Add a V8FP instruction 'vcvt{b,t}' to convert between half and double precision.
[oota-llvm.git] / test / MC / ARM / invalid-v8fp.s
1 @ RUN: llvm-mc -triple armv7 -show-encoding < %s | FileCheck %s
2
3 @ VCVT{B,T}
4
5   vcvtt.f64.f16 d3, s1
6 @ CHECK-NOT: vcvtt.f64.f16 d3, s1      @ encoding: [0xe0,0x3b,0xb2,0xee]
7   vcvtt.f16.f64 s5, d12
8 @ CHECK-NOT: vcvtt.f16.f64 s5, d12     @ encoding: [0xcc,0x2b,0xf3,0xee]
9
10