)]}'
{
  "commit": "c1c201200a359cf3b6e2e36a4236cdca77a3cd8e",
  "tree": "e214a2773f20b814cb8cc8c5509066cd9ec98a1a",
  "parents": [
    "5955c7a2cfb6a35429adea5dc480002b15ca8cfc"
  ],
  "author": {
    "name": "Boaz Harrosh",
    "email": "bharrosh@panasas.com",
    "time": "Tue Feb 03 07:47:29 2009 +0100"
  },
  "committer": {
    "name": "Jens Axboe",
    "email": "jens.axboe@oracle.com",
    "time": "Wed Feb 18 10:32:00 2009 +0100"
  },
  "message": "bsg: Fix sense buffer bug in SG_IO\n\nWhen submitting requests via SG_IO, which does a sync io, a\nbsg_command is not allocated. So an in-Kernel sense_buffer was not\nset. However when calling blk_execute_rq() with no sense buffer\none is provided from the stack. Now bsg at blk_complete_sgv4_hdr_rq()\nwould check if rq-\u003esense_len and a sense was requested by sg_io_v4\nthe rq-\u003esense was copy_user() back, but by now it is already mangled\nstack memory.\n\nI have fixed that by forcing a sense_buffer when calling bsg_map_hdr().\nThe bsg_command-\u003esense is provided in the write/read path like before,\nand on-the-stack buffer is provided when doing SG_IO.\n\nI have also fixed a dprintk message to print rq-\u003eerrors in hex because\nof the scsi bit-field use of this member. For other block devices it\ndoes not matter anyway.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\nAcked-by: FUJITA Tomonori \u003cfujita.tomonori@lab.ntt.co.jp\u003e\nSigned-off-by: Jens Axboe \u003cjens.axboe@oracle.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "d414bb5607e89f4e3a05466b2a488db5e79c98a3",
      "old_mode": 33188,
      "old_path": "block/bsg.c",
      "new_id": "0ce8806dd0c1a3786b6cd86ba8a90b743ca81b8d",
      "new_mode": 33188,
      "new_path": "block/bsg.c"
    }
  ]
}