)]}'
{
  "commit": "feea87cd6b645d5166bdd304aac88f47f63dc2ef",
  "tree": "d39b837c1e1b17d9fd539c997d9b592ac41ed5ce",
  "parents": [
    "fbc8fb36e3636854195705cba13278befb94158d"
  ],
  "author": {
    "name": "Paolo Bonzini",
    "email": "pbonzini@redhat.com",
    "time": "Fri Jul 25 01:10:12 2025 +0200"
  },
  "committer": {
    "name": "Paolo Bonzini",
    "email": "pbonzini@redhat.com",
    "time": "Fri Jul 25 14:51:11 2025 +0200"
  },
  "message": "target/i386: fix width of third operand of VINSERTx128\n\nTable A-5 of the Intel manual incorrectly lists the third operand of\nVINSERTx128 as Wqq, but it is actually a 128-bit value.  This is\nvisible when W is a memory operand close to the end of the page.\n\nFixes the recently-added poly1305_kunit test in linux-next.\n\n(No testcase yet, but I plan to modify test-avx2 to use memory\nclose to the end of the page.  This would work because the test\nvectors correctly have the memory operand as xmm2/m128).\n\nReported-by: Eric Biggers \u003cebiggers@kernel.org\u003e\nTested-by: Eric Biggers \u003cebiggers@kernel.org\u003e\nCc: Ard Biesheuvel \u003cardb@kernel.org\u003e\nCc: \"Jason A. Donenfeld\" \u003cJason@zx2c4.com\u003e\nCc: Guenter Roeck \u003clinux@roeck-us.net\u003e\nCc: qemu-stable@nongnu.org\nFixes: 79068477686 (\"target/i386: reimplement 0x0f 0x3a, add AVX\", 2022-10-18)\nSigned-off-by: Paolo Bonzini \u003cpbonzini@redhat.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "853b1c8bf95e884110d2a7b890f694ae5e08db5c",
      "old_mode": 33188,
      "old_path": "target/i386/tcg/decode-new.c.inc",
      "new_id": "51038657f0f99d508164e6d6abb321f7af71981b",
      "new_mode": 33188,
      "new_path": "target/i386/tcg/decode-new.c.inc"
    }
  ]
}