)]}'
{
  "commit": "7ad7a345a4f17c08a1bb9bfdbb62f7793d84aa36",
  "tree": "83839c0af2da4529d5fdec1b9c961b541d572c66",
  "parents": [
    "6d23ea4b1906f28f5d99ad6aeef7207c48be6bfd"
  ],
  "author": {
    "name": "Lothar Waßmann",
    "email": "LW@KARO-electronics.de",
    "time": "Thu Dec 08 09:15:44 2011 +0100"
  },
  "committer": {
    "name": "Vinod Koul",
    "email": "vinod.koul@linux.intel.com",
    "time": "Fri Dec 23 20:54:55 2011 +0530"
  },
  "message": "dma: mxs-dma: Don\u0027t use CLKGATE bits in CTRL0 to disable DMA channels\n\nThis is how the original Freescale code (unintentionally) worked,\nbecause the code path which would have asserted the CLKGATE bit was\nnever actually reached in their code.\nThis fixes the nefarious \"DMA timout\" bug when multiple DMA channels\n(e.g. GPMI NAND and MMC) are used at the same time.\nIf a better fix for this problem should be found, the clkgate handling\ncould be reinstated.\nSee http://lists.infradead.org/pipermail/linux-arm-kernel/2011-September/065228.html\n\nAlso reverse the order of mxs_dma_disable_chan() and\nmxs_dma_reset_chan() in mxs_dma_control() because mxs_dma_reset_chan()\ncan only work when the DMA channel is enabled.\n\nSigned-off-by: Lothar Waßmann \u003cLW@KARO-electronics.de\u003e\nSigned-off-by: Vinod Koul \u003cvinod.koul@linux.intel.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "1b4c6be3aacb071224002f2864a723cea5a8b363",
      "old_mode": 33188,
      "old_path": "drivers/dma/mxs-dma.c",
      "new_id": "6548595c26dcfafc4aeb66c2c0b1c395e0cc08a9",
      "new_mode": 33188,
      "new_path": "drivers/dma/mxs-dma.c"
    }
  ]
}