)]}'
{
  "commit": "dc8d76cac942e7344a72ad18afb90fa46cf20bb4",
  "tree": "0516fdca404bc9a87271710a93159bfda5a04c77",
  "parents": [
    "1959ca80e1f88b82c1cb7227f437910768ab0c94"
  ],
  "author": {
    "name": "Ben Skeggs",
    "email": "bskeggs@redhat.com",
    "time": "Wed Jan 06 12:00:02 2010 +1000"
  },
  "committer": {
    "name": "Dave Airlie",
    "email": "airlied@redhat.com",
    "time": "Mon Jan 11 14:41:16 2010 +1000"
  },
  "message": "drm/nv50: prevent a possible ctxprog hang\n\nThe below is mainly an educated guess at what\u0027s going on, docs would\nsure be handy...  NVIDIA? :P\n\nIt appears it\u0027s possible for a ctxprog to run even while a GPU exception\nis pending.  The GF8 and up ctxprogs appear to have a small snippet of\ncode which detects this, and stalls the ctxprog until it\u0027s been handled,\nwhich essentially looks like:\n\n\tif (r2 \u0026 0x00008000) {\n\t\tr0 |\u003d 0x80000000;\n\t\twhile (r0 \u0026 0x80000000) {}\n\t}\n\nI don\u0027t know of any way that flag would get cleared unless the driver\nintervenes (and indeed, in the cases I\u0027ve seen the hang, nothing steps\nin to automagically clear it for us).  This patch causes the driver to\nclear the flag during the PGRAPH IRQ handler.\n\nSigned-off-by: Ben Skeggs \u003cbskeggs@redhat.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "370c72c968d186ce779306ae20552ef4a6cab715",
      "old_mode": 33188,
      "old_path": "drivers/gpu/drm/nouveau/nouveau_irq.c",
      "new_id": "919a619ca7fa376c70f0a935df029935203e924d",
      "new_mode": 33188,
      "new_path": "drivers/gpu/drm/nouveau/nouveau_irq.c"
    }
  ]
}