Fix uses of invalid LLVM intrinsics by sayantn · Pull Request #1807 · rust-lang/stdarch (original) (raw)

• name: "vand{neon_type.no}" doc: Vector bitwise and arguments: ["a: {neon_type}", "b: {neon_type}"] return_type: "{neon_type}" attr:
  • *neon-v7
  • FnCall: [cfg_attr, [*test-is-arm, {FnCall: [assert_instr, [vand]]}]]
  • FnCall: [cfg_attr, [*neon-target-aarch64-arm64ec, {FnCall: [assert_instr, [and]]}]]
  • *neon-not-arm-stable
  • *neon-cfg-arm-unstable safety: safe types:
  • int8x8_t
  • int8x16_t
  • int16x4_t
  • int16x8_t
  • int32x2_t
  • int32x4_t
  • uint8x8_t
  • uint8x16_t
  • uint16x4_t
  • uint16x8_t
  • uint32x2_t
  • uint32x4_t
  • int64x1_t
  • int64x2_t
  • uint64x1_t
  • uint64x2_t compose:
  • FnCall:
    • simd_and
    • • a
      • b