)]}'
{
  "commit": "2d431d0139493cdefb49f662b5df295d8bc32f65",
  "tree": "987649287594d9f2c631124eb5e235a50460d22e",
  "parents": [
    "dd0ccd82d7b8c56200c71de8f1f51c6c2865cff2"
  ],
  "author": {
    "name": "Indu Bhagat",
    "email": "indu.bhagat@oracle.com",
    "time": "Tue Dec 23 14:57:58 2025 -0800"
  },
  "committer": {
    "name": "Indu Bhagat",
    "email": "indu.bhagat@oracle.com",
    "time": "Tue Dec 23 14:58:05 2025 -0800"
  },
  "message": "Revert \"gas: Don\u0027t skip SFrame FDE if .cfi_register specifies RA w/o tracking\"\n\nThis reverts commit df174a6b760.\n\n    commit df174a6b760ac5c66ee62ecb872c760ae21fd4cd\n    Date:   Thu Jul 4 10:34:12 2024 +0200\n\n    gas: Don\u0027t skip SFrame FDE if .cfi_register specifies RA w/o tracking\n\n    Do not skip SFrame FDE if .cfi_register specifies RA register without\n    RA tracking being actually used. Without RA tracking the register\n    contents can always be restored from the stack using the fixed\n    RA offset from CFA.\n\nEven for ABI/arch without RA tracking, there may be instances where user\nmay specify \u0027.cfi_register RA, reg\u0027.  This needs to be caught, skipping\nthis from SFrame generation may not be correct.  This may be done in\ncertain hand-written asm sequences where the user needs to manipulate\nthe return to a certain function.\n\nNo testcase is being added ATM because in SFrame V3, a new FDE type can\nbe used to represent such cases (A new test case will be added then).\n\nReviewed-by: Jens Remus \u003cjremus@linux.ibm.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "f31cbaee060e19d5d8af0cf4a98fd1943fde5a62",
      "old_mode": 33188,
      "old_path": "gas/gen-sframe.c",
      "new_id": "5b1b1ba23ea8ac7420a3eb61dc562c6df23d2f7e",
      "new_mode": 33188,
      "new_path": "gas/gen-sframe.c"
    }
  ]
}