",
    "time": "Thu Mar 19 13:21:44 2009 +0100"
  },
  "committer": {
    "name": "Ingo Molnar",
    "email": "mingo@elte.hu",
    "time": "Thu Mar 19 13:38:35 2009 +0100"
  },
  "message": "symbols, stacktrace: look up init symbols after module symbols\n\nImpact: fix incomplete stacktraces\n\nI noticed such weird stacktrace entries in lockdep dumps:\n\n[    0.285956] {HARDIRQ-ON-W} state was registered at:\n[    0.285956]   [\u003cffffffff802bce90\u003e] mark_irqflags+0xbe/0x125\n[    0.285956]   [\u003cffffffff802bf2fd\u003e] __lock_acquire+0x674/0x82d\n[    0.285956]   [\u003cffffffff802bf5b2\u003e] lock_acquire+0xfc/0x128\n[    0.285956]   [\u003cffffffff8135b636\u003e] rt_spin_lock+0xc8/0xd0\n[    0.285956]   [\u003cffffffffffffffff\u003e] 0xffffffffffffffff\n\nThe stacktrace entry is cut off after rt_spin_lock.\n\nAfter much debugging i found out that stacktrace entries that\nbelong to init symbols dont get printed out, due to commit:\n\n  a2da405: module: Don\u0027t report discarded init pages as kernel text.\n\nThe reason is this check added to core_kernel_text():\n\n-       if (addr \u003e\u003d (unsigned long)_sinittext \u0026\u0026\n+       if (system_state \u003d\u003d SYSTEM_BOOTING \u0026\u0026\n+           addr \u003e\u003d (unsigned long)_sinittext \u0026\u0026\n            addr \u003c\u003d (unsigned long)_einittext)\n                return 1;\n\nThis will discard inittext symbols even though their symbol table\nis still present and even though stacktraces done while the system\nwas booting up might still be relevant.\n\nTo not reintroduce the (not well-specified) bug addressed in that\ncommit, first do a module symbols lookup, then a final init-symbols\nlookup.\n\nThis will work fine on architectures that have separate address\nspaces for modules (such as x86) - and should not crash any other\narchitectures either.\n\nAcked-by: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nCc: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nLKML-Reference: \u003cnew-discussion\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "e136ed8d82ba56ab81283a48fe3d6cfb41c23c79",
      "old_mode": 33188,
      "old_path": "kernel/extable.c",
      "new_id": "c46da6a47036708dc373f074503efe6483bcaf2c",
      "new_mode": 33188,
      "new_path": "kernel/extable.c"
    }
  ]
}